Modeling Biocemented Sands: Hypoplastic Model for MICP-Based Ground Improvement

Abstract

Bio-cementation offers a sustainable and low-carbon alternative to conventional ground improvement methods, aligning with the growing demand for green geotechnical solutions. To enable reliable numerical simulation of bio-cemented soil behavior in engineering applications, advanced constitutive models are required to account for the effects of microbially induced bonding and the resulting changes in mechanical response. In this study, an existing hypoplastic model that accounts for intergranular strain and semi-fluidized state is adopted as a benchmark. The model is further modified and implemented in finite element code to investigate the spatial heterogeneity characteristic of bio-cemented soils, arising from nonuniform bacterial activity and localized calcite precipitation. An extension of the model is introduced by incorporating a generalized critical state surface based on the Matsuoka–Nakai criterion. The model's performance is validated against a series of monotonic and cyclic triaxial tests on both untreated and bio-cemented sand samples. Key capabilities of the model include its ability to accurately capture the enhancement of strength and stiffness due to cementation, the evolution of pore pressure, the manifestation of cyclic mobility, and the increased resistance to liquefaction. To account for spatial variability in bio-cementation, a boundary value finite element model was developed in PLAXIS using cement content measurements obtained at nine locations within the specimen. The resulting model more accurately captured the nonuniform stress and strain distributions and demonstrated superior agreement with experimental results compared to the homogenized sample.