Professional Documents
Culture Documents
Aging Management of Concrete Structures in Nuclear Power Plants
Aging Management of Concrete Structures in Nuclear Power Plants
A numerical model for concrete strength change under neutron and gamma-ray irradiation
Ippei Maruyama, Kazumoto Haba, Osam Sato, Shunsuke Ishikawa, Osamu Kontani, Masayuki Takizawa 144-162
Outline of the guidelines for maintenance and management of structures in nuclear facilities
Yoshito Umeki, Shohei Sawada, Shiro Mitsugi, Toshinobu Maenaka, Katsuki Takiguchi 643-663
Numerical strategy for forecasting the leakage rate of Inner containments in double-wall nuclear reactor
buildings
Mehdi Asali, Bruno Capra, Jacky Mazars, Jean-Baptiste Colliat 408-420
Residual static strength of concrete cylinder specimen and stud anchor specimen after cyclic loadings
Osamu Kontani, Naoki Ishitobi, Junya Kawada, Noriaki Taogoshi, Makoto Koge, Yoshito Umeki 634-642
Accelerated aging of concrete dry cask storage systems for nuclear waste
Arezou Attar, Bora Gencturk, Mohammad Hanifehzadeh, Kaspar Willam 299-310
Review of research on the high temperature resistance of concrete structures in Chinese NPP
Jianzhuang Xiao, Wengang Xie, Qinghai Xie 335-343
Experimental study on hydraulic and macro-mechanical property of a mortar under heating and cooling
treatment
Zhinan Lin, Weiya Xu, Wei Wang, Jiuchang Zhang, Huanlin Wang, Rubin Wang 261-270
Even after the accident of the Fukushima Daiichi Nuclear Power Plant, the Japanese government has
positioned nuclear power as an important base load power source in terms of energy policies and has
focused on safety operation of the nuclear power generation. This is because it is an indispensable power
source for the use of renewable energy in Japan with limited resources. On the other hand, the first
generation commercial reactors in the world generally have a designed life of 30 to 40 years, but most of
them have been in existence for more than 20 years. Japan is no exception, and several NPPs aged already
more than 40 years.
Analysis and research show that many commercial reactors leave enough capacity to operate beyond the
designed service period. Replacement of the concrete members used in the reactor buildings is expensive
making it difficult in many cases. Therefore, in order to continue nuclear power generation over the long
term, it is necessary to maintain the performance of the concrete member during the designed service
period. From this point of view, it is indispensable to conduct research on maintenance management,
performance evaluation, and prediction of deterioration of concrete structures.
Research on concrete in the field of civil engineering and building science has been conducted for a long
time, while reevaluation of concrete performance, physical properties and various problems in terms of
nuclear field needs, i.e. Plant Life Management and Ageing Management, and addressing issues peculiar
to the nuclear power field will be an important contribution from the field of concrete research to the field
of nuclear power engineering. As discussed at OECD / NEA, alkali silica reactions and radiation influence
are important issues for the maintenance of nuclear power plants, and many research on them has been
reported.
Also in the past, research investment on concrete in the field of nuclear power has been carried out many
times in Japan as well as in other developed countries. However, the international publication of the research
results has been limited. In this ACT special issue, especially on advanced research in recent years, I widely
urged stakeholders to make efforts to allow research trends in Japan to be forecast from other countries.
Based on these backgrounds, we applied for papers focusing on Plant Life Management at nuclear power
plant facilities and related technologies in the research related to concrete. I think that it became a State of
the Art report, which highlights the current problems. I also anticipate it could lead to a trust from the society
concerning future nuclear power generation technology and a development of the concrete research field.
Ippei Maruyama
Professor, Nagoya University
Journal of Advanced Concrete Technology Vol. 14, 368-383 July 2016 / Copyright © 2016 Japan Concrete Institute 368
Invited paper
Abstract
A review of the current state of knowledge on the effects of radiation on concrete in nuclear power production applica-
tions is presented. Emphasis is placed on the effects of radiation damage, as reflected by changes in engineering proper-
ties of concrete, in the evaluation of the long-term operation and for plant life or aging management of nuclear power
plants (NPPs) in Japan, Spain, and the United States. National issues and concerns are described for Japan and the
United States followed by a discussion of the fundamental understanding of the effects of radiation on concrete. Spe-
cifically, the effects of temperature, moisture content, and irradiation on ordinary Portland cement paste and the role of
temperature and neutron energy spectra on radiation-induced volumetric expansion (RIVE) of aggregate-forming min-
erals are described. This is followed by a discussion of the bounding conditions for extended operation; the significance
of accelerated irradiation conditions; the role of temperature and creep; and how these issues are being incorporated into
numerical and meso-scale models. From these insights on radiation damage, analyses of these effects on concrete struc-
tures are reviewed, and the current status of work in Japan and the United States is described. Also discussed is the re-
cent formation of a new international scientific and technical organization, the International Committee on Irradiated
Concrete, to provide a forum for timely information exchanges among organizations pursuing the identification, quanti-
fication, and modeling of the effects of radiation on concrete in commercial nuclear applications. The paper concludes
with a discussion of research gaps, including (1) interpreting test-reactor data, (2) evaluating service-irradiated concrete
for aging management and to inform radiation damage models with the Zorita NPP (Spain) serving as the first compre-
hensive test case, (3) irradiated-assisted alkali-silica reactions, and (4) RIVE under constrained conditions.
levels for concrete integrity might not be suitable be- replaced is crucial for long-term operation of NPPs.
cause data were not representative of concrete mixtures, Before the Fukushima Daiichi accident, the Japanese
temperatures, and radiation fields typically seen in government’s PLM program inspected NPPs to evaluate
LWRs. Furthermore, a comprehensive review and re- their integrity after an operation period of 30 years and
analysis of the literature by Field et al. (2015) has every 10 years thereafter. This process required data on
greatly expanded the database and confirmed the pre- the integrity of structures under present and future con-
dominant role of radiation-induced volumetric expan- ditions (Japan Nuclear Energy Safety Organization
sion (RIVE) in the degradation of irradiated concrete.. 2013; Architectural Institute Japan 2015). To evaluate
In this report, a review of the current state of knowl- concrete structures in PLM, two measures of deteriora-
edge on the effects of radiation on concrete in nuclear tion were chosen: reduction in concrete strength and
applications is presented. Emphasis is placed on the reduction in shielding performance. The following fac-
effects of radiation damage as reflected by changes in tors arising from the environmental conditions and ma-
engineering properties of concrete for PLM and the terials used in reactor buildings adversely affect the
evaluation of the long-term operation of NPPs in Japan, concrete strength: elevated temperature, irradiation,
Spain, and the United States. carbonation, the alkali-silica reaction (ASR), and ma-
Following a discussion of issues and concerns in Ja- chine vibration.
pan and the United States, this report focuses on the After the Fukushima Daiichi nuclear accident, new
current status of the fundamental understanding of ra- nuclear reactor regulations came into effect on 8 July
diation effects on concrete. It begins with the effects of 2013, and an approval system for extending the opera-
radiation on hardened cement paste (HCP), including tion period of NPPs was introduced by the Nuclear
drying and temperature effects and radiolysis, and RIVE Regulatory Authority of Japan (NRA). Generally, a 40-
of aggregate-forming minerals. This is followed by a year operation period is allowed for every NPP although
discussion of the combined effects of temperature, dry- this period can only be extended to a maximum of 20
ing, and radiation on concrete, including the signifi- years after the facility passes a special inspection.
cance of accelerated irradiation conditions, the role of In PLM and the inspection for extending the opera-
temperature and creep, as well as variations in energy tion period, the integrity of concrete affected by radia-
spectra and damage energy, and how these factors are tion is evaluated with reference levels taken from data
being incorporated into numerical and mesoscale mod- reported in Hilsdorf et al. (1978) and interpreted by
els of concrete structures. Kontani et al. (2010): 1 × 1020 n/cm2 for fast neutrons
From these insights on radiation damage and the and 2 × 105 kGy for gamma rays. However, in a national
bounding fluences at extended lifetimes, analyses of project organized by the Japan Nuclear Energy Safety
these effects on concrete structures are reviewed, and Organization (partly a former organization in the NRA),
the current status of work in Japan and the United States the Hilsdorf compiled data were found to be unsuitable
is described. Furthermore, the recent formation of a for this purpose because they did not cover the type of
new international scientific and technical organization concrete used in LWRs (Kontani et al. 2010). A national
[the International Committee on Irradiated Concrete project has been launched to elucidate the effects of
(ICIC)] to provide a forum for timely information ex- radiation on concrete properties and to develop a system
changes among organizations pursuing the identification, for evaluating the integrity of concrete affected by ra-
quantification, and modeling of the effects of radiation diation (Maruyama et al. 2012).
on concrete in commercial nuclear applications is de-
scribed. US conditions, issues, and concerns:
The paper concludes with a discussion of current re- As described in Rosseel et al. (2014), extending the
search gaps, including (1) interpreting test-reactor data, operating lifetimes of current NPPs beyond 60 years and
(2) evaluating service-irradiated concrete for aging making additional improvements in their productivity is
management and to inform radiation damage models essential to meeting future United States national energy
with the Zorita NPP (Spain) serving as the first compre- needs while reducing greenhouse gas emissions. To
hensive test case, (3) irradiated-assisted alkali-silica meet these goals, a critical evaluation of the knowledge
reactions, and (4) RIVE under constrained conditions. gaps of the effects of exposure time to temperature,
stress, coolant, and radiation fields on materials that
Japanese conditions, issues and concerns: comprise the structures and components of an NPP is
LWRs have been identified as an important component required. And although it is expected that the vast ma-
of electric power generation as established by extensive jority of concrete structures will continue to meet their
discussions following the Fukushima Daiichi nuclear functional and performance requirements during future
accident (METI 2014). Ensuring the safe, long-term licensing periods, there may be isolated examples where
operation of NPPs is necessary for the public and for structures may not exhibit the desired durability, primar-
electric power suppliers, particularly because construc- ily due to environmental effects such as radiation.
tion of new NPPs is difficult in Japan. For these reasons, To address these extended lifetimes effects, the Elec-
aging management of concrete structures that cannot be tric Power Research Institute (EPRI), through the Long-
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 370
Term Operations (LTO) Program, and Oak Ridge Na- where water is present (i.e., in the HCP). Because irra-
tional Laboratory (ORNL), through the US Department diation results in energy-deposition-induced heating,
of Energy (DOE) Light Water Reactor Sustainability temperature and moisture transport must be considered
(LWRS) Program, established a research plan to inves- concurrently in the study of irradiation effects. The fol-
tigate the aging and degradation processes associated lowing sections provide a summary of the state of
with NPP concrete. The basis for the plan arose in part knowledge of the effects of temperature, moisture con-
from the Expanded Materials Degradation Analysis tent and irradiation on ordinary Portland cement paste
(EMDA) report on the aging of concrete, an effort and aggregate.
jointly supported by DOE and the US Nuclear Regula-
tory Commission (NRC), and performed by expert pan- 2.2. Hardened cement hydrates
els from US national laboratories, industry, academia, 2.2.1 Drying and temperature effects:
and international organizations. A key finding of the The shrinkage and strength properties of HCP affect the
EMDA was the urgent need to develop a consistent physical properties of concrete. The aggregates restrain
knowledge base on irradiation effects in concrete the shrinkage of cement paste during drying or heating
(Graves et al. 2014). Although not part of the (Carlson 1938; Hansen and Nielsen 1965; Hobbs 1974),
EPRI/LWRS plan, the NRC has initiated confirmatory and the large difference in volume change, due to envi-
research for the developing the technical basis for the ronmental conditions, between the aggregates and the
regulatory framework of subsequent license renewal cement paste matrix produces cracks in the cement paste
(see for example Section 6.2). that act as pores. These cracks affect the strength and
The LWRS program is focusing on (1) assessing the Young’s modulus of the concrete (Bažant et al. 1982;
radiation environment in concrete biological shields Son and Hosoda 2010; Maruyama et al. 2014b; Lin et al.
(CBSs), defining the upper bound of the neutron and 2015). Therefore, the differential shrinkage is the main
gamma dose levels expected at extended operation, and factor in the monotonic reduction of the Young’s
estimating adsorbed dose; (2) harvesting and testing modulus of the concrete (Maruyama et al. 2014b).
service-irradiated concrete to validate models; (3) irra- The strength of the mortar and concrete decreases
diating prototypical concrete and its components under from high relative humidity (RH) to intermediate RH
accelerated ion, neutron and gamma dose levels to es- (40-50% RH at room temperature) and increases below
tablish upper bounds to inform radiation damage mod- the intermediate RH range under drying or heating con-
els; (4) developing improved models to enhance the ditions (Pihlajavaara 1974; Maruyama et al. 2014b).
understanding of the effects of radiation on concrete; This behavior is attributed to the change in strength of
and (5) establishing an international collaborative re- the HCP, and the cracks due to drying are important in
search and information exchange to leverage capabili- reducing the compressive strength above 60% RH
ties and knowledge including developing cooperative through limiting the load bearing path in concrete (Ma-
test programs to improve confidence in data obtained ruyama et al. 2014b). This complex behavior of strength
from various concretes and from accelerated irradiation of HCP is explained by the colloidal nature of C-S-H
experiments (Rosseel et al. 2015a). EPRI is now focus- (Jesser 1927; Powers and Brownyard 1946; Tomes et al.
ing on the structural analyses of select US NPPs with 1957; Thomas and Jennings 2006; Jennings 2008). Dur-
CBS that perform structural support and whose esti- ing the initial drying process, the C-S-H particles form-
mated fluence at ~ 80 years of operation approaches the ing colloidal C-S-H agglomerate, producing large pores
bounding fluence of 6 × 1019 n/cm2 at E >0.1 MeV (Sec- in the cement paste. The increase in the volume of large
tion 4.1). pores reduces the paste strength. In contrast, drying
strengthens the C-S-H by tightening the layered struc-
2. Fundamental understanding of radiation ture. The increase in the strength of the solid structure
effects on concrete overcomes the reduction of strength caused by large
pores, resulting in the strength increase of hardened
2.1 Introduction cement paste at lower RHs (Maruyama et al. 2014a).
Concrete is by nature and constitution, a composite ma-
terial made of HCP and aggregate. Water is present in 2.2.2 Radiolysis effects:
the HCP under various forms (i.e., chemically bound, Gamma rays affect material through electronic excita-
adsorbed and free water). Gamma and neutron irradia- tions. Metastable calcium peroxide octahydrates can be
tion interact differently with the concrete compounds produced in HCP with high water content (Bouniol and
[See Table 3 in Kontani et al. (2010)]. In particular, the Aspart 1998; Vodák et al. 2005; Lowinska-Kluge and
poorly crystallized structure of the calcium-silicate- Piszora 2008), altering the strength and pore size distri-
hydrates (C-S-H) and the HCP porosity makes them less bution and, especially, resulting in subsequent carbona-
susceptible to neutron-induced lattice defect accumula- tion. It is also suggested that gamma radiation-induced
tion and distortion(Kontani et al. 2010) as opposed to carbonation increases the strength of HCP (Vodák et al.
well-crystallized aggregate-forming minerals. Concur- 2005; Vodák et al. 2011). Underlying these phenomena
rently, gamma-induced radiolysis occurs predominantly is the radiolysis of water. Radiolysis is the physico-
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 371
chemical consequence of ionizations along the trajec- stresses in the cement paste [e.g., Altoubat and Lange
tory of radiation. Several groups, including Bouniol and (2001) for shrinkage cracking, Giorla, Scrivener et al.
Aspart (1998), Bouniol and Bjergbakke (2008), and (2015) for ASR].
Kontani et al. (2013), have studied the radiolysis of pore
water in cement. 2.3. Aggregate-forming minerals
For example, Kontani et al. (2013) performed 2.3.1 Radiation-induced volumetric expansion:
gamma-ray irradiation experiments on cement paste and With about 70% of the volume fraction of concrete con-
collected all the gases from the samples by argon career sisting of aggregate, understanding aggregate-forming
gas. By analyzing time-dependent gas composition mineral sensitivity to radiation is of the utmost impor-
changes during gamma irradiation, they found the fol- tance. Alpha decay of certain radionuclides is known to
lowing phenomena of gas production in cement paste: produce a metamict state in rock-forming minerals (Ew-
2)
• The hydrogen production G value is linear with ing et al. 1988). Ion-beam irradiation of minerals gener-
gamma-ray dose rate (1~10 kGy/h), and this relation ally results in crystalline-to-amorphization (CA) transi-
is not affected by the sample temperature (20~60ºC). tion (Wang et al. 1991; Eby 1992; Douillard and
• The linearity between gamma-ray dose rate and hy- Duraud 1996; Harbsmeier and Boise 1998) and, in some
drogen production G value implies a validation of ac- cases, in the precipitation of oxides (Meldrum et al.
celerated gamma-ray irradiation experiments with 1997) or the formation of cavities (Templier et al. 1997).
gamma-ray dose rates less than10 KGy/h. Amorphization is predominantly observed in silicate-
• Only a few percent of the chemically bound water based minerals. For example, while quartz presents
decomposed even when subjected to the reference strong long-order radiation-induced disordering, irradia-
level of 2 × 105 kGy, since the G values were very tion of calcite causes primarily the rotation of the car-
low at around 0.033). This means that hydration prod- bonate groups (Pignatelli et al. 2016). The critical CA
ucts can maintain their integrity. dose in silicates depends on several factors: (1) the
• The hydrogen generation rate decreased as the evap- melting temperature; (2) a structural factor (e.g., degree
orable water content in the sample decreased. of SiO4 polymerization, elastic properties); and (3) the
Further scientific research is necessary to validate the proportion of Si-O bonding (Eby et al. 1992). Amorphi-
integrity of concrete and find a possible degradation zation results in changes in physical, optical and me-
mechanism by radiolysis in concrete. chanical properties (e.g., Wittels 1957; Primak 1958;
Mayer and Lecomte 1960; Wong 1974). Initial anisot-
2.2.3 Interaction between creep and irradiation: ropic crystalline elastic properties are gradually modi-
Creep of the cement paste is a key aspect of the me- fied with the irradiation dose to converge toward an
chanical behavior of concrete, especially considering isotropic stiffness tensor (Mayer and Lecomte 1960;
the time scales involved in the LTO of NPPs. On the one Zubov and Ivanov 1967). Amorphization also induces
hand, creep may affect the response of the structure it- significant density change, especially in silicates. For
self, for example, by relaxing the pre-stress in rebars example, the maximum volumetric expansion of quartz
(e.g., Magura et al. 1963), and on the other hand, it and feldspars has been shown to be as large as ~18%
strongly affects crack propagation in concrete (Rusch (Primak 1958) and ~8% (Krivokoneva 1976), respec-
1960; Bazant and Gettu 1992). Gray (1971) measured tively, while the change of density in calcite remains
an increase in the creep and shrinkage of a cement grout rather low (~0.3% according to Wong). Depending on
specimen under neutron irradiation even though the the mineralogical content, considerable variations in
various changes of hygro-thermal conditions during the aggregate RIVE have been observed as described in the
experiments, and the lack of control specimen, make comprehensive review by Field et al. (2015). Moreover,
this specific experiment difficult to interpret. some observed post-irradiation expansions exceed what
McDowall (1971) found that under gamma irradiation, is considered as detrimental by ASR researchers (e.g.,
creep of concrete was reduced, while autogeneous Fournier and Berube 2000; Rajabipour et al. 2015).
shrinkage was increased. This is consistent with creep
experiments at low relative humidity (Whittmann 1970), 2.3.2 Role of temperature in radiation damage:
and therefore might be a manifestation of the radiolysis Increased irradiation temperature delays the onset of the
in the cement paste. Understanding how creep is af- critical dose for amorphization and reduces the expan-
fected by radiation is critical because creep controls the sion kinetics of quartz (Bykov et al. 1981) and silicates
long-term crack propagation by relaxation of the such as feldspars (Krivokoneva 1976) and pyroxenes,
yet it does not seem to affect the total RIVE at the end
2)
of the amorphization process. The effect of temperature
The G value refers to the number of specified chemical on quartz expansion is explained by the annealing of
events produced in an irradiated substance per 100 eV point defects during irradiation occurring at an early and
of energy absorbed. intermediate stage of amorphization. However, no sig-
3)
This value is estimated on an assumption that only nificant annealing or anti-annealing (Primak 1958, Yano
hydroxyl groups in the cement paste sample absorb the et al. 2007) of RIVE is expected from temperature
gamma-ray energy.
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 372
variations occurring in the test reactors and LWRs (Le bution of gamma rays to the total absorbed dose is simi-
Pape et al. 2016). lar and may exceed the neutron-induced absorbed dose
(Remec et al. 2013). However, preliminary results indi-
2.3.3. Energy spectra and damage energies: cate that the gamma-ray contribution to the atom dis-
As previously noted, a review of existing data on irradi- placements does not exceed a few percent of the neu-
ated concrete reveals that neutron fluence is predomi- tron-induced displacements and will therefore be insig-
nately used to characterize the radiation environment nificant in most cases. In accelerated irradiation experi-
(Hilsdorf et al. 1978; Field et al. 2015), but there is little ments, the conditions may be considerably different, and
consistency in the neutron energy cutoff applied for the contributions of gamma rays need to be carefully evalu-
neutron fluence (Remec et al. 2016b). However, in the ated for each specific experiment.
CBS of pressurized water reactors (PWRs), the neutron
fluence for a cutoff energy of E >0.1 MeV is about 8-15 3. Irradiated concrete
times higher than the neutron fluence at E >1 MeV (and
the neutron fluence for E >0 eV can be 20-40 times 3.1 Test reactor data
higher) (Field et al. 2015). Using consistent and appro- The essential reference used to systematize the analysis
priate energy cutoff is therefore crucial for the assess- on irradiated concrete structural properties is Hilsdorf et
ment of concrete degradation. al. (1978). Later, Fujiwara et al. (2009) limited their
RIVE is currently considered the dominant cause for analysis to ordinary Portland structural concrete and
degradation of irradiated concrete (Field et al. 2015). It found no data subject to any significant effect of neutron
is, therefore, of interest to investigate neutron-induced irradiation on concrete for fluence less than 1 × 1019
atom displacements in aggregates. Remec et al. (2016b) n/cm2. More recently, Field et al. (2015) revisited his-
showed that, for the neutron spectra in the biological torical test reactor irradiation data and collected 307
shield of two-loop and three-loop (the number of reactor compression strength data, 62 tensile strength data, 138
cooling circuits) PWRs, neutrons with energies above elastic modulus data and 114 linear expansion data. The
0.1 MeV contribute more than 95% of all atom dis- data suggest that concrete compressive and tensile
placements in several widespread aggregate minerals, strengths as well as the Young’s modulus decrease
while neutrons with energies above 1 MeV cause only above fluences around 1 × 1019 n/cm2. This threshold
about 20 to 25% of the total atom displacements, as fluence value, however, is subject to variations with the
shown in Fig. 1 (Remec et al. 2016a). The conclusions concrete composition and, possibly, the exposure condi-
were similar when, instead of atom displacements, neu- tions. A strong correlation with the radiation-induced
tron-induced absorbed dose in minerals was considered. (and thermal expansion) volumetric expansion of aggre-
These observations suggest that neutron fluence with a gate and concrete is observed, suggesting that RIVE is a
cutoff of 0.1 MeV is the preferred parameter for radia- primary degradation factor (Seeberger and Hilsdorf
tion field characterization in irradiation experiments and 1982). However, Field et al. (2015) noted important
for accessing radiation-induced concrete degradation. uncertainties limiting the interpretation of the gathered
Work is in progress to develop a unified irradiation data, particularly for the fluence levels of interest (i.e.,
parameter that will account for neutron and gamma-ray above 1 × 1019 n/cm2): (1) The absence of energy cutoff
contributions to concrete degradation. In the typical normalization (as described in Section 2.3.3); (2) the
radiation field of the PWR biological shield, the contri- absence of comparable mechanical testing procedures
Fig. 1 Relative contribution of neutrons with energies E > E0 to the dpa (displacements per atom) rate for the neutron
spectrum in the cavity of a three-loop PWR, for several selected minerals.
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 373
(age, geometry and size of specimens); (3) the effects of to the previous numerical calculation results of concrete
conditioning (moisture content) before and during irra- shrinkage (Maruyama and Sugie 2014). Based on an
diation are often not reported; and (4) most data corre- input of mortar shrinkage, damage in concrete produced
spond to testing temperatures above 100°C leading to by the restraint of coarse aggregate was evaluated ex-
enhanced damage in the HPC and at the interface with plicitly. Owing to the modeling of ITZ, shrinkage of
aggregate. Finally, although reliable data on the effects concrete under different drying conditions can be repro-
of gamma irradiation on concrete mechanical properties duced by the proposed model. After drying, specimens
are scarce (Alexander 1963; Kelly et al. 1969; Gray were loaded and strength changes calculated. The re-
1971), they seem to suggest limited impact. Further re- sults are shown as the change in the relative compres-
search is needed. sive strength (Fc/Fco) in Fig. 3. From the calculation, it
A correct interpretation of irradiation experiments re- is elucidated that compressive strength of cylinder con-
quires a physics-based mechanistic understanding and crete is determined by both damage in concrete due to
modeling of the temperature, moisture content and irra- the volume change difference between mortar and
diation effects in the concrete compounds. The next coarse aggregate and strength change of cement paste
section provides some insights regarding the recent pro- (in this calculation, it is strength mortar) due to colloidal
gress made using mesoscale models of irradiated con- alteration under heating and drying.
crete. A preliminary simulation to evaluate the impact of
aggregate expansion due to radiation was conducted. In
3.2 Numerical models the preliminary simulation, expansion of coarse aggre-
Mesoscale model: gate in addition to the presented model discussed above
The impact of the expansion of aggregates due to radia- cannot simulate the expansion of concrete and strength
tion and the resultant damage observed as changes in reduction due to damage in concrete. In the case of ag-
physical properties can be understood through numeri- gregate expansion, impact of damage in concrete is
cal calculations. This approach has two objectives. The overestimated, and further study is required of the rela-
first is to find a mechanism that describes changes in tionship of damage in concrete in large expansion to the
physical properties of concrete components due to heat physical properties of concrete.
and drying, as well as radiation. This mechanism will A 2D mesoscale model for irradiated concrete (Giorla,
contribute to the interpretation of experimental data and Vaitová et al. 2015) was developed at ORNL based on
to formulating a universal conclusion. The second is to
extrapolate the phenomena from a laboratory specimen
size to a structural size that cannot be obtained experi-
mentally.
In the mesoscale, or aggregate-size scale, it is impor-
tant to consider the role of aggregate explicitly in a nu-
merical calculation. For this objective, the finite-
element method (FEM) and rigid body spring model
(RBSM) are applied to the evaluation of concrete under
drying, heating and radiation.
Originally developed by Kawai (1978), RBSM has
been applied extensively for structural analysis. RBSM
deals with crack propagation of concrete directly since it
represents a continuum material as an assembly of rigid
particle elements interconnected by zero-size springs (a) Voronoi mesh (b) Aggregate mesh
along their boundaries. Being nonlinear, these zero-size Fig. 2. RBSM mesh (Ogawa and Maruyama 2016).
springs can simulate the cracking behavior of a contin-
uum material. The nonlinearity and discrete behavior of
the continuum material is emulated by cracks develop-
ing at the interfaces of the rigid particles. For this reason,
crack patterns and the resultant nonlinear behavior of
the target model are significantly affected when a mesh
design is employed. To solve this problem, random ge-
ometry using Voronoi diagrams was applied (Bolander
and Saito 1998).
Recently, RBSM has been applied to simulate volume
change and compressive strength of concrete concur-
rently (Ogawa and Maruyama 2016). Coarse aggregate
is explicitly modeled in the mesh (Fig. 2), and the inter-
facial transition zone (ITZ) role is considered according Fig. 3. Comparison of calculated Fc/Fco with experimental
data (Ogawa and Maruyama 2016).
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 374
the AMIE finite element framework (Dunant and Scriv- transport calculations of the attenuation through the
ener 2010, Giorla et al. 2014). The model aim is to pro- pressure vessel and CBS (Remec et al. 2013). An ex-
vide the concrete expansion and internal degradation as ample of the attenuation for a Japanese PWR is shown
a function of the aggregate RIVE and the mechanical in Fig. 5.
properties of the paste and the aggregates, notably ac- Using the transport calculations as guidance, a multi-
counting for the coupling between creep and damage in step process was implemented by Esselman and Bruck
the cement paste (Giorla, Le Pape and Dunant 2016). (2013) to determine the expected neutron fluence values
The model was validated in an irradiation experiment on the CBS for the current US boiling water reactor
with serpentine concrete (Elleuch et al. 1972) since it is (BWR) and PWR fleet for up to 80 years of operation
one of the very few experiments in the literature in assuming 92% operating efficiency. The results shown
which the properties of the concrete, the aggregates, and in Fig. 6 highlight the importance of plant design and
the cement paste were independently characterized (see fuel loading scheme in determining the maximum ex-
Fig. 4). Under variable temperature and flux, the model pected fluence on the surface of the CBS at 80 years of
predicts a higher damage and expansion than under a operation. For example, 2- and 3-loop PWR reactors
constant temperature and flux (Le Pape et al. 2016). The show significantly higher fluence values than the 4-loop
model also shows an increase and orientation of internal plants. It should be noted that the BWR fluence at 80
cracking under external restraint, which is similar to the years of operation is generally lower than 4-loop PWR
effect observed for ASR (Giorla, Le Pape, and Dunant fluence.
2016).
Even though these models (RBSM and FEM) are 4.2 Structural analysis
based on different representations of the material (nota-
bly in terms of material continuity and mechanical be- Summary of concrete structural analysis litera-
havior), they both greatly overestimate the damage in ture:
the cement paste. This effect might be caused by the fact The primary function of the CBS is to provide radio-
that aggregates are represented as homogeneous parti- logical protection from the radiation emitted from the
cles that swell uniformly, and could therefore be re- nuclear steam supply system (NSSS), in particular, the
duced with a more accurate description of the aggregate RPV. In some PWR designs, the structural function of
mineral microstructure. Accounting for different RIVE the CBS is to transfer the static or seismic load of the
within the aggregate might cause damage to initiate reactor to the foundation system and to mitigate the ef-
within the aggregates rather than at their surface, as it is fects of accident scenarios. The CBS is generally made
for example the case in ASR (Ben Haha et al 2007; of conventional [i.e., normal weight aggregate (Essel-
Dunant and Scrivener 2010). Doing so requires an ex- man and Bruck 2013)] carbon steel reinforced concrete
perimental investigation and characterization of the ag- (Hookham 1991). Some designs also include a steel
gregates themselves, as well as a well-informed data- plate liner.
base of minerals and their properties. Limited irradiated concrete structural analyses have
been published in the open literature using different
4. Significance of irradiation for concrete models: analytical axisymmetric models (Andreev and
structures in NPP Kapliy 2014; Le Pape 2015), finite-element membrane
models (Mirhosseini et al. 2014), finite-element contin-
4.1 Bounding fluence at extended lifetimes and
radiation transport in concrete
As previously discussed, (Remec et al. 2016b; Rosseel
et al. 2014; Rosseel et al. 2015b), the onset of radiation-
induced concrete degradation has been correlated to
neutron fluence and gamma ray dose based on data as-
sembled by Hilsdorf et al. (1978). Obtaining accurate
estimates of the bounding fluence and dose values for
the US NPP fleet at the projected extended service is
clearly important. Due to variations in plant operations
and differences in reactor designs, however, this effort is
not simple. Fortunately, because operating NPPs are
required to implement a reactor pressure vessel (RPV)
surveillance program, a wealth of fluence information is
available. In the United States, the information is avail- Fig. 4 Comparison of calculated concrete RIVE with ex-
able from the US NRC Agency-wide Documents Access perimental data from Elleuch et al (1972), and example
and Management System (ADAMS) database (US NRC of simulated damage pattern at a neutron fluence of
19 2
2015). 1.3×10 n/cm ). The gray area represents the scatter in
The critical first step is the neutron and gamma ray the experimental data. Figure adapted from Giorla, Vai-
tova, et al (2015).
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 375
Fig. 5 Example of Japanese PWR radiation field developed from neutron transport calculations (Fukuya et al. 2002). Dr.
O. Sato (Mitsubishi Research Institute Inc.) and Dr. T. Igari (MRI Research Associates Inc.) confirmed this calculation.
Individual US PWRs
Fig. 6 Summary of US PWR fleet fluence at the outer diameter of the RPV extrapolated to 80 years (E >0.1 MeV) and
grouped by design (heat transfer loops). Graph reinterpreted from Esselman et al. (2013).
uum models (no RIVE) (Pomaro et al. 2011), and finite- the development of cracking (Giorla, Le Pape, and
element mesoscale models (Salomoni et al. 2014; Giorla, Dunant 2016). The extent of radiation-induced concrete
Le Pape, and Huang 2016). damage appears to be limited to a depth of about <20
In the absence of RIVE (or swelling), radiation dam- cm (Le Pape 2015; Giorla, Le Pape, and Huang 2016).
age [i.e., induced loss of mechanical properties (Hils- The consequences of this damage to the structural per-
dorf et al. 1978; Field et al. 2015)] has very limited ef- formance under seismic conditions (e.g., impact on the
fects on the CBS (Pomaro et al. 2011). Moreover, the RPV supports, or accident conditions; e.g., sudden in-
radiation field’s strong attenuation produces a high crease of temperature in the cavity, or seismic events)
RIVE gradient (Le Pape 2015) causing high biaxial still remain to be investigated.
compressive elastic stresses in the vertical and hoop
directions near the reactor cavity and important tensile Ongoing structural analysis work in the United
hoop stresses toward the back of the CBS (Andreev and States:
Kapliy 2014; Le Pape 2015). Simultaneously, the pro- Structural analyses currently being performed by US
longed moderate temperature exposure (<65°C by de- industry are motivated by recent research and develop-
sign) and strong internal moisture content gradient (Ox- ment gap identification and prioritization reviews for
fall 2013) affect the degree of hydration of concrete, and LWR concrete structures (Graves et al. 2014; Wall et al.
thus, its mechanical properties. In particular, it leads to 2012) and, more recently, the US NRC Draft Standard
the development of lower strengths toward the reactor Review Plan for Review of Subsequent License Re-
cavity (Maruyama and Igarashi 2015). The dissipation newal Applications for Nuclear Power Plants (NRC
of the RIVE-induced elastic stresses implies a potential 2015). The US industry approach, as described in Sec-
relaxation in the cement paste, but, more significantly, tion 4.1, has been to first determine the predicted neu-
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 376
tron fluence in the concrete biological shields for each radiation impacts show a distribution corresponding to
NPP in the US fleet for neutrons of E >0.1 MeV ex- the fluence variation. To evaluate the time-dependent
trapolated to 80 calendar years of operation (Esselman spatial distribution of concrete properly, DEVICE
and Bruck 2013). This was done using reactor-specific (Damage Evaluation for Irradiated Concrete) a numeri-
surveillance data reported to the NRC, information that cal code, which takes into account the heat, moisture,
is publically available in the NRC ADAMS database and radiation transport coupled with cement hydration,
(US NRC 2015). The bounding (maximum) neutron was proposed by Maruyama et al. (2016). This code is
fluence was found to be ~6 × 1019 n/cm2, which is composed of the established computational cement-
higher than the threshold for damage in concrete, as based material model (CCBM) (Maruyama et al. 2015)
described previously. To determine the effects of and the one-dimensional deterministic transport SN
chronic neutron irradiation on the structural margin in code ANISN (Engle 1967). DEVICE predicts the prop-
biological shields that also perform a structural support erty of concrete and creates the input for further struc-
function, a 3D wedge FEM is being developed. This tural analysis. The rigid-body spring network, which
model is based on the Westinghouse Type 1 PWR vessel was originally proposed by Kawai (1978), was applied
support configuration (Lapay et al. 2000). A schematic for the structural analysis, and the impact of expansion
of a typical PWR reactor vessel and biological shield is of the aggregate and the resultant change of seismic
shown in Fig. 7 for reference. The finite element mesh performance was investigated. The preliminary result,
was developed from a set of design drawings obtained which is shown in Fig. 9, indicates that there is a risk of
from a recently shut down 2-loop PWR. In this model,
the biological shield in the vicinity of a hot or cold leg
feed-through, under which the reactor vessel support sits,
and the embedded steel rebar geometry are meshed
separately and merged. The biological shield concrete
elements were segregated into annular layers to allow
variation of mechanical properties and variations in
RIVE as a function of depth into the concrete (Fig. 8).
The model will be run to failure under live and seismic
loading conditions to estimate the loss of structural
margin and changes in the internal stress morphology
due to chronic in-service neutron irradiation.
(a) (b)
Table 1 Exposure conditions in irradiation testing facilities and in-service PWRs. (Remec et al. 2013), (Maruyama et
(c) (d) (e) (f) (g) (h)
al. 2013), (Field et al. 2015), (Gray 1971), (Elleuch et al. 1972), (Dubrovskii 1967), (Kontani et al. 2010),
(i) (j) (k) (l)
(Seeberger and Hilsdorf 1982), (McDowall 1971), (Vodák et al. 2011), (Łowińska-Kluge and Piszora 2008), (Kelly
(m) (n)
et al. 1969), (Kitsutaka and Matsuzawa 2010), (Oxfall 2013).
PWR Test reactor/Gamma facility
Fast neutron flux (n/cm2/s) 1 to 2 × 1010 (ID) (a) 5 × 1011 to 2 × 1014 (b)
Fast neutron fluence (n/cm2) <6 × 1019 (ID) at 80 yr (c) <1020 (d,e,f)
Gamma flux (kGy/h) 5 to 20 (a,g) 0.02 to 200 (h,i,j,k,l,m)
Gamma dose (MGy) 50 to 200 at 80 yr 0.04 to 1.3
Temperature (oC) <65 (design) 40 to >250 (d,e)
Relative Humidity (-) Strong gradient
~0.5 at 50 mm, ~0.7 at 150 mm, >0.9 at 400 mm at 30 yr (n)
spalling due to radiation-induced expansion of the ag- tion on concrete is solely based on test reactor data (i.e.,
gregate. radiation under accelerated conditions). Table 1 pro-
vides a detailed comparison of the differences in radia-
5. International forum for collaborations: tion flux and dose (E >0.1 MeV), temperature, and rela-
The International Committee on Irradiated tive humidity. Accelerated testing raises questions of
Concrete (ICIC) possible rate-effect mechanisms and, consequently, the
significance of test reactor data. Interestingly, it is not
Understanding the effects of radiation on concrete is clear that accelerated testing leads to higher damage for
important for PLM and in determining long-term or a given dose. For example, higher temperatures may
extended operating performance of concrete structures increase the damage in the HCP and at its interface with
in existing NPPs. Not surprisingly, this issue is being an aggregate, while the RIVE rate is reduced, thereby
addressed by research organizations and utilities across potentially reducing the damage in the HCP.
the globe. In the last three years, researchers have been Another example is the effect of low/fast-RIVE on
actively working to build international partnerships and the stress relaxation in polycrystalline minerals, aggre-
collaborations in an effort to better define the issues, gate and HCP, which has received little attention so far.
develop a sound approach to resolving the major ques- In particular, the effects of gamma and neutron irradia-
tions, and maximize resources. The result is the crea- tion on the HCP creep properties are not clearly under-
tion of the ICIC (Rosseel et al. 2015b). stood (McDowall 1971; Gray 1971).
The purpose of the ICIC is four-fold. First, it provides Thorough fundamental understanding of both sepa-
a forum for timely information exchange among organi- rated and combined effects of irradiation, temperature
zations pursuing the identification, quantification, and and internal moisture on aggregate and HCP is still
modeling of the effects of radiation on concrete in needed and should be incorporated into physics-based
commercial nuclear applications (specifically, determi- modeling to derive irradiation consequences on LWRs.
nation of the effects of radiation damage as reflected by
changes in engineering properties). Second, it promotes 6.2 Evaluating service irradiated concrete for
the broad application of highly specialized or unique PLM and to inform/Confirm radiation damage
investigation techniques to relevant research materials models: Harvesting zorita concrete
of wide interest to develop improved understanding and Harvesting of concrete cores from decommissioned
predictions of the effects of radiation on concrete, ce- NPPs may provide not only an opportunity to generate
ment composites, and constituents. Third, it facilitates data from concrete that has experienced typical radia-
the conception, planning, and guidance of new coopera- tion fields, but it may also provide guidance to acceler-
tive research programs, for example, facilitating the ated irradiation studies. Moreover, the coupling of ac-
collection (harvesting) and distribution of pedigreed celerated or laboratory-irradiated concrete with har-
materials suitable for concrete irradiation damage inves- vested cores from NPP is expected to facilitate the effort
tigations. And fourth, it promotes cross-institutional use to develop an understanding of the damage mechanisms
of resources, where possible, that reflects mutual re- in irradiated concrete, including understanding potential
search interests or needs. Additional details may be rate effects of accelerated irradiation studies. These re-
found at the ICIC web page: sults may also be used to confirm models of radiation-
http://web.ornl.gov/sci/psd/mst/ICICFGM/index.shtml. induced degradation at extended lifetime such as the
ones proposed previously and for aging management
6. Discussion of research gaps programs.
The Jose Cabrera nuclear power station or “Zorita
6.1 Interpretation of test reactor data NPP” consists of a single-loop, Westinghouse PWR
Without an examination of service-irradiated LWR con- reactor with 160-MWe output. It was commissioned in
crete, the current understanding of the effects of radia- 1968 and operated until 2006, accumulating 26.5 effec-
tive full power years of operation. In February 2010,
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 378
Empresa Nacional de Residuos Radiactivos, S.A (EN- tion of US NPPs. The second issue is whether the accu-
RESA), the Spanish national company for nuclear waste, mulated fluence is sufficient to create observable dam-
took responsibility for decommissioning the station. In age in the concrete and aggregate. Although the Zorita
2009, the Consejo de Seguridad Nuclear (CSN), the CBS has accumulated the highest radiation of any reac-
Spanish Nuclear Safety Council, convened a meeting tor currently scheduled for decommissioning, the flu-
with national organizations to evaluate the feasibility of ence must reach the threshold level required to produce
obtaining concrete from Zorita to research the effects of microstructural changes in the cement paste and the
high radiation (neutron and gamma) and temperature on onset of RIVE in the aggregates. The third issue is the
concrete structures. Similar to the Zorita Internals Re- type of aggregate. Research has shown that siliceous
search Project (ZIRP), Zorita’s concrete structures could aggregates exhibit higher swelling (damage) than calcif-
be of great value to developing an improved understand- erous aggregates (Field et al. 2015; Rosseel et al.
ing of the effects of radiation on aging concrete because 2015b). The greater the ratio of siliceous to limestone
the neutron fluence on the Zorita CBS was approaching aggregates, the larger the observable effects of radiation
the onset for possible concrete deterioration as esti- on concrete. It is estimated that the evaluation will be
mated by the Hilsdorf curve (Hilsdorf et al. 1978). In completed in late 2016.
January 2010, the CEIDEN (Spanish Strategic Platform
on Nuclear Research and Development) established a 6.3 Irradiation-assisted alkali-silica reaction
working group on Zorita’s concrete cores. The current Ion-implantation-induced disorder (amorphization) of
partners of the project are ENRESA, ENDESA (utility), the three-dimensional framework silicates, α-quartz
IBERDROLA (utility), Gas Natural Fenosa (utility), (Ichikawa and Koizumi 2002; Pignatelli et al. 2016) and
licensee during operation of Zorita NPP, IETcc (Instituto plagioclase (Ichikawa and Kimura 2007) results in an
Eduardo Torroja de Ciencias de la Construcción), CSN, increase by several orders of magnitude of the dissolu-
and recently EPRI. tion rate in contact with a highly alkaline solution close
The structures being considered are the CBS (radia- to that of unirradiated amorphous silica (Pignatelli et al.
tion and temperature effects), the spent fuel pool/trans- 2016). Irradiated-amorphous silica also shows, to a
fer channel (boric acid degradation) and the contain- lesser extent, a greater susceptibility to dissolution
ment building (as reference). The building also will be (Ichikawa and Koizumi 2002). Conversely, irradiation-
inspected through nondestructive testing of the steel induced density change in calcite is essentially caused
liner under the concrete slab. The irradiated concrete by rotations and distortions of the carbonate groups with
tasks include: calculation of neutron and gamma fluence respect to the Ca atom positions leading to minor
and temperature by numerical modeling; obtaining test change of the dissolution rate (Pignatelli et al. 2016).
specimens for chemical, mechanical, and microstruc- The increasing reactivity of irradiated aggregate is one
tural characterization of irradiated and non-irradiated factor driving the formation of alkali-silica reaction gels.
concrete. Specifically, core samples from the CBS will Two other aspects must be considered before any defini-
be drilled from areas of high irradiation, high tempera- tive conclusion can be made. ASR gels are highly hy-
ture, and a combination of high irradiation and tempera- drophilic but only internal moisture resulting from mix
ture. From each core sample, several test coupons will excess water is available for their formation, which
be obtained: the closest to the reactor vessel, one from could rapidly become a limiting factor. The stability of
the opposite side, and some intermediate coupons to alkali-silica hydrates under gamma irradiation (radioly-
verify the attenuation effects. sis) is unknown and requires further research.
On behalf of the NRC, ORNL is reviewing informa-
tion and reports provided by the Zorita Concrete Work- 6.4 RIVE in restrained conditions
ing Group and ENRESA to evaluate the concrete and Although no literature data were found on the effect of
conditions to determine whether it would serve as a mechanical restraint on the development of RIVE, it is
suitable surrogate for US PWR NPPs. To determine the likely that restrained RIVE is subject to the same effects
applicability to long-term aging degradation of US as alkali-silica reaction (ASR) in concrete due to the
NPPs, three issues are being evaluated. The first issue similarity of both mechanisms (i.e., expansion of a spe-
is whether the concrete, and especially the aggregate, cific phase contained in the aggregates). ASR expansion
are the same or similar to concrete and aggregate found is reduced in the direction of the load and increased in
in the US NPPs. A careful study of concrete cores har- the lateral directions (although the total volumetric ex-
vested from any decommissioned NPP would provide pansion is not constant with the load level), while dam-
an opportunity to generate data from concrete that has age is oriented preferably in the direction of the load
experienced typical radiation fields and, therefore, and increases with the load level (e.g., Larive 1997;
would facilitate the effort to develop an understanding Multon and Toulemonde 2006; Dunant and Scrivener
of the damage mechanisms in irradiated concrete. If the 2012). Preliminary mesoscale simulations suggest that
concrete and aggregate were not similar to concrete and RIVE is indeed subject to such anisotropy of expansion
aggregate found in some US NPPs, questions would and degradation when confined by external restraints
remain as to its general applicability to extended opera- (Giorla, Le Pape and Dunant 2016) although experimen-
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 379
tal verification is needed to conclusively determine this search gaps including evaluating service-irradiated con-
effect. crete from decommissioned NPPs, such as Zorita
(Spain), to inform degradation models and aging man-
7. Conclusions agement programs, evaluation of test reactor data, irra-
diated-assisted alkali-silica reactions, and RIVE under
In most countries with operating NPPs, LWRs have constrained conditions.
been identified as an important component of electric
power generation. For that reason, ensuring the safe, Acknowledgments
long-term operation of NPPs is critical to maintaining a The authors gratefully thank the following sponsors: US
reliable energy supply. Moreover, extending the operat- DOE, Office of Nuclear Energy, Light Water Reactor
ing lifetimes of current NPPs as allowed by national Sustainability Program (TMR, YLP, ABG, and IR); the
regulations, and making additional improvements in Nuclear Regulation Authority, Japan (IM and OK); the
productivity using PLM or aging management is essen- EPRI Long Term Operations Program (JJW); the US
tial to meeting future national energy needs while reduc- NRC (MS); Consejo de Seguridad Nuclear (CA and
ing greenhouse gas emissions. To meet these goals, a MO); IETcc-CSIC (CA); and ENRESA (MO) for sup-
critical evaluation of the knowledge gaps of the effects port of work on the effects of radiation on concrete. The
of exposure time to temperature, stress, coolant, and authors also express their appreciation to Drs. Kevin
radiation fields on materials such as concrete is required. Field and Elena Tajuelo Rodriguez for their review of
And although it is expected that the vast majority of the document and their helpful comments and sugges-
concrete structures will continue to meet their functional tions.
and performance requirements during licensing periods,
there may be isolated examples where structures may References
not exhibit the desired durability, primarily due to envi- Alexander, S., (1963). “Effects of irradiation on
ronmental effects such as radiation (Rosseel et al. 2014; concrete final results.” Technical report HL.63/6438.
Rosseel et al. 2015a). Harwell, Oxfordshire, UK, Atomic Energy Research
In this paper, a review of the current state of knowl- Establishment.
edge on the effects of radiation on concrete in nuclear Altoubat, S. A. and Lange, D. A., (2001). “Creep,
applications is presented. Emphasis is placed on the shrinkage, and cracking of restrained concrete at early
effects of radiation damage as reflected by changes in age.” ACI Materials Journal, 98(4), 323-331.
engineering properties of concrete in the evaluation of Andreev, V. and Kapliy, D., (2014). “Stress-state of a
the LTO of NPPs and for PLM in Japan, Spain, and the thick-walled cylindrical shell under the combined
United States. action of radiation and temperature field.” Advanced
Following a discussion of issues and concerns in Ja- Materials Research, 1006-1007, 177-180
pan (post Fukushima Daiichi) and the United States Architectural Institute Japan, (2015). “Guidelines for
(LTO and anticipated subsequent license renewal), this maintenance and management of structures in
report focuses on the current status of the fundamental nuclear facilities.” Tokyo: Maruzen.
understanding of radiation effects on concrete and its Bažant, Z. P. and Gettu, R., (1992). “Rate effects and
components. Specifically, it addresses drying and tem- load relaxation in static fracture of concrete.” ACI
perature effects and radiolysis on HCP and RIVE of Materials Journal, 89(5), 456-468.
aggregate-forming minerals. This is followed by a dis- Bažant, Z. P., Chern, J. C. Abrams, M. S. and Gillen, M.
cussion of the combined effects of temperature, drying, P., (1982). “Normal and refractory concretes for
and radiation on concrete, including the significance of LMFBR application, Vol. 1: Review of literature on
accelerated irradiation conditions, the role of tempera- high-temperature behavior of Portland cement and
ture and creep, as well as variations in energy spectra refractory concretes.” EPRI Report NP-2437.
and damage energy, and how these factors are being Chicago, Illinois: Northwestern University and
incorporated into numerical and mesoscale models of Portland Cement Association.
concrete structures. Using these insights, attempts to Ben Haha, M., Gallucci, E., Guidoum, A., and Scrivener,
evaluate the impact of radiation-induced degradation by K. L. (2007). “Relation of expansion due to alkali
numerical simulations of concrete structures are re- silica reaction to the degree of reaction measured by
viewed, and the current status of work in Japan and the SEM image analysis.” Cement and Concrete
United States described. Research, 37(8), 1206-1214.
The recent formation of a new international scientific Bolander, Jr., J. E. and Saito, S., (1998). “Fracture
and technical organization (the ICIC) to provide a forum analyses using spring networks with random geome-
for timely information exchanges among organizations try.” Engineering Fracture Mechanics, 61(5-6), 569-
pursuing the identification, quantification, and modeling 91.
of the effects of radiation on concrete in commercial Bouniol, P. and Aspart, A., (1998). “Disappearance of
nuclear applications is also described. The paper con- oxygen in concrete under irradiation: The role of
cludes with a discussion of current, high-priority re- peroxides in radiolysis.” Cement and Concrete
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 380
silica reaction of aggregates.” Journal of Nuclear mécaniques.” PhD Thesis. Ecole Nationale des Ponts
Science and Technology (Tokyo, Japan), 39, 880-88. et Chaussees.
Ichikawa, T. and Kimura, T., (2007). “Effect of nuclear Le Pape, Y., Giorla, A. and Sanahuja, J., (2016).
radiation on alkali-silica reaction of concrete.” “Combined effects of temperature and irradiation on
Journal of Nuclear Science and Technology, 44, concrete damage.” Journal of Advanced Concrete
1281-128. Technology, 14, 70-86.
Japan Nuclear Energy Safety Organization, (2013). Le Pape, Y. , (2015). “Structural effects of radiation-
“Review manual for aging management technical induced volumetric expansion on unreinforced
evaluation.” Tokyo: JNES-RE-2013-9012. concrete biological shields.” Nuclear Engineering
Jennings, H. M., (2008). “Refinements to colloid model and Design, 295, 534-548.
of C-S-H in cement: CM-II.” Cement and Concrete Lin, M., Itoh, M. and Maruyama, I., (2015).
Research, 38(3), 275-289. “Mechanism of change in splitting tensile strength of
Jesser, L., (1927). “Kolloide chemische reaktionen des concrete during heating or drying up to 90°C.”
tonerdezementmörtels.” Zement, 16, 741. Journal of Advanced Concrete Technology, 13(2), 94-
Kawai, T., (1978). “New discrete models and their 102.
application to seismic response analysis of Łowińska-Kluge, A. and Piszora, P., (2008). “Effect of
structures.” Nuclear Engineering and Design, 48(1), gamma irradiation on cement composite observed
207-229. with XRD and SEM methods in the range of radiation
Kelly, B., Brocklehurst, J., Mottershead, D. and dose 0-1409 MGy.” Acta Physica Polonica, 114(2),
McNearney, S., (1969). “The effects of reactor 399-411.
radiation on concrete.” In: Proceedings of the second Magura, D. D., Sozen, M. A. and Siess, C. P., (1963). “A
information meeting on pre-stress concrete and study of stress relaxation in prestressing reinforce-
reactor pressure vessels and their thermal isolation, ment.” University of Illinois.
volume EUR-4531, Brussels, 237-265. Majorana, C., Salomoni, V., Gramegna, F., Xotta, G. and
Kitsutaka, Y. and Matsuzawa, K., (2010). “The effect of Pomaro, B., (2014). “Macro- and meso-scale analysis
gamma radiation on the fracture properties of of concrete as a multiphase material for biological
concrete.” In: B. Oh, Ed., Fracture mechanics of shields against nuclear radiation.” International
concrete and concrete structures—Recent Advances Journal for Numerical and Analytical Methods in
in Fracture Mechanics of Concrete, FramCoS-7, 23- Geomechanics, 38, 518-535
28 May, 2010, 61-64. Maruyama, I., Kontani, O., Ishizawa, A., Takizawa, M.
Kontani, O., Ichikawa, Y., Ishizawa, A., Takizawa, M. and Sato, O., (2012). “Development of system for
and Sato, O., (2010). “Irradiation effects on concrete evaluating concrete strength deterioration due to
structure.” In: International symposium on the ageing radiation and resultant heat.” In: Proceedings of 3rd
management and maintenance of nuclear power international conference on NPP life management for
plants, 2010, 173-182. long term operations. IAEA-CN-194-096, Salt Lake
Kontani, O., Ichikawa, Y., Ishizawa, A., Takizawa, M. City, May 2012.
and Sato, O., (2011). “Irradiation effects on concrete Maruyama, I., Kontani, O., Sawada, S., Sato, O.,
durability of nuclear power plants.” Proceedings of Igarashi, G. and Takizawa, M., (2013). “Evaluation of
ICAPP 2011, Nice, France, 2-5 May 2011. irradiation effects on concrete structure—background
Kontani, O., Sawada, S., Maruyama, I., Takizawa, M. and preparation of neutron irradiation test.” In:
and Sato, O., (2013). “Evaluation of irradiation Proceedings of the ASME 2013 power conference
effects on concrete structure—Gamma-ray irradiation POWER2013, number 98114. Boston, USA, 29 July-
tests on cement paste.” In: Proceedings of the ASME 1 August 1, 2013.
2013 power conference: Power2013, Boston, 29 July- Maruyama, I., Nishioka, Y., Igarashi, G., and Matsui, K.,
1 August, 2013. New York: American Society of (2014a). “Microstructural and bulk property changes
Mechanical Engineers, Vol. 2, paper no. 98099, in hardened cement paste during the first drying
V002T07A002. process.” Cement and Concrete Research, 58(0), 20-
Krivokoneva, G., (1976). “Structural changes in 34.
feldspars under impact of radiation,” In: Crystal Maruyama, I., Sasano, H., Nishioka, Y. and Igarashi, G.,
Chemistry and Structural Features of Minerals, (2014b). “Strength and Young's modulus change in
Leningrad: Nauka, 75-79. concrete due to long-term drying and heating up to
Lapay, W., Wang, S. C. Y. and Kim, C., (2000). “License 90°C.” Cement and Concrete Research, 66(0), 48-63.
renewal evaluation: aging management for reactor Maruyama, I. and Sugie, A., (2014). “Numerical study
coolant system supports.” Westinghouse, WCAP- on drying shrinkage of concrete affected by aggregate
14422, Rev. 2-A. size.” Journal of Advanced Concrete Technology,
Larive, C., (1997). “Apports combinés de 12(8), 279-288.
l'expérimentation et de la modélisation à la Maruyama, I. and G. Igarashi, (2015). “Numerical
compréhension de l'alcali-réaction et de ses effets approach towards aging management of concrete
T. M. Rosseel, I. Maruyama, Y. Le Pape et al. / Journal of Advanced Concrete Technology Vol. 14, 368-383, 2016 382
Scientific paper
Abstract
Aggregate radiation-induced volumetric expansion (RIVE) is a predominant mechanism in the formation of mechanical
damage in the hardened cement paste (hcp) of irradiated concrete under fast-neutron flux (Giorla et al. 2015). Among the
operating conditions difference between test reactors and light water reactors (LWRs), the difference of irradiation flux
and temperature is significant. While a temperature increase is quite generally associated with a direct, or indirect (e.g., by
dehydration) loss of mechanical properties (Maruyama et al. 2014), it causes a partial annealing of irradiation amorphi-
zation of α-quartz, hence, reducing RIVE rate. Based on data collected by Bykov et al. (1981), an incremental RIVE
model coupling neutron fluence and temperature is developed. The elastic properties and coefficient of thermal expansion
(CTE) of irradiated polycrystalline quartz are interpreted through analytical homogenization of experimental data on
irradiated α-quartz published by Mayer and Lecomte (1960). The proposed model, implemented in the meso-scale
simulation code AMIE, is compared to experimental data obtained on ordinary concrete made of quartz/quartzite ag-
gregate (Dubrovskii et al. 1967). Substantial discrepancy, in terms of damage and volumetric expansion developments, is
found when comparing irradiation scenarios assuming constant flux and temperature, as opposed to more realistic test
reactor operation conditions.
Neutron irradiation experiments of quartz at different (inverse of a fluence) govern the shape of the sigmoid. In
temperatures have been solidly reported by Bykov et al. particular, the inflexion point of the sigmoid, i.e.,
(1981). Dubrovskii et al. (1967) and Pedersen (1971) ∂ ε / ∂Φ = 0 , corresponds to a fluence of
2 * 2
also measured the irradiated properties of concrete made ln(ε max / κ ) / δ and a swelling of (εmax − κ ) / 2 (Le Pape
with quartz aggregates. Notably, Dubrovskii et al. pro- 2015). At the same inflexion point, the rate of swelling is
vide the temperature history for some of their samples, given by the expression (ε max − κ )δ / 4 . Hence, the main
thus allowing a detailed analysis of that particular ex- characteristics of Zubov's curve, except for the final
periment. The data collected by Bykov et al. is inter- amplitude of swelling, do not show a one-to-one de-
preted in this article to develop a RIVE model for quartz pendency on Zubov's model parameters. Eq. (1) can also
accounting for temperature and neutron flux. This model be interpreted in a slightly modified form, which corre-
is complemented with an analysis of the evolution of the sponds, mathematically speaking, to the form adopted by
mechanical properties of quartz as a function of its den- several researchers, e.g., (Larive 1997; Saouma and
sity, using experimental measurements by Mayer and Perotti 2006), to model the expansion of concrete sub-
Lecomte (1960) and analytic homogenization schemes. jected to alkali-silica reaction (ASR) as a function of
This provides the sufficient information to perform a time. Substitution of time by the fluence Φ leads to:
preliminary numerical analysis of Dubrovskii et al. ex-
periment with the mesoscale model previously devel- −
Φ
As noted by Bykov et al., "the very strong temperature From Eq. (2), the incremental variation of volumetric
effect in the range of 30 °C to 100 °C and the reduction of expansion at a given neutron fluence, temperature, neu-
this effect at a higher temperature indicates a nonuni- tron flux and temperature increment reads:
formity of distribution of the defects with respect to the
energy of activation, in particular, to a reduction of the d * ∂ε * dΦ ∂ε * dT
fraction of defects with an increased energy of activa- ε (Φ, T ) = ⋅ + ⋅ (4)
dt ∂Φ T dt ∂T dt
tion." Hence, instead of trying to interpret the data with Φ
3.1. Elastic properties slightly increases from 73 GPa to 76 GPa. Hence, there is
Trigonal α-quartz is characterized by 6 independent still a significant gap between the Young moduli of
elastic constants, i.e., cij stiffness coefficients, including a α-quartz and silica at the end of Mayer and Lecomte's
non-zero c14 term, while isotropic silica only requires the experiments preventing any extrapolation. It appears that
determination of 2 constants. The values of cij for varied Voigt-Reuss bounds tend to narrow with irradiation. This
polymorphs of crystalline silica and silica glass have suggests that the behavior of α-quartz crystals evolves,
been obtained experimentally in the temperature range of with irradiation, towards isotropy.
0 °C to ≈ 1, 000 °C by several authors, e.g., see review
articles (Ballato 2008; Pabst and Gregorová 2013). Poisson’s ratio of irradiated α-quartz.
Elastic constants of irradiated α-quartz and silica have The Poisson ratio of irradiated α-quartz, v* , shows a
been obtained by Mayer and Gigon (1956); Mayer and rapid increase from its initial value v0 = 0.08 to reach
Lecomte (1960); Zubov and Ivanov (1967) in the cor- about 0.18 at ≈ 4.5% volume expansion - See Fig. 7. In
responding range of RIVE below ≈ 6%, i.e., at about a this range of RIVE, the Poisson ratio estimated by the
third of the maximun expansion. Tentative extrapolation selfconsistent scheme is fittted by an exponential func-
for RIVE beyond 6% is discussed hereafter. The elastic tion. In the same range of fluence, the Poisson ratio of
properties of polycrystalline quartz can be theoretically silica increases from 0.17 and seem to level off at about
by derived by homogenization, i.e., Voigt-Reuss bounds vs* ≈ 0.21, which will be considered here, as the
(Voigt 1889; Reuss 1929), or self-consistent scheme maximum limit value of irradiated α-quartz at full
(Dormieux et al. 2006). The results can either be plotted amorphization - See Fig. 8. The proposed evolution of
as a function of RIVE or density. While the implemen- irradiated α-quartz Poisson ratio reads:
tation in a finite element frame (See section 5) is done
considering properties function of the volumetric ex-
pansion (internal variable), plotting the irradiated prop-
( ( b
) )
v* = min 1 − v0 + exp a ( ε * ) ; vs* (11)
Fig. 6 Young modulus of irradiated α-quartz and silica as Fig. 7 Poisson ratio of irradiatedα-quartz and silica as a
a function of the irradiated materials density. (s) amor- function of the irradiated materials density. (s) amor-
phous silica; (r) irradiated quartz, Reuss bound; (v) irra- phous silica; (r) irradiated quartz, Reuss bound; (v) irra-
diated quartz, Voigt bound; (SC) irradiated quartz, diated quartz, Voigt bound; (SC) irradiated quartz,
self-consistent scheme. self-consistent scheme.
Y. Le Pape, A. Giorla, J. Sanahuja / Journal of Advanced Concrete Technology Vol. 14, 70-86, 2016 75
Table 1 Irradiation parameters and volumetric expansion (assumed as the cube of the measured diameter expansion) for
Dubrovskii et al.'s data. Table assembled from (Dubrovskii et al., 1967, Original Tab. 1 & 3). †: estimated values.
Group No.
1 2 3 4
20 -2
integrated flux (×10 n.cm ) 3-3.3 2-2.8 1.2-1.45 0.4-0.6
13 -2 -1
maximum flux (×10 n.cm s ) 1.9-2.1 1.3-1.8 0.75-0.9 0.25-0.4
average flux (×1013 n.cm-2 s-1) 0.62†
maximum heat deposition (Wcm-3) 0.19-0.21 0.12-0.17 0.07-0.085 0.025-0.04
maximum temperature (°C) 300-350 225-275 100-150 20-50
average temperature (°C) 115 †
89 57 †
35†
concrete vol. expansion (%) 16.2 >17.4 9.3 2.4
> 19.2 >18.9 5.1 1.5
> 17.7 >15.9 5.7
> 18.0
0.4-0.6 μm m-1 °C-1, and, remains one order of magnitude n.cm-2 s-1). Assuming that the found linear correlation is
below that of pristine α-quartz CTE, the effects of tem- applicable in the range of temperatures corresponding to
perature on α 0* can be neglected ( α 0* ≈ 0.5 μm the BR-5 reactor operation, the flux history of group 2
m-1 °C-1). experiments, plotted in Fig. 11, can be derived from the
history of temperature provided in (Dubrovskii et al.
