Nuclear Engineering and Design 342 (2019) 20–28

Contents lists available at ScienceDirect

Nuclear Engineering and Design

journal homepage: www.elsevier.com/locate/nucengdes

Influence of flow velocity and temperature on flow accelerated corrosion T

rate at an elbow pipe
⁎
Y. Utanohara , M. Murase
Institute of Nuclear Safety System, Inc., 64 Sata, Mihama-Cho, Mikata-Gun, Fukui 919-1205, Japan

A R T I C LE I N FO A B S T R A C T

Keywords: Flow accelerated corrosion (FAC) is one of the important issues that must be considered for aging fossil and
Flow accelerated corrosion nuclear power plants. To understand the effect of thermal flow field on FAC, FAC rates at an elbow pipe were
Elbow pipe measured under different flow velocity and temperature conditions. The elbow test section was made of stainless
Wall thinning rate steel (diameter D = 49.5 mm) and FAC rates were measured using corrosion sensors made of carbon steel. The
Electric resistance method
dissolved oxygen concentration was kept under 0.1 μg/kg, and pH was nearly neutral (about 7.0) at room
temperature. The mean cross-sectional velocity was changed from 0.39 to 5.74 m/s (Reynolds number, about
1.0e6). When the water temperature was about 150 °C, the FAC rate was smaller at the intrados of the elbow pipe
than at other circumferential locations. This tendency continued downstream. The FAC rates at the elbow pipe
were larger than those upstream and downstream from the elbow pipe and the FAC rates downstream from the
elbow pipe decreased along the flow direction. FAC rates increased as flow velocity increased and their re-
lationship was not linear. The ratios of the maximum FAC rate at the elbow to the FAC rate in the upstream
straight pipe ranged from about 1.7 to 2.9. When temperature decreased to 100 °C, FAC rate at the intrados
became the largest of the other circumferential sensors. When temperature decreased further to 50 °C, FAC rate
also decreased, but the value was not negligible. The influence of flow velocity was negligibly small at 50 °C and
remarkable at 100 °C and 150 °C. The combination effect of flow velocity and temperature was different from
place to place and particularly strong at the elbow.

1. Introduction
Flow accelerated corrosion (FAC) is an important issue for aging
management of fossil and nuclear power plants. FAC occurs at the pipe
geometry where flow is strongly disturbed such as downstream from an
orifice and at an elbow. FAC causes thinning of the pipe wall which
occasionally leads to a pipe rupture accident. Thinning of pipe wall
thickness had been managed based on non-destructive inspection of
pipe wall thickness and evaluation of remaining lifetime following each
electric utility’s own guideline. The Japanese Society of Mechanical
Engineers published its guideline on pipe wall thinning management in
2005 (The Japan Society of Mechanical Engineers, 2006). Since then,
the management method for pipe wall thinning has been unified under
this guideline in Japan.
For prediction of FAC rates, it is important to understand factors
influencing FAC. Generally, influencing factors are the water tem-
perature, pH, dissolved oxygen concentration, material composition
and fluid dynamics factor (Dooley and Chexal, 2000). FAC studies from
the viewpoint of the fluid dynamics factor (which the authors focus on)
have mainly been evaluated by using the mean cross-sectional velocity
(Bignold et al., 1981; Heitmann and Schub, 1983), but recently the
relationship between FAC and a local flow field has been reported
(Yoneda et al., 2009; El-Gammal et al., 2010; Pietralik and Schefski,
2011; Fujisawa et al., 2012; Tsuneyoshi et al., 2015). The essential fluid
dynamics factor influencing FAC is recognized to be the mass transfer
coefficient near a pipe wall, and studies have been conducted based on
this concept. To validate prediction methods for FAC rates, FAC data
are needed. There are many power plant data but uncertainties for
operation conditions and water chemistries are generally large.
Therefore, the authors have measured FAC rates under well-controlled
conditions in a test loop for the region downstream from an orifice
(Utanohara et al., 2012, 2013; Kamahori et al., 2013) and elbow
(Utanohara et al., 2016). The goal of these studies has been to develop a
numerical simulation method by which effects of the local flow field on
FAC can be evaluated.
In this paper, the authors focused on FAC rates at an elbow pipe,
because there are many elbows in a power plant but FAC rates at an
elbow pipe have not been directly measured in experiments, although

