Energy and Buildings 41 (2009) 721731

Contents lists available at ScienceDirect

Energy and Buildings

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

Durability of 20-year-old external insulation and assessment of various types of retrotting to meet new energy regulations
`a F. Stazi a,*, C. Di Perna b, P. Munafo
a b

Department of Architecture, Construction and Structures (DACS), Faculty of Engineering, Polytechnic University of Marche, Via Brecce Bianche, 60131 Ancona, Italy Department of Energy, Faculty of Engineering, Polytechnic University of Marche, Via Brecce Bianche, 60131 Ancona, Italy

A R T I C L E I N F O

A B S T R A C T

Article history: Received 3 November 2008 Received in revised form 26 January 2009 Accepted 12 February 2009 Keywords: External insulation Monitoring Laboratory tests Thermo-graphic survey Hygrothermal performance Chemicalphysical conservation Parametric analyses

The exterior envelope of some social housing scheme buildings constructed at the beginning of the 1970s without thermal insulation has proved to be the cause of great thermal loss and condensation. At the start of the 1980s, in order to resolve these problems, our research group carried out a study which led to the introduction of external thermal insulation (on the basis of previously developed performance specications) and verication of the thermal performance achieved. With the aim of verifying the efcacy of the intervention after 20 years and in order to assess the thermalhygrometric performance and the state of conservation of the exterior envelope we carried out a two stage study: 1. Performance analysis carried out through monitoring and laboratory tests. 2. Formulation of hypotheses for retrotting, assessed through simulations and parametric analysis. The results showed the efcacy and durability from the thermalhygrometric and mechanical point of view of the external insulation applied in the 1980s. It was also possible to verify energy saving for the different types of retrot scenarios and to identify the correct positioning of the thermal insulation on the brickwork and on the oors so as to increase the surface temperatures in winter phase. 2009 Elsevier B.V. All rights reserved.

1. Introduction At the start of the 1980s, our research group realized the retrotting of social housing buildings through the introduction of external thermal insulation in order to reduce either the risk of condensation or energy consumption. The present study is carried out with the aim of verifying whether after 20 years this type of intervention had maintained its efcacy. In the rst phase of the research we carried out a series of monitoring activities that allowed us to evaluate the thermal hygrometric performance and the state of conservation of the exterior insulation. Studies focusing on long-term assessment of thermalhygrometric behaviour of external insulation coatings regard mainly component recently introduced into building technology (vacuum insulation panel, transparent insulation) and are about experimental analyses on laboratory-induced deteriorations [1,2]. Very

* Corresponding author. Tel.: +39 071 2204783; fax: +39 071 2204783. E-mail address: f.stazi@univpm.it (F. Stazi). 0378-7788/$ see front matter 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.enbuild.2009.02.008

interesting the experimental studies that focus on the effects of atmospheric deposition on the cement plaster [3], the durability of the nishing [4,5] and the degradation risk assessment of external envelopes [6]. In the second phase we simulated the buildings in dynamic conditions and we calibrated the virtual model by comparison with the experimental results. Several researchers studied the importance of the monitoring activities in order to calibrate virtual models as close as possible to the reality [79]. This made it possible to achieve a correct, realistic and reliable virtual model of the as-built case in which parameters could be varied by changing the characteristics at every new level of investigation (parametric analysis): it was possible to predict the effect of various types of retrotting actions which respect the increasingly strict performance criteria laid down by new energy regulations and to compare the different conformations of exterior envelope to be used if designing new buildings. The new regulations on energy saving, including Italian legislative decree No. 311/2006, aim at maximum reduction in wintertime heat dispersion and consequently at thermal superinsulation.

722

F. Stazi et al. / Energy and Buildings 41 (2009) 721731

above ground level (Figs. 1 and 2). It is made up of four apartments, two on the rst oor and two on the second. The ground oor houses a cellar space. The exterior envelope is made up of the layers shown in Table 1 veried using endoscopic inspection. The envelope of the ground oor does not have exterior insulation. The apartments studied are on the rst oor: one faces the north-east (squared hatch in Fig. 2) and the other the southwest (lined hatch in Fig. 2). 3. Methodologies Monitoring and simulations were carried out in semi-stationary conditions so as to measure the heat transmission rate, and in dynamic conditions to study any aspects connected with the thermal inertia of the envelopes (time lag, decrement factor) by considering the variability over time of the thermal ow. In particular measurements were carried out inside the apartments in two ways:
Fig. 1. View of the building from the south-east.

