COMPARISON OF SAMPLING METHODS FOR AIR TIGHTNESS MEASURMENTS IN NEW FRENCH RESIDENTIAL BUILDINGS B. Moujalled 1 ; F. Richieri 1 ; R. Carrié 2 ; A. Litvak 3. 1: CETE du Sud-Ouest, Rue Pierre Ramond, CS 60013, 33166 Saint-Médard-en-Jalles Cedex 2: INIVE, Lozenberg 7, BE-1932 Sint-Stevens-Woluwe 3: CDPEA, 58 rue Jean Duvert, Ecoparc, Le Fiducia, 33295 Blanquefort Cedex andres.litvak@cdpea.fr ABSTRACT With the increasing need for higher energy efficiency in buildings, thermal regulations are evolving toward more stringent rules in buildings at the European and national level. In France, new constructed buildings will have to comply with mandatory regulations that promote low-energy buildings from 2012 and positive-energy at the horizon of 2020. In this context, air tightness becomes a major performance issue in buildings. In fact, the part of heat loss due to air leakage in well insulated buildings becomes more important. Besides, air leakage can lead to cold draught and construction pathologies. The French thermal regulation sets a compliance scheme to guaranty good air tightness in new constructed buildings based on the evaluation of the overall building air tightness, through field measurements. However, there are often practical, technical and economical limitations to measure the global air tightness in large or multi-family buildings. In this case, building air tightness is evaluated through partial measurements in few apartments of the building using a sampling method. This method does not account for the leakages in the untested zones, including other apartments, halls, stairwells, etc. Actually, a research project (MININFIL) associating national institutions, research laboratories, resource centres and building professionals is conducted in France in order to enhance the knowledge on the air tightness and its impact on the energy performance in buildings, and to support building professionals to achieve much better envelope air tightness. MININFIL includes a campaign of air tightness measurements in ten residential buildings in order to identify the sampling methods that characterize the air tightness of a whole building through limited set of measurements in apartments, hall, stairway, etc. This paper presents a bibliographical review of the sampling methods followed by the measurement campaign that was carried out in ten new buildings. In each building, the air tightness was measured on the whole building and in each apartment of the building. The results have served to identify the typical air leakage paths, and the appropriate sampling method for the air tightness evaluation in large residential buildings. INTRODUCTION As recommended by the EPBD, the new French thermal regulation imposes that new constructed buildings must be low energy buildings from 2012. Different studies have shown that the impact of the infiltration losses becomes more important on the energy efficiency in the case of high performance buildings [1, 2]. Therefore, air tightness requirements have been included in the French regulation with an obligation for air tightness measurements in residential buildings, especially in multi-family buildings.
The European standard EN ISO 13829 [3] describes the measurement method of air permeability of buildings. This measurement is planned to be done on the whole building. In the case of multi-family buildings, there are often practical limitations to measure the air permeability of the whole building (main reasons: building is too large, floors are not connected with an internal airflow path, or very leaky elements in the stairway). For these buildings, the standard allows separate measurements on individual parts of the building; e.g. apartments of multi-family buildings can be measured individually. In many European countries, the measurement is done on a sample of apartments for practical and technical reasons. A sampling method is given to specify the sample size, the selection criteria, and the airtightness requirements for the sample. Walther [2] has presented a review of the different sampling methods in Europe. Table 1 gives the German, UK, and French methods. The French sampling method has been included in the EN ISO 13829 implementation guide [4]. Sampling Method Germany UK France (GA P50-824) Sample size At least 20% of the total number of apartments At least 20 % of the buildings envelope area = 3 apartments if the building has 30 units or less = 6 apartments otherwise Selection criteria Apartments located at the top, in-between and ground floor. Areas representative of the whole envelope construction. 1. Apartments with the largest length of floors and windows. AND 2. Apartments located at the top, in-between and ground floor. Extrapolation method for the whole building 1. The weighted average based on the volume must be lower than the limit value. 