Hygrothermal performance of building envelopes: uses for 2D and 1D simulation

A building’s durability depends on controlling heat and moisture within its envelope. Moisture diffuses through porous materials that may suffer mould growth and decay (if wood-based) when left moist and warm for too long. Designers try to keep vulnerable components dry, but materials can start to deteriorate before reaching the dew point temperature. Researchers use two-dimensional hygrothermal modeling to calculate time-varying moisture content and temperature at points on a plane through the building envelope thickness. One-dimensional versions of several research programs have been written to assist designers and other building envelope specialists in their work. This paper compares moisture and temperature histories in two building envelopes exposed to a variety of climatic conditions over three years, as calculated by 2D and 1D versions of one such computer program. The 2D calculations come from reports on a methodology for moisture management of wood-frame walls, published in 2003 by a consortium of industry and research partners. Results for a face-sealed stucco wall with rain entry by diffusion only (no seal deficiencies) showed reasonable agreement between 2D and 1D, whereas those for a brick wall with a ventilated air space diverged considerably. With due respect for limitations, 1D simulation can give a first indication of the differences in performance of a wall exposed to different climates, or between different wall assemblies. In some cases the user should consider following up with 2D simulation or field monitoring. INTRODUCTION Acceptable hygrothermal performance of building envelopes cannot always be taken for granted. In the mid 1990's, residential designs and construction techniques with adequate track records in relatively dry climates began to fail prematurely in great numbers in wet coastal climates. Investigators reported that rain leaks at through-wall penetrations caused 90% of the damage found in their study of 37 condos on the lower mainland of coastal BC. Three rain management strategies (face seal, weather barrier with drainage plane, rain screen, each of which exhibited failures) had been used, in a climate with 5 times the rain, 4 times the wind-driven rain index, and only 1/3 the hours of sunshine of most inland communities in Canada. [1] Designers, building code officials and inspectors, developers, contractors and materials suppliers, even research institutions and lending agencies, all came under fire over the heavy losses suffered by owners and tenants in the Vancouver area. Mould and wood rot were common features of the damage resulting from extended periods of wetness inside walls. Designers sometimes use steady-state calculations to predict dew point temperatures at critical points through the thickness of the wall, on the assumption that condensation would signal the onset of mould or wood rot. It now appears that such organisms will germinate and grow at relative humidities (RH) over 80 percent for temperatures (T) over 5 °C. [2] To predict time-varying RH and T throughout a wall section, researchers use hygrothermal modeling programs that require input files of material properties, hourly weather records, and considerable expertise to operate. [3, 4, 5] In many investigations two-dimensional (2D) grids are required as well. “User-friendly” 1D versions of research programs are now available for practitioners to become familiar with the hygrothermal performance of building envelopes. [6, 7] Material properties and weather files are provided, which simplifies the preparation of a simulation run. The simulation takes only a fraction of the time needed for a 2D run, and output quantities such as RH and T can be displayed graphically for easy comparisons of performance between different wall assemblies, or for different climates. 1 Corresponding author, consultant, 237 Valley Ridge Green NW, Calgary AB T3B 5L6, phone 403-286-2177 alandalg@shaw.ca 2Research Officers, Institute for Research in Construction, National Research Council of Canada, 1200 Montreal Road, Ottawa, ON, K1A 0R6 32 10th Canadian Conference on Building Science and Technology Ottawa, May 2005 A practitioner will find such a program easier to use than a full-blown research program (1D or 2D), but still has to acquire a certain level of expertise before applying the results to practical problems, and must learn which problems can be handled by a 1D simulation. In 2003, a research consortium published a study of moisture management for building envelopes in a range of North American climates. [8] The authors of this paper selected some of the consortium’s 2D simulations and redid them, using a 1D version of the same program. Our purpose was to find out the extent to which 1D simulation might be considered useful for generating insights consistent with the original findings, but in addition, we would like to comment on some of the uncertainties facing users of both 1D