Some researchers stressed that these kind of regulations, which set limits exclusively on transmittance and consumption values, brought about within the same climatic zone the use of envelopes seemingly with the same performance, but with very different behaviours as concerns comfort and internal surface temperature [1012]. Our study made it possible to verify energy saving and internal surface temperatures for the different types of retrot scenarios and to identify the correct positioning of the thermal insulation on the brickwork and on the oors in order to increase the surface temperatures in winter phase. 2. The case study The study was carried out on a building located in the city of Pievetorina (central Italy), characterized by an Apennine climate and by 2189 degree days. The building is compact with two oors

- asking the users to keep the heating system continuously on during the data acquisition period with the internal temperature set-point at 20 8C. This apartment is indicated with squared hatch in Fig. 2. - leaving the occupier freedom to select intermittent use of the heating system: this apartment is indicated with lined hatch in Fig. 2. The occupiers switched on the heating system between 7.00 a.m. and 9.00 p.m. with the thermostat temperature set at 18 8C. 3.1. Experimental facility The following investigations were carried out on the building between 13 February and 10 March: 1. survey of the outdoor environmental conditions using an external weather station that included: hygro-thermal probe, wind speed and direction probe, solar radiation probe.

Fig. 2. Plan of the rst oor.

F. Stazi et al. / Energy and Buildings 41 (2009) 721731 Table 1 Layers making up the exterior envelope of the building in the as built condition. Material (from inner to outer side) Cement plaster Perforated brick Air gap Cement mortar Perforated brick Plaster Expanded polystyrene Plaster s [cm]

723

l [W/m2 K]
0.900 0.510 1.400 0.400 0.044 0.300

C [W/m2 K]

R [m2 K/W]

- hygrometric performance (water content). - chemical and physical state of conservation (analysis of the composition and presence of chemical, physical and biological deterioration). 3.2. Numerical investigations Simulations were carried out in semi-stationary and dynamic conditions. By comparing the values measured on the building during the monitoring activities with the calculated values it was possible to verify the efcacy of the simulation tools and to develop virtual models which are closer to the real situation. The calibration of the model involved: - development of a climate le indicating the real outdoor environmental conditions monitored. - denition of programmes which reect reality as regards the particular use of the heating system, the air inltration (also connected with the opening of the windows) and the internal gains due to the occupants and the indoor lighting. Using the calibrated model described above some input parameters were varied in order to verify several hypotheses for energy retrotting and identify the best solutions to use when designing the structure of the envelope in new buildings. The types of retrotting analysed concern the insulation of the vertical envelope, the insulation of the oors and the insulation of the roof, the substitution of single with double glazing, the introduction of a condensing boiler and high performance glasses. With a view to new design, different thicknesses of thermal insulation placed in different positions on the walls and oors were compared.

1.00 8.00 9.00 1.00 12.00 1.00 4.00 0.50

5.00 3.22 -

0.01 0.2 0.18 0.01 0.31 0.03 1.11 0.02 K = 0.536 W/m2 K

2. survey of microclimate conditions in the apartments monitored using indoor microclimate stations: in particular we recorded the dry-bulb temperature (this probe was shaded for radiation heat ux), the wet-bulb temperature, the mean radiant temperature. 3. endoscopic inspection using a exible tube optical bre endoscope in order to identify the layers making up the wall (Table 1). 4. detailed thermo-graphic survey of the external envelope: using an infrared thermo-camera in order to verify the presence of thermal bridges and the efcacy of the anchoring of the insulation panels to the brickwork; using a ow-meter to highlight the thermal ows through the wall; using a set of thermo-resistances (Fig. 3) to measure the external and internal surface temperatures at the centre of the wall, on the thermal bridge close to the window frame and on the thermal bridge of the supporting column (Fig. 4). 5. a series of laboratory tests on samples of indoor and outdoor plaster so as to identify:

Fig. 3. Thermo-resistance for measuring the surface temperature.

Fig. 4. Plan of a room indicating the position of the probes.

724

F. Stazi et al. / Energy and Buildings 41 (2009) 721731

Fig. 5. Graph showing the trend in the surface temperature values monitored.