⁎
Corresponding author.
E-mail addresses: utanohara@inss.co.jp (Y. Utanohara), murase@inss.co.jp (M. Murase).

https://doi.org/10.1016/j.nucengdes.2018.11.022
Received 31 July 2018; Received in revised form 7 November 2018; Accepted 20 November 2018
Available online 03 December 2018
0029-5493/ © 2018 Elsevier B.V. All rights reserved.
Y. Utanohara, M. Murase
the distribution of the mass transfer coefficient was measured experi-
mentally (El-Gammal et al., 2010) and wall thinning data in a power
plant were examined (Pietralik and Schefski, 2011). Previously
(Utanohara et al., 2016), though the authors had measured FAC rates at
an elbow pipe with the diameter of 49.5 mm, the experiment was
conducted under a constant condition, flow velocity was 5.07 m/s and
fluid temperature was about 150 °C. The authors have also conducted
measurement of mass transfer coefficient and numerical simulation of
the elbow flow field (Tsuneyoshi et al., 2017). Through the comparison
between FAC rate and mass transfer coefficient of the same elbow
geometry, the authors have tried to clarify their quantitative correlation
between them. Hence, to understand the influence of thermal flow field
on FAC rate, more detailed experiment is necessary with flow velocity
and temperature as a parameter. Particularly, it is well known that FAC
usually occurs between 100 °C and 280 °C and FAC rate becomes
maximum near 130–150 °C (Dooley and Chexal, 2000; Bignold and
Garbett, 1982). However, Yoneda et al. (Yoneda et al., 2016) confirmed
that FAC rate was considerably large at the relatively low temperature
of 50 °C. Hence, in this study, temperature was decreased to 50 °C and
FAC rate was measured. Here, FAC rate of an elbow flow was measured
under different velocity and temperature conditions and the findings
are reported.
2. Experimental method
2.1. Test loop
The test loop used in the FAC rate measurement experiment is
shown in Fig. 1. The experimental method was almost the same as in
previous studies (Utanohara et al., 2012, 2013, 2016; Kamahori et al.,
2013). The recirculation pump was equipped downstream from the test
section to avoid pump-induced disturbance effects on the section. Dis-
solved oxygen concentration of the feed water was decreased using the
degasifier before heating. Dissolved oxygen concentration was normally
below 0.1 μg/kg. The pH was not controlled and was nearly neutral at
room temperature with no additives. Iron concentration was kept low
using the demineralizer. The loop temperature was automatically con-
trolled by the heater power and cooling unit equipped in a small branch
pipe downstream from the test section. Under the steady-state opera-
tion, cooling was needed to keep the water temperature constant due to
Joule heating of the recirculation pump. To avoid cavitation in the test
loop, water pressure was also controlled to 1.5 MPa by an automatic
pressure control system using nitrogen gas. The pressure in the hot
Fig. 1. Test loop for measurements of FAC rate.
21
Nuclear Engineering and Design 342 (2019) 20–28
Table 1
Specifications of measurement instruments.
Item measured Manufacturer Instrument model Method
Dissolved oxygen Nikkiso 7115–22 Membrane
polarography
Conductivity Nikkiso 9792
pH Horiba D-51 Glass electrode
Iron concentration Hach Pocket colorimeter TPTZ
II
water tank, the water temperature and flow rate in the loop were
measured. In the sampling system, water was depressurized to atmo-
spheric pressure and cooled to room temperature to monitor the dis-
solved oxygen concentration and conductivity, then water returned to
the main loop. Additionally, small amount of water was periodically
taken from the main loop in the nitrogen atmosphere and then pH and
iron concentration were measured manually under the room tempera-
ture. The measurement instruments for water chemistries are listed in
Table 1.
The test section for FAC rate measurement is shown in Fig. 2. The
test section was made of stainless steel (SUS304) pipe which had an
inner diameter of D = 49.5 mm. The corrosion sensors were made of
carbon steel plate (STPT42: Ni, 0.02 wt%; Cr, 0.04 wt%; Mo, 0.01 wt%)
and implanted on the pipe inner surface as shown in Fig. 3. The cor-
rosion sensors were insulated from the stainless steel pipe by a covering
resin and were implanted on the pipe inner surface so that one surface
of the plate was exposed to the fluid (hot water).