2. Air permeability of tested apartments can be 30% greater than the limit value. 1. Air permeability of tested apartments must be 10% smaller than the limit value. 2. No requirements for the whole building. 1. The weighted average based on the envelope area must be lower than the limit value. 2. No requirements for the tested apartments. Table 1: The sampling methods for the air tightness measurements in multi-family buildings. In France, the research project MININFIL has been conducted since 2008 in order to enhance the knowledge of professionals on the air tightness and its impact on the energy performance in buildings. Under the task 3 of the project, an extensive campaign of airtightness field measurements has been carried out in ten new multi-family buildings. In each building, the air tightness was measured for all the apartments and for the whole building. The results have enabled the evaluation of the sampling methods for the measurement of airtightness in mutifamily buildings. This paper presents the results of this work. METHOD The air permeability measurements have been realised with the fan pressurisation method according to the standard EN ISO 13829 [3] using Minneapolis Blower Door systems. The aim of the measurements is to identify separately the air permeability of each apartment, the common areas, and the whole building. Therefore, three types of air permeability measurements have been carried out in each building:
1. Individual measurements of each apartment of the building with a blower door positioned on the entrance door of the apartment. The entrance door of the building is fully opened. 2. A global measurement of the whole building, including leakages in apartments and common areas. This measurement is realised with a blower door using a single or double fans. The blower door is positioned on the entrance door of the building. The doors between apartments and common areas are fully opened 3. A measurement of the common areas. This measurement is similar to the previous, but this time the doors between apartments and common areas are closed, in order to eliminate the leakage in apartments from the measurement. This requires that the doors are air-tight. Table 2 presents the multi-family buildings characteristics. The number of apartments per building varies between 12 and 38, and the number of the levels between 2 and 7. The volume of buildings varies between 2365 m 3 and 5704 m 3, with 2 buildings larger than 4000 m 3. Building code B01 B02 B03* B04 B05* B06 B07 B08 B09 B10 # of levels 4 3 2 4 3 5 5 4 7 5 # of flats 17 12 17 20 16 17 17 16 38 36 Area (m²) 1325 956 1266 1486 1150 1455 1248 1375 2256 2246 Volume (m 3 ) 3280 2365 3150 3700 2893 3544 3446 3031 5704 5175 Table 2: The description of the assessed buildings. The measurements have been done at the end of the building construction. All apartments were unoccupied, in order to facilitate the access to all the parts of the building. However, the global measurements of buildings B03* and B05* have been disturbed by the presence of workmen during the tests. The global measurements for these building will not be considered in the analysis. RESULTS In France, the air permeability (Q 4Pa-Surf ) is calculated as the ratio between the infiltration airflow rate at 4 Pa with the envelope area of the building except the floors area (A T bat ). The new thermal regulation sets the limit value required of air permeability to 1.0 m 3 /h/m 2 for the case of multi-family buildings. This value is based on the French low-energy building standard. In this paper, the results will be presented with the French air permeability indicator. The ten buildings represent a total of 208 apartments. For the individual measurements, more than half of the apartments (52%) show lower results than the limit value 1.0 m 3 /h/m 2. For the global measurement of the whole buildings, only three buildings (over eight) are lower than the limit value. The major part of the leakage in the apartments (40%) has occurred across the fenestration (joints at window sash, window sill, and shutter box), while 30% of the leakage occurs at the joints of hatch and ducts, and 25% across the electricity plugging. The leakage across the joints between walls and slabs are negligible. Figure 1 presents the results of the individual and global measurements for each building. We can observe that the individual measurements of air permeability are very heterogeneous between buildings, and between apartments in the same building in some cases. The buildings can be classified into two categories: Buildings B05*, B06, B07 and B08 having the global measurement and the individual measurements globally below the required limit value (1.0 m 3 /h/m 2 ). For these buildings, the individual measurements are uniform and vary in a narrow range.