and 2D programs. MOISTURE MANAGEMENT – FINDINGS OF A RESEARCH CONSORTIUM: The consortium worked out a method for predicting the moisture management performance of walls as a function of climate severity, in which hygrothermal modeling played a central role. A 2D program was used to solve the equations of conservation of mass and energy governing the flow of heat, air and moisture, to track temperature and moisture content in four generic building envelope assemblies. Two of the four wall assemblies are shown in Figure 1. The consortium's final report [9] gives designers and other non-researchers insight on what moisture-handling performance to expect from a few different wall designs exposed to a range of North American climates, but cautions readers to consult the research team before applying the information to building designs. Interested readers should consult the report for a full description; the purpose here is limited to identifying those 2D simulation cases suitable for comparison with 1D simulation. Modeling rain leakage – not feasible for 1D simulation: The findings of the B.C. condo study made leakage paths a major factor in the research project. Full-scale walls with deliberately created paths were exposed (in lab tests) to water sprays and air pressure differences simulating driving rain. Water entering the stud space was expressed as a function of spray rate (amount of rain hitting the wall/hour) and air pressure (average wind pressure during the hour). The fitted relations found for each wall and deficiency type were then used to calculate quantities of water that were injected, as point sources, into the stud spaces for most of the hundreds of 2D simulations. Moisture from point sources can diffuse up or down in a 2D simulation as well as horizontally, which is the only option for 1D. With the possible exception of a vertical array of uniform point sources, the authors believe that simulation of rain leakage is not feasible for 1D. This is the first major difference in capability between 2D and 1D programs. FIGURE 1 Arrangement of material layers for 2D and 1D wall types: stucco (left) and brick (right), not to scale 10th Canadian Conference on Building Science and Technology Ottawa, May 2005 33 SELECTION OF FOURTEEN 2D CASES FOR COMPARISON TO 1D: Several different configurations of material properties were simulated for each of the four wall systems. One set of material properties was selected as the reference wall, and with no injections of leakage moisture, functioned as a base case. In the base cases, liquid and vapour entered by diffusion processes only, making them candidates for 1D simulation. Even without leakage paths, rain striking the wall played a major role in the ingress of moisture for the base cases. Of the four generic wall assemblies, the most suitable was the stucco wall, for two reasons: 1) it contained no ventilated air space, and 2) in the early stages of the parametric study, histories of moisture content and temperature averaged over vertical grid points for the bottom 600 mm of the interior face of the OSB sheathing were available. Unfortunately for our investigation, however, the region of interest was later switched to the grid points on 50 mm of the top of the bottom plate, which is spruce and not OSB. The brick wall (results available for the spruce plate only) was selected to show the greatest differences between 2D and 1D, thus rounding out the study to cover 14 base case simulations: two wall assemblies exposed to seven climates. Weather records and material properties: 1D hygIRC includes a database containing weather files, with the same records used in the 2D simulations for the consortium (see Table 1). In 2D, a preliminary year of conditioning with the so-called “wet” year was followed by a repetition of the wet year and the average year. The latter two years were used to examine RH and T, and those 2D results will be compared to 1D simulations using the same combinations of years of weather record. TABLE 1 Weather Records for 2D and 1D Simulations City wet year average year City wet year average year Phoenix AZ 1978, East 1979, East Winnipeg MB 1968, North 1978, North Fresno CA 1982, East 1980, East Ottawa ON 1969, East 1959, East San Diego CA 1983, South 1982, South Seattle WA 1990, South 1984, South Wilmington NC 1984, North 1988, North East The algorithm recommended by Straube [10] is used to convert hourly rainfall on the ground (Rh) to rain hitting the wall (Rv). For current developments in this area, see Blocken and Carmeliet’s review of wind-driven research and its implications for hygrothermal modeling [11] : Rv = RAF•DRF•cos( )•Vz • Rh, where Rain Admittance Factor RAF = 0.4 for the mid-height of a low building wall (~2m); Driving Rain Factor DRF = 1/Vt, the terminal velocity of raindrops, a function of their diameter; is the angle between the normal to the wall and the wind direction; Vz = V•(1.8/10), in which V is the weather file wind speed at a height of 10 m in open terrain. The