Twenty corrosion sensors were installed: four at 9.35D upstream
from the entrance of the elbow in the straight pipe, four at the center of
the elbow pipe, four each at 2.68D and 3.68D downstream, and two
each at 4.68D and 5.68D downstream from the exit of the elbow. It
should be noted that the elbow pipe was bent in the horizontal plane,
namely, the upstream pipe, the elbow pipe and the downstream pipe
were placed in the horizontal plane.
The detailed geometry of the elbow pipe used in the FAC rate
measurement experiment is shown in Fig. 4. Its geometry was different
from that of normally used elbow types, because of the difficulty to
install the corrosion sensors on the wall surface. In the previous studies
(Utanohara et al., 2012, 2013), the inner surface of the test section was
smoothed using a boring machine to remove undesirable bumps be-
tween the corrosion sensors and the pipe wall and to get a tight fit.
However, such machining cannot be applied on the elbow surface.
Hence, the corrosion sensors were installed in a straight part of 54 mm
Y. Utanohara, M. Murase
Fig. 2. Test section for FAC rate measurement, overhead view (unit: mm).
that was sandwiched between two 45-degree elbows with the curvature
radius of Re = 76.2 mm. The equivalent curvature radius of the elbow
pipe was Rep = 114.4 mm.
The layout of corrosion sensors as viewed from upstream is shown
in Fig. 4. Four corrosion sensors were installed in the circumferential
direction at the center of the elbow pipe. The directions were the upper,
the lower, the extrados and the intrados of the elbow pipe.
2.2. Measurement method for FAC rate
The FAC rate was measured by the electric resistance method. The
electric resistance of the sensor plate R is expressed by Eq. (1):
L L
R = ρe = ρe
S d·w (1)
where ρe, S, L, d and w are the electric resistivity, cross-sectional area,
length, thickness and width of the sensor plate, respectively. The
thickness d can be measured from the resistance R and then the FAC
rate of the sensor plate can be obtained from the gradient of the time
history of metal loss.
Since ρe depends on temperature, the measured resistances are also
influenced by temperature. In the experiment, temperature was con-
trolled to a constant value (150 °C) and temperature change during the
experiment was within about ± 1.0 °C. The influence of the tempera-
ture change on the final result (FAC rate) were under 0.3%, then ρe
could be treated as a constant value in this study.
Fig. 3. Corrosion sensors.
22
Nuclear Engineering and Design 342 (2019) 20–28
2.3. Experimental conditions
The experimental conditions are shown in Table 2. The experiment
had been conducted from Run 1 to Run 9 without interruption and the
conditions of flow velocity and temperature were changed when the
experiment moved to next run. The experimental period of each run
was over a few days. Before Run 1, the surfaces of the corrosion sensors
were polished to remove any oxide films that had formed. Velocity was
gradually increased through Run 1 to Run 3. The 2014 Run 4 was a
previously conducted run and it was used as a reference in comparing
the present results. Temperature was nominally controlled to 150 °C
through Run 1 to Run 5, but there was variation because heat balances
of the test section were different; in other words, the inlet enthalpy rate
increased as velocity increased, while heat release from the test section
was constant. The details are described in Section 3.1. After Run 6,
temperature was decreased step by step to 100 °C (Run 6 and Run 7)
and 50 °C (Run 8 and Run 9). The water was pressurized to 1.5 MPa to
avoid cavitation in the test loop. The dissolved oxygen concentration
was almost always below 0.1 μg/kg. Generally, FAC occurs when the
dissolved oxygen concentration is low and FAC rate decreases as the
dissolved oxygen increases. This is because the dissolved oxygen in-
fluences the composition of oxide films. In this study, by decreasing the
dissolved oxygen below 0.1 μg/kg, the influence of the dissolved
oxygen on FAC rate seemed to be reduced to negligibly low level.
Fig. 5 shows pH and iron concentration during Run 1 to Run 9. pH
and iron concentration were measured intermittently to confirm that
they could be kept at constant values. The pH was not controlled and
was nearly neutral at room temperature. Iron concentration was kept
Y. Utanohara, M. Murase Nuclear Engineering and Design 342 (2019) 20–28