4. Results 4.1. Results of monitoring In order to verify the performance of the external thermal insulation after 20 years the heat transmission rate values measured during the 1987 experimentation were compared with the measurements made in 2008 (Table 2). Theoretical values of the heat transmission rates were calculated (according to UNI EN ISO 6946:1999) before and after the application of the external thermal insulation to the outside wall: the result was a considerable reduction in the initial value (from 1.05 to 0.54 W/m2 K). During the inspection carried out following the 1987 experimentation the heat transmission rate of the insulated wall was measured in semi-stationary conditions (using two measuring points), resulting in a value of 0.50 W/m2 K at one point and 0.55 W/ m2 K at the other, therefore conrming the value obtained through calculation. The recent experimentation also involved the measurement of the heat transmission rate: this value was 0.410.42 W/m2 K in the apartment studied in semi-stationary conditions (i.e. with the heating system continuously on) and 0.38 W/m2 K in the apartment studied in dynamic conditions (i.e. with the heating plant intermittently on). The results show that the heat transmission rate after 20 years has decreased. In our opinion this is probably due to the progressive reduction in water content of the materials which make up the envelope. Monitoring of the surface temperatures (Fig. 5) showed that in the apartment studied in dynamic conditions where the occupier switched on the heating system between 7.00 a.m. and 9.00 p.m. with a thermostat temperature set at 18 8C, the surface temperatures of the wall (Ts1) always stayed above 14 15 8C and did not follow the changes in the outside surface temperature (Ts2). Even the thermal bridge corresponding to the window (Ts PT2) was reasonably correct. These probes are

positioned in the central part of the wall where the outside insulation is continuous. The surface temperatures recorded on the thermal bridges corresponding to a corner column (probes TsPt3 and TsPt4 respectively on the extrados of the oor and on the lower part of the vertical wall) the surface temperatures go down to 13 8C. In fact the measuring points in this case were located at the bottom of the wall (almost at oor level) and were inuenced by the fact that the outside wall of the ground oor (used as a garage) had not been insulated. This shows the importance of the presence of an external insulation coating. 4.2. Measurement uncertainty The tool used to study the thermal ows works by acquiring two types of data: the value q of the thermal ow in W/m2 K and the values of the indoor and outdoor surface temperatures in 8C. The error in the percentage of heat transmission rate is a combination of the error in the data acquired by the ow-meter and the errors made by the thermometric probes: we were interested in nding out which of the two values had greater inuence on the calculated heat transmission rate error. Uncertainty is evaluated as a percentage of accuracy and a percentage of resolution declared by the manufacturers (Table 3, Figs. 6 and 7). For the thermometric probes the acquired data had a tolerance of 0.15 8C distributed equally for all the values measured, while in the case of the ow-meter the accuracy was determined by the formula 5%VL + 1 W/m2 (VL = read value) and results in a mean square error which uctuates over time. 4.3. Results of the thermo-camera survey The measurements carried out with the thermo-camera (Fig. 8) conrmed the results obtained through the supercial thermal

Table 2 Heat transmission rate values measured during the 1987 experimentation compared with the measurements made in 2008. Heat transmission rate of the exterior envelope K [W/m2 K] calculated (UNI EN ISO 6946:1999) before the application of the external insulation calculated after the application of the external insulation measured during the 1987 experimentation measured in the recent experimentation with the heating system continuously on (southern wall) 0.41 measured in the recent experimentation with the heating system continuously on (northern wall) 0.42 measured in the recent experimentation with the heating system intermittently on (northern wall) 0.38

1.05

0.54

0.50.55

F. Stazi et al. / Energy and Buildings 41 (2009) 721731

725

deterioration was conrmed by the thermo-graphic images and it was possible to note that the problem is present above all facing south where the changes in temperature are greater and more frequent. The vertical cracks create some areas exposed to the external environment characterized by very high temperatures with values up to 4045 8C (see Point 2 in Fig. 10 and Temperature Prole line in Figs. 11 and 12) during the winter time. The thermal daily uctuation makes the situation critical. The temperatures at the same hour (11:00 a.m) in the northern wall are lower: we recorded values of about 5 8C (Figs. 13 and 14). 4.4. Results of the laboratory tests 4.4.1. Water content The moisture content was determined by drying at a high temperature according to UNI EN ISO 12570:2001. This regulation states that the constant mass of the sample is reached when the variation in mass over 3 successive weightings carried out 24 h from each other is less than 0.1% of the initial mass (Table 4). The laboratory test obtained a mass moisture content per unit of mass of: u m mo 10:2325 g 10:1326 g 0:0099 mo 10:1326 g 0:99%

Figs. 6 and 7. Uncertainty compared with the surface temperatures and ows.

