1、2009 ASHRAE 111ABSTRACTNumerical models of heat, air and moisture (HAM) trans-fer in building components continue to advance, but there remains a need for accurate and well-documented experimen-tal data for model validation. In the framework of IEA/ECBCS Annex 41 a round robin experiment on the dete
2、rmination of the hygric properties of porous building materials was combined with a transient heat and moisture transfer experiment to generate a data set for benchmarking numerical models. In the round robin experiment fourteen laboratories measured the vapor permeability and sorption isotherm of c
3、oated and uncoated gypsum board. In the second part, ten different numerical models used the mean values of the measured mate-rial properties to simulate the dynamic behaviour of a bed of gypsum board as measured in a transient moisture transfer facility. The combination of round robin and dynamic e
4、xperi-ments also allowed a sensitivity study of the numerical simu-lations, in which the influence of the uncertainty in material properties and boundary conditions were investigated.INTRODUCTIONAs computing power increases and numerical models for whole building heat, air and moisture (HAM) transfe
5、r advance, there remains a general need for more experimental data that quantify HAM transport in porous building materi-als. For example, recent benchmark data for validating 1-D HAM simulation models produced in a large international project (Hagentoft et al. 2004) rely solely on numerical and ana
6、lytical data because well-documented and accurate 1-D data are scarce. An important part of the research in IEA/ECBCS Annex 41 has been on heat and moisture transfer between indoor air and hygroscopic materials during transient changes in indoor humidity because research has shown that moisture buff
7、ering may improve comfort, air quality and energy consumption in buildings (Rode et al., 2004, Holm et al, 2004, Simonson et al., 2002, 2004a, 2004b and Osanyintola and Simonson, 2004). To validate models that simulate mois-ture buffering of hygroscopic materials, new experimental data are needed th
8、at accurately quantify heat and moisture transfer between humid air and hygroscopic materials during transient changes in the air humidity. Experimental data are available in the literature, but most data are not well suited to benchmark detailed numerical models because carefully planned laboratory
9、 experiments are best suited for model vali-dation (Holm et al., 2004, Simonson et al., 2004b, Tariku and Kumaran, 2006, Svennberg et al., 2007, Kalamees and Vinha, 2006, Jenssen et al., 2002, Hens, 2004, Qin et al, 2005 and Talukdar et al., 2007a and b). In addition many of the experi-ments in the
10、literature are conducted on non-hygroscopic materials, where a majority of the moisture accumulation is due to condensation and frosting near a cold surface. Further-more, in many cases the thermal transients dominate the prob-lem. To benchmark models that intend to consider moisture buffering of hy
11、groscopic materials in contact with indoor air, experimental data are needed where the air humidity is changed in a transient manner as presented in this paper.The objective of this paper is to compare experimental and numerical data for 1-D heat and moisture transport in a bed of gypsum boards, whi
12、ch was conducted as part of Subtask 2 of IEA/ECBCS Annex 41 (Roels, 2008). Comparing the numerical and experimental data serves a dual purpose of vali-dating the numerical models as well as confirming the exper-imental data. However, as previous research (e.g. BCR 1992, Time and Uvslokk 2003, Roels
13、et al. 2004) showed that apart from the lack of good benchmark models, also attaining Reliability of Transient Heat and Moisture Modeling for Hygroscopic BufferingStaf Roels, PhD Chris James Prabal Talukdar, PhD Carey J. Simonson, PhD, PEngStudent Member ASHRAE Member ASHRAEStaf Roels is a professor
14、 in the Department of Civil Engineering, K.U. Leuven, Heverlee, Belgium. Chris James is a departmental assistant and Carey J. Simonson is a professor in the Department of Mechanical Engineering, University of Saskatchewan, Saskatoon, Saskatchewan, Canada. Prabal Talukdar is an assistant professor in
15、 the Department of Mechanical Engineering, Indian Institute of Technology, Hauz Khas, New Delhi, India.LO-09-009 2009, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2009, vol. 115, part 2. For personal use only. Add
16、itional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission.112 ASHRAE Transactionsuniformity in measuring material properties remains difficult; therefore, it was decided to perform first an interlaboratory comparison
17、 of the measurement of the hygric properties of porous materials by means of a round robin test. Not all mate-rial properties have been measured, but the round robin focussed on those properties relevant for indoor moisture buff-ering: the water vapor transmission properties and the sorption isother