Fig. 4. Detailed geometry of the elbow pipe (unit: mm).

Table 2
FAC experimental conditions.
Run 1 2 3 4 5

Experimental period [h] 444 175 141 332 78

Mean cross-sectional velocity [m/s] 0.39 1.40 4.34 5.74 4.73
Temperature [°C] 145.2 148.8 149.2 149.2 149.2
Reynolds number 9.7 × 104 3.6 × 105 1.1 × 10 6
1.4 × 106 1.2 × 106
Dissolved oxygen [μg/kg] below 0.1
pHa 6.8 (6.4–7.3) 6.8 (6.7–7.0) 7.3 7.5 7.1 (6.9–7.3)
Iron concentration [ppm] 0.02 to 0.06 0.03 to 0.07 0.05 0.05 0.05
Corrosion sensor property Carbon steel plates STPT 42 (Ni; 0.02%, Cr; 0.04%, Mo; 0.01%)

Run 6 7 8 9 2014 Run 4

Experimental period [h] 145 113 580 1624 503

Mean cross-sectional velocity [m/s] 4.73 1.37 1.34 4.78 3.64
Temperature [°C] 99.2 98.9 49.0 49.2 149.2
Reynolds number 8.0 × 105 2.3 × 105 1.2 × 105 4.3 × 105 9.0 × 105
Dissolved oxygen [μg/kg] below 0.1
pHa 7.3 (7.2–7.4) 6.6 7.5 – 6.9 (6.5–7.2)
Iron concentration [ppm] 0.05 0.06 0.02–0.03 – 0.03
Corrosion sensor property Carbon steel plates STPT 42 (Ni; 0.02%, Cr; 0.04%, Mo; 0.01%)
a
The average during the corrosion measurement. The numbers in parentheses are the minimum and the maximum of the measurement values.

3. Experimental results

3.1. Time history of metal loss

Fig. 6 shows time histories of metal loss of corrosion sensors in four

Fig. 5. pH and iron concentration.
constant at about 0.04 ± 0.02 μg/g, except at the start of Run 5. The
reason of the increase of iron concentration in Run 5 was not clear, but
the authors thought that it did not have great influence on the experi-
mental results.
runs. They are the values along the extrados of the elbow test section.
The origin of the longitudinal direction is the inlet of the elbow pipe.
The metal losses linearly increased with time, but their curves had
fluctuations. The range of fluctuations became smaller as velocity in-
creased. These fluctuations in metal losses were caused by thickness
measurement error and this error varied due to temperature fluctuation
of water. The measured FAC rates are related to the water temperature,
because the electric resistivity in Eq. (1) depends on temperature. The
main reason of the fluctuation of the water temperature was energy
balance between heat release to the outside and heat input from the
heater. The amount of the heat release was constant without depending
on the flow velocity. On the contrary, the amount of inflow enthalpy
decreased as the flow velocity decreased and temperature at the test
section also became lower. As the flow velocity increased, the amount
of inflow enthalpy increased and temperature drop became smaller. As
a result, the temperature fluctuation also became smaller. The fluc-
tuation of room temperature also influenced the water temperature as
shown in Fig. 6(b)–(d). Since the influence of the temperature drop
became smaller, the effect of the daily variation of the room