This value is better than acceptable, showing the real soundness of the system even after 20 years. 4.4.2. X ray diffractometry The diffraction study on the sample of outside plaster did not reveal any type of chemical deterioration (Fig. 15). In fact, in the diffraction spectrum shown below no peaks can be noted to indicate the presence of harmful substances which can be attributed for example to ettringite. The main substances which make up the sample are silicon dioxide (SiO2) and calcium carbonate (CaCO3) which are evident on the spectrum with very high reections.

probes: the external envelope is characterized by a grate difference of temperature between the ground oor and the rst oor. The insulation panel in the rst plan is still effective and anchored to the supporting masonry. Visual inspection indicated the presence of vertical cracks in the panels (Fig. 8) probably due to thermal dilatation. In fact this deterioration is found on the south-west wall which has greater exposure to sunlight (Fig. 9). The hypothesis of physical

Table 3 The probes used and resolution and accuracy values.

Fig. 8. Thermo-graphic image of the joint between the ground oor with no insulation and the rst oor with external insulation.

726

F. Stazi et al. / Energy and Buildings 41 (2009) 721731

Fig. 9. Thermo-graphic image of the south-west wall.

Fig. 10. Detail of the thermo-graphic image.

The thermo-gravimetric analysis (Fig. 16), carried out at the same time as the differential thermal analysis, consists in heating the sample in the air and recording the changes in weight compared with a sample of material which is thermally stable and inert (calcinated kaolin) subjected to the same thermal treatment. The changes in weight induced by heating the sample may be a consequence either of endothermal transformations such as dehumidication (loss of free moisture), dehydration (loss of bound, combined or crystallized moisture), decomposition (loss of volatile substances produced by thermal decomposition), or of exothermic transformations such as combustion (loss of volatile combustion products). Moreover, any peaks in the differential thermal analysis which are not associated with weight loss are caused by so-called allotropic (or crystalline phase) transformations, which may be endothermic or exothermic, as happens for the exothermic transformation of quartz from a to b shape at a temperature of about 570 8C.

The samples of plaster examined show signs of both biological deterioration in terms of organic carbon present, and of endothermic or exothermic reactions which have developed. In particular, the sample analysed, taken from the exterior plaster of the external insulation, showed peaks corresponding to: - Exothermic peaks at %325 8C Organic carbon attributable to biological deterioration of unknown origin; - Endothermic peaks at %550 8C Phase transition (not clearly visible) of the crystalline quartz; - Endothermic peaks at %785 8C Decomposition of carbonates which corresponds to the loss of CO2. 4.5. Results of the retrotting simulations Starting from the case study in the as built condition, the following parametric variations were carried out:

Figs. 11 and 12. Detail of the thermo-graphic image with the temperature prole line.

F. Stazi et al. / Energy and Buildings 41 (2009) 721731

727

Figs. 13 and 14. Thermo-graphic image of the south-east wall.

Table 4 Determination of the moisture content by drying at a high temperature according to UNI EN ISO 12570:2001: the constant mass of the sample is reached when the variation in mass over 3 successive weightings carried out 24 h from each other is less than 0.1% of the initial mass. initial mass = 10.2325 g; 0.1% total mass = 0.0102 g Date and hour 01/04/08 12.00 02/04/08 12.00 03/04/08 12.00 Box weight (g) 2288 2288 2288 Box + sample (g) 12,4214 12,4208 12,4206 Sample-weight (g) 10,1334 10,1328 10,1326 0.0102>

D (initial mass-sample weight) (g)

0.006 0.0002 0.0002

1) introduction of thermal insulation in the internal side of the roof (8 cm of EPS) in order to reach a heat transmission rate of 0.32 (W/m2 K); insulation of the internal side of the oor (4 cm of EPS) in order to reach a heat transmission rate of 0.38 (W/m2 K). 2) substitution of single glazing with double glazing 4/10/4 with10 cm air cavity. 3) insulation of the walls of the cellar situated on the ground oor in order to reach a heat transmission rate of 0.8 (W/m2 K). 4) insulation of the walls of the cellar situated on the ground oor and insulation of the walls of the stairs cell. 5) a combination of variations 2 and 3. 6) a combination of variations 1, 2, 3. 7) a combination of variations 1, 2, 3 and increased thickness of the existing external insulation.

