A reliable model for predicting the settling of soft soil subgrade to ensure highway safety

This paper introduces a new approach to managing subgrade settlement through finite element inverse analysis, focusing