4. Interpretation of irradiation experiments 1967, original Fig. 1). Note that kinetic energies above
on quartz aggregate concrete 10 keV account for about 99.5% of the total energy (See
energy distributions in (Dubrovskii et al. 1967, Table 2.,
While quartz and quartzite sand is commonly used in p 1054)). Hence, the empirical parameters ac, bc, aL , and
concrete, coarser aggregate are usually made of other bL, derived from the interpretation of Bykov et al.'s data
rocks, e.g., limestone, granite, basalt, etc. . . Hence, ir- are assumed valid to model Dubrovskii et al.'s data. The
radiation experiments on concrete are rarely conducted average temperature and average flux calculated from
with pure quartz aggregate. We found only to sets of these plots are 89 °C and 6.2 × 1012 n.cm-2 s-1, i.e, re-
irradiation experiments on quartz aggregate concrete in spectively, 2.5-3 times, and 2-2.5 times, lower that the
the literature (Dubrovskii et al. 1967; Pedersen 1971). maximum temperature and flux. Figure 12 illustrates the
The details provided by Pedersen are unfortunately quite potential implication of different temperature and flux
limited to envision a thorough interpretation through scenarios in terms of volumetric expansion evolutions of
modeling. Dubrovskii et al. (1967) tested the effects of quartz, assuming the proposed model hold - Eqs. (2- 10):
neutron irradiation of ordinary concrete (density 2.31) (a), i.e., the “reference scenario”: variable temperature
cylinder of 40 mm in height and diameter, made of
Portland cement (w=c = 0.5). The mix contains quartz
river sand (0.15 to 0.6 mm, c : a = 1 : 1.7) and sandstone
gravel (5 to 10 mm, c : a ≈ 1 : 3). The volume fraction
of quartz sand is about 25%. The total volume fraction of
aggregate is close to 70%. Using the chemical composi-
tion of the mix provided by Dubrovskii et al., and as-
suming a typical oxide composition for the Portland
cement, we found that the molar composition of the
sandstone is about 31% of silicon and 66% of oxygen,
proving the highly-siliceous nature of the tested sand-
stone containing also traces of calcium, iron, magnesium
and aluminum, supporting Dubrovskii et al. statement
that “the total quartz content in this concrete was 70%.”
Table 1 presents the irradiation fluxes, the temperature,
and the post-irradiation volumetric expansions for the Fig. 11 Estimated time-evolution of the integrated energy
four tested groups. flux (black solid line) and fluence (thick solid red line)
Quasi-linear correlations between the maximum flux, over time - Group 2. Dashed line: fluence evolution as-
Φ max , the maximum heat deposition and maximum
12 -2 -1
suming a constant flux at 6:2×10 n.cm s . The error
temperature, Tmax can be observed ( Φ ≈ 5.83 ×
max bar at 501 d corresponds to the intervale of fluences
-3
10 Tmax + 0.106 with Tmax in °C and Φ 13
max in ×10 provided in Table 1.
Y. Le Pape, A. Giorla, J. Sanahuja / Journal of Advanced Concrete Technology Vol. 14, 70-86, 2016 77
Fig. 12 Time-evolution of the volumetric expansion of Fig. 13 Time-evolution of the volumetric expansion of
quartz for group 2 specimens assuming different tem- concrete for group 2 specimens assuming a variable
perature and flux scenarios. (a): variable temperature temperature (Dubrovskii et al. 1967, original Fig. 1) and
(Dubrovskii et al. 1967, original Fig. 1) and variable flux variable flux (Fig. 11). The boxes indicate the ranges of
(Fig. 11); (b): constant temperature (89 °C) and variable maximum fluences and expansions provided by Du-
flux; (c): constant temperature and constant flux 6.2 × brovskii et al. (data in Table 1). Dashed lines indicate the
12 -2 -1
10 n.cm s ; and, (d): variable temperature and con- uncertainty of the expansion measurement because the
stant flux. The black solid line corresponds to the flux and irradiated samples were wedged in the capsule (see
temperature correlated histories. text).
Y. Le Pape, A. Giorla, J. Sanahuja / Journal of Advanced Concrete Technology Vol. 14, 70-86, 2016 78
numerical model, as analytical micromechanical models pansion and degradation of the sand and cement paste
have limited capabilities to capture these effects. mixture not only as a function of the fluence and tem-
perature, but also as a function of the level and direction
5. Numerical modeling of the current stress. Mechanical restraint is likely to
affect the relation between expansion and degradation,
We simulate the experiments of Dubrovskii et al. (1967) notably making them both anisotropic as it is the case in
with the model previously developed by the authors for similar phenomena like ASR (Larive 1997; Multon and
irradiated concrete. The material is considered at the Toutlemonde 2006; Dunant and Scrivener 2012). The
meso-scale with a two-dimensional finite element rep- only experiment of neutron radiation under load in the
resentation of its microstructure. The model calculates literature (Gray 1971) lacks the sufficient supporting data
the deformation as well as the localization and extent of and control measurements to be conclusive in that re-
the micro-mechanical damage in the sample, accounting spect. This effect is therefore neglected here due to lack
for the different properties and behaviors of the cement of knowledge about irradiation expansion under load.
paste and the aggregates. The interfacial transition zone (ITZ) between the ag-
Post-irradiation deformations are the only experi- gregates and the cement paste is also neglected. As-
mental result available for this set of temperature and signing cohesive properties to the ITZ might control the
fluences, as Dubrovskii et al. were unable to measure debonding between the cement paste and the aggregate
mechanical properties due to the samples being wedged (e.g., Maruyama and Sugie (2014)). However, me-
in the irradiation rig. The model provides an estimation chanical properties of the ITZ need to be calibrated on
of the history leading to that deformation. Rigorous experimental data, which were not readily available for
validation of the simulation results cannot be achieved this experiment. In particular, there is no available data in
due to the absence of monitoring during the experiments the open literature on the effects of irradiation on the ITZ
and the uncertainties related to the data provided by bonding properties.
Dubrovskii et al.. The principal purpose of the simula- The microstructure is discretized with a conforming
tions is to give insights on the influence of the tempera- unstructured mesh using a Delaunay triangulation. The
ture and neutron flux variations on the expansion and calculations were run on a mesh containing ≈ 14,000
damage variations, and assess the importance of consid- linear elements. The two-dimensional representation was
ering combined or separated effects of temperature and chosen as the model does not present strong anisotropy,
irradiation in further modeling efforts. in terms of mechanical properties, loadings, or geometry.
The main components of the model are summarized However, it is recognized that this hypothesis might lead
below, and described in greater details in Giorla et al. to an earlier crack percolation compared to a
(2015). The model is implemented in the C++ finite three-dimensional analysis.
element framework AMIE, which was initially devel- The displacements of the bottom side of the sample are
oped in the context of mesoscale simulations of ASR fixed in the vertical direction. The central node of the
(Dunant and Scrivener 2010), and later applied for the bottom side is fixed in the horizontal direction to prevent
study of creep (Giorla et al. 2014) and irradiation (Giorla global sliding of the sample. The microstructure and
et al. 2015). mechanical boundary conditions are presented in Fig. 14.
The temperature, relative humidity and neutron fluence
5.1. Microstructure fields are assumed homogeneous through the sample.
We simulate a two-dimensional vertical slice of a con- The four cases detailed in section 4 (See also Fig. 12) are
crete cylindrical sample. Aggregates are represented with simulated: (a) variable temperature and flux, (b) variable
circles randomly placed in a square box. Dubrovskii et al. temperature and constant flux, (c) constant temperature
used two different aggregate classes in their experiments: and variable flux, and, (d) constant temperature and flux.
sand (0.15-0.6 mm) and gravels (5-10 mm), without
intermediate sizes. In the simulation, only gravels are 5.2. Constitutive behaviors
accounted for, to avoid using a very fine finite element Cement paste. The hcp is considered as a viscoelastic
mesh. While sand is not explicitely accounted for in the material with isotropic linear damage and thermal ex-
simulation, the surface fraction of gravels is increased to pansion (with α p the coefficient of thermal expansion).
58% in order to partly compensate for its absence (It is The creep deformation is divided into two compo-
reminded that the actual aggregate volume fraction is nents: a short-term recoverable creep, modeled with a
70% in Dubrovskii et al.'s experiment). The particle size Kelvin-Voigt unit, and a long-term irrecoverable viscous
distribution follows a Füller curve between the given flow, modeled with a time-dependent dashpot placed in
minimum and maximum gravel diameters in absence of series (see Fig. 15). With this model, the long-term creep
further data (Fuller and Thompson 1907). behavior of the paste follows a logarithmic function of
The RIVE of the missing sand fraction is neglected for time. The model relies on four parameters: the Young
this set of simulations. The properties of the cement paste modulus E p , the Poisson ratio v p , the creep viscosity
could be adapted in order to account for that additional η p (measured as the slope of the creep curve in the
expansion. However, to do so, one would need the ex- logarithmic scale), and the creep characteristic time τ p
Y. Le Pape, A. Giorla, J. Sanahuja / Journal of Advanced Concrete Technology Vol. 14, 70-86, 2016 79
Recoverable Non-Recoverable
Elasticity
100
creep creep Thermal expansion
η /τ
α(T-T 0 )
E η(1+t/ τ )
95
Effective properties [%]
90
η
250
continued
85
loading for
t > 100
200
80
Young's modulus
strain [arbitrary units]
Tensile strength
150
75
unloading at
100
0
ier-Cabot 1989), and is therefore taken as the radius of
strain [arbitrary units]
the largest sand particle neglected in the model. The
Fig. 15 Rheological model for the creep of cement paste elastic modulus and tensile strength of a cement paste
(top), typical creep curves for a loading at t = 0 showing element are reduced with the maximum temperature
the influence of the different components of the models reached in that element following the experimental re-
(using arbitrary parameter values) (middle) and sults of Fu et al. (2004) obtained on thin cement paste
stress-strain curve for a paste element (bottom). On the samples with a water/cement ratio of 0.5 (see Fig. 16).
creep curve E can be read from the initial deformation, The ultimate strain at rupture is dynamically changed
η from the asymptotic slope in the logarithmic scale, through the simulation to keep the fracture energy of
and τ as the intersection between the initial and final each element constant in absence of further data.
asymptotes. Aggregates. The mechanical properties of the aggre-
gates are taken as function of the irradiation, as described
in the previous sections. They are assumed to be inde-
5.3. Material properties
pendent of the temperature in the temperature range
The material properties of each phase are gathered in
recorded during the experiments. The Young's modulus
Table 2. Notations are explained in greater details in
(Giorla et al. 2015). of the quartz is taken as constant due to lack of data for
Cement paste. The material properties are taken from higher fluences. In a previous study (Giorla et al. 2015),
the authors found that the Young's modulus of the ag-
experimental results in the literature for cement pastes
gregate had no significant impact on the damage forma-
with a water/cement ratio of 0.5 as in Dubrovskii et al.
tion and propagation in the hcp during irradiation.
experiments whenever possible.
The elastic and creep properties at ambient tempera-
ture are provided from the basic creep experiments on 5.4. Results and discussion
sealed cement paste samples with a water/cement ratio of Two main outputs are derived from the simulations: (1)
Y. Le Pape, A. Giorla, J. Sanahuja / Journal of Advanced Concrete Technology Vol. 14, 70-86, 2016 81
the expansion of concrete, measured from the maximum expected from analytical models. The gap between the
displacements variation between the opposite edges of sample and the irradiation capsule wall should account
the sample, and, (2) the damage of the cement paste, for this effect in order to avoid wedging the samples into
measured as the surface average of the damage scalar the capsule, as what happened in Dubrovskii et al. ex-
variable over the elements representing the cement paste. periments.
The corresponding results are plotted as a function of the Damage. Damage in the cement paste (Fig. 18) in-
fluence on Figs. 17 and 18, respectively. The simulations creases rapidly in the sample after initiation. The neutron
at variable temperature show an earlier increase of the fluence at which this rise occurs is significantly reduced
expansion and damage as a function of the fluence. When when the complete history of temperature is simulated.
the temperature is maintained constant, the expansion Furthermore, the final stage of the damage propagation
and damage do not seem to depend on the neutron flux, (from 50% of damage in the cement paste to 65%) is
as both curves (constant and variable flux) coincide. All much sharper in the case of variable temperature than
simulations reach a plateau value in term of damage, constant temperature, as if it ranged from stable propa-
≈ 65%, which corresponds to the loss of integrity of the gation (constant temperature) to unstable (variable tem-
sample: at that point the cement paste provides no me- perature).
chanical restraint and the aggregates are free to swell. A closer analysis to the first damage increase reveals
Expansion. The volumetric expansion reaches values that, when considering a variable temperature and neu-
between ≈ 15% and ≈ 18% depending on the studied tron flux, the initial damage increase occurs during the
scenarios, which is consistent with the experimental first cooldown of the reactor, as shown in Fig. 19. During
observations from Dubrovskii et al., considering that the that cooldown, the flux remains strictly positive due to
difference of actual aggregate volume fraction in the the hypothesis of linearity between flux and temperature
experiment (0.70) and the numerical model (0.58). The (see section 4): the small increase in neutron fluence,
final expansion is about 30% higher than the integral of coupled with the low temperature, leads to a sharp in-
the average strains in the sample, the later corresponding crease of the RIVE. This behavior is also observed, albeit
to Eq. (13) assuming a volume fraction of aggregate of with a smaller extent, during the second cooldown. Note
58%. that this effect is not a modeling artifact as
This difference is explained by the severe mi- quasi-identical results are obtained when assuming a
cro-cracking and heterogeneities of the sample micro- neutron zero flux for temperatures below 30°C, i.e.,
structure, notably induced by the presence of large ag- considering a complete reactor shutdown.
gregates near its surface. It should be noted that gradient
effects are not considered in the present model, and may
further increase this difference. This has a direct practical
70
40
30
15
Concrete volumet ric expansion [%]
variable
constant T
20
constant Phi
constant T, Phi
variable
constant T
10
10
constant Phi
constant T, Phi
0
0 5 10 15 20 25
5
19
Neutron fluence [10 n/cm 2]
temperature
RIVE beyond ≈ 30% of the maximum swelling, which
100
100
variable Phi
constant Phi
limits the predictive capabilities of the proposed model.
However, it should be noted that because of the impor-
tant rigidity contrast between the aggregate and the ce-
80
80
Damage in cement paste [%]
Temperature [oC]
60
60
modification of the damage level in the hcp, e.g., (Giorla
et al. 2015), as the cracking development is primarily
40
40
governed by the aggregate expansion amplitudes.
Based on the extensive data collected by Bykov et al.,
a model accounting for the coupled effects of tempera-
20
20
ture and irradiation irradiation on quartz RIVE has been
developed: RIVE rate is described as a function of tem-
0
0
0 20 40 60 80 100
perature. In the absence of neutron flux, thermal an-
Time [d]
nealing can be neglected in the range of temperature
observed in test reactors. After implementation in the
Fig. 19 Average damage in the cement paste as a func-
code AMIE, the simulation of Dubrovskii et al.'s ex-
tion of time for the Group 2 at variable temperature and
periments on ordinary concrete made of quartz/quartzite
neutron flux (thick solid lines) and flux (dashed). Tem-
aggregate showed consistency with the observed expan-
perature variations in red.
sion, and pointed out that: (1) Measuring the dimensional
change of the diameter and/or height may lead to a sig-
The difference between simulations at variable or nificant overestimation of the actual average volumetric
constant temperature has a critical impact on the analysis expansion in the specimen; (2) Assuming constant neu-
of degradation of concrete structures exposed to irradia- tron flux and temperature can lead to important dis-
tion. Indeed, it indicates that the remaining service life of crepancies in term of post-irradiation damage and
such structure might be overestimated if the temperature volumetric expansion when compared to more realistic
is averaged over time. Operation cycles, and most nota- scenarios accounting for periodic reactor shutdowns or
bly temperature cycles should be accounted for in the other power cycles. Both effects certainly contribute to
analysis. Cooldown phases should be examined with care, the experimental scatter observed in data collected in the
as a residual flux at low temperature might still be able to open literature, and should be considered when analyzing
initiate damage in the microstructure. The fact that the these data, as well as for future simulations of irradiated
evolution of the damage does not seem to depend on the structures.
flux might lead to the hypothesis that creep does not play While the coupled effects of temperature and neutron
a significant role in degradation caused by irradiation. irradiation of quartz expansion have been thoroughly
However, these simulations correspond to short-term studied by Bykov et al. (1981), the same effects on other
experiments compared to actual structures, with very physical and mechanical properties need to be studied to
high deformations and without mechanical restraints, and obtain a thorough understanding of the potential effects
in which most of the damage occurs in the first months. of differential RIVE in polyphasic, i.e., poly-minerals,
This may not be applicable for conditions representative aggregates.
of the operation of nuclear power plants and requires Nevertheless, qualitative outcomes in terms of engi-
further analysis (Giorla and Le Pape 2015). neering applications can be made: (1) Tempera-
ture-induced point-defect annealing in irradiated miner-
6. Conclusions als affects greatly the RIVE kinetics and occurs even at
low fluence. This effect questions the validity of using
The irradiation of concrete biological shields in LWRs directly radiation-induced aggregate expansion curve
occurs at relatively low temperature (< 65 °C) while the obtained in test rector for LWRs conditions. In that par-
range of irradiation temperature in test reactor varies ticular perspective, it suggest that some test reactor data,
importantly from “low” temperature, i.e., < 40 °C) to e.g., (Gray 1971), may be overly conservative because
“high” temperature (> 150 °C). In that range of tem- obtained at about 45 °C, while other data should be re-
peratures, the amorphization rate, and thus, the RIVE considered as relevant for LWRs operation because ob-
rate of quartz is significantly affected. Based on data tained at quite high temperatures. (2) Because of the
collected on quartz, RIVE rate appears to be significantly effects of temperature on RIVE kinetics, LWR operation
accelerated when the irradiation temperature decreases. cycles can potentially affect radiation-induced damage
The evolution of the mechanical properties (elastic development in the biological shield.
properties and CTE) of irradiated α-quartz are expressed
as a function of the current density of the material (that is, Acknowledgements
the current level of RIVE) instead of the current fluence, This research is sponsored by the U.S. Department of
using analytic homogenization schemes. These proper- Energy (DOE) Light Water Reactor Sustainability Pro-
ties have not been characterized in the literature for
Y. Le Pape, A. Giorla, J. Sanahuja / Journal of Advanced Concrete Technology Vol. 14, 70-86, 2016 83
gram. This manuscript has been authored by UT-Battelle, uniaxial stress on alkali-silica reaction induced
LLC under Contract No. DE-AC05-00OR22725 with the expansion of concrete.” Cement and Concrete
U.S. Department of Energy. The United States Gov- Research, 42, 567-576.
ernment retains and the publisher, by accepting the arti- Elleuch, L., Dubois, F. and Rappeneau, J., (1972).
cle for publication, acknowledges that the United States “Effects of neutron radiation on special concretes and
Government retains a non-exclusive, paid-up, irrevoca- their components.” Special Publication of The
ble, world-wide license to publish or reproduce the pub- American Concrete Institute, 43, 1071-1108.
lished form of this manuscript, or allow others to do so, Eshelby, J., (1957). “The determination of the elastic
for United States Government purposes. The Department field of an ellipsoidal inclusion and related problems.”
of Energy will provide public access to these results of Proceedings of the Royal Society, 241, 376-396.
federally sponsored research in accordance with the DOE Field, K., Remec, I. and Le Pape, Y., (2015). “Radiation
Public Access Plan, effects on concrete for nuclear power plants, Part I:
(http://energy.gov/downloads/doe-public-access-plan). Quantification of radiation exposure and radiation
effects.” Nuclear Engineering and Design, 282,
References 126-143.
Amatuni, A. and Shevchenko, E., (1966). “Linear Fu, Y. F., Wong, Y. L., Poon, C. S., Tang, C. A. and Lin,
thermal expansion on monocrystalline quartz and P., (2004). “Experimental study of micro/macro crack
aluminum oxide.” Measurement Techniques-USSR, development and stress-strain relations of
10, 1256-60. cement-based composite materials at elevated
Ballato, A., (2008). “Basic material quartz and related temperatures.” Cement and Concrete Research, 34,
innovations.” In: Piezoelectricity., Springer Berlin 789-797.
Heidelberg. volume 114 of Springer Series in Fuller, W. B. and Thompson, S. E., (1907). “The laws of
Materials Science, 9-35. proportioning concrete.” Transactions of the
Bažant, Z. and Steffens, A., (2000). “Mathematical American Society of Civil Engineers, 59, 67-143.
model for kinetics of alkali-silica reaction in Giorla, A. and Le Pape, Y., (2015). “Computing
concrete.” Cement and Concrete Research, 30, creep-damage interactions in irradiated concrete.”
419-428. Journal of Nanomechanics and Micromechanics,
Bazant, Z. P. and Pijaudier-Cabot, G., (1989). Submitted.
“Measurement of characteristic length of nonlocal Giorla, A., Scrivener, K. and Dunant, C., (2014). “Finite
continuum.” Journal of Engineering Mechanics, 115, elements in space and time for the analysis of
755-767. generalised visco-elastic materials.” International
Beattie, J. A., Blaisdell, B. E., Kaye, J., Gerry, N. T. and Journal for Numerical Methods in Engineering, 97,
Johnson, C. A., (1941). “An experimental study of the 454-472.
absolute temperature scale. VIII. The thermal Giorla, A., Vaitová, M., Le Pape, Y. and Štemberk, P.,
expansion and compressibility of vitreous silica and (2015). “Meso-scale modeling of irradiated concrete
thermal dilation of mercury.” Proceedings of the in test reactor.” Nuclear Engineering and Design, 295,
American Academy of Arts and Sciences, 74, 371-388. 59-73.
Bykov, V., Denisov, A., Dubrovskii, V., Korenevskii, V., Graves, H., Le Pape, Y., Naus, D., Rashid, J., Saouma, V.,
Krivokoneva, G. and Muzalevskii, L., (1981). “Effect Sheikh, A., and Wall, J., (2014). “Expanded material
of irradiation temperature on the radiation expansion degradation assessment (EMDA) Volume 4: Aging of
of quartz.” Atomnaya Energiya, 51, 593-595. concrete.” Technical Report NUREG/CR-7153,
Dormieux, L., Kondo, D. and Ulm, F. J., (2006). ORNL/TM-2011/545. U.S. Nuclear Regulatory
“Microporomechanics.” John Wiley & Sons. Commission.
Dubrovskii, V., Ibragimov, S., Kulakovskii, M. Y., Gray, B., (1971). “The effects of reactor radiation on
Ladygin, A. Y. and Pergamenshchik, B., (1967). cement and concrete.” In: Proceedings of an
“Radiation damage in ordinary concrete.” Atomnaya Information Exchange Meeting on 'Results of
Énergiya, 23, 310-316. Concrete Irradiation Programmes', Commission des
Dunant, C. and Bentz, E., (2015). “Algorithmically Communautés Européennes, Brussels, Belgium.
imposed thermodynamic compliance for material Hilsdorf, H., Kropp, J. and Koch, H., (1978). “The
models in mechanical simulations using the AIM effects of nuclear radiation on the mechanical
method.” International Journal for Numerical properties of concrete.” Special Publication of The
Methods in Engineering, 104, 963-982. American Concrete Institute, 55, 223-254.
Dunant, C. and Scrivener, K., (2010). Johnson, W. and Parsons, W., (1944). “Thermal
“Micro-mechanical modelling of alkali-silica expansion of concrete aggregate materials.” Journal of
reaction-induced degradation using the AMIE Research of the National Bureau of Standards, 32,
framework.” Cement and Concrete Research, 40, 101-126.
517-525. Kelly, B., Brocklehurst, J., Mottershead, D. and
Dunant, C. and Scrivener, K., (2012). “Effects of McNearney, S., (1969). “The effects of reactor
Y. Le Pape, A. Giorla, J. Sanahuja / Journal of Advanced Concrete Technology Vol. 14, 70-86, 2016 84
radiation on concrete.” In: Proceedings of the Second Multon, S. and Toutlemonde, F., (2006). “Effect of
Information Meeting on Pre Stress Concrete and applied stresses on alkali-silica reaction-induced
Reactor Pressure Vessels and their Thermal Isolation, expansions.” Cement and Concrete Research, 36,
Brussels. 237-265. 912-920.
Keppens, V. and, Laermans, C., (1996). “Influence of Naus, D., (2005). “The effect of elevated temperature on
heat treatment on the tunneling states in concrete materials and structures - A literature
neutron-irradiated quartz.” Physical Review, B 53, review.” Technical Report ORNL/TM-2005/553. Oak
14849-61. Ridge National Laboratory. Oak Ridge, TN.
Kosinski, J., Gualtereri, J. and Ballato, A., (1991). Naus, D., (2010). “A compilation of elevated
“Thermal expansion of alpha quartz.” In: IEEE temperature concrete mmaterial property data and
Forty-Fifth Annual Symposium on Frequency Control, information for use in assessments of nuclear power
Piscataway NJ. plant reinforced concrete structures.” Technical
Kröner, E., (1977). “Bounds for effective elastic moduli Report NUREG/CR-7031, ORNL/TM-2009/175. Oak
of disordered materials.” Journal of the Mechanics Ridge National Laboratory.
and Physics of Solids, 24, 137-155. Otto, J. and Thomas, W., (1963). “Die thermische
Larive, C., (1997). “Apports combinés de Ausdehnung von Quarzglas im Temperaturbereich
l'expérimentation et de modélisation à compréhension von 0 bis 1060 °C.” Zeitschrift für Physik, 175,
de l'alcali-réaction et de ses effets mécaniques.” Thesis 337-344. (in German)
(PhD). Laboratoire Central des Ponts et Chaussées. (in Pabst, W. and Gregorová, E., (2013). “Elastic properties
French) of silica polymorph - a review.” Ceramics – Silikáty,
Le Pape, Y., (2015). “Structural effects of 57, 167-184.
radiation-induced volumetric expansion on Pedersen, A., (1971). “Radiation damage in concrete -
unreinforced concrete biological shields.” Nuclear measurements on miniature specimens of cement
Engineering and Design. 295, 534-548. mortar.” In: Benzler, H. Ed., Proceedings of an
Le Pape, Y., Field, K. and Busby, J., (2014). Information Exchange Meeting on 'Results of
“Degradation of concrete for nuclear structures: Concrete Irradiation Programmes, Commission of the
Identified mechanisms and knowledge gaps.” European Communities, Brussels, Belgium.
Transactions of the American Nuclear Society, Reno Primak, W., (1958). “Fast-neutron-induced changes in
NV. quartz and vitreous silica.” Physical Review, 110,
Le Pape, Y., Field, K. and Remec, I., (2015). “Radiation 1240-1254.
effects in concrete for nuclear power plants - Part II: Reuss, A., (1929). “Berechnung der Fließgrenz von
Perspective from micromechanical modeling.” Mischkristallen auf Grund der Plasticitätsbedingung
Nuclear Engineering and Design, 282, 144-157. für Einkristalle.” Zeitschrift für Angewandte
Le Roy, R., (1995). “Déformations instantanées et Mathematik und Mechanik 9, 49-58. (in German)
différées des bétons à hautes performances.” Thesis Rosseel, T., Wall, J., Field, G., Le Pape, Y., Naus, D.,
(PhD). Ecole Nationale des Ponts et Chaussées. Remec, I., Busby, J. and Bruck, P., (2014). “Radiation
Luikov, A., (1975). “Systems of differential equations of damage in reactor cavity concrete.” In: Fontevraud 8 -
heat and mass transfer in capilary-porous bodies Contribution of Materials Investigations and
(review).” International Journal of Heat and Mass Operating Experience to LWRs' Safety, Performance
Transfer, 18, 1-14. and Reliability, Avignon, France.
Maruyama, I., Nishioka, Y., Igarashi, G. and Matsui, K., Saouma, V. and Perotti, L., (2006). “Constitutive model
(2014). “Microstructural and bulk property changes in for alkali-aggregate reactions.” ACI Materials Journal,
hardened cement paste during the first drying 103, 194-2002.
process.” Cement and Concrete Research, 58, 20-34. Seeberger, J. and Hilsdorf, H., (1982). “Einuß von
Maruyama, I. and Sasano, H., (2014). “Strain and crack radioactiver Strahlung auf die Festogkeit and Struktur
distribution in concrete during drying.” Materials and von Beton.” Technical Report NR 2505. Institut für
Structures, 47, 517-532. Massivbau and Baustofftechnologie, Universität
Maruyama, I. and Sugie, A., (2014). “Numerical study Karlsruhe. (in German)
on drying shrinkage of concrete affected by aggregate Voigt, W., (1889). “Über die Beziehung zwischen den
size.” Journal of Advanced Concrete Technology, 12, beiden Elasticitätskonstanten isotroper Körper.”
279-288. Annalen der Physik, 274, 573-587. (in German)
Mayer, G. and Gigon, J., (1956). “Effets des neutrons Weissmann, S. and Nakajima, K., (1963). “Defect
rapides sur quelques constantes physiques du quartz structure and density decrease in neutron-irradiated
cristallin et de la silice vitreuse.” Le Journal de quartz.” Journal of Applied Physics, 34, 611-618.
Physique et le Radium, 18, 109-114. Welche, P., Heine, V. and Dove, M., (1998). Negative
Mayer, G. and Lecomte, M., (1960). “Effet des neutrons thermal expansion in beta-quartz.” Physics and
rapides sur le quartz cristallin et la silice vitreuse.” Le Chemistry of Minerals, 26, 63-77.
Journal de Physique et le Radium, 21, 846-852. Wittels, M., (1957). “Structural behavior of neutron
Y. Le Pape, A. Giorla, J. Sanahuja / Journal of Advanced Concrete Technology Vol. 14, 70-86, 2016 85
irradiated quartz.” Philosophical Magazine, 2, The macroscopic stress Σ being the spatial average
1445-1461. of the microscopic stress field σ , the effective behav-
Wittmann, F., (1968). “Surface tension shrinkage and iour of the REV is found to be linear elastic:
strength of hardened cement paste.” Materials and
Structure, 1, 547-552. Σ = ^ eff : E with ^ eff = ^( z ) : A( z ) (16)
Ω
Wittmann, F., (1970). “Einuß des Feuchtigkeitsgehaltes
auf das Kriechen des Zementsteines.” Rheologica Acta,
9, 282-287. (in German) where Ω
denotes the spatial average of a field over
Yano, T., Fukuda, K., Imai, M. and Miyazaki, H., (2007). the REV Ω . The local behaviour is now considered as
“Physical property changes of crystalline and linear thermoelastic:
non-crystalline SiO2 due to neutron irradiation and
recovery by subsequent annealing.” Journal of σ ( z ) = ^( z ) : (ε ( z ) − α ( z )ΔT ) (17)
Nuclear Materials, 367-370, 730-735. Proceedings of
the Twelfth International Conference on Fusion where the field of temperature increment ΔT is as-
Reactor Materials (ICFRM-12). sumed to be homogeneous across the REV. Alternatively,
Zhang, Q. and Ye, G., (2012). “Dehydration kinetics of this behaviour can be written using the eigenstrain ε L
Portland cement paste at high temperature.” Journal of
Thermal Analysis and Calorimetry, 110, 153-158.
σ ( z ) = ^ ( z ) : (ε ( z ) − ε L ( z )) with ε L ( z ) = α ( z )ΔT (18)
Zubov, V. and Ivanov, A., (1966). “Expansion of quartz
caused by irradiation with fast neutrons.” Soviet
Physics Crystallography, 11, 372-374. or, introducing the eigenstress σ p :
Zubov, V. and Ivanov, A., (1967). “Elasticity of quartz
irradiated with fast neutrons.” Soviet Physics σ ( z ) = ^( z ) : ε ( z ) + σ p ( z )
Crystallography, 12, 313-314. (19)
with σ p ( z ) = −^( z ) : ε L ( z ) = −^( z ) : α ( z )ΔT
Appendices
The effective behaviour of the REV is found to be
A. Appendix: Micromechanical Estimates of
linear thermoelastic (Levin's theorem):
Thermoelastic Properties of Quartz Polycrystal
Numerical computations at the concrete scale require the ∑ = ^eff : E + ∑ effp
thermomechanical properties of aggregates as a function
of irradiation. These properties are well documented for with ^eff = ^( z ) : A( z ) Ω
(20)
the quartz monocrystal Mayer and Lecomte (1960). The and ∑ effp = σ p ( z ) : A( z )
aim of this appendix is to upscale these properties from Ω
expansion α cr (θ , φ , ψ ) .
Voigt and Reuss Bounds and Estimates
The classical Voigt and Reuss bounds on the poly-
crystalline stiffness ^ pc are obtained from an average
over the whole rev of the stiffness or compliance tensor:
−1
^ cr (θ , φ , ψ ) = ^ ≤ ^ pc ≤ ^ = ^ cr (θ , φ , ψ )
−1 R V
or or .
(22)
The Reuss estimate considers that the stress field is A cr (θ , φ ,ψ ) = [I + PSC : ...(^ cr (θ , φ ,ψ ) − ^ SC )]−1 : ...
homogeneous in the whole rev, σ ( z ) = ∑ : −1 (29)
[I + PSC : (^ cr (θ , φ ,ψ ) − ^ SC )]−1
or .
A (θ , φ ,ψ ) = ^ cr (θ , φ ,ψ ) : ^
R −1 R
(26)
The self consistent estimate of the effective stiffness is
then obtained from (16):
Due to isotropy of crystal orientation, the tensor of
thermal expansion is also isotropic, and its estimates ^ SC = ^ cr (θ , φ ,ψ ) : [I + PSC : (^ cr (θ , φ , ϕ ) − ^ SC )]−1 :
or .
read: −1
(30)
: (^ cr (θ , φ ,ψ ) − ^ )]
−1
... [I + P
SC SC
or .
α V 1 = ^V −1 : ^ cr (θ , φ ,ψ ) : α cr (θ , φ ,ψ ) or .
(27) The latter is a non linear tensorial equation, as the Hill
and α R 1 = α cr (θ , φ ,ψ ) or . tensor PSC depends on the sought self-consistent stiff-
ness ^ SC . As the latter is isotropic, and as the average
where the averaging operator ' or .
' is also defined by operator yields isotropic tensors, this tensorial equation
(23) for second order tensors. can be reduced to two non linear equations over κ SC
and g SC , by projection over J and K . The
Self-Consistent Estimate self-consistent estimate α SC of the coefficient of ther-
Contrary to Voigt and Reuss bounds, so-called ho- mal expansion is obtained from (21) and (29).
mogenization schemes provide estimates of the effective
Journal of Advanced Concrete Technology Vol. 14, 144-162 April 2016 / Copyright © 2016 Japan Concrete Institute 144
Scientific paper
Abstract
For performance evaluation of existing reinforced concrete members under irradiation conditions, a numerical code called
“DEVICE” (Damage EValuation for Irradiated ConcretE), which takes into account the heat, moisture, and radiation
transport coupled with cement hydration, is proposed. This code is composed of the established computational ce-
ment-based material (CCBM) model and the one-dimensional deterministic transport Sn code “ANISN”. In the proposed
model, temperature-dependent irradiation-induced expansion of aggregate minerals and resultant strength deterioration of
concrete are introduced. Currently, the knowledge and modeling of irradiation-induced expansion of aggregate mineral is
limited only for α-quartz. DEVICE was used for evaluating the strength distribution of the decommissioned plant Japan
Power Demonstration Reactor (JPDR). Compressive strength distribution in a concrete biological shielding (CBS) wall of
the JPDR was obtained by core sampling, and the compressive loading test results were compared with the calculation
results. This comparison proved the practicality potential of DEVICE to predict the concrete strength distribution in a
CBS. In addition, concrete strength change and its distribution in a CBS of an anonymous two-loop pressurized water
reactor was simulated by DEVICE. The contributing factors for the change in the distribution of concrete strength at the
inner surface of the CBS are discussed. Furthermore, the ways of integrity evaluation other than the existing allowable
fast neutron fluence method are proposed and discussed as follows: 1) mineral composition-based allowable fast neutron
fluence; 2) strength prediction at the inner surface based on the expansion of mineral composition of aggregates and the
lower limit curve of the ratio of compressive strength of the specimen after irradiation (Fc) to that of the reference
specimen (Fco) as a function of concrete expansion; and 3) direct numerical calculation for seismic performance by
considering irradiation-induced volume expansion and degradation of concrete.
Table 1 Examples of expected fast neutron fluence and gamma-ray dose of Japanese NPPs after 60-year operation. (This
table is based on the information at 24th Jan. 2016.)
Type Name of plant Location Fast neutron (n/cm2) Gamma-ray dose (Gy)
BWR Fukushima 1st -1 Outside of reactor 3.0 × 1018 *1
< 2.0 × 108
Fukushima 1st -2 pressure vessel. 3.0 × 1018 *1 < 2.0 × 108
Shimane 1 3.0 ×1018 *1 8.0 × 107
Tsuruga 1 1.1 × 1017 *2 < 2.0 × 108
Fukushima 1st -5 Reactor pressure vessel - 2.8 × 1014 *1 2.3 × 104
Hamaoka 1 pedestal structure. 5.7 × 1013 *1 1.6 × 104
Fukushima 1st -3 1.6 × 109 *1
2.2 × 105
19 *3
PWR Mihama -2 Internal surface of the 6.2× 10 1.8 × 108
19 *3
Ikata -1 first biological shielding 6.0× 10 2.8 × 108
19 *3
Genkai -1 wall. 4.8× 10 2.8 × 108
19 *3
Mihama -3 4.7× 10 2.5 × 108
Takahama -1 4.5× 1019 *3
2.3 × 108
Takahama -2 4.5× 1019 *3
2.4 × 108
19 *3
Mihama -1 3.6× 10 1.5 × 108
: Decommisioning sites. *1 : ≥ 0.1 MeV, *2 : ≥ 1.0 MeV, *3 : ≥ 0.11 MeV.
although a decreasing trend above a dose of 5 × 105 kGy operation of an NPP. NRA recognized 1.0 × 1020 n/cm2
was reported. This result is almost consistent with the for fast neutrons and 2.0 × 1010 rad (= 2.0 × 105 kGy) for
results of Dubrovskii et al. (1966, 1967, 1968, 1970), gamma-ray irradiation as reference levels. The American
which concluded that compressive strength might be Concrete Institute (2012) Committee cites Hilsdorf et
affected above a total fluence of ~5 × 1019 n/cm2. On the al.’s results and suggests a fluence of 1.0 × 1021 n/cm2 at
contrary, Fujiwara et al. (2009) and Kontani et al. (2010) which loss of mechanical properties is observed; it also
pointed out that Hilsdorf et al.’s (1978) experimental recommended a conservative fluence limit of 1.0 × 1017
results included concrete or mortar that are not suitable n/cm2 to prevent lifetime radiation-induced degradation of
for LWRs. Therefore, Maruyama et al. (2013) summa- concrete. The British Standards Institute (1990) noted a
rized the relationship between neutron fluence and rela- neutron radiation level of 5.0 × 1017 n/cm2 below which its
tive compressive strength ratio (Fc/Fco, ratio of the effect on concrete structures of an NPP is not significant.
compressive strength of irradiated concrete to that of the Table 1 lists the expected fast neutron fluence of
reference concrete) for concrete or mortar applicable to Japanese NPPs after 60 years of operation, as obtained
concrete members in LWRs. They confirmed clear de- from the PLM document. Clearly, although fast neutron
creasing trend of Fc/Fco above 1–3 × 1019 n/cm2. Re- fluence levels of all plants will not exceed NRA’s refer-
cently, Field et al. (2015) summarized more compre- ence level, the fluence levels for some NPPs are expected
hensive data of irradiated concrete and its components, to exceed Hilsdorf et al.’s (1978) value. Consequently,
which also support the decreasing trend of physical prop- scientific knowledge is required for addressing this issue.
erties of irradiated concrete as neutron fluence increases. NRA launched a national project (Maruyama et al. 2012)
Several studies (Dubrovskii et al. 1967; Elluech et al. with this objective. Irradiation acceleration experiments
1971; Pedersen 1971; Hilsdorf et al. 1978; Kontani et al. have been limited with respect to specimen size, tem-
2010; Maruyama et al. 2012; Field et al. 2015) have perature control of the specimen, and ratio of gamma-ray
suggested that the mechanism of concrete deterioration to neutron flux (Maruyama et al. 2013). Moreover, large
due to irradiation is aggregate expansion that originates gaps exist between the specimens and the actual concrete
from metamictization or amorphization of silicious rock structures in a plant. In the present study, a numerical
minerals due to irradiation (Primak et al. 1955; Primak simulation model is proposed to compensate part of such
1958; Elluech et al. 1971; Eby et al. 1992; Bonnet et al. gaps. A new numerical code called DEVICE (Damage
1994; Douillard and Duraud 1996; Bolse 1999; Ewing et EValuation for Irradiated ConcretE) is developed by cou-
al. 2000). Due to aggregate expansion, the mortar matrix pling the computational cement-based material model
in concrete is damaged, resulting in deterioration of (CCBM) (Maruyama and Igarashi 2015), and moisture,
concrete’s physical properties, such as compressive heat, and radiation transport codes for predicting com-
strength and Young’s modulus, such as in the case of pressive strength changes and their distribution in a mas-
alkali–silica reaction in concrete (Giorla et al. 2015; Le sive concrete member subjected to irradiation on one side.
Pape et al. 2015). This model succeeds the current PLM process and aims to
The Nuclear Regulatory Authority (NRA) of Japan has contribute to the material basis or material potential
recommended a concrete integrity evaluation process for evaluation for integrity of reinforced concrete structures
NPPs, called “plant life management (PLM),” which during PLM. However, radiation impact on the structural
needs to be conducted for every 10 years after 30 years of performance is not addressed through this model and is a
I. Maruyama, K. Haba, O. Sato, S. Ishikawa, O. Kontani and M. Takizawa / Journal of Advanced Concrete Technology Vol. 14, 144-162, 2016 146
major concern (Pomaro et al. 2011; Salomoni et al. 2014; node, and phase composition of cement paste from the
Le Pape 2015). Therefore, this study focuses on material previous step.
strength, while similar approaches focused on the damage Considering radiation conditions on concrete, a de-
in CBS walls (Pomaro et al. 2011; Salomoni et al. 2014). terministic transport theory is used for the present study.
The proposed model is used for evaluating the de- In the ANISN code, originally developed in Oak Ridge
commissioned plant Japan Power Demonstration Reactor National Laboratory (ORNL) (Engle Jr. 1967), the
(JPDR). The strength of concrete core samples taken Boltzmann transport equation, which is an inte-
from a CBS of the JPDR is evaluated and discussed by gro-differential balance equation for the conservation of
comparing with the results obtained by DEVICE. Finally, radiative energy, is calculated using the discrete coordi-
concrete strength and related properties of concrete in a nates (Sn) approximation by discretizing spatial, angular,
CBS of an anonymous PWR are calculated and the results energy, and variables. This 1-D ANISN code is coupled
are discussed from the viewpoint of integrity evaluation. with the CCBM in the present study. The neutron and
gamma-ray flux distributions in concrete are calculated
2.Numerical modeling with the ANISN-ORNL by using the spectra of radiation
energy at the inner surface of the concrete and the
2.1 Coupling of the CCBM and neutron and cross-sections of the concrete member based on the ma-
gamma-ray transport calculation terial compositions and temperature. The energy spectra
In the present study, concrete behavior is calculated by at the inner surface of the concrete must be calculated
using the CCBM which covers the models for the rate of separately using a nuclear reactor model. In this system,
hydration of cement minerals, phase composition, re- the ANISN calculation region is divided into five zones,
sultant hygro-thermo-mechanical properties of cement in each of which the material compositions and tem-
paste, i.e., compressive strength, Young’s modulus, perature are averaged. The ANISN calculation involves
Poisson’s ratio, thermal expansion coefficient, autoge- P3-S16 approximation and a multi-group cross-section
nous shrinkage, drying shrinkage, heat capacity, heat library V7-200N47G in SCALE6 that has 200 neutron
transfer coefficient, water vapor sorption isotherms, and groups and 47 gamma-ray groups.
water transfer coefficient. In addition, the model em- The detailed procedure is described as follows:
pirically evaluates changes in the concrete compressive 1) Create input data “GIP” (Rhoades 1978), which is a
strength under heating or drying and takes into account function to calculate the cross-section of a concrete
cement paste strength changes due to its colloidal feature member on the basis of the atomic composition at each
as well as micro-defects produced around aggregates by FDM node.
the volume difference between the aggregate and mortar. 2) Calculate the cross-section of a concrete member at
For modeling the target concrete member, heat and each FDM node from the temperature and atomic
moisture transfer problems are calculated using one composition data using “GIP”.
dimensional finite differential method (1D-FDM) with 3) Calculate the radiation flux at all FDM nodes, which
an implicit method and a time step of 0.001 hour. All the correspond to the calculation nodes for the ANISN
parameters such as moisture capacity, moisture transfer code, based on the cross-section of the concrete
coefficient, water consumption, heat capacity, heat member.
transfer coefficient, and heat consumption are calculated 4) Calculate heat production rate at each node from the
by the information of equilibrium and history of tem- corresponding ANISN calculation results.
perature and relative humidity of the concrete at each 5) Calculate heat and moisture transfer and hydration
process of cement, and then go to step 1). calcium silicate hydrates (C–S–H), which are formed
Figure 1 shows the process of introduction of the from alite and belite. Portlandite and the Ca–O layer in
ANISN codes to the CCBM. Figure 2 shows the cou- C–S–H are ionically bonded and may not be largely
pling process and the calculation flow. affected by gamma rays without the interaction of water.
The Si–O covalent bonds in C–S–H may be affected
2.2 Modeling of concrete deterioration due to slightly, but the structure of C–S–H is sustained by the
irradiation Ca–O layer (Gartner 1997; Gartner et al. 2000). Even
2.2.1 Impact of gamma rays though the atoms in C–S–H are dislocated by secondary
First, the impact of gamma rays is discussed based on electron collision, since the Ca–O layer has a winding or
literature (Kontani et al. 2012; Maruyama et al. 2012; wavy shape and is not ordered because of the attachment
Kontani et al. 2013). of silicates (Gartner 1997), such dislocations do not
Gamma rays affect material through electronic exci- largely impact the physical properties of hardened ce-
tations. The excess electronic energies of electrons in the ment paste.
excited states are transformed to vibrational energies of Metastable calcium peroxide octahydrates can be
atoms composing the materials, resulting in breaking of produced in hardened cement paste with high water
the covalent bonds and producing radicals. From this content (Bouniol and Aspart 1998; Vodák et al. 2005;
viewpoint, the covalent nature of the concrete materials Lowinska-Kluge and Piszora 2008), altering the strength
can be affected by gamma rays. In case of the Si–O bonds and pore size distribution and especially resulting in
in α-quartz, the impact is significant beyond ~1.0 x subsequent carbonation.
1012 Gy for gamma rays (Douillard and Duraud 1996; The drying and heating associated with gamma-ray
Ichikawa and Kimura 2007). Hydroxy groups are also irradiation affects the properties of hardened cement
affected by gamma rays. Water is decomposed as hy- paste (Maruyama et al. 2014a), because C–S–H formed
droxyl and hydrogen radicals, forming hydrogen, oxygen, from Portland cement has a colloidal feature, and the
and hydrogen peroxide (Bouniol and Aspart 1998). morphology of C–S–H is affected by the drying and
As the energy of gamma rays are absorbed by the heating process (Pihlajavaara 1974; Litvan 1976, Parrott
target material during the decay process, heating always et al. 1980; Parrott and Young 1981; Cong and
accompanies gamma-ray irradiation. In addition to ele- Kirkpatrick 1995; Thomas and Jennings 2003; Jennings
vated temperature due to gamma-ray irradiation, de- 2008; Thomas et al. 2008; Maruyama et al. 2014a; Ma-
composition of water enhances drying of the concrete. ruyama et al. 2015). Therefore, gamma heating largely
Therefore, the combined impacts of drying and heating impacts the physical property of hardened cement paste.
alter the physical properties of concrete. Natural rock and chemically stable aggregates are
Hardened cement paste, which is a product of hydra- generally used in concrete for LWRs. Unlike hardened
tion of Portland cement and water, is the main binder cement paste, an aggregate is composed of
matrix in concrete. Cement is generally composed of well-crystallized minerals, and the major bonds of rock
alite, belite, aluminate phase, ferrite phase, and gypsum. minerals are covalent and ionic bonds. Therefore, ag-
Major cement hydrates that are responsible for the gregates can be considered to be not largely affected by
strength of hardened cement paste are portlandite and gamma rays through electronic excitations.
Fig. 2 Calculation flow of CCBM and ANISN-ORNL, their interaction, and prediction of physical properties of concrete.
I. Maruyama, K. Haba, O. Sato, S. Ishikawa, O. Kontani and M. Takizawa / Journal of Advanced Concrete Technology Vol. 14, 144-162, 2016 148
Collision of gamma rays on a material surface results because of lack of scientific data (Kontani et al. 2013).
in a photoelectric effect, Compton scattering, or pair
production, and thereby causes electrons, and occasion- 2.2.2 Impact of neutrons
ally positrons, to be ejected or produced. These electrons Since neutrons are electrically neutral, they do not have
collide with the material and eject more electrons, the an ionizing effect. Neutrons always interact with a nu-
secondary electrons; this process continues until the cleus by scattering and absorption. Two types of scat-
energies of the secondary electrons become lower than tering occur: inelastic and elastic scattering. Both change
the ionization energy of the material. Only when the first the direction and speed of neutrons and the number of
gamma-ray energy is very high, the secondary electrons protons and neutrons in the interacted nucleus remains
can collide and move atoms in target materials. the same as before. In elastic scattering, a neutron col-
Reproducible facts indicate that α-quartz, which is one lides with a lighter nucleus such as that of H, C, or O and
of the major rock minerals for concrete aggregates, may a neutron is ejected, thus preserving the total kinetic
undergo expansion due to irradiation. Fully metamictized energy of the nucleus. On the contrary, in inelastic scat-
or amorphized condition is attained at ~6 × 1027 eV/g (= tering, a neutron collides with a heavier atomic nucleus,
9.6 × 1012 Gy) by electron irradiation. Electron damages which attains an excited state owing to internal rear-
are determined by the ionization cross-section; the rangement and subsequently releases radiation. When a
damage decreases with increasing electron energy. neutron is absorbed by a nucleus, a wide-energy-range
However, if the electron energy becomes more than 1.5 radiation can be released, or fission can be initiated.
MeV, the irradiation process is accelerated by a knock-on Cement hydrates always coexist with evaporable wa-
displacement mechanism (Inui et al. 1990). ter under the environmental condition around the con-
Maximum ejected electron energy of photoelectric crete in LWRs, and the dissolution and precipitation of
effect and Compton scattering is almost the same as the these hydrates are in equilibrium. Atoms in the cement
energy of irradiated gamma-ray, when energy of an in- hydrates can be collided by neutrons, dislocating atoms
cident electron is more than MeV, while it is 0.51 MeV in from the hydrates. However, the equilibrium state plays a
case of pair production. The possibility of those events is healing role for this neutron damage in cement hydrates.
a function of atomic mass and therefore, the Compton In addition, C–S–H generally has a short-period recur-
scattering event which will be brought by gamma-ray sive structure that cannot be detected by powder X-ray
energy ranging about 1.8 ~ 20 MeV should contribute to diffraction in the layer-piling direction; therefore, the
the amorphization of α-quartz through recoiled electrons. defects in the structure may not largely impact its
In case of a concrete structure and neighboring surfaces physical properties. The only possible damage by the
of an LWR, the gamma-ray levels are 100 KeV to 10 collision of neutrons is that the portlandite, with its
MeV (Fukuya et al. 2002), therefore, gamma rays are maximum crystal size being approximately 1 micron,
considered to have some effect on the concrete compo- may expand because of the dislocation of atoms and
nents. This impact is not fully understood quantitatively defects in its structure (Seeberger and Hilsdorf 1982).
and needs further investigation for clarity, but based on However, the change in hardened cement strength under
the experiment by Inui et al. (1990), the impact of irradiation conditions is not confirmed experimentally
Compton recoiled electron seemed almost the same order (Gray 1971; Elleuch et al. 1972; Seeberger and Hilsdorf
of that by electronic excitations, therefore, the rate of 1982).
amorphization due to gamma-ray irradiation is negligible As stated in the previous section, aggregates are well
in case of aggregate or hardened cement paste in LWR. crystalized, but there is a facts that α-quartz is expanded
In the present study, the impact of gamma rays on due to irradiation. Previous neutron irradiation experi-
hardened cement paste and concrete is assumed to be the ments of α-quartz showed almost consistent density of
same as the impact of heating and drying on these mate- the final state (Wittels and Sherrill 1954; Primak et al.
rials, which has been already modeled in previous studies 1955; Primak 1958; Bykov et al. 1981; Bonnet et al.
(Maruyama et al. 2014a, 2014b; Maruyama and Igarashi 1994), with its reduction value in the range 15%–18%
2015). The existing data (Vodák et al. 2005) also support and a linear expansion of approximately 5%-6%. Neu-
this assumption. Production of calcium peroxide octa- tron irradiation damages are generally described ac-
hydrate must be related to the portlandite and water cording to three fluence levels (Douillard and Duraud
content in the cement paste. However, in the case of 1996). For low neutron fluence (φn ≤ 1 × 1018 n/cm2), the
mature cement paste under a sealed condition, only a irradiation induces point defects, mainly E’ centers (≡Si•)
small amount of capillary water exists (Muller et al. and oxygen hole centers. The point defect concentration
2012), and the possibility of a reaction may be very lim- increases as a power function of the fluence. For medium
ited due to lack of space for precipitation. In addition, neutron fluence (1 × 1018 ≤ φn ≤ 1 × 1020 n/cm2), α-quartz
due to lack of experimental experiences, we will not undergoes an α–β transition, immediately followed by a
consider this phenomenon at the present research stage. complete amorphization for a neutron fluence greater
The production of gases by the decomposition of capil- than 1 × 1020 n/cm2 (Douillard and Duraud 1996).
lary water may also impact the water vapor movement in Note that to the best of authors’ knowledge, the final
concrete, but this phenomenon is also not considered here amorphized conditions attained after electron irradiation
I. Maruyama, K. Haba, O. Sato, S. Ishikawa, O. Kontani and M. Takizawa / Journal of Advanced Concrete Technology Vol. 14, 144-162, 2016 149
or ion bombardment have not been compared with those for NPPs, a smaller tangent than that for the overall trend
after neutron irradiation in terms of α-quartz density. is confirmed. Therefore, based on this, the following
Therefore, we could not conclude that gamma rays have equation is formulated:
the same impact as neutrons.
α-Quartz is the major component of rocks and is pre- ⎛ − Ea , quartz ⎞
K (T ) = 3.3 ×1023 ⋅ exp ⎜ ⎟ (2)
sent in many kinds of aggregate. Therefore, aggregate is ⎝ RT ⎠
likely expanded under the neutron irradiation. This ag-
gregate expansion in concrete leads to expansion of where Ea,quartz: activation energy for the nucleation and
concrete, accumulation of damage in concrete, and re- growth of the amorphized region in α-quartz, 22500
sultant degradation of the physical properties of concrete (J/mol), and R: gas constant (J mol-1 K-1). During this
(Dubrovskii et al. 1967; Elluech et al. 1971; Hilsdorf et formulation, minimum energy of the neutron flux that
al. 1978; Field et al. 2015; Le Pape et al. 2015). There- affects damage accumulation of α-quartz is 0.1 MeV, as
fore, the modeling of aggregate expansion is necessary. obtained from the flux information of reactors used in the
Damage accumulation and volume expansion of α-quartz experiment by Bykov et al (1981).
are well described by the nucleation and growth model Assuming simplified homogenization, aggregate ex-
proposed by Avrami (1941) (Weber 1990; Harbsmeier pansion due to irradiation is expressed by the following
and Bolse 1998; Bolse 1999; Field et al. 2015). Our equation:
proposed model is also based on this model. In terms of
the temperature condition, the result of a distinguished ε n , agg = ∑ ε n ,i vi (3)
experiment reported by Bykov et al. (1981) is shown in i
Fig. 3. Based on this figure, the nucleation and growth where εn,i : expansion (or shrinkage) of mineral i caused
model, which is shown in the following equation, is used by neutron irradiation, and vi: volumetric ratio of mineral
to evaluate the temperature condition and expansion i in the aggregates. Rock minerals other than α-quartz
strain εn(n). may possibly expand by the irradiation (Ichikawa and
Kimura 2007); however, in the present study, due to the
⎛ ⎛ −n ⎞
d
⎞
ε n , quartz ( n ) = ε n , quartz ,∞ ⋅ ⎜1 − exp ⎜ ⎟ (1) limitation of reliable data, only α-quartz is considered.
⎜ ⎟ ⎟
⎝ K (T ) ⎠ The role of the matrix around α-quartz grains is also
⎝ ⎠
ignored. For example, a chlorite matrix produced in
where εn,quartz, ∞ : maximum expansion strain of α-quartz sandstone by diagenesis shows large shrinkage under
(18% in case of volume change and 6% in case of linear drying and heating (Igarashi et al. 2015), but such prop-
strain); n: neutron fluence (n/cm2) with given energy erties of the cement mineral are not considered directly
criteria, indicating the minimum energy level for amor- here. In addition, stress re-distribution between the ma-
phization; K(T): temperature-dependent factor (n/cm2) trix and many types of mineral grains due to α-quartz
for nucleation and growth of the amorphized phase; T: grain expansion is not considered. This phenomenon
absolute temperature (K); and d: dimensionality factor, should play an important role in predicting the aggregate
normally ranging from 2 to 5 and is 2.38 in this study expansion at the stage of point defect concentration.
(Field et al. 2015). These are the limitations of the present model and need to
The neutron fluence at which half the volume expan- be improved in future work for further precise prediction.
sion occurs at the fully amorphized condition is evalu- Based on aggregate expansion, concrete expansion is
ated by an Arrhenius plot as shown in Fig. 4. As the
figure shows, the overall trend conforms with the Ar-
rhenius law; however, in the realistic temperature range
Fig. 3 Temperature-dependent behavior of the expansion Fig. 4 Arrhenius plot for α-quartz expansion due to neu-
of α-quartz due to neutron irradiation, described by tron irradiation as evaluated by the nucleation and growth
Bykov et al. (1981). model.
I. Maruyama, K. Haba, O. Sato, S. Ishikawa, O. Kontani and M. Takizawa / Journal of Advanced Concrete Technology Vol. 14, 144-162, 2016 150
estimated for predicting the physical properties of the Table 2 Operation summary of JPDR-I and JPDR-II (Idei
concrete. As the concrete properties are dominated by et al. 1990).
cracks due to mismatched volume change between the JPDR-I JPDR-II
aggregate and matrix (paste or mortar) and due to the Total operation time 14096 hour 3014 hour
colloidal nature of hardened cement paste (Maruyama et Reactor thermal power 18501 MWD 2951 MWD
al. 2014b), concrete expansion strain must be a key fac- output
tor that reflects the damage and cracks of concrete caused Total power generation 12236 hour 1992 hour
by aggregate expansion (Dubrovskii et al. 1967; Le Pape time
et al. 2015). Total power generation 1.22 × 108 kWh 1.47 × 104 kWh
Equivalent rated power 1.13 EFPY 0.09 EFPY
There exist models regarding the concrete volume
operation time
change, especially under drying, based on the informa-
tion of mixture proportion and properties and volume
change of the components. This topic has been discussed stantaneous loading is applied in order to consider
since the early 20th century (Carlson 1938 ; Pickett cracking in the mortar and long-term compressive creep
1956). Composite laws based on the continuous theory due to aggregate expansion.
are developed by many researchers (Hansen and Nielsen In the present study, the effect of fine aggregate ex-
1965; Hobbs 1974; Kishitani and Baba 1975; Hansen pansion in mortar is also considered in the same way as
1987). However, for the prediction of concrete volume for coarse aggregate and mortar expansion.
change, cracking around the aggregate as a result of The expansion strain is a key parameter for deter-
mismatched volume change between the aggregate and mining the deterioration of physical properties of con-
matrix should be taken into account. This impact is crete, and the compressive strength is a representative
generally implicitly considered as a result of a lower property for concrete structures. The relationship be-
Young’s modulus of matrices than the measured value tween the ratio of compressive strength after irradiation
under an instantaneous loading experiment, which is (Fc) to that of the reference specimen (Fco) and the linear
attributed to creep or invisible cracking (Hansen 1987). expansion strain of the specimen due to irradiation is
Recently, the impact of the interfacial transition zone obtained as shown in Fig. 5. As shown in Fig. 5, Fc/Fco
(ITZ) is also emphasized for evaluating concrete decreases with the expansion of the specimen. The ex-
shrinkage, because such continuous theory-based on perimental results by Dubrovskii et al. (1967), Du-
prediction cannot precisely explain the concrete shrink- brovskii et al. (1970), Pedersen (1971), and Elleuch et al.
age affected by aggregate size (Maruyama and Sugie (1972) showed relatively the same trend, while the ex-
2014). Therefore, in the present study, a composite law perimental results by Dubrovskii et al. (1967) showed an
proposed by Hobbs (1971) is applied with slight modi- upper trend compared with those by others. To evaluate
fication regarding the ITZ impact, as given below. these trends, Fc/Fco of concrete specimens affected by
alkali–silica reaction (ASR) and the linear expansion
ε c n + 1 + ( 2mn − n − 1) VAgg ,c strain of the samples are plotted as shown in Fig. 5 (Na-
= (4)
εm n + 1 + ( n − 1) VAgg ,c
n = E Agg / Em (5)
m = ε Agg ε m (6)
kano et al. 1987; Swamy and Al-Asali 1988; Tomita et al. JPDR-I attained self-sustaining nuclear chain reaction
1988; Kubo et al. 2006; Ueda et al. 2011; Itoh and Ma- in August 1963, and continued intermittent operation
ruyama 2014). This is because the deterioration mecha- until September 1969. It was then upgraded to JPDR-II,
nism due to aggregate expansion is common to those which operated from May 1972 to March 1976. The
concrete degradations (Graves et al. 2011). As a result, operation summary is given in Table 2 and the operation
the data by Dubrovskii et al. (1970), Pedersen (1971), history of the reactors is presented in Fig. 6 (Idei et al.
Elleuch et al. (1972) almost correspond to the minimum 1990).
values of ASR data and the data by Dubrovskii et al. The concrete strength of cores extracted from a CBS
(1967) almost correspond to the maximum values of of the JPDR was evaluated by the proposed numerical
ASR data. The trends are consistent. A lower boundary method. During the construction of the JPDR, concrete
curve is represented by the following equation: for the CBS walls was casted from March to November
1962. The concrete was composed of ordinary Portland
Fc / Fco = 0.3 − 0.7exp ( −1.4 ⋅104 ε n,c ) (9) cement from Hitachi Cement Co. Ltd., Naka River gravel,
and sea sand taken from a place near the site. Chemical
Note that some differences exist in the experimental composition and physical properties of the cement used
conditions of the abovementioned previous studies. For a are shown in Tables 3 and 4. The timing of these records
specimen of young concrete at the commencement of does not correspond to the date of construction of the
exposure to irradiation or ASR environment, an increase CBS, but the variation in cement quality was small, and
in the strength occurs because continuous hydration hence, these data seem applicable to the cement of the
increases the compressive strength of the specimen. High CBS. The mixture proportion of concrete for the CBS is
temperature (~100 ºC) also causes an increase in the shown in Table 5. The strength for mixture proportion
C–S–H strength and additional hydration of residual was 35 MPa and the slump was 7.5 cm (Idei et al. 1990;
cement minerals. These impacts yield variation in the
experimental results. Using the lower boundary curve as
the safety values, such unexplained phenomena are
evaluated.