23
Y. Utanohara, M. Murase
Fig. 6. Time histories of metal loss of corrosion sensors.
temperature became visible. Anyway, the FAC rates can be obtained
from the gradient of the metal loss and these fluctuations can be re-
moved by the time-average of the piecewise gradients of the metal loss.
In Run 1, the metal losses changed with time and gradients became
constant after a certain period of time. In Fig. 6(a), the gradients be-
came relatively constant after around 190 h. A similar tendency was
observed in the previous study (Utanohara et al., 2016). On the other
hand, in Run 2 to Run 4, gradients were relatively constant from the
early period. This was because the surfaces of the corrosion sensors
were not covered by any oxide film at the beginning of Run 1 and FAC
rate decreased with the formation of the oxide film. When the gradient
of the metal loss curve became constant, thickness of the oxide film was
presumed to be steady. Since Run 2, Run 3 and Run 4 started from the
condition that the oxide films were already formed on the surface,
thickness of oxide film became steady relatively sooner than in Run 1.
Hence in Run 1, the time-averaged increasing rate of the metal loss
from 192 to 384 h was used to obtain FAC rates after reaching the
steady state.
3.2. FAC rate distribution
Fig. 7 shows the distribution of FAC rates. As mentioned in Section
3.1, time-averaged values are plotted in the figure. Tendency of the
distribution was the same as the previous study (Utanohara et al.,
2016): the FAC rate was smaller at the intrados of the elbow pipe than
at other circumferential locations. The FAC rates at the elbow pipe were
larger than those upstream and downstream from the elbow pipe and
the FAC rates downstream from the elbow pipe decreased along the
flow direction. The possible reason for the low FAC rate at the intrados
was the low velocity near the wall surface due to separation in the
elbow. In the previous study (Tsuneyoshi et al., 2017) numerical si-
mulations for the elbow geometry employed in this study was con-
ducted as shown in Fig. 8. In the figure, the simulated flow velocities
are compared between the different elbow geometries, short elbow,
24
Nuclear Engineering and Design 342 (2019) 20–28
long elbow and the elbow test section in this study. As a result, rela-
tively lower velocity distribution appeared near the intrados side and
this velocity distribution continued downstream from the elbow.
Hence, mass transfer near the intrados side might be also lower. Flow
separation occurred most strongly in the short elbow and did not occur
in the long elbow. Regarding the elbow geometry in this experiment,
relatively weak separation occurred at the first elbow and flow velocity
near the extrados side became higher than those of short and long el-
bows. In this way, the elbow geometry influences flow field and prob-
ably the mass transfer coefficient near a pipe wall. For more details on
the influence of the geometry difference, comparison between the
measured FAC rates and simulated thermal flow fields remains as future
work.
The FAC rate distributions under different flow velocities showed
the same tendencies. The ratios of the maximum FAC rate at the elbow
to the FAC rate in the upstream straight pipe were about 2.4 in Run 1,
1.7 in Run 2, 2.9 in Run 3 and 2.14 in Run 4.
3.3. Influence of the flow velocity
Fig. 9 shows the influence of flow velocity on FAC rates along the
extrados and intrados. Except for the extrados in Run 3, as velocity
increased, FAC rates also increased at all locations. The shape of the
distribution was not changed and FAC rates at the elbow were larger
than those at the other locations.
To clarify further the influence of flow velocity on FAC rate, FAC
rates at the extrados and intrados are shown in terms of flow velocity in
Fig. 10. The figure shows that FAC rates increased as flow velocity in-
creased while the tendency of Run 3 was different from the others. The
reason of the difference was not clear. If Run 3 is left aside here, the
relationship was monotonic increase, but it was not linear. Previous
studies also showed the same tendency (Bignold et al., 1981; Heitmann
and Schub, 1983; Utanohara et al., 2013; Yoneda et al., 2016). How-
ever, the relationship between the FAC rate and flow velocity depends
Y. Utanohara, M. Murase Nuclear Engineering and Design 342 (2019) 20–28

Fig. 7. FAC rate distributions (145–149.2 °C).