18、m. In total fourteen laboratories participated in the round robin test. The results of the round robin test are presented first. Then, the experimental facility is explained and the different test cases and applied numerical models are presented. Finally, a comparison is made between the experi-ment
19、al and numerical results. The large number of partici-pants in the round robin tests, also allowed a sensitivity study of the numerical simulations, in which the influence of the uncertainty in material properties and boundary conditions were investigated.INTERLABORATORY COMPARISON OF MATERIAL DATAT
20、he main goal of the round robin testing was twofold: on the one hand to assess the reliability of material property measurements relevant for Annex 41, on the other hand to collect reliable data that could be used in the benchmark case. The first goal was partly inspired by the European HAMS-TAD-pro
21、ject which showed that even commonly used measurement methods could not always determine the mate-rial properties with an acceptable level of precision and repeat-ability (Roels et al. 2004). BCR (1992) reported a similar observation when describing the results of a round robin test in which 13 labo
22、ratories performed water vapor permeability tests on two different materials. Better results were obtained in (Time and Uvslokk, 2003), a project in which six Nordic coun-tries performed cup tests on three different materials. In the framework of Annex 41, to gain time, no real round robin was perfo
23、rmed, but samples of the same batch were randomly distributed among the participating laboratories. A finishing material (gypsum board) was selected as the test material. Though, the material was measured both uncoated and finished with a priming coat and two kind of finishing coat. The gypsum board
24、 and primer have been supplied by GYPROC BPB, Belgium. The gypsum board is from the type GYPROC A ABA-board with overall dimensions of 1200x2400 mm and a thickness of 12.5 mm. For the finishing coats two types, labelled A and L, have been used. Both were supplied by BOSS Paints, nv Bossuyt, Belgium.
25、 Finishing coat A is an acrylic paint (type Decomat, BOSS paints) and was coloured light blue. Finishing coat L is a latex paint (Bolatex, BOSS paints) and was coloured light yellow.Three test series have been measured: 1) the uncoated gypsum board; 2) the gypsum board covered with the priming coat
26、and acrylic finishing coat A; 3) the gypsum board covered with the priming coat and latex coat L. These three series allow to asses the reliability to measure the vapor permeability of a layer of paint. In total 20 plates of gypsum board (2400x1200x12.5 mm) have been cut in different parts of 500x30
27、0 mm. All parts were labelled with a letter (G for uncoated board, A for board covered with the acrylic finishing coat, L for board covered with the latex finishing coat), followed by the number of the board and the number of the part. To randomise the test samples as much as possible parts form dif
28、ferent boards were sent to each partner. Each partner had to cut specimens out of these parts in a way that for each measurement set-up specimens of different boards are measured. For series 2 and 3 the priming coat and finishing coat were applied with a paint roller by the University of Leuven and
29、the painted gypsum board samples were distrib-uted among all participants.Water Vapor Transmission TestThe water vapor transmission properties have been deter-mined in accordance to EN ISO12572:2001 Hygrothermal performance of building materials and products determina-tion of water vapor transmissio
30、n properties. A complete description of the test method and procedure can be found in (EN ISO 12572:2001). All three main test series (the uncoated gypsum board, the gypsum board covered with primer and paint A and the gypsum board covered with primer and paint L) had to be measured at three test co
31、nditions as given in Table 1. Test condition C1 and C2 correspond respectively to set A and set C in chapter 7.1. of (EN ISO 12572:2001). Test condition C3 was included to apply an analytic fit to the measured data, which could be used in numerical simulations. If CaCl2-flocks were used as desiccant
32、 in test condition C1, it was asked to make sure to start from fully dried (at 200-250C Table 1. Test Conditions for the Determination of the Water Vapor Transmission PropertiesSet Conditions Temperature Relative HumidityC RH Climatic chamber CupC1 23 0/5023 0.5 C (73.4 0.9F)Aqueous solution 53%Mg(N
33、O3)2DesiccantCaCl2-flocks or Mg(ClO4)2C2 23 50/9323 0.5 C(73.4 0.9F)Aqueous solution 53%Mg(NO3)2Aqueous solution 94%KNO3C3 23 86/9323 0.5 C(73.4 0.9F)Aqueous solution 85%KClAqueous solution 94%KNO3ASHRAE Transactions 113392-482F) flocks. Specific prescriptions for the preparation and more details on