For the other buildings, both the global measurement and the median of the individual measurements are greater than the limit value. The individual measurements in each building are very heterogeneous and vary in a wide range. In B09, the upper value of the individual measurements is almost ten times greater than the lower value. Figure 1: Box plot of the measured air permeability values in each building: the box lines indicate the statistic results of individual measurements and the red marks indicate the global measurement of the whole building in each case. The global measurements of B03* and B05*have been excluded. In order to evaluate the sampling methods, the GA P50-784 method has been tested and compared to the results of the measurements. Table 3 presents a comparison of the GA P50-784 method with other methods in terms of the sample size criterion. Only one building has a sample with more than 20% of the total number of apartments as required by the German method [2]. Whereas, 7 buildings have samples with more than 20% of the building envelope as required by the UK method. This can be explained by the fact that the selection criterion of the GA P50-784 method privileges the apartments with the largest envelope areas. Building code B01 B02 B03* B04 B05* B06 B07 B08 B09 B10 # of flats 17 12 17 20 16 17 17 16 38 36 # of tested flats 3 3 3 3 3 3 3 3 6 6 % of tested flats 18% 25% 18% 15% 19% 18% 18% 19% 16% 17% % of tested A T bat 20% 22% 22% 20% 29% 21% 18% 24% 19% 16% Table 3: Comparison of the GA P50-784 sampling method against the other methods. The left panel of Figure 2 presents a comparison between the weighted average air permeability of the sample of apartments and the weighted average of all the apartments for each building. For the buildings with uniform individual measurements lower than the limit value (B05*, B06, B07, and B08), the results of the samples are very close to those obtained with all the apartments. For the other buildings, the difference is more significant. This figure gives also the statistical values of the weighted average air permeability that can be obtained with the different possible random samples of apartments for each building. The
random samples are selected according to the required size of the sampling method with the apartments being at the top, in-between and ground floors. The median values are globally close to the weighted average of all the apartments. For buildings B01, B02, B04, B09 and B10, the difference between the minimal and maximal values that can be obtained with a random sample is very important. The minimal values can be as low as the half of all the apartments in the case of B10. The sample size and the selection criterion are critical for the sampling method. Figure 2: Comparison of the weighted average air permeability of the samples of apartments against the weighted average air permeability of all apartments on the left panel, and against the global measurement of whole building on the right panel. The GA P50-784 method evaluates the air permeability of the whole building through the weighted average of the sample of apartments. Besides it doesn t impose any requirement on the individual measurements. The use of a selection criterion is important to guarantee that the air permeability of the sample is at least equivalent to all the apartments. The GA P50-784 method uses the ratio of the length of floor and windows per unit of floor area as a selection criterion. The method selects the apartments with the largest value of this ratio, as they are considered to be potentially the leakiest apartments. Figure 3 shows on the right panel the variation of the air leakage rates at 4 Pa against this ratio. As we can see, there is no correlation between these two parameters, and besides the air leakage rate shows a decreasing tendency with this ratio. The correlation is more significant with the envelope area as we can see on the right panel of figure 3. The envelope area seems to bee more relevant as a selection criterion than the ratio of the length of floor and windows per unit of floor area. Now we will compare the results of individual measurements against the global measurement. The right panel of Figure 2 presents the comparison of the weighted average air permeability of the samples of apartments (the sample of GA P50-784 method, and the sample of all the apartments) against the global measurement of the whole building. For both samples, the weighted average air permeability of apartments is always lower than the air permeability of the whole building as it doesn t account for the leakage in the common areas caused by the lift shaft, the parking basement and other shafts and hatches. The greatest difference was found in the case of buildings B08, B09 and B10 with lift shaft and basement parking in the common areas.
Figure 3: The variation of the measured air leakage rates at 4 Pa as a function of the sampling criteria (the GA P50-784 sampling criteria on the left panel and the envelope area excluding floor on the right panel). CONCLUSION A detailed campaign of air permeability measurements has been carried out in ten multifamily buildings in France. For each building, the air permeability of individual apartments and the whole building have been measured, and the sampling method of the implementation guide GA P50-784 has been evaluated. The results showed that the sampling method gives good results only in the case of buildings with uniform individual measurements. The method is as good as a random sampling. Nevertheless, in the case of buildings with heterogeneous individual measurements, the result of a random sampling varies in a very wide range. The selection criterion of the GA P50-784 sampling method doesn t identify the apartments with greater risk of leakage. The use of another criterion based on the envelope area can be more relevant. The results have shown that the leakage in the common areas are significant and can have an important impact on the air permeability of the whole building in the case of common areas with lift shaft and basement parking. These leakages should be considered in the measurement method if it extrapolates the individual measurements to the whole buildings. ACKNOWLEDGEMENTS This work has been conducted under the task 3 of the research project MININFIL that is financed by the French Mininstry of Housing and the French Agency for the Energy Efficiency and the Environment ADEME. REFERENCES 1. Erhorn-Kluttig, H, Erhorn, H, Lahmidi, H: Airthightness requirements for high performance building envelopes. ASIEPI information paper #157, 2009. 2. Walther, W, Rosenthal, B: Airtightness testing of large multi-family buildings in an energy performance regulation context. ASIEPI information paper #165, 2009. 3. NF EN 13829 : February 2001 - Thermal performance of buildings - Determination of air permeability of buildings - Fan pressurization method. 4. GA P50-784 : February 2010 Thermal performance of Buildings Implementation guide for NF EN 13829 : 2001.