Fig. 8. Contour of time-averaged streamwise velocity of the elbow for

Re = 45,000 (Tsuneyoshi et al., 2017). The geometry of INSS elbow is the same
as that employed in this experiment.

on experimental facilities. The previously obtained FAC rates were

proportional to the power of flow velocity with an exponent value from
1.0 to 2.3 (Yoneda et al., 2016). This was because the FAC rate is also
influenced by turbulence near the wall as well as flow velocity (The
Japan Society of Mechanical Engineers, 2014) and turbulence depends
on the scale of the experimental facilities.
Hence, to understand the influence of the flow velocity on FAC rate,
an accurate flow field is required, particularly near the wall. To do this,
numerical simulation for the elbow test section employed in this study
has been being carried out (Tsuneyoshi et al., 2017) to simulate the
flow field and mass transfer in the elbow test section. Comparison be-
tween simulated flow fields and FAC rates for various flow velocities
remains as future work.

3.4. Influence of temperature

Fig. 9. FAC rate distributions under different flow velocities.

Fig. 11 shows FAC rate distributions under different temperature

25
Y. Utanohara, M. Murase Nuclear Engineering and Design 342 (2019) 20–28

conditions. In the case of 150 °C (Run 3), as described in Section 3.2, the
FAC rate was smaller at the intrados of the elbow pipe than that at other
circumferential locations. However, when temperature decreased to
100 °C (Run 6), the tendency changed and FAC rate at the intrados
became the largest of the four circumferential sensors. Since flow ve-
locities in Run 3 and Run 6 were the almost same, the reason of the
change seemed to be due to temperature decrease from 150 °C to
100 °C. Though the detailed reason was not clear at this time, one of the
possible reasons was that flow structure in the elbow changed such as
separation point on the intrados due to the change of physical prop-
erties.
When temperature decreased further to 50 °C, FAC rates decreased
overall and FAC rate at the intrados of the elbow was still the largest. As
mentioned in the Section 1, FAC usually occurs over 100 °C with a peak
near 130–150 °C (Bignold and Garbett, 1982). However, FAC rate at
50 °C was not negligible, 0.09 mm/y at the intrados of the elbow. The
ratio of the maximum FAC rates at 50–100 °C was 0.2. The tendency of
FAC at low temperature agreed with the experimental results by
Yoneda et al. (2016).
Fig. 12 shows combined effect of flow velocity and temperature on
FAC rate. Compared between results of same flow velocities, FAC rates
at 100 °C were larger than those of 150 °C and 50 °C. This result was
slightly different from previous study (Bignold and Garbett, 1982), in
which FAC rate became peak near 130–150 °C. FAC rates at 50 °C in
Run 8 and Run 9 were comparable with the values in Run 2, 150 °C and
lower flow velocity of 1.40 m/s.
To clarify the combination effect of flow velocity and temperature,
the increase ratios of FAC rates due to flow velocity increase are sum-
marized in Table 3. It should be noted that ratios of 50 °C were under
1.0 and it means FAC rates decreased as flow velocity increased.
However, the authors thought that this was unreasonable and these
values could be treated as 1.0, that is, the change of FAC rates were
negligibly small. Anyway, the influence of flow velocity was small at
Fig. 10. Influence of the flow velocity on FAC rates.
50 °C and remarkable at 100 °C and 150 °C. Similar tendency was

Fig. 11. FAC rate distributions under different temperatures.