Japan (Iwata 1981) and using this value, the air content in Table 6 Mineral composition of cement for calculation.
concrete was calculated as 5.5%, which is relatively large Alite Aluminate Ferrite Gypsum
for the air content required for resistivity to freezing and Belite (%)
(%) phase (%) phase (%) (%)
thawing reaction. Therefore, the air content and density 53.9 23.1 9.0 9.6 4.4
of river gravel and sea sand were assumed to be 4.5% and
2.61 g/cm3, respectively. perature, water content per cement paste in concrete,
No information regarding the quartz content in ag- degree of hydration of cement, and compressive strength
gregates is available in literature. However, the estimated distribution in CBS at arbitrary concrete ages are shown
density is similar to that of α-quartz (2.65 g/cm3), and to in Fig. 11. Due to its size, the heat of hydration accu-
take into account the risk of neutron irradiation in the mulated up to ~60 °C. In case of some aggregates, con-
calculation as much as possible, the aggregates were crete strength development was stagnated or deteriorated
considered as composed of α-quartz only. Therefore, the because of increase in number of large pores by forma-
shrinkage of aggregates is zero and thermal expansion tion of densified outer C-S-H due to elevated temperature
coefficient is 10.0 × 10-6 /K (Igarashi et al. 2015). (Sugiyama and Masuda 1999; Gallucci et al. 2013) and
For the construction process, we assumed 1 hour for the difference between the volume changes in the ag-
transport from a ready-mixed concrete plant to the CBS gregate and matrix and resultant damage accumulation
location. The mold was assumed to be made of steel and around aggregates (Son and Hosoda 2010; Maruyama et
the heat transfer coefficient to be 14 W/m2/K. The inside al. 2014b). However, in this calculation, silicate content
steel plate liner was considered to be used for the mold. was high and the thermal expansion of the aggregate was
Therefore, the inner part of the CBS was not exposed to 10 × 10-6 /K. Therefore, the damage does not have a large
the dry atmospheric air. Environmental conditions were impact on this calculation and the strength distribution is
293 K and 60% RH, which are the average values of the similar to that of the hydration degree. After 1 year, the
Japanese climate. The timing of demolding was not re- degree of hydration was about 0.87 and the distribution
corded. Based on the history of concrete casting, the age of hydration was almost homogenous except for the outer
of demolding of concrete was assumed to be 28 days surface where the cement hydration was stagnant owing
after casting. Then, immediately before the operation and to water loss after demolding and resultant moisture
11 months after concrete casting, the resin coating was evaopration to the surrounding air.
finished at the outer side of the CBS. The resin coating After about 16 years, operation of the JPDR was
was assumed to have 1/100 of the permeability of con- stopped. During the operation period of 6.3 years, the
crete. maximum temperature was observed. In this case, the
A 1500-mm-thick member was used for this calcula- maximum temperature was recorded at a 400 mm depth
tion because this was the minimum thickness of the CBS. from the inner surface. This temperature was attained
During the operation, the temperature of the air on the mainly by gamma heating accompanied by water cooling
surface of the steel plate liner was assumed to be 43 °C, from the embedded pipes. Therefore, we can estimate
and the concrete temperature around the pipe was as- that the water cooling pipe largely impacts the reduction
sumed to be 25 °C. This is because the data of tempera- of temperature in the CBS caused by gamma heating. In
ture of cooling water and ventilating air are not available general, water movement is derived by the temperature
and existing experimental data and former calculations gradient in the thick concrete wall (England and Ross
are reliable. In the off-operation state, environmental 1972), but the CBS of the JPDR does not have a large
condition of both sides of the CBS was assumed to be temperature difference between its inner and outer sur-
298 K and 50% RH. face parts, because the water cooling pipe was effective.
Regarding the radiation condition, neutron and Resultantly, although a small variation in the water con-
gamma-ray flux was assumed to have a linear relation-
ship with the output power, as shown in Fig. 6. In the
present calculation, the impact of the water cooling pipe
has been studied beforehand, because water content and
iron within the cooling pipe may affect the shielding and
resultant secondary gamma rays and heat production.
Our prior calculation, which is shown in Fig. 10, exhibits
little impact on secondary gamma rays. Therefore, the
calculation mesh did not consider these water cooling
pipes.
As for meshing for the calculation, a 7.5 mm discrete
mesh was applied.
tent was observed around the cooling pipe, it is generally liner plate as well as at the outer surface by resin
almost constant. This is the main reason that the concrete coating also have a homogenizing effect on the con-
strength is almost homogenous in the CBS. In addition, crete strength of the CBS. In the present calculation,
the fast neutron fluence does not exceed 1.0 × 1018 n/cm2 the depth affected by drying is limited to 100 mm from
after the operation. Within this value, the α-quartz is the surface.
almost inert and we cannot see any impact on the con-
crete strength. The experimental results of the concrete 4. Calculation of an anonymous PWR
strength in the BSW were compared with the numerical reactor
results as shown in Fig. 12. The comparison shows that
the calculation results are comparable to the experimen- In the present study, a two-loop pressurized water reactor
tal data. From the experimental data and supporting (PWR) is studied for the discussion of the integrity
numerical simulation, the following points can be drawn, evaluation of a CBS. This is because a two-loop PWR
which can contribute to the understanding of aging imparts a high risk of neutron fluence on the CBS. The
management. target CBS is assumed to be 2.0 m thick and the concrete
1) A fast neutron fluence of less than 1.0 × 1018 n/cm2 at component is assumed to be the same as that of the JPDR.
room temperature (<40 °C) does not significantly The radiation boundary condition is calculated from
impact the concrete strength. Fukuya et al. (2002), and the results of neutron and
2) The control of temperature at the inner part of the CBS gamma-ray flux at the inner surface of the CBS are
has an impact to homogenize the concrete strength in shown in Fig. 13. With regards to the boundary condition
the CBS. This tendency is favorable to evaluate the of moisture transfer at the inner and outer surfaces, the
representative concrete strength in thick concrete walls. conditions of sealed from and open to the ventilated air
3) The sealing conditions at the inner surface by a steel are assumed. For temperature conditions at operation,
64 °C and 53 °C are considered at the inner surface,
based on the code regulation and our personal interview
for the two-loop PWR condition, respectively, while
35 °C is assumed for the outer surface. The temperatures surface are summarized in Fig. 14. The fast neutron
at both sides of the CBS during the periodical inspection, fluence at the surface position was about 6.2 × 1019 n/cm2
corresponding to the absence of heat from the reactor and after 60 years of operation (Fig. 14 (b).) During the op-
continuous 91 days (25% of a year) allocated for this eration, temperature at the surface of the concrete was
period, is 20 °C. The conditions and calculation pa- about 64 °C, but due to neutron scattering and
rameters are summarized in Table 7. In the construction gamma-ray heating, maximum temperature was ob-
process, demolding is after 28 days of casting, and 1 year served at 40 mm from the inner surface. In addition, a
and 3 months is assumed as the construction period. The negative gradient of temperature was confirmed from the
construction period is relatively short compared with the position of peak temperature to the outer surface of the
real construction process. Moreover, after the construc- CBS (2000 mm from the inner surface) (Fig. 14 (d)).
tion process, the first operation was started. Due to this temperature gradient, the resulting accumu-
The calculation results of 337s (Notation and pa- lation of vapor pressure and a little, but not negligible,
rameters are shown in Table 7) in which the CBS was water transport at the outer surface of the CBS, water
sealed on both surfaces and irradiated from the inner slightly moved from the inner part to the outer part (Fig.
Fig. 14 Calculation results for 338s until 60-year operation. (a) Compressive strength, (b) fast neutron (>0.1 MeV) fluence,
(c) evaporable water per cement paste, and (d) temperature.
I. Maruyama, K. Haba, O. Sato, S. Ishikawa, O. Kontani and M. Takizawa / Journal of Advanced Concrete Technology Vol. 14, 144-162, 2016 156
14 (c)). In particular, from 0 to about 500 mm from the evaporable water at the inner surface.
inner surface, the water was slightly reduced. Although Furthermore, the temperature impact on strength
the evaporable water tended to move to the outer surface, change at the inner surface of the CBS of 337s and 326s
a sufficient amount of water remained in the inner part of is discussed. Calculation results are shown in Fig. 16.
the CBS and cement hydration gradually continued (Fig. The distribution of evaporable water content is not dif-
14 (a)). ferent from the others. The largest impact was deteriora-
Theoretically, water loss should relate to the reduction tion of concrete at the inner part. In the case of 337s, the
of both heat capacity and shielding performance; how- strength of concrete at the inner surface was 16 MPa,
ever, in this case, significant temperature difference while it was 14 MPa in the case of 326s. The depth of
could be found. The temperature difference was within concrete affected by irradiation was 12 cm in the case of
0.02 °C during the operation period. 326s, while it was 10 cm in the case of 337s. This dif-
After 20 years, a substantial drop in the concrete ference was caused by the temperature difference of
strength was observed at the inner surface. The fast neu- 11 °C at the inner surface, which will result in the dif
tron fluence was 1.61 × 1019 n/cm2 at the surface, and the ference in neutron-induced expansion of α-quartz (see
concrete strength was deteriorated by neutron irradiation Fig. 3). The changes in compressive strength at the inner
at the region from 0 mm to 40 mm from the inner surface. surfaces for 337s and 326s are almost the same. However,
At this point, the concrete expansion in a hypothetical the difference of depth of concrete strength deterioration
free restraint condition at the surface was 0.32% and at might considerably affect the structural performance of
the region 40 mm from the surface was 0.066%. reinforced concrete members.
After 60 years operation, the depth at which concrete Water, which is the main source of hydrogen atom in
strength was deteriorated by neutron irradiation was concrete, is considered to have a large impact on
about 100 mm. The expansion strain of concrete at the shielding performance of concrete. Therefore, the evap-
surface was 3.2% and at 100 mm from the inner surface, orable water content distribution and the ratio of fast
where the main rebars are generally embedded, was neutron fluence after 60 years of operation are shown in
0.12%. Therefore, in this case, the structural integrity is Fig. 17. In this figure, the ratios of fast neutron fluence
questionable, because we do not have enough knowledge for 326s and 337o to that of 337s as well as the evapor-
of the bond behavior with such expanded concrete con- able water content distributions for 337s, 337o, and 326s
ditions. are plotted. The figure shows that increase (or preserv-
Next, the case of the CBS with open surfaces is dis- ing) of water content in the CBS reduces the neutron
cussed. The comparison of calculation results of 337s fluence in the CBS, while decrease of water content at
and 337o is shown in Fig. 15. A large difference was the inner surface increases the neutron fluence in the
confirmed in the evaporable water content, because wa- CBS. However, concrete deterioration caused by irra-
ter was evaporated from the surfaces of the CBS. Espe- diation-induced aggregate expansion was confirmed
cially, near the inner surface, which was heated by irra- within 100 mm from the inner surface. Considering the
diation and attained about 64 °C, a large amount of water range up to which neutrons have the impact, the sealing
was evaporated from the surface after operation began. condition at the inner surface has a large effect. At a
The depth of the dried zone was approximately 500 mm. depth of 100 mm, the fluence varied about 10%. This
This dried zone reflected the water vapor desorption value is linearly related to the time span of concrete
isotherm model considered in the CCBM. In the model, deterioration; therefore, the steel liner is very effective
water vapor BET surface area is a function of tempera- for neutron-induced concrete deterioration in the CBS.
ture and humidity, and water amount at 98% RH is de-
termined by the temperature. This model was based on 5. Discussion
our experimental results of the colloidal behavior of
calcium silicate hydrates (Maruyama et al. 2014a). Un- 5.1 Limits of DEVICE
der the heating and drying conditions, the concrete sorp- While the DEVICE takes into account many aspects of
tion amount at the inner surface decreased with increas- concrete phenomena, there are limitations. Moisture
ing temperature, and water vapor was accumulated in the transport in DEVICE is validated by the experiments
pore space, which then tended to attain the equilibrium (Maruyama and Igarashi 2015), but long-term drying
condition. This process is strongly affected by the ele- process in massive concrete under the relatively high
vated temperature rather than the relative humidity. After temperature condition needs more validation. Because
the drying process, concrete strength development was data are very scarce, we need more scientific data of the
stagnant immediately after the operation began. In addi- moisture transport in massive concrete under the high
tion to this stagnation of compressive strength devel- temperature conditions as well as the temperature gra-
opment, deterioration due to neutron irradiation was dient conditions.
confirmed after several years of operation. Reduction of concrete due to neutron irradiation is
The sealing condition at the inner surface of the CBS modeled by a statistical way, but the background data
will contribute to attain the strength margin because of have a large variation in the relationship between ex-
continuous cement hydration through conservation of pansion strain of concrete and Fc/Fco. This variation is
I. Maruyama, K. Haba, O. Sato, S. Ishikawa, O. Kontani and M. Takizawa / Journal of Advanced Concrete Technology Vol. 14, 144-162, 2016 157
Fig. 15 Comparison of calculation results for 337s and 337o for 60-year operation.
Fig. 16 Comparison of calculation results for 337s and 326s. (a) (top) strength distribution and (bottom) temperature dis-
tribution of 337s. (b) (top) strength distribution and (bottom) temperature distribution of 326s. The elevated temperature at
the inner surface strongly affects the distribution of evaporable water and resultant strength development.
I. Maruyama, K. Haba, O. Sato, S. Ishikawa, O. Kontani and M. Takizawa / Journal of Advanced Concrete Technology Vol. 14, 144-162, 2016 158
larger than that of loading experiment for compressive Based on previous accelerated irradiation experiments
strength, therefore, there should be a mechanism to ex- of concrete and mortar, the results were numerically
plain this variation. Concrete’s physical property changes extrapolated to the problem in a CBS by using DEVICE.
affected by aggregate expansion depend on the crack The calculation results indicate that the depth of concrete
propagation in it. This implies that the deterioration deterioration is approximately 100 mm from the inner
might be anisotropic due to concrete stress caused by surface of the CBS after 60 years of operation of the
dead load. This is not considered in DEVICE. In addition, reactor. We discuss how to evaluate this result consider-
a combination of dead load and aggregate expansion may ing the limitations of our knowledge.
cause compaction of mortar which might increase the An argument about the aggregate type or the mineral
strength by closing the crack width and altering C–S–H composition in the aggregate is addressed here. In the
by the compaction effect (Sereda et al. 1966). The dif- present calculation, the aggregate is considered com-
ference in strength from the lower boundary curve shown posed of 100% α-quartz. If this amount is half, the
in Fig. 5 can possibly be explained via compaction of volume of the elements that can expand in the concrete
mortar or cement paste in a capsule that is used in irra- clearly becomes half and thus the damage in concrete and
diation experiments along with the irradiation-induced the degree of resultant deterioration should be decreased.
aggregate expansion. The ionic minerals, such as calcite, probably have high
The temperature and duration of irradiation highly resistivity to irradiation, and concrete with pure lime-
affect α-quartz expansion, as shown in Fig. 3. A ther- stone may show favorable performance under irradiated
modynamic healing impact of silicate crystal structures conditions. Even though we do not have complete data-
exists, which is a function of temperature and time. If the base for the volume stability of aggregate minerals under
temperature is elevated, the healing process is enhanced, irradiated conditions, we can conclude that the aggregate
and consequently, the expansion rate under irradiation type and its mineral composition play an important role
conditions is low. This implies that the accelerated ex- in irradiation-induced concrete degradation. Therefore, it
periment result might underestimate the healing process is possible to rebuttal with a viewpoint of mineral com-
of silicate structure, because the realistic time might be position in the used aggregate.
more than 10 times the experimental time span. There- Secondary integrity evaluation is possible after the
fore, defect accumulation and consequent macroscopic fluence limit is evaluated. First, the aggregate type and
volume change must be scientifically studied and the the mineral composition must be specified, and then, the
healing process must be independently understood for aggregate expansion strain needs to be evaluated by
achieving a comprehensive understanding of irradia- considering the temperature at irradiation condition,
tion-induced expansion of α-quartz. In addition, a more which requires data on the resistivity of rock minerals.
accurate estimation may be obtained by considering this The allowable fast neutron fluence can be then proposed
healing impact. In addition, impacts of other aggregate according to the mineral composition of the aggregate.
minerals also should be considered, but the current code Alternatively, the strength can be evaluated, and an al-
can not take into account due to lack of data. lowed reduction of concrete strength can be determined
such that it satisfies the required design strength. The
5.2 Integrity evaluation lower boundary curve shown in Fig. 5 and Eq (9) give
The CBS that is exposed to radiation develops a risk of
irradiation-induced concrete deterioration. The expected
degradation can be categorized into two phenomena:
large expansion (Le Pape 2015) and mechanical property
deterioration. Undoubtedly, these two phenomena are
strongly interconnected through stress and resultant
cracks in the concrete of the CBS; however, to the best of
author’s knowledge, there is no experimental report
about reinforced concrete member affected by irradiation
induced concrete deterioration. As shown in the intro-
duction, the integrity of a CBS is normally evaluated by
determining the limit of fast neutron fluence and/or
gamma-ray dose, which depend on the authorities or
countries. Previous studies have suggested that the con-
crete strength deteriorates above a fast neutron (>0.1 Fig. 17 Impact of water content on the fast neutron flu-
MeV) fluence of 1019 n/cm2 under the temperature con- ence after 60-year operation. Water content distribution
dition of light water reactors (Hilsdorf et al. 1978; Field and the ratio of fast neutron fluence after 60-year opera-
et al. 2015), while the minimum neutron energy that tion are considered. Water content distribution changes
affects α-quartz is 0.01 MeV (Primak 1958; Weissmann during operation and the neutron fluence are affected by
and Nakajima 1963; Bykov et al. 1981). Further scien- this water content distribution. The data are accumulated
tific research is needed to confirm this criterion. results from the onset of operation to the final state of the
calculation.
I. Maruyama, K. Haba, O. Sato, S. Ishikawa, O. Kontani and M. Takizawa / Journal of Advanced Concrete Technology Vol. 14, 144-162, 2016 159
the worst-case value of Fc/Fco. Therefore, based on the can produce the strength margin owing to continuous
DEVICE calculation with a calibration of the obtained cement hydration.
data, we can evaluate the strength at the inner surface as a 2) The temperature difference at the inner surface of the
safe-side value, which can be compared with the required CBS has large impact on irradiation-induced concrete
compressive strength, and thus we can judge the integrity degradation through temperature dependence of irra-
of a CBS through material-basis evaluation. diation-induced expansion of α-quartz.
The last option is a more comprehensive approach. 3) However, the difference in compressive strength due
The integrity of a CBS with regards to the seismic per- to temperature-dependent irradiation-induced con-
formance can be evaluated by omitting a part of a section crete degradation is almost the same order as that by
in the CBS by estimating it to be deteriorated by the the strength margin built by the sealing condition.
gamma-ray dose. In Japan, although a gamma-ray limit 4) While the volume stability of aggregate minerals,
of 2.0 × 105 kGy exists, some PWRs have experienced minimum neutron flux energy for the damage in
more than this limit. An electrical power company has α-quartz, healing effect of the silicate structure as a
supported this condition using a calculation that consid- function of temperature, and degradation process un-
ered the deteriorated part (The Nuclear Regulation Au- der stress in concrete are limitations of our current
thority, Japan 2015). This type of approach can be used evaluation method, and they need more scientific
for evaluating the integrity of a CBS. The impact of background.
concrete expansion, which may cause spalling on the In addition, possible integrity evaluation methods
surface and deterioration of bond in the rebars, should be other than allowable fast neutron fluence evaluation for a
considered for seismic performance evaluation. biological shielding wall are proposed as follows.
The validation relating to important phenomena such 1) Mineral composition-based allowable fast neutron
as healing process of α-quartz, stress-related physical fluence.
property change, bond performance affected by volume 2) Strength prediction at the inner surface on the basis of
expansion of concrete, and irradiation-induced deterio- mineral composition of aggregate and the lower limit
ration of concrete can be carried out only on decommis- curve of Fc/Fco as a function of concrete expansion.
sioned plants. Therefore, such research projects using a 3) Direct numerical calculations for the seismic per-
decommissioned plant are necessary for long-term op- formance by considering irradiation-induced volume
eration of aging plants. expansion and concrete degradation.
6. Conclusion Acknowledgement
This research was supported by the Nuclear Regula-
A numerical code called “DEVICE” (Damage EValua- tion Authority (NRA, Japan), and we thank the NRA for
tion for Irradiated ConcretE), which takes into account their sponsorship. The information presented in this
heat, moisture, and radiation transport coupled with paper is the sole opinion of the authors and does not
cement hydration, is proposed. This code is based on the necessarily reflect the views of the sponsoring agencies.
computational cement-based material model (CCBM)
and the one-dimensional deterministic transport Sn code References
ANISN. First, DEVICE was used for evaluating the American Concrete Institute, (2012). “Code
decommissioned plant JPDR. The compressive strength requirements for nuclear safety related concrete
distribution in a biological shielding wall of the JPDR is structures.” ACI 349 - 12.
obtained by core sampling. The compressive loading test Avrami, M., (1941). “Granulation, phase change, and
results were compared with the calculation results, which microstructure kinetics of phase change. III.” The
presented the potential of DEVICE to predict concrete Journal of Chemical Physics, 92, 177-184.
strength distribution in a concrete biological shielding Bogue, R. H., (1929). “Calculation of the compounds in
(CBS) wall. The fast neutron fluence level of the CBS Portland Cement.” Industrial & Engineering Chemis-
under which significant impact on concrete strength is try Analytical Edition, 14, 192-197.
not observed is 1.0 × 1018 n/cm2 at room temperature Bolse, W., (1999). “Amorphization and recrystallization
(<40 °C). The comparison results also showed that the of covalent tetrahedral networks.” Nuclear Instru-
sealing conditions at the inner and outer surfaces of the ments and Methods in Physics Research, Section B:
CBS by a steel liner plate and resin coating, respectively, Beam Interactions with Materials and Atoms 148,
have a homogenizing effect on the concrete strength of 83-92.
the CBS. Bonnet, J. P., Boissier, M. and Ait Gherbi, A., (1994).
In addition, simulation results for the concrete strength “The amorphization process of neutron-irradiated
change and its distribution in a biological shielding wall of crystalline quartz studied by Brillouin scattering.”
an anonymous two-loop pressurized water reactor up to 60 Journal of Non-Crystalline Solids, 167, 199-204.
years’ operation indicated the following conclusions: Bouniol, P. and Aspart, A., (1998). “Disappearance of
1) A steel liner plate and resin coating, which restrain oxygen in concrete under irradiation: the role of
water vapor evaporation from the surface of a CBS, peroxides in radiolysis.” Cement and Concrete Re-
I. Maruyama, K. Haba, O. Sato, S. Ishikawa, O. Kontani and M. Takizawa / Journal of Advanced Concrete Technology Vol. 14, 144-162, 2016 160
Ichikawa, T. and Kimura T., (2007). “Effect of nuclear Institute 28(1), 1691-1696.
radiation on alkali-silica reaction of concrete.” Journal Le Pape, Y., (2015). “Structural effects of
of Nuclear Science and Technology, 44(10), radiation-induced volumetric expansion on unrein-
1281-1284. forced concrete biological shields.” Nuclear
Idei, Y., Kamata, H., Akutsu, Y., Onizawa, M., Nakajima, Engineering and Design, 295, 534-548.
N., Sukegawa T. and Kakizaki, M., (1990). “Strength Le Pape, Y., Field, K. G., and Remec, I., (2015).
of Biological Shield Concrete of JPDR.” Technical “Radiation effects in concrete for nuclear power plants,
Report JAERI-M 90-205. Part II: Perspective from micromechanical modeling.”
Igarashi, G., Maruyama, I., Nishioka Y. and Yoshida, H., Nuclear Engineering and Design, 282, 144-157.
(2015). “Influence of mineral composition of siliceous Litvan, G. G., (1976). “Variability of the nitrogen surface
rock on its volume change.” Construction and area of hydrated cement paste.” Cement and Concrete
Building Materials, 94, 701-709. Research, 6(1), 139-143.
Inui, H., Mori, H., Sakata T. and Fujita, H., (1990). Lowinska-Kluge, A. and Piszora, P., (2008). “Effect of
“Electron irradiation induced crystalline-to-amor- gamma irradiation on cement composites observed
phous transition in quartz single crystals.” Journal of with XRD and SEM methods in the range of radiation
Non-Crystalline Solids, 116(1), 1-15. dose 0-1409 MGy.” Acta Physica Polonica-Series A
Itoh, M. and Maruyama, I., (2014). “Evaluation of General Physics, 114(2), 399.
compressive strength of concrete expanded by Maruyama, I. and Igarashi, G., (2015). “Numerical
alkali-silica reaction.” Summaries of the annual CAJ approach towards aging management of concrete
meeting of cement and concrete engineering, 68, structures: Material strength evaluation in a massive
250-251. (in Japanese) concrete structure under one-sided heating.” Journal
Iwata, S., (1981). “Transition of aggregate demand and of Advanced Concrete Technology 13(11), 500-527.
supply and current status of survey of natural Maruyama, I., Igarashi, G. and Nishioka, Y., (2015).
resources.” Cement & Concrete, 415(9), 6-13. (in “Bimodal behavior of C-S-H interpreted from
Japanese) short-term length change and water vapor sorption
Jennings, H. M., (2008). “Refinements to colloid model isotherms of hardened cement paste.” Cement and
of C-S-H in cement: CM-II.” Cement and Concrete Concrete Research, 73(7), 158-168.
Research, 38(3), 275-289. Maruyama, I., Kontani, O., Ishizawa, A., Takizawa, M.
Kakizaki, M., Idei, Y., Sukegawa, T., Akutsu, Y., Hatano and Sato O., (2012). “Development of system for
H. and Kurioka, H., (1996). “Study on environmental evaluating concrete strength deterioration due to
and mechanical properties of irradiated concrete.” radiation and resultant heat.” In: Proceedings of 3rd
Journal of structural and construction engineering, International Conference on NPP Life Management
488, 1-10. (in Japanese) for Long Term Operations, IAEA-CN-194-096, Salt
Kishitani, K. and Baba, A., (1975). “An essay on the Lake City, May.
composite mechanisms of drying shrinkage of Maruyama, I., Kontani, O., Sawada, S., Sato, O., Igarashi
concrete.” Transactions of the Architectural Institute G. and Takizawa M., (2013). “Evaluation of irradiation
of Japan, 234, 1-11. (in Japanese) effects on concrete structure-background and prepara-
Kontani, O., Y. Ichikawa, A. Ishizawa, M. Takizawa and tion of neutron irradiation test.” In: POWER2013,
O. Sato (2010). “Irradiation effects on concrete American Society of Mechanical Engineers, Power
structure.” In: Int. Symp. On the Ageing Manag. of Division.
Nucl. P. P., 173-182. Maruyama, I., Nishioka, Y., Igarashi, G. and Matsui, K.,
Kontani, O., Ichikawa, Y., Ishizawa, A., Takizawa M. (2014a). “Microstructural and bulk property changes
and Sato, O., (2013). “Irradiation effects on concrete in hardened cement paste during the first drying
structures.” Infrastructure Systems for Nuclear Energy, process.” Cement and Concrete Research, 58(4),
459-473. 20-34.
Kontani, O., Ishizawa, A., Maruyama, I., Takizawa M. Maruyama, I., Sasano, H., Nishioka Y. and Igarashi, G.,
and Sato, O., (2012). “Evaluation of irradiation effects (2014b). “Strength and Young's modulus change in
on concrete structure.” In: Proceedings of the 2012 concrete due to long-term drying and heating up to
International Congress on Advances in Nuclear Power 90 °C.”Cement and Concrete Research, 66(4), 48-63.
Plants, Chicago, Maruyama, I. and Sugie, A., (2014). “Numerical study
Kontani, O., Sawada, S., Maruyama, I., Takizawa, M. on drying dhrinkage of concrete affected by aggregate
and Sato, O., (2013). “Evaluation of irradiation effects size.” Journal of Advanced Concrete Technology,
on concrete structure - Gamma ray irradiation tests on 12(8), 279-288.
cement paste.” In: Proceedings of the ASME 2013 Muller, A. C. A., Scrivener, K. L., Gajewicz, A. M. and
Power Conference POWER2013. McDonald, P. J., (2012). “Densification of C–S–H
Kubo, Y., Ueda, T., Kuroda, T. and Nomura, M., (2006). measured by 1H NMR relaxometry.” The Journal of
“Influence of ASR expansion on mechanical proper- Physical Chemistry C, 117(1), 403-412.
ties of concrete.” Proceedings of the Japan Concrete Nakano, K., Kobayashi, S. and Nakagami, A., (1987).
I. Maruyama, K. Haba, O. Sato, S. Ishikawa, O. Kontani and M. Takizawa / Journal of Advanced Concrete Technology Vol. 14, 144-162, 2016 162
“Degree of damage due to alkari-silica reaction and Research and Development Effectiveness in the Light
physical properties of concrete.” Cement Association of the Accident at the Fukushima daiichi Nuclear
Japan (CAJ) Review of the 41st General Meeting / Power Plant, Vienna, Austria, International Atomic
Technical Session 41, 419-422. Energy Agency.
Parrott, L. J., Hansen, W. and Berger, R. L., (1980). Sereda, P. J., Feldman, R. F. and Swenson, E. G., (1966).
“Effect of first drying upon the pore structure of “Effect of sorbed water on some mechanical properties
hydrated alite paste.” Cement and Concrete Research, of hydrated portland cement pastes and compacts.”
10(5), 647-655. Highway Research Board Special Report 90, 58-73.
Parrott, L. J. and Young, J. F., (1981). “Effect of Son, H. N. and Hosoda A., (2010). “Detection of
prolonged drying upon the silicate structure of microcracking in concrete subjected to elevated
hydrated alite pastes.” Cement and Concrete Research, temperature at very early age by acoustic emission.”
11(1), 11-17. Journal of Advanced Concrete Technology, 8(2)
Pedersen, A., (1971). “Radiation damage in concrete - 201-211.
Measurements on miniature specimens of cement Sugiyama, H. and Masuda, Y., (1999). “Strength
mortar.” In: Proceedings of an Information Exchange development of concrete cured under high temperature
Meeting on 'Results of Concrete Irradiation conditions in an early age.” Journal of structural and
Programmes'. construction engineering, 515, 23-30. (in Japanese)
Pickett, G., (1956). “Effect of aggregate on shrinkage of Swamy, R. N. and Al-Asali, M. M., (1988). “Engineering
concrete and a hypothesis concerning shrinkage.” ACI properties of concrete affected by alkali-silica reac-
Journal proceedings, 52(1), 581-590. tion.” Materials Journal, 85(5), 367-374.
Pihlajavaara, S. E., (1974). “A review of some of the The Nuclear Regulation Authority Japan, (2015).
main results of a research on the ageing phenomena of “Approval of operational safety programs for Kyushu
concrete: Effect of moisture conditions on strength, electric power company's Sendai nuclear power
shrinkage and creep of mature concrete.” Cement and station units 1 and 2.” Tokyo, NRA.
Concrete Research, 4 (5), 761-771. Thomas, J. J., Allen, A. J. and Jennings, H. M., (2008).
Pomaro, B., Salomoni, V. A., Gramegna, F., Prete, G. and “Structural changes to the calcium–silicate–hydrate
Majorana, C. E., (2011). “Radiation damage evalua- gel phase of hydrated cement with age, drying, and
tion on concrete within a facility for Selective resaturation.” Journal of the American Ceramic
Production of Exotic Species (SPES Project), Italy.” Society, 91(10), 3362-3369.
Journal of Hazardous Materials, 194, 169-177. Thomas, J. J. and Jennings, H. M., (2003). “Changes in
Primak, W., (1958). “Fast-neutron-induced changes in the size of pores during shrinkage (or expansion) of
quartz and vitreous silica.” Physical Review, 110(6), cement paste and concrete.” Cement and Concrete
1240-1254. Research, 33(11), 1897-1900.
Primak, W., Fuchs, L. and Day, P., (1955). “Effects of Tomita, Y., Kousa, K., Nakano K. and Nakagami, A.,
nuclear reactor exposure on some properties of (1988). “Diagnosis method using core sampling for
vitreous silica and quartz.” Journal of the American concrete structure suffering from ASR.” Review of the
Ceramic Society, 38(4), 135-139. 42st General Meeting / Technical Session, Cement
Rhoades, W. A., (1978). “The GIP program for Association Japan (CAJ), 42, 335-338. (in Japanese)
preparation of group-organized cross-section librar- Ueda, N., Nakamura, H., Kunieda, M., Maeno, H.,
ies.” Informal notes (RSIC CODE PACKAGE Morishit, N. and Asai, H., (2011). “Evaluations on
PSR-229 MICRO). ASR damage of concrete structure and its structural
Salomoni, V. A., Majorana, C. E., Pomaro, B., Xotta G. performance.” Doboku Gakkai Ronbunshuu E 67 1
and Gramegna, F., (2014). “Macroscale and mesoscale 28-47. (in Japanese)
analysis of concrete as a multiphase material for Vodák, F., Trtík, K., Sopko, V., Kapičková, O. and Demo,
biological shields against nuclear radiation.” Interna- P., (2005). “Effect of γ-irradiation on strength of
tional Journal for Numerical and Analytical Methods concrete for nuclear-safety structures.” Cement and
in Geomechanics, 38(5), 518-535. Concrete Research, 35(7), 1447-1451.
Seeberger, J. and Hilsdorf, H., (1982). “Einfluß von Weber, W. J., (1990). “Radiation-induced defects and
Radioactiver Strahlung auf die Festogkeit and amorphization in zircon.” Journal of Materials
Struktur von Beton.” Technical Report NR 2505, Research, 5(11), 2687-2697.
Institut für Massivbau and Baustofftechnologie, Weissmann, S. and Nakajima, K., (1963). “Defect
Universität Karlsruhe. Structure and density decrease in neutron‐irradiated
Sekimura, N., (2015). “Roadmap on R&D and Human quartz.” Journal of Applied Physics ,34(3), 611-618.
Resource for Light Water Reactors Safety in Japan, Wittels, M. and Sherrill, F. A., (1954). “Radiation
-Nuclear Safety Visions and Technical basis damage in SiO2 structures.” Physical Review, 93(5),
Reconstruction after the Fukushima Accident -.” In: 1117-1118.
International Experts Meeting on Strengthening
Journal of Advanced Concrete Technology Vol. 14, 706-716, November 2016 / Copyright © 2016 Japan Concrete Institute 706
Scientific paper
Abstract
We investigated changes in the density of natural rock minerals following high-energy electron irradiation, using the
plasmon peak shift of electron energy-loss spectra and transmission electron microscopy. The target materials were the
natural rock minerals α-quartz, orthoclase, anorthite, albite, biotite, muscovite, and chlorite. These crystalline minerals
can be classified into three groups based on their Si-network geometries: 3-dimensional 6-member ring; 4-member ring +
6-member ring; and planar 6-member ring. The metamictization rates and changes in relative density are discussed using
a phenomenological model, which we used to identify the physical parameters that describe the metamictization process
as a function of the volume density of Si and Al atoms, or Si atoms alone, in the crystal structures. The relative densities
following metamictization all decreased by more than a few percent, except for albite, which became denser. These results
suggest that radiolysis damage causes initial compaction, then metamictization, characterized by the expansion of the Si-
and Al-polyhedra in the aggregate. The stability of concrete containing α-quartz, orthoclase, and anorthite should be
further investigated in the light of the present results.
Table 1 Chemical composition, origin, density, molar mass, and silicate structure of each natural rock mineral analyzed.
Chemical composition Density Molar mass Silicate (includ-
Mineral Origin
(General form) (g/cm2) (g/mol) ing Al) structure
3-dimensional
α-quartz SiO2 Nellore, Andhra Pradesh, India 2.65 60.08
6-ring
Orthoclase KAlSi3O8 Nellore, Andhra Pradesh, India
2.56 278.33 4 ring + 6 ring
2.73
Anorthite CaAl2Si2O8 Yoichi, Hokkaido, Japan 277.41 4 ring + 6 ring
(2.72 - 2.75)
Governador Valadares, Minas, Gerais, 2.62
Albite NaAlSi3O8 263.02 4 ring + 6 ring
Brazil (2.61 - 2.63)
3.09
Biotite K(Mg,Fe)3(AlSi3O10)(F,OH)2 Nellore, Andhra Pradesh, India 433.53 Planate 6 ring
(2.8 - 3.4)
2.82
Muscovite KAl2(AlSi3O10)(F,OH)2 Ishikawayama, Ishikawa, Ishiwaka, Japan 398.71 Planate 6 ring
(2.77 - 2.88)
2.65
Chlorite (Mg,Fe)5Al(Si3Al)O10(OH)8 Miwa, Nagasaki, Japan 595.22 Planate 6 ring
(2.55 - 2.75)
α-quartz; orthoclase; anorthite; albite; biotite; musco- tional to its atomic density, and the relative volume
vite; and chlorite (Table 1). All samples were supplied change can be estimated from the displacement of the
by Nichika Co., Japan. The powder X-ray diffraction plasmon peak. Drude model was originally developed
patterns of the samples confirmed that each sample con- by assuming the free-electron gas model for the valence
sisted of a single phase (Fig. 1).
2.2 Method
Pristine samples were ground into powders with parti-
cles < 25μm in diameter, using a planetary ball mill. The
ground powder was sprinkled onto a carbon microgrid
supported by a copper mesh disc, which was mounted to
a transmission electron microscope (TEM) sample
holder. The holder was inserted into a Jeol JEM2100
TEM equipped with a Gatan ENFINA 1000 electron
energy-loss spectrometer (EELS). The TEM was oper-
ated at 100 kV, mostly below the threshold for knock-on
damage to these minerals. Electron flux was measured
using an electron current meter placed in the TEM fluo-
rescent screen.
Changes in the low-loss spectrum of α-quartz with Fig. 1 X-ray diffraction data for the rock minerals ana-
irradiation time are shown in Fig. 2. The strongest peaks lyzed.
at zero energy loss are the zero-loss peak, and the
broader peaks at ~22-25 eV correspond to the plas-
mon-loss peaks. During electron irradiation, amorphiza-
tion was confirmed by the occurrence of halo rings in
the electron diffraction pattern, as shown for α-quartz in
Fig. 3.
Plasmon energy Ep is expressed by Eq. (1), based on
the Drude model (Raether 1982, Egerton 2009):
ne 2 N
Ep = = ∝ , (1)
ε0m V
electrons in metals, though it has been well established Table 2 Change in density of rock minerals and the
that the model is applicable to materials having covalent electron dose required for metamictization.
and/or ionic bonds when the plasmon peak energy is Minerals Change in relative density Electron dose
significantly larger than the band gap energy (Egerton
α-quartz −4% 6 × 1020 e/cm2
2011) and Eq. (1) provides reasonable agreement with
experimental values. It is thus considered that the pre- Orthoclase −2% 2 × 1021 e/cm2
sent analyses can be effective unless a significant com- Anorthite < −3% >2 × 1021 e/cm2
positional change or appreciable phase separation is Albite +4% 8 × 1021 e/cm2
observed during the irradiation. Biotite −8%. 1 × 1022 e/cm2
Muscovite −4% 1 × 1023 e/cm2
3. Results Chlorite < −6% >7 × 1023 e/cm2
Fig. 3 TEM images and electron diffraction images of α-quartz before irradiation and after amorphization, under 100 KeV
electron irradiation.
I. Maruyama and S. Muto / Journal of Advanced Concrete Technology Vol. 14, 706-716, 2016 709
(g) Chlorite
Fig. 4 Change in relative density of rock minerals as a function of electron dose at 100 KeV. Inset images show the electron
differentiation patterns used to identify the completion of metamictization.
I. Maruyama and S. Muto / Journal of Advanced Concrete Technology Vol. 14, 706-716, 2016 710
Table 3 Derived parameters for equations (2), (3), and (4), based on the metamictization process.
Minerals b σ1 σ2 R2 R3
α-quartz 1.4E−02 5.30E−21 1.20E−20 1.01E+00 9.55E−01
Orthoclase 1.5E−04 5.00E−20 6.00E−22 1.01E+00 9.60E−01
Anorthite 5.0E−04 1.80E−21 4.00E−23 1.04E+00 9.60E−01
Albite 1.0E−02 7.00E−20 7.00E−21 9.00E−01 1.05E+00
Biotite 2.3E−07 1.50E−23 1.50E−22 1.01E+00 9.10E−01
Muscovite 1.3E−04 1.50E−21 5.00E−22 9.70E−01 9.55E−01
Chlorite 7.5E−06 1.00E−22 6.00E−24 1.04E+00 9.40E−01
structures. The metamictization process can be ex- thus be used as a measure of damage tolerance. All the
pressed using the following reaction rate equations: data points except that for α-quartz exhibit a linear rela-
tionship. Muscovite, chlorite, and biotite consist of a
⎧ dC1 planar 6-ring Si structure (as does Al to some extent and
⎪ dt = −φσ 1C1 + bC2 in some cases), which is thought to be more tolerant of
⎪
⎪ dC2 irradiation than a 6-member ring and 4-member ring
⎨ = φσ 1C1 − φσ 2 C2 − bC2 (4) structure or a 3-dimensional 6-ring structure. Biotite is
⎪ dt characterized by 6-member Si ring wedges and K+ and
⎪ dC3
⎪ dt = φσ 2 C2 Mg/Al-O layers interpenetrating each other. Chlorite, on
⎩ the other hand, has a 6-member Si ring wedge between
a Mg/Al-O layer and interlayer Mg2+/Al2+ cations, and
where φ is electron flux (e/cm2), and σ 1 and σ 2 are muscovite includes a 6-member Si ring wedge between
the effective cross sections for damage production from interlayer K+ cations and Al-O layers. Of these three
the initial crystalline phase to the intermediate phase, minerals, muscovite shows the largest cross section σ 1 ,
and from the intermediate phase to the metamict phase, probably because of the larger number of Al-O-Al,
respectively. We assume that the intermediate phase can Al-O-Si, and Si-O-Si networks it contains.
be thermally reversed to the initial phase, whereas the Albite, orthoclase, and anorthite have 4-member Si
metamict phase is irreversible once reached. This as- ring and 6-member Si ring structures, with larger cations
sumption is drawn from the facts that (neutron-) in the channels between the networks, and have larger
metamicted structure showed consistent volume change σ 1 values than the minerals with planar 6-member ring
regardless temperature during irradiation (Bykov et al. structures. Among these three minerals, albite has the
1981). It should be noted that this model is quite simpli- largest σ 1 , which can be explained by the Si-O-Si an-
fied but this has been successfully applied to predict the gles linking the Si-polyhedra being more flexible due to
physical property of electron-irradiated materials the small cations (Na+) in the spacious channels of the
(Nakano et al. 2005). The set of equations (4) can be Si-networks. In this context, α-quartz is an outlier (Fig.
solved analytically, and the experimental data were fit to 7(a)), probably because it has a greater capacity for the
the solutions using least-squares (solid lines in Fig. 4) to Si-O-Si angles to change, since it has no other cations in
yield the parameters σ 1 , σ 2 , b, R2, and R3. (Table 3). the channels to resist local deformation.
The thermal recovery parameter b can take arbitrary The relationship of the thermal recovery parameter, b,
small values for orthoclase and biotite, where the cross relative to the density of Si and Al atoms is similarly
sections for damage formation, σ 1 and σ 2 , are not linear, and α-quartz is not as much of an outlier (Fig.
significantly different from each other. This suggests 7(b)). We can thus conclude that both σ 1 and b are
that metamictization proceeds according to a single ra- governed by the flexibility inherent in their Si-O-Si,
diolysis process in those minerals. On the other hand, Si-O-Al, and Al-O-Al networks and the location of
the other minerals exhibited considerable thermal re- other cations in these networks. Larger cations probably
covery and different values for σ 1 and σ 2 , which play a role in restricting the change in angle possible in
suggests that metamictization follows a two-step proc- these networks.
ess in those minerals. The derived cross section values The value of σ 2 , the effective cross section of trans-
are also reasonable for radiolysis (Ugurlu et al. 2011). formation from the intermediate phase to the metamict
These results are obviously supporting that the applied phase, is thought to be related to the collapse of Si-O-Si
eq. (4) is good enough as the first approximation of the bonds. Plotting σ 2 as a function of the density of Si
observed phenomenon and not far from the real phe- atoms (Fig. 7(c)) yielded a strong linear relationship,
nomenon. although this parameter did not show a good correlation
The σ 1 s values for the seven minerals are plotted as with the density of Si and Al atoms. We therefore con-
a function of the density of Si and Al atoms in Fig. 7(a), clude that the final metamictization process is mainly
since the bonding of the oxygen atom in Si-O-Si, attributable to the collapse of Si-O-Si bonds. However,
Si-O-Al, or Al-O-Al is likely the most susceptible point α-quartz was an outlier in this respect as well, and ex-
for irradiation damage, and the density of Si and Al can hibited different tendencies than the other minerals.
I. Maruyama and S. Muto / Journal of Advanced Concrete Technology Vol. 14, 706-716, 2016 712
The relative density of the final metamictized phase rings tended to show significant expansion despite hav-
was closely related to the extent of volume expansion ing a higher tolerance for electron irradiation. We note
after metamictization, which is crucial for concrete per- that among the minerals that have similar Si and Al
formance under radiolysis impact (such as gamma-ray atom densities, anorthite and orthoclase exhibited an
irradiation). When plotting the corresponding parameter, almost identical reduction in density, whereas albite
R2, as a function of the density of Al and Si atoms, all underwent a substantial increase in density. This distinct
tested minerals except albite appear to lie on a curve difference is probably explained by the size of the
(Fig. 8). The structures containing planar 6-member cations in the spacious channels of the Si (or Al) net-
works. In albite, Na+ ions are small enough to allow the
Si-O-Si units to shrink, and stabilize the structure fol-
lowing shrinkage. If there were no cations within the
Si-network, the SiOx polyhedra would expand at
metamictization and then be stable, as in the case of
α-quartz.
It has been already suggested that understanding the
metamictization process due to neutron irradiation of
α-quartz should take into account both the radiolysis
impact and the ballistic impact. And both of them is not
clarified for other rock minerals. Even in the case of
neutron irradiation experiment, we always have an im-
pact of gamma-rays (radiolysis) from reactor or secon-
dary effect of neutron. Therefore, better understanding
of radiolysis impact on rock forming minerals is impor-
tant for volume expansion under neutron and
(a) σ 1 (cm2/e) gamma-ray irradiation. From a such view point, the
present study is significant. And our findings are that
nominal sections for 2 different phases under metamic-
tization and thermal recovery/annealing parameter are
determined based on simplified phenomenalogical
equations and the obtained nominal section and thermal
recovery parameters are discussed through the indices
obtained by crystal structures. As a results, the density
of Si and Al atoms are key parameters for those.
The final density change of radiolysis impact (4% for
α-quartz) is far from the value of neutron-irradiated
samples (15% for α-quartz). This supports that ballistic
impact has a large impact on the final density change of
metamictized condition and radiolysis impact is less,
while the incubation process, in which thermal anneal-
ing has a large impact, should take into account the ra-
(b) b (1/s) diolysis impact.
(c) σ 2 (cm2/e)
Fig. 7 Relationships between parameters for rate of
metamictization by electron irradiation and the index Fig. 8 Relative density of the metamictization phase of the
obtained for the crystal structure. seven minerals tested.
I. Maruyama and S. Muto / Journal of Advanced Concrete Technology Vol. 14, 706-716, 2016 713
Appendix
The crystal structure of used minerals are summarized
here. The crystal structure of used minerals are summa-
rized here. The figures were drawn with VISTA
(Momma and Izumi 2011).
(b) Orthoclase (4-ring + 6-ring) (Viswanathan and Kielhorn (d) Albite (4 ring + 6 ring) (Harlow 1982)
1983)
(c) Anorthite (4 ring + 6 ring) (Wenk et al. 1980) (e) Biotite (Planate 6-ring) (Takeda and Ross 1975)
I. Maruyama and S. Muto / Journal of Advanced Concrete Technology Vol. 14, 706-716, 2016 716
(f) Muscovite (Planate 6-ring) (Richardson and Richard- (g) Chlorite (Planate 6-ring) (Zanazzi et al. 2007)
son 1982)
Fig. A. 1 Structure of each mineral. Blue tetrahedral represent the Si or Al atom.
Journal of Advanced Concrete Technology Vol. 14, 691-705, November 2016 / Copyright © 2016 Japan Concrete Institute 691
Invited paper
Abstract
Based on a review from a view point of multi-scale perspective, it is concluded that possible mechanisms of natural
frequency change of concrete structures under an ordinary drying condition are reduction of Young’s modulus of concrete
and stiffness reduction of reinforced concrete member due to shrinkage-induced cracking. Both two mechanisms can be
provoked by drying, which is natural environmental condition for all the building structures. This natural frequency
change is important for the integrity analysis of nuclear power plants because seismic response change and sympathetic
vibration between structure and installed implementations are expected. In addition, to understand the mechanism this
mechanism should contributed to post-earthquake integrity evaluation which is necessary to efficient and continuous
operation of nuclear power plants. The lost parts of scientific knowledge and needs of experimental data for whole picture
of this natural frequency change have been pointed out.
tributable to the soil-structure interaction problem solely of stiffness of reinforced concrete member, and two
unlike the cases found in European buildings. mechanisms for reduction of stiffness of reinforced
In case of nuclear power plant, Unit No.1 of Onagawa concrete member are derived; reduction of Young’s
Nuclear Power Plant has been analyzed since 1985, and modulus of concrete and cracking in reinforced concrete
they found the continuous reduction of natural frequency member. Now, the common environmental condition,
of the plant and it was 22% after 25 years of construction. drying, is discussed to emphasize the importance of
This means that even the RC members are thick enough, scientific research for prediction of performance of
it has a natural trend to decrease the stiffness of the concrete structures to contribute to aging management of
members (Nuclear and Industrial Safety Agency 2011; concrete structures. The whole structure of the manu-
Ogata et al. 2011). In addition to this phenomenon, there script is shown in Table 1.
were some cases that the natural frequency showed
sudden drop during or after the earthquakes but there is 3. Concrete components and concrete
some recovering mechanism and it turned back to the
natural decreasing trend. Similar phenomenon have been 3.1 Cement paste properties under the first
reported by Saito (2012). In case of very large size of drying
earthquake, it is possible that this recovering mechanism Hardened cement paste (hcp) made from cement paste
can not bring the frequency back to the natural trends with relatively high water cement ratio (W/C = ~0.55)
(Takahashi et al. 2012; Uebayashi and Nagano 2012). has notable colloidal nature. This nature is attributed to
Summary of natural frequency of concrete related calcium silicate hydrate (C-S-H) with low crystallinity
buildings are shown in Fig. 1. which is hydration product of alite and belite. There are
Based on these previous researches, it is concluded as many alibi of the colloidal nature of C-S-H. For example,
follows; almost all the reinforced concrete structures BET surface from nitrogen sorption isotherm (Tomes et
show decrease of natural frequency and this attributes to al. 1957; Hunt et al. 1960; Parrott et al. 1980; Litvan and
the reduction of stiffness of concrete structure, because Myers 1983), water vapor sorption isotherm (Tomes et al.
there is little difference of mass of structure during the 1957; Maruyama et al. 2014a), and surface area obtained
monitoring in some cases. Consequently, there is a by SAXS (Winslow and Diamond 1974; Kropp et al.
common environmental condition which cause reduction 1985; Völkl et al. 1987) are affected by long-term drying
Fig. 1 Natural frequency change as a function of time, the relative natural frequency is the ratio of natural frequency to the
natural frequency of buildings just after construction.
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 693
condition. A shape of water vapor sorption isotherms is modulus change of hardened cement paste under the first
also dramatically affected by the drying (Maruyama et al. desorption process and it showed concave curve in the
2014a). It has been reported that irreversible shrinkage relationship between relative humidity and Young’s
after the first drying is large when W/C of paste is large modulus, and the minimum value was observed around
(Helmuth and Turk 1967). A part of these phenomena is 30 ~ 40% RH. The variation was within 18% of the
considered to attribute to the low density (LD) C-S-H, Young’s modulus at the saturated state. This Young’s
which is quantitatively proposed by Tennis and Jennings modulus is obtained by compressive loading test.
(2000). A picture obtained by field emission (FE) scan- In our previous experiments (Horiguchi et al. 2011;
ning electron microscopy (SEM) of hydrated cement Maruyama et al. 2014a), similar trends are confirmed by
paste is shown in Fig. 2. In this picture, the low density using the cement paste made from white cement by
C-S-H is clearly confirmed in white Portland cement bending test and cement paste made from white cement
paste with W/C = 0.55 and it is produced outer area of and ordinary Portland cement by ultrasonic pulse veloc-
original boundary of cement particle. This LD C-S-H is ity measurement. The summary is shown in Fig. 4. In
almost similar definition of that of outer product firstly case of Young’s modulus determined by bending test and
addressed by Goto et al. (1976) ultrasonic pulse velocity measurement, they show 50%
This LD C-S-H is believed to shrink and be compacted and 25% variation under the first drying, and around 40 ~
by drying and the nitrogen accessible area is reduced. A 50% RH the minimum values were observed. It should
morphology change of LD C-S-H is observed by envi- be mentioned that there is a discrepancy in the range
ronmental SEM (Fonseca and Jennings 2010) and this is from 100% to 80% RH between bending test data shown
also reproduced by FE-SEM as they were shown in Fig. in Fig. 4 and ultra-sonic measurement results or Witt-
3. In wet condition, LD C-S-H has indeterminate mor-
phology while it has fibrillar shape in the sample slowly
dried under 40% RH for 1.5 years. These observations
clearly show the dynamic microstructural change in
hardened cement paste under the first drying process.
And thus, cement paste has a colloidal nature (Jesser
1927; Tomes et al. 1957; Jennings 2000, 2008).
Based on these facts, modulus of elasticity and
shrinkage strain of hardened cement paste should cause
the change of Young’s modulus and volume of concrete.
Sereda et al. (1966) substantially studied the relation
between relative humidity and Young’s modulus of hcp,
but they focused mainly on the Young’s modulus change
after strongly dried condition and avoided to discuss
about the unstable calcium silicate hydrate under the first
desorption process because their objective is to clarify
the role of well-stabilized calcium silicate hydrate in Fig. 2 FE-SEM observation of LD C-S-H and HD C-S-H
cement paste. Wittmann (1973) showed the Young’s in white Portland cement paste with W/C=0.55.
Fig. 3 FE-SEM observation of morphology change of LD C-S-H in white Portland cement paste with W/C=0.55. In case of
lime-water cured sample with rapid drying for pre-drying showed indeterminate morphology of C-S-H on the crushed
surface (left). Fibrillar LD C-S-H shown on the surface of cracked surface of matured white Portland cement paste dried
after 40% RH for 1.5 year (right). The sample detail is shown in references (Maruyama et al. 2014a, Maruyama et al.
2015), and they are SDS100 and SDS40 respectively. These SEM data are provided by Drs. Watanabe and Matsui (Asahi
Kasei Co.).
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 694
(a) Young’s modulus by bending test (b) Young’s modulus by ultrasonic pulse velocity
Fig. 4 Young’s modulus of hardened cement paste as a function of relative humidity for long-term drying affected by ce-
ment type, water to cement ratio, and loading type. W: white cement, N: ordinary Portland cement, XX: water to cement
ratio of cement paste. All the cements are categorized as Type-I. Data are from Horiguchi et al. (2011) and Maruyama et
al. (2014a).
was suggested that another shrinkage mechanism is ac- layers, and randomly attached silicate dimers on the CaO
tive rather than surface energy shrinkage mechanism in layer should inflect the structure, crystal growth in the
this range (Maruyama et al. 2015), and concerning this 002 direction is difficult (Gartner 1997). In addition,
another shrinkage mechanism, statistical thickness of water sorption process should be initiated in the space
adsorption is the key parameter. surrounded by C-S-H monolayers, because surface po-
Now, the important issue for long-term properties tentials from each C–S–H monolayer are overlapped.
change of hcp is laid in the first drying process, and it is These structural characteristics must change the distance
related to colloidal behavior and atomic-scale structural between C–S–H monolayers during the sorption process
change of C-S-H. All the phenomena related water con- like montmorillonites (Cases et al. 1992). In other words,
tent change under drying can be interpreted different gel water is re-interpreted as interlayer water in movable
ways as shown in Fig. 7. In the previous models, C-S-H C-S-H monolayers.
is considered to have the fixed interlayer distance and Complete picture of C-S-H contributes the time de-
water sorption process in the interlayer space is saturated pendent properties change of concrete structures under
under 11 ~ 30% RH. This assumption is based on an drying, further investigations are required for this pur-
interpretation by Brunauer et al. (1967) from water vapor pose.
sorption isotherms. However, it is possible that distance
of C-S-H mono layers and number of adsorbed water in 3.2 Aggregate behavior under the first drying
this space depend on the drying process. Previous papers Aggregate generally occupied about 60~70% of volume
(Maruyama et al. 2015, 2016a) proposed a hypothesis of normal strength concrete. Therefore, the property of
that there is a movable C–S–H monolayer and the dis- aggregate has a large impact on mechanical performance
tance between the C–S–H monolayers is dynamically of concrete. Generally, the chemical and physical sta-
changed by drying and re-humidifying. Because there is bilities for required property as a concrete aggregate are
no structural limitation in maintaining the distance be- evaluated by national or international standard before
tween C–S–H monolayers, such as Si-Q3 sites which concrete mixing, however, still many variations are pos-
should be a role of pillar to fix the distance between CaO sible. One of them the shrinkage performance of aggre-
(c) Relationship between incremental statistical thickness of adsorption and drying shrinkage
Fig. 6 (a) Shrinkage strain of ordinary Portland cement paste with water to cement ratio of 0.55 (N55), 0.40 (N40), and
0.30 (N30). (b) Water content in hcp for N55, N40, and N30. (c) Relationship between incremental statistical thickness of
adsorption and drying shrinkage above 40% RH.
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 696
gate is not fully checked. Drying shrinkage of concretes will be treated in the next section through restraint per-
which were made by Japanese design for mixture pro- formance for shrinkage of mortar or cement paste.
portion showed high correlation of shrinkage of coarse
aggregate (Fujiwara 2008). The example is shown in Fig. 3.3 Young’s modulus, shrinkage, and creep of
8 made from the data by Horiguchi et al. (2011) and concrete under the first drying
Teranishi et al. (2011). From this aspect, the inherent Shrinkage strain of matured mortar is more than 5 times
shrinkage property of aggregate is significant. In our as large as that of aggregate, and Young’s modulus of
previous study, shrinkage of sandstone aggregates mortar is less than quarter of that of aggregate (Maru-
available in Japan is found to be a function of amount of yama et al 2014b). Consequently, large tensile stress is
chlorite which was formed in diagenesis (Igarashi et al. produced in mortar of concrete under the first drying
2015), as it is shown in Fig. 9. This is because chlorite process brought by the restraint roll of coarse aggregate
forms in matrix located between quartz or feldspar grains and cracks can be formed around coarse aggregate
and the clay minerals can shrink without large restraint (Carlson 1938; Hansen and Nielsen 1965; Hobbs 1974;
by the grains. The relationship between mineral compo- Bisschop and van Mier 2008; Idiart et al. 2011;
sition of aggregate and physical properties are important Maruyama and Sasano 2014). The width of cracks
issue for aging management of concrete. formed around aggregate becomes wider as mortar
Particle size and its distribution, and interfacial zone shrinks. Therefore, the crack opening volume accumu-
Calcium silicate
mono layer
Water content
Water in
space between Wet Dry
C‐S‐H monolayers
Relative humidity
(b) Proposed model
Fig. 7 Comparison of (a) Feldman and Sereda microstructure model (Feldman and Sereda 1968) and (b) Proposed model
in which water is trapped between calcium silicate mono layers and their distance and number of trapped water molecules
are dramatically changed under the first desorption process (Maruyama et al. 2015).
Fig. 8 Relationship between shrinkage of coarse aggre- Fig. 9 Relationship between shrinkage strain of sand-
gate and that of concrete (Horiguchi et al. 2011; Teranishi stone aggregate and amount of chlorite (Igarashi et al.
et al. 2011). 2015).
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 697
lated in concrete is a function of the difference of role. As it is seen in Fig. 10, concrete with shrinking
shrinkage strain between mortar and coarse aggregate. sandstone showed less reduction of Young’s modulus
The increase in voids in concrete reduces the Young’s after drying.
modulus of concrete because load bearing area is reduced Consequently, Young’s modulus change of concrete is
(Maruyama et al. 2014b) (see Fig.10). Resultantly, dif- tightly related to the shrinkage of concrete and mortar.
ference of shrinkage between concrete and mortar rep- There are plenty of researches on relationship between
resent the volume of accumulated spaces by opening of concrete shrinkage and mixture proportion, therefore,
cracks around aggregate. Therefore, the reduction ratio some important notes are addressed here.
of Young’s modulus of concrete after drying has linear Relationship between the shrinkage of mortar and the
relation with the difference of shrinkage between con- shrinkage of cement paste is not fully understood. Be-
crete and mortar as it is shown in Fig. 11. With this re- cause the shrinkage of mortar is only a function of
gards, shrinkage of aggregate is related to the reduction volume of aggregate (Fig. 12) and not the size of ag-
of Young’s modulus through restraint role to the mortar gregate (Thierry et al. 2008; Sakaida et al. 2014) while,
shrinkage. Shrinking aggregate has a smaller restraint the size of coarse aggregate has a large impact on
shrinkage of concrete (Carlson 1938; Maruyama and
Sasano 2014). This phenomenon can be partially ex-
plained by a role of interfacial transition zone (ITZ)
formed between cement paste matrix and aggregate
(Neubauer et al. 1996; Maruyama and Sugie 2014) but
the restraint role of small grains on the shrinkage of paste
remains to be solved.
ITZ is a one of the key factor of concrete property. It is
Fig.12 Shrinkage of paste and mortar with different fine aggregate volume and sizes (Sakaida et al. 2014).
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 698
possible to explain the shrinkage of concrete affected by the concept proposed by Neville and Meyers (1964). In
the aggregate size by using the presence of ITZ case of Young concrete and concrete block with a large
(Maruyama and Sugie 2014). The large size of aggregate section, the Neville’s model is applicable, but in matured
reduces the shrinkage of concrete. ITZ is generally con- stage, the impact of increase in Young’s modulus due to
sidered to be produced by the “wall effect” of the cement additional hydration after demolding can be negligible
particle packing process on the surface of the aggregate and drying impact may overcome it. To confirm this, the
(Scrivener et al. 2004). And the weak property on the components of creep coefficient is evaluated by using the
surface of aggregate is believed by the facts of porosity data reported by Raymond and Harmer (1931) in Fig. 14.
distribution findings (Breton et al. 1993; Scrivener et al. In Fig. 14 right, reduction of Young’s modulus, which is
2004; Cwirzen and Penttala 2005; Herve et al. 2010), an based on the Fig. 10 is considered for one of the creep
SEM analysis (Scrivener and Gartner 1988; Scrivener et coefficient component. It should be noted that this as-
al. 2004), and physical properties testing (Xie et al. 1991; sumption is applicable only to the condition under the
Ping and Beaudoin 1992; Mitsui et al. 1994; Ollivier et al. equilibrium, and this is due to re-distribution of stress in
1995; Hearing 1997). As the porosity in the ITZ is more concrete under equilibrium condition. Schematic stress
than three times that of normal mortar located far from re-distribution in concrete is shown in Fig. 15(a). The
the surface of the aggregate (Scrivener and Gartner 1988; residual components is caused by stress concentration
Scrivener et al. 2004), resistivity to shrinkage of mortar mechanism after drying. One is based on material char-
matrix is low. Therefore, even there is a report that the acters. After formation of cracks around aggregate, stress
thickness of ITZ depends on the size of aggregates (Xie bearing passes are limited and resultantly, the stress per
et al. 1991; Ollivier et al. 1995; Tasong et al. 1999; Rao unit area of cement paste in concrete is increased and
and Raghu Prasad 2004), the resistivity of aggregate to creep behavior is enhanced due to this stress localization.
the shrinkage of mortar matrix can be evaluated by the The other is observed in member size. After drying,
ratio of thickness of ITZ to the radius of aggregate. The surface part of concrete exhibit larger shrinkage than
larger size of aggregate has a large restraint role to the inner concrete. Therefore, the stress is re-distributed and
mortar shrinkage, while the smaller size of aggregate has localized in central part of concrete (shown in Fig. 15(b)).
a small restraint role. And this large restraint role is un- The peripheral concrete part shows shrinkage rather than
deniably related to the reduction of Young’s modulus. creep, therefore, the creep behavior should be governed
Reduction of Young’s modulus also contributes to the by central part concrete. This consequence indicates that
increase of creep under the drying. Larger creep coeffi- measurement procedure of creep deformation, especially
cient under the drying environment than that of concrete in position, needs special care and drying creep is
without any moisture exchange with environment is structural performance and not material performance.
known as Pickett effect (Pickett 1942). If creep is in-
herent nature of cement paste, creep strain at the cement
paste in concrete should be definable. Therefore, this
phenomenon should be explained by the structural nature.
One is the structural condition of mortar and coarse ag-
gregate. Shrinkage strain of cement paste is larger than
the average creep strain of concrete specimen. This leads
that cracks are formed around aggregate and reduction of
Young’s modulus of dried part of concrete is observed
even under the sustained load. This reduction of Young’s
modulus increase the elastic strain under the drying Fig. 13 Schematic of creep coefficient components of
condition as it is shown in Fig. 13. This is different from matured concrete proposed by Neville (A) and proposed
concept (B).
Fig.14 Creep coefficient data of concrete in wet, 50% RH, and 70% RH conditions (left) reported by Raymond and Harmer
(1931) and calculated components of creep coefficients (right).
