In-depth temperature profiles in pyrolyzing wood

The move towards performance-based design of the fire resistance of structures requires more accurate design methods. An important variable in the fire performance of timber structures is the in-depth temperature distribution, as wood is weakened by an increase of temperature, caused by exposure to high heat fluxes. New construction techniques of timber structures use new types of metallic connectors that have poor fire performance, and present evidence of low temperature failure. The temperature distribution is also an important variable in the performance of these connections. Thus, a proper prediction of temperature profiles in wood structural elements has become an essential part of timber structural design. Current design methods use empirically determined equations for the temperature distribution of the uncharred section of the wood member, but these assume constant charring rates (i.e. steady-state conditions), do not account for changes in the heating conditions, and were obtained under poorly defined boundary conditions in fire resistance furnaces. Another approach consists of calculating the temperature profiles by modelling wood pyrolysis. The energy conservation equation for the timber element is solved numerically, and several models have been created in the past. However, there has been no clear indication whether each term included in the different models adds enough accuracy to justify the increased computational cost to solve a more complicated model. Finally, comparisons of predicted and measured results show good initial agreement, but greater inaccuracy as the pyrolysis process progresses and the temperatures rise. As part of this research project, a series of experimental in-depth temperature measurements were done in wood samples exposed to various intensities of radiant heat fluxes, with clearly defined boundary conditions that allow a proper input for pyrolysis models. The imposed heat fluxes range from 10 kWċm−2, which generates an almost inert behaviour, to 60 kWċm−2, where spontaneous flaming is almost immediately observed. Mass loss measurements for all