Thermal Performance of Lightweight Earth: From Prediction to Optimization through Multiscale Modeling

Abstract

This study investigates the prediction of the thermal conductivity of lightweight earth and raw earth blocks incorporating plant aggregates. Given the high variability of raw materials, it is not currently possible to predict the thermal performance of this type of material before sample production. This is a major obstacle to using these eco-materials, although their use is widely encouraged to improve building performance under evolving regulatory frameworks such as The French RE2020 standard. The incorporation of plant aggregates into earth-based materials offers improved insulation properties without compromising their mechanical integrity, positioning them as promising sustainable alternatives. Mean-field homogenization techniques, including the Mori-Tanaka as well as double inclusion models, are used to develop predictive tools for thermal behavior, using rigorously selected experimental data. The selected methods are particularly relevant. The Mori-Tanaka model appears to be better suited when the proportion of aggregates is limited, whereas the double inclusion scheme proves its worth when a higher proportion of aggregates is incorporated. This study emphasizes the influence of aggregate types and processing methods on thermal conductivity, highlighting the need for precise formulation and processing techniques to optimize performance. This paper demonstrates the relevance of the applied homogenization techniques applied. It enables the real morphology of the materials studied, such as aggregate shape and intrinsic cracking, to be taken into account. It contributes to the advancement of eco-material modeling toward predictive digital twins, with the goal of simulating and optimizing complex material behavior under various environmental conditions.