34、 the applied test assemblies can be found in (Roels, 2008).The results of the water vapor transmission tests are presented as equivalent air layer thickness (sd-values in m). Figure 1 shows the probability density functions of the mean values as measured by the different participating laboratories,
35、when no corrections for the air layers or for the masked edge are performed. The uncoated gypsum board is found to be very vapor open. In fact, only the dry cup value (test condition C1) exceeds an equivalent air layer thickness of 0.1 m, the criterion used in the standard as minimum value to make c
36、up tests appli-cable due to the increasing uncertainty on the measurement results for more vapor open materials. Applying a primer and finishing coat on the gypsum board changes the vapor perme-ability considerably, but the type of coat seems to be very important. Applying an acryl finishing coat tr
37、iples the equiv-alent air layer thickness (mean value of all laboratories for test condition C1), while the latex coat multiplies the equivalent air layer thickness with almost a factor 30! But, the dependency Figure 1 Cumulative distribution function of the mean values of the equivalent air layer t
38、hickness as determined by the participating laboratories for (a) uncoated, (b) acrylic coated and (c) latex coated gypsum board.114 ASHRAE Transactionsof the vapor permeability on relative humidity is far more pronounced for the coated gypsum board than for the uncoated gypsum board. Put otherwise,
39、the vapor resistance of the finishing coat strongly decreases with increasing relative humidity. This feature is far more pronounced for the latex coat than for the acrylic coat. Compared to the uncoated gypsum board where the wet cup value of the equivalent air layer thickness (test condition C3) i
40、s 70% of the dry cup value, a reduction with almost a factor three is found for the acryl coat and a reduction with a factor seven for the latex coat. Comparing the results of the different laboratories, the data show remarkable high differences. Knowing that the measurements are carried out accordi
41、ng to the prescriptions of an existing standard, unacceptable deviations are found. Furthermore, the more vapor tight the specimen is (dry cup values of gypsum board covered with latex coat), the more pronounced the differences become. However, when compar-ing the individual results of each of the l
42、aboratories the devi-ations are much smaller than the deviations between the laboratories. Note that these data are not included here, but can be found in (Roels, 2008). This suggests that there could be systematic differences between the participating laboratories. However, when plotting the data o
43、f the three test series versus one another, no indication for systematic differences between the participating laboratories could be found.The results presented in Figure 1 correspond to uncor-rected data. Depending on the test set-up assembly two types of corrections have to be applied. According t
44、o Annex G of (EN ISO 12572:2001) a correction for the vapor resistance of the layer of air in and above the cup has to be applied for very permeable materials. Very permeable materials are defined as materials with an equivalent air layer thickness less than 0.2 m. This is the case for the uncoated
45、gypsum board and for the gypsum board with acrylic finishing coat under test condition C2 and C3. (note that for uncoated gypsum board only test condition C1 is within the scope of the standard). For the air layer in the test cup the measured permeance has to be corrected with the permeance of the s
46、tagnant air layer in the cup. Since most of the laboratories followed the prescriptions very well, which stated that the air space between the desiccant or aqueous solution and the specimen had to be within the range of 155 mm, all correction are in the same order of magnitude. In this way applying
47、a correction for the air layer in the cup just reduces the calculated equivalent air layer thick-ness (mean value after correction 0.107 m instead of 0.128 m), but has hardly any difference on the spread in results. To ensure that the resistance of the air layer above the cup is negli-gible, EN ISO1
48、2572:2001 requests that the air in the test chamber be well mixed and to ensure an air velocity above each specimen of at least 2 m/s. Furthermore it is advised to use cups without a high rim. For most laboratories the air velocity in the test chamber is much lower than 2 m/s (Roels, 2008). Moreover
49、, different types of cups have been used, some with a rim and others without a rim. However, no correlation could be found between the measured values and the air veloc-ity above the cups or the presence (and height) of the rim.Some participants use cup types with masked edges, which means that the edge of the specimen overlaps the edge of the cup. This overlap zone is a possible route for two dimen-sional vapor transfer, leading to an overestimate of the vapor permeance, since the total flow through the specimen in the test assembly is greater tha