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 699
4. Impact of volume and Young’s modulus 20 degree Celsius and 60% relative humidity was con-
change of concrete on RC member ducted (Maruyama and Igarashi 2015) and then rela-
performance tionship between water content and concrete shrinkage
(Maruyama et al. 2014b) is used for input of shrink-
Shrinkage of concrete induces stiffness of reinforced
concrete structure (Okada et al. 1986), and it decrease in
a flexural cracking moment and increase in curvature and
maximum crack width of beam (Tanimura et al. 2007;
Maruyama and Teshigawara 2010). This is explained by
tensile stress of concrete induced by shrinkage restraint
and strain difference before and after cracking of con-
crete. From this regards, creep strain of concrete in RC
column under axial load has the same role as that of
shrinkage. Creep strain and shrinkage strain in column (a) Stress re-distribution in local concrete after drying
make re-distribution of stress among concrete and rebars,
and in case of un-loading process such as seismic re-
sponse, crack is easily produced in tension (Komuro et al.
2008). This re-distribution of stress caused by shrinkage
also reduces the moment capacity at the failure of cover
concrete (Collins et al. 1993). Releasing of compressed
stress of rebar due to concrete shrinkage restraint causes
the additional bending movement and crack opening. In
addition, deflection at the moment capacity at failure of (b) Stress re-distribution in concrete member after drying
concrete is increased by creep strain based on our pre- Fig. 15 Schematic of stress re-distribution in concrete
liminary experiment for impact of sustained load on total and concrete member after drying.
stress-strain curve of concrete shown in Fig.16.
A diagonal cracking moment of RC beam as well as a
shear capacity of RC beam is also affected by shrinkage
(Maruta and Yamazaki 1990; Katayose et al. 2006; Sato
and Kawakane 2008; Gebreyouhannes and Maekawa
2011). This is simply because the stress, position, and
direction of diagonal cracking are altered by shrink-
age-induced stress and enlarged crack width. The wider
diagonal cracks also affect the performance of RC col-
umn, the loading capacity in cyclic loading process after
opening of diagonal cracking is decreased by this width
of diagonal cracks (Maruyama et al. 2011).
Decreasing of Young’s modulus of concrete leads re-
duction of stiffness of RC member. In addition, shrinkage
induced cracking has a large impact on stiffness of RC
member. It should be noted that the both two impacts are Fig. 16 Impact of sustained load on total stress-strain
brought by the shrinkage of cement paste in concrete curve of concrete. (Concrete specimens were the same
mainly. One is by way of cracks around aggregate and as G3 specimens in Maruyama et al. (2014b) with more
they are almost invisible, and the other is by way of than 3-year-curing period. W/C was 0.55.)
visible cracks which are localized cracks due to restraint
of concrete by rebars or outer structure. This difference is
induced by aggregate and its ITZ performance. Concrete
containing (pure) limestone coarse aggregate is known to
reduce the shrinkage-induced visible cracking (Mitani et
al. 2016), but this concrete showed relatively large de-
crease in Young’s modulus of concrete under drying (as it
is shown in Fig. 10). This is the trade-off relationship
(Maruyama et al. 2016b). To compare the impacts of
decrease in Young’s modulus and shrinkage-induced
cracking, simplified numerical analysis has been con-
ducted (Sugie et al. 2014). The calculation was made by
2-dimensional finite element method (FEM). The target
RC wall is shown in Fig. 17. In this moisture transfer at
Fig.17 Target concrete wall structure for FE-analysis.
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 700
age-induced stress calculation in concrete structure. the risk of base line trend of natural frequency. In case of
Shrinkage strain is considered as equivalent nodal force high-rise building or structure, secondary or other modes
to this system. In addition to this, Young’s modulus should be considered.
change of concrete is considered by using the relation- The base line trend is also crucial to evaluate the
ship between water content and Young’s modulus of damage during the earthquake. Although many moni-
concrete (Maruyama et al. 2014b). In FE-analysis, cracks toring sensors, such as accelerometer and velocimeter,
are considered as smeared model. The results of structure are installed to analyze the performance curve (dis-
made from concrete containing limestone are shown in placement versus load) of the structure, damage due to
Fig. 18(b). The reduction of stiffness (Fig.18(a)) as well seismic response can not be identified without baseline
as typical calculated cracks in concrete wall (Fig. 18(b)) trend due to drying. These sensors and obtained analysis
is shown. Behavior of structure made from concrete results should contribute to post earthquake performance
containing limestone is compared with that from con- evaluation of the structure for efficient operation of NPP.
crete containing shrinking sandstone concrete and the To develop the automated system for post earthquake
results are shown in Fig. 18(b). integrity analysis, the mechanism of natural frequency
In Fig. 19, the ratio of natural frequency to that change and its prediction are necessary.
without drying is estimated under the condition that there It should be noted that the strength and Young’s
is no mass change, and the primary mode is intended. modulus of concrete has no longer strong correlation
Figure 19 indicated that the shrinkage-induced cracking after matured state (Maruyama et al. 2014b), because
has larger impact than that of decrease in Young’s drying condition changes strength and Young’s modulus
modulus due to drying. Based on the comparison with a of concrete by different mechanisms. Direct evaluation
record of Onagawa NPP-3, the reduction trend obtained of Young’s modulus of concrete is desired, but using the
by FE-analysis seems comparable to that of real record
and it suggests that the reason of stiffness reduction of
real concrete structure is caused by drying environment.
(a) (b)
Fig. 18 Calculation results of relationship between load and deformation of concrete wall structure after drying (a) and
typical calculated crack patterns in concrete wall (b).
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 701
data obtained from core-sample needs special care, be- plants by identifying the damage due to seismic response
cause massive concrete member has a distribution of from baseline trend of natural frequency change as well
water content and resultant Young’s modulus. In this as reduction of stiffness of structure.
sense, numerical evaluation might have potential to The lost parts of scientific knowledge and desired
evaluate the average concrete stiffness. experimental data to obtain the whole picture of the
The most import knowledge is that the reduction ratio natural frequency change under ordinary drying condi-
of natural frequency after equilibrium and the rate of tion have been pointed out.
reduction of natural frequency to achieve the equilibrium.
Because aging management is based on both integrity at Acknowledgement
the present state and in future until the target life span, The present research has been partly sponsored by JSPS
those two parameters are crucial. Regarding the reduc- KAKENHI Grant Number 15H04077.
tion ratio, it is possible to estimate the value by using
drying shrinkage strain and Young’s modulus of concrete References
under the given environment which is equivalent to that Appa Rao, G. and Raghu Prasad, B. K., (2004). “Influ-
in nuclear power plant. But the rate of reduction of ence of type of aggregate and surface roughness on the
natural frequency needs more scientific background of interface fracture properties.” Materials and Struc-
concrete under drying process. The first drying process tures, 37(5), 328-334.
in concrete is always accompanied with colloidal altera- Bažant, Z. P., (1972). “Thermodynamics of hindered
tion of C-S-H, and this process is not considered in ex- adsorption and its implications for hardened cement
isting research and they always assumes that concrete is paste and concrete.” Cement and Concrete Research, 2,
just a porous material, resultantly, the drying process 1-16.
under high temperature or temperature gradient needs Beaudoin, J. J., Raki, L., Alizadeh, R. and Mitchell, L.,
more data (Shiire and Cheong 1988; England and (2010). “Dimensional change and elastic behavior of
Khoylou 1995; Lien and Wittmann 1998). layered silicates and portland cement paste.” Cement
and Concrete Composites, 32, 25-33.
6. Summary Beltzung, F. and Wittmann, F. H., (2005). “Role of
disjoining pressure in cement based materials.”
It is concluded that possible mechanisms of natural fre- Cement and Concrete Research, 35, 2364-2370.
quency change of concrete structures under an ordinary Bisschop, J. and van Mier, J. G. M., (2008). “Effect of
drying condition are reduction of Young’s modulus of aggregates and microcracks on the drying rate of
concrete and stiffness reduction of reinforced concrete cementitious composites.” Cement and Concrete
member due to shrinkage-induced cracking. Both two Research, 38, 1190-1196.
mechanisms can be provoked by drying, which is natural Breton, D., Carles-Gibergues, A., Ballivy, G. and
environmental condition for all the building structures. Grandet, J., (1993). “Contribution to the formation
While the Young’s modulus change of concrete is ne- mechanism of the transition zone between
glected in the present design codes, it is possible that it rock-cement paste.” Cement and Concrete Research,
shows more than 20% reduction in general drying con- 23, 335-346.
dition. In addition, drying condition changes strength and Brunauer, S., Mikhail, R. S. and Bodor, E. E., (1967).
Young’s modulus differently, it should be kept in mind “Some remarks about capillary condensation and pore
that there is no longer correlation between strength and structure analysis.” Journal of Colloid and Interface
Young’s modulus of concrete in matured state, unlike Science, 25(3), 353-358.
young concrete. From this point of view, direct evalua- Carlson, R. W., (1938). “Drying shrinkage of concrete as
tion of Young’s modulus of concrete, but not from a affected by many factors.” American Society Testing &
calculation based on compressive strength and relation- Materials Proceedings, 38(2), 419-437.
ship between strength and Young’s modulus, is desired. Cases, J. M., Berend, I., Besson, G., Francois, M., Uriot, J.
Concrete, however, has a distribution in its properties in P., Thomas, F. and Poirier, J. E., (1992). “Mechanism
the reinforced concrete member under drying condition, of adsorption and desorption of water vapor by
therefore, data treatment of core-sample should take into homoionic montmorillonite. I. The sodium-exchanged
account this nature of concrete member. form.” Langmuir, 8(11), 2730-2739.
The natural frequency change is important for the in- Chiang, W.-S., Fratini, E., Baglioni, P., Liu, D. and Chen,
tegrity analysis of nuclear power plants because seismic S.-H., (2012). “Microstructure determination of
response change and sympathetic vibration between calcium-silicate-hydrate globules by small-angle
structure and installed implementations are expected. neutron scattering.” The Journal of Physical
Second mode of structure is a potential risk for sympa- Chemistry C, 116( 8), 5055-5061.
thetic vibration of installed implementations. To under- Chiang, W.-S., Fratini, E., Ridi, F., Lim, S.-H., Yeh, Y.-Q.,
stand the mechanism should contributed to Baglioni, P., Choi, S.-M., Jeng, U. S. and Chen, S.-H.,
post-earthquake integrity evaluation which is necessary (2013). “Microstructural changes of globules in
to efficient and continuous operation of nuclear power calcium–silicate–hydrate gels with and without
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 702
additives determined by small-angle neutron and x-ray on the shrinkage of concrete.” ACI Journal, 71(9),
scattering.” Journal of Colloid and Interface Science, 445-450.
398, 67-73. Horiguchi, N., Igarashi, G. and Maruyama, I., (2011a).
Collins, M. P., Mitchell, D. and Macgregor, J. G., (1993). “Fundamental study on volume change of aggregate
“Structural design considerations for high-strength due to loss of evaprable water.” Proceedings of Japan
concrete.” Concrete International, 15(5), 27-34. Concrete Institute, 33(1), 131-136. (in Japanese)
Cwirzen, A. and Penttala, V., (2005). “Aggre- Horiguchi, N., Igarashi, G. and Maruyama, I., (2011b).
gate–cement paste transition zone properties affecting “Young's modulus and poisson's ratio of hardened
the salt–frost damage of high-performance concretes.” cement paste under different relative humidity.” J.
Cement and Concrete Research, 35, 671-679. Struct. Constr. Eng. , 76(660), 231-236. (in Japanese)
England, G. L. and Khoylou, N., (1995). “Moisture flow Hunt, C. M., Tomes, L. A. and Blaine, R. L., (1960).
in concrete under steady state non-uniform tempera- “Some effects of aging on the surface area of portland
ture states: Experimental observations and theoretical cement paste.” Journal of Research of the National
modelling.” Nuclear Engineering and Design, Bureau of Standards, 64A(2), 163-169.
156(1–2), 83-107. Idiart, A., López, C. and Carol, I., (2011). “Modeling of
Feldman, R. F. and Sereda, P. J., (1968). “A model for drying shrinkage of concrete specimens at the
hydrated portland cement paste as deduced from meso-level.” Materials and Structures, 44(2),
sorption-length change and mechanical properties.” 415-435.
Matériaux et Construction, 1(6), 509-520. Igarashi, G., Maruyama, I., Nishioka, Y. and Yoshida, H.,
Fonseca, P. C. and Jennings, H. M., (2010). “The effect (2015). “Influence of mineral composition of siliceous
of drying on early-age morphology of C–S–H as rock on its volume change.” Construction and
observed in environmental SEM.” Cement and Building Materials, 94, 701-709.
Concrete Research, 40, 1673-1680. Japanese Government, (2013). “Basic act for national
Fujiwara, T., (2008). “Effect of aggregate on drying resilience contributing to preventing and mitigating
shrinkage of concrete.” Journal of Advanced Concrete disasters for developing resilience in the lives of the
Technology, 6(1), 31-44. citizenry.” Act No. 95 of 2013.
Gartner, E. M., (1997). “A proposed mechanism for the Jennings, H. M., (2000). “A model for the microstructure
growth of C–S–H during the hydration of tricalcium of calcium silicate hydrate in cement paste.” Cement
silicate.” Cement and Concrete Research, 27, 665-672. and Concrete Research, 30, 101-116.
Gebreyouhannes, E. and Maekawa, K., (2011). “Nu- Jennings, H. M., (2008). “Refinements to colloid model
merical simulation on shear capacity and post-peak of C–S–H in cement: CM-II.” Cement and Concrete
ductility of reinforced high-strength concrete coupled Research, 38, 275-289.
with autogenous shrinkage.” Journal of Advanced Jesser, L., (1927). “Kolloide chemische reaktionen des
Concrete Technology, 9(1), 73-88. tonerdezementmörtels.” Zement, 16, 741.
Goto, S., Daimon, M., Hosaka, G. and Kondo, R., (1976). Kashima, T. and Kitagawa, Y., (2006). “Dynamic
“Composition and morphology of hydrated tricalcium characteristics of a building estimated from strong
silicate.” Journal of the American Ceramic Society, motion records using evolution strategy.” Journal of
59(7-8), 281-284. structural and construction engineering, 602, 145-152.
Grassl, P., Wong, H. S. and Buenfeld, N. R., (2010). (in Japanese)
“Influence of aggregate size and volume fraction on Katayose, N., Takamori, N., Nishida, H. and Teraoka, M.,
shrinkage induced micro-cracking of concrete and (2006). “Mechanical properties and autogenous
mortar.” Cement and Concrete Research, 40, 85-93. shrinkage behavior of high strength concrete at early
Hansen, T. C. and Nielsen, K. E. C., (1965). “Influence of age.” Proceedings of the Japan Concrete Institute,
aggregate properties on concrete shrinkage.” ACI 28(1), 497-502. (in Japanese)
Journal, 62(7), 789-794. Kobayashi, K., (2000). “Konkuri-to ga abunai [concrete
Hearing, B., (1997). “Fracture behavior of crisis].” Tokyo, Japan, Iwanami Shoten. (in Japanese)
mortar-aggregate interfaces in concrete.” Thesis Komatsu, Y., (2000). “Life span of dwellings (Special
(MSc), Massachusetts Institute of Technology. issue 1: Dwelling stock analysis).” Housing review,
Helmuth, R. A. and Turk, D. H., (1967). “The reversible 16(2), 5-20. (in Japanese)
and irreversible drying shrinkage of hardened portland Komatsu, Y., (2008). “Life time estimations of Japanese
cement and tricalcium silicate pastes.” Journal of the buildings and houses at the years of 1997 and 2005.”
PCA Research and Development Laboratories, 9(2), 8 Journal of architecture and planning, 73(632),
- 21. 2197-2205. (in Japanese)
Herve, E., Care, S. and Seguin, J. P., (2010). “Influence Komuro, T., Imai, K., Korenaga, T. and Watanabe, F.,
of the porosity gradient in cement paste matrix on the (2008). “Study on long term performance of
mechanical behavior of mortar.” Cement and Concrete reinforced concrete columns using high strength
Research, 40, 1060-1071. concrete.” Proceedings of the Japan Concrete Institute,
Hobbs, D. W., (1974). “Influence of aggregate restraint 30(3), 223-228. (in Japanese)
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 703
Korb, J. P. (2009). “NMR and nuclear spin relaxation of Maruyama, I., Sasano, H., Nishioka, Y. and Igarashi, G.,
cement and concrete materials.” Current Opinion in (2014b). “Strength and young's modulus change in
Colloid & Interface Science, 14(3), 192-202. concrete due to long-term drying and heating up to
Kropp, J., Gräfenecker, T. and Hilsdorf, H. K., (1985). 90 °C.” Cement and Concrete Research, 66, 48-63.
Characterization of the microstructure of hydrated Maruyama, I. and Sugie, A., (2014). “Numerical study
cement paste by small angle x-ray scattering. In: on drying shrinkage of concrete affected by aggregate
Principles and Applications of Pore Structural size.” Journal of Advanced Concrete Technology,
Characterization. J. W. Arrowsmith. Bristol, 83-96. 12(8), 279-288.
Lagier, F., Jourdain, X., De Sa, C., Benboudjema, F. and Maruyama, I. and Teshigawara, M., (2010). “Effect of
Colliat, J. B., (2011). “Numerical strategies for autogenous shrinkage of UHSC on bending behavior
prediction of drying cracks in heterogeneous of RC column.” Sixth International Conference on
materials: Comparison upon experimental results.” Concrete under Severe Conditions: Environment and
Engineering Structures, 33(3), 920-931. Loading Merida, Mexico, CRC press, 2, 879-888.
Li, L., Nakamura, A., Kashima, T. and Teshigawara, M., Maruyama, I., Teshigawara, M., Suwada, H., Tanimura,
(2014). “Earthquake damage evaluation of an 8-story M., Mitani, Y., Nakamura, A. and Teramoto, A.,
steel-reinforced concrete building using Sa-Sd curves.” (2011). “Effect of autogenous shrinkage on the
J. Struct. Constr. Eng., AIJ, 79(702), 1107-1115. structural performance of short rc column.”
Lien, H. P. and Wittmann, F. H., (1998). “Mass transfer in Summaries of Technical Papers of Annual Meeting
inhomogeneous porous media under thermal Architectural Institute of Japan., C-2, Structures IV,
gradients.” Nuclear Engineering and Design, 179(2), Reinforced concrete structures prestressed concrete
179-189. structures masonry wall structures, 107-108. (in
Litvan, G. G. and Myers, R. E., (1983). “Surface area of Japanese)
cement paste conditioned at various relative McDonald, P. J., Rodin, V. and Valori, A., (2010).
humidities.” Cement and Concrete Research, 13, “Characterisation of intra- and inter- C–S–H gel pore
49-60. water in white cement based on an analysis of NMR
Maruta, M. and Yamazaki, M., (1990). “Study on the signal amplitudes as a function of water content.”
shear behavior of RC beams subjected to hight Cement and Concrete Research, 40, 1656-1663.
temperature for long time.” Proceedings of the Japan Mitani, Y., Ishii, Y., Tanimura, M. and Maruyama, I.,
Concrete Institute, 12(1), 603-608, (in Japanese). (2016). “Evaluation on shrinkage cracks of
Maruyama, I., (2010). “Origin of drying shrinkage of expansive/shrinkage reducing concrete.” Cement
hardened cement paste: Hydration pressure.” Journal Science and Concrete Technology, 69, 341-348. (in
of Advanced Concrete Technology, 8(2), 187-200. Japanese)
Maruyama, I. and Igarashi, G., (2015). “Numerical Mitsui, K., Li, Z., Lange, D. A. and Shah, S. P., (1994).
approach towards aging management of concrete “Relationship between microstructure and mechanical
structures: Material strength evaluation in a massive properties of paste-aggregate interface.” ACI
concrete structure under one-sided heating.” Journal Materials Journal, 91(1), 30-39.
of Advanced Concrete Technology, 13(11), 500-527. Muller, A. C. A., Scrivener, K. L., Gajewicz, A. M. and
Maruyama, I., Igarashi, G. and Nishioka, Y., (2015). McDonald, P. J., (2012). “Densification of C–S–H
“Bimodal behavior of C–S–H interpreted from measured by 1H NMR relaxometry.” The Journal of
short-term length change and water vapor sorption Physical Chemistry C, 117, 1, 403-412.
isotherms of hardened cement paste.” Cement and Muller, A. C. A., Scrivener, K. L. Gajewicz, A. M. and
Concrete Research, 73, 0, 158-168. McDonald, P. J., (2013). “Use of bench-top NMR to
Maruyama, I., Nishioka, Y., Igarashi, G. and Matsui, K., measure the density, composition and desorption
(2014a). “Microstructural and bulk property changes isotherm of C–S–H in cement paste.” Microporous
in hardened cement paste during the first drying and Mesoporous Materials, 178, 99-103.
process.” Cement and Concrete Research, 58, 20-34. Muller, A. C. A., Scrivener, K. L., Skibsted, J., Gajewicz,
Maruyama, I., Sakamoto, N., Matsui, K. and Igarashi, G., A. M. and McDonald, P. J., (2015). “Influence of silica
(2016a). “Microstructural changes in white Portland fume on the microstructure of cement pastes: New
cement paste under the first drying process evaluated insights from 1H NMR relaxometry.” Cement and
by WAXS, SAXS, and USAXS.” Cement and Con- Concrete Research, 74, 116-125.
crete Research, doi: 10.1016/j.cemconres.2016.10.002. Neubauer, C., Jennings, H. M and Garboczi, E., (1996).
Maruyama, I. and Sasano, H., (2014). “Strain and crack “A three-phase model of the elastic and shrinkage
distribution in concrete during drying.” Materials and properties of mortar.” Advanced Cement Based
Structures, 47(3), 517-532. Materials, 4(1), 6-20.
Maruyama, I., Sasano, H. and Lin, M., (2016b). “Impact Neville, A. M. and Meyers, B. L., (1964). “Creep of
of aggregate properties on the development of shrink- concrete: Influencing factors and prediction.” Special
age-induced cracking in concrete under restraint con- Publication, 9, 1-33.
ditions.” Cement and Concrete Research, 85, 82-101. Nicoleau, L., Gadt, T., Chitu, L., Maier, G. and Paris, O.,
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 704
(2013). “Oriented aggregation of calcium silicate Advanced Concrete Technology, 6(1), 45-67.
hydrate platelets by the use of comb-like copolymers.” Scrivener, K. L., Crumbie, A. and Laugesen, P., (2004).
Soft Matter, 9(19), 4864-4874. “The interfacial transition zone (ITZ) between cement
Nuclear and Industrial Safety Agency, (2011). paste and aggregate in concrete.” Interface Science,
“Numerical simualtion of seismic responce of 12(4), 411-421.
buildings.” Tohoku Electric Power Co., Inc. Onagawa Scrivener, K. L. and Gartner, E. M., (1988).
Nuclear Power Plant 2 and 3., Nuclear and Industrial “Microstructural gradients in cement paste around
Safety Agency. Buildings/Structure, Tokyo, 4-3-1. aggregate particles.” Materials Research Society
Ogata, Y., Hirotani, K., Aizawa, N., Inoue, N. and Symposium Proceedings, 114, 77-85.
Nozawa, T., (2011). “A study of initial stiffness on the Sereda, P. J., Feldman, R. F. and Swenson, E. G., (1966).
building changed by passing age at small and “Effect of sorbed water on some mechanical properties
medium-sized earthquake ground motion by arx of hydrated portland cement pastes and compacts.”
model.” Summaries of technical papers of Annual Highway Research Board Special Report, 90, 58-73.
Meeting Architectural Institute of Japan. B-2, Setzer, M. J., (2007). “The solid-liquid gel-system of
Structures II, Structural dynamics nuclear power hardened cement paste.” In: M. J. Setzer Ed. Transport
plants, 2011, 983-984, (in Japanese). in concrete : nano- to macrostructure, TRANSCON 07,
Okada, T., Akino, K., Yoshizaki, S., Inada, Y. and proceedings of 5th International Essen Workshop.
Sugano, S., (1986). “Model tests for evaluation of Essen, Freiburg : Aedificatio Publishers, 3-23.
restoring force characteristics of reactor buildings : Shiire, T. and Cheong, S., (1988). “Experimental study
Part 6 scale effect test.” Summaries of Technical on strength characteristics of massive concrete
Papers of Annual Meeting Architectural Institute of structures subjected to high temperature from on side :
Japan. B, Structures I, 1986, 1089-1090. (in Japanese) In case of subject to temperature of 65℃ from on
Ollivier, J. P., Maso, J. C. and Bourdette, B., (1995). side.” Journal of structural and construction
“Interfacial transition zone in concrete.” Advanced engineering, 387, 8-14. (in Japanese)
Cement Based Materials, 2(10), 30-38. Sugie, A., Maruyama, I. and Teshigawara, M., (2014).
Parrott, L. J., Hansen, W. and Berger, R. L., (1980). “Numerical simulation for stiffness change of RC wall
“Effect of first drying upon the pore structure of due to drying.” Summaries of Technical Papers of
hydrated alite paste.” Cement and Concrete Research, Annual Meeting, 2014, 1175-1176. (in Japanese)
10, 647-655. Takahashi, K., Morohashi, N., Kudo, K. and Sakurada, T.,
Pickett, G., (1942). “The effect of change in (2012). “Vibration characteristics of school buildings
moisture-content on the crepe of concrete under a in the campus of the CIT , Nihon University as inferred
sustained load.” Journal Proceedings, 38, 333-356. from the strong-motion records from the 2011 Off the
Pickett, G., (1956). “Effect of aggregate on shrinkage of Pacific Coast of Tohoku Earthquake.” Summaries of
concrete and a hypothesis concerning shrinkage.” ACI Technical Papers of Annual Meeting, 2012, 683-684.
Journal proceedings, 52(1), 581-590. (in Japanese)
Ping, X. and Beaudoin, J. J., (1992). “Effects of Tanimura, M., Sato, R. and Hiramatsu, Y., (2007).
transition zone microstructure on bond strength of “Serviceability performance evaluation of rc flexural
aggregate-portland cement paste interfaces.” Cement members improved by using low-shrinkage
and Concrete Research, 22, 23-26. high-strength concrete.” Journal of Advanced
Powers, T. C., (1965). “Mechanism of shrinkage and Concrete Technology, 5(2), 149-160.
reversible creep of hardened cement paste.” The Tasong, W. A., Lynsdale, C. J. and Cripps, J. C., (1999).
structure of concrete and its behaviour under load, “Aggregate-cement paste interface: Part I. Influence of
319-344. aggregate geochemistry.” Cement and Concrete
Raymond, E. D. and Harmer, E. D., (1931). “Flow of Research, 29, 1019-1025.
concrete under the action of sustained loads.” ACI Tennis, P. D. and Jennings, H. M., (2000). “A model for
Journal Proceedings, 27(3), 837-901. two types of calcium silicate hydrate in the
Saito, T., (2012). “Change in dynamic characteristics of microstructure of portland cement pastes.” Cement
a seismically isolated building caused by the 2011 Off and Concrete Research, 30, 855-863.
the Pacific Coast of Tohoku Earthquake.” Research Teranishi, K., Maruyama, I., Saito, K., Hiraiwa T. and
Insitute of Shimizu Corporation. Tokyo, Japan, 89, Mori, K., (2011). “Estimation of drying shrinkage
13-22. (in Japanese) strain of concrete and selection of methods for
Sakaida, T., Sasano, H. and Maruyama, I., (2014). shrinkage crack control : Research committee report
“Effect of fine aggregate size on drying shrikage of on shrinkage crack of concrete.” Concrete Journal,
mortar.” Proceedings of the Japan Concrete Institute, 49(11), 8-16. (in Japanese)
36(1), 526-531. (in Japanese) Thierry, J. P., Chanvillard, G., Schlangen, E. and De
Sato, R. and Kawakane, H., (2008). “A new concept for Schutter, G., (2008). “Autogenous and drying
the early age shrinkage effect on diagonal cracking shrinkage modeling: From paste to concrete.”
strength of reinforced hsc beams.” Journal of International RILEM Symposium on Concrete
I. Maruyama / Journal of Advanced Concrete Technology Vol. 14, 691-705, 2016 705
Modelling-ConMod'08, RILEM Publications SARL, Valori, A., McDonald, P. J. and Scrivener, K. L., (2013).
337-344. “The morphology of C–S–H: Lessons from 1H nuclear
Tomes, L. A., Hunt, C. M. and Blaine, R. L., (1957). magnetic resonance relaxometry.” Cement and
“Some factors affecting the surface area of hydrated Concrete Research, 49, 65-81.
portland cement as determined by water-vapor and Winslow, D. N. and Diamond, S., (1974). “Specific
nitrogen adsorption.” Journal of Research of the surface of hardened portland cement paste as
National Bureau of Standards, 59(6), 357-364. determined by small-angle x-ray scattering.” Journal
Toyobe, R., Kojima, H., Tobita, J. and Fukuwa, N., of the American Ceramic Society, 57(5), 193-197.
(2013). “Long-term seismic response observation for Wittmann, F. H., (2008). “Heresies on shrinkage and
evaluation of dynamic properties of low and creep mechanisms.” Proceedings of the 8th
medium-rise buildings.” Tokai-shibu Kenkyu International Conference on Creep, Shrinkage and
Houkokusyu, 51, 177-180. (in Japanese) Durability Mechanics of Concrete and Concrete
Trapote-Barreira, A., Porcar, L., Cama, J., Soler, J. M. Structures (CONCREEP 8), Sept, 3-9.
and Allen, A. J., (2015). “Structural changes in C–S–H Wittmann, F. H., (1968). “Surface tension skrinkage and
gel during dissolution: Small-angle neutron scattering strength of hardened cement paste.” Matériaux et
and Si-NMR characterization.” Cement and Concrete Construction, 1(6) 547-552.
Research, 72, 76-89. Wittmann, F. H., (1973). “Interaction of hardened cement
Uebayashi, H. and Nagano, M., (2012). “Statistical paste and water.” Journal of the American Ceramic
estimation of stiffness and damping of high-rise Society, 56(8), 409-415.
buildings based on the reports concerning the response Xie, P., Beaudoin, J. J. and Brousseau, R., (1991). “Effect
analyses of strong motion observation records.” of aggregate size on transition zone properties at the
Summaries of technical papers of annual meeting, portland cement paste interface.” Cement and
2012, 697-698. (in Japanese) Concrete Research, 21, 999-1005.
Völkl, J. J., Beddoe, R. E. and Setzer, M. J., (1987). “The Zhang, W., Zakaria, M. and Hama, Y., (2013). “Influence
specific surface of hardened cement paste by of aggregate materials characteristics on the drying
small-angle x-ray scattering effect of moisture content shrinkage properties of mortar and concrete.”
and chlorides.” Cement and Concrete Research, 17, Construction and Building Materials, 49, 500-510.
81-88.
Journal of Advanced Concrete Technology Vol. 14, 643-663, October 2016 / Copyright © 2016 Japan Concrete Institute 643
Technical report
Abstract
“Guidelines for Maintenance and Management of Structures in Nuclear Facilities” was published by the Architectural
Institute of Japan in 2008 and revised in 2015. This is the first standard that covers the maintenance and management of
structures in Japanese Nuclear Facilities. This has been contributing to the broad activities related to the life-time man-
agement of Nuclear Facilities. This paper outlines these guidelines.
verifying satisfaction of the required performance level structures at nuclear facilities and the performance
of the structural components and members necessary to characteristics required for the maintenance of these
fulfill the required structure functions, and shall be se- functions.
cured by taking appropriate countermeasures and veri- (2) While taking account of the degradation mechanisms
fying their effectiveness. in concrete and steel-framed structures under the
Information on the relation between required functions, conditions specific to structures at nuclear facilities,
related performance characteristics, and soundness such as environmental, siting, structural, and mate-
evaluations is based on the Nordic Five Level System rial, the degradation phenomena and degradation
(Fig. 2). factors to be monitored shall be defined.
When maintenance of structures at nuclear facilities is (3) Prior to maintenance, a maintenance plan shall be
applied to this hierarchy, the purpose is to “ensure public established. This plan shall describe the basic ap-
safety through the maintenance of structures at nuclear proach to inspections, soundness evaluations, coun-
facilities.” Specific requirements to achieve this purpose termeasures and their effectiveness verification, and
are functional requirements, which correspond to, e.g., records.
functions to support important equipment (supporting (4) An inspection plan shall be established based on the
function) and to prevent radioactive material leakage to maintenance plan to specify the inspection method,
outside the structure (leak-proof function). Performance the parts to be inspected, the timing of inspection, the
requirements are specific requirements necessary to inspection implementation scheme, etc. Inspections
satisfy functional requirements. Hence, performance shall be performed according to this inspection plan.
requirements refer to, for example, liquid-tightness per-
formance required of the leak-proof function. A per-
formance verification method is for determining the
attainment of a required performance level through
Purpose
comparison of inspection results, appropriate analysis,
and experimental results with evaluation criteria. A Functional
deemed-to-satisfy provision is a means of securing per- requirements
formance without the need of performance verification
and is established to specify specific structural dimen- Performance
sions, materials, construction methods, etc., deemed to requirements
satisfy the required performance level.
The principle of maintenance and management shown Verification Deemed-to-satisfy
in the present Guidelines is to conduct the following methods provisions
activities in accordance with the above concept. The
interrelationship between the activities is shown in Fig. 3. Fig. 2 Concept of hierarchical performance design of
(1) It is necessary to identify the functions required of buildings.
Inspection Records
(Chapter 6) (Chapter 9)
(5) An evaluation item shall be set for each degradation newly prescribed level. The maintenance covered in the
phenomenon and degradation factor to be monitored, Guidelines must ensure that the structure’s performance
so that the structure can be evaluated for soundness varied by degradation due to aging or degradation due to
at the time of inspection and during its in-service sudden event meets the required level from the present
period. into the future. In case the required performance level is
(6) In accordance with the results of soundness evalua- changed and some of the inspection results fail to meet
tion, a countermeasure shall be taken, such as con- the evaluation criteria, it is reasonably acceptable if the
tinuing or improving the current inspection practices structure’s soundness as a whole is maintained.
and executing repair work. If any repair work is Figure 4 shows a schematic of changes in perform-
executed, the repaired portion shall be checked by ance of a structure affected by degradation due to aging
periodical inspection as well as by additional in- considered in the Guidelines. The performance of a
spection as necessary to ensure that the repaired hard-to-replace concrete may show different changes
portion is restored to the performance level expected depending on the evaluation item (e.g. The compressive
at the time of repair planning. strength increases in the early stage and stops increasing
(7) The results of inspection, soundness evaluation, in the long term. The carbonation depth proceeds fast in
countermeasure implementation and effectiveness the early stage and slowly in the long term), but basically
verification shall be recorded accurately and in detail show gradual changes with maintaining the required
and stored with easy access for future maintenance performance level. Meanwhile, inherently replaceable
references. materials like a coating are specified prescriptive re-
(8) The cycle from maintenance planning to counter- quirements (a deemed-to-satisfy provision) such as the
measure implementation and effectiveness verifica- material specifications and the construction methods.
tion shall be repeated for continuous improvement of Their performance is maintained by replacing at periodic
maintenance practices to ensure more effective intervals throughout its in-service period. The mainte-
maintenance. nance covered in the Guidelines primarily focuses on
ensuring the non-occurrence of rapid performance deg-
3.2 Structure performance changes radation in hard-to-replace structures, and is not intended
The structure’s performance affected by degradation due to restore or improve the structure’s performance, with
to aging tends to show a gradual change over time the exception of replaceable materials such as coatings.
throughout its in-service period. The degree of this In other words, in principle, structures shall always be
change depends on the degree of initial performance, kept in a sound condition through inspection, soundness
environmental conditions, and maintenance practices. evaluation, and repairs or other countermeasures.
Meanwhile, the structure’s performance affected by In the seismic diagnosis of a general structure, seismic
degradation due to sudden event tends to change within a performance of the structure is evaluated with consid-
short time. Structures at nuclear facilities should be de- eration of the degradation due to aging. The seismic
signed to satisfy their required performance level performance of a structure is estimated by multiplying a
throughout their in-service period. Even if the required coefficient according to aging level to its initial per-
performance level is changed according to a revision of formance. Since structures at nuclear facilities are con-
relevant codes and standards, structures must satisfy stantly maintained, severe deterioration sometimes ob-
Performance
Performance decreases fast in the early Performance
stage and slowly in the long term (e.g.,
carbonation) Repair (replacement)
Initial Initial
performance performance
Required level of Required level of
performance* performance*
Performance increases in
the early stage and stops in- Decrease in performance
creasing in the long term (e.g., (without replacement)
compressive strength)
In-service period In-service period
Time Time
Start of Inspection Start of Inspection
operation Soundness operation Soundness
evaluation evaluation
(1) Hard-to-replace part (2) Inherently replaceable part
*: The level of performance, which reflects the latest knowledge, required at the point of soundness evaluation (The level of per-
formance required at the time of design is not necessarily used throughout the in-service period of the building.)
Fig. 4 Conceptual graph of changes in performance characteristics of a building prone to degradation due to aging as-
sumed in the maintenance in accordance with the present Guidelines.
Y. Umeki, S. Sawada, S. Mitsugi, T. Maenaka and K. Takiguchi / Journal of Advanced Concrete Technology Vol. 14, 643-663, 2014 646
served in general structures is not expected. However, it and for long-term soundness evaluation will differ in
is reported that natural frequency of structures decreases specific procedure and/or significance. In any case, the
by earthquake effects or by degradation due to aging. In use of local destructive methods shall be kept to the very
case the dynamic characteristics of structures at nuclear minimum to avoid damage to structures as much as pos-
facilities show noticeable changes caused by earthquake sible.
effects or degradation due to aging, their structural
soundness should be evaluated with taking account of 3.4 Required functions and related performance
such changes. Since the principal cause of these de- characteristics for nuclear facilities
creases in natural frequency due to earthquake effects has It is necessary to identify the functions required of
not been clarified, how to specify the dynamic charac- structures at nuclear facilities and the performance
teristics of structures remains a challenge for analytical characteristics required for the maintenance of such
evaluation. It will be necessary in future to promote functions in accordance with the basic concept described
studies on soundness evaluation of structures affected by above in Section 3.1. The required functions of struc-
earthquake and to accumulate scientific knowledge. tures at nuclear facilities are shown in Fig. 5. These are
classified into two functions, one for ordinary structures
3.3 Minimization of damage to a structure dur- and the other specially required for structures at nuclear
ing maintenance facilities.
In an inspection for current soundness evaluation, the The functions required of a structure generally include
whole structure should be inspected. Accordingly, in supporting function for its own dead weight and live and
principle, damage to concrete shall be checked by visual seismic loads; comfort use-related functions such as
method. For visually unobservable changes in concrete vibration isolation and sound proofing; functions to
strength, their tendencies are grasped using withstand secular changes and to resist fire and other
non-destructive methods. There may be cases where high-temperature phenomena. These functions are simi-
strength is assessed by core sampling for verification and larly required of nuclear facilities. In the Guidelines,
accuracy enhancement of a non-destructive method. these general functions shall be collectively called the
Core sampling not only causes damage to structures but usage function.
also does not necessarily provide more information than The following additional functions are specially re-
visual methods. Hence, it is assumed that soundness quired for structures at nuclear facilities.
evaluation will remain valid even if the use of core ・Supporting function
sampling is limited. A function with which to support major installations
Meanwhile, inspection for long-term soundness and equipment during normal operation or an earth-
evaluation is required to measure the degree of effect of quake.
each degradation factor in order to predict the progres- ・Pressure-proof function
sion of degradation phenomena. Accordingly, there may A function with which to withstand pressures. In the
be cases that require use of a local destructive method, case of a concrete containment vessel, a function re-
such as core sampling or drilling. quired of its concrete portion.
From the above, even in the case of core sampling and ・Shielding function
other inspection methods classified as local destructive A function with which to reduce the radiation dose to
methods, inspections for current soundness evaluation a predetermined level to prevent radiation exposure
Usage function
3 3
Usability 3*1 - - - -
(liquid-tightness) (airtightness)
Shielding
- - - 3 - - -
performance
impact on the general public and occupationally ex- ture and shall also include load-bearing capacity for
posed personnel. temperature loads, fatigue capacity against cyclic loads,
・Leak-proof function and impact resistance capacity against impact loads.
A function with which to prevent leakage of liquid Usability is a generic term for performance charac-
radioactive materials from a specified area to outside. teristics required for comfortable use of structures and
・Negative-pressure holding function refers to, among other things, resistance capacity against
A function in which some areas in a nuclear facility increase in deflection under permanent load due to creep,
are kept under negative pressure by ventilation vibration isolation performance against harmful vibra-
air-conditioning equipment to reduce external dis- tion, thermal insulation performance against heat, wa-
charge of radioactive materials. terproof performance against ordinary rainfall and water
・Secondary-accident prevention function usage, liquid- and air-tightness of concrete skeletons
A function with which to prevent any damage, falling, constructed by conventional methods, and surface pro-
or overturning of a structure from compromising the tection of coatings and surface finishes. In some cases,
functions of installations and equipment. aesthetic appearance not directly related to performance
While the supporting and secondary-accident preven- characteristics of structural components and members
tion functions can be regarded here as part of the usage may have to be taken into consideration and hence shall
function required also of general structures, these func- be included in usability.
tions are intentionally defined as those specifically re- Shielding performance is specifically required of some
quired of nuclear facilities, emphasizing that nuclear areas of each structure in nuclear facilities and refers
facilities are required to have a higher level of safety mainly to neutron and gamma ray shielding performance.
against seismic and similar loads than general structures. Durability shall be regarded as a generic term for re-
Whether a structure’s required functions can be sistance to degradation due to aging in various per-
maintained depends on the performance of its structural formance characteristics and shall be included in per-
components and members. Therefore, the required func- formance characteristics related to structural safety, us-
tions must be related to the required performance char- ability, and shielding performance rather than as an in-
acteristics to define the levels of performance necessary dependent performance category.
to keep the structure functional. For fire resistance, performance characteristics, such
Table 1 shows the relation of the required functions as damage resistance, heat shielding performance, and
covered in the Guidelines to the performance character- flame shielding performance, and the specification code
istics of structural components and members. Supporting, for their satisfaction are stipulated in, e.g., the Structure
pressure-proofing, shielding, leak-proofing, nega- Standards Act and the Order for Enforcement of the same,
tive-pressure holding, and secondary-accident preven- the provisions of which shall be complied with for nu-
tion functions are defined as those required of structures clear facilities. In the Guidelines, fire damage resistance
at nuclear facilities. shall be considered as part of structural safety
Structural safety characteristics shall include (load-bearing capacity for special loads) while heat and
load-bearing capacity for general loads (structure’s dead flame shielding performance shall be considered as part
weight, live loads, seismic loads, wind loads, and snow of usability (protection of the internal environment
loads) taken into consideration in the design of a struc- against heat).
Y. Umeki, S. Sawada, S. Mitsugi, T. Maenaka and K. Takiguchi / Journal of Advanced Concrete Technology Vol. 14, 643-663, 2014 648
Thus, the functions required of structures at nuclear performance manifest itself, and that this shall be
facilities shall be related to structural safety, usability, achieved by repairs and other countermeasures. Ac-
and shielding performance in the Guidelines. cordingly, the Guidelines assume the following two
categories of maintenance:
3.5 Two maintenance methods ・Maintenance for ensuring current soundness
The maintenance in the Guidelines assumes that struc- In order to check structures for current soundness,
tures shall be checked for performance by inspection and their current condition shall be assessed by grasping
soundness evaluation and that they shall be kept in sound the presence/absence or degree of any degradation
condition without letting any noticeable decrease in phenomenon, and an appropriate countermeasure(s)
shall be taken.
Chapter 5 Preparation of maintenance plans ・Maintenance for ensuring long-term soundness
Evaluation of In order to ensure that structures’ functions will be
Maintenance planning validity of maintained into the future, the degrees of effect of
maintenance plan degradation factors shall be grasped before the oc-
currence of any degradation phenomenon. At the same
Chapter 6 Inspection time, the long-term effect shall be evaluated by a
Preparation of periodical and extra method such as progression prediction to take an ap-
inspection plans propriate countermeasure(s).
Maintenance for ensuring current soundness shall be
Implementation of periodical and practiced as the basic mode of maintenance and shall be
extra inspections for current
soundness evaluation supplemented by maintenance for ensuring long-term
(mainly visual method) soundness to make it possible to demonstrate that current
structural soundness is maintained and to ensure that no
Chapter 7 Soundness evaluation noticeable decrease in performance will occur during the
Current soundness evaluation rest of the in-service period. Table 2 compares the two
(Evaluation item: presence/absence or maintenance categories.
degree of degradation phenomenon)
(1) Maintenance for ensuring the current
Chapter 8 Countermeasures and verification of effectiveness soundness
The purpose of this maintenance is to ensure that current
Countermeasure implementation
structural soundness is maintained. Hence, this mainte-
and effectiveness verification
nance shall be performed relatively frequently and the
specific activities involved shall include investigating the
Chapter 9 Records presence/absence of degradation likely to lead to reduced
Recording and storage structural performance and taking an appropriate coun-
termeasure(s) in response to the evaluation results.
Figure 6 shows the flow of the specific maintenance
Fig. 6 Implementation flow for the maintenance for en-
activities to be implemented.
suring the current soundness.
Y. Umeki, S. Sawada, S. Mitsugi, T. Maenaka and K. Takiguchi / Journal of Advanced Concrete Technology Vol. 14, 643-663, 2014 649
In the Guidelines, the causes of a direct decrease in contributory to any degradation phenomenon that may
structural performance are regarded as degradation affect the structural safety and shielding performance
phenomena. that are critical for the safety of structures. Degradation
Degradation due to aging that occurs over time shall phenomena are expected to proceed at points likely to be
be investigated by periodical inspection for current significantly affected by degradation factors. Therefore,
soundness evaluation. Meanwhile, degradation due to by selecting such points as representative parts, the
earthquakes, typhoons, fires, and other sudden events number of inspection points can be reduced. It should be
shall be investigated by extra inspection. These degra- noted that degradation factors that affect the usability of
dation phenomena often occur in the form of damage structures should be deemed as having already been
such as cracking in concrete surfaces and coated surfaces, assessed because the related degradation phenomena are
except for some phenomena, such as reduced concrete checked in the maintenance for ensuring current sound-
strength, and can therefore be grasped by visual methods. ness.
This is performed to determine the presence/absence of The degrees of effect of degradation factors shall be
any current structural damage and hence in principle grasped based on data collected through inspection. At
shall be performed for structures in their entirety. the same time, a progression prediction shall be made by
Inspection results shall be checked against evaluation checking the predicted values against the evaluation
criteria to perform structure soundness evaluation and criteria to perform soundness evaluation and consider
consider countermeasures. If it is necessary to investigate countermeasures. After any repair is performed, its ef-
degradation factors contributory to any degradation phe- fectiveness shall be verified. At the same time, the results
nomenon in order to consider repair methods and other of inspection, soundness evaluation, and countermeasure
countermeasures, non-destructive, local destructive, and implementation and effectiveness verification shall be
other methods shall be used to collect detailed data. After recorded and stored to review the maintenance plan as
any repair, its effectiveness shall be verified. At the same necessary. Since this maintenance is intended for deg-
time, the results of inspection, soundness evaluation, and radation that does not proceed rapidly, the required in-
countermeasure implementation and effectiveness veri- spection and soundness evaluation may be performed at
fication shall be recorded and stored to review the intervals longer than in maintenance for ensuring current
maintenance plan as necessary. soundness. Note that use of longer intervals is effective
in minimizing damage to skeletons due to the local de-
(2) Maintenance for ensuring long-term structive method.
soundness
The purpose of this maintenance is to check structures
for soundness now and into the future. Even if no likely Chapter 5 Preparation of maintenance plans
Evaluation of
degradation phenomenon manifests itself now, the de- Maintenance planning validity of
grees of effect of degradation factors contributory to maintenance plan
degradation phenomena shall be grasped and specific
activities shall include predicting whether structural Chapter 6 Inspection
performance would be maintained above the required
level during the assumed period. Figure 7 shows the flow Preparation of periodical
inspection plans
of the specific maintenance activities to be implemented.
Since this maintenance is intended for degradation that Implementation of periodical inspection
proceeds over time, its scope does not include degrada- for long-term soundness evaluation
tion due to sudden event. Hence, necessary data shall be (mainly local destructive method)
collected by periodical inspection for long-term sound-
ness evaluation. This data collection shall precede the Chapter 7 Soundness evaluation
manifestation of any degradation phenomenon. Accord- Long-term soundness evaluation
ingly, a local destructive method may become necessary (Evaluation item: degrees of
rather than a visual method or a non-destructive method. effect of degradation factors)
The parts to be inspected must be specified in advance
and must be representative of each structure. Moreover, Chapter 8 Countermeasures and verification of effectiveness
from the perspective of the importance of structures at
nuclear facilities, precautions must be taken to avoid Countermeasure implementation
damaging the skeleton due to sampling of specimens. It and effectiveness verification
is desirable to minimize the number of inspection points.
This maintenance uses techniques, such as local de- Chapter 9 Records
structive methods that cause damage to skeletons.
Recording and storage
Therefore, this maintenance should not be performed for
all degradation factors. Instead, the degradation factors to
be monitored should be extracted with focus on those Fig. 7 Implementation flow for maintenance for ensuring
long-term soundness.
Y. Umeki, S. Sawada, S. Mitsugi, T. Maenaka and K. Takiguchi / Journal of Advanced Concrete Technology Vol. 14, 643-663, 2014 650
Collection and collation of existing knowledge on degradation phenomena and degradation factors
• Degradation phenomena and degradation factors
• Relation between performance characteristics and degradation phenomena
• Degradation mechanisms
Table 3 Degradation phenomena and degradation factors specified by existing guidelines to be monitored (for concrete
structures).
Degradation factor
(elevated temperatures)
Heat
Radiation exposure
Carbonation
Chloride penetration
Alkali-aggregate reaction
Machine vibration
Freeze-thaw
Chemical attack
Drying shrinkage
Creep
Insolation
Weathering
Differential settlement
Overloads
typhoons, etc.
Loads due to earthquakes,
Reduced concrete
3 3 3 3
strength
Cracking 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
Spalling and falling 3 3 3 3 3 3 3 3 3
Rebar corrosion 3 3 3 3 3 3
Deflection 3 3
Vibration 3 3
Abrasion 3
Explosive spalling 3
Water leakage 3
Moisture diffusion 3 3 3
Table 4 Degradation phenomena and degradation factors specified by existing guidelines to be monitored (for steel
framed structures).
Degradation factor
Plating degradation
Steel corrosion 3 3
Reduced strength due to heat 3
Yielding, buckling, crack, fracture, and excessive deformation of steel 3 3
Slipping, crack, and fracture of joint
3 3 3
Looseness and fracture of bolt
Reduced strength due to fatigue 3
Guidelines as a degradation phenomenon. are coating degradation and plating degradation (loss of
Relations between degradation phenomena and deg- galvanized plating). Cyclic loads due to wind, etc. (fa-
radation factors specified by existing guidelines are tigue) cause reduced strength due to fatigue, a degrada-
summarized for concrete and steel-framed structures as tion phenomenon.
shown in Tables 3 and 4. The selected degradation phenomena may decrease
Degradation factors leading to degradation phenom- the performance characteristics of structures at nuclear
ena of reduced concrete strength are heat (elevated facilities. Tables 5 and 6 show the degradation phe-
temperatures), radiation exposure, weathering, and heat nomena that may decrease performance characteristic
due to fire. Concrete cracking is regarded, as mentioned values (structural safety, usability, and shielding per-
above, as a degradation phenomenon. It has several formance) for concrete and steel-framed structure, re-
degradation factors including heat (elevated temperature), spectively. Degradation phenomena related to each of the
radiation exposure, carbonation, chloride penetration, performance characteristics are classified into degrada-
and machine vibration. tion due to aging and degradation due to sudden event, to
For steel-framed structures, the degradation factors be consistent with the maintenance plan, where earth-
leading to the degradation phenomenon of steel corrosion quake, typhoon, fire, etc. are assumed to cause degrada-
Y. Umeki, S. Sawada, S. Mitsugi, T. Maenaka and K. Takiguchi / Journal of Advanced Concrete Technology Vol. 14, 643-663, 2014 652
Table 5 Relation of performance characteristics of concrete structures at nuclear facilities to degradation phenomena.
Degradation phenomenon
Performance characteristic Degradation due to sudden event
Degradation due to aging
(earthquake, typhoon, fire, etc.)
Structural safety Reduced concrete strength Reduced concrete strength
Cracking Cracking
Rebar corrosion Spalling and falling
Spalling and falling Explosive spalling
Usability Cracking Cracking
Spalling and falling Spalling and falling
Deflection Explosive spalling
Vibration
Abrasion
Water leakage
Shielding performance Moisture diffusion Moisture diffusion
Cracking Cracking
Spalling and falling Spalling and falling
Explosive spalling
Table 6 Relation between performance characteristics and degradation phenomena in steel framed structures at nuclear
facilities.
Degradation phenomenon
Performance char-
Degradation due to sudden event
acteristic Degradation due to aging
(earthquake, typhoon, fire, etc.)
Structural safety Reduced strength due to heat
Steel corrosion
Yielding, buckling, crack, fracture, and excessive
Slipping, crack, and fracture of joint
deformation of steel
Looseness and fracture of bolt
Slipping, crack, and fracture of joint
Reduced strength due to fatigue
Looseness and fracture of bolt
Usability Yielding, buckling, crack, fracture, and excessive
Steel corrosion
deformation of steel
Slipping, crack, and fracture of joint
Slipping, crack, and fracture of joint
Looseness and fracture of bolt
Looseness and fracture of bolt
tion due to sudden event. cessive deformation of any steel member, slipping,
The relations between performance characteristics and cracking or fracturing of any joint, looseness or frac-
degradation phenomena are as follows: turing of any bolt, or any reduced strength due to fa-
a) The structural safety of a concrete structure may be tigue.
reduced due to cross-sectional area reductions and e) The usability of a steel-framed structure relates to,
cracks resulting from reduced concrete strength, rebar among other things, vibration isolation performance
corrosion, explosive spalling, or spalling and drop- against harmful vibration or resistance capacity
ping. against increases in deflection. These performance
b) Usability is a generic term for performance charac- characteristics may be affected by steel corrosion,
teristics required for comfortable use of structures and yielding, buckling, cracking, fracturing, or excessive
refers to, among other things, liquid-tightness, deformation of any steel member; slipping, cracking
air-tightness, resistance capacity to increases in de- or fracturing of any joint; or looseness or fracture of
flection, and vibration isolation performance against any bolt.
harmful vibration. Liquid-tightness and air-tightness It is important for selection of degradation phenomena
may be reduced due to cracking, while resistance ca- and degradation factors, evaluation, and progression
pacity against increases in deflection may be reduced prediction, to understand degradation mechanisms for
due to creep and overloads. each degradation factor. Degradation mechanisms are
c) Shielding performance relates to concrete density and summarized for each degradation factor specified by
member thickness. Density reductions due to moisture existing guidelines to be monitored. The degradation
diffusion, member thickness reductions due to explo- factors are shown in Tables 3 and 4 for concrete and
sive spalling or spalling and dropping, and excessive steel-framed structures, respectively.
cracking may result in decreases in shielding per-
formance. (3) Selection of degradation factors based on
d) Structural safety of a steel-framed structure may de- characteristics of nuclear facilities
crease due to steel corrosion, reduced strength due to Structures at nuclear facilities exist in a unique envi-
heat, yielding, buckling, cracking, fracturing or ex- ronment. Degradation mechanics need to be considered
Y. Umeki, S. Sawada, S. Mitsugi, T. Maenaka and K. Takiguchi / Journal of Advanced Concrete Technology Vol. 14, 643-663, 2014 653
when degradation factors are selected according to con- e) Concrete surface protection conditions
ditions such as environmental conditions and site condi- ・ Outdoor coverings in structures at newer nuclear
tions. Moreover, degradation mechanics are also impor- facilities are thicker than those in general structures
tant for evaluation of degradation progression processes and are protected with a finishing material such as a
and prediction of progression. coating.
Degradation factors to be monitored are selected ・ Moreover, many parts have indoor concrete surfaces
through summarization of characteristics of nuclear fa- coated for easier removal of radioactive materials
cilities. The following conditions of nuclear facilities are adhering to them (decontamination).
used as selection criteria: a) Environmental conditions, ・ Furthermore, parts in contact with liquids are cov-
b) Site conditions, c) Structural conditions, d) Material ered with liners or coating films.
conditions and e) Concrete surface protection conditions. For degradation factors selected according to mainte-
nance-related guidelines, degradation mechanisms are
a) Environmental conditions summarized, and ones to be monitored are selected based
・ Parts of structures in nuclear facilities are exposed to on the characteristics of nuclear facilities. Table 7 and
elevated temperatures. Table 8 show the selected degradation factors to be
・ Nuclear power plants use the nuclear fission reaction monitored for concrete and steel-framed structures at
of radioactive materials, and parts of their structures nuclear facilities.
are exposed to radiation. Chemical attack, differential settlement, and overloads
・ Many machines are installed in structures at nuclear are not selected as degradation factors for degradation
facilities, and many parts of structures are exposed due to aging of concrete structures. The reasons include
to vibration. the facts that no existing nuclear facilities are located in
・ Structures in nuclear facilities include vicinities of hot-spring areas or acidic river basins and
bulk-receiving docks with floors, etc., exposed to that differential settlement is unlikely because the struc-
loads due to inbound and outbound vehicular traffic tures at nuclear facilities are supported directly on bed-
and equipment deliveries. rock or indirectly via piles. Freeze-thaw is selected in
・ With fewer people inside than in general structures, areas with a minor or higher frost damage risk as a deg-
structures at nuclear facilities tend to be low in radation factor to be monitored only for usability.
carbon dioxide gas concentration. All the degradation factors for steel-framed structures
b) Site conditions are selected to be monitored.
・ Nuclear power plants in Japan are located in coastal
areas, and their facilities may be exposed to airborne (4) Summary of degradation phenomena and
sea salt particles. degradation factors to be monitored
・ Existing nuclear facilities are located in either Based on the results mentioned above, Tables 9 and 10
semi-frost damage prone areas or general frost show the relations between degradation phenomena and
damage prone areas with a minor or higher frost degradation factors to be monitored for concrete and
damage risk. steel-framed structures, respectively, at nuclear facilities.
・ No existing nuclear facilities are located in vicinities In the tables, degradation phenomena are grouped by
of hot-spring areas or acidic river basins. performance characteristics, and combinations with
・ Nuclear facilities are not different from general degradation factors are shown. Degradation due to aging
structures regarding factors (ultraviolet rays, heat, and degradation due to sudden event are shown in re-
moisture, blown sand, mechanical stress, etc.) spective tables.
leading to carbonation (outdoors), insolation,
weathering, drying, or coating degradation. 5. Inspection and soundness evaluation
c) Structural conditions
・ Many parts of steel reinforced concrete structures at Inspection and Soundness evaluation follows the PDCA
nuclear facilities have larger cross-sections and cycle (Plan, Do, Check and Action) and plans for im-
covering thicknesses than those in general struc- provement through the in-service period. Inspection and
tures. Soundness in the PDCA cycle includes collecting basic
・ In principle, structures at nuclear facilities are sup- information on maintenance, predicting soundness and
ported directly on a rock mass or indirectly via piles. contributing to Nuclear Safety.
・ Structures at nuclear facilities are equipped with Inspection should use a general way and the latest way.
many pieces of heavy equipment. Therefore, larger This section describes a basic way of thinking about the
live loads are taken into consideration than in gen- choice of an inspection method. The criteria in Sound-
eral structures. ness evaluation reflects the latest knowledge and a re-
d) Material conditions cord.
・ When structures at nuclear facilities are built, This section, in addition to the Inspection and
stringent control is in place for the selection and Soundness evaluation, also describes “Countermeasures
mixing of concrete materials. and verification of effectiveness” and “Records”.
Y. Umeki, S. Sawada, S. Mitsugi, T. Maenaka and K. Takiguchi / Journal of Advanced Concrete Technology Vol. 14, 643-663, 2014 654
Table 7 Selection results of degradation factors to be monitored for concrete structures at nuclear facilities (Legend: 3: To be
selected as degradation factor and −: Not to be selected as degradation factor)
Selection
Degradation factor Applicable part Remarks
results
Heat (elevated tem- Reactor body foundation,
3
peratures) internal concrete, etc.
Radiation exposure Reactor body foundation, shell
3
wall, and internal concrete
Carbonation Concrete in general • In applicable parts with coating, degradation progression is
3
inhibited.
Chloride penetration Buildings’ external wall • In applicable parts with coating, degradation progression is
3
surfaces inhibited.
Alkali-aggregate Concrete in general
3
reaction
Machine vibration Turbine generator pedestal,
emergency generator
3
foundations, pump and motor
foundations, etc.
Freeze-thaw Local parts around openings • In areas with a minor or higher frost damage risk, only usabil-
3*
and parapets on external walls ity shall be monitored.
Chemical attack • No existing nuclear facilities are located in vicinities of
− − hot-spring areas or acidic river basins.
• Indoor parts are unlikely to undergo chemical attack.
Drying shrinkage Walls, columns, beams, floors, • Structural safety and shielding performance do not seem
3 etc. likely to be compromised. Accordingly, only usability shall be
monitored.
Creep Beams and floors • The effect of creep has been taken into design consideration.
3
Accordingly, only usability shall be monitored.
Insolation Rooftop, etc. • The effect of temperature rise is minor even on a rooftop, with
a waterproof layer and its protective material provided on it.
3
Accordingly, only usability shall be monitored.
• The walls restraining a rooftop floor shall also be assessed.
Weathering Buildings’ external wall sur- • Considering the large concrete covering thicknesses in nuclear
faces facilities, weathering has a minor effect on the structural safety
3 and shielding performance of buildings. Accordingly, only
usability shall be monitored.
• Applicable parts with coating are less prone to degradation.
Vehicular traffic, etc. Floors, etc. • Degradation will be limited to the surface layers of applicable
3
parts. Accordingly, only usability shall be monitored.
Differential settle- • In principle, nuclear facilities are supported directly on a rock
ment − − mass or indirectly via piles on it and are unlikely to undergo
differential settlement.
Overloads • Prior to installing any heavy equipment, etc., exceeding the
load assumed at the design phase, an evaluation is performed
− −
by an architect(s)/engineer(s) as necessary to ensure the
soundness of structural components and that of members.
Loads due to earth- Walls, columns, beams, floors,
3
quakes, typhoons, etc. etc.
Heat due to fire 3 Parts in rooms on fire
*: Dependent on site conditions.
Table 8 Selection results of degradation factors to be monitored for steel-framed structures at nuclear facilities (Legend: 3:
To be selected as degradation factor).
Degradation factor Results of selection Typical applicable part Remarks
Coating degradation 3 Stacks and roof trusses
Plating degradation (loss of gal-
3 Stacks and roof trusses
vanized plating)
Cyclic loads due to winds, etc.
3 Stacks
(fatigue)
Loads due to earthquakes, ty-
3 Stacks and roof trusses
phoons, etc.
Heat due to fire 3 Stacks and roof trusses
Y. Umeki, S. Sawada, S. Mitsugi, T. Maenaka and K. Takiguchi / Journal of Advanced Concrete Technology Vol. 14, 643-663, 2014 655
Table 9(a) Summary of degradation phenomena and degradation factors in degradation due to aging of concrete
structures at nuclear facilities.
Degradation factor
peratures)
Heat (elevated tem-
Radiation exposure
Carbonation
Chloride penetration
action
Alkali-aggregate re-
Machine vibration
Freeze-thaw
Drying shrinkage
Creep
Insolation
Weathering
Table 9(b) Summary of degradation phenomena and degradation factors in degradation due to sudden event of concrete
structures at nuclear facilities.
Table 10(a) Summary of degradation phenomena and degradation factors in degradation due to aging of steel-framed
structures at nuclear facilities.
Degradation factor
Coating degradation
Table 10(b) Summary of degradation phenomena and degradation factors in degradation due to sudden event of
steel-framed structures at nuclear facilities.
Performance Degradation factor
Degradation phenomenon
characteristic Loads due to earthquakes, typhoons, etc. Heat due to fire
Structural safety Reduced strength due to heat 3
Yielding, buckling, crack, fracture, and
3 3
excessive deformation of steel
Slipping, crack, and fracture of joint
3 3
Looseness and fracture of bolt
Usability Yielding, buckling, crack, fracture, and
3 3
excessive deformation of steel
Slipping, crack, and fracture of joint
3 3
Looseness and fracture of bolt
concrete structure are indicated in Table 13. non in a concrete structure often manifests itself as sur-
When the relevant part is difficult to access and inspect face damage and therefore can be grasped by a visual
by a direct method due to the problem of high altitude or method in extra inspections, similar to periodical in-
equipment layout, an indirect method shall be used. spections. If damage is detected by a visual method, a
more detailed inspection shall be performed by the
b) Steel-framed structures method shown in Table 15 to grasp effects on related
Periodical inspections for long-term soundness evalua- performance characteristics.
tion of a steel-framed structure are indicated in Table 14.
