Multiscale modelling of thermal conductivity in fumed silica VIPs for building energy conservation

Vacuum insulation panels (VIPs) with ultra-low thermal conductivity are widely studied for building applications to enhance energy efficiency, with fumed silica (FS) being one of the most effective core materials. However, existing analytical models are inadequate for FS due to its multiscale structure. This study develops a highly adaptable predictive model for analyzing and predicting the thermal conductivity of FS VIP cores. A three-dimensional physical model is first established to enable adaptable solid heat transfer analysis based on characterization data, such as particle diameter, porosity, and coincidence angle. A novel classification of coexisting micro- and nano-porosities is introduced to model gaseous heat transfer accurately, achieving high predictive accuracy at lower pressures. Validation results show that the predictive model performs well across different pressures. The predicted values closely match experimental data, even at near-atmospheric pressure. In the pressure range where gaseous conduction is significant (p > 1 kPa), the model achieves an average deviation of 3.7 % and a maximum deviation of 15 %. Overall, the proposed model reliably predicts thermal conductivity, offering valuable insights for designing, manufacturing, and maintaining FS cored VIPs.