MODELING OF THERMAL CONDUCTIVITY OF BIO-BASED BUILDING COMPOSITES

The paper presents a numerical method of calculating the effective thermal conductivity of bio-based building composites. It is developed applying the microscale approach in which the heat conduction equation is solved in the three-dimensional control domain, which accounts for detailed real complex composite morphology. The microstructures of the samples were obtained by using X-ray micro-computed tomography. Using the averaging technique, the macroscale (effective) composite thermal properties are obtained from the micro-scale temperature distributions. The method is tested and validated by calculating effective thermal conductivities of ecological building composites made of hemp shives and a magnesium binder. The samples of three compositions, i.e., of different bio-fibers to binder ratios and composite densities, and therefore significantly different thermal conductivities, were considered. In the first step, computational domains were generated from micro-computed tomography scans, and their sizes were found, i.e., the representative elementary volumes for each composite were estimated. Then in the second phase, the model parameters were tuned using the thermal conductivity measurement for one of the considered composites. The model predictions were validated in the final step by comparing them with thermal conductivity measurements for all composites. It was found that the model predicted the effective thermal conductivities with good accuracy, i.e., the relative errors between predictions and measurements were up to 11%.