Experimental and Numerical Modeling of Heat Transfer in Wall Assemblies
It is critical for the construction industry to ensure that new building designs and materials, including wall and floor assemblies, provide an acceptable level of fire safety. A key fire safety requirement that is specified in building codes is the minimum fire resistance rating. A manufacturer of building materials (e.g., insulation or drywall) is currently required to perform full-scale fire furnace tests in order to determine the fire resistance ratings of assemblies that use their products. Due to the cost of these tests, and the limited number of test facilities, it can be difficult to properly assess the impact of changes to individual components on the overall fire performance of an assembly during the design process. It would be advantageous to be able to use small-scale fire tests for this purpose, as these tests are relatively inexpensive to perform. One challenge in using results of small-scale fire tests to predict full-scale fire performance is the difficulty in truly representing a larger product or assembly using a small-scale test specimen. Another challenge is the lack of established methods of scaling fire test results. Cone calorimeter tests were used to measure heat transfer through small-scale specimens that are representative of generic wall assemblies for which fire resistance ratings are given in the National Building Code of Canada. Test specimens had a surface area of 111.1 mm (4.375 in.) by 111.1 mm (4.375 in.), and consisted of single or double layers of gypsum board, stone wool insulation and spruce-pine-fir (SPR) studs. As the specimens were designed to represent a one-quarter scale model of a common wall design, with studs spaced at a centre-to-centre distance of 406.4 mm (16 in.), the wood studs were made by cutting nominal 2x4 studs (38 mm by 89 mm) into 9.25 mm by 89 mm (0.375 in. by 3.5 in.) pieces. The scaled studs were then spaced at a centre-to-centre distance of 101.6 mm (4 in.). Three types of gypsum board were tested: 12.7 mm (0.5 in.) regular and lightweight gypsum board, and 15.9 mm (0.625 in.) type X gypsum board. Temperature measurements were made at various points within the specimens during 70 min exposures to an incident heat flux of 35, 50 and 75 kW/m2 using 24 AWG Type K thermocouples and an infrared thermometer. Temperature measurements made during cone calorimeter tests were compared with temperature measurements made during fire resistance tests of the same generic assemblies and the result show a very good agreement for the first 25 min of testing at the unexposed side. A one-dimensional conduction heat transfer model was developed using the finite difference method in order to predict temperatures within the small-scale wall assemblies during the cone calorimeter tests. Constant and temperature-dependent thermal properties were used in the model, in order to study the effects of changes to materials and thermal properties on fire performance. A comparison of predicted and measured temperatures during the cone calorimeter tests of the generic wall assemblies is presented in this thesis. The model had varying degrees of success in predicting temperature profiles obtained in the cone calorimeter tests. Predicted and measured times for temperatures to reach 100C and 250C on the unexposed side of the gypsum board layer closest to the cone heater were generally within 10%. There was less agreement between predicted and measured times to reach 600C at this location, and the temperature increase on the unexposed side of the test specimen. The model did not do a good job in predicting temperatures in the insulated double layer walls. Sensitivity studies show that the thermal conductivity of the gypsum board has the most significant impact on the predicted temperature.
DegreeMaster of Science (M.Sc.)
SupervisorTorvi, David A.
CommitteeOdeshi, Akindele G.; Simonson, Carey J.; Sparling, Bruce F.
Copyright DateApril 2014
Numerical Heat Transfer