8) variation 7 and substitution of the 24 kW boiler with a condensing boiler. 9) variation 8 and argon gas ll in the air cavity of the glass. After retrotting the building has an EP (energy performance index for winter heating) limit of 98.66 kWh/m2 year. From the table and the graph (Table 5 and Fig. 17) it can be seen that the thermal insulation of the oors (variation 1) has a greater inuence on energy saving than the other variations. The roof and the rst oor (leaning to the unheated garage) are the building elements characterized from the bigger heat dispersal. This type of intervention is better than the increase of the thickness of the existing external insulation in the vertical walls. Quite interesting is the substitution of single glazing with double glazing, the adoption of a super-glazing (with argon) and the replacement of heating system with a Condensation boiler. 4.6. Results of the simulations for the external envelope The following parametric variations were carried out on the exterior envelope: 1) Thermal insulation both on the internal (s = 3 cm) and on the external side (s = 4 cm) of the wall. 2) Super thermal insulation placed as a coat on the outside surface (s = 10 cm). 3) Traditional wall with internal insulation (s = 9 cm). This casestudy result from a traditional wall realized before the energetic saving regulations without any type of thermal insulation. The improving of the thermal performance could be resolved by the injection of polyurethane foam in the existing internal air gap.

Fig. 15. Result of X ray diffractometry.

728

F. Stazi et al. / Energy and Buildings 41 (2009) 721731

Fig. 16. Result of the thermo-gravimetric analysis.

Table 5 Parametric variations carried out starting from the case study in the as built condition in order to verify several hypotheses for energy retrotting. Retrotting simulations EP (energy performance index for winter climatization) (kWh/m2 anno) 227.89 145.18 212.56 227.04 225.80

As built 1introduction of thermal insulation in the internal side of the roof (8 cm EPS) and in the internal side of the oor (4 cm EPS) 2substitution of single glazing with double glazing 3insulation of the walls of the cellar 4insulation of the walls of the cellar situated on the ground oor and insulation of the stairs walls Combination 5a combination of variations 2 and 3 6a combination of variations 1, 2, 3 7a combination of variations 1, 2, 3 and increased thickness of the existing external insulation 8variation 7 and substitution of the 24 kW boiler with a condensing boiler 9variation 8 and argon gas ll in the air cavity of the glass

210.10 128.22 118.37 105.18 93.87

A preliminary thermo-hygrometric assessment was carried out to evaluate any formation of condensation in the layers of the solutions hypothesized. As a result of this analysis, the hypothesis 3 concerning the traditional wall, was rejected since interstitial condensation was found at the insulationbrick interface (Fig. 18). The comparison between the as built condition (dotted grey line in Fig. 19) and the two alternative solutions hypothesized shows that it is preferable to place the insulating material on the outside surface. In fact thermal insulation on the inside surface prevents the inertial mass from stabilising the uctuations in temperature. On the contrary, in the case of

Fig. 17. Energy saving as a result of different types of retrotting.

F. Stazi et al. / Energy and Buildings 41 (2009) 721731

729

Fig. 18. Formation of condensation between layers 3 and 4 of the traditional wall.

Fig. 19. Surface temperatures of the as built wall and of the two alternative walls studied.

Fig. 20. Glaser diagram for the wall in the as built condition.

730

F. Stazi et al. / Energy and Buildings 41 (2009) 721731

Fig. 21. Glaser diagram for the wall studied with insulation on the inside surface.

