Abstract
In building simulation codes, conduction transfer functions or finite difference methods are generally implemented to model the heat transmitted through the opaque components. Thermal conductivity is conventionally assumed constantly equal to its nominal value, even if it depends on both temperature and moisture content, especially for porous building materials, like timber. Therefore, the adoption of constant nominal values can bring inaccuracies on the calculated heat flux, affecting the predicted building energy performance. In this research, the magnitude of such inaccuracies is quantified for a spruce cross-lam (i.e. XLAM) wall, by comparing hourly specific heat fluxes calculated with variable thermal conductivity and those with nominal one. To achieve such a goal, a MATLAB® 1D heat and mass transfer model was developed according to the finite difference method. Material´s thermal conductivity was characterised as a function of moisture content and temperature. Finally, the developed model was calibrated and validated against experimental data collected in the Building Physics Laboratory of the Free University of Bozen-Bolzano. Annual simulations were run considering 110 Italian climates and different orientations and inclinations of the wall. Constant internal boundary conditions and hourly variable external conditions were adopted; in particular, external solicitations were expressed through the sol-air temperature definition.