26
Y. Utanohara, M. Murase
Fig. 12. Influence of the flow velocity and temperature on FAC rates.
Table 3
Increase ratios of FAC rate due to flow velocity increase.
Temperature [°C] Flow velocity [m/s] Upstream
Extrados
150 1.40 to 4.73 1.5
100 1.37 to 4.73 1.7
50 1.34 to 4.78 0.7
reported in the previous study (Bignold and Garbett, 1982), that is, the
influence of flow velocity was remarkable near 130–150 °C. The com-
bination effect of flow velocity and temperature was slightly different
from those at upstream and downstream from the elbow. The ratios at
the elbow were relatively larger than the others. It indicated that the
combination effect of flow velocity and temperature was different from
place to place and particularly strong at the elbow.
Generally, FAC process involves at least two steps, corrosion pro-
cesses (steel oxidation, oxide dissolution, etc.) and mass transfer of the
corrosion products to the bulk solution (Dooley and Chexal, 2000).
Temperature affects the corrosion processes and flow field (flow velo-
city and turbulence) affects the mass transfer. Hence, it can be said from
the experimental results that the solubility of ferrous ion increased with
temperature rise and the mass transfer also increased with flow velocity
increase. Additionally, pipe geometry such as elbow enhanced the
turbulent mixing and the mass transfer was also enhanced. On the
contrary, when temperature was 50 °C, the solubility might be too small
and the change of flow field (flow velocity and pipe geometry) was not
affect the FAC rate so much. Hence in the low temperature FAC, the
bottleneck of the FAC process might be the corrosion processes.
4. Conclusions
In this study, FAC rates at an elbow pipe were measured under
different flow velocity and temperature conditions to understand the
effects of thermal flow field on FAC rate.
The FAC rate was smaller at the intrados of the elbow pipe than at
other circumferential locations when fluid temperature was 150 °C. The
FAC rates at the elbow pipe were larger than those upstream and
downstream from the elbow pipe and the FAC rates downstream from
the elbow pipe decreased along the flow direction. FAC rates increased
as flow velocity increased and their relationship was not linear. The
ratios of the maximum FAC rate at the elbow to the FAC rate in the
upstream straight pipe were about 1.7–2.9.
When temperature decreased to 100 °C, FAC rate at the intrados
became the largest of the four circumferential sensors. When tem-
perature decreased further to 50 °C, FAC rate decreased, but not neg-
ligible, 0.09 mm/y at the intrados of the elbow. Compared between
results of same flow velocities, FAC rates at 100 °C were larger than
Nuclear Engineering and Design 342 (2019) 20–28
Elbow Downstream
Intrados Extrados Intrados Extrados Intrados
1.1 2.1 2.8 2.0 1.8
1.8 1.8 1.9 1.7 1.7
0.8 0.9 0.9 0.7 0.7
those of 150 °C and 50 °C. Regarding the combination effect of flow
velocity and temperature, the influence of flow velocity was negligibly
small at 50 °C and remarkable at 100 °C and 150 °C. Additionally, the
combination effect of flow velocity and temperature was different from
place to place and particularly strong at the elbow. The combination
effect of flow velocity and temperature can be explained by the cor-
rosion processes and the mass transfer. When temperature was 50 °C,
the solubility might be too small and the change of flow field (flow
velocity and pipe geometry) did not affect the FAC rate so much.
To understand more clearly the influence of the thermal flow field
on FAC rate, numerical simulation for the elbow test section employed
in this study have been conducted. Comparison between measured FAC
rates and simulated thermal flow fields remains as future work.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.nucengdes.2018.11.022.