Fig. 22. Comparison between the two types of oor insulation.

external insulation (green) the mass determines a reduction in surface temperature oscillations. Moreover the insulation of the inside surface of the envelope determines an increase in the vapour pressure, as can be seen in the Glaser diagram (Figs. 20 and 21). The following parametric variations were carried out on the internal oor (Fig. 22):

4) oor insulation placed on the extrados. 5) oor insulation placed on the intrados. A 0.5 8C increase in temperature is recorded after adding insulation to the extrados of the oor. On the contrary tting the insulating material on the inside surface leads to great changes in temperature, thereby causing discomfort (Fig. 23).

Fig. 23. Inside surface temperatures of the as built oor (without insulation) and the two alternative oors studied (with insulation on the intrados and extrados).

F. Stazi et al. / Energy and Buildings 41 (2009) 721731

731

The thermal mass is better to be positioned in the internal side of the envelope, to have comfort during the winter time. Positioning the mass in the internal side of the wall is preferable also during the hot season to avoid the overheating due to the internal gains.

and the oors so as to obtain indoor comfort throughout the year. Acknowledgement We thank Prof. G. Moriconi for his support on laboratory tests.

5. Conclusions An experimental study was made on a building constructed as part of a social housing scheme in order to assess the durability of retrotting with external insulation carried out 20 years before. The building was simulated using software in dynamic conditions. Using parametric analyses it was possible to predict the effect of various types of retrotting and to compare different structures of exterior envelope to be used when designing new buildings. The study demonstrated that the retrotting carried out 20 years before was still effective from the point of view of thermohygrometric performance (guaranteeing the heat transmission rate values and the elimination of thermal bridges) and from the point of view of mechanical performance (the system used for anchoring the insulation panels to the brickwork continues to guarantee that the insulation panels are bound to the supporting masonry). There is no evidence of chemical deterioration although the surface nishing shows the presence of cracks, peeling and aking. The cracks, which are present corresponding to the joints between the insulation panels, are due to overheating and differential temperature dilatation; the extent of the damage is greater on the south face where there are more temperature changes. The use of panels without staggered joints has contributed to the formation of the cracks. Finally, the parametric analysis allowed the identication of the correct positioning of thermal insulation on the walls

References
[1] H. Simmler, S. Brunner, Vacuum insulation panels for building application: basic properties, aging mechanism and service life, Energy and Buildings 37 (2005) 11221131. [2] G.M. Wallner, R.W. Lang, Aging of polymeric lms for transparent insulation wall applications, Solar Energy 79 (2005) 603611. [3] F. Tittarelli, G. Moriconi, A. Bonazza, Atmospheric deterioration of cement plaster in a building exposed to a urban environment, Journal of Cultural Heritage 9 (2008) 203206. ` and dur[4] V. Fassina, M. Favaro, A. Naccari, M. Pigo, Evaluation of compatibilita ability of hydraulic lime-based plaster applied on brick wall masonry of historical buildings affected by rising damp phenomena, Journal of Cultural Heritage 3 (2002) 4551. [5] E. Grinzato, P.G. Bison, S. Marinetti, Monitoring of ancient buildings by the thermal method, Journal of Cultural Heritage 3 (2002) 2129. [6] V. Vydra, Degradation risk assessment of external envelopes: a practical engineering approach, Building and Environment 42 (1) (2007) 344348. [7] J. Kosny, E. Kossecka, Multi-dimensional heat transfer through complex building envelope assemblies in hourly energy simulation programs, Energy and Buildings 34 (2002) 445454. [8] F. Miranville, H. Boyer, T. Mara, F. Garde, On the thermal behaviour of roofmounted radiant barriers under tropical and humid climate conditions: modelling and empirical validation, Energy and Buildings 35 (2003) 9971008. [9] F. Stazi, C. Di Perna, L. Del Prete, A. Stazi, Comparison between monitoring and simulating. An important step forward for model reliability, Proceedings of the10th International Building Performance Simulation Association Conference and Exibition, September 36, 2007, Tsinghua University, Beijing, China, ISBN: 09771706-3-2, pp. 14701476. [10] Y. Feng, Thermal design standards for energy efciency of residential buildings in hot summer/cold winter zones, Energy and Buildings 36 (2004) 13091312. [11] P.T. Tsingiliris, Thermal ywheel effects on the time varying conduction heat transfer through structural walls, Energy and Buildings 35 (2003) 10371047. [12] E. Kossecka, J. Kosny, Inuence of insulation conguration on heating and cooling loads in a continuously used building, Energy and Buildings 34 (2002) 321331.