In periodical inspections for long-term soundness b) Steel-framed structures
evaluation of steel-framed structures, points at which a Extra inspections of a steel-framed structure shall be
relatively large stress occurs and points that may be performed to check the structure in its entirety for any
affected by fatigue under cyclic loads due to winds, etc. degradation phenomenon for loads due to earthquake,
shall be checked by a visual method or a non-destructive typhoon, etc., or to check around a fire-damaged point
method. for any degradation phenomenon. A degradation phe-
nomenon in a steel-framed structure often manifests
4) Extra inspection itself as surface damage and therefore can be grasped by
a) Concrete structures a visual method in extra inspections, similar to periodical
Extra inspections of a concrete structure shall be per- inspections. If any damage is detected by a visual method,
formed to check the structure in its entirety for any deg- a more detailed inspection shall be performed by the
radation phenomenon for loads due to earthquake, ty- method shown in Table 16 to grasp the effects on the
phoon, etc., or to check around a fire-damaged point for related performance characteristics.
any degradation phenomenon. A degradation phenome-
Y. Umeki, S. Sawada, S. Mitsugi, T. Maenaka and K. Takiguchi / Journal of Advanced Concrete Technology Vol. 14, 643-663, 2014 659
tion results meet the evaluation criteria, and A3 (To be cussed). Maintenance flow for long-term soundness
discussed) when the inspection results fail to meet the evaluation is indicated in Fig. 11.
evaluation criteria. B1 (Sound) shall be given when the results of a pro-
The evaluation criteria shall be set based on, among gression prediction meet the evaluation criteria. B2
other things, existing guidelines, latest knowledge, and (Follow-up observation) shall be given when the results
evidential data supported by actual measurement results, of an inspection meet the evaluation criteria at the mo-
from the perspective of securing the level of performance ment but are unlikely to meet the evaluation criteria
required to maintain the necessary function(s). during the in-service period. B3 (To be discussed) shall
Evaluation ratings and evaluation criteria for current be given when the results of an inspection and the like at
soundness evaluation are indicated in Table 17. Evalua- the moment do not already meet the evaluation criteria.
tion ratings and evaluation criteria for cracking are in- Progression predictions can be made using inspection
dicated in Table 18. results or existing formulae for prediction of progression.
Evaluation ratings are judged by the presence of deg- As for degradation factors for which a formula for pre-
radation phenomena. Especially, evaluation ratings and diction of progression is available, it is possible to predict
evaluation criteria for cracking are based on degradation the progression of degradation during the in-service
factors. period based on the applicable formula. It is also possible
to predict the tendency of progression of degradation
c. Evaluation of long-term soundness from the results of inspections performed in the past.
In a long-term soundness evaluation, the results of a Formulae have been proposed for predicting progres-
progression prediction shall be checked against a set of sion for carbonation, chloride penetration, and cyclic
predetermined evaluation criteria and rated as B1 loads due to winds, etc. (fatigue). When there are nec-
(Sound), B2 (Follow-up required), or B3 (To be dis- essary data available, it is possible to make a progression
A1 A2 A3
Sound Follow-up To be
required discussed
B1 B2 B3
Sound Follow-up To be
required discussed
Table 17 Evaluation ratings and evaluation criteria for current soundness evaluation.
Evaluation rating and evaluation criteria
Degradation phenomena A2 (Follow-up obser-
A1 (Sound) A3 (To be discussed)
vation)
compressive concrete compressive concrete
Reduced concrete strength strength is equal to or - strength is below the design
above the design strength strength
Moisture diffusion a soundness evaluation
Cracking Refer to Table 18
No rebar corrosion Any rebar corrosion (rust
Rebar corrosion -
(rust fluid) fluid)
Spalling and falling No spalling and falling - Any spalling and falling
Concrete No deflection - Any deflection
structures Deflection deflection span ratio is less deflection span ratio is 1/300
-
than 1/300 or more
Vibration No vibration - Any vibration
No exposed coarse Any exposed coarse aggre-
Abrasion -
aggregate gate
No water leakage or water Any water leakage or water
Water leakage -
leak mark leak mark
Degradation phenomena due to
No effect of the fire - Any effect of the fire
fire
No damage on coating or Any damage on coating or
Reduced strength due to fatigue -
plating plating
Any damage (tiny rust
spots on all over the Any steel member is cor-
little rust or few coating
Steel corrosion surface , large rust roded and reduced in
plating quality defects
spots) on coating or cross-sectional area
plating
Steel-framed
Yielding, buckling, crack,
structures
fracture, or excessive
deformation of steel members No damage on the coating Any damage on the coating
-
Yielding, buckling, crack, or plating or plating
fracture, or excessive
deformation of steel members
Degradation phenomena due to
No effect of the fire - Any effect of the fire
fire
prediction. Note, however, that coefficients or the like etc. fatigue. Therefore, such a prediction shall be based
used in an existing formula for prediction of progression on the results of an inspection by the visual method on
are often formulated based on the results of accelerated the condition of the coating or plating and the pres-
degradation of specimens. Hence, the results of a pre- ence/absence of excessive deformation. The evaluation
diction on an actual structure may differ from those of its ratings shall be in line with current soundness evaluation.
inspection. Therefore, a formula for predicting progres- The rating shall be B1 (Sound) for damage-free coating
sion must be checked in advance for its applicability. or plating and shall be B3 (To be discussed) when any
With a collation of inspection results accumulated coating/plating quality defect or any excessive deforma-
over years, even in the absence of a formula for predic- tion is detected by the visual method.
tion of progression, it is possible to predict the tendency
of the effect due to the degradation factor. Generally, (3) Countermeasures and verification of effec-
degradation due to aging does not proceed rapidly. tiveness
Therefore, if the results of inspections performed in the To maintain the functions required of nuclear facilities, it
past reveal that no degradation phenomenon affecting is necessary to implement countermeasures based on the
performance has occurred, the risk of rapid performance results of current or long-term soundness evaluations so
degradation in future shall be estimated as low. Even if that the current performance of the structures during the
there is a paucity of accumulated inspection results, this in-service period will exceed the required level. Figure
prediction method will be effective if the future period to 12 shows the flow of maintenance countermeasure im-
be covered is relatively short. plementation and effectiveness verifications.
Evaluation criteria for degradation factors are indi- Methods of implementing countermeasures shall in-
cated in Table 19. It is difficult to make a direct predic- clude continuation of the current inspection, enhance-
tion of the progression of any effect due to machine ment of inspections, and repairs. Repairs shall include
vibration and alkali aggregate reaction. Therefore, such a countermeasures implemented alternatively, such as
prediction shall be based on the crack occurrence situa- consideration for members or structural components,
tion. Accordingly, in a soundness evaluation for machine when repairs are unavailable, or when the required per-
vibration and alkali aggregate reaction, A1 (Sound) and formance level cannot be achieved by repairs. Any al-
A3 (To be discussed), the evaluation ratings for cracks ternative countermeasure implemented shall be checked
affecting structural safety, shall be read as B1 (Sound) for its effectiveness.
and B3 (To be discussed), respectively.
In steel-framed structures, the condition of coating or (4) Records
plating is checked by periodical inspections for current For continuous implementation of maintenance, the
soundness evaluation and mainly by the visual method. following items shall be recorded:
The evaluation ratings for coating degradation and plat- ・ Maintenance plan
ing degradation (loss of galvanized plating) in the present ・ Contents and results of inspections
Guidelines shall be B1 (Sound) for damage-free coating ・ Methods and results of soundness evaluations
or plating and B3 (To be discussed) for any coat- ・ Contents and results of countermeasures and methods
ing/plating quality defect. To ensure long-term preven- and results of effectiveness verifications
tion of steel member corrosion, it is important to main- To enhance the reliability and effectiveness of main-
tain the soundness of the coating or plating. The need for tenance, it is necessary to record and store information on
periodic recoating shall be determined based on useful maintenance while analyzing records and continuously
life evaluation. making improvements.
It is difficult to make a direct prediction of the pro- The results of inspections, soundness evaluations, and
gression of any effect due to cyclic loads due to winds, countermeasures and effectiveness verifications shall be
Y. Umeki, S. Sawada, S. Mitsugi, T. Maenaka and K. Takiguchi / Journal of Advanced Concrete Technology Vol. 14, 643-663, 2014 663
A1 or B1 A2 or B2 A3 or B3
Sound Follow-up required To be discussed
Chapter 8
recorded accurately and in detail, using predetermined buildings at nuclear facilities has been left to the indi-
forms, to allow easy understanding of maintenance ac- vidual operators in the absence of standardized criteria
tivities. Even if the methods of recording and storage are and guidelines.
changed due to improvements in the maintenance Under such circumstances, the previous version of the
method and technological advancements, arrangements present Guidelines was established in 2008 as a set of
shall be made to allow continued use of past records. guiding principles that set standards for maintenance
When records are electronically stored, data leakage or work to be performed to maintain the functions required
loss shall be borne in mind. As for data handling, it is of buildings in nuclear power plants over a long period.
recommended to specify an appropriate control method In addition, the Guidelines were revised to reflect the
and implement measures for ensuring information secu- latest knowledge in 2015. Since then, the present
rity. Guidelines have been useful in a wide range of mainte-
nance activities from daily inspection to building
6.Conclusion soundness evaluation with focus on aging.
This paper has outlined the Guidelines applied to nu-
Safe long-term use of nuclear facilities depends heavily clear facilities in Japan. An English version of the
on proper design, construction, and maintenance. Guidelines will be released by the Architectural Institute
There are an increasing number of nuclear power of Japan.
plants in service for more than 30 years since the start of
operation. Accordingly, the importance of maintenance Reference
of nuclear facilities is increasing, in addition to the tra- AIJ, (2015). “Guidelines for maintenance and manage-
ditionally recognized importance of their design and ment of structures in nuclear facilities.” Architectural
construction. So far, however, the maintenance of Institute of Japan, Tokyo. (in Japanese)
Journal of Advanced Concrete Technology Vol. 15, 1-9, January 2017 / Copyright © 2017 Japan Concrete Institute 1
Technical report
Abstract
Nuclear concrete structures are typically passive under normal operating conditions. They are, however, play a key role
in mitigating impacts of extreme/abnormal operating and environmental events. As structures age, changes in material
properties arise from continuing microstructural changes and environmental influences. Changes in environmental con-
ditions anticipated during the original design can sometimes lead to unpredicted effects, which may jeopardize the in-
tegrity of the structures.
In order to continue successful operation, continuous integrity of the safety-related concrete structures shall be ensured
by controlling and mitigating aging related degradation. A key element of aging management is the systematic and rig-
orous assessment of structures most commonly referred to as Condition Assessment (CA). Over the years, SNC-Lavalin
personnel have assessed containment structures and safety related structures of Nuclear Power Plants (NPPs), research
reactor and waste storage facilities.
This paper discusses requirements and approach to aging management for existing as well as new NPP concrete struc-
tures, and presents the methodology used to perform the CA.
Standard (2015) was issued in 2015. The standard pro- programs for concrete structures refer to IAEA TEC-
vides aging management requirements to ensure that DOC 1025 (1998) issued in 1998 to address assessment
concrete containment structures satisfy their functional and aging management of concrete containment struc-
and performance requirements in all different phases of tures. The TECDOC was revised recently as discussed
their life-cycle including design, construction, commis- in Moore et al. (2015) to include state of the art infor-
sioning, operation, and decommissioning. Moreover, the mation regarding aging management of concrete struc-
requirements provided in this standard are intended to tures. The document was also expanded to address other
ensure that for new concrete containment structures, NPP concrete structures besides containment. The re-
aging is addressed during design and construction vised document (IAEA 2016) was issued in 2016.
phases.
The International Generic Aging Lessons Learned 3. Effective aging management of existing
(IGALL) project was initiated by IAEA a few years ago structures
in order to provide detailed information on specific pro-
grams to manage aging and degradation of SSCs. This The documents listed in Section 2 require a systematic
initiative followed the United States Nuclear Regulatory approach to aging management to be implemented as
Commission (USNRC) Generic Aging Lessons Learned illustrated in Fig. 1. As can be seen from this figure,
(GALL) activity (USNRC 2010) but was extended to understanding aging lies in the heart of the effective
summarize international practices for aging manage- aging management.
ment of mechanical, electrical and civil SSCs of NPPs. Other components of effective aging management in-
Phase 2 of the project completed in 2015 provides clude development of the Aging Management (AM)
collection of proven aging management programs and plan, operating to minimize expected degradation, per-
Time Limited Aging Analysis (TLAAs). Phase 3 of the forming inspections and monitoring activities to check
project is currently underway. The aging management for degradation, and implementing mitigative measures.
The lessons learned in implementation of AM plan monly the process is referred to as Condition Assess-
along with the applicable Operating Experience (OPEX) ment). Figure 2 illustrates schematics of the process.
and Research and Development (R&D) activities should The methodology was specifically tailored to civil
be considered for optimisation of the AM plan. The ma- structures and includes the following:
jor components of effective aging management as it
pertains to existing nuclear concrete structures are dis- • Gathering and reviewing of design, manufacturing,
cussed in this Section. construction, commissioning, operation, inspection,
and maintenance history, and any other relevant
3.1 Understanding aging and planning aging documentation;
management plan • Defining the design basis and identifying any changes
Prior to the development of an AM plan, evaluation for made to the structure during construction and opera-
aging management and condition assessment shall be tion;
performed in order to ensure understanding of aging and • Establishing the structure's physical and functional
to establish base line condition of the structure. Re- boundaries and associated components;
quirements for condition assessment and evaluation for • Undertaking an aging assessment of structures or
aging management are defined by CSA N287.8 (CSA components remaining after screening. This includes
2015). an assessment of ARDMs, aging evaluation, and es-
tablishing the structure’s health prognosis; and
3.1.1 Methodology for evaluation for aging • Providing recommendations for a structure in terms of
management and condition assessment inspections, monitoring, repair, and maintenance ac-
As stated in IAEA Safety Guide NS-G-2.12 (IAEA tivities to ensure plant safety and production goals
2009), an aging management review (or an evaluation over the plant life.
for aging management) forms an integral part of a CA. In addition to current condition and future health pre-
Thus, the condition assessment and evaluation for aging dictions, changes necessary to address issues related to
management are typically performed in combination aging effects are identified as a result of CA; economic
and constitute Life Assessment (although, more com- improvement opportunities might also be included.
1
Fig. 2 Schematic Condition Assessment Process (Tcherner and Aziz 2013)
1
The term Condition Assessment, as more commonly used, refers to Life Assessment as well.
J. Tcherner, L. Vaithilingam and M. Han / Journal of Advanced Concrete Technology Vol. 15, 1-9, 2017 4
The comprehensive data is required to ensure the to prevent forced outages, etc.;
depth and quality of the CA. Since the results of CA are • Medium means the degradation mechanism is known
largely based on the engineering judgement, qualified for this structure or component, either at this facility
and experienced personnel shall be performing the CAs. or a similar one, and is being managed or mitigated;
Engineering judgment comprises of the knowledge of and
the design and functional requirements of the structure • Low means the mechanism is possible for the struc-
being assessed as well as the understanding of how the ture, but is easily managed with the current programs
stressors can influence the onset and development of or has no impact on achieving target life.
degradation mechanisms and how these mechanisms ARDMs not included in the matrix, or those having
may affect ability of the structure to meet its functional no marks against them, are considered unlikely for the
requirements. While there are standards and guidelines structure given the environmental conditions or materi-
documents available to evaluate influence of various als or a combination thereof and are not considered fur-
parameters on performance of structures e.g. (CSA ther.
2015; IAEA 2016), the value of experience should not Conclusions and health prognoses are provided based
be underestimated. on a thorough review of available information with par-
Interviews with the station staff and contractors in- ticular attention to aging related degradation. Where
volved in design, construction, commissioning and op- sufficient information to provide a health prognosis is
eration of the structure as well as the walkdowns are not available, activities to support the CA results are
required in order to identify information not formally recommended. Additionally, recommendations related
documented and locations of deterioration of the struc- to existing plant programs are provided to ensure con-
ture. tinuing structure health.
Based on a review of available information, design AM plan can then be developed based on the under-
and operational exposure conditions of a structure are standing of the design, construction, material properties,
also compared and engineering judgement is used to operating conditions, aging mechanisms and stressors,
document all potential aging mechanisms that might consequences of degradation, inspection, operation and
influence performance of the structure being assessed. maintenance history and taking into consideration rec-
Evaluation process shown in Fig. 3 is used to assess and ommendations for inspection, monitoring and mitigative
prioritize ARDMs. An ARDM matrix is then created to activates. Refer to CSA N287.8 (2015) for contents of
record evaluation results. A structure’s susceptibility to an effective AM plan.
each ARDM is evaluated as high, medium, or low,
based on the following criteria: 3.1.2 Implementation example
• High means the degradation mechanism is occurring Table 1 contains example of ARDM matrix created for
or has occurred, either in this structure or in a similar a safety related reinforced concrete structure with a non-
structure under similar conditions, and that steps have metallic liner containing demineralised water and ex-
not been taken, or it is unclear if any such steps are posed to elevated temperatures and radiation. Based on
adequate to mitigate degradation for the target life or materials and environment of exposure for the structure,
Hydrogen Embrittlement
Freeze-Thaw Attack
General Corrosion
Fatigue/Vibration
Acid/Base Attack
Creep/Shrinkage
Sulphate Attack
Delamination
Carbonation
Weathering
Settlement
Irradiation
Wear
Concrete M M L M
Reinforcing Steel M
Embedded Parts M
Joint Sealant M M M M
Non-metallic Liner M L L L M
J. Tcherner, L. Vaithilingam and M. Han / Journal of Advanced Concrete Technology Vol. 15, 1-9, 2017 5
the following ARDMs were considered potentially rele- gate Reaction (AAR), elevated temperatures/thermal
vant: cycling, and shrinkage;
• For concrete – leaching/efflorescence, Alkali Aggre- • For reinforcing steel – corrosion;
• For embedded parts – corrosion; tion may not necessarily be manifesting itself on the
• For joint sealant – elevated temperature/thermal cy- surface of the structure. Therefore, monitoring instru-
cling, irradiation, weathering, and wear. mentation is often relied upon to provide assurance with
• For non-metallic liner - elevated temperature/thermal the integrity of the structure and to confirm its predicted
cycling, irradiation, weathering, and wear. response to the applied loading (e.g. during containment
Each ARDM listed above was systematically evalu- leakage rate test).
ated following the process illustrated in Fig. 3. For ex- Monitoring of environmental and operating condition
ample, evidences of efflorescence and leaching were enables assessing and trending the degradation stressors.
observed on the outside (dry) surface of the structure. While it is imperative for inaccessible structures to en-
There was no program in place to detect and manage sure that design assumptions are valid, it has also
leaching. Leaching increases porosity and the perme- proven to be useful to monitor environment of exposure
ability of concrete, thus lowering its strength and mak- of accessible structures particularly in between periodic
ing concrete more vulnerable to aggressive environ- inspections. The change in exposure environment might
ments. However, since there were no evidences of a facilitate existing degradation mechanisms and may
widespread degradation, this was considered a local cause origination of the new ones, i.e. the aging man-
phenomenon that was not likely to affect facility target agement activities may need to be refocused on other
life. Therefore, medium ranking was given. It should be mechanisms than those identified as per the original
noted however that as part of the life assessment, rec- condition assessment. Depending on the structure, its
ommendations are provided including those related to location, geometry, function and possible degradation
enhancements to the facility’s aging management pro- mechanisms, monitoring of the following parameters
gram. Thus, although it is stated that the mechanism is has proven to be simple (as far as monitoring instrumen-
not expected to affect facility’s target life, this may tation involved) and informative: ambient and operation
change in the long term as local phenomenon may be- temperature and humidity, chemistry of the ground wa-
come a widespread phenomenon if aging management ter and soil, elevation of the ground water table and its
is not implemented. fluctuations, monitoring chemistry, temperature, level of
As far as corrosion of reinforcement and embedded the liquid inside the structure (for liquid retaining struc-
parts is concerned, there were no evidences of corrosion tures) and the amount of a make-up water.
damage observed during inspection of the structure. It is As aging is a dynamic process, it is important to trend
however considered possible given the material and performance of the structure in order to establish the
environment of exposure including exposure of concrete rate of degradation. Experience shows that in order to
to demineralised water (discontinuities in the liner were track performance of the structure, condition indicators
evidenced by observed efflorescence) and elevated tem- – a characteristic of the structure that may infer or di-
peratures. Leaching may cause local reduction in alka- rectly indicate the current and future ability of the struc-
linity of concrete that might lead to corrosion of rein- ture to function within acceptance criteria. Examples of
forcement and embedded parts. Furthermore, elevated condition indicators include cracks, spalls, delaminations,
temperature tends to accelerate the rate of corrosion deflections, ingress of harmful substances (e.g. chlorides),
reaction. Corrosion has a potential to impact facility’s depth of carbonation, leak rate, prestressing force).
target life, therefore medium ranking is given. Furthermore, environmental and operating conditions
(e.g. temperature, humidity, radiation, chemistry of the
3.2 Minimizing expected degradation water and soil that the structure is exposed to) should be
In order to minimize expected degradation, the opera- trended and any excursions from the parameters consid-
tors should understand the importance of aging man- ered in the design should be noted to establish the rela-
agement and be involved in the implementation of AM tionship between degradation and associated stressors.
plan. Operating conditions that affect aging assumptions Predictions for future performance of the structure may
should be monitored and trended to ensure that the as- then be made based on the analysis of past performance,
sumptions made during design are valid. Any excursions the rate of degradation and the changes in associated
or events causing the change in operating conditions stressors. Knowledge of the rate of degradation and in-
should be evaluated and their effects on aging of con- fluencing parameters will also allow selecting appropri-
crete structures should be assessed. ate mitigative measures if necessary.
Aging must be considered when implementing any
change in operations or modifications to the structure or 3.4 Mitigating degradation
its components. Maintenance activities may be required to mitigate deg-
radation. These might include preventative maintenance,
3.3 Checking the degradation corrective maintenance, spare parts management, and
Periodic visual inspections are typically used to ensure replacement. Typically, corrective maintenance is used
adequate performance of the structures. However, nu- for nuclear concrete structures.
clear concrete structures often have limited accessibility OPEX shows that majority of the issues with concrete
for visual inspection. In addition, the internal degrada- structures are either directly caused by or facilitated by
J. Tcherner, L. Vaithilingam and M. Han / Journal of Advanced Concrete Technology Vol. 15, 1-9, 2017 7
the inadequate design and construction practices (refer formance in similar application and environment should
to Fig. 4). The design challenges include: be specified. The pre-qualification tests should be per-
- Inadequate structural design; formed to ensure proper application of the specified
- Poor detailing; material in the anticipated conditions. Mock-up tests
- Improper material selections; and should be considered for first-of-a-kind repair activities.
- Inadequate specifications. Additional information related to waterproofing of nu-
Construction issues include: clear concrete structures is provided (Tcherner 2015b).
- Not following specifications/ procedures;
- Lack of attention to details; and 3.5 Optimising aging management
- Poor QC / QA. As illustrated in Fig. 6, the inputs for optimisation of the
This should be considered when designing and im- AM plan include the lessons learned from implementa-
plementing repairs to mitigate the effects of degradation. tion of the AM Plan including operating, inspection and
Specifically, the importance of adequate technical speci- mitigation activities.
fication cannot be overemphasized. Although the indus- Repeating evaluation for aging management and con-
try is moving towards performance based specification, dition assessment would determine if there are any
specifications for repair of nuclear concrete structures changes in the condition of the structure that may neces-
tend to contain some prescriptive requirements as well sitate changes to operation, inspection/monitoring and
to ensure that specific restrictions as required by the maintenance activities. It would also help ensuring that
standards or dictated by the OPEX are incorporated. operating and environmental conditions that were ac-
Figure 5 illustrates the list of necessary inputs and out- counted for in design have not changed. As aging is a
puts of the specification including both performance dynamic process, the benefits of trending the stressors
based and prescriptive requirements. as well as the performance of the structures to improve
Inadequate repairs may facilitate degradation of the understanding of aging as discussed in Section 3.3 can-
structure and are often impractical to mitigate. There- not be overemphasised. This information needs to be
fore in order to ensure satisfactory and long-lasting re- addressed when optimising AM plan.
pair, the importance should be placed on the pre- Furthermore, it is important to consider applicable
qualification tests as opposed to the verification tests. OPEX and R&D activities thus help addressing any
Materials qualified by tests and satisfactory field per- “surprises” that tend to manifest themselves more often
Fig. 5 Technical Specification for Repair – Performance Based and Prescription Based (Tcherner 2015b).
J. Tcherner, L. Vaithilingam and M. Han / Journal of Advanced Concrete Technology Vol. 15, 1-9, 2017 8
as the knowledge and experience in NPP operation and their effect should be evaluated and analyses need to be
nuclear concrete structures’ performance grows. updated so that the structures can perform their func-
Thus, optimising includes improvements/revisions to tions safely. Material properties used for analysis should
AM Plan addressing the following: consider aging of the structure. Specifically for concrete
• Evaluation of changes in performance of the structure structures, samples should be obtained if possible and
via monitoring and trending of condition indicators; tested to determine actual properties that account for
• Evaluation of changes in the environmental and oper- aging. Concrete, unless subjected to degradation is
ating conditions via monitoring and trending of the known to gain strength with age. This should be con-
stressors and comparing them to those considered in firmed by physical testing and should be accounted for
the design; when analysis of mature structures are performed.
• Evaluation for aging management and condition as- An evaluation for aging management and condition
sessment (i.e. life assessment); assessment for nuclear safety-related structures per-
• Relevant OPEX including OPEX with implementa- formed as part of periodic safety review, in preparation
tion of AM Plan as well as OPEX with performance of life extensions, to establish a baseline condition for
of other structures subjected to similar environmental periodic inspection, prior to modification of the struc-
and operating conditions; and ture, or for other purposes provide an effective tool in
• Relevant R&D. assessing aging of the structures, evaluating their cur-
rent condition, providing a health prognosis, recom-
4. Conclusions mending remedial measures and identifying opportuni-
ties for improvements including design, operation and
OPEX indicates that concrete structures have a history aging management.
of reliability and durability, but there have been occur- It is important to consider instrumentation as well as
rences of degradation. As OPEX increases, more “sur- ability to inspect and monitor the structures during de-
prises” are to be expected. Degree of degradation de- sign and construction. Concrete containment structures
pends on quality of the structure as constructed and the particularly those with grouted post-tensioning system
aggressiveness of environment. Analysis and design should be monitored to ensure that they will continue
assumptions should be confirmed during the life of the meeting their functional requirements and have ade-
structure to account for changes in operating environ- quate margins available during the life of the plant (in-
ment and material properties. If changes are identified cluding life extensions). Monitoring of the stressors
(such as higher operating temperatures, greater loading), (environmental and operating conditions) and perform-
J. Tcherner, L. Vaithilingam and M. Han / Journal of Advanced Concrete Technology Vol. 15, 1-9, 2017 9
Scientific paper
Abstract
A change in distance between the generator rotor axis and the bearing of the table deck in the turbine generator (TG)
foundation of Unit #1 of the Ikata nuclear power plant was observed during an inspection in 1979 after the start of op-
eration. This led to measurements of the structure expansion (from 1981), the cracking condition (from 1982), and core
sampling of the concrete member (from 1986), based on which deformation and cracking was determined to be the result
of an alkali–silica reaction (ASR). However, as the strength of the TG foundation was still in excess of its design value,
ASR was not considered to have affected its ability to support equipment. Subsequent analysis of the structural stability of
the TG foundation found no change in deformation or concrete strength over a period of continued monitoring, and hence,
the TG foundation has been used up to the present time. The aim of this paper is to establish a maintenance management
method for similar structures exhibiting cracks caused by ASR.
3. Maintenance management
1
Construction Project Group, Civil & Architectural
3.1 Concept
Engineering Dept., Shikoku Electric Power Co., Inc.,
The current condition of the TG foundation was assessed
Japan. *Corresponding author,
(to understand the extent of ASR) through measurements
E-mail: manabe16561@yonden.co.jp
2 of its deformation, the mechanical properties of the RC
Subleader of Construction Project Group, Civil &
member, and a structural analysis. These results have
Architectural Engineering Dept., Shikoku Electric Power
suggested that its support function can be maintained;
Co., Inc., Japan.
3 however, a long-term evaluation has also been imple-
Leader of Construction Project Group, Civil &
mented that will continue these measurements so as to
Architectural Engineering Dept., Shikoku Electric Power
predict the future should any change in the ASR process
Co., Inc., Japan.
4 occur (Table 1, Fig. 2).
Leader of Equipment and technology Group (PLM),
Nuclear Power Dept., Shikoku Electric Power Co., Inc.,
Japan.
R. Manabe, H. Kawae, K. Ogawa and M. Matsuura / Journal of Advanced Concrete Technology Vol. 14, 590-606, 2016 591
3.2 History change in the distance between the generator rotor axis
A change in the distance between the generator rotor axis and bearing. Observation of these cracks and measure-
and bearing was discovered in the table deck during an ment of the deformation of the table deck therefore
equipment overhaul in 1979, but no abnormality was commenced. In order to clarify this, strain gauges were
found in the axis vibration and the deformation was able installed on the TG foundation in 1986 to measure the
to be corrected by adjusting the equipment. Cracks later strain in the reinforcing bars. Concrete core samples
appeared on the lateral surface of the beam of the table were also taken and subjected to compressive strength
deck in 1982, and it is thought that these may have been testing. These studies concluded that the deformation of
caused by deformation of the concrete resulting in the the TG foundation was caused by ASR based on the
Sea Side
50
Low-Pressure Steam
11
Axle High-Pressure Housing
Generator
2.5
No.2 Low-
Pressure Turbine
No.1 Low-
4.0 Pressure Turbine
3.0 15
2.5 High-Pressure High-Pressure
3.0 Turbine Steam
2.5 Foundation Table
5.0 Deck
2.5
4.76
2.5 18.7
Mountain Side
Reinforcing bar
・ Deformed bar SD345 (Former JIS SD 35) Unit : m
Concrete
・ Specified Strength 2
: 20.6 N/mm (210 kg/cm )
2 ・ Length : 50.0 m
・ Cement : moderate heat portland cement ・ Width (Turbine Side) : 15.0 m
・ Fine Aggregate : washed sea sand ・ Width (Generator Side) : 11.0 m
: Ehime-Pref. Imabari-City ・ Height : 18.7 m
・ Coarse Aggregate : crashed stone (from top of foundation mat)
: Yamaguchi-Pref. Yanai-City Andesite
Fig. 1 Schematic of turbine generators and their foundation.
R. Manabe, H. Kawae, K. Ogawa and M. Matsuura / Journal of Advanced Concrete Technology Vol. 14, 590-606, 2016 592
developing table deck expansion, distribution of concrete might affect its operation. Furthermore, the compressive
cracks with directionality along the main reinforcement strength of the concrete has remained in excess of its
direction, a decrease in the static elastic modulus, and the design strength and ongoing structural analysis com-
fact that andesite containing silicate mineral was used as menced in 1988 showed enough supporting function.
a coarse aggregate. A gelled substance extracted from the As ASR is a progressive process, it is necessary to
surface of the core samples was identified as alkaline evaluate the long-term soundness of the TG foundation
silica gel, with the alkali content of the concrete as much based on the measurements that have been taken up to the
as 9 kg/m3. Despite this, it was concluded that there was present time. In 2002, the development of ASR was
no problem in terms of the TG foundation providing verified to have settled based on the fact that table deck
sufficient support, as there was no induced vibration that expansion stopped around 1991, no new cracks indica-
Cracks appeared
on the lateral surface
of the beam of table deck
(1982)
Start operation judging that ASR settled
(1977) (2002) year
(1988) (2005)
structuralanalyses
tive of ASR were observed, and that the rate of remaining to understand the current ASR conditions and determine
expansion had been decreasing since 1989. The structure whether they had settled, as well as the results of a sub-
was therefore considered to have a low possibility of sequent structural analysis. Chapter 5 outlines the state
expansion in the future (Takakura and Ishikawa et al. monitoring that has been continued or added since de-
2005). Also in 2005, structural analysis by a termining that the ASR conditions had settled.
three-dimensional frame model of the TG foundation
deformation that took into account the mechanical 4. Soundness evaluation
properties of its concrete verified that its supporting
function would be maintained. Structural analysis as- 4.1 Analysis of ASR condition
suming redevelopment of ASR was carried out to show 4.1.1 Table deck expansion
enough supporting function in future. Since then, no Expansion of the table deck was measured to determine
major changes in the deformation conditions or com- the extent of any deformation of the TG foundation after
pressive strength of the concrete that would indicate a measuring the change of the distance between the gen-
change in ASR conditions have been observed through erator rotor axis and its bearing. Ten benchmarks were
ongoing column inclination measurements, rebound set for this measurement on both the mountain side and
hardness tests and reinforcing bar rupture investigations. sea side of the TG foundation, as shown in Fig. 3.
As a result of this verification of the current and Starting in December 1981, the distances between these
long-term soundness of the TG foundation, it has re- benchmarks were measured using a micrometer and steel
mained in continuous use up to the present time. rod, and the total of these distances was used to deter-
Chapter 4 presents the main measurement results used mine the displacement. As shown in Fig. 4, expansion in
L10 L9 L8 L7
L6 L5 L4 L3 L2 L1
2×5,600 5,100
3×5,600 2×5,600
Low-Pressure Turbine High-Pressure
Generator Section Section Turbine Section
(21,778.14) (16,800) (11,200)
(49,778.14)
Mountain Side Unit : mm
40
35
Expansion Values of Axial Direction (mm)
20
15
10
5
0
-5 Sea Side Mountain Side
-10
'78 '80 '82 '84 '86 '88 '90 '92 '94 '96 '98 '00 '02 '04
(Year)
the axial direction due to ASR increased up until the function, and to better understand the extent of ASR and
early 1980s, but appears to have stopped around 1991. the expansion rate. The compressive strength of the
Since then, the expansion has remained at the same level concrete, its static elastic modulus, observations of the
as the bench mark measurement, which fluctuates by cut surface of the cores and the results of accelerated
approximately 10 mm as a result of temperature varia- expansion tests carried out up until 2002 are all presented
tions due to seasonal change and turbine generator op- here.
eration.
(1) Compressive strength and static elastic
4.1.2 Crack conditions modulus
Visual inspection of concrete cracks in the TG founda- The locations on the TG foundation from which concrete
tion has been undertaken since 1982 to monitor the de- core was sampled are shown in Fig. 6, and the compres-
formation of the TG foundation. In 2002, major cracks sive strength of these cores is shown in Fig. 7. This re-
were observed in the main reinforcing bar direction on veals that the actual compressive strength of the table
the mountain side surface of the TG foundation (Fig. 5),
and these were accompanied by cracks developed in a
perpendicular direction to the reinforcing bar at the lower
portion of the turbine support beams. The maximum
local crack width was approximately 3.0 mm. As only a
few cracks were observed in the main reinforcing bar
direction and top of the column during the early stage of
crack monitoring (1982), these later cracks are assumed
to have developed as a result of expansion of the TG
foundation. Once ASR was determined to have settled,
and crack repair work was carried out, no new cracks
have been observed.
4.2 Structural evaluation sidered in order to fully examine the long-term sound-
The structural stability of the TG foundation then was ness of the structure (Shimizu and Asai et al. 2005; Ta-
verified in 2005 by an analytical evaluation that took into kakura et al. 2009).
account the compressive strength and static elastic
modulus values obtained from the table deck expansion 4.2.1 Analysis model
and concrete core tests conducted once it was determined A three-dimensional frame model with beam elements,
that ASR had settled. Although it is quite unlikely that as shown in Fig. 12, was adopted to analyze the TG
ASR may reoccur in the future, this possibility was con- foundation so as to
0.15 0.15
Column・
intermediate 0.10 0.10
beam (%) (%)
0.05 0.05
M4
M6
M10 0.00 0.00
0 3 6 9 12 0 3 6 9 12
M11 (M)
(M)
0.15 0.15
Foundation
0.10 0.10
mat
(%) (%)
0.05 0.05
M9 0.00 0.00
0 3 6 9 12 0 3 6 9 12
(M) (M)
(1) Take into account the difference in elongation be- took into account the material property reduction ratio,
tween the sea side and mountain side of the table i.e., the ratio of the material properties of concrete af-
deck and the nonlinearity of some of the applied load. fected by ASR to those of unaffected concrete. This ratio
(2) To model the entirety of the TG foundation structure was defined on the basis of a literature review (Koyanagi
above the basemat in order to properly evaluate the et al. 1998, Tanahashi et al. 1996, Yamura et al. 1994),
stresses in the columns and intermediate beams. model testing (Murazumi and Watanabe et al. 2005,
In the beam element shown in Fig. 12, the mechanical Murazumi and Hosokawa et al. 2005), vibration testing,
properties of the concrete and reinforcing bar were de- measurements of elastic velocity (Takakura and Wata-
fined individually. This was then used to simulate the nabe et al. 2005) and testing of core specimens taken
stress states in the reinforced concrete caused by expan- from the actual TG foundation, as shown in Table 2. The
sion due to ASR. The expansion behavior of the concrete relationship between the reinforcement ratio, compres-
was simulated to fit the measured deformation on sive strength ratio and modulus ratio obtained using these
horizontal bidirection by applying a linear thermal ex- methods is shown in Figs. 13 and 14.
pansion coefficient and an equivalent temperature in- The material properties used for analysis of the con-
crement. Here, it was assumed that the thermal expansion crete were those obtained from core samples of the TG
of concrete can be used to simulate expansion by ASR foundation based on an average value from five different
(Takakura and Shimizu et al. 2005). methods, as shown in Table 2 (Shimizu and Watanabe et
al. 2005). In order to simulate long-term forced defor-
4.2.2 Material properties mation by elongation of the table deck section, the static
From previous studies of material properties obtained elastic modulus in the sound section was reduced by 1/3
through investigations of core specimens taken from from that of the core sample (Takakura and Shimizu et al.
structures affected by ASR, it is known that the strength 2005). The material properties of concrete were defined
and static elastic modulus deteriorate greatly when using these values for the uniformly sound section and
compared to unaffected concrete (Kobayashi et al. 1991). ASR-affected section.
However, it has also been reported that there is very little The material properties of the reinforcing bars were
change in the stiffness and strength of concrete affected defined using average values for the uniformly obtained
by ASR if it contains reinforcing bar restraints resulting through tensile testing of reinforcing bars taken from the
in chemical prestress (Kobayashi 1986). The material TG foundation (Table 3).
properties of the table deck used in the analysis therefore For the stress-strain curve of the concrete and rein-
on the structures with and without ASR. Elastic modulus of the structures have Ratio of Elastic
Vibration Test
been estimated by parametric analysis to obtain eigen value and mode to coincide modulus
the test results.
Measurements of elastic wave velocity(Logging Method, Transmission Method)
Measurement of Ratio of Elastic
have been performed on the structures with and without ASR. Elastic modulus
elastic velocity modulus
has been obtained from the velocity.
Test of Core
Ratio of Compressive
specimens taken Material property tests of core specimens taken from the structures with and
Strength and Elastic
from actual foun- without ASR have been performed.
modulus
dation
forcing bars, the non-linear curve shown in Fig. 15 was 50% that of the respective sound portion of the structure.
used so as to appropriately reflect the nonlinearity on These values were considered the lower limit based on
compression side of the concrete, the cracks produced on the study results shown in Figs. 13 and 14 (Shimizu and
the tensile side, and the yielding of the reinforcing bars. Watanabe et al. 2005).
For structural analysis assuming the redevelopment of
ASR, the compressive strength and static elastic modulus 4.2.3 Applied load and deformation
of the ASR-affected region were assumed to be 80 and Dead load, operating load (live load, equipment load,
(Axial direc-
- - - Sc - - - - ○ - -
tion)
Horizontal Z direction + - - - Sc - - - - ○ -
③ seismic load
(Perpendicular
(static)*2
to axis direc- - - - - - Sc - - - - ○
tion)
+ - - - - - - ○ - ○ - ○ - ○ -
Vertical Y direction
- - - - - - - - ○ - ○ - ○ - ○
④ short circuit torque*3 - - - - - T - - - -
*1 TD : Table Deck
*2 Sc : seismic load equivalent to design level
○ : gradually increasing seismic load
Load applied direction X direction (axial direction) : plus direction is from generator to turbine
Y direction (Vertical direction) : plus direction is from lower to upper
Z direction (perpendicular direction) : plus direction is from sea side to mountain side
*3 T : short circuit torque
etc.), seismic load and short circuit torque load were all equivalent to the average expansion (42 mm) in the axial
considered in this analysis (Japan Association for direction should occur in the perpendicular direction.
Building Research Promotion 2006; Shimizu and Asai et The average value for the expansion rate of the table deck
al. 2005). Expansion in the axial direction of the table in 2002 was determined to be 0.02% based on the ac-
deck due to ASR was assumed to be 50 mm on the sea celerated expansion tests of core samples discussed in
side of the TG foundation and 34 mm on the mountain 4.1.3(3). The expansion value was assumed to be 10 mm
side. These values were determined by adding the dis- (i.e., 0.02% of the total length of the table deck), which
placement at the axial center of the turbine generator increased the length of expansion on the sea side from 50
obtained from adjustment records of the turbine genera- to 60 mm. Thus, for structural evaluation assuming the
tor from the start of its operation until December 1982, at redevelopment of ASR, the expansion value was set to
which point the expansion difference between the sea 1.2 times the total value.
side and mountain side was 4 mm (Fig. 16).
Expansion in the perpendicular direction was assumed 4.2.4 Evaluation methods
to be 6.7 mm for the generator side and 9.7 mm for the To evaluate the margin of safety with regards to the
turbine side based on the notion that a strain (840 μ) aseismic strength of the structure, the material strengths
of the reinforcing bar and concrete obtained through 4.2.5 Results of analysis
testing were used to determine an appropriate reference Analytical evaluations of the TG foundation based on
(i.e., a standard value for margin evaluation). For this, the monitoring of its condition up until 2004, and assuming a
following values were adopted as standards: recommencement of ASR, were undertaken for two
2
• Reinforcing bar yield strength = 409 N/mm (as de- separate cases. In Case 1, conditions under normal op-
termined by tensile strength testing) eration (a dead load + operating load and expansion of
• Concrete compressive strength (determined from core the table deck) were combined with seismic load or load
specimens taken in 2002) due to short circuit torque to determine if the stress state
2
• ASR-affected region = 28.8 N/mm remained within designed levels. In Case 2, the seismic
2
• Unaffected concrete = 30.4 N/mm load was gradually increased to determine the safe mar-
To calculate the shear strength capacity, the gin. A list of the analytical conditions used to evaluate
Ono–Arakawa lower bound equation was used to get the safety margin for strength in both cases is provided in
conservative results, as outlined by the standards of the Table 4.
Architectural Institute of Japan (AIJ). The influence of
twisting was taken into account by using an equation (1) Evaluation of structure based on measurements
proposed by Article 22 of “AIJ Standard for Structural a) Case 1
Calculation of Reinforced Concrete Structures”. The tensile stress of the reinforcing bar, compressive
stress of the concrete, and occurrence of stress due to
shear force are shown in Fig. 17 in relation to the stan-
60
50
40
Displacement (mm)
30
20
10
0
'75'76'77'78'79'80'81'82'83'84'85'86'87'88'89'90'91'92'93'94'95'96'97'98'99'00'01'02'03'04
Sea Side Year
Mountain Side
Difference between Sea Side and Mountain Side
Sea Side ( on annual inspection )
Mountain Side ( on annual inspection )
Difference between Sea Side and Mountain Side ( on annual inspection )
End Play
Fig. 16 Expansion of turbine generator foundation.
R. Manabe, H. Kawae, K. Ogawa and M. Matsuura / Journal of Advanced Concrete Technology Vol. 14, 590-606, 2016 601
dard value. A representative set of results for this analysis A comparison of the deformation between these two
is given by the deformation mode of Case 1-1 and Case cases suggests that ASR was more critical than seismic
1-3 shown in Fig. 18. Note that in Case 1-1, the maxi- load (Fig. 18).
mum ratio of working stress to the standard value of
tensile stress for the reinforcing bar is 60%, which pro- b) Case 2
vides a sufficient safety margin relative to the reference Figure 19 compares the seismic loads corresponding to
value (Fig. 17). In Case 1-3, this ratio increases to 70%, yielding of the reinforcing bars, compressive fracture of
but this is still considered within an appropriate margin. the concrete and shear failure to the designed seismic
40
35
10
5
0
-5 Sea Side Mountain Side
-10
'78 '80 '82 '84 '86 '88 '90 '92 '94 '96 '98 '00 '02 '04 '06 '08 '10 '12 '14
(Year)
Fig. 20 Expansion of table deck in axial direction (until 2014).
b) Case 2
An analysis similar to Case 2-3, where member safety
margin to seismic load was minimum, for the recom-
mencement of ASR using the measured conditions up to
2004 found that the strength of the structure remained 4.4
times that of the gradually increasing seismic load, which Fig. 21 Expansion between benchmark distances in each
section.
was considered a sufficient margin.
approximately 40°C on the sea side and approximately sea side was lower (maximum temperature of approxi-
45°C on the mountain side. These temperatures dropped mately 40°C) than on the mountain side, and that the unit
by approximately 10°C during winter. A similar seasonal expansion of the mountain side was greater. As the unit
change was observed in other sections, but during regular expansion on the mountain side was highest in areas of
inspection, the internal temperature was generally ap- low internal concrete temperature (maximum tempera-
proximately 25°C all year. The positions at which the ture was approximately 40°C), the temperature most
internal temperature was measured are shown in Fig. 22, conducive to promoting ASR is considered to be in the
while the time-dependent behavior of each section for a vicinity of 40°C. The shrinkage observed in the table
given temperature is shown in Fig. 23. deck axial direction can also be explained by a decrease
From the relationship between the table deck expan- in the internal temperature of the concrete after opera-
sion (Fig. 21) and internal temperature of the concrete tions were stopped.
(Fig. 23), it is evident that the internal temperature on the
(2) Reinforcing bar strain
The strain in the reinforcing bars has been measured
since 1986 using meters that were installed following
cutting of the main reinforcement of the columns and
beams, where the initial strain at the cut-off was intro-
duced to the meter. The measurement points are shown in
Fig. 24 and the time-dependent behavior of the main
reinforcing bar (M-1) in the sea side of the high-pressure
turbine section is shown in Fig. 25. We see from this that
there was a downward trend in the reinforcing bar strain
following a peak in the early 1990s, which is consistent
with the change in expansion behavior between the
benchmarks.
by measuring the difference between the displacement of 5.2 Compressive strength of concrete
the table deck relative to the position of the column base In order to determine the compressive strength of the
(a difference in height of approximately 11 m) using concrete, rebound hardness has been measured since 2008
pendulums situated at each corner of the TG foundation. near the portion where concrete cores were taken from the
The measurement points and starting displacement is TG foundation in 2002. These tests have verified that there
shown in Fig. 26, and the time-dependent behavior of the has been no major change in this time (Fig. 28).
table deck displacement along its axis is shown in Fig. 27.
Note that prior to 2011, the displacement on the moun- 5.3 Reinforcing bar rupture
tain side varied between 56 and 62 mm depending on The possibility of reinforcing bar rupture by ASR was
temperature, while the sea side displacement was be- considered to be low in the TG foundation because of its
tween 50 and 56 mm. Once operations ceased, however, indoor environment. However, as reinforcing bars of
this displacement decreased owing to the decrease in the similar diameter have ruptured in outdoor structures, the
internal temperature of the concrete. magnetic flux density of the bars has been measured at
the corners of the columns and beams of the TG founda-
tion since 2015. This has been achieved using a mag-
netizing unit based on a permanent magnet, with rupture
being determined by whether the peak magnetic flux
density exceeds a predetermined value obtained through
a series of experiments (Hirose et al. 2012). No rupture
has been detected since measurements began, with an
example of the peak value for the magnetic flux density
measured in the TG foundation provided in Fig. 29.
silica reaction based on non-linear structure analysis.” concrete structure with alkali-silica reaction.” 18th
18th International Conference on Structural International Conference on Structural Mechanics in
Mechanics in Reactor Technology (SMiRT18), Reactor Technology (SMiRT18), SMiRT18-H03-1,
SMiRT18-H03-4, Beijing, China, August 7-12. Beijing, China, August 7-12.
Shimizu, H., Watanabe, Y., Ishikawa, T., Sekimoto, H., Takakura, T., Ishikawa, T., Mitsugi, S., Matsumoto, N.,
Oshima, R., Takiguchi, K., Masuda, Y. and Nishiguchi, Takiguchi, K., Masuda, Y. and Nishiguchi, I., (2005).
I., (2005). “Study on material properties in order to “Investigation on the expansion value of turbine
apply for structural analysis of turbine generator generator foundation affected by alkali-silica
foundation affected by alkali-silica reaction.” 18th reaction.” 18th International Conference on Structural
International Conference on Structural Mechanics in Mechanics in Reactor Technology (SMiRT18),
Reactor Technology (SMiRT18), SMiRT18-H03-5, SMiRT18-H03-7, Beijing, China, August 7-12.
Beijing, China, August 7-12. Takakura, T., Shimizu, H., Masuda, Y. and Matsumoto,
Tanahashi, K., Iwanaga, T., Koyanagi, W. and Asano, Y., N., (2009). “Consideration of deformation for
(1996). “Structural behavior of RC beam and columns reinforced concrete structure affected by alkali-silica
deteriorated by ASR.” Proceedings of the Japan reaction.” Concrete Research and Technology, 20(3),
Concrete Institute, 18(1), 843-848. (in Japanese) 1-10. (in Japanese)
Taniguchi, T., Ogawa, K., Kawae, H. and Takagi, S., Takakura, T., Shimizu, H., Masuda, Y., Matsumoto, N.,
(2015). “Investigation on the long-term changes of Oshima, R. and Sekimoto, H., (2005). “Investigation
turbine generator foundation affected by alkali-silica of safety margin for turbine generator foundation
reaction.” Proceedings of the Japan Concrete Institute, affected by alkali-silica reaction (Part3) Simulation of
37(1), 871-876. (in Japanese) model and actual structure.” Preprints of Annual
Takakura, M., Watanabe, Y., Hosokawa, T., Ishi, T., Convention of the Architectural Institute of Japan,
Takiguchi, K. and Masuda, Y., (2005). “Vibration Osaka, Japan, September 1-3. (in Japanese)
measurement and simulation analysis on a reinforced
Journal of Advanced Concrete Technology Vol. 14, 464-474 August 2016 / Copyright © 2016 Japan Concrete Institute 464
Technical report
Abstract
The main feature of a prestressed concrete containment vessel (hereinafter referred to as “PCCV”) is that unbonded
tendons, which are not bonded to the structure, are used because the tension of the tendon needs to be measured in pe-
riodical in-service inspection (hereinafter referred to as “ISI”) to ensure that the performance of the power plant is
maintained during the service life. This report describes the results of measurement of the tension of the tendon, which has
been performed as an activity to maintain the PCCV in Ohi Nuclear Power Plant (hereafter referred to as Ohi “NPP”)
Units 3 and 4 of Kansai Electric Power in the past twenty some years since the construction of the PCCV, the changes in
the standards and guidelines for the maintenance of the PCCV in Japan, as well as the discussion of non-destructive
test-based methods for estimating the strength of high-strength concrete used for the PCCV and the development of a
hydraulic shim-type load cell-based system for the measurement of the tension of the tendon, taking into account the ease
of non-destructive testing, etc.
SCV by prestressing the concrete and thus can be more prestressing members (hereinafter referred to as “ten-
resistant to the pressure in a LOCA. The steam leakage don”) is 10 MN (1000 tf).
can be prevented by applying a steel liner plate to the Figure 3 shows the overall arrangement of the tendons
inner surface of the concrete. In addition, the SCV re- and a schematic cross-sectional plan view of the PCCV.
quires a reinforced concrete outer shielding wall, which The tendons consist of hoop tendons (90 in the cylin-
has the function to shield radiation, on the outer side of drical part and 18 in the dome part), which apply a force
the containment. The concrete of the PCCV has the in a manner to hoop the PCCV, and inverted U vertical
function to shield radiation. Thus, the PCCV is a rea- tendons (90), which apply a force in a manner to press
sonable structure with the functions of the SCV and the down the PCCV. The hoop tendons, which circulate
outer shielding wall combined. around the circumference of the PCCV, are anchored at
Table 2 shows the basic specification of the PCCV. both ends to the anchorages on the buttresses. There are
Figure 1 shows a schematic cross-sectional elevation two buttresses, one each at diagonal positions on the
view of the PCCV and Fig. 2 shows the details of the circumference. The hoop tendons are arranged so that
elevation. The PCCV consists of about 10 m thick base they are anchored alternately to each buttress. A group of
mat and about 65 m high cylindrical shell and a dome 45 inverted U tendons are arranged in an orthogonal grid
with a diameter of about 45 m. The cylindrical shell and pattern in a plan view and anchored to the ceiling of the
the dome, the upper structure, are prestressed concrete.
The steel liner plate on the inner surface has the function
to prevent leakage, and the load capacity of the
Dome part
Inverted
U tendon
Cylindrical part
Equipment hatch
Shell
Piping
penetration
Bottom
Fig. 1 Schematic cross-sectional elevation of the PCCV. Fig. 2 Wall structure of PCCV (Yonezawa et al. 2003).
Y. Yamaguchi, T. Kitagawa, H. Tanaka, Y. Koshiro, H. Takahashi, et al. / Journal of Advanced Concrete Technology Vol. 14, 464-474, 2016 466
tendon gallery in the base mat. During prestressing, the bonded to the structure, are used and the sheath for the
prestressing force applied to the ends of the tendon is tendon is filled with grease as the corrosion protection
7.19 MN (733 tf) for the hoop tendon and 7.05 MN (719 medium.
tf) for the inverted U tendon, so that the membrane stress The prestressed system uses a 10 MN-class VSL
during a LOCA is compressive, even taking into account (Vorspann System Losinger) system. One tendon con-
the loss of the prestressing force due to the creep of sists of 55 PC steel strands with a diameter of 12.7 mm
concrete and the relaxation of PC steel strands during the and has a rated tensile strength of 10.09 MN (1028.5 tf).
service life of the plant. The anchorage is such that the reaction is taken by the
To ensure that the performance of the PCCV is main- anchor head; the tendons are pulled by a jack; and the PC
tained during the service life, periodical ISI needs to be steel strands are anchored to the anchor head by wedges.
performed to measure the prestressing force of the ten- Figure 4 shows a schematic of the VSL system. As a
don. For this reason, unbonded tendons, which are not compressing force, the prestressing force applied to the
tendons is transferred to the concrete structure via the
anchor heads and the bearing plates. The inner diameter
of the trumpet sheath is extended at one end into the
Inverted U concrete structure so that the anchor head moves into the
tendon Dome sheath to release the prestressing force. Shims are in-
Hoop tendon serted between the anchor head and the bearing plate to
of dome transfer the stress during the service life. The anchorage
Cylinder
is covered by the end cap and filled with the corrosion
protection medium to prevent entry of moisture and
Equipment Hoop tendon protect the system from corrosion.
hatch of cylinder
3. In-service Inspection (ISI)
3.1 Outline
A plan for ISI of the PCCVs of Ohi NPP Units 3 and 4
Tendon gallery has been developed in accordance with the Technical
Base mat Guidelines for Prestressed Concrete Containment Ves-
sels for Ohi NPP Units 3 and 4 (February 1987 (partially
revised in 1988), Agency for Natural Resource and En-
Tendon arrangement ergy, Ministry of International Trade and Industry). (1)
Visual inspection, (2) inspection of the prestressed sys-
Hoop tendon
tem and (3) inspection of the corrosion protection me-
dium were performed in the 1st, 3rd and 5th year of op-
eration (the liner is excluded from this report). 4% and 4
Buttress or more of the total number of tendons in each group
were selected for (2) inspection of the prestressed system
from each group of tendons (hoop and inverted U ten-
dons). To evaluate the behavior, etc. of the tendons dur-
Hoop tendon ing the service life of the plant and find a correlation
between the observation data from the previous per-
Schematic cross-sectional plan view of the PCCV
formed ISIs, more than one “common” tendon (history
inspection tendon) to be inspected in every ISI was se-
Fig. 3 Overall arrangement of the tendons and a sche- lected from each group.
matic cross-sectional plan view of the PCCV (Ozaki et al. Based on the results of the inspection performed in the
2004). 1st, 3rd and 5th year of operation, Regulatory Guide 1.35
of USNRC and ASME Section XI, it was decided to measurement of the lift-off load. In a lift-off test, it is
perform the inspection every five years between the 10th verified that the prestressing force of the tendon is more
to the 15th year of operation. The quantity of tendons to than the design requirement.
be inspected was reduced to 2% and 3 or more of the total
number of tendons in each group. For the common ten- 3.2 Summary of the observed results and the
dons, tendons same with the ones inspected in the 1st, 3rd evaluation
and 5th year of operation were inspected. In the Ohi NPP, the 25th year ISI of Unit 3 PCCV was
Some of the tendons to be inspected in the 20th and performed in August 2015 and the 20th year ISI of Unit 4
subsequent years of operation were changed based on the was performed in February 2012.
plan to incorporate the Rules on Fitness-for-Service for In each ISI, as a visual inspection, a general visual
Nuclear Power Plants into the Codes for Nuclear Power inspection of the concrete surface and a detailed visual
Generation Facilities - Rules on Concrete Containment inspection of the controlled portions were performed;
Vessels for NPPs (hereinafter referred to as “the CCV and as testing of the prestressed system, a tension test of
Rules”) enacted by Japan Society of Mechanical Engi- the tendons, a detailed visual test of the anchorages of the
neers. More specifically, there is no change to the plan to tendons and a test of the corrosion protection medium
inspect 2% and 3 or more of the total number of tendons were performed.
in each group in the 5th and subsequent years of operation. The results and evaluation of the ISI of Ohi NPP Units
However, regarding hoop tendons, a total of two com- 3 and 4 are summarized below.
mon hoop tendons, one dome hoop tendon and one cyl-
inder hoop tendon, were selected. In the plan to incor- (1) Visual inspection
porate the Fitness-for-Service Rules into the CCV Rules, i) General visual inspection of the concrete surface
one common tendon will be selected from all hoop ten- For a general visual inspection of the concrete surface, a
dons and two arbitrarily selected tendons (hereinafter general visual inspection was performed on the outer
referred to as “random sample tendon”) will be selected surface of the cylindrical part and the buttresses. The
as tendons to be inspected. Considering the above, as for results of the previously performed ISIs showed no
hoop tendons to be inspected, a total of four tendons (two cracks and damage that could adversely affect the
common tendons and two random sample tendons) were structure of the containment vessel.
selected. As for the common tendons, two tendons con-
tinued to be inspected, taking into account the accumu- ii) Detailed visual inspection of the controlled areas
lated data. For the random sample tendons, tendons that For a detailed visual inspection of the controlled areas,
had not been inspected were selected from the perspec- distinctive areas of the surface of the containment vessel
tive of avoiding duplication, which is based on the plan were selected and defined as control areas, and a visual
to incorporate the Fitness-for-Service Rules into the inspection of the concrete surface at the following areas
CCV Rules. was performed in each ISI. The results of the previously
The inspection of the prestressing force is called a performed ISIs showed only very small cracks in the
lift-off test to verify the residual tension of the tendon. concrete surface.
More specifically, a filler gauge (a 0.3 mm thick stainless i Cylindrical part near the base mat
steel sheet) is used to measure the lift-off load. In this ii General cylindrical part with no pipe penetration in
measurement, a jack is placed at the end of an anchorage; the vicinity
reaction is taken by the bearing plate; and the tendons are iii Boundary between the cylindrical part and the
pulled by the jack; and the lift-off load is the load when dome
the filler gauge between the anchor head and the bearing iv Boundary between the cylindrical part and the but-
plate can be removed. Figure 5 shows the method for the tress
v Near the air lock
vi Near the equipment hatch
vii Local area of the cylindrical part with a pipe pene-
tration
9.0
9.0
8.0 8.0
7.0 7.0
Lift-off load (MN)
5.0 5.0
3.0 3.0
1 10 100 1 10 100
ii) Detailed visual test of the anchorages of the ten- under Article 3 of Ministerial Ordinance No. 62) were
dons individually made by the Agency for Natural Resources
In a detailed visual test of the anchorages of the tendons, and Energy, Ministry of International Trade and Industry.
for the tendon which is the subject of inspection, a visual The aforementioned Technical Guidelines for
inspection for cracks in concrete at the anchorages of Prestressed Concrete Containment Vessels for Ohi NPP
the tendons, a visual inspection of the anchorage (anchor Units 3 and 4 were used as the internal review rules for
head, bearing plate, etc.), and an inspection of the num- the Ohi NPP.
ber of effective strands are performed. The results of the The internal rules provide technical guidelines for the
previously performed ISIs showed no problem. material, design, construction, pre-service structural
integrity testing (hereinafter referred to as “SIT”) and ISI
(3) Inspection of the corrosion protection medium of a PCCV. The guidelines provide only a brief descrip-
For an inspection of the corrosion protection medium, a tion of periodic ISIs: visual inspection for cracks in the
visual inspection of the extracted corrosion protection surface of concrete and corrosion of PC steel members,
medium, a chemical composition analysis of the ex- liner plates, etc. tension testing, and material properties
tracted corrosion protection medium, and a volume testing of tension member fittings. As reference literature,
measurement of the newly injected corrosion protection Regulatory Guide 1.90 Rev.1 “Inservice Inspection of
medium were performed. In the visual examination, it Prestressed Concrete Containment Structures with
was checked that there was no infiltration of water from Grouted Tendons”, Aug. 1977 and Proposed Revision 3
the outside, and water-soluble impurities etc. were in- to Regulatory Guide 1.35 “Inservice Inspection of Un-
spected in the chemical component analysis of the ex- grouted Tendons in Prestressed Concrete Containments”,
tracted corrosion protection medium. The results of the Apr. 1979 are cited.
previously performed ISIs showed all clear with the Subsequently, a reinforced concrete containment ves-
results of the visual inspection of the extracted corrosion sel (hereinafter referred to as “RCCV”) was added to the
protection medium. Also, for the results of the chemical guidelines, and it became necessary to develop technical
composition analysis of the extracted corrosion protec- standards for CCVs. In September 1987, the Ministry of
tion medium, water-soluble impurities (Chloride ion etc.) International Trade and Industry made an inquiry to the
were below the detection limit. The results of the volume Thermal and Nuclear Power Engineering Society con-
measurement of the newly injected corrosion protection cerning “Development of Technical Standards for Con-
medium showed that the amount of the newly injected crete Containment Vessels.” In response to the inquiry,
corrosion protection medium was more than the amount the Committee on the Development of Standards for
of the removed corrosion protection medium. Concrete Containment Vessels prepared the report
“Proposed Technical Standards for Concrete Contain-
4. Changes over the year in the standards ment Vessels” in February 1989. The above standards
and guidelines for the maintenance of specify technical standards for the design of concrete
concrete containment vessels in Japan components, the materials used in reinforced and
prestressed concrete, liners, knuckles, body anchors, and
The PCCV of Tsuruga NPP Unit 2 of The Japan Atomic pressure testing, etc. The Committee also presented
Power Company is the first concrete containment vessel proposed guidelines for the verification of the structural
(hereinafter referred to as “CCV”). Since at that time, performance and ISI of CCVs as well as for the design
there were no technical standards for the CCV, internal and construction of diaphragm floors in the containment
review rules (Approval of Specially Designed Facilities vessel.
Y. Yamaguchi, T. Kitagawa, H. Tanaka, Y. Koshiro, H. Takahashi, et al. / Journal of Advanced Concrete Technology Vol. 14, 464-474, 2016 469
The proposed technical standards were incorporated Code (JSME S NE1-2014), and the Code was issued in
into the Technical Standards for the Structure of Concrete 2015.
Containment Vessels (Ordinance No. 452 of the Ministry Structures in nuclear facilities have been maintained
of International Trade and Industry enacted in October by nuclear operators. Since there were no unified codes
1990), but those for SIT and ISI were not. and guidelines for nuclear facilities, in 2008, the Archi-
In an effort to develop commercial standards in a tectural Institute of Japan (AIJ) issued the Guidelines for
systematic manner, the Power Generation System Stan- Maintenance and Management of Structures in Nuclear
dards Committee was established in October 1996 by the Facilities (hereinafter referred to as “the AIJ Mainte-
Japan Society of Mechanical Engineers (JSME), and nance Guidelines”), which can be also used for the Plant
rules such as “Rules on Design and Construction for Life Management (hereinafter referred to as PLM), to
Nuclear Power Plants (JSME S NC1-2001)”, “Rules on provide standards to maintain the required function of
Fitness-for-Service for Nuclear Power Plants (JSME S structures in nuclear facilities during the service life.
