A bottom-up hierarchical model for MICP-treated soil based on DEM

Microbially induced calcite precipitation (MICP) is a promising soil improvement technique. This study presents a novel discrete element model (DEM) for the MICP-treated soil, featuring a bottom-up hierarchical approach encompassing three different levels. At level I (the CaCO₃-sand system), the model incorporates realistic representations of CaCO₃ crystals on a soil particle surface, capturing their size, quantity, and distribution. At level II (representative sand pair with CaCO₃ cementation), the model investigates the influence of interparticle gaps, soil particle sizes, and CaCO₃ bond strength and mass fraction on the CaCO₃ cementation properties between soil particles. This is achieved by upscaling the model from level I to a representative pair. The analysis at level II leads to the introduction of generalised equivalent cementation bonds (ECB) that effectively capture the mechanical behaviour of interparticle CaCO₃ crystals and their dependence on sand properties. At level III (MICP-treated soils), large-scale simulations of soil particles are utilised to examine the effects of CaCO₃ mass fraction and soil void ratio on the mechanical behaviours of soil samples, based on the generalised ECB developed at level II. Additionally, to account for the CaCO₃ heterogeneity of MICP-treated soil, random field theory is utilised to generate spatially varied CaCO₃ distribution at this level. The results of DEM analysis indicate that CaCO₃ mass fraction significantly enhances soil strength, while its influence on elastic modulus is less pronounced. Furthermore, a reduction in void ratio increases interparticle bonds, leading to improved mechanical performance. This effect is further amplified by higher CaCO₃ mass fractions.