References
Bignold, G.J., Garbett, K., Garnsey, R., Woolsey, I.S., 1981. Erosion-corrosion in nuclear
steam generators. In: Water Chemistry of Nuclear Reactor Systems 2. British Nuclear
Engineering Society, pp. 5–18.
Bignold, G.J., Garbett, K., Woolsey, I.S., 1982. Mechanistic aspects of temperature de-
pendence of erosion-corrosion. Paper No. 12 In: Berge, Ph., Kahn, F. (Eds.),
Corrosion-Erosion of Steels in High Temperature Water and Wet Steam. Electricite de
France, Les Renardieres, France.
Dooley, R.B., Chexal, V.K., 2000. Flow-accelerated corrosion of pressure vessels in fossil
plants. Int. J. Pressure Vessels Pip. 77 (2–3), 85–90.
El-Gammal, M., Mazhar, H., Cotton, J.S., Shefski, C., Pietralik, J., Ching, C.Y., 2010. The
hydrodynamic effects of single-phase flow on flow accelerated corrosion in a 90-
degree Elbow. Nucl. Eng. Design 240, 1589–1598.
Fujisawa, N., Yamagata, T., Kanno, S., Ito, A., Takano, T., 2012. The mechanism of
asymmetric pipe-wall thinning behind an orifice by combined effect of swirling flow
and orifice Bias. Nucl. Eng. Design 252, 19–26.
Heitmann, H.G., Schub, P., 1983. Initial experience gained with a High pH value in the
secondary system of PWRs. In: Water Chemistry of Nuclear Reactor Systems 3. vol. 1.
British Nuclear Engineering Society, pp. 243–252.
Kamahori, K., Utanohara, Y., Nakamura, A., Murase, M., Nagaya, Y., 2013. Flow ac-
celerated corrosion downstream from an Orifice (1. Measurement of corrosion rate).
In: Proceedings of the Conference on Flow Accelerated Corrosion (FAC2013),
Avignon, France, May 21-24, Paper No.1234.
Pietralik, M.J., Schefski, S.C., 2011. Flow and mass transfer in bends under flow-
27
Y. Utanohara, M. Murase
accelerated corrosion wall thinning conditions. ASME J. Eng. Gas Turbines Power
133, 012902.
The Japan Society of Mechanical Engineers, Rules on Pipe Wall Thinning Management,
JSME S CA1-2005 (2006) (in Japanese).
The Japan Society of Mechanical Engineers, Technical Report, Research Committee on
Improvement and Practical Use of Pipe-Wall-Thinning Management, P-SCCII-4,
(2014), (in Japanese).
Tsuneyoshi, T., Kamiya, K., Ito, T., Tsuji, Y., 2015. Measurement of mass transfer coef-
ficients and analyses of turbulent scalar transport with LES behind the orifice in a
circular pipe. Trans. JSME 81 (828) Paper No. 15–00162 (in Japanese).
Tsuneyoshi, T., Ito, T., Utanohara, Y., Tsuji, Y., 2017. Study on effects of wall mass
transfer condition for measurement of geometry factor. Trans. JSME 83 (847) Paper
No. 16–00417 (in Japanese).
Utanohara, Y., Nagaya, Y., Nakamura, A., Murase, M., 2012. Influence of local flow field
on flow accelerated corrosion downstream from an orifice. JSME J. Power Energy
28
Nuclear Engineering and Design 342 (2019) 20–28
Syst. 6 (1), 18–33.
Utanohara, Y., Nagaya, Y., Nakamura, A., Murase, M., Kamahori, K., 2013. Correlation
between flow accelerated corrosion and wall shear stress downstream from an orifice.
JSME J. Power Energy Syst. 7 (3), 138–147.
Utanohara, Y., Kamahori, K., Nakamura, A., Murase, M., 2016. Measurement of flow
accelerated corrosion rate at an elbow pipe and combination effect of an upstream
orifice. E-J. Adv. Maintenance 8 (1), 1–12.
Yoneda, K., Ohira, T., Tanji, K., Akiba, S., Niiyama, K., Morita, R., Inada, F., 2009.
Evaluation of hydraulic factors affecting flow accelerated corrosion and its verifica-
tion with power plant data. In: Proceedings of the ASME 2009 Pressure Vessels and
Piping Division Conference (PVP2009), Prague, Czech Republic, July 26–30, Paper
No. PVP2009-77486.
Yoneda, K., Morita, R., Fujiwara, K., Inada, F., 2016. Development of flow-accelerated
corrosion prediction method (1) acquisition of basic experimental data including low
temperature condition. Mech. Eng. J. 3 (1) 15-00232.