NA1-2000)” and “Rules on Welding for Nuclear Power CCVs were excluded from the application of the AIJ
Plants (JSME S NB1-2001)” were developed under the Maintenance Guidelines for the reason that they are
Nuclear Power Expert Committee. However, the provi- within the scope of application of JSME’s CCV Code.
sions for class-MC vessels in JSME S NC1-2001 were Since the CCV Code is applicable only to ISI, subse-
applicable only to SCVs and not applicable to CCV. quently, the need arose to develop guidelines for the
Therefore, it was necessary to develop technical stan- maintenance of CCVs as with other structures in nuclear
dards for CCV as the academic society and association facilities. Since response to the PLM was required, par-
standards. Thus, the Concrete Containment Vessel ticularly for PCCVs, a provision for the maintenance of
Standards Subcommittee was established in September PCCVs was added to the Appendix when the AIJ Main-
2002 under the Nuclear Power Expert Committee, and tenance Guidelines were revised in 2015. The provision
“Rules on Concrete Containment Vessels for Nuclear for the maintenance of PCCVs, which does not apply to
Power Plants (JSME S NE1-2003)” consistent with ISI, states that the ISI set forth in the CCV Code should
JSME S NC1-2001 were developed. To make the CCV be used as needed.
Code consistent with the international standards, the Thus, domestic standards and guidelines were put in
Codes were developed based on Boiler and Pressure place nearly 30 years after the first NPP that uses a CCV
Vessel Code, Section III, Division 2, Code for Concrete went into operation.
Containments formulated by American Society of Me-
chanical Engineers (hereinafter referred to as “ASME 5. Development of new technology
Sec. III, Div. 2”), which was used to prepare Ordinance
No. 452. 5.1 Strength estimation methods based on
In accordance with the policy presented on July 22, non-destructive testing of high-strength con-
2002 in the report, “Performance Specification in Tech- crete (Tsuzuki et al. 2014)
nical Standards for Nuclear Power Generation Facilities A non-destructive testing method based on the rebound
and Use of Commercial Standards” prepared by the Nu- number is often used to evaluate the compressive
clear Reactor Safety Subcommittee of the Nuclear and strength of concrete. The commonly used formula of the
Industrial Safety Subcommittee, Advisory Committee Society of Materials Science, Japan (JSMS) for strength
for Natural Resources and Energy, the Ministerial Ordi- estimation and the formula defined in AIJ’s
nance Partially Revising the Ministerial Ordinance of non-destructive testing manual tend to underestimate the
Establishing Technical Standards for Nuclear Power actual strength of a high-strength concrete structure. The
Generation Equipment (Ministerial Ordinance No. 68 of compressive strength of concrete used in PCCVs is ex-
October 31, 2005 of the Ministry of International Trade pected to be more than 50 N/mm2 from the results of
and Industry) was put into effect; and, as a result, the long-term material age strength testing of control
Technical Standards for Structure etc. of Nuclear Power specimens and monitoring specimens prepared at the
Generation Equipment (Ministerial Ordinance No. 501 time of construction. Therefore, the current formula to
of the Ministry of International Trade and Industry) and estimate the strength of concrete from the rebound
the Technical Standards for Structure, etc. of Concrete number is not applicable. To estimate the compressive
Containment Vessels (Ministerial Ordinance No. 452 of strength of concrete used in PCCVs, it is desirable to use
the Ministry of International Trade and Industry) were a strength estimation formula that takes into account the
removed and the CCV Code (JSME S NE1-2003) be- effects of aggregate used in PCCVs, mixing, concrete
came the technical standards for CCVs after the Code placement, and ambient conditions. Methods for esti-
was endorsed in 2005. mating the compressive strength of concrete that can be
However, the CCV Code (JSME S NE1-2003) pro- applied to high-strength mass concrete structures were
vided technical standards only for the material and de- evaluated using a mass concrete member simulating a
sign of CCVs. Subsequently, in 2011, the CCV Code was PCCV. Table 3 shows the preparation of the concrete
revised. Construction, testing and inspection during used in the full-scale simulated member. Figure 7 shows
construction, the SIT and SIS were added to the CCV a schematic of the simulated member and its shape.
Y. Yamaguchi, T. Kitagawa, H. Tanaka, Y. Koshiro, H. Takahashi, et al. / Journal of Advanced Concrete Technology Vol. 14, 464-474, 2016 470
Table 3 Preparation of concrete for the full-scale simulated member (Tsuzuki et al. 2014).
Unit Quantity (kgf/m3)
Mixing Ratio Percentage of Fine
Water Cement Fly ash
(W/B) Aggregate (%) Fine aggregate Coarse aggregate
Admixture
(W) (B)
50% 49.7 165 264 66 889 921 2.145
40% 46.6 165 330 83 800 937 2.478
30% 44.5 170 454 113 698 888 7.371
Water: groundwater, Cement: moderate-heat Portland cement (density: 3.21 g/cm3)
Fly ash: JIS Type II (density: 2.33 g/cm3, specific surface area: 3890 cm2/g)
Fine aggregate: mixture of crushed sand (density in saturated surface-dry condition: 2.66 g/cm3) and mountain sand (density in
saturated surface-dry condition: 2.59 g/cm3)
Coarse aggregate: crushed tight sand (density in saturated surface-dry condition: 2.70 g/cm3), Admixture: polycarboxylic
acid-based high-performance AE water-reducing agent (standard type)
The non-destructive testing methods used in the that can estimate the compressive strength (About 30 to
evaluation were the rebound hammer, ultrasonic (surface 90 N/mm2) of a high-strength mass concrete structure
wave velocity) and elastic wave methods (P-wave ve- that meets the conditions for the full-scale simulated
locity, S-wave velocity). The results obtained by these member.
methods were compared with the compressive strength
of the core samples taken from the full-scale simulated F = 16.4 × exp(2.91×10−2 × R)
member. Figure 7 shows the position where the core
samples were taken and the position where the where F is estimated compressive strength (N/mm2) and
non-destructive test was conducted. R is rebound number determined by a rebound hammer.
A single regression analysis was performed on the Figure 8 shows the correlation between the rebound
results obtained by each non-destructive test-based number and the compressive strength of the core samples.
strength estimation method and the compressive strength The figure shows the evaluated strength estimation
of the core samples. The square of the coefficient of formula, the 95% confidence interval, the formula of the
correlation (= r) expressed by an exponential function JSMS, and the formula in AIJ’s non-destructive testing
was defined as the “percentage of contribution,” and the manual. At the lower limit of 95% confidence interval
correlation between the compressive strength of the core calculated by the evaluated strength estimation formula,
samples and each non-destructive testing method was the compressive strength calculated by the evaluated
evaluated based on the percentage of contribution. As a method is higher than the one calculated by the formula
result, the percentage of contribution of the rebound of the JSMS and the formula in AIJ’s non-destructive
number was the highest, and the material age had the testing manual when the rebound number is 40 or higher.
least effect on the percentage of contribution. A rebound Figure 8 also shows the results of the compressive
number-based strength estimation formula was evaluated strength test performed at a material age of 22 years on
1300mm
2325mm :コア採取位置
Core sampling point
Mate- :反発度測定位置
Rebound number measurement point
材齢
rial
age
44週 :表面波速度測定位置
Surface wave velocity measurement point
13
:P波,S波速度測定位置
P-wave and S-wave velocity measurement point
13週
Arrangement
配筋: of reinforcement:
26週
26
横筋:D38@225(かぶり厚:96㎜)
Horizontal reinforcement: D38@225 (covering depth: 96 mm)
2250mm
1年
1 Vertical reinforcement: D38@310 (covering depth: 186 mm)
縦筋:D38@310(かぶり厚:186mm)
2年
2 Placement: winter (mid-January, 2006)
打設:冬期(平成18年1月中旬)
During the formwork installation, 200 mm insu-
反発度測定面およびその裏面以
3年
3
lation was placed on surfaces other than the
外は,型枠時に断熱材(200㎜)を
5
5年 surface for rebound number measurement and its
設置している。
backside.
L C R
*※材齢4年目に,
The top of the simulated member
採取位置
Sampling point was cut off at a material age of 4 years.
上部を切断撤去。
Fig. 7 Schematic and shape of the full-scale simulated member (Tsuzuki et al. 2014).
Y. Yamaguchi, T. Kitagawa, H. Tanaka, Y. Koshiro, H. Takahashi, et al. / Journal of Advanced Concrete Technology Vol. 14, 464-474, 2016 471
the PCCV monitoring specimens (Monitoring Specimen A full-scale model test was performed to verify the
B) that were fabricated with the same mixing ratio on the accuracy of the prestressing force measurement method
site during the construction of Ohi NPP Units 3 and 4 and based on the change in the amount of oil. Figure 13
exposed to the same ambient conditions. This indicates shows the test specimen.
that the evaluated strength estimation formula gives a In the test, a 10 MN-class VSL system, which is the
rough estimate of the compressive strength of the actual
PCCV.
same as the one used in the actual PCCV, was reproduced Table 4 Results of the tests.
in a concrete block (1.1 × 1.1 × 1.6 m) and a hydraulic Prestressing Evaluated
shim load cell was installed at the tendon anchorage. B/A
force A (MN) load B (MN)
Three levels of prestressing force, 5.5 MN, 6.5 MN and No.1 5.52 5.62 1.02
7.3 MN, were applied in the test. Three measurements 5.5MN No.2 5.55 5.53 1.00
were taken by the hydraulic shim load cell at each load No.3 5.58 5.68 1.02
level and compared with the reference load cell load. No.1 6.42 6.56 1.02
The displacement of the hydraulic booster cylinder 6.5MN No.2 6.37 6.42 1.01
(the amount of oil) increased nearly linearly until the No.3 6.38 6.40 1.00
reference load cell load was exceeded and thereafter No.1 7.34 7.40 1.01
7.3MN No.2 7.38 7.22 0.98
significantly increased. This trend was observed in all
No.3 7.38 7.21 0.98
tests performed, and thus the measurement method based
on the amount of oil was determined to be valid. Table 4
shows the results of the tests. tember 2001 during ISI of Ohi NPP Unit 3. The lift-off
As a result of the tests, the ratio of the evaluated force load is the load of the intersection of the two straight
to the reference load cell load was between 0.98 and 1.02, lines when the data is separated on the point where the
with a maximum evaluation error of 2%. tangent stiffness of each step becomes a half initial
Figure 14 shows an example of measurements of the stiffness, and linear approximation of the first slope and
change in the amount of oil pumped into the lift-off jack the second slope is carried out by the least-squares
and the lift-off load in a lift-off test performed in Sep- method.
Fig. 12 Concept of the prestressing force measurement based on the change in the amount of oil.
Y. Yamaguchi, T. Kitagawa, H. Tanaka, Y. Koshiro, H. Takahashi, et al. / Journal of Advanced Concrete Technology Vol. 14, 464-474, 2016 473
end of December 2015. Before the passing of 30 years containment vessels.” The first fib Congress, Session
from the start of operation, a technical evaluation of 15, Monitoring, 233-238.
degradation of the PCCV due to aging needs to be per- Ozaki, M., Takeuchi, K. and Kitagawa, T., (2004).
formed in accordance with the PLM, taking into account “Construction and maintenance of a Nuclear Power
the results of ISI and the results of technology devel- Plant Prestressed Concrete Containment Vessel
opment, and the evaluation needs to be reviewed. The (PCCV).” Journal of Prestressed Concrete, 46(6),
PCCV has been maintained appropriately as described 41-48. (in Japanese)
above. To ensure the integrity of the PCCV, the mainte- Tsuzuki, M., Koshiro, Y., Fushimi, M. and Nakano, M.,
nance activities will be continued with the understanding (2014). “Study on estimating strength of high strength
of the purpose of the new regulatory standards. concrete − Experimental studies of several
non-destructive inspection for concrete and concrete
References core strength using mock-up mass concrete specimen
Abe, T., Kawai, I., Ohashi, H., Murazumi, Y. and at 5 years age −.” AIJ J. Technol. Des. 20(45), 487-490.
Takezaki, S., (2003). “New prestressing force meas- (in Japanese)
urement system for prestressed concrete containment Yonezawa, K., Imoto, K., Ohba, M., Ikeuchi, T., Kozuma,
vessels.” SMiRT-17, August, Paper #H03-3. M., Murazumi, Y. and Sato, K., (2003) “Analytical
Ozaki, M., Kawai, I., Ohashi, H., Murazumi, Y. and study on structural failure mode of 1/4 PCCV test
Takezaki, S., (2002). “New prestressing force model.” SMiRT-17, August, Paper #H02-1.
measurement system for prestressed concrete
Journal of Advanced Concrete Technology Vol. 14, 384-396 July 2016 / Copyright © 2016 Japan Concrete Institute 384
Scientific paper
Abstract
In the Hokuriku district in Japan, large numbers of concrete structures have been suffering from damage caused by alkali
silica reaction (ASR). Some huge volcanoes are located within this district, and the headwaters of main rivers prompt the
outflow and spreading of volcanic rocks such as andesite, rhyolite and tuff stones, which are the main volcanic reactive
stones causing the serious damage of ASR in the entire area. To solve this problem effective countermeasures such as
repair and strengthening methods should be established.
One intake tower in this area had deteriorated due to ASR, and deformation had occurred as a consequence of ASR ex-
pansion of the concrete. Countermeasures were carefully considered by academic experts, and post-tensioned tendons
were inserted into the intake tower concrete (vertically oriented) so that the power station can continue to operate safely in
the future. This is thought to be the first challenge of its type anywhere in the world for which the deformation of a real
structure caused by ASR expansion must be controlled.
In this paper, the results obtained from laboratory tests using reactive aggregate and the overview of the investigation and
the method of the reinforcement including the effect of the countermeasures will be discussed.
1
Professor, Kanazawa University, Kanazawa, Japan.
*Corresponding author, E-mail: torii@t.kanazawa-u.ac.jp
2
Manager, Hokuriku Electric Power Company Co. Ltd,
Ushijima-machi, Toyama, Japan.
3
Assistant manager, Hokuriku Electric Power Company
Co. Ltd, Ushijima-machi, Toyama, Japan.
4
Hokuriku Electric Power Company Co. Ltd, Fig. 1 Typical features of seriously ASR-deteriorated
Ushijima-machi, Toyama, Japan. bridge piers.
K. Torii, T. Kubo, C. Sannoh and M. Kanitani / Journal of Advanced Concrete Technology Vol. 14, 384-396, 2016 385
Table 1 Results of chemical method according to JIS A1145 and accelerated mortar bar method according to ASTM C1260.
Chemical method (JIS A1145) Mortar bar method (ASTM C1260)
Name of river
Sc*1 Rc*2 Sc/Rc ASR Reactivity Expansion at 14day(%) ASR Reactivity
Kurobe 22 37 0.59 Innocuous 0.137 Innocuous or Deleterious(Unclear)
Jouganji 242 72 3.36 Not Innocuous 0.720 Deleterious
Jinzu 27 41 0.66 Innocuous 0.306 Deleterious
Shou 26 54 0.48 Innocuous 0.139 Innocuous or Deleterious(Unclear)
*1: Amount of Dissolved Silica (m mol/l), *2: Amount of Alkalinity Reduction (m mol/l)
Pl
PX
PX
Crs
Pl Crs
Op
0.1mm 0.1mm
(a) Opl from altered Andesite: plane polarized light (b) Cristobalite from Andesite: plane polarized light
PX
Gls
Gls
Gls Pl PX
Gls Pl
Trd Trd Pl
Gls
Trd
0.1mm Pl 0.1mm
Gls
(c) Tridymite from Andesite: crossed polarized light (d) Volcanic glass from Andesite: plane polarized light
[Op : opl; Crs : cristobalite; Trd : tridymite; Gls : volcanic glass; Pl : plagioclase; Px : pyroxene]
Fig. 2 Photomicrographs of reactive minerals and non-reactive minerals in the river gravel used in the intake tower taken
with a polarizing microscope observation.
K. Torii, T. Kubo, C. Sannoh and M. Kanitani / Journal of Advanced Concrete Technology Vol. 14, 384-396, 2016 386
constituted of granitic rocks and volcanic stones, in power demand. Usually it reaches the highest level
which the reactive minerals in typical andesite stones around May, and decreases to about 1070m in the sum-
were cristobalite and/or tridymite, opl, and small mer months. Then it decreases further to about 1060m in
amounts of volcanic glass, as shown in Table 2. The the winter (February - March), returning to the highest
composition ratio of 30% andesite stones was roughly water level due to snowmelt in the spring. Temperatures
equivalent to the most pessimistic results gained by the during the summer rise to about 25 degrees, and in winter
mortar bar or concrete bar test of andesite stones in the they fall to about minus 10 degrees. In the winter, there is
river. Thus, it has become very clear that the reactive snow cover of about 5m. Therefore, between November
river sand and river gravel contain reactive minerals, and April it is not possible to go to the site area.
especially opl, the most reactive mineral, and that its Twenty years after the construction of the tower, an
amount may be near the most pessimistic volume. A inclination with concomitant expansion of the structure
recent survey shows that in the bridges and buildings was discovered due to the formation of a clearance of
using this river sand and gravel, a severe ASR has still 40mm between the top of the intake tower and the con-
occurred after 2010 even in the case of total alkali con- necting bridge. After the detection of this separation,
tent of 2 kg/m3 and less, which is considerably smaller investigations were conducted to find the cause.
than the JIS A5308 required value of less than 3 kg/m3.
3.2 Displacement measurement
3. Overview of the investigation Distance from vertical line to both the interior wall and
the exterior wall were measured at each elevation. Fig-
ure 5 shows the results of the investigation of the interior
3.1 History of the intake tower wall and the exterior wall. Measured distance shows the
This intake tower is located at a dam in the upstream average of measured values on the interior wall and the
region of Jouganji river in Hokuriku District. Figure 3 exterior wall. With the bottom of the tower’s base located
shows the actual site. 74m3/s of water can pass through at 1,045 meters above sea level, we noted that from about
the intake tower and 400 thousands kW can be generated an elevation of 1,065 meters the intake tower wall began
by using this water at 3 power stations located down- to exhibit an inclination in the direction of the lake. At
stream of the tower. This electricity corresponds ap- the top of the tower (elevation 1,090.7 meters above sea
proximately to the power needs of about 130 thousand level), there was a deviation from the vertical of 80mm. It
households. Thus, this intake tower is a critical structure was not clear at the time of construction how accurately
for the area and the company. the vertical orientation of the tower wall was, so it is
The intake tower is a SRC structure. Figure 4 shows a unclear exactly to what degree the wall’s configuration
cross-sectional view and a side view of the intake tower. has deformed.
The water level of the dam varies depending on the Figure 6 shows the horizontal change in distant over
time at the top of the tower, and Fig. 7 shows the vertical
change over time at the top of the tower, which horizontal
and vertical measurements were both conducted from 22
Intake tower
years after construction. During the 8-year period of 22
to 30 years following construction the rate of horizontal
High water level
change at the top of the tower was 5.6mm per year. As-
Connecting bridge suming that the average rate of change was the same
prior to the commencement of measurements and taking
into account the currently measured change of 80mm, we
estimate that the beginning of this deformation started
Front gate
about 15 years following construction. During the same
8-year period, the rate of vertical change was calculated
to be 2.1mm per year. Thus, assuming that the deforma-
tion along the vertical axis of the tower also began 15
years after construction and assuming that the rate of
Fig. 3 Overview of actual site of the intake tower. deformation was uniform over time, we estimate that the
K. Torii, T. Kubo, C. Sannoh and M. Kanitani / Journal of Advanced Concrete Technology Vol. 14, 384-396, 2016 387
overall height of the tower at its top has increased by El 1060m after 30 years of the construction. As shown in
33mm on the side of the connecting bridge. Fig. 8, a slightly expansive tendency was observed at the
heights of El 1070m and El 1060m.
3.3 Results of the residual expansion test Therefore, it has been determined that it will be nec-
Figure 8 shows the results of the residual expansion test essary to watch for the future expansion.
according to the JCI DD2 method . Each core was taken
from the side of the tower at the heights of El 1070m and
2,500 9,500
9,000 500
H-beam or BH-beam
EL.1,090,700
300
3,900 300
1,500
2,200
中間壁柱断面
側壁柱断面
2,000
7,850
2,500
5,000
H-582×300×12×17
H-582×300×12×17
Built-up-H-beam(BH-beam)
7,500
24,700
① BH-1400×400×16×28mm
H-582×300×12×17
H-588×300×12×20
② BH-1400×400×19×32mm
13@2,500=32,500
6@3,750=22,500
7,500
37,700
③ H-588×300×12×20mm
A
A A
A
FPL-17×350 WPL-12
FPL-20×350 WPL-12
④ H-582×300×12×17mm
7,500
FPL-25×350 WPL-16
FPL-28×350 WPL-16
10,000
62,700
5,500
B
B B
B
FPL-28×350 WPL-16
FPL-32×350 WPL-19
8,500
EL.1,051,500
for construction
8@2,750=22,000
C
C
C
C
EL.1,047,500
8,500
EL.1,045,500
3,750
7,500
FPL-32×400 WPL-19
FPL-28×400 WPL-16
EL.1,043,000
3,750
9,000
3,000
EL.1,038,000
7,500
2,500
19,500
EL.1,034,000
4,500
8,200
5,500
6,000
Direction
3,000
550
EL.1,028,000
2,400 2,200 2@1,050=2,100
7@1,000=7,000 5@900=4,500 6@800=4,800 500
6,500 17,000
Gate metal material North
23,500
(Channel steel:200×90×2~8mm)
A-A B-B C-C
9,000 Rebar 10,000 11,500
2,500
2,500
④ ① ③
7,600
7,600
7,600
②
22,700
22,700
22,700
2,500
2,500
2,500
③ ② ③
7,600
7,600
7,600
①
1,400 2,500
1,400 2,500
1,400 2,500
④ ① ③
(measured in 2007.8)
41mm 5.6mm/year
Intake tower
Vertical change
Connecting bridge
Expansion(%) Expansion(%)
(EL 1070m) (EL 1060m)
0.07 0.07
0.06 0.06
Judgment line Judgment line
0.05 0.05
残存膨張量(%)
残存膨張量(%)
0.04 0.04
0.03 0.03
0.02 0.02
0.01 0.01
0.00 0.00
-0.01 -0.01
0 7 14 21 28 35 42 49 56 63 70 77 84 91 0 7 14 21 28 35 42 49 56 63 70 77 84 91
日 日
Time (days) Time (days)
(immersed in the temperature of 40℃ and the humidity of 95% or more)
Mountain
Lake side
side
Cores were taken from the inside of the mountain side
× Cores were taken from the outside of the mountain side
Cores were taken from the inside of the lake side
inside outside
3.4 Cause of the inclination mountain side wall has more exposure to sunlight which
Figure 9 shows the condition of the surface of the intake is believed to be a large influence on the ease of ASR
tower’s exterior wall. On the surfaces of the exterior and development. What’s more, on the mountain side in the
interior walls extensive mapping cracks were observed. vertical direction, the structure is not reinforced. Thus, it
The cracks were approximately 0.2mm in width at 20 to is believed that this permits a large ASR influence on this
30 cm intervals and were not severe. side of the tower.
The intake tower concrete (at the levels of EL1035, 1047, The amount of ASR expansion is said to be affected by
1060, 1070, and 1087) was investigated using an the percentage of the steel that constitutes the cross sec-
SEM-EDS. Figure 10 shows an example of an image at tion. Table 3 shows the value of the cross-sectional area
the level of EL1060. This investigation discovered ASR of the steel (steel, rebar, gate metal material) as a fraction
gel in all samples taken from all locations. of the cross-sectional area of the concrete. Comparing
As a result of the research, ASR was observed in the the mountain side wall to the lake side wall, the propor-
entirety of the intake tower, causing expansion. Fur- tion of steel in the mountain side wall is small, thus it is
thermore, from the elevation of 1065m, the tower was presumed that the expansion of the mountain side is
leaning towards the lake. The height of the tower is about larger than that of the lake side. Comparing the fractional
63m, and in the cross section of the tower the water in- value of steel in each section, the value of the A-A sec-
take gates are shown. Due to the gate there is less ex- tion is small, thus it is presumed that the expansion of the
posure to sunlight on the lake side. On the mountain side upper portion is larger than that of the lower portion.
there is a solid wall. Compared to the lake side, the Furthermore, in March the lake level is at its lowest, so
ASR gel
1m
Fig. 9 Condition of the surface of the intake tower’s exte- Fig. 10 Example of the observation of the concrete by
rior wall. SEM-EDS.
K. Torii, T. Kubo, C. Sannoh and M. Kanitani / Journal of Advanced Concrete Technology Vol. 14, 384-396, 2016 390
Table 3 Values of the steel as a fraction of the cross-sectional area of the concrete (%).
Values of the steel as a fraction of the cross-sectional area of the concrete (%)
Section (See Fig. 4)
Lake side wall Mountain side wall
A-A 0.0065 0.0028
B-B 0.0079 0.0038
C-C 0.0079 0.0039
between May and August as the lake fills, the upper whether or not the connecting bridge might possibly drop.
portion of the tower is exposed periodically to more In fact, the stopper installed to prevent the such bridge
sunlight followed by lake water inundation which we collapses has already been repaired once due to the de-
believe promotes the formation of ASR gel. formation of the tower.
In this way the comparative differences between the As described above, concerning the overall usage of
concrete environments and the different degrees of re- the intake tower, there have been no safety or usage
inforcement on the mountain side and lake side result in problem up to now except repairing the bridge’s con-
unequal amounts of ASR expansion on opposite sides of necting members. Although after 30 years of ASR, the
the tower. This is the fundamental cause of the tower’s alkalinity and the residual expansion tests do not indicate
increasing inclination in the direction of the lake. Figure the faster development of the tower inclination, in fact
11 is a graphic illustration of the primary causes of the the intake tower change is continuing so we must con-
ASR. Incidentally, this figure is obtained by charting the sider that it may continue. If this occurs the concrete and
estimated, qualitative effects taking into account the metal gates could separate resulting in the gates being
basic mechanisms of the ASR (Larive,C., et al., 2000). rendered unusable. Thus, even if there is little physical
deformity or expansion, we believe ASR countermea-
4. The Method of the reinforcement sures are valid.
mind, we think it is proper to consider, even if just a little, (Pre-stressed Concrete) developed by SEEE
the control of the ASR expansion and deformation. Corporation.
Table 4 shows the applicability of normal repairs and Figure 12 shows the cross-section of the hole after
reinforcement for this intake tower given the ASR situa- construction and Fig. 13 shows the situation of the
tion (Torii et al. 2004; Kubo et al. 2004; Daidai et al. construction. The anchors are inserted into the drilled
2012). Generally, deformation bound by PC wire inside holes. The anchor rods consist of 19 cables, which are
the precast concrete or the attachment of a steel plate to each made from 7 PC steel wires. They have a diameter
the concrete surface are the countermeasures of choice. of 89mm and a length of 45.7m. The cable has a flexible
However, as this intake tower has a complicated shape structure. That is to say the cable (see cross sectional
due to the gate on the front, restraining further figure) is coated with anti-rust oil. This together with the
deformation in the horizontal direction is difficult. polyethylene resin is inserted into the outer slide pipe.
Further to this, since the inclination of the intake tower is The lower parts of the rods are made from steel to fix the
due to the vertical expansion of the ASR, this time it was anchors at the bottoms of the holes. Throughout the holes
decided to adopt a method for restraining the further cement milk is injected. First the upper parts of the rods
vertical expansion using the PC steel wire. Of course,
complete rebuilding of the structure would be the most
reliable method. However, since such a construction Cement milk
project is exorbitantly large in scope and requires Polyethylene resin
long-term shut-down of the power plant’s operation, that Slide pipe
option was not adopted. Cable (19×φ12.7mm)
Hydraulic
system
Tendon(φ89mm)
Anchor fixing part
(φ120mm, 4.7m)
During the insertion of the anchor During the tensioning work
Fig. 13 Photographs documenting the construction.
K. Torii, T. Kubo, C. Sannoh and M. Kanitani / Journal of Advanced Concrete Technology Vol. 14, 384-396, 2016 392
are hydraulically pulled, and then they are fixed in place research to date concerning the expansive pressures of
with a nut. In other words, it is possible to create a ASR gels and the required forces required to suppress
compression force on the concrete by introducing a that expansion there are reports about such topics. (e.g.,
tensioning force to the rods. The anchors adopted this Kawamura, M.)
time have a screw on their tops. They are a type of anchor According to some results of the investigations,
with adjustable tensioning force. Additionally, after pressure of 0.2 ~ 0.3MPa is said to be effective to
considering margins corresponding to potential ASR suppress the expansion due to ASR (Torii et al. 2000;
development in the future, the largest anchor available in Ishii et al. 2005; Okuyama, et al. 2007; Ishii et al. 2008).
our country was adopted. However, the actual phenomena observed in the field
Figure 14 shows the construction location plan view, have been much more complex. To this end we are
and Fig. 15 shows the drilling depth view. Drilling on the conduct creep tests on cores taken from the intake tower
mountain side was, as much as possible, evenly spaced by introducing a small, long-term load to the test sample
with 10 places, each place spaced at a distance of about in order to confirm the long-term influences of such
2m from the next. Furthermore, in order to balance the loads on deformability(Torii et al. 2010).
whole structure, two anchors were placed on the Furthermore, the behavior of the actual structure is
lake-side. The anchors’ base positions of immobilization unclear. Thus, the monitoring is being continued, and,
are set at near EL1045m close to the bottom portion of further to the behavior observed, we will adjust the
the tower structure, and in this way tension was applied anchors’ introduced forces accordingly.Incidentally,
to the entire structure. Thus, 45.7m drilling depth was when the tension in the longitudinal direction is
needed. introduced, there is a possibility that deformation will
Construction was first carried out for only one of the proceed in the transverse direction (Gravel et al. 2000),
candidate locations in 2009. This was a test construction so follow-up measurements and visual inspections of the
for the purpose of confirming the concrete properties and structure’s appearance become increasingly important.
the workability of anchors in advance. Then, the re-
maining construction for the rest of the eleven locations 5. Testing of the boring core concrete
was carried out in 2011. Drilling was accurate to within ±
0.5°. Tension work was carried out in the spring of 2012.
The initially applied force was set at 1500kN (50% of the 5.1 Compressive strength and static modulus of
yield point load, 0.26N /mm2). The value of 0.26MPa elasticity
was derived from the following calculation. Figure 16 shows the test results for the compressive
Tension per anchor : 1500kN strength of a core sample from the tower, and Fig. 17
Total number : ten shows the test results for static modulus of elasticity.
The cross-sectional area of the intake tower’s Please note that in Figs.16 and 17, the green outline of
mountain side : 2.5 × 22.7m = 56.75m2 the tower on the left illustrates the relative locations in
Thus, 1500 × 10 ÷ 56.75 = 0.26Mpa. In the history of the tower structure where the various compressive
strength measurements were taken. The compressive
6.5m 2.5m
EL 1,090.7m
45.7m Tendon
Metal gate about 2m
plate
EL 1,066m
EL 1,056m
Mountain
Lake side side
EL 1,045m
Fig. 14 Top view of the locations of the holes for the Fig. 15 Side view indicating the depth of the holes for the
post-tensioned tendons (holes represented by red dots). post-tensioned tendons.
K. Torii, T. Kubo, C. Sannoh and M. Kanitani / Journal of Advanced Concrete Technology Vol. 14, 384-396, 2016 393
strength results were on average higher than the planned lower part of the tower below 1066m differs from the
average standard strength(σ28:24N/mm2). Also, the re- upper part is the lower part is under water most of the
sults for compressive strength and static modulus of year.
elasticity were both even higher at the 1,090 meter level.
The compressive strength of the core from the intake 5.2 Observation of the thin section by the po-
tower’s center showed a relatively high value. This is larizing microscope
because the restraining force of the steel is stronger at the Some thin pieces made from cutting the core (40mm ×
center of the tower. 25mm) were observed under a polarizing microscope.
Although variations of the compressive strength and These pieces were made one by one, each from the core
the static modulus of the concrete in which ASR occurs in the vicinity of EL1086.5m, EL1080.6m, EL1076.1m,
tends to be large, the durability of the structure is be- EL1071.3m, EL1068.4m, EL1061.9m, EL1057.1m,
lieved to be no problem, because the actual concrete EL1052.4m, EL1050.5m and EL1046.4m. Figure 18
interior is bound with re-bars. In the lower part of the shows an example of a micrograph (EL1061.9m). As a
tower below 1066m, the fact that the static modulus of result, since cracks through the fine aggregate were
elasticity is reduced indicates that ASR should have observed in slices from all the elevations, it has become
progressed. However, the constraining conditions are clear that ASR has occurred throughout the whole
significant, and we believe this is why not so much ex- interior of the intake tower.
pansion was observed. Furthermore, it is believed that an
additional reason why the progress of the ASR in the
1090
(m)
1080
Elevation
1070
標高(m)
1060
1050
1040
1030
0 10 20 30 40 50
圧縮強度(N/mm2)
Compressive strength (N/mm2)
Fig. 16 Compressive strength measurements of the cores taken at various locations in the tower.
1090
Elevation (m)
1080
1070
標高(m)
1060
1050
1040
1030
-10 0 10 20 30 40 50
静弾性係数(kN/mm2)
Static modulus of elasticity (N/mm2)
Fig. 17 Static modulus of elasticity measurements of the cores taken at various locations in the tower.
K. Torii, T. Kubo, C. Sannoh and M. Kanitani / Journal of Advanced Concrete Technology Vol. 14, 384-396, 2016 394
20
20
00
2000.8
H12.8.10 H13.9.14 H14.10.19 H15.11.23 H16.12.27 H18.1.31
2008.8
H19.3.7
H20.4.10 2015.11
H21.5.15 H22.6.19 H23.7.24 H24.8.27 H25.10.1 H26.11.5 H27.12.10
Time (years)
Fig. 19 Horizontal change in distant over time at the top of the tower.
00
2000.8
H12.8.10 H13.9.14 H14.10.19 H15.11.23 H16.12.27 H18.1.31 2008.8
H19.3.7
H20.4.10 2015.11
H21.5.15 H22.6.19 H23.7.24 H24.8.27 H25.10.1 H26.11.5 H27.12.10
Time (years)
Fig. 20 Vertical change in distant over time at the top of the tower.
K. Torii, T. Kubo, C. Sannoh and M. Kanitani / Journal of Advanced Concrete Technology Vol. 14, 384-396, 2016 395
lake side of the tower. Although the results of this rein- Grvel, C., Ballivy, G., Khayat, K., Quirion, M. and
forcement have had a small inhibiting effect on the Lachemi, M., (2000). ”Expansion of AAR concrete
amount of vertical displacement, the difference in verti- under triaxial stress: Simulation with instrumented
cal expansion rate between the mountain side and the concrete block.” Proc. 11th Inter. Conf. on
lake-side of the tower has been reduced, and this has Alkali-aggregate Reaction in Concrete, 959-968.
effectively suppressed the horizontal displacement. Hashimoto, T. and Torii, K., (2013). ”The development
Please note that the primary purpose of the rein- of highly durable concrete using classified fine fly ash
forcement is to suppress the horizontal deformation. in Hokuriku district.” Journal of Advanced Concrete
Therefore, it can be said that the effect of the Technology, 11, 312-321.
countermeasure has been confirmed. Huang, M. and Pietruszczak, S., (1996). ”Numerical
analysis of concrete structures subjected to
7. Concluding remarks alkali-aggregate reaction.” Mechanics of
Cohesive-frictional Materials, 1, 305-319.
Ishii, K., Okuda, Y. and Torii, K., (2005). “ASR
In this paper the assessment of alkali-silica reactivity of mitigation-effect of repair and retrofitting methods
volcanic gravel from Jouganji river in the Hokuriku against ASR damaged reinforced concrete columns.”
district in Japan, which is considered to be the most Concrete Journal, 43(7), 42-50. (in Japanese)
reactive one in Japan, has been introduced. Secondly, the Ishii, K., Okuda, Y., Wasada, S. and Torii, K., (2008).
investigation of ASR-caused deformities to a water in- “Countermeasure to the ASR-affected bridge piers.”
take tower which was constructed by using river aggre- Concrete Journal, 46(1), 55-60. (in Japanese)
gates produced from Jouganji river and the attempts to Jensen, V., (2000). “In-situ measurement of relative
reinforce the tower using post-tensioned tendons has humidity and expansion of cracks in structures
been introduced. Although an inclination was confirmed damaged by AAR.” Proc. 11th Inter. Conf. on
on the intake tower, it has been shown that this is because Alkali-aggregate Reaction in Concrete, 849-858.
of the difference in the ASR expansion due to the Kawamura, M. and Iwahori, K., (2004). ”ASR gel
difference in environmental conditions and structural composition and expansive pressure in mortars under
conditions. As a method for controlling the deformation restraint.” Cement and Concrete Composites, 26,
of the intake tower caused by ASR, the method of using a 47-56.
PC anchor was adopted. Consequently it was possible to Kubo, Y., Yamanashi, T. and Torii, K., (2004). “Study on
effectively suppress the deformation from the applied repair method with FRP sheet for spalled concrete
force of 0.2 ~ 0.3N/mm2. structure caused by alkali-silica reaction.” Proc. 12th
Inter. Conf. on Alkali-aggregate Reaction in Concrete,
References 1254-1261.
Bianchi, R. H. and Bruce, D. A., (1992). ”The use of Larive, C., Laplaud, A. and Coussy, O., (2000). “The role
post-tensioned tendons on Stewart Mountain dam, of water in alkali-silica reaction.” Proc. 11th Inter.
Arizona; A case study involving precision drilling.” Conf. on Alkali-aggregate Reaction in Concrete,
Second Interagency Symposium on Stabilization of 61-69.
Soils and Other Materials, Metairie, LA, 15 p. Multon, S. and Toutlemonde, F., (2006). ”Effect of
Binal, A., (2004). ”A new experimental method and applied stress on alkali-silica reaction-induced
device for measuring alkali silica gel pressure in expansions.” Cement and Concrete Research, 36,
mortar.” Proc. 12th Inter. Conf. on Alkali-aggregate 912-920.
Reaction in Concrete, 266-272. Nomura, M., Komatsubara, A., Kuroyanagi, M. and Torii,
Blaikie, N. K. and Bowling, A. J., (1996). “The K., (2012). “Evaluation of the residual expansion of
assessment and management of alkali-silica reaction cores due to alkali-silica reaction in Hokuriku
in the gordon river power development intake tower.” district.” Proc. 14th Inter. Conf. on Alkali-aggregate
Proc. 10th Inter. Conf. on Alkali-aggregate Reaction Reaction in Concrete, 10 pages
in Concrete, 500-507. Okuyama, K., Ishii, K., Okuda, Y. and Torii, K.,
Cavalcanti, A. and Silveira, J., (1989). “Investigation on (2007). ”The strengthening and monitoring of an
the Moxoto powerhouse concrete affected by ASR-affected bridge footings.” Concrete Research
alkali-silica reaction.” Proc. 8th Inter. Conf. on and Technology, 29 (1), 1263-1268. (in Japanese)
Alkali-aggregate Reaction in Concrete, 797-802. Pan. J., Feng. Y. T., Jin. F. and Zhang. C.,
Daidai, T., Andrade, O. and Torii, K., (2012). ”The (2013). ”Numerical prediction of swelling in concrete
maintenance and rehabilitation techniques for arch dams affected by alkali aggregate reaction.”
ASR-affected bridge piers with fracture of steel bars.” European Journal of Environmental and Civil
Proc. 14th Inter. Conf. on Alkali-aggregate Reaction Engineering, 17 (4), 231-247.
in Concrete, 10 p. Salles, F., Pinfari, J., Kuperman, C., Mizumoto, C. and
Diamond, S., (1989). ”ASR-another look at Bernardes, H., (2012). ”Jaguari hydropowerplant-
mechanisms.” Proc. 8th Inter. Conf. on Alkali- evaluation, diagnosis and control of a structural
aggregate Reaction in Concrete, 83-94.
K. Torii, T. Kubo, C. Sannoh and M. Kanitani / Journal of Advanced Concrete Technology Vol. 14, 384-396, 2016 396
affected by alkali-aggregate reaction.” Proc. 14th Inter. (2000). “Strengthening method for ASR affected
Conf. on Alkali-aggregate Reaction in Concrete, 10 concrete piers using prestressing steel wire.” Proc.
pages 11th Inter. Conf. on Alkali-aggregate Reaction in
Sannoh, C. and Torii, K., (2008). ”Suppressing effect of Concrete, 1225-1233
various mineral admixtures on combined deterioration Torii, K., Sannoh, C., Kubo, Y. and Ohashi, Y., (2004).
caused by ASR and chloride attack.” Proc. 13th Inter. “Serious damages of ASR affected RC bridge piers
Conf. on Alkali-aggregate Reaction in Concrete, 9 and their strengthening techniques.” Proc. 12th Inter.
pages. Conf. on Alkali-aggregate Reaction in Concrete,
Seignol, J. F., Boldea, L. I., Leroy, R., Godart, B. and 1283-1288.
Hammerschlag, J. G., (2009). ”Hydro-power structure Torii, K., Inagaki, T., Wakita, K. and Sannoh, C., (2010).
affected by alkali-aggregate reaction: a case-study “Creep behavior of ASR deteriorated concrete under
involving numerical re-assessment.” International the humid environment.” Concrete Research and
Conference on Computational Technologies in Technology, 32(1), 971-976. (in Japanese)
Concrete Structures, 21p. Torii, K., (2010). “The characteristic feature of fracture
Sellier, A., Bourdarot. E., Multon. S. and Grimal. L. E., of steel reinforcement in ASR-deteriorated concrete
(2009). ”Combination of structural monitoring and structures.” J. Corrosion Engineering, 59 (4), 59-65.
laboratory tests for assessment of alkali-aggregate Torii, K., Prasetia, I., Minato, T. and Ishii, K., (2012).
reaction swelling: Application to Gate structure dam.” “The feature of cracking in prestressed concrete bridge
ACI Materials Journal, 106 (3), 281-290. girders deteriorated by alkali-silica reaction.” Proc.
Shayan, A., Wark, R. E. and Moulds, A., (2000). 14th Inter. Conf. on Alkali-aggregate Reaction in
“Diagnosis of AAR in Canning dam, characterisation Concrete, 10 p.
of the affected concrete and rehabilitation of the Torii, K., Hashimoto, T., Kubo, T. and Sannoh, C.,
structure.” Proc. 11th Inter. Conf. on Alkali-aggregate (2016). ”The alkali-silica reactivity of volcanic
Reaction in Concrete, 1383-1392. aggregates and its preventive measure using fine fly
Silveira, J., Degaspare, J. and Cavalcanti, A., (1989). ashes.” Proc. 14th International Conference on
“The opening of expansion joints at the Moxoto Structural Engineering and Construction, 310-318.
powerhouse to counteract the alkali-silica reaction.” Zoilo, C., Bernardes, H., Salles, F. and Kuperman, C.,
Proc. of 8th Inter. Conf. on Alkali-aggregate Reaction (2012). “Monitoring of displacements due to
in Concrete, 747-751. alkali-aggregate reaction.” Proc. 14th Inter. Conf. on
Torii, K., Kumagai, Y., Okuda, Y., Ishii, K. and Sato, K., Alkali-aggregate Reaction in Concrete, 8 p.
Journal of Advanced Concrete Technology Vol. 14, 408-420 August 2016 / Copyright © 2016 Japan Concrete Institute 408
Scientific paper
Abstract
In the context of life extension of nuclear reactor buildings while ensuring safety and regulatory requirements, a nu-
merical strategy is proposed to compute and forecast the air leakage rate evolution of inner containments in double-wall
reactor buildings under standard long-term operation. In order to reduce the numerical cost of such complex computa-
tions, the proposed strategy is based on a coarse but adapted mesh together with a chained weakly-coupled thermo-
hydro-mechanical modeling. The total leakage rate is then computed with a specially-designed tridimensional finite
element based on a non-linear thermal analogy.
This methodology is used to model the behavior of the VeRCoRs mock-up, a simplified nuclear reactor building at scale
1:3, built and monitored by EDF. Results obtained until the first pre-operational pressurization test have been discussed
in a dedicated benchmark organized by EDF. The proposed methodology provides delayed strains and leakage results in
good agreement with available experimental data. A blind prolongation until the first decennial test of the mock-up is
presented and analyzed.
- Leakage, with the prediction of the total air leakage 3D numerical strategy is proposed to compute the air
rate at the end of the pressurization plateau. leak rate. The underlying philosophy is to use simple
The results of the benchmark (Corbin and Garcia physically representative models to describe accurately
2016) have been presented and discussed during a dedi- those complex structures. At that stage, no early age
cated workshop, held beginning of March 2016 at EDF behavior of concrete is considered.
R&D center “Les Renardières”, in France. Despite the weak coupling of phenomena, computa-
To predict accurately the air leakage rate of NRB, the tions are chained in the following order to take into ac-
coupling of thermal, hydric and mechanical (THM) ef- count their relative influence.
fects has to be taken into account. Although a wide
range of more or less sophisticated models are already 2.1 Thermal model
available to represent each physical phenomenon occur- First, the spatiotemporal temperature field T ( x, t ) is
ring in concrete (such as drying, delayed strains, crack- computed (in K) with the classical non-linear heat equa-
ing or evolution of permeability), global strategies aim- tion,
ing at computing the final leakage rate of a full-size
construction are less common. Indeed, a good compro- ∂T ( x, t )
ρb C p + ∇ ⋅ [−λ (T )∇T ( x, t )] = 0 (1)
mise between the refinement of discretization, the com- ∂t
plexity of behavior laws and the strength of couplings
has to be reached to tackle the global problematic of air where ∇ is the differential “nabla” operator ( ∇ being
leakage rates for industrial purposes. From the point of the gradient operator and ∇ ⋅ the divergence operator),
view of computational cost, this leads either to fully- ρb is the concrete density (in kg·m-3), C p is the con-
coupled non-linear strategies (Wang and Hutchinson crete thermal capacity (J·kg-1·K-1) and λ is the concrete
2005; Nicklash et al. 2005; Jason and Masson 2014) thermal conductivity field (in W·m-1·K-1).
only applicable to structural elements or to full-structure
computations based on weakly-coupled elastic strategies 2.2 Hydric model
and post-treatments with the help of design codes for- In a second step, the water saturation field Sl ( x, t ) is
mulae (Mozayan et al. 2012). In those strategies, the computed (values between 0 and 1). Considering that
representation of localized cracking is often one of the drying of concrete is mainly due to liquid water transfer,
most complex steps. the hydric problem reduces to a single diffusion equa-
In that context, the present paper introduces a nu- tion in (Eq. 2a) (Verdier 2001). Thermal dependency of
merical strategy aiming at forecasting the leakage rate the hydric diffusion coefficient D ( Sl ( x, t ), T ( x, t )) is
evolution of inner containments during operation. It modeled with an Arrhenius-type law as proposed by
focuses on a chained weakly-coupled THM modeling Granger (1995) in (Eq. 2b). For the sake of concision in
used to represent, at lower numerical cost, the physical all equations, the notation D( Sl , T ) for fields depending
behavior of the structure (Fig. 1). First, the models im- on other computed fields is now used.
plemented for each physical phenomenon are described, ∂Sl ( x, t )
with a 3D finite element (FE) specially designed for air + ∇ ⋅ [ D ( Sl , T )∇Sl ( x, t )] = 0 (2a)
leakage computation through unsaturated porous and ∂t
cracked concrete. Then, the proposed strategy is applied E ⎛ 1 1⎞
to the VeRCoRs mock-up. Main obtained results are
l
K int krl ( Sl ) ∂Pc T ( x, t ) Ra ⎜⎝ T ref − T ⎟⎠
D ( Sl , T ) = e (2b)
analyzed and compared with experimental data when μl φ ∂Sl T ref
available and benefit from discussions and feedback l
from the first VeRCoRs benchmark. where K int is the water intrinsic permeability of con-
crete (in m2), krl is the water relative permeability field
2. Modeling principles of the numerical (values between 0 and 1), μl is the water viscosity (in
strategy Pa·s), φ is the concrete porosity, Pc is the capillary
pressure field (in Pa), T ref is the reference temperature
After having justified (Asali et al. 2014) that the main (in K) at which all hydric properties of (Eq. 2) are de-
assumption of a weak coupling between THM effects is fined or measured, Ea is the thermal activation energy
valid for inner containments under standard operation, a of concrete (in J·mol-1), R is the universal gas constant
(in J·mol-1·K-1).
The water retention model of Van Genuchten (1980)
is used to define the evolution of the capillary pressure
1. Temperature 2. Saturation 3. Strains field with the saturation degree,
(section 2.1) (section 2.2) (section 2.3)
Pc ( Sl ) = Pr ( Sl−1/ m − 1)
1n
(3)
4. Air Leak where Pr (in Pa) and n (> 1) are model parameters and
(section 2.4)
m = 1 − 1/ n . Mualem’s law (Mualem 1976) gives the
Fig. 1 Representation of the THM strategy. evolution of the water relative permeability field with
M. Asali, B. Capra, J. Mazars and J-B. Colliat / Journal of Advanced Concrete Technology Vol. 14, 408-420, 2016 410
saturation, evolution of creep strains with time (Eqs. 11, 13) and
2
biaxial creep effects with a creep Poisson’s ratio differ-
krl ( Sl ) = Sl ⎡⎢1 − (1 − Sl1/ m ) ⎤⎥
m
(4) ent from elasticity (Eqs. 11, 12). Evolution laws of the
⎣ ⎦ model are the following,
ε bc ( x, t ) = ε kv ( x, t ) + ε am ( x, t ) (9)
2.3 Mechanical model
The third step of the methodology aims at computing σ α ( x, t ) = α bc σ + σ (10)
+ −
the total strain state of concrete. Prestressing losses in
tendons are modeled according to ETC-C formulae ηˆam (t )ε am = (1 +ν bc )σ α −ν bc tr (σ α )I 3 (11)
(AFCEN 2010) with a perfect bond between concrete
and steel cables. Passive reinforcement bars are not ex- ηˆkv ε kv + kˆkv εkv = (1 + ν bc )σ −ν bc tr (σ )I 3 (12)
plicitly taken into account.
In order to model delayed effects in prestressed con- where ε kv is the reversible basic creep strain field, ε am
crete over time, the concrete total strain tensor field is the non-reversible creep strain field, σ α is the stress
ε ( x, t ) is split into five main components, field taking into account a decrease or an increase of
creep strain in tension compared to compression
ε ( x, t ) = ε el + ε bc + ε dc + ε ds + ε th (5)
through the use of the coefficient α bc and vbc is the
where ε el is the elastic strain tensor, ε bc is the basic basic creep Poisson’s ratio. Moisture and temperature
creep tensor, ε dc is the drying creep tensor, ε ds is the impact creep strains as follows,
drying shrinkage tensor and ε th is the thermal expan-
kam t ⎛E ⎛1 1 ⎞⎞
sion tensor. ηˆam (t , T , Sl ) = exp ⎜ a ⎜⎜ − ⎟⎟ ⎟ (13)
Sl ⎜ R ⎟
⎝ ⎝ T Tref ⎠⎠
(a) Thermal expansion model
The thermal expansion tensor is considered as isotropic ⎛ Ew ⎛ 1 ⎞⎞
η kv 1
and proportional to the temperature, ηˆkv (T , Sl ) = exp ⎜ a ⎜ − ⎟⎟ ⎟ (14)
Sl ⎜ ⎜ ⎟
⎝ R ⎝ T Tref ⎠⎠
ε th ( x, t ) = α th (T ( x, t ) − T ( x, 0))I 3 (6)
k ⎛ Ew ⎛ 1 1 ⎞⎞
where α th is the thermal expansion coefficient (in K-1) kˆkv (T , Sl ) = kv exp ⎜ a ⎜ − ⎟⎟ ⎟ (15)
Sl ⎜ ⎜ ⎟
and I 3 is the 3D identity matrix. ⎝ R ⎝ T Tref ⎠⎠
(b) Drying shrinkage model where kam is the stiffness associated to the non-
The drying shrinkage tensor is considered as isotropic reversible ageing dashpot (in Pa), kkv is the stiffness
and is proportional to the saturation degree, associated to the reversible creep strains (in Pa), η kv is
the viscosity associated to the reversible creep strains
ε ds ( x, t ) = κ ds ( Sl ( x, 0) − Sl ( x, t ))I 3 (7) (in Pa·s) and Eaw is the thermal activation energy asso-
ciated to water viscosity (in J·mol-1).
where κ ds is the drying shrinkage coefficient (without
unit).
(e) Elastic-damageable model
Isotropic damage evolution is described with the μ-
(c) Drying creep model
model (Mazars et al. 2015) in order to take into account
Concrete’s total creep is higher when drying compared
the unilateral effect (crack closing and reopening) oc-
to basic creep in autogenous conditions (Pickett 1942).
curring with cycling inflating tests. Main evolution laws
In order to take this Pickett’s effect into account, the
are,
drying creep strain is considered as proportional to the
stress and hydric states of concrete, E ε el ( x , T ) = (1 + v)σ − vtr(σ )I 3 (16)
ε dc ( x, t ) = κ dc Sl ( x, t ) σ ( x, t ) (8)
σ ( x, t ) = (1 − d )σ ( x, t ) (17) (in m2), krg its air relative permeability field (values
between 0 and 1), μ g is the air viscosity (in Pa·s), M g
(1 − A)Y0 B (Y −Y ( ε )) is the air molar mass (in kg·mol-1) and Pg is the air
d ( x, t ) = 1 − − Ae 0 eq (18) pressure field (in Pa). The evolution of the air relative
Y (ε eq )
permeability field follows Mualem’s law as modified by
ε eq ( x, t ) = ε el + χ (ε bc + ε dc ) (19) Verdier (2001),
θ
Q
Pint we Pext
Strain
Fig. 3 Principle of the cracked FE for leakage computa-
tion. Fig. 4 1D principle to define the crack-effective strain.
M. Asali, B. Capra, J. Mazars and J-B. Colliat / Journal of Advanced Concrete Technology Vol. 14, 408-420, 2016 412
QM expressed in kg·s-1. In order to compare leakage 3.3 Boundary conditions and loadings
rates of all NRB independently of each ILR test tem- As the overall structure is subject to a changing envi-
perature and pressure conditions, EDF converts mass ronment as well as to a complex construction phasing
flow rates into volumetric flow rates expressed at nor- process, the on-site evolutions of loadings and boundary
mal temperature and pressure conditions, conditions (BC) are simplified before being applied
over the whole considered time frame. The main con-
R T norm struction planning of the mock-up is considered as such:
QVnorm = 3600 QM (26)
M g P norm - Erection is done at once, the starting date of all com-
putations is on 30 November 2014;
where QVnorm is the volumetric total leakage rate (in - Due to water aspersion on the mock-up’s walls, the
Nm3·h-1) at normal temperature T norm (0 °C) and normal structure begins to dry only from 1 April 2015;
pressure P norm (101315 Pa). - Prestressing occurs between 6 May and 12 August
2015 with a reduced number of tensioning sequences;
3. Application to the VeRCoRs mock-up - Two main life phases of the structure are occurring:
pre-operational until 28 January 2016 during which
3.1 Used tools the mock-up is subject to its external environment,
The numerical strategy presented in section 2 is imple- then operational starting on 1 April 2016. The transi-
mented within the general framework of Code_Aster tion between both phases is linear.
(http://code-aster.org/). Code_Aster v12.3 is used as the - The first three ILR tests of the mock-up are consid-
FE solver for all THM and air leak computations. The ered in the present study: two pre-operational tests
cracked FE of section 2.4 is developed as a new non- (PT0 and PT1) occurring between 10-14 November
linear thermal model within Code_Aster. The concrete 2015 and 24-28 January 2016 respectively, and the
delayed strain model of section 2.3 is implemented as an first decennial test (DT1) between 20-24 April 2017.
external user behavior law through MFront v2.0.1
(http://tfel.sourceforge.net/). (a) Thermal BC and loadings
Soil temperature applied at the bottom of the raft is kept
3.2 Mesh constant during the whole computation at 10.5 °C (ini-
The coarser mesh proposed for the benchmark is lin- tial value). During pre-operation, measurements within
earized before being used for all computations. It in- concrete are available to calibrate imposed temperatures.
cludes 34694 nodes in 3D concrete elements (penta- and No daily temperature variations are considered, only
hexahedra) and 16066 nodes in 1D cable elements linear evolutions between extremal values.
(bars). 295 cables (horizontal, vertical, gamma and Operating temperature conditions are 35 °C at the in-
dome) and their deviations are taken into account (Fig. ner side and 15 °C at the outer side of the wall. Tem-
5(a)). Despite vertical and circumferential coarse mesh- perature is kept constant during each decennial test at
ing of concrete (Fig 5(b)), the 10-element progressive 10 °C (Fig. 6).
fineness within the 40 cm thickness is suitable to repre-
sent phenomena linked to drying (Fig. 5(c)). (b) Hydric BC and loadings
Hydric conditions are naturally expressed in terms of
40 cm
relative humidity (RH, value between 0 and 1). In order ing decennial tests (Fig. 7).
to convert RH to saturation, Van Genuchten’s model (Eq.
3) is completed with Kelvin’s law, (c) Mechanical BC and loadings
As no early age behavior is considered, initial stress and
ρl RT strain states are null. The bottom of the raft is embedded
Pc ( RH , T ) = − ln( RH ) (27)
Ml during the whole computation. Reinforcement bars are
not explicitly meshed, but participate in the dead weight
where ρl is the water density (in kg·m-3) and M l is the of the structure (extra 100 kg·m-3 for concrete density).
water molar mass (in kg·mol-1). The raft and anchoring ribs remain visco-elastic without
As for temperature, soil RH is kept constant at 0.98 damage. Prestressing is simplified compared to the real
(initial value). During pre-operation weather station construction planning to only 14 grouped steps (instead
measurements are imposed at the inner and outer sides of 41 sequences in 16 phases), with an initial tension of
with linearization between extremal values. Operating 848 kN for each cable. All ILR tests are identical: they
HR conditions are 0.45 at the inner side and 0.6 at the last four days, with a 24-hour ramp from standard atmos-
outer side of the containment, with no interruption dur- phere to maximal pressure (5.2 bar abs.), a 24-hour pla-
teau at maximal pressure and 48 h of deflating (Fig. 8).
Pre‐op Operational phase
40
(d) Air leakage BC and loadings
35
During ILR tests, inner and outer air pressure follows
30
the evolution of Fig. 8. The raft being submerged during
ILR tests, it is not considered as a possible leakage
Temperature (°C)
25
20 pathway (null flux).
15
10 3.4 Modeling and material parameters
5 Inner side Concerning concrete, the mean value of parameters
0 Outer side measured on-site for each lift is used when available
‐5 (density, porosity, Young’s modulus, tensile strength,
thermal expansion coefficient, air apparent permeabil-
ity). Other concrete parameters are either provided
Fig. 6 Evolution of imposed temperature. (thermal capacity, linear thermal conductivity with tem-
perature, activation energy, fracture energy) or identified
1
on specific lab tests (water intrinsic permeability, Van
Inner side
Genuchten’s model parameters, drying shrinkage coeffi-
0,9
Outer side cient, basic and drying creep parameters) (Fig. 9).
Relative humidity
0,8
Due to lack of some characterization tests, biaxial
creep effects, distinction between creep in tension and in
0,7 compression and coupling of creep with damage have
0,6 not been taken into account.
Cable parameters come from design and manufac-
0,5
turer values. The complete set of parameters used for
0,4 the present study is available in Appendix A.
Outer side dard areas (far from any change in concrete geometry or
4 0,3 tendon deviation). Points H1, H2, H5 and H6 are moni-
Stress (MPa)
47,4 0
Strain (µm/m)
Weight (kg)
46,8 ‐200
46,6 ‐300
46,4
‐400
46,2
46,0 ‐500
0 30 60 90 120 150 180 210 240 90 120 150 180 210 240 270 300 330 360 390
Time (days) Time (days)
(a) Specimen weight loss (b) 1D drying shrinkage
0 0
Strain (µm/m)
‐300 ‐600
‐400 ‐800
‐500 ‐1 000
‐600 ‐1 200
‐700 ‐1 400
90 120 150 180 210 240 270 300 330 360 390 90 120 150 180 210 240 270 300 330 360 390
Time (days) Time (days)
(c) 1D basic creep (d) 1D total creep
Fig. 9 Identification procedure for some concrete parameters.
H6
tests DT2 to DT6 occur in May 2018, June 2019, July
H5
2020, August 2021 and September 2022 respectively.
As for temperature, saturation at the inner and outer
sides of the mock-up is following seasonal variations
during the first year. With the identified Van Genuch-
ten’s parameters, imposed RH values of 0.6 and 0.45
correspond to imposed saturation values of 0.64 and
0.57 respectively.
In the dome and cylindrical part, the saturation level
at the core of concrete thickness is decreasing with little
impact of imposed pre-operational RH cycles, meaning
global drying of the structure.
H2
Lacking concrete sorption isotherms data to identify
H1 Van Genuchten’s parameters independently of the con-
Fig. 10 Localization of observation points H1, H2, H5 crete water intrinsic permeability on Fig. 9(a) and some
and H6. RH measurements within the mock-up, drying kinetics
of the model is not yet experimentally confirmed.
4.1 Temperature in concrete
Temperature of concrete follows external seasonal 4.3 Evolution of concrete total strains
variations during the first year of the structure. Slowly Figure 12 shows the evolution of total strains within the
changing BC and loadings associated to the used ther- mock-up at the four observation points. All measured
mal parameters lead to an almost homogenous tempera- strains of Fig. 12 are corrected from temperature effects
ture of concrete within the cylindrical part and within according to the following formula (with null initial
the dome (less than 1 °C-difference between mid- strains),
thickness and outer sides).
ε cor (t ) = ε raw (t ) + α s (T (t ) − T (0)) (28)
4.2 Drying of concrete where ε cor is the corrected measured strain, ε raw is the
Figure 11 shows the evolution of the saturation degree raw measured strain and α s is the wire thermal expan-
along line H1-H2 in the thickness of the cylinder sion coefficient (in K-1, steel material). Until the first
(40 cm) before each ILR test. During operation, leakage
M. Asali, B. Capra, J. Mazars and J-B. Colliat / Journal of Advanced Concrete Technology Vol. 14, 408-420, 2016 415
0 0
End of tensioning
DT1
DT1
PT0
PT1
PT0
PT1
Strain (µm/m)
‐300
‐300 ‐400
‐500
‐400
End of tensioning
‐600
‐500
‐700
Operation Operation
‐600 ‐800
(a) PT0
(b) PT1
(c) DT1
3 -1
Table 1 Evolution of total air leakage rate and distribution (in Nm ·h ) at the inner side (starred values are measured at
the outer side of the wall).
VeRCoRs measurements Computation results
ILR test Total no crack
Total Hatch Gusset Dome Cylinder Total Hatch Gusset Dome Cylinder
PT0 7.7 0.67* 2.44* 0.17* 1.08* 6.69 0.09 0.03 1.45 5.12 6.58
PT1 9.5 - - - - 9.25 0.09 0.02 3.30 5,84 7.79
DT1 - - - - - 55.6 0.09 7.14 2.74 45.6 7.71
from the early-age behavior of the structure and, despite from the inner side, with a maximal opening of 160 µm.
being closed by prestressing, reopens during ILR tests. According to Fig. 13(c) and contrary to previous pre-
As the present computation starts with a null initial state operational tests, two cracks connected through the
and a structure built in only one step, this phenomenon thickness are computed at the two junctions between the
could not be represented in associated results. But the gusset and anchorage ribs. The crack opening within the
capacity of the model to close and reopen cracks en- gusset and the first lift above it between 200 and
ables to take into account another initial stress-strain- 400 grades evolves from 150 µm at the inner side to
damage state, coming from previous early-age computa- 40 µm at the outer side.
tion for instance.
During PT1, the first thin layer of FE at both sides of 4.5 Air leakage of the mock-up
the containment is damaged at a low average level Table 1 shows the evolution of air leakage rate com-
(value around 0.4). This generalized surface behavior, puted at the end of maximal pressure plateau for the first
not encountered in other computed ILR tests, may be three ILR tests (PT0, PT1 and DT1). The total leakage
due to temperature and humidity gradients that are still is compared to in-situ measurements for the two pre-
evolving during the pre-operational phase. The maximal operational tests. The distribution of air leakage by area
crack opening at mid-thickness within the material is also shown. However this distribution is provided at
hatch increases to 90 µm. According to Fig. 13(b), more the inner side for the computation while experimental
highly damaged areas on the inner and outer sides (val- results are only collected at the outer side of the inner
ues greater than 0.7) are linked to geometrical accidents containment. It can be noticed that only 57 % of the
in concrete and to tendons deviations: anchorage ribs, total internal leak is measured at the outer side: it is in-
material and personal hatches. As it is not connected deed difficult to inspect 1100 m2 of concrete during the
through the concrete thickness, this diffuse surface 24 h of maximal pressure and the focus is put on visu-
cracking does not impact significantly the total leakage ally leaking areas (soap aspersion). The influence of
rate of the mock-up. modeled cracking on the total leakage rate is quantified
During DT1, the first layer of FE at the inner side of by comparing computed results to an extra computation
the containment is damaged at a medium average level taking into account the unsaturated-concrete darcean
(value around 0.5), which may be due to the tempera- flow only.
ture gradient imposed between the internal and external A first computation has been launched using the mean
sides during operation. Cracking progresses in the mate- value of air apparent permeabilities (3.13·10-16 m2) as
rial hatch: 70 % of its thickness is continuously cracked the intrinsic permeability of concrete. In that case, no
Fig. 14 Visual inspection of the outer side after PT0 (cracks in green, Corbin and Garcia 2016).
M. Asali, B. Capra, J. Mazars and J-B. Colliat / Journal of Advanced Concrete Technology Vol. 14, 408-420, 2016 418
Klinkenberg’s effect (Klinkenberg 1941) is taken into cially-designed 3D FE to compute air leakage through
account leading to overestimating the total leakage rate unsaturated porous and cracked concrete, the proposed
during PT0 (42.9 Nm3·h-1, +557 % compared to the ex- methodology is applied and benchmarked to the scale
perimental measure). Thus the intrinsic permeability 1:3 NRB mock-up VeRCoRs built by EDF. After show-
value for the proposed computation is calibrated on this ing good overall agreement with experimental observa-
first ILR test (to 5·10-17 m2). For PT0 and PT1, this tions until the first pre-operational ILR test, the compu-
value enables to compute total air leak rates that are tation is extended to the first decennial test of the mock-
very close to experimental measurements (respectively - up. Under the considered assumptions of operating con-
13 % and -3 %). ditions with no early-age behavior effects, the proposed
For the proposed computation, air leakage is mainly methodology predicts a proper physical evolution of the
due to concrete porosity during PT0 and PT1 (98 % and total leakage rate coupled to delayed strains, drying and
84 % of the total leak respectively). Even if cracking cracking of concrete.
occurs at that stage, it is not sufficiently going through Compared to available experimental results on the
concrete thickness to participate significantly to the total VeRCoRs mock-up, some improvements of the pro-
leakage, as stated in the previous section. This is cor- posed strategy are already identified: (1) A drying
roborated by analyzing the leakage rate distribution by shrinkage model able to represent short-term kinetics as
areas, which is globally proportional to the surface of well as long-term delayed drying strains levels could be
each zone. The influence of the hatch and gusset does introduced (based on a Burger rheological chain for
not evolve between the first two tests and all the leakage instance). The same applies to the drying creep model if
increase is due to the cylinder and to the dome, which the drying shrinkage improvement is not sufficient. (2)
are thinner and on which drying has more effect during In all cases, the drying kinetics of core concrete needs to
those three months. be validated with future experimental data. (3) The
During DT1, cracking of concrete is contributing sig- mesh refinement in the thickness may need to be further
nificantly to the total leakage rate. Indeed, leakage customized if temperature and humidity effects are in-
through porosity does not evolve between PT1 and DT1 deed the cause of surface cracking during ILR tests. (4)
(and only represents 14 % of the total leak), contrary to The identification procedure, based on lab results for the
leakage through the gusset and the cylinder which in- mock-up, could be adapted to in-situ total strains meas-
creases strongly (respectively 357 and 7.8 times). Asso- urements, for instance to define a representative creep
ciated to the damage (Fig. 13(c)) and air flux mapping, Poisson’s coefficient. As only few characterization tests
the leaking increase is mainly caused by the continuous are available on full-size NRB compared to VeRCoRs,
cracks at the junctions between the gusset and the two this feature seems necessary. (5) The 3D FE designed
anchorage ribs. for air leak computations is currently modeling one
The computed total leakage value of 55.6 Nm3·h-1 crack related to the mechanical state of concrete. But its
may be high due to the conservative assumption of not formulation enables patching as many independent
considering any friction coefficient for the air flow cracks as desired in the same FE. This feature could also
within cracks. For instance, using a classical reduction be used to consider initial defects of the structure such
factor of 0.1 (Picandet 2001) on the Poiseuille’s part of as porosity lines or leaking lift joints. (6) Finally, the no-
the flow during DT1 would lead to a total leakage rate early-age effect assumption of the strategy may need to
of 12.5 Nm3·h-1. In that case, the leak through concrete be reconsidered in light of experimental observations.
porosity would still be prevailing (62 %). This coeffi- Lack of initial early-age cracks leads to underestimating
cient could be identified on experimental measures the gusset leaking during PT0, which would be prob-
when available. lematic only if crack openings in this area evolve with
Considering that air leak distribution measured during time. If not, the simplest way to take into account the
PT0 at the outer side could be extrapolated at the inner initial damage of the gusset could be to locally increase
side of the containment, experimental data show a local the intrinsic permeability to recover experimental results.
behavior different from the one predicted by the compu- If crack openings increase in the gusset, an initial early-
tation. The gusset area participates significantly to the age computation would be necessary to define a non-
total leak. This observation confirms that cracks identi- null strain-stress-damage state for the proposed strategy
fied at the outer side of the gusset are most certainly which is able to close them during prestressing and re-
radial and going through the whole concrete thickness. open them during ILR tests.
Considering conservative assumptions, using simple
5. Conclusions and perspectives modeling techniques (non-linear thermal analogy for
drying and for air-leak computations) and a delayed
This paper presents a numerical strategy that aims at strain model representative of important phenomena
computing and forecasting the leakage rate of double- with few parameters to identify (compared to other
wall NRB (without steel liner) under operating condi- models available in the literature and FE codes), the
tions. Based on (1) a coarse but adapted mesh, (2) a proposed methodology is numerically cost-effective and
chained weakly-coupled THM modeling and (3) a spe- thus enables taking into account variabilities and uncer-
M. Asali, B. Capra, J. Mazars and J-B. Colliat / Journal of Advanced Concrete Technology Vol. 14, 408-420, 2016 419
tainties of main parameters. It could be used as a quali- Cement and Concrete Research, 6, 773-782.
tative and quantitative decision-support tool to help op- Jason, L. and Masson, B., (2014). “Comparison between
erators: (1) Pre-empting and optimizing leak mitigation continuous and localized methods to evaluate the
actions before each decennial ILR test and consequently flow rate through containment concrete structures.”
avoiding outage extensions and associated losses of Nuclear Engineering and Design, 277, 146-153.
income; (2) Forecast the behavior of their assets in the Jourdain, X., (2014). “Etude numérique méso-macro des
context of long-term operation and life extension be- propriétés de transfert des bétons fissurés.” Thesis
yond design in order to demonstrate their compliance (PhD). Ecole Normale Supérieure de Cachan.
with safety and regulatory requirements. Klinkenberg, L. J., (1941). “The permeability of porous
media to liquids and gases.” API Drilling and
Acknowledgements Production Practice, 200-213.
Carried out within the RSNR MACENA project, this Matallah, M., La Borderie, C. and Maurel, O., (2010).
study received government assistance handled by the “A practical method to estimate crack openings in
French National Research Agency (ANR) under the concrete structures.” International Journal for
“investments for the future” program, part number Numerical and Anaytical Methods in Geomechanics,
ANR-11-RSNR-0012-01. 34, 1615-1633.
The authors gratefully acknowledge the VeRCoRs Mazars, J., Hamon, F. and Grange, S., (2015). “A new
project team for analyzing and sharing large amounts of 3D damage model for concrete under monotonic,
data as well as the AMA team of EDF Lab Paris-Saclay cyclic and dynamic loadings.” Materials and
for numerous discussions and for helping with Structures, 48, 3779-3793.
Code_Aster and MFront modeling. Mozayan, M., Davenne, L. and Ghavamian, S., (2012).
“Modelling of leak tightness degradation of
References reinforced containment vessels due to aging and
AFCEN, (2010). “EPR technical code for civil works.” cracking mechanisms.” In: SSCS 2012 – Strategies
Lyon: Association Française pour les règles de for Sustainable Concrete Structures, Aix-en-Provence
Conception, de construction et de surveillance en 29 May – 1 June 2012, France.
exploitation des matériels des Chaudières Electro Mualem, Y., (1976). “New model for predicting
Nucléaires. hydraulic conductivity of unsaturated porous media.”
Asali, M., Capra, B., Colliat, J.-B. and Mazars, J., Water Resources Research, 12 (3), 513-522.
(2014). “Elements for computing and forecasting the Nicklasch, C., Coudert, L., Heinfling, G., Hervouet, C.,
leakage rate of the inner containment of reactor Masson, B., Herrmann, N. and Stempniewski, L.,
buildings.” In: Fontevraud 8 – Contribution of (2005). “Numerical investigation of the leakage
Materials Investigations and Operating Experience to behavior of reinforced concrete walls.” In:
LWR’s Safety, Performance and Reliability, Avignon NURETH11 – International Topical Meeting on
14-18 September 2014, France. Nuclear Reactor Thermal-Hydraulics, Avignon 2-6
CEA, “Mfront, a code generation tool dedicated to October 2005, France.
material knowledge.” released under GNU GPL and Picandet, V., (2001). “Influence d’un endommagement
CECILL-A license. Website: mécanique sur la perméabilité et sur la diffusivité
http://tfel.sourceforge.net/ [accessed on 15 March hydrique des bétons.” Thesis (PhD). Université de
2016]. Nantes.
Corbin, M. and Garcia, M., (2016). “International Pickett, G., (1942). “The effect of change in moisture
benchmark VeRCoRs 2015 – Overview, synthesis and content on the creep of concrete under a sustained
lessons learnt.” Villeurbanne: EDF SEPTEN. load.” ACI Journal Proceedings, 38, 333-355.
EDF Lab, “Code_Aster, Strucutres and VeRCoRs project website. https://fr.amiando.com/EDF-
Thermomechanics Analysis for Studies and vercors-project.html [accessed on 15 March 2016].
Research.” released under GNU GPL. Website: Wang, T. and Hutchinson, T. C., (2005). “Gas leakage
http://code-aster.org/ [accessed on 15 March 2016]. through reinforced concrete shear walls: numerical
Granger, L., (1995). “Comportement différé du béton study.” Nuclear Engineering and Design, 235, 2246-
dans les enceintes de centrales nucléaires : analyse et 2260.
modélisation.” Thesis (PhD). Ecole Nationale des Van Genuchten, M. T., (1980). “A closed-form equation
Ponts et Chaussées. for predicting the hydraulic conductivity of
Hilaire, A., (2014). “Etude des déformations différées unsaturated soils.” Soil Science Society of America
des bétons en compression et en traction, du jeune Journal, 44, 892-898.
âge au long terme.”Thesis (PhD). Ecole Normale Verdier, J., (2001). “Contribution à la caractérisation de
Supérieure de Cachan. l’évolution du taux de fuite des enceintes de
Hillerborg, A., Modeer, M. and Petersonn, P. E., (1976). confinement du parc nucléaire.” Thesis (PhD).
“Analysis of crack formation and growth in concrete Université Paul Sabatier Toulouse.
by mean of fracture mechanics and finite elements.”
Journal of Advanced Concrete Technology Vol. 14, 408-420 August 2016 / Copyright © 2016 Japan Concrete Institute 420
Scientific paper
Abstract
The residual strength tests after cyclic loadings on concrete cylinder specimens and stud anchor specimens were per-
formed to obtain data regarding effects of cyclic loadings on concrete and stud anchor specimens for proposing a new
method capable of evaluating the soundness of equipment support using the S-N curves. More specifically, cyclic load-
ings may lead to a reduction in the strength and stiffness of concrete and also lead to the pullout and a reduction of the
strength of stud anchor sections, which may reduce the rigidity of the equipment support, and this situation is not pref-
erable from the viewpoint of long-term soundness evaluation. Then, in order to investigate whether or not cyclic loadings
reduce the strength and the rigidity of the equipment support, cyclic compressive loadings with constant amplitude were
applied to cylinder specimens, then static compression tests were performed. Also, cyclic tensile loadings with constant
amplitude were applied to stud anchor specimens, then static pullout tests were performed. As a result, it was found that
the cyclic loading would not lead to the pullout of stud anchor and a reduction of the strength of stud anchor if the
maximum stress ratio is less than the long-term allowable stress.
tained loading with the maximum stress ratio of 0.9 at life load cycles was increased by 5% compared with the static
ratio of 0.3 which means at 30% cycles of fatigue life or tensile strength without cyclic loadings. The maximum
30% of duration to cause the creep failure. and minimum stress ratio was 0.5 and 0.1 respectively
Ballatore and Bocca (1997) measured the residual with loading speed of around 1Hz.
strength after cyclic loadings using the plain concrete core
specimens of 45.5 mm diameter and 140 mm long with 2. Concrete cylinder specimen
compressive strength of 19MPa and 59MPa. The maxi-
mum stress ratios were at 0.1 and 0.2 and the minimum This section reports the residual strength test results of the
ratio was 0.02 of the static strength without cyclic load- concrete cylinder specimens in order to obtain the very
ings. The cyclic loadings of at most 7200 cycles were basic information of concrete under cyclic compressive
applied with loading speed of 1Hz. The increases in the stresses.
residual strength and stiffness were 0.11 and 0.87 respec-
tively for concrete with 19MPa and were 0.14 and 0.22 2.1 Outline of tests
respectively for concrete with 59MPa. Table 1 shows the mix proportion of concrete. The con-
Bocca and Crotti (2003) measured the residual strength crete was manufactured using gravel and sand produced
after cyclic loadings using the plain concrete core speci- in Shizuoka Prefecture and ordinary Portland cement with
mens of 60 mm diameter and 120 mm long with com- water to cement ratio of 0.59. Compressive strength of
pressive strength of 24MPa, which were similar tests 35MPa was the target.
performed by Ballatore and Bocca (1997). The maximum The cylinder specimens were 10 cm in diameter and 20
stress ratios were at 0.2, 0.6 and 0.8 of the static strength cm long and were cast in steel molds. The specimens were
without cyclic loadings. The cyclic loadings of 3600 cy- demolded the next day, and were cured and sealed in
cles were applied with loading speed of 1Hz. The in- aluminum bags at 30°C for 4 months to prevent further
creases in the residual strength and stiffness were 5% and strength development during fatigue tests. After 4 months
6% respectively for the maximum stress ratio of 0.2, were of curing, cylinder specimens were still kept in aluminum
7.6% and 36% respectively for the maximum stress ratio bags until those were submitted to tests. It took 3 months
of 0.6, and were 11.4% and 36.1% respectively for the to finish all the tests. During cyclic loadings and residual
maximum stress ratio of 0.8. strength tests, the specimens were exposed to an atmos-
Ballatore and Bocca (1997) stated that dynamic loading phere in the testing laboratory.
reduces pore space to improve material compaction and Table 2 shows the maximum stress ratios and number
strain hardening takes place all the same time as is re- of load cycles before residual strength tests for each
flected in an increase in strength and stiffness after cyclic specimen. Three levels of the maximum stress ratios (0.75,
loadings. Bocca and Crotti (2003) stated that the test 0.65 and 0.33) were considered in the fatigue tests. The
pieces subjected to preloading yielded more homogenous minimum stress ratio was 0.05 and the loading speed was
results, with a reduction in standard deviation, compared 5 Hz in all cases. Figure 1 shows test cases in terms of the
to non-preloaded test pieces probably because of a de- number of load cycles and the maximum stress ratios in
crease in the stress peaks occurring in the material and a the S-N curves. During tests, the compressive strains were
redistribution of stresses due to cyclic load induced creep.
For the headed stud anchor, only one datum of the re- Cyclic Loading+Static Test
1 (10 specimens)
sidual tensile strength after cyclic loadings was found in Fatigue Test Max. Stress Ratio
Matsuzaki, et al. (1981). With the stud diameter of 19 mm, 0.8
(5 specimens) 0.75
Maximum Stress Ratio
0.65
0.75
embedded depth of 80 mm and concrete strength of Fatigue Test
(5 specimens) 0.33
0.65
23.5MPa, the residual tensile strength after two million of 0.6 Cylinder Specimen
load cycles was increased by 27% compared with the Cyclic Loading+Static Test (Uni-Axial Compression)
(10 specimens)
static tensile strength without cyclic loadings. The 0.4 Stud Anchor Specimen
0.33 (Corn Failure)
maximum and minimum stress ratio was 0.75 and 0.11
Cyclic Loading+Static Test
respectively with loading speed of 10Hz. 0.2 (15 specimens)
The residual bond strength between deformed bar and
0
concrete after cyclic loadings was found in Rehm and 10 0 10 4 10 8 10 12 10 16 1020
Eligehausen (1979). With the rebar diameter of 14 mm, Nunber of Load Cycles
embedded length of 42 mm and concrete strength of Fig. 1 Test Cases in terms of number of load cycles and
23.5MPa, the residual tensile strength after two million of the maximum stress ratio.
O. Kontani, N. Ishitobi, J. Kawada, N. Taogoshi, M. Koge and Y.o Umeki / Journal of Advanced Concrete Technology Vol. 14, 634-642, 2016 636
measured by two strain gages with a gage length of 9cm der uniaxial compression as shown in Eq. (1) (Kishitani
attached on the opposite side of cylinder specimens. The and Nishizawa 1987). The number of load cycles was set
loads applied to specimens were also measured. at two million for the maximum stress ratio of 0.33 be-
The maximum stresses were determined based on the cause the stress level of 0.33 was too low to expect fatigue
compressive strength obtained from the static tests of the failure within feasible time frame.
same cylinder specimens. In the tests described in Table 2,
the number of load cycles was applied to the cylinder s = 1 − log N /18 (1)
specimens with constant amplitude and then residual
where s: maximum stress ratio (Strass ratio to fatigue
compression tests were performed in order to investigate
failure at number of load cycles of N) and N:fatigue life
whether or not cyclic loadings reduce strength and stiff-
(number of load cycles to fatigue failure at the maximum
ness of concrete. The number of load cycles before com-
stress ratio of s)
pression tests was set to be 1/10 and 1/100 of the fatigue
life estimated from S-N curve of cylinder specimens un-
O. Kontani, N. Ishitobi, J. Kawada, N. Taogoshi, M. Koge and Y.o Umeki / Journal of Advanced Concrete Technology Vol. 14, 634-642, 2016 637
2.2 Test results the stress peaks occurring in the material and redistribute
Table 3 shows the list of results of residual compressive stresses due to cyclic load induced creep. These might be
tests after cyclic loadings. The compressive strength was the reasons of increase in strength and stiffness after cy-
37.3MPa, and the elastic modulus was varying between clic loadings.
31.1 kN/mm2 and 34.4 kN/mm2 from the compression The average elastic moduli were subjected to little ef-
tests without cyclic loadings. The fatigue life was evalu- fects of cyclic loadings at the maximum stress ratio of
ated to be 31,600 and all specimens exhibited fatigue 0.33, and were decreased as the number of load cycles
failure for the maximum stress ratio of 0.75, and to be increased at the maximum stress ratios of 0.65 and 0.75.
2,000,000 and three specimens out of five exhibited fa- Figure 2 shows the stress-strain relationships of the
tigue failure for the maximum stress ratio of 0.65. The static test without cyclic loadings, the cyclic loadings and
cylinder specimens that did not exhibit fatigue failure residual tests after cyclic loadings. The only representa-
during cyclic loadings showed a larger average residual tive cycles are shown for the cyclic loadings in Fig. 2. The
compressive strength than the compressive strength of figure shows the tendency that the residual strength after
specimens without cyclic loadings. The increase ratios are cyclic loadings is higher than that in the static test without
from 1% to 9%. These tendencies are similar to the pre- cyclic loadings. Even at the stress level of 0.33, there is
vious research found in Ballatore and Bocca (1997) and inelastic compressive strain. This can be explained by
Bocca and Crotti (2003). The explanation of the phe- cyclic load induced creep and material compaction.
nomena can be also found in Ballatore and Bocca (1997) Figure 3 shows the changes of the compressive strains
and Bocca and Crotti (2003). The cyclic loadings reduce at the maximum stresses with respect to the number of
pore space to improve material compaction and strain load cycles. The figure shows that although the compres-
hardening takes place. The cyclic loadings also decrease sive strain of cyclic loadings is proportional to logarithm
(CL-033-B-4) (CL-033-C-4)
(CL-065-B-4) (CL-065-C-4)
(CL-075-B4) (CL-075-C-4)
Fig. 2 Compressive load-displacement relationship.
O. Kontani, N. Ishitobi, J. Kawada, N. Taogoshi, M. Koge and Y.o Umeki / Journal of Advanced Concrete Technology Vol. 14, 634-642, 2016 638
of the number of load cycles up to 1/100 of fatigue life at It is preferable that the failure mode of the anchor sec-
the maximum stress ratios of 0.65 and 0.75, it suddenly tion is not brittle fracture relating to concrete fracture but
increases when the number of load cycles exceeds 1/10 of yielding of the anchor enabling the expectation of
fatigue life. Increasing number of load cycles increase toughness. Therefore, the concrete of the stud anchor
compressive strain, especially in cases with the same section seems to hardly be damaged by cyclic loadings
maximum stress levels. Compared with cases of the under the long-term allowable tensile load during the
maximum stress levels of 0.65 and 0.75, compressive normal operation of a nuclear power plant.
strains of the maximum stress levels of 0.65 are greater However, the pullout of anchors and the reduction of
than those of the maximum stress levels of 0.75 because the anchor strength due to cyclic loadings may reduce
the number of load cycles of 0.65 cases is much larger rigidity of the equipment support, and is not preferable
than 0.75 cases. For cases with fatigue failure, the com- from the viewpoint of long-term soundness evaluation.
pressive strain might not be measured up to the last load Therefore, the cyclic tensile loadings with constant am-
cycles because the strain gage came off due to fatigue plitude were applied to the stud anchor section specimens,
damages on the surface. In case of the maximum stress and then static pull-out tests were performed in order to
ratio at 0.33, the compression strain increases very investigate the reduction of rigidity and anchor strength of
gradually compared with cases of 0.65 and 0.75. the equipment support section due to cyclic loadings.
Figure 4 shows the changes of the elastic modulus ra-
tios with respect to the number of load cycles. The 3.1 Outline of tests
modulus ratios were obtained from the hysteresis curves Figure 5 shows a stud anchor specimen. No reinforce-
of cyclic loadings. The stiffness of the cylinder specimens ment was arranged in the cone region with an angle of 30°
from the second hysteresis curve is higher than that the in reference to the top surface of the specimen so that the
first hysteresis curve. Although the stiffness decreases as reinforcement doesn’t disturb formations of conical
the number of load cycles increases, the rate of decreasing cracks. The diameter of the headed stud anchor (material:
stiffness at the maximum stress ratio of 0.33 is very low JIS SCM435) was 19 mm, and the embedment depth of 90
compared with cases of 0.65 and 0.75. mm was adopted so that the failure mode due to tensile
load became the cone failure of concrete. Table 4 shows
3. Stud anchor specimen the mix proportion of concrete. The concrete was manu-
factured using gravel and sand produced in Shizuoka
This section reports the results of the residual strength Prefecture and ordinary Portland cement with water to
tests conducted using the stud anchor specimens in order cement ratio of 0.6.
to establish the method for evaluating the soundness of The anchor specimens were cast in the wood form and
equipment support against machine vibration. cured at 20°C for one month and then cured at ambient
temperatures varying between 3°C and 28°C for 6 months Since the influence of loading speed on fatigue strength is
to prevent further strength development during fatigue very small (Inoue et al. 1992), the speed of 10 Hz was
tests. The specimens were cured in sealed condition until adopted only for the case of 10 million load cycles to
those were submitted to tests. It took one month to finish reduce the test time. The pullout displacement of the stud
all the tests. During cyclic loadings and residual strength anchor was measured by displacement transducers. The
tests, the specimens were exposed to an atmosphere in the tensile loads applied to the anchor were also measured.
testing laboratory. Compressive strength of 35MPa was S-N curve of the stud anchor specimen can be found in
the target, and the compressive strength was 32–33MPa at Eligehausen et al. (2006) and was expressed in Eq. (3) in
the test. the same manner as Eq. (1).
0.4
jected area of concrete cone (mm2) expressed by 0.22
0.2
Ac = π ⋅ lc ⋅ (lc + D) , lc : embedment depth (mm) and D:
diameter of head of stud anchor (mm) 0.0
The minimum stress ratio of 0.05 was adopted for all 100
1.E+00 104
1.E+04 108
1.E+08 1012
1.E+12 1016
1.E+16 1020
1.E+20
ranging the maximum stress ratio and the horizontal axis Figures 9 and 10 show the test results of the specimen
for arranging the logarithm of load cycles, and is used for #1 at the maximum stress ratio of 0.75. In this case, the
assessing fatigue life corresponding to the maximum inelastic displacement is also observed from the first cycle
stress ratio. Figure 6 shows the maximum stress ratio and as shown in the load-displacement relation and the pullout
the residual strength ratio against the static test of the displacement of the anchor monotonously increases as the
specimen #7. In addition, the cyclic loading is expressed number of load cycles increases. Although cyclic loadings
by the arrow in the X-direction and the residual anchor stopped just before fatigue failure judging from the pro-
strengths are shown by the arrow in the Y-direction. All gress of the pullout displacement, the residual strength of
the residual anchor strength after cyclic loadings is equal the stud anchor was still higher than the static strength of
or greater by up to 17% than the static strength of the the specimen #7. The inelastic displacement of 1.1mm
specimen #7. This phenomenon has a tendency similar to was remained and the displacement at the peak load is
the results of compression tests of cylinder specimens higher than that of the static test of the specimen #7. In
after cyclic loadings. It is considered according to Bocca Fig. 9, the stiffness of residual anchor strength is greater
and Crotti (2003) that there must be conical cracking than the last hysteresis curve of cyclic loading. The reason
surface will be developed by pulling stud anchor and that might be recovery of remaining strain occurring in 15
the stress singularities in crack tips may be decreased and hours of rest period between the cyclic loadings and the
redistribution of stress occur due to cyclic loading in- residual strength test. The same kind of strain recovery of
duced creep. In this way, the residual anchor strength plain concrete can be found in Bennett and Muir (1967).
might be greater than the static strength without any cyclic In case of the maximum stress ratios of 0.65 and 0.75,
loadings. The residual anchor strength is getting greater although the anchor strengths after cyclic loadings are
with increasing number of loading cycles. higher than the static strength of the specimen #7, the
Figures 7 and 8 show the results of tests of the speci- pullout displacement monotonously increases as the
men #3 at the maximum stress ratio of 0.65 (the number of number of load cycles increases, which might cause the
load cycles of 719,413). In this case, the inelastic dis- reduction in the rigidity of the equipment support section.
placement is observed from the first cycle as shown in the Figures 11 and 12 show the results of tests of the
load-displacement relationship and the pullout displace- specimen #2 at the maximum stress ratio of 0.22 (load
ment of stud anchor monotonously increases as the cycles of 10 million). In this case, no inelastic displace-
number of load cycles increases. The inelastic displace- ment is observed in the load-displacement relationships in
ment of stud anchor of 0.5 mm is remained in the end of spite of a large number of load cycles, and moreover the
load cycles. The residual anchor strength and the pullout load-displacement relationships up to 40kN is similar to
displacement at peak load after cyclic loadings are greater that of the static test of the specimen #7.
than those of the static test of the specimen #7. The anchor specimen was designed so that cone failure
Fig. 7 Tensile load vs. pullout of stud anchor. (Specimen Fig. 9 Tensile load vs. pullout of stud anchor. (Specimen
#3) #1)
Fig. 8 Pullout of stud anchor vs. number of load cycles. Fig. 10 Pullout of stud anchor vs. number of load cycles.
(Specimen #3) (Specimen #1)
O. Kontani, N. Ishitobi, J. Kawada, N. Taogoshi, M. Koge and Y.o Umeki / Journal of Advanced Concrete Technology Vol. 14, 634-642, 2016 641
of concrete occurred when tensile loads are applied. The long-term allowable tensile load level, the fatigue life of
cone angle is approximately 35° on average according to the equipment support of would be 3.98 x 1015, which is
Eligehausen et al. (2006). In these tests, the some concrete much greater that of turbine support explained the above.
cones were broken into two to four pieces. It seems that In this way, soundness of equipment support can be ac-
the cone angle became smaller when conical cracking cessed.
surface is extended outward with increasing number of
load cycles. When pullout displacement is increased at 4. Summary
residual strength test, the concrete cone might be broken
into pieces because bending moment might be generated The residual compressive strength tests are performed on
between pullout of anchor and unseparated periphery of cylinder specimens with the number of load cycles and the
the concrete cone. maximum stress ratio as parameters. The residual strength
When the maximum stress ratio is less than the of cylinder specimens after cyclic loadings is higher by
long-term allowable tensile load defined by (AIJ 2010), 1% to 9% than the static strength without cyclic loadings.
the pullout displacement of the anchor is not observed at The stiffness of the cylinder specimens from the second
all even if subjected to load cycles up to 10 million. Then, hysteresis curve is higher than that the first hysteresis
it is considered that no reduction of rigidity and strength curve. The stiffness decreases as the number of load cy-
will be observed in the anchor section. cles increases. The compressive strain of cyclic loadings
The intercept of the X-axis of the fatigue strength is proportional to logarithm of the number of load cycles
formula of the anchor shown in Fig. 6 demonstrates that up to 1/100 of fatigue life, it suddenly increases when the
fatigue failure occurs at the number of load cycles of 1020 number of load cycles exceeds 1/10 of fatigue life. In-
even if the maximum stress ratio is zero. This is not real- creasing number of load cycles increase compressive
istic. However, according to Bennett (1967) concrete has strain, especially in cases with the same maximum stress
no fatigue limit under 10 million of load cycles differing levels.
from steel materials. The fatigue limit of concrete wasn't The residual strength tests are performed on headed
confirmed in the tests also. It is impossible to set fatigue stud anchor specimens with the number of load cycles and
limit for concrete. However, it might be reasonable to the maximum stress ratio as parameters. All the residual
employ S-N curve for predicting fatigue life of the anchor strength after cyclic loadings is equal or greater by
equipment support if S-N curve is extrapolated in the up to 17% than the static strength without cyclic loadings.
region of load cycles more than 10 million as shown in The residual anchor strength is getting greater with in-
Fig. 6. creasing number of loading cycles.
For example, the turbine generator at 60Hz region in In case of the maximum stress ratios of 0.65 and 0.75,
Japan will generates load cycles of 5.68 x 1010 for 60 although the anchor strengths after cyclic loadings are
years of operation. If the stress level is 0.22 equal to the higher than the static strength without cyclic loadings, the
pullout displacement increases monotonously as the
number of load cycles increases.
In case of the maximum stress ratio of 0.22, the anchor
strengths after cyclic loadings are higher than the static
strength without cyclic loadings and no residual dis-
placement is observed in the load-displacement relation-
ships in spite of a large number of load cycles and fur-
thermore the load-displacement relationships of the re-
sidual strength tests is similar to that of the static test
without cyclic loadings.
The results of residual strength tests of anchor speci-
mens after cyclic loadings revealed that no reduction of
Fig. 11 Tensile load vs. pullout of stud anchor. (Specimen rigidity and anchor strength is observed in the anchor
#2) section if the maximum load is not more than the
long-term allowable tensile load.
Acknowledgment
This research was performed as a part of the cooperative
research performed by electric industries for the common
purpose. We sincerely thank Professor Emeritus Katsuki
Takiguchi of Tokyo Institute of Technology, Dr. Hideo
Kasami, Professor Shigemitsu Hatanaka of Mie Univer-
sity, Professor Ichizo Kishimoto of Kinki University and
Professor Ippei Maruyama of Nagoya University who
Fig. 12 Pullout of stud anchor vs. number of load cycles. instructed us in the implementation of this study.
(Specimen #2)
O. Kontani, N. Ishitobi, J. Kawada, N. Taogoshi, M. Koge and Y.o Umeki / Journal of Advanced Concrete Technology Vol. 14, 634-642, 2016 642
Scientific paper
Abstract
Dry cask storage systems (DCSS) are widely used worldwide for storage of spent nuclear fuel (SNF). Particularly, in
the United States, other than the SNF pools, DCSS are the only means for storage of SNF. In the United States, the
DCSS are licensed for an initial 20 years (with a possible extension of 40 years). The absence of a long-term (or perma-
nent) storage facility has brought up concerns regarding the long-term performance of DCSS, which may now have to
be used for extended durations reaching over 100 years. The DCSS with an exposed concrete overpack account for ap-
proximately 61% of the DCSS inventory in the United States. The corrosion of the steel reinforcing bars (rebar) and the
alkali-silica reactivity (ASR) of concrete have been identified as two of the main degradation mechanisms. In this paper,
the accelerated aging of reinforced concrete (RC) overpacks of a vertical DCSS is evaluated experimentally at the struc-
tural scale. Three 1/3-scale specimens were fabricated. The first specimen was built using a conventional self-
consolidating concrete to serve as a control. The second and third specimens were prepared using special concrete mix-
tures, designed to accelerate the corrosion of rebar and ASR. All three casks were observed for 2 years for aging-
induced deterioration using various non-destructive approaches including visual inspection, half-cell potential, Schmidt
hammer, and ultra-sonic pulse velocity (UPV) measurements. The RC overpacks have been observed to exhibit signifi-
cant distress due to these aging mechanisms. The overall conclusion is that accelerating ASR and corrosion through use
of reactive aggregates and/or addition of chemicals (NaOH and CaCl2 in this particular case) is a viable and practical
approach for large-scale studies. Although accelerated aging of concrete structures have been extensively studied in the
literature, this is one of the first studies on the long-term degradation in DCSS due to corrosion and ASR.
Table 1 Options for dry storage of spent nuclear fuel (IAEA 2007a, b).
Option Heat Transfer Containment Shielding Feature
Double lid metal gasket
Metal cask Conduction through cask wall Metallic wall Dual purpose
(inert gas)
Air convection around Cavity lining / seal
Concrete cask/silo Concrete and steel overpack Vertical
canister welding (inert gas)
Concrete Air convection around Canister sealing
Concrete wall Horizontal
module canister (inert gas)
Air convection around
Vault Thimble tube (inert gas) Concrete wall Several cases
thimble tube
Drywell/ Heat conduction
Canister (inert gas) Earth Below ground
tunnel through earth
approximately 140,000 metric tons of waste in DCSS the long-term performance of DCSS, particularly, the
(U.S. GAO 2012), see Fig. 2. This will increase the impacts of material aging. The majority of commercial
number of DCSS to 9,300. DCSSs may be categorized SNF is stored in concrete storage casks or modules. As
into two: non-canistered designs where the fuel assem- of 2012, 60.6% of casks had an exposed reinforced con-
blies are directly loaded into the cask and canister-based crete (RC) overpack. Therefore, this study focused on
storage systems where fuel assemblies are contained in DCSS with an exposed RC overpack. Out of all DCSS
large multi-assembly canisters and inserted into an over- with an exposed RC overpack, 62.84% are in a horizon-
pack (Chopra et al. 2013). Certain cask configurations tal and 37.26% are in a vertical configuration (Lambert
are designed for dual purpose (both storage and trans-
portation) while others can only be used for storage. 4
x 10
These storages rely on concrete or steel, or both for 15
shielding the nuclear radiation and heat. Both vertical
Spent Nuclear Fuel (ton)
Power Temporary
Generation Storage
6 7
Fresh Fuel
Uranium Fuel
Mining Refining Conversion Enrichment Fabrication
Interim
1 Uranium Ore 2 Concentrate 3 Natural UF6 4 Enriched 5 8 Storage
Uranium
Recovered
Plutonium
Recovered
Uranium
11 9 Reprocessing
Recovered Plutonium
& Uranium
10
Permenant Storage &
Disposal Facilty
Fig. 1 A typical cycle of nuclear fuel. Note that the dashed lines indicate the cycle for SNF.
A. Atta, B. Gencturk, M. Hanifehzadeh and K. Willam / Journal of Advanced Concrete Technology Vol. 14, 299-310, 2016 301
Table 2 Degradation factors that can impact the performance of safety related RC structures for NPPs (IAEA 1998; Naus
2007).
Aging stressors Aging mechanism Aging effect
Percolation of fluid through concrete due Increased porosity and permeability;
Leaching and efflorescence
to moisture gradient lower strength
Exposure to alkali and magnesium sulfates
Sulfate attack Expansion and irregular cracking
present in soils, seawater or groundwater
Conversion of hardened cement to
Exposure to aggressive acids and bases Increased porosity and permeability
soluble material that can be leached
Combination of reactive aggregate, Alkali-aggregate reactions Cracking; gel exudation; aggregate
high moisture levels, and alkalis leading to swelling pop-out; disintegration
Cyclic loads/vibrations Fatigue Cracking; strength loss
Exposure to thermal cycles
Freeze/thaw Cracking; scaling; spalling
at relatively low temperatures
Exposure to flowing gas or liquid carrying
Abrasion; erosion; cavitation Section loss
particulates and abrasive components
Moisture content changes and
Cracking; spalling; strength loss; reduced
Thermal exposure/ thermal cycling material incompatibility due different
modulus of elasticity
thermal expansion values
Irradiation Aggregate expansion; hydrolysis Cracking; loss of mechanical properties
Exposure to water containing dissolved salts
Salt crystallization Cracking; loss material
(e.g., seawater)
Depassivation of steel due to carbonation or Compression or concentration Concrete cracking and spalling; loss of
presence of chloride ions cells leading to corrosion reinforcement cross section
et al. 2012; Leduc 2012). In this study, we focused on a DCSS due to higher temperatures inside the concrete
vertical configuration for its higher vulnerability to resulting from the decay heat of the SNF and being ex-
other events such as tip-over impact in the case of a posed to high levels of relative humidity in outdoor
hazard. conditions. As mentioned earlier, these two degradation
Like any other structure, DCSS should be capable of mechanisms are investigated in this paper for RC over-
performing their intended function (i.e., safe storage of packs of DCSS.
SNF) throughout their lifetime. DCSS are subjected to ASR in concrete were studied extensively using ac-
harsh environmental conditions including high internal celerated aging methods (Diamond and Ong 1994;
temperatures, radiation, and exposure to weathering Smaoui et al. 2005; Poole 2007; Temuujin et al. 2009;
effects from outside; therefore, aging management is Garci Juenger and Jennings 2011; Nuruddin et al. 2011;
necessary to maintain a high level of safety, and contin- Bahadure and Naik 2013; Gao et al. 2013; Kupwade-
ued operation. The degradation of cask materials de- Patil and Allouche 2013; Memon et al. 2013). However,
pends on the environmental conditions. As concrete these studies were conducted at the material level (or in
matures, its microstructure changes constantly, starting other words on small-scale specimens). In these studies,
with cement hydration and continuing with the crystalli- in addition to use of high levels of alkali, the specimens
zation of amorphous constituents, and reactions between were conditioned in high temperature with or without
cement, paste, aggregate and reinforcing steel. Concrete the application of sodium hydroxide solutions. Similarly,
may also deteriorate because of its interaction with the to accelerate corrosion, at the material level, different
environment and exhibit distress due to shrinkage, creep, method such as submerging small specimens in saline
corrosion of reinforcing steel, and alkali-silica reactivity solutions, passing a current or mixing salts with con-
(ASR) of aggregates. These mechanisms affect the con- crete were used (Diamond 1986; Helmuth et al. 1993;
crete, the steel and the interface properties, and reduce Pakshir and Esmaili 1998; Kelestemur and Yildiz 2006;
structural performance under mechanical loading such Wang et al. 2006; Anacta 2013). However, except for
as impact. IAEA (1998) and Naus (2007) address poten- the last one, these methods are not applicable for large-
tial aging mechanisms, age-related degradation, and scale investigations. There have been a handful of stud-
aging management for the concrete containment struc- ies in literature on accelerated aging of structural mem-
tures for NPPs, which can be extended to DCSS. The bers for ASR and corrosion (Folliard et al. 2006;
degradation factors and their effects are shown in Table Deschenes et al. 2009; Talley et al. 2016). These studies
2 (IAEA 1998; Naus 2007). were focused on structural components (e.g., bridge
One of the most common degradation mechanism in bent caps and columns). Therefore, the significance of
RC DCSS is the corrosion of steel reinforcement (rebar) this study is in the assessment of an important structural
(Naus et al. 1996). Most NPP (hence the sites for type under accelerated aging. To our knowledge, there
DCSS) are located close to a marine environment and are no other similar studies performed on DCSS over-
the wind carries the chloride ion into the land (Cicek packs where aging induced deterioration due to corro-
2013). ASR may also cause serious safety problems for sion and ASR were studied at this scale.
A. Atta, B. Gencturk, M. Hanifehzadeh and K. Willam / Journal of Advanced Concrete Technology Vol. 14, 299-310, 2016 302
Lid
Air
Outlet
Liner
Air
Inlet
Base
Plate
a) b) c)
Fig. 3 Cask model (a) steel components, (b) concrete shield, and (c) cask.
A. Atta, B. Gencturk, M. Hanifehzadeh and K. Willam / Journal of Advanced Concrete Technology Vol. 14, 299-310, 2016 303
Table 4 Properties of the concrete mixtures. models is shown in Fig. 5(c). The temperature and rela-
Mix SCC NaOH-8 CaCl2-4 tive humidity were continuously measured during the
Slump (mm) 258 229 275 testing period of 1.5 years, the temperature varied be-
Density (g/cm3) 2.328 2.320 2.349 tween 4°C and 37°C, and the relative humidity varied
Setting time (min) 285 240 180 between 30% and 100%.
Hydration temperature (°C) 60.71 59.21 63.95 The self-consolidating concrete mixtures were de-
signed to be cohesive enough to fill spaces of almost
ing was conducted. They were placed around the steel any size and shape without segregation or bleeding. The
liner in the Laboratory, see Fig. 5(a), before the con- results of the tests performed on the fresh concretes are
crete molds are assembled as shown in Fig. 5(b). The shown in Table 4. The results presented here are the
three concrete mixtures: two identical SCC mixtures average of at least three measurements. The slump, den-
with 25% flay ash replacement of cement and one SCC sity and setting time were measured based on ASTM
mixture with no fly ash replacement of cement were C1611 (2009c), ASTM C1688 (2013a), and ASTM
delivered separately in three ready-mix trucks. NaOH C403 (2008), respectively. The differences in the slump
powder was added to the fresh SCC mixture with no fly values were mainly attributed to differences in the con-
ash to accelerate ASR, and mixed inside the truck for 15 crete batches and mixture proportions; however, the
min before casting. Similarly, CaCl2 flakes were added addition of chemicals may have also altered the slump.
to one of the SCC mixtures with fly ash to accelerate It is noted here though that the latter is not explicitly
corrosion. Thereafter, casks were poured and formwork studied in this research. The density of the three mix-
was sealed with plywood and plastic sheets to prevent tures was found to be very similar. It was seen that the
rapid loss of moisture. Twenty-eight days after casting, setting time decreased substantially for the CaCl2 mix-
the formwork was removed. A picture of the three cask ture. The effect of NaOH on the setting time was less
#10
#10 76.2 mm
101.6
mm #10
#10
1016 mm
1828.8
mm
1930.4 mm
1143 mm
a) b) c)
Fig. 5 (a) Steel parts and rebar cage, (b) mold for concrete casting, and (c) three concrete casks.
A. Atta, B. Gencturk, M. Hanifehzadeh and K. Willam / Journal of Advanced Concrete Technology Vol. 14, 299-310, 2016 304
drastic. During casting of the concrete, the temperature ity index test ASTM C 1611(2014) was performed and
inside the concrete (at mid-point along the wall thick- all the mixtures were determined as highly stable. A
ness and close to the bottom of the cask) was measured, summary of material mechanical properties for the three
see Fig. 6. Adding calcium chloride to the mixture concrete mixtures are presented in Fig. 7.
slightly increased maximum hydration temperature from The aging-induced deterioration of the casks was
60.71°C to 63.95°C in comparison to SCC cask. The monitored through a series of non-destructive tests: vis-
sodium hydroxide did not have a significant effect on ual inspection, ultrasonic pulse velocity (UPV), and
the hydration temperature. Furthermore, a visual stabil- half-cell potential over a period of 1.5 years. The test
plan is given in Table 5.
CaCl2
the height of the casks, a Schmidt hammer test was per-
50
formed at 370 days. The Schmidt hammer measure-
40
ments were calibrated using the actual compressive
strengths obtained from material testing such that the
30
mean value along the height be the same as the average
of the material test results at 365 days. It is noted here
20 that the purpose of the Schmidt hammer test was to do a
0 1000 2000 3000 4000 5000 6000
relative comparison of the strength along the height of
Time (min)
the cask. To convert the Schmidt hammer index ob-
Fig. 6 Hydration temperature.
50 5
Splitting Tensile Strength (MPa)
Compressive Strength (MPa)
SCC
40 NaOH-8
4
CaCl -4
2
30 3
20 2
SCC
10 1 NaOH-8
CaCl2-4
0 0
a) 1 3 7 28 210 365 b) 1 3 7 28 210 365
Age (Days) Age (Days)
6
Modulus of Rupture (MPa)
2 SCC
NaOH-8
1 CaCl -4
2
0
c) 28 365
Age (Days)
Fig. 7 (a) Compressive strength, (b) splitting tensile strength, and (c) modulus of rupture for the three concrete mixtures.
A. Atta, B. Gencturk, M. Hanifehzadeh and K. Willam / Journal of Advanced Concrete Technology Vol. 14, 299-310, 2016 305
tained from the instrument, the cylinder strength was the forms for 28 days after pouring and therefore, the
used. It is not intended here to study the accuracy of loss of moisture was been prevented during this period,
Schmidt hammer testing or to relate the cylinder tests to which significantly reduces the drying shrinkage. (2)
actual field strength. The strength using a Schmidt The concrete was reinforced with steel at a ratio higher
hammer was measured at ten equally spaced locations than minimum shrinkage reinforcement requirement
along the height of the cask. Ten measurements were which reduces the likelihood of the cracks due to
collected around each location and the results were av- shrinkage. (3) SEM/EDS analysis were performed on
eraged. The results of these measurements are shown in concrete samples and gel formation was observed
Fig. 8. Before calibration, the SCC cask showed the around the aggregates. Observation of ASR cracks at 90
highest strength and NaOH-8 cask shows the lowest days of age indicates that the ASR process was acceler-
strength, which was in line with the material tests. In all ated considerably given that in regular structures it takes
three casks, the strength was found to decrease along the approximately 10 years until the appearance of first
height of the casks from bottom to top. This is attributed cracks (IStructE, 2010). The crack maps for ASR accel-
to the settlement of the coarse aggregate towards the erated cask at 517 and 591 days are shown in Figs. 9(a)
bottom of the casks since the casting was done from the and (b), respectively. It is seen that the density of crack-
top and a self-consolidating mixture was used. Although ing increased with time and higher crack density was
some variation along the height is observed, the strength observed towards the top of the cask potentially due to
at the bottom was approximately 17%, 28% and 30% lower concrete strength in that region as explained
higher than that at the top for the SCC, NaOH-8 and above. The crack widths remained between 0.05 mm
CaCl2-4 casks, respectively. and 0.1 mm over the measurement period.
Cracks have been observed on the NaOH cask ap- On the CaCl2-4 cask, the first cracks were visually
proximately 90 days after casting. Similar crack pattern observed after around 120 days. The crack propagation
to those from ASR can also be seen from different proc- over time on the CaCl2-4 cask is shown in Fig. 10. After
esses such as drying shrinkage, freezing/thawing cycles less than a year, there was extensive cracking on the
and sulfate attack. For the given exposure conditions in casks and several of the cracks had coalesced. The
this study, the only possible mechanisms left for such propagation of the cracks were in the same direction
cracks are the drying shrinkage and ASR (no presence with the rebar, which indicates that the cracks were due
of sulfates or temperature below freezing). To ensure to pressure generated by the corrosion products. The
and verify that the source of cracks on the casks is ASR, crack width was measured at 23 randomly chosen points
the following was performed. (1) The casks were kept in on the casks, and the change over time in the average
SCC NaOH-8 CaCl2-4 Height from the SCC Cask NaOH-8 CaCl2-4
base (mm) (MPa) Cask (MPa) Cask (MPa)
2000 2042 37.3 24.3 24.7
Height from the Base (mm)
a) b)
Fig. 9 Crack propagation on cask with NaOH-8 mixture (a) 517 days (b) 591 days.
A. Atta, B. Gencturk, M. Hanifehzadeh and K. Willam / Journal of Advanced Concrete Technology Vol. 14, 299-310, 2016 306
crack width is shown in Fig. 11. It is seen that at 569 probability of corrosion in each of the casks and the
days (i.e., the last measurement) the cracks were 0.33 results are presented in Fig. 12. At each measurement
mm wide on average with the largest and smallest date, more than 150 points were read on one-half of the
measurements being 0.1 mm and 1.0 mm. cask along a grid of lines. The measurements were con-
ASR and corrosion crack propagation is affected by sistent at different locations on the casks; therefore, in
the size, distribution and amount of the reinforcement. Fig. 12, only the average of these measurements (in a
Lollipop specimens were prepared with both #10 (bar 1.5 year time period) is shown. Half-cell potentials in
diameter of 9.52 mm) and #19 (bar diameter of 19.05 SCC and NaOH-8 casks were similar and always higher
mm) reinforcing bars with concrete covers of 38.1 mm than -200mV/CSE, which indicates that the probability
and 63.5 mm, respectively. In both cases of ASR and of corrosion is less than 10%. On the other hand, this
corrosion, the crack density was lower in specimens measurement for CaCl2-4 cask was always lower than -
with more concrete cover and larger size of reinforce- 350mV/CSE that shows that the probability of corrosion
ment. The increase in crack width due to corrosion in is more than 90%. As an overall trend, the half-cell po-
specimens with #19 rebar was less compared to those tential has decreased over time, which is an indicator of
with a #10 rebar. These findings are in line to what has increased corrosion activity in the CaCl2-4 cask.
been reported in the literature (Houston et al. 1972; The UPV was measured at three different levels and
Rasheeduzzafar et al. 1992; Du et al. 2014). on two sides of each cask as shown in Fig. 13. The av-
The half-cell potential was measured to evaluate the erage of UPV measurements over time is shown in Fig.
a) b)
c) d)
e) f)
Fig. 10 Crack propagation on cask with CaCl2-4 mixture (a) 163 days (b) 186 days (c) 225 days (d) 295 days (e) 512
days (f) 569 days.
0.7 -600
0.6 -500
Potential (mV/CSE)
Crack Width (mm)
0.5
-400
SCC
0.4
-300 NaOH-8
0.3 CaCl2-4
-200
0.2
-100
0.1
0
0 168 199 248 291 334 469 499 532 569
163 186 225 195 512 569 Age (Days)
Age (Days) Fig. 12 Half-cell potentials of reinforcing steel in three
Fig. 11 Crack width growth on CaCl2 cask. casks.
A. Atta, B. Gencturk, M. Hanifehzadeh and K. Willam / Journal of Advanced Concrete Technology Vol. 14, 299-310, 2016 307
14 along with the temperature and relative humidity approximately 36% at 28 days compared to control
measurements during the same period. These results self-consolidating concrete (SCC).
indicated a higher velocity in the SCC cask compared to • Similarly, it was observed that the direct addition of
NaOH-8 and CaCl2-4 casks, which had similar UPV CaCl2 is an effective way of accelerating the corro-
values. These findings were consistent with the material sion-induced aging. In this study, corrosion cracks
testing and Schmidt hammer measurements. These were observed at approximately 120 days of age with
measurements also indicate that there is a strong corre- 4% addition of CaCl2 by weight of cement into the
lation between the temperature and the UPV values. fresh concrete. However, it was seen that the concrete
However, since the variation of relative humidity was strength reduced by 24% at 28 days in comparison to
almost constant through time, its effect could not be the control SCC mixture. The CaCl2 and SCC mix-
quantified. The UPV results were not able to capture tures were identical except for the addition of chlo-
any degradation in concrete properties due to aging at ride; therefore, all of this reduction was due to the
this scale in contrast to what was initially thought be- presence of chloride.
cause of their sensitivity to moisture content, tempera- • The average width of cracks in CaCl2 cask increased
ture and the presence of rebar and other minor defects from about 0.1 mm to 0.3 mm in approximately 1 year,
inside the concrete around the measurement locations. and the maximum crack width reached to 1.0 mm.
• Rebound hammer tests showed an increase in strength
4. Conclusions towards the bottom of the cask. This is attributed to
the casting of the self-consolidating concretes from
Based on the findings of this study, the following con- the top, which resulted in a higher coarse aggregate
clusions are drawn: content towards the bottom due to settlement. This
• Adding NaOH to the concrete mixture, removing any variation in strength has also resulted in the observa-
supplementary cementitious materials with alkali- tion of higher crack density at the top of both NaOH
silica reactivity (ASR) suppression characteristics and CaCl2 casks.
(e.g., fly ash) and using reactive fine aggregate could • Adding chloride or alkaline to the mixture decreased
be used as an effective method to accelerate ASR at the setting time of the concrete by 37% and 16%, re-
the structural level. An 8% addition of NaOH by spectively in comparison to the control SCC mixture.
weight of cement into the fresh concrete reduced the Addition of chloride also increased the hydration tem-
time to first observation of ASR distress from tens of perature by 5%.
years to approximately 90 days. However, the con- • Although not studied directly as a part of this paper, it
crete strength with this special mixture reduced by is expected that environmental aging due to ASR and
corrosion observed in these tests will impact the struc-
tural performance of the casks, particularly, in the
Side 1
Level 1
case of impact (e.g., a tip-over event). On the other
hand, it is not anticipated that the thermal or radiation
shielding performance will be effected from environ-
mental aging. Testing and quantitative analysis on
Level 2 combined effects of environmental aging and me-
chanical or shielding performance remain as a future
45° research topic.
Level 3
Acknowledgements
Side 2
The financial support for this project was provided by
the United States Department of Energy through the
Fig. 13 Locations of UPV measurement on casks. Nuclear Energy University Program under the Contract
60
Temperature (C)
Ultrasonic Pulse Veloity (m/sec)
4600 SCC 40
NaOH-8
4400
CaCl2-4 20
4200
0
100 200 300 400 500 600
4000 b)
Age (Days)
Relative Humidity
100
3800
3600 50
3400
100 200 300 400 500 600
a) 0
Age (Days) 100 200 300 400 500 600
c) Age (Days)
Fig. 14 (a) UPV measurement on casks (b) temperature from 100 to 600 days (c) relative humidity from 100 to 600 days.
A. Atta, B. Gencturk, M. Hanifehzadeh and K. Willam / Journal of Advanced Concrete Technology Vol. 14, 299-310, 2016 308
No. 00128931. The findings presented herein are those S.-W., Fabian, R. R., Han, Z. and Liu, Y. Y., (2013).
of the authors and do not necessarily reflect the views of “Managing aging effects on dry cask storage systems
the sponsor. for extended jong-term storage and transportation of
used fuel.” FCRD-UFD-2013-000294, Argonne
References National Laboratory (ANL), Department of Energy
Anacta, E. T., (2013). “Effect of salt-contaminated (DOE), Lemont, IL.
mixing water and aggregates on time-to-initiate rebar Cicek, V., (2013). “Cathodic protection: Industrial
corrosion in concrete.” International Journal of solutions for protecting against corrosion.” Wiley-
Scientific & Engineering Research, 4(4), 1524-1527. Scrivener, Beverly, MA.
Andrews, A., (2006). “Radioactive waste streams: Deschenes, D. J., Bayrak, O. and Folliard, K. J., (2009).
Waste classification for disposal.” Report No. “ASR/DEF-damaged bent caps: Shear tests and field
RL32163, Congressional Research Service (CRS), implications.” Report No. 12-8XXIA006, Texas
Report for Congress. Department of Transportation (TxDOT), Austin, TX.
Andrews, A., (2008). “Nuclear fuel reprocessing: U.S. Diamond, S., (1986). “Chloride concentrations in
Policy Development.” Report No. S22542, concrete pore solutions resulting from calcium and
Congressional Research Service (CRS), The Library sodium chloride admixtures.” Cement, Concrete and
of Congress. Aggregates, 8(2), 97-102.
ASTM, (2008). “Standard test method for time of setting Diamond, S. and Ong, S,. (1994). “Effects of added
of concrete mixtures by penetration pesistance.” alkali hydroxides in mix water on long-term SO42-
ASTM C403/C403M-08, American Society for concentration in pore solution.” Cement & Concrete
Testing of Materials (ASTM), West Conshohocken, Composites, 16(3), 219-226, DOI: 10.1016/0958-
PA. 9465(94)90019-1.
ASTM, (2009a). “Standard test method for corrosion Du, X., Jin, L. and Zhang, R., (2014). “Modeling the
potentials of uncoated reinforcing steel in concrete.” cracking of cover concrete due to non-uniform
ASTM C876-09, American Society for Testing of corrosion of reinforcement.” Corrosion Science, 89,
Materials (ASTM), West Conshohocken, PA. 189-202.
ASTM, (2009b). “Standard test method for pulse Ferraris, C. F., (1995). “Alkali-silica reaction and high
velocity through concrete.” ASTM C597-09, performance concrete.” NISTIR 5742, National
American Society for Testing of Materials (ASTM), Institute of Standards and Technology (NIST),
West Conshohocken, PA. Gaithersburg, MD.
ASTM, (2009c). “Standard test method for slump flow Folliard, K. J., Barborak, R., Drimalas, T., Du, L.,
of self-consolidating concrete.” ASTM Garber, S., Ideker, J., Ley, T., Williams, S., Juenger,
C1611/C1611M-09b, American Society for Testing of M., Fournier, B. and Thomas, M. D. A., (2006).
Materials (ASTM), West Conshohocken, PA. “Preventing ASR/DEF in new concrete: Final report.”
ASTM, (2013a). “Standard test method for Density and FHWA/TX-06/0-4085-5, Center for Transportation
Void Content of Freshly Mixed Pervious Concrete.” Research at The University of Texas at Austin, Austin,
ASTM C1688/C1688M-13, American Society for TX.
Testing of Materials (ASTM), West Conshohocken, Gao, X. X., Multon, S., Cyr, M. and Sellier, A., (2013).
PA. “Alkali-silica reaction (ASR) expansion: Pessimum
ASTM, (2013b). “Standard test method for rebound effect versus scale effect.” Cement and Concrete
number of hardened concrete.” ASTM C805/C805M- Research, 44, 25-33.
13a, American Society for Testing of Materials Garci Juenger, M. C. and Jennings, H. M., (2011).
(ASTM), West Conshohocken, PA. “Effects of high alkalinity on cement pastes.” ACI
ASTM, (2014). “Standard test method for slump flow of Materials Journal, 98(3), 251-255.
self-consolidating concrete.” C1611/C1611M − 14, Helmuth, R., Stark, D., Diamond, S. and Moranville-
American Society of Testing and Materials Academic Regourd, M., (1993). “Alikali-silica reactivity: An
& Science (ASTM), West Conshohocken, PA. overview of research.” SHRP-C-342, Strategic
Bahadure, B. M. and Naik, N. S., (2013). “Effect of Highway Research Program (SHRP), Washington,
alkaline activator on workability and compressive DC.
strength of cement concrete with RHA.” Houston, J. T., Atimtay, E. and Ferguson, P. M., (1972).
International Journal of Computational Engineering “Corrosion of reinforcing steel embedded in
Research, 3(5), 15-20. structural concrete.” 112-1F, Texas Highway
CBO, (1977). “Nuclear reprocessing and proliferation: Department, US Department of Transportation
Alternative approaches and their implications for the Federal Highway Administration (FHWA), Austin,
federal budget.” Report No. O-89-442, The Congress TX.
of the United States, Congressional Budget Office Howard, R. and Akker, B. v. d., (2014). “Considerations
(CBO), Washington, D.C. for disposition of dry cask storage system materials at
Chopra, O. K., Diercks, D., Ma, D., Shah, V. N., Tam, end of storage system life.” Symposium on Recycling
A. Atta, B. Gencturk, M. Hanifehzadeh and K. Willam / Journal of Advanced Concrete Technology Vol. 14, 299-310, 2016 309
of Metals arising from Operation and Nuruddin, M. F., Demie, S., Ahmed, M. F. and Shafiq,
Decommissioning of Nuclear Facilites, April 8-10, N., (2011). “Effect of superplasticizer and NaOH
Nyköping, Sweden, D-4. molarity on workability, compressive strength and
IAEA, (1998). “Assessment and management of ageing microstructure properties of self-compacting
of major nuclear power plant components important geopolymer concrete.” International Journal of
to safety: Concrete sontainment buildings.” IAEA- Environmental, Chemical, Ecological, Geological
TECDOC-1025, International Atomic Energy Agency and Geophysical Engineering, 5(3), 187-194.
(IAEA), Vienna, Austria. Pakshir, M. and Esmaili, S., (1998). “The effect of
IAEA, (2007a). “Operation and maintenance of spent chloride ion concentration on the corrosion of
fuel storage and transportation casks/containers.” concrete.” Scientia Iranica, 4(4), 201-205.
IAEA-TECDOC-1532, International Atomic Energy Poole, J. L., (2007). “Modeling temperature sensitivity
Agency (IAEA), Vienna, Austria. and heat evolution of concrete.” Thesis (PhD). The
IAEA, (2007b). “Selection of away-from-reactor University of Texas at Austin, Austin, TX.
facilities for spent fuel storage, a guidebook.” IAEA- Rasheeduzzafar, Al-Saadoun, S. S. and Al-Gahtani, A.
TECDOC-1558, International Atomic Energy Agency S., (1992). “Corrosion cracking in relation to bar
(IAEA), Vienna, Austria. diameter, cover, and concrete quality.” Journal of
IStructE, (2010). “Addendum, The Institution of Materials in Civil Engineering, 4(4), 327-342.
Structural Engineers.” London, UK. Rigby, D. B., (2010). “Evaluation of the technical basis
Jones, F. E. and Tarleton, R. D., (1958). “Reaction for extended dry storage and transportation of used
between aggregates and cements.” Parts I-IV, DSIR nuclear fuel.” United States Nuclear Waste Technical
National Building Studies Research, HMSO, London, Review Board (NWTRB), Arlington, VA,
UK. http://www.nwtrb.gov/reports/eds_rpt.pdf.
Kelestemur, O. and Yildiz, S., (2006). “Effect of various Smaoui, N., Bérubé, M. A., Fournier, B., Bissonnette, B.
NaCl concentration on corrosion of steel in concrete and Durand, B., (2005). “Effects of alkali addition on
produced by addition of styrofoam.” Gazi University the mechanical properties and durability of concrete.”
Journal of Science, 19(3), 163-172. Cement and Concrete Research, 35(2), 203-212.
Kupwade-Patil, K. and Allouche, E., (2013). “Impact of Stanton, T. E., (1943). “Studies to develop an
alkali silica reaction on fly ash-based geopolymer accelerated test procedure for the detection of
concrete.” ASCE Journal of Materials in Civil adversely reactive cement-aggregate combination.”
Engineering, 25(1), 131-139. American Society of Testing and Materials
Lambert, J. D., Bakhtiari, S., Bondar, I., Kot, C. and Proceeding, 43, 875-893.
Pence, J., (2012). “NRC job code V6060: Extended Talley, K. G., Kapitan, J. G. and Breen, J. E., (2016).
In-situ and real time monitoring.” ANL/NE‐12/18, “Methods for approximation of ASR/DEF damage in
Argonne National Laboratory (ANL), Lemont, IL. concrete columns.” ACI Structural Journal, 113(1),
Leduc, D. R., (2012). Dry Storage of Used Fuel 105-110.
Transition to Transport.” FCRD-UFD-2012-000253, Temuujin, J., Williams, R. P. and Riessen, A. v., (2009).
Savannah River National Laboratory (SRNL), Aiken, “Effect of mechanical activation of fly ash on the
SC. Properties of gopolymer cured at ambient
Memon, F. A., Nuruddin, M. F., Khan, S., Shafiq, N. temperature.” Journal of Materials Processing
and Ayub, T., (2013). “Effect of sodium hydroxide Technology, 209(12-13), 5726-5280.
concentration on fresh properties and compressive U.S. GAO, (2012). “Spent Nuclear fuel, accumulating
strength of selfcompacting geopolymer concrete.” quantities at commercial reactors present storage and
Journal of Engineering Science and Technology, 8(1), other challenges.” GAO-12-797, United States
44-56. Government Accountability Office (GAO),
Naus, D. J., (2007). “Primer of durability of nuclear Washington, D.C.
power plant reinforced concrete structures - A review U.S. NRC, (2015a). “Backgrounder on dry cask storage
of Pertinent Factors.” NUREG/CR- of spent nuclear fuel.” United States Nuclear
6927, ORNL/TM-2006/529, Oak Ridge National Regulatory Commission (U.S. NRC), Available from:
Laboratory (ONRL), Washington, DC. http://www.nrc.gov/reading-rm/doc-collections/fact-
Naus, D. J., Oland, C. B. and Ellingwood, B. R., (1996). sheets/dry-cask-storage.html, Accessed on February
“Report on aging of nuclear power plant reinforced 10, 2016.
concrete structures.” NUREG/CR-6424 BNL- U.S. NRC, (2015b). “Backgrounder on licensing Yucca
NUREG-13148, Oak Ridge National Laboratory Mountain.” United States Nuclear Regulatory
(ONRL), Oak Ridge, TN. Commission (U.S. NRC), Available from:
NEI, (2014). “Guidance for operations based aging http://www.nrc.gov/reading-rm/doc-collections/fact-
management for dry cask storage.” Report No. NEI sheets/yucca-license-review.html, Accessed on
14-03, Nuclear Energy Institute (NEI), Washington February 10, 2016.
D.C. U.S. NRC, (2016). “NRC regulations, title 10 Code of
A. Atta, B. Gencturk, M. Hanifehzadeh and K. Willam / Journal of Advanced Concrete Technology Vol. 14, 299-310, 2016 310
Scientific paper
Abstract
In line with the strategic energy goals and sustainable development, considerable investment has been made on nuclear
energy, leading to construction of a growing number of nuclear power plants (NPPs) across China. Concrete is the main
structural materials applied in many key elements of a typical NPP including reactors. Among other stringent require-
ments, resistance to thermal loads both at early ages and during operating life is crucial for the concrete used in nuclear
facilities. This paper provides an overview of high temperature resistance of concrete structures in Chinese NPP, in-
cluding information on construction and operation. Early-age thermal cracking of mass concrete structures due to dif-
ferential thermal stresses induced by hydration heat has been highlighted in literature as a major issue. The common
approaches to tackling this issue are based on limiting the maximum temperature in the concrete and temperature dif-
ference between hotter interior and cooler exterior of the concrete. These approaches include the optimum mix ratio
method and the reasonable design of the construction technology. During operation, the temperature control of con-
tainment under steady or accidental case is an important prerequisite to ensure the safety of NPP, especially after a loss of
coolant accident (LOCA). In addition, various coatings used for thermal insulation in China’s NPP are introduced and
compared. Based on the existing literature, further studies involving longtime monitoring of temperature, strain and
displacement are found necessary to obtain a better understanding of thermal resistance, long-term performance and
safety of concrete containment used in NPP.
2. Temperature control during construction (NPPs) (Zhao and Zhou 2012). In order to limit the
maximum temperature and the temperature difference
The hydration of Portland cement is an exothermic re- between inside and outside of the concrete within the
action. In mass concrete, (e.g., concrete containment and prescribed standard, the concrete temperature should be
foundation), where the surface area of concrete is rela- reduced from two aspects as the optimum mix design and
tively small compared to its total volume, the hydration the reasonable construction technology.
heat may result in considerable temperature difference
between interior and exterior of concrete. The thermal 2.1 Concrete mix design
stresses caused by such temperature gradients can readily Hydration heat is the main reason that the volume of
induce the development of early-age thermal cracks in concrete grows larger when the temperature increases.
concrete, which is highly undesired from both aesthetic Medium or low heat cement can be chosen to control the
and durability points of view (Acker and Ulm 2001; hydration heat of binding materials in concrete strictly.
Wang 2011). Temperature control of concrete after its Meanwhile, adding mineral admixture and chemical
casting is considered to be the key to limit the concrete additive should also be helpful (Xu 2010). The mineral
thermal cracking (Ling et al. 2005; Schindler 2004). The admixture can partly replace the Portland cement to
delayed ettingite formation (DEF) because of high tem- lower and delay the peak temperature of hydration.
perature and the resulting stresses generated by the ex- Furthermore, some chemical additive such as Polycar-
panding ettringite have been particularly identified as a boxylate water-reducing agent (PCA) can retard cement
cause of concrete cracking, especially in European hydration and increase the dosage of slag or fly ash in
pressurized-water reactor (EPR) nuclear power plants concrete. The concrete proportion of the reactor building
J. Xiao, W. Xie and Q. Xie / Journal of Advanced Concrete Technology Vol. 14, 335-343, 2016 337
raft of Taishan and Yangjiang NPPs can be taken as a section of concrete and stress orientation in this section
typical example, which is listed in Table 3. The concrete are determined to provide the guidance of the position
unit weights of both Taishan and Yangjiang NPPs meet design of sensors. The data from the sensors could pro-
the standard of normal-weight aggregates in Table 4, vide useful information for the effective adjustment of
which have been used in the biological shields of pres- maintaining measures. When the temperature in concrete
surized-water reactors (PWRs) (Kaplan 1989; Willam et is overhigh, some cooling measures can be adopted like
al. 2013). ice bag cooling, rebar covering, spraying on pouring pipe
during stratified slope concreting and adjusting the cov-
2.2 Reasonable construction scheme ering layer thickness based on the real-time monitoring
During the maintaining process of the mass concrete, data in the conservation process. The research shows that
temperature measurement is the main gist of temperature there can be no obvious cracks even when the tempera-
control to decide the maintenance time of concrete, and ture reaches 80 °C.
to adjust the layers of the proof material. This procedure Considering the enormous pouring volume of con-
can limit the temperature difference and cooling rate tainment of Taishan NPP with a height of 48.95 m, the
inside the concrete, and thereby can prevent the devel- area is divided into 16 layers, where the thickest one is
opment and growth of cracks. During the construction of 1.8 m. The outer containment has an inner diameter of 53
the foundation in Hongyanhe NPP, a temperature moni- m and an outer diameter of 55.6 m, with the C45 massive
toring and controlling system was applied to avoid the concrete of 1237.6 m³ in all (Liu 2014). Through the
negative influence of winter weather and to assure the numerical simulation, Liu (2014) found that the center of
program on schedule. In this strategy, 54 temperature containment wall had the highest temperature over 70 °C
sensors and 57 strain sensors were set to monitor the while the surfaces were much cooler. The temperature of
temperature and stress development, along with an alert inside surface was the same as the outside surface, and
system being triggered when stress got close to the ten- they were symmetrical by the center. According to the
sile strength of concrete (Yang et al. 2009; Wei et al. existing construction layering arrangement, the thickness
2008). of construction layer rather than the height could have a
In the southern coastal area of China, the high tem- significant influence on the temperature field. For a 1.8
perature and humid environment brings many difficulties m thick layer, the center temperature of containment wall
to the pouring and maintaining process of NPP founda- could reach 75 °C, which was 5 °C higher than that of the
tion, such as overhigh central temperature and great 1.3 m one. The maximum vertical temperature gradient
temperature gradient, which are very difficult for con- would reach 71.8 °C/m, and the maximum radial tem-
trolling cracks. Gu et al. (2010) undertook an intensive perature gradient would reach 50.3 °C/m. These posi-
research on controlling cracks during the maintaining tions with the maximum temperature gradient in two
period by taking some NPPs in south China as examples. directions should be the key point during the construction.
A finite element numerical simulation model was built by It is advised that maintaining time should be increased to
ANSYS to compare with the monitoring data of tem- 10 days since increasing the maintaining time is one of
perature and strain during the pouring of concrete. Ac- the effective ways to reduce the temperature difference.
cording to the numerical calculation results, the control The worse the insulation is, the much bigger the differ-
J. Xiao, W. Xie and Q. Xie / Journal of Advanced Concrete Technology Vol. 14, 335-343, 2016 338
ence between the inside and outside surface before Qinshan NPP and established the differential equations.
hardening will be. Because the thermal conductivity of The temperature field was established according to the
concrete suddenly rises after hardening, the surface calculated temperature at each time point. By compari-
temperature reduces more rapidly with a higher tem- son, the calculated temperature were consistent with the
perature before hardening. The better the insulation ma- measured temperature. For example, both the measured
terial is, the much bigger the difference between the temperature and the calculated temperature of 62.5 m
inside and outside surface after hardening will be (Liu elevation on the surface of containment fluctuated in the
2014; Qian et al. 2014). Therefore, adopting proper in- range of 20~45 °C.
sulation material can help to prevent the temperature According to weather records of several years in the
difference from exceeding the limit in the China stan- local experiment place and the existing data of air tem-
dards of GB 50496-2009. perature in the containment, Lin and Yan (2003) revealed
As the temperature rises after pouring, it is necessary the variation, distribution pattern and the magnitude of
to cure the concrete by wet maintenance methods to the effects of temperature field based on the finite dif-
guarantee the hydration reaction of binding material ference method. When the maximum of temperature
inside concrete. This can reduce the peak temperature difference reached 20 °C on the surface, the temperature
inside the concrete and avoid the loss of strength in the varied within 1 °C in the middle thickness. Due to the
later period. When concrete begins to cool down, the influence of projection angle, duration of sunlight and
structure shrinks and the compression state of concrete other factors, the surface temperature on the top of con-
turns into tension state. Penetrating cracks may generate tainment was higher than that of the tube wall, and the
due to the restraints, which can break the integrality and variation range was larger (Fig. 1). The calculation re-
durability of structures (Ariyawardena et al. 1997). sults of temperature field were then introduced to the
Therefore, heat and humidity preservation measures structure analysis program, and results showed that the
should be taken to avoid the occurrence of penetrating structure deflection was less than 0.3 ~ 0.5 mm under the
cracks during the cooling stage. atmospheric environment for less than 10 days. In the
atmospheric environment, the containment temperature
3. Thermal safety of containment decreases rapidly and distributes nonlinearly along
thickness. The temperature difference of containment
The containment is the retaining structure of a nuclear surface has little effect on the temperature and thermal
reactor, which is the third safety barrier in addition to fuel stress in the inner (Singh and Heller 1980).
rod cladding and pressure vessel. According to the With the increasing operation time of NPP, the con-
composition of materials, the containment structure crete structure and facility are bound to deteriorate.
system of nuclear reactor can be classified as steel con- Therefore, the reactor must undergo in-depth inspection
tainment, reinforced concrete containment and and test at intervals while the containment is evaluated
prestressed concrete containment. The function of con- by monitoring and inspecting. In order to ensure the
tainment is to restrict and eliminate the release of ra- security and safety of NPP, the research on the behavior
dioactive fission products to ensure the safety of NPP and of containment after withstanding longtime radiation and
the public when the loss of coolant accident (LOCA) temperature variation need to be conducted by meas-
caused by the rupture of primary coolant pipe happens urements of temperature, strain, and displacement.
(Zhao et al. 2003; Costaz and Danisch 1997). The tem-
perature effect has great influence on overall mechanical 3.2 Thermal resistance of containment after
properties of the containment structure, and calculation LOCA
of the containment temperature field is an important LOCA releases high temperature steam and water, in-
prerequisite to ensure the safety of NPP under steady and ducing an equivalent pressure of fluid with 150 °C acting
accident situation.
50
3.1 Thermal response of containment in at- In the top of containment
mospheric environment At 47.8 m of containment
40
Temperature (°C)
on the containment surface. Then the temperature rises The distribution of temperature field in the concrete
suddenly in the containment, resulting in a greater tem- containment becomes obviously nonlinear after LOCA.
perature difference between inside and outside, which Based on the analysis by Zhao et al. (2003) about the
brings about considerable temperature stress and influ- containment temperature change for PWR, Wu et al.
ences the distribution of temperature and stress (Li and (2010) and Sun et al. (2015) respectively carried out the
Wang 2010; Kanzleiter 1976). Therefore, the design of finite element simulation of concrete containment after
containment should take the effect of LOCA into con- LOCA. The parameters are listed in Table 5 and the
sideration, especially in the type of PWR NPP which is results are shown in Fig. 3. Due to the differences of
applied widely in China. factors including containment wall thickness, thermal
Li and Wang (2010) from the Central Research Insti- conductivity of concrete and temperature difference
tute of Building and Construction of Metallurgical Cor- inside and outside the containment, the extent of tem-
poration of China Group analyzed the temperature field perature increase in outer wall of containment is quite
and stress field of a 1.2 m-thick containment structure different when it is 6 h after the occurrence of LOCA,
under the temperature effect of the LOCA by ANSYS. while the outside temperature of containment with
Assuming that the inner surface of containment was thickness of 1.1 m does not obviously change. After
adiabatic and the heat transmits from the inner surface to obtaining the temperature field in containment, the inner
the outer surface, the instantaneous temperature of force distribution and the deformation of containment in
150 °C was applied to node on the inner wall of the different time can be derived from theoretical analysis or
containment. According to the analysis results, the tem- numerical simulation:
perature of each node at different distance from the inner (1) 6 h after the occurrence of LOCA, the bending mo-
wall is shown in Fig. 2. It can be found in Fig. 2 that the ment reaches its maximum value according to the
temperature only influences a thin part of the tube wall. theory of elasticity, ignoring the confinement effect
Affected by accidental instantaneous temperature, the on structure bottom. Thus the temperature field at
high stress is mainly concentrated at the fixed bottom of this time can act as a benchmark for internal force
the cylinder, at the top of the dome and around the design. In practical engineering, a reduction coeffi-
opening, which does not reach the tensile strength of cient ranging from 0.3 to 0.5 should be adopted for
concrete. So the effect of accidental instantaneous tem- the calculation of thermal stress to consider the
perature does not cause destructive damage to concrete concrete cracking and creep effect (Wu et al. 2010).
structures. (2) The temperature effect has a great influence on the
160 160
0m Thickness 1.1m/10min(Wu et al. 2010)
140
Thickness 1.1m/6h (Wu et al. 2010)
120 0.1m 120 Thickness 0.9m/10min(Sun at al. 2015)
Temperature(°C)
Temperature (°C)
80 0.6m 80
60
40 40
20
0 0
0 1800 3600 5400 7200 9000 10800 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Time(s) Relative thickness
Fig. 2 Distance and temperature curve (Li and Wang Fig. 3 Variation of temperature along relative thickness at
2010). different time after LOCA.
J. Xiao, W. Xie and Q. Xie / Journal of Advanced Concrete Technology Vol. 14, 335-343, 2016 340
Table 6 Main types of coatings for Qinshan I NPP Project (Gong 2009).
Category Concrete structure Steel structure Steel equipment
Red polyurethane+ Inorganic zinc silicate + Inorganic zinc silicate +
Containment
polyurethane polyurethane polyurethane
Epoxy primer +
Radiation control area Epoxy alkyd
polyurethane
Epoxy zinc-rich +
Security equipment Epoxy, polyurethane Alkyd, epoxy asphalt etc.
alloprene
deformation pattern of the overall containment de- Qinshan I NPP Project. Based on the successful applica-
formation. When a LOCA occurs, the maximum tion of these coatings, corresponding standard of special
expansion displacement of containment structure coating for NPP is compiled referring to relevant con-
increases nearly 14 mm and the overall expansion tents in the American Society for Testing and Materials
displacement also increases 6 ~ 10 mm under the and Norme Française standards. At present, there are few
combination of internal temperature and pressure researches on the aging and life prediction of the coating
(Sun et al. 2015). on the NPP in China, while the focus is placed on the
If the pressure vessel fails, the integrity of containment production and application of the coating (Liu 2015).
will be jeopardized. The task of top priority is to cool In the test of the performance of the coating, simulated
down and seal the corium inside the pressure vessel. As Design Basis Accident (DBA) is included as well as
an important material used in core catch, sacrificial con- coating irradiation test and coating decontamination test.
crete can melt and be mixed with the corium, which will DBA simulates the effect of reactor on the coating for
bring the benefits of reducing solidus and liquids tem- containment under the condition of design basis accident.
perature of corium. Therefore, research on the prepara- The coating sample is placed in a high temperature and
tion, function mechanism of nuclear sacrificial concrete high pressure vessel sprayed with boric acid aqueous
is of great significance to ensure that containment re- solution with pH=9 ~ 10. The temperature, pressure,
mains intact. The sacrificial concrete designed by period and other conditions can be determined according
Southeast University in China has been applied in to different standards or specific design parameters. The
Taishan NPP (Chu et al. 2016). typical condition is a 4 ~ 7 d period experiment under
153 °C with a certain pressure. After the test, the sample
4. High temperature resistant coating for coating sample will be considered as not qualified if the
NPP cracking, delamination, bubble or breakage of the coat-
ing occurs (Zhang 2007).
Coating is a continuous film formed by spraying all paint Chi et al. (2006) made a comprehensive check and
at once to the surface of solid, usually referred to as the evaluation on the aging condition of No.1 reactor factory
substrate, such as metal, fabric and plastic. It is a plastic in Daya Bay NPP after an operation time of over 10 years.
thin layer for protection, thermal insulation, decoration Through field test, the coating adhesion of steel structure,
etc. Coating of NPP is mainly used for steel and concrete walls, floors and ceilings surface conformed to the re-
structure, pipes and other components on the nuclear quirement of the original design of the NPP. The coatings
island and the conventional island. Although the corro- of the steel panels and concrete blocks had a poor con-
sion resistance is one of the most basic requirements of dition with destructive defects caused by scratch, grind,
the coating performance, strict process control has hit, etc. Those test boards were prepared during the con-
guaranteed the basic requirements of anticorrosion for struction of Daya Bay NPP, which were later used as
the special nuclear power coating. specimens for LOCA tests. Specimens were examined at
During the operation period, the temperature of the 2 hours and 2 weeks after LOCA tests respectively, and
concrete containment is about 50 °C, and the radiation is the adhesion test were performed after 2 weeks. Ex-
0.1 Gy/h. However, the temperature is likely to rise to cluding the effect of preexisting defects, the steel panels
more than 150 °C under accidental conditions with nu- and concrete blocks reached the requirements for ac-
merous of radiation particles generating. Therefore, the ceptance in Table 7.
coating will experience a continuous aging process due The passive containment used in the third-generation
to the high temperature and radiation, which may jeop- NPP is composed of reinforced concrete structure and
ardize its original function in the normal operation of steel containment. The inner and outer surfaces of steel
NPP or accidental conditions (Chen 2013). containment are covered by inorganic zinc coating with
A great deal of research and development work have excellent heat transfer performance (Kindred and Wright
been done by the research institutes and suppliers in 2014). Different from the aging characteristics of ordi-
China. The coatings for Qinshan I NPP Project are pro- nary epoxy coatings, the inorganic zinc coating will
vided by the Shanghai Kailin paint factory and the generate zinc oxide during aging, and the oxidation sub-
Changzhou Coating Chemical Research Institute (Ni et stances can continue to fill pores in the coatings. Because
al. 2005). Table 6 gives the main types of coatings for the thermal conductivity of zinc oxide is better than that
J. Xiao, W. Xie and Q. Xie / Journal of Advanced Concrete Technology Vol. 14, 335-343, 2016 341
of coating pores, the heat transmission of inorganic zinc (3) To complete further analysis (such as the research on
coating will not be affected adversely by aging (Liu et al. the whole life appraisal of containment), the longtime
2015). According to the physical environment that might measurements of temperature, strain, and displace-
form in the LOCA accident, Fan et al. (2015) conducted ment are advised to adopt for the safety of contain-
a simulation experiment on test panel with inorganic zinc ment.
coating. Dry film thickness and adhesion before and after (4) In China, some institutions already have the ability to
the LOCA accident are shown in Table 8. A single research and produce the special coating for NPP.
simulation of the LOCA accident does not have a sig- They have compiled the norms of NPP coating by
nificant impact on the dry film thickness and adhesion of referring to the norms of France and USA to provide
inorganic zinc coating, and the coating still meets the technical support for the safe application of the NPP.
requirements of serviceable capacity, such as the integ- Moreover, the damage of coating is inevitable and
rity, heat transfer and radiation resistance. needs alleviation during the operation of generator
units. Therefore it is necessary to compile the norms
5. Summary of coating maintenance and assessment based on the
situation of the coating of NPP.
This paper reviewed the researches and technical meas-
ures for minimizing the risk associated with concrete in Acknowledgements
Chinese NPP exposure to high temperatures. The main The authors would like to gratefully acknowledge all the
conclusions are summarized in the following: national and international investigators related to the
(1) The concrete component should be optimized for the references and the research grants from the Chinese
construction of foundation and main structure of NPP. National 973 Plan (2012CB719703).
While temperature and strain sensors are used for
real-time monitoring of the temperature of each Reference
measuring point, appropriate insulation measures can Acker, P. and Ulm, F., (2001). “Creep and shrinkage of
be taken to prevent cracking due to excessively high concrete: physical origins and practical
temperature or excessive temperature gradient in measurements.” Nuclear Engineering and Design,
concrete. 203(2), 143-158.
(2) Finite element analysis can be used to predict the Ariyawardena, N., Ghali, A. and Elbadry, M., (1997).
temperature field of the containment vessel under the “Experimental study on thermal cracking in
conditions of normal temperature and LOCA accident, reinforced concrete members.” ACI Structural
which can provide reference for the internal force Journal, 94(4), 432-441.
analysis and the structural design of the containment. Chen, Z., (2013). “Study of aging and lifetime for the
J. Xiao, W. Xie and Q. Xie / Journal of Advanced Concrete Technology Vol. 14, 335-343, 2016 342
epoxy coatings in nuclear power plant.” Shanghai Liu, S. H., (2014). “Research on construction
Jiao Tong University. (in Chinese) temperature field characteristics and effect of nuclear
Chi, Z. H., Tang, M. L. and Cui, L., (2006). “Evaluation island outer containment.” Harbin Institute of
of aging of inner coating of containment vessel of Technology. (in Chinese)
Daya Bay nuclear power plant.” National Nuclear Liu, X. Q., Xu, X. L. and Meng, F. J., (2015).
Material Academic Communication 2006. (in “Reliability and design analysis of coatings used in
Chinese) passive containment vessel of nuclear power plant.”
Chu, H., Jiang, J., Sun, W., and Zhang, M., (2016). Paint & Coatings Industry, 45(4), 74-78. (in Chinese)
“Thermal behavior of siliceous and ferro-siliceous Ni, A. B., Ding, W. Z. and Wang, L. F., (2005).
sacrificial concrete subjected to elevated tempera- “Development of heat-resistant coatings for nuclear
tures.” Materials & Design, 95, 470-480. facilities and equipment.” Paint & Coatings Industry,
Costaz, J. L. and Danisch, R., (1997). “Discussion on 35(11), 7-9. (in Chinese)
recent concrete containment designs.” Nuclear Qian, H. L., Qin, F. and Fan, F., (2014). “Analysis on
Engineering and Design, 174(2), 189-196. temperature filed of outer containment of the Taishan
Fan, H. C., (2015). “Study on the characteristic of nuclear power plant.” Journal of Building Structures,
inorganic zinc coating for passive containment.” 35(7), 115-122. (in Chinese)
Surface Technology, 6, 7-10. (in Chinese) Schindler, A. K., (2004). “Effect of temperature on
GB 50496-2009, (2009). “Code for construction of mass hydration of cementitious materials.” ACI Materials
concrete.” Beijing: Standards Press of China. (in Journal, 101(1), 72-81.
Chinese) Schneider, M. and Froggatt, A., (2015). “World nuclear
Gong, Y. Z., (2009). “Review on the research and industry status report 2015.”
development of China nuclear power coatings.” Sekimura, N., (2012). “Securing the stability of ageing
China Paint, 24(6), 5-8. (in Chinese) NPP considering Fukushima Dai-Ichi lessons
Gu, H. M., Cheng, D. Y. and Zhang, X. B., (2010). learning: improving inspection system and ageing
“Monitoring of temperature strain during pouring raft management and planning.” 3rd International
foundation for a nuclear island of a nuclear power Conference on NPP Life Management (PLIM) for
plant in hot and humid environment.” Industrial Long Term Operations (LTO), Salt Lake City, U.S.A.
Construction, 40(1), 17-20. (in Chinese) Singh, M. P. and Heller, R. A., (1980). “Random
Kanzleiter, T. F., (1976). “Experimental investigations thermal stress in concrete containments.” Journal of
of pressure and temperature loads on a containment the Structural Division, 106(7), 1481-1496.
after a loss-of-coolant accident.” Nuclear Sun, F., Pan, R. and Chai, G. H., (2015). “Preliminary
Engineering and Design, 38(1), 159-167. study on response law of prestressed concrete
Kaplan, M. F., (1989). “Concrete radiation shielding: containment after LOCA in certain NPP.” Atomic
nuclear physics, concrete properties, design and Energy Science and Technology, 10, 1815-1820. (in
construction.” Essex: Longman Scientific & Chinese)
Technical. Wang, G. T., (2011). “Preparation of self-compacting
Kindred, T. A. and Wright, R. F., (2014). “Effects of concrete used in nuclear power plant and its
aging on the thermal conductivity of the AP1000® application in cut-off cone structure.” South China
containment vessel inorganic zinc coating.” 2014 22nd University of Technology. (in Chinese)
International Conference on Nuclear Engineering, Wang, H. C., Chen, L. S. and Lv, G. Y., (2012).
American Society of Mechanical Engineers, “Preparation and construction technology of nuclear
V001T03A028-V001T03A028. power plant high-performance mass concrete.”
Li, J. W. and Wang, Y. H., (2010). “Calculation and Nuclear Power Engineering, 32(z2), 47-50. (in
analysis of temperature and stress fields for nuclear Chinese)
power plants under temperature effect of LOCA Wei, J. G., Zhang, Z and Zhang, X. B., (2008).
accident.” Industrial Construction, S1, 300-305. (in “Temperature-controlling monitoring and emulation
Chinese) analysis of mass-concrete of raft foundation of
Lin, Z. S. and Yan, J. H., (2003). “The temperature field nuclear power plant.” Industrial Construction, z1,
of PCCV of nuclear power station under atmosphere 1033-1035. (in Chinese)
circumstance.” Journal of Building Structures, 36(6), Willam, K., Xi, Y. and Naus, D., (2013). “A review of
12-17. (in Chinese) the effects of radiation on microstructure and
Ling, D. S., Xu, D. S. and Shen, Y. Y., (2005). “Model properties of concrete used in nuclear power plants.”
for hydration heat of cement in concrete and its US NRC, NUREG/CR-7171.
application.” Journal of Zhejiang University, 39(11), Wu, C., Meng, S. P. and Zhou, Z., (2010). “Analyses on
1695-1698. (in Chinese) temperature field distribution and internal force of
Liu, G. Z., (2015). “Research on immersed thermal concrete containment vessel after LOCA.”
aging of coatings in nuclear power plant.” Shanghai Engineering Mechanics, 27(12), 206-212. (in
Jiao Tong University. (in Chinese) Chinese)
J. Xiao, W. Xie and Q. Xie / Journal of Advanced Concrete Technology Vol. 14, 335-343, 2016 343
Wu, Y., Shi, L. and Liu, J. Z., (2013). “Gas tightness Chinese Society for Electrical Engineering, 32(11),
evaluation of structural concrete applied to nuclear 1-8. (in Chinese)
containment vessels.” Applied Mechanics & Zhang, Y., (2007). “Analysis on application of special
Materials, 405, 19-23. (in Chinese) coating for nuclear power plant.” 1st Symposium on
Xie, Y. J., Xue, J. L. and Fu, B. L., (1993). “The study Special Coating and Application, 210-214. (in
of the structural integrity test on security containment Chinese)
volumn of Qinshan nuclear power station.” Industrial Zhao, J., Yu, H. X. and Li, F., (2003). “Calculation and
Construction, 3, 36-41. (in Chinese) analyses of containment pressure and temperature for
Xu, C., (2010). “Temperature control measure used in PWR1000XL.” Nuclear Power Engineering, 24(5),
HRA reactor building of Taishan NPP.” Concrete, 12, 409-411. (in Chinese)
114-117. (in Chinese) Zhao, P. and Zhou, C. G., (2012). “Optimization of
Yang, Z. Y., Wei, J. G. and Ji, L. B., (2009). “Integral construction mix proportion based on raft foundation
pouring construction technology of raft foundation A concrete of DEF nuclear power plant.” Henan
and B layers in Hongyanhe nuclear power station Technology, 17, 88-90. (in Chinese)
reactor workshop.” Construction Technology, 38(8), Zhao, S. M., Lin, S. T. and Wang, Y. H., (2003). “The
90-93. (in Chinese) structural integrity test of containment for Qinshan
Ye, Q. Q., (2012). “China’s nuclear power development nuclear power plant.” Industrial Construction, 33(9),
after Fukushima nuclear power plant accident.” 38-40. (in Chinese)
Journal of Advanced Concrete Technology Vol. 14, 261-270 May 2016 / Copyright © 2016 Japan Concrete Institute 261
Scientific paper
Abstract
A research program was carried out to investigate the effect of heating and cooling treatment upon hydraulic property and
macro-mechanical property of cement based material. Those treatments were performed on cylindrical mortar specimens
(water to cement ratio of 0.5), subjected to 5, 10, 15 heating-cooling cycle times and three levels of maximum heating
temperatures of 105°C, 145°C and 200°C. The hydraulic property tests were performed on a gas permeability and po-
rosity test system. The initial permeability and porosity value of the material were measuring before the heating and
cooling treatment under slightly confine (3MPa) and show the material has a good homogeneity. The damage of mortar
due to the effect of heating and cooling treatment were evaluated the variation of permeability, porosity and
macro-mechanical parameters by compare before and after heating and cooling treatment. Significant changes in the
macro-mechanical property and hydraulic property before and after heating and cooling treatment in mortar were ob-
served. The compressive strength and elastic modulus gradually decrease with the increase of maximum heating tem-
perature and the number of heating cycles, while the porosity and permeability increases gradually.
1. Introduction used to isolate the waste and fill the concrete containers
in the nuclear wastes repository. As the last protection
The safety of radioactive waste storage has become a barrier, its role is to isolate the waste from water or hy-
major concern in the development of nuclear energy drologic isolation (Plecas 2010), so it plays a significant
since it had been generated in the early 1950s. Mortar has role on the protection system (Plecas et al. 2015). But on
been widely used for radioactive waste storage because account of the porousness of mortar, its permeability and
of its low permeability and could be further used for the porosity should be minimized in the long-term operation
formation of disposal engineering barriers (McCulloch et of the nuclear waste storage. So it is important to predict
al. 1985; Wilding 1992; Glasser et al. 1994). It is usually the damage of mortar in the operation of long-term waste
storage. During its long service life, mortar damage is
developed both under the change of internal parameters
1 (e.g. the chemical decompositions as a result of hydration
Ph.D. candidate, Research Institute of Geotechnical
processes (Mainguy et al. 2000; Vydra et al. 2001), and
Engineering, Key Laboratory of Ministry of Education
also under the change of external conditions (e.g. tem-
for Geomechanics and Embankment Engineering, Hohai
perature (Lim and Mondal 2014; Vodák et al. 2004),
University, Nanjing, China.
2 moisture (Saemann et al. 2000), leading to the significant
Professor, Research Institute of Geotechnical
degradation of material properties. Those processes can
Engineering, Key Laboratory of Ministry of Education
seriously affect durability of the material (Chen et al.
for Geomechanics and Embankment Engineering, Hohai
2010) so as to reduce the safety of radioactive wastes
University, Nanjing, China.
3 storage.
Assistant Professor, Research Institute of Geotechnical
It is generally acknowledged that high temperature can
Engineering, Key Laboratory of Ministry of Education
increase mortar’s porosity due to the generation of mi-
for Geomechanics and Embankment Engineering, Hohai
cro-cracks in the structure (Chen et al. 2009; Gao et al.
University, Nanjing, China.
* 2002; Vydra et al. 2001). The effect of temperature on
Corresponding author, E-mail: wwang@hhu.edu.cn
4 strength properties of concrete has been extensively
Postdoc, Research Institute of Geotechnical
investigated (Bastami et al. 2011; Chan et al. 2000; Cül-
Engineering, Key Laboratory of Ministry of Education
fik and Özturan 2002; Ergün et al. 2013; Fall and Samb
for Geomechanics and Embankment Engineering, Hohai
2009; Gardner et al. 2005; Lin et al. 2015), and it was
University, Nanjing, China.
5 concluded that the reduction in strength of a concrete is
Professor, Key Laboratory of Coastal Disaster and
disastrous if heated up to 800°C (Chan et al. 2000). As
Defence, Ministry of Education, Hohai University,
reported in many publications(Caré 2008; Castellote et al.
Nanjing, China.
6 2004; Farage et al. 2003; Lim and Mondal 2014; Rostásy
Assistant Professor, Research Institute of Geotechnical
et al. 1980; Zhang et al. 2013), the change of mi-
Engineering, Key Laboratory of Ministry of Education
cro-structure and pore structures of mortar subjected to
for Geomechanics and Embankment Engineering, Hohai
high temperatures was studied, revealing that the port-
University, Nanjing, China.
Z. Lin, W. Xu, W. Wang, J. Zhang, H. Wang and R. Wang / Journal of Advanced Concrete Technology Vol. 14, 261-270, 2016 262
landite in mortar was decomposed abruptly and trans- Table 2 Mass of the mortar constituents.
formed into lime when it was heated up to 620°C before Constituent Mass (g)
cooling (Castellote et al. 2004). The transport properties Cement 720
of mortar subjected to temperature loading have been Standard sand 1220
investigated (Caré 2008; Chen et al. 2010; Chen et al. Water 310
2010; Pulkrabek and Ibele 1987; Vydra et al. 2001). The Bubble reducing agent 5
correlation between heating-induced crack porosity (and
crack aspect ratio) and permeability has been investi- clinker and gypsum, and 0-5% limestone. Its 3-day and
gated (Chen et al. 2010). The variation of porosity and 28-day compressive strengths are 31.6 MPa and 65.6
permeability for mortar under heat cycling were studied MPa, respectively. Its 3-day and 28-day tensile strengths
(Chen et al. 2010; Chen et al. 2009). Fire tests were are 6.4 MPa and 9.8 MPa, respectively. The chemical
carried out to study the mechanical behavior of rein- compositions and physical properties are presented in
forced concrete frame at elevated temperatures (Raouf- Table 1.
fard and Nishiyama 2015). Damage evaluation in mortar In order to simulate the cement-based material cov-
with history of high temperature was investigated by ering the nuclear waste, the material whose compositions
Henry (Henry et al. 2016). Furthermore, the durability of are similar to those existing in real nuclear waste re-
an Ultra High Performance Fiber Reinforced Concrete pository was adopted. According to Ilija Plecas (Plecas et
after accelerated aging was studied through gas perme- al. 2015), mortar samples were prepared with the 0.5
ability test (Wang et al. 2014c). W/C ratio and the 1/1.694 cement-to-sand ratio, which
Recently, many researches focused on the strength could minimize the permeability and leaching rate and
properties, the micro-structure and the permeability of maximize the mechanical strength. The standard sand
mortar subjected to high temperature effects and heat (mud mass content<0.2%, mass content of SiO2 is
cycling (Bastami et al. 2011; Cülfik and Özturan 2002), 96.0%) was used with 0-1 mm in diameter. The mass of
to a certain extent, they can provide useful information the various constituents are presented in Table 2.
related to the macro-mechanical property and hydraulic Firstly, the mixing procedure consisted in the mixing
property of mortar subjected to heating treatment. sand and cement during 30 s. Then water and bubble
However, few of them studied the relationship between reducing agent were added and the mixing process was
the variation of macro-mechanical, permeability and maintained for further 3 minutes. Subsequently, the
porosity of mortar subjected to heat cycling. Besides, mixtures were put into a cylindrical organic glass mold
those studies often aimed at a few heating-cooling cycles (50 mm high, 50 mm in diameter) until no air bubbles
or only once, and they did not study the influence of were produced. After about 3 h, they were sealed into a
multiple heating-cooling cycles on the mortar during the plastic bag with some water at the bottom. The sealed
long-term operation of engineering. In this work, a spe- mold remained in the laboratory at the temperature of 24
cial mortar was selected as the experiment object, dif- ± 1°C. After 24-h curing, the mortars were demount and
ferent multiple heating-cooling cycles were adopted in stored in a freshwater tank with water temperature at 20±
order to simulate the real heating treatment during the 5°C. After 28 days curing, the saturated mortars were
long-term operation of engineering. And during the taken out from the water tank.
heating process, different maximum temperatures were
selected in order to simulate the different heating treat- 2.2 Heating and cooling cycle procedure
ments. During the tests, the variations of permeability, The experiment was designed to simulate the in-situ
porosity and macro-mechanical of mortar were measured. conditions of the long-term storage of the radioactive
The purpose of the test was to study the damage of wastes, therefore multiple heating-cooling cycles and
mortar subjected to different heating treatments and different maximum temperatures during the heating
heating-cooling cycles. process could be used. Based on the methods by Lion et
al. (2005), the reference temperature was set at 60ºC,
2. Testing equipment and methods since it could cause evaporation of free porous water
without any notable effect on the CSH gel. In this study,
2.1 Specimen preparation and curing each reaction cycle included a heating and cooling
The experiments were undertaken on a normalized process. The heating was performed between 105°C and
mortar with a water/cement (w/c) ratio of 0.5. The ce- 200°C. During this process, the specimens were divided
ment was made of China commercial PⅡ52.5 Portland into three batches for different tests, which were heated
cement. According to the Chinese National Standard up to 105, 145 and 200°C, respectively. The initial tem-
(GB 175-2007), this cement comprised more than 95% perature was 24ºC (room temperature) and the incre-
Z. Lin, W. Xu, W. Wang, J. Zhang, H. Wang and R. Wang / Journal of Advanced Concrete Technology Vol. 14, 261-270, 2016 263
mental rate was set at 5ºC/minute by a high temperature Science and Technology. The diagram of the test system
oven. When the peak temperature was attained, samples is shown in Fig. 1 (Wang et al. 2014a). The confining
were maintained in the oven for more 12 h until constant pressure Pc is generated and controlled by an oil pump
weights were accessed (Lion et al. 2005). During the and its maximum capacity is up to 60MPa. During the
cooling process, samples were transferred into a fresh- gas permeability test, the interstitial pressure is con-
water tank with a temperature of 20± 5°C for 3h. In order trolled and monitored by two manometers, which lo-
to simulated the real heating treatment during the cated at the ends of both side of samples. In this test, the
long-term operation of engineering, the multiple heat- argon with 99.9% purity was used as permeable media
ing-cooling cycles was adopted. The specimens were instead of water. The test system can be also used to
processed with 5, 10 and 15 times of heating and cooling perform the porosity test through adding a pipe and a
cycles, respectively. The experiment design of heating precision pressure gauge with the pressure range of 0~
treatment for each batch of specimens are presented in 2 MPa and the resolution of 0.0001 MPa (Wang et al.
Table 3. 2014a).
The porosity φ of a porous medium describing the
2.3 Testing equipment of macro-mechanical fraction of void space in the material, it is defined by the
parameters following formula:
The macro-mechanical parameters tests were performed
on the automatic triaxial servo-controlled instrument. φ = ⎛⎜ Vn V ⎞⎟ ×100% (1)
The lateral pressure is controlled by a high-precision ⎝ ⎠
pump and automatically regulated by the computer. The
where Vn is the effective volume of the pores (m3), and
measurement system consists of a pair of LVDTs (linear
V is the total volume of sample (m3).
variable displacement transducer) and a chain transducer.
In this study, the porosity of mortar is measured
Axial displacement is measured by the LVDTs installed
through the gas expansion method. The procedures is
inside the triaxial chamber. During the experiments, the
described as follows. A sample with a known bulk vol-
deformation data was recorded by the corresponding
ume is enclosed in a container of known volume, as
sensors in an acquisition system.
shown in Fig. 1. It is connected to another container with
a known volume which is evacuated (i.e., near vacuum
2.4 Measurements of permeability and porosity
pressure). When a valve connecting the two containers is
The permeability and porosity subjected to the heat-
opened, gas passes from the first container to the second
ing-cooling cycles were evaluated by using a gas per-
until a uniform pressure distribution is attained. By using
meability and Porosity Test System which was jointly
the ideal gas law, the volume of the pores can be calcu-
developed by our laboratory and Lille University of
lated,
Z. Lin, W. Xu, W. Wang, J. Zhang, H. Wang and R. Wang / Journal of Advanced Concrete Technology Vol. 14, 261-270, 2016 264
Fig.1 Schematic diagram of test system for permeability and porosity measurements.
a b
Fig. 2 Variation of porosity of sample under different heating and cooling cycle curing conditions (confining pressure=3
MPa).
heating) as an example, it shows that with the increase in ferent heating-cooling cycles treatment (T=105ºC). It
gas pressure, the permeability of mortar decreases. This shows that the permeability gradually increases with the
may be explained by the Klinkenberg effect which de- number of temperature heating-cooling cycle increased.
scribes the collision phenomenon between gas molecules For example, under the same fluid pressure of 0.5MPa,
and pore walls. The increasing pressure leads to smaller for the 5, 10, 15 times of heating-cooling cycle treatment,
effective permeability, showing the good mechanical the permeability increases by 61.40%, 258.20% and
state of the samples: damaged specimens tend to give 391.34%, respectively.
increasing pressure (Farage et al. 2003). Figure 5 shows the variation of permeability during
It also shows that the permeability of sample is higher the loading and unloading process for the samples q-2
after the heating-cooling cycle treatment than before. and q-3 under two different fluid pressures (0.5MPa and
The permeability increase of mortar indicates a modifi- 1MPa). Regardless of different fluid pressures, during
cation of the connected porous network happened. Under the loading and unloading process of confining pressure,
the same fluid pressure of 0.5MPa, the permeability the permeability was bigger in the loading stage than that
increases by 61.4%, 136.03% and 149.80% for the sam- in the unloading stage for the same confining pressure.
ples q-1, q-2, and q-3, respectively. It means that with the Take the sample of q-2 for example, under the confining
maximum temperature of heating process (N=5) in- pressure of 15MPa, the permeability was 5.2·10-17 in the
creased, the permeability of mortar increases continu- loading stage while it was 4.35·10-17 in the unloading
ously. As can be seen from Figs. 2 (a) and 3, the poros- stage. It means that when the confining pressure is re-
ity and permeability of mortar consistently increased. leased, the deformation of its pores cannot be completely
Compare with Fig. 2, the permeability change is much restored, and the mortar sample underwent the plastic
larger than the porosity change in the same samples. In deformation during the loading and unloading process.
means that permeability is more sensitive to mi- This phenomenon was owing to the stress equilibrium
cro-cracks aperture and connection than porosity. state of mortar granule structure was destructed when the
Figure 4 shows the variation of porosity under dif- confining pressure increase after the heating and cooling
cycles curing, then the granule structure of mortar hap-
8
q-1(before heating cycle) q-1(after heating cycle)
pen to irreversible damage.
q-8(before heating cycle) q-8(after heating cycle) It is shown from those figures that the permeability of
7 q-9(before heating cycle) q-9(after heating cycle)
mortar decreases sharply at the beginning of the loading
6 stage, and reduces slowly with the increase of pressure
Permeability(1E-17m2)
7 7
Permeability(1E-17m2)
5
lo
ad
ngi
lo
pr
ad
5
o
4
in
ce
g
ss
pr
oc
es
s
3
4
unloading process 2
unloading process
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Confining pressure(MPa) Confining pressure(MPa)
a b
Fig. 5 Variation of permeability during loading and unloading under different fluid pressures.
Z. Lin, W. Xu, W. Wang, J. Zhang, H. Wang and R. Wang / Journal of Advanced Concrete Technology Vol. 14, 261-270, 2016 267
Table 4 Compressive strengths of mortar samples under different heating and cooling cycle times.
Number of heating cycles Compressive strength (MPa) Mean compressive strength (MPa)
50.25
Initial (60°C) 55.49 50.37
45.38
43.89
0 times 40.5 41.80
41.02
36.56
10 times 34.51 34.94
33.75
33.19
15 times 33.94 33.5
33.38
Table 5 Compressive strengths of mortar samples under 105°C, 145°C and 200°C maximum heating temperatures.
Maximum heating temperature Compressive strength (MPa) Mean compressive strength (MPa)
50.25
Initial (60°C) 55.49 50.37
45.38
43.89
105°C 40.5 41.80
41.02
29.25
145°C 31.13 29.81
29.06
25.88
200°C 20.25 24.42
27.13
19
divided into five stages: compaction stage, elastic stage,
elastic-plasticity stage and failure stage. However, it can
un
lo
18
ad
17
macro-mechanical properties of typical sample under
different heating-cooling cycle treatments (T=105°C). It 0 105 15 20 25 30
is shown that with the increasing number of heat- Confining pressure (MPa)
ing-cooling cycles (T=105°C), more severe deterioration Fig. 6 Variation of porosity during loading and unloading
in macro-mechanical parameters (peak strength and process under fluid pressure of 0.5MPa.
Z. Lin, W. Xu, W. Wang, J. Zhang, H. Wang and R. Wang / Journal of Advanced Concrete Technology Vol. 14, 261-270, 2016 268
60 60
S-1 (initial) S-12(105°, 5times) S-1 (initial) S-12(105°, 5times)
S-22(105°, 10times) S-32 (105°, 15times) S-22(105°, 10times) S-32 (105°, 15times)
50 50
40 40
Stress (MPa)
Stress (MPa)
30 30
20
20
10
10
0
0 0 3 6 9 12 15 18
0 3 6 9 12 15 18
Strain(‰)
Strain(‰)
Fig. 9 Typical stress-axial strain curves of mortar sam-
Fig.7 Typical stress-axial strain curves of mortar under
ples under 105°C, 145°C and 200°C maximum heating
different heating-cooling cycle treatments.
temperatures.
gradually decreases. The average compressive strength mum temperature of heating process.
respectively decreases by 17.01%, 40.81%, and 51.61% Figure 10 shows the comparisons of compressive
for 105°C, 145°C and 200°C in comparison with the strength and elastic modulus for four different maximum
referenced sample strength. It is noted that under dif- heating temperatures (60°C, 105°C, 145°C and 200°C).
ferent maximum heating temperatures, different physical With the increase of maximum heating temperature, the
reactions occurred, resulting in different damage degrees, compressive strength and the elastic modulus gradually
so that the macro-mechanical properties showed differ- decrease. In contrast to the referenced value, the elastic
ent degrees of deterioration. Compared with Fig. 7, Fig. modulus and compressive strength for 200°C decrease
9 shows that the mortar is more ductile with the maxi- by 66.45% and 50.30%, respectively.
Z. Lin, W. Xu, W. Wang, J. Zhang, H. Wang and R. Wang / Journal of Advanced Concrete Technology Vol. 14, 261-270, 2016 269
Scientific paper
Abstract
The ferro-siliceous sacrificial concrete is an important protection material to Nuclear power plant, high density aggregate
and high fluidity can lead to occurrence of shrinkage crack. In this paper, the early-age crack and restrained shrinkage
crack of sixteen types of sacrificial concrete containing different content of cementitious materials and two kinds of fibers
were researched. Results showed that the workability of sacrificial concrete decreased with increasing addition of fiber,
and the slump loss resulted from polypropylene (PP) fiber was much higher than that of basalt fiber. The fibers had less
influence on the compressive strength, while had greater impact on the flexural strength, which was improved by 10-20%.
The reduction of total cementitious materials and replacement ratio increment of mineral admixtures including fly ash and
Ground Granulated Blastfurnace Slag (GGBS) in sacrificial concrete could decrease its early-age crack obviously. The
utilization of 1.0-1.5kg/m3 PP fiber could extend the cracking time and reduce the crack width of sacrificial concrete and
total cracking area. The improving crack resistance capacity of sacrificial concrete resulted by basalt fiber was weaker
than PP fiber. The optimized mix proportion of sacrificial concrete mixture was 450kg/m3 total cementitious materials,
50% replacement ration of mineral admixtures and 1.0kg/m3 PP fiber.
When molten core penetrates the reactor pressure crete through shrinkage crack and induce the corrosion of
vessel and reacts with sacrificial concrete, the concrete in steel bar and result in seriously durability problem. Ad-
the floor and sidewalls of the cavity starts to melt. Nor- ditionally, these shrinkage cracks in sacrificial concrete
mally the erosion rate of the concrete is limited by the will also promote the thermal conduction and the leakage
heat transfer rate from the melt to the concrete. During of radioactive during the melting of the core. Therefore,
the first pouring of the melt, spalling of concrete might crack control is very important for durability and safety
take place (Kalifa et al. 2000), and it can cause faster of sacrificial concrete.
erosion of the sacrificial concrete than expected and Mineral admixtures including fly ash and GGBS are
damage the containment vessel concrete. Therefore, commonly used to replace cement in high performance
sacrificial concrete possesses high temperature stability concrete to improve the durability of concrete under
and excellent melting properties during a nuclear severe environments (Ouda 2015; Seleem et al. 2010).
incident. Molten organic fibers can generate release Additionally, low pozzolanic reactivity of fly ash and
passage for water vapor in concrete, and thus reduce GGBS decreases the autogenous shrinkage at early age,
burst crack of concrete (Jin et al. 2010). The basalt fiber and the drying shrinkage of self-compacting concretes
can keep about 90% of the normal temperature strength incorporating fly ash, GGBS and metakaolin is lessened
after exposure at 600 °C for 2 h(Sim et al. 2005; Fiore et (Mermerdas and Arbili 2015; Guneyisi et al. 2010). The
al. 2015). The steel fiber can improve the strength and use of fibers is effective in resisting shear forces in con-
crack-resistance capacity, but decrease the workability of crete structures, and the addition of fibers into conven-
concrete (Pikus 2016). Therefore, the PP and basalt fiber tional concrete increases the ductility and energy dissi-
were often used to improve the high temperature resis- pation of wall concrete structures subjected to strong
tance capacity of sacrificial concrete. ground motions (Choun and Park 2015). And fibers
But as time goes on, the protective properties of these increase the resistance to crack formation and propaga-
sacrificial concrete will degrade due to environmental tion, therefore reduce crack width and length (Afrough-
loads including thermal effect, irradiation, neutralization sabet, and Ozbakkaloglu 2015; Saeid et al. 2012). There
and penetration of salinity (Kitsutaka and Tsukagoshi is very scarce experimental data about the shrinkage
2014). Safety guide for the monitoring of nuclear power crack behavior of this special concrete type. This is one
stations and ageing management for nuclear power plants of the reasons for optimizing crack resistance capacity of
has been published by IAEA and French Electricity sacrificial concrete by use of mineral admixtures and
Board (Stephan and Salin 2012). As mentioned, Fe2O3 different type of fibers.
and SiO2 in sacrificial concrete will react with corium The main objective of the current research is to inves-
active reducers and reduce the explosion risk. Therefore, tigate the suitability of some concrete components for
the sacrificial concrete in the EPR reactor is a special producing EPR sacrificial concrete using mineral ad-
type “FeSi” concrete, which includes both high density mixture and fibers that could enhances its shrinkage
aggregates, hematite (Fe2O3) and siliceous (SiO2). And crack resistance capacity.
the density of sacrificial concrete used in EPR is around
2600 kg/m3 (Sevón et al. 2010). High density aggregate 2. Experiment and materials
prone to sinking in fresh mortar and induced dry
shrinkage in the surface layer of sacrificial concrete. 2.1 Materials and specimen preparation
Additionally, the size of sacrificial concrete core catcher P.I.52.5 Portland cement in accordance with Chinese
zone is φ6m×0.45m, the workability of fresh concrete standard GB175-2007, with a compressive strength of
will be great in order to cast in small space (Chu et al. 59.8 MPa at an age of 28 days, was used in this study.
2016). Consequently, early-age autogenous shrinkage Class I fly ash (as per Chinese standard GB1596-2005)
cracking and dry shrinkage crack will easily generate in and S95 GGBS (Chinese standard GB/T18046-2008)
EPR sacrificial concrete due to segregation and high were employed to replace Portland cement. The chemical
content of mixing water. When excessive shrinkage composition of cement, fly ash, and GGBS is shown in
cracks occur in sacrificial concrete, its durability will be Table 1. The specific gravity of cement, Fly ash and
threatened seriously. From an aging management per- GGBS is 3.13, 2.30 and 2.20, respectively.
spective, the presence of concrete cracks is an impor- Considering the various phenomena that occur during
tance threat because they provide possible avenues of core melt localization, it is important to minimize the
access for corrosion ions (Naus et al. 1999; Dai et al. highly exothermic corium oxidation by water/air and to
2010; Jin et al. 2012). In containment building, CO2, limit additional hydrogen production to ensure hydrogen
water steam and O2 will penetrate into sacrificial con-
Z. Jin, X. Zhao, T. Zhao and Y. Cao / Journal of Advanced Concrete Technology Vol. 14, 502-510, 2016 504
1.Lateral side panel; 2. Vertical side panel; 3.Bolt; 4. Reinforcing rib; 5. Inducing groove; 6. Bottom plate
Fig. 2 Diagram of test equipment for early-age crack.
Z. Jin, X. Zhao, T. Zhao and Y. Cao / Journal of Advanced Concrete Technology Vol. 14, 502-510, 2016 506
the concrete matrix, in fact as the fiber content increased, given in Fig. 4. The compressive strength of sacrificial
the slump slightly decreased. Due to the density of ba- concrete with total cementitious materials content of
salt fiber and PP fiber are 2650kg/m3 and 910kg/m3, 500kg/m3 is about 50MPa, and 37-42MPa for sacrificial
respectively, the specific surface area of basalt fiber is concrete with 400kg/m3cementitious materials. In gen-
smaller than that of PP fiber, therefore, concrete slump eral, the influence of fibers on compressive strength of
loss resulted from PP fiber is much greater than that of sacrificial concrete can be ignored. The flexural strength
basalt fiber. Test results showed that, the application of of sacrificial concrete was tested and shown in Fig. 5.
fibers decreases the unit weight of all tested concrete The flexural strength values of fiber reinforced sacrifi-
slightly. And the density of sacrificial concrete is about cial concrete increased by 10-20%, comparing to the
2600-2700kg/m3. respective plain concrete specimens. Additionally, the
increase of the concrete flexural strength by basalt fiber
3.2 Compressive strength and flexural strength is stronger than 1.5kg/m3 PP fiber. And the flexural
The part results of the compressive strength tests are strength of sacrificial concrete containing 1kg/m3 PP
60 50
(a)
3d 28d (b)
3d 28d
Compressive strength (MPa)
50
Compressive strength (MPa)
40
40
30
30
20
20
10 10
0 0
MF0 MF1 MPF1 U0 UF0 UPF0
3 3
Fig. 4 Compressive strength of fiber reinforced sacrificial concrete (a) 500kg/m cementitious materials, (b) 400kg /m
cementitious materials.
Z. Jin, X. Zhao, T. Zhao and Y. Cao / Journal of Advanced Concrete Technology Vol. 14, 502-510, 2016 507
fiber is higher than that containing 1.5kg/m3 PP fiber, from 0.62mm to 0.49mm and the total cracking area on
because increased content of fiber would result in de- unit area also decreased by 50%. For the sacrificial con-
creased dispersion and more defects in concrete. crete with 400kg/m3 cementitious materials, the re-
placement ratio of mineral admixture was increased from
3.3 Early-age crack of sacrificial concrete 20% (U0) to 35% (T3), and the total cracking area on unit
The crack number, length and width of each crack were area was also decreased by about 6.7%. Therefore, with
observed and recorded after concrete were cast and cur- the increment of the replacement ratio of mineral ad-
ing for 24h. And then the maximum crack length and mixture in cementitious materials, when the hydration
width, the mean crack area, the crack number and the rate of fly ash and GGBS was slow and hydration heat
total crack area on unit area were calculated and sum- was still low, the cracking risk of the sacrificial concrete
marized in Table 6. was reduced accordingly.
The cracking degree of the sacrificial concrete in- For sacrificial concrete with 500 kg/m3cementitious
creased with the amount of cementitious material as it is materials, 6kg/m3basalt fiber decreased the maximum
400~500kg/m3. When the amount of cementitious mate- crack width by 0.33mm, reduced the total cracking area
rials is 500kg/m3, the maximum crack width of the sac- on unit area by 24.2%. 1.5kg/m3 PP fiber reduced the
rificial concrete is 1.6 times as much as that of the maximum crack width and the total cracking area on unit
400kg/m3, and the total cracking area on unit area in- area by 56% and 45%, respectively. However, because
creased by over 20% than that of 450 and 400kg/m3. This the amount of cementitious materials is too high, even if
is due to the increase amount of cementitious materials, the basalt fiber and PP fiber were mixed in sacrificial
the hydration heat would aggravate, and then, the risk of concrete, the cracking risk of the sacrificial concrete was
cracking improved. In addition, comparing Y0 with N1, much higher than that of the others.
when the replacement rate of mineral admixture in ce- For the sacrificial concrete with 450kg/m3 cementi-
mentitious materials increased from 20% to 50%, the tious materials, the cracking risk decreased with incre-
maximum crack width of sacrificial concrete decreased ment content of PP fiber. When the content of PP fiber
reached 1.5kg/m3, early cracking would not occur. For
5 the basalt fiber, the crack resistance ability for the sacri-
ficial concrete is not significantly increased with its fu-
4 ture increasing fiber addition. As the basalt fiber content
Flexural strength (MPa)
0
MF0 MF1 MPF1 Y0 Y4 Y5 U0 UF0 UPF0
3.4 Restrained shrinkage of sacrificial concrete
Fig. 5 Flexural strength of fiber reinforced sacrificial con- The cracking behavior of PP fiber reinforced sacrificial
crete. concrete in restrained shrinkage conditions was tested
Z. Jin, X. Zhao, T. Zhao and Y. Cao / Journal of Advanced Concrete Technology Vol. 14, 502-510, 2016 508
1.0 60
3
0kg/m PP fiber (a)
0.9 (b)
3
1.0kg/m PP fiber 40
0.8 3 3
Maxmium crack width (mm)
3
0.6 1.5kg/m PP fiber
0
0.5
0.4 -20
0.3
-40
0.2
0.1 -60
10 12 14 16 18 20 22 24 26 28 30 32 0 4 8 12 16 20 24 28 32
Time (days) Time (days)
Fig. 6 Cracking behavior of sacrificial concrete in restrained shrinkage conditions, (a) crack width, (b) strain in inner steel
ring.
Z. Jin, X. Zhao, T. Zhao and Y. Cao / Journal of Advanced Concrete Technology Vol. 14, 502-510, 2016 509
120
with molten core for1 hour.
100
4. PP fiber with the content range from 1.0 kg/m3 to
80
1.5 kg/m3 can reduce the early-age crack of sacrificial
60
concrete, and the compressive strain in the steel ring of
40
the concrete with PP fiber was reduced by 50%, the total
20
crack area decreased by 37-48% comparing to the plain
0
0 1 2 3 4 5 6 7 sacrificial concrete.
Time (h ) 5. The slump loss of fresh sacrificial concrete caused
by basalt fiber is less than that of PP fiber, 3-6 kg/m3
Fig. 8 Influence of crack on molten depth of sacrificial
basalt fiber decreased the cracking area on unit area of
concrete.
sacrificial concrete by 30-50%, but the improving crack
resistance capacity of basalt fiber was weaker than that of
1 hour ahead of that of the non-crack sacrificial concrete
PP fiber.
layer. Therefore, the crack control of sacrificial concrete
is very important for the design of sacrificial concrete
Acknowledgments
layer in core spreading compartment and core catcher
This work is part of a series project financially supported
zone.
by the Chinese National Natural Science Foundation
(NSF) Grant No. 51378269 and No.51420105015, and
4 Conclusions the Chinese National 973 project Grant No.
2015CB655100. The authors gratefully appreciate the
The workability, density, compressive strength, flexural financial support provided by the NSF and other Foun-
strength, early-age crack and restrained shrinkage crack dations.
behavior of sacrificial concretes with total cementitious
materials of 400-500kg/m3, 20-50% replacement ration References
of mineral admixtures and two types of fiber were in- Afroughsabet, V. and Ozbakkaloglu, T., (2015). “Me-
vestigated. Based on the findings of this study, the fol- chanical and durability properties of high-strength
lowing conclusions were drawn: concrete containing steel and polypropylene fibers.”
1. The influence of basalt fiber and PP fiber on the Construction and Building Materials, 94, 73-82.
density and compressive strength of sacrificial concrete Choun, Y. S. and Park, J. H., (2015). “Evaluation of
could be ignored. The slump of fresh concrete decreased seismic shear capacity of perstressed concrete con-
with the volume fraction of fibers, and the flexural tainment vessels with fiber reinforcement.” Nuclear
strength of fiber reinforced sacrificial concrete was in- Engineering Technology, 47, 756-765.
creased by 10-20% comparing to plain sacrificial con- Chu, H. Y., Jiang. J. Y., Sun, W. and Zhang, M. Z.,
crete. (2016). “Thermal behavior of siliceous and
2. The compressive strength of sacrificial concrete ferro-siliceous sacrificial concrete subjected to
increased with total cementitious materials, while the elevated temperatures.” Materials and Design, 95,
early-age crack risk obviously increased with total ce- 470-480.
mentitious materials content. Replacement of cement by Dai, J. G., Akira, Y., Wittmann, F. H., Yokota, H. and
high volume of fly ash and GGBS caused a significantly Zhang, P., (2010). “Water repellent surface impregna-
reduction of early-age shrinkage crack of sacrificial tion for extension of service life of reinforced con-
concrete.
Z. Jin, X. Zhao, T. Zhao and Y. Cao / Journal of Advanced Concrete Technology Vol. 14, 502-510, 2016 510
crete structures in marine environments: The role of Mermerdas, K. and Arbili, M. M., (2015). “Explicit
cracks. “ Cement and Concrete Composites, 32, formulation of drying and autogenous shrinkage of
101-109. concretes with binary and ternary blends of silica
Fiore, V., Scalici, T., Di Bella, G. and Valenza, A., fume and fly ash.” Construction and Building
(2015). “A review on basalt fibre and its composites.” Materials , 94, 371-379.
Composites Part B, 74, 74-94. Naus, D. J., Oland, C. B. and Ellingwood, B. R., (1999).
Fischer, M., (2004). “The severe accident mitigation “Summary and conclusions of a program addressing
concept and the design measures for core melt aging of nuclear power plant concrete structures.”
retention of the European Pressurized Reactor Nuclear Engineering and Design, 194, 73-96.
(EPR).” Nuclear Engineering and Design, 230(1-3), Ouda, A. S., (2015). “Development of high-perform-
169-180. ance heavy density concrete using different aggre-
Guneyisi, E., Mehmet, G. and Ozbay, E., (2010). gates for gamma-ray shielding.” Progress in Nuclear
“Strength and drying shrinkage properties of Energy, 79, 48-55.
self-compacting concretes incorporating multi-system Pikus, G. A., (2016). “Steel fiber concrete mixture
blended mineral admixtures.” Construction and workability.” Procedia Engineering, 150, 2119 -
Building Materials, 24, 1878-1887. 2123.
Jin, Z. Q., Sun, W., Hou, B. R. and Jiang, J. Y., (2010). Stephan, P. and Salin, J., (2012). “Ageing management
“Deformation and microstructure evolution of of concrete structure: Assessment of EDF
concrete subjected to elevated temperature.” Journal methodology in comparison with SHM and AIEA
of Southeast University (Natural Science Edition) guides.” Construction and Building Materials, 37,
40(3), 619-623. 924-933.
Jin, Z. Q., Zhao, T. J., Chen, H. S. and Zhuang, Q. C., Seleem, H., Rashad, A. M. and El-Sabbagh, B. A.,
(2012). “Chloride corrosion of splitting cracked (2010). “Durability and strength evaluation of
concrete under marine environment.” Journal of high-performance concrete in marine structures.”
Central South University (Science and Technology), Construction and Building Materials, 24, 878-884.
43, 2821-2826. Saeid, K., Hazizan, M. A., Morteza, J. and Jalal, R.,
Kalifa, P., Menneteau, F.-D. and Quenard, D., (2000). (2012). “The effects of polypropylene fibers on the
“Spalling and pore pressure in HPC at high properties of reinforced concrete structures.” Const-
temperatures.” Cement and Concrete Research, 30, ruction and Building Materials, 27(1), 73-77.
1915-1927. Sevón, T., Kinnunen, T., Virta, J., Holmström, S.,
Komlev, A. A., Almjashev, V. I. and Bechta, S. V., Kekki, T. and Lindholm. I., (2010). “HECLA experi-
(2015). “New sacrificial material for ex-vessel core ments on interaction between metallic melt and
catcher.” Journal of Nuclear Materials, 467, 778-784. hematite-containing concrete.” Nuclear Engineering
Kitsutaka, Y. and Tsukagoshi, M., (2014). “Method on and Design, 240, 3586-3593.
the aging evaluation in nuclear power plant concrete Sim, J., Park, C. and Moon, D. Y. (2005). “Characteris-
structures.” Nuclear Engineering and Design, 269, tics of basalt fiber as a strengthening material for con-
286-290. crete structures.” Composites: Part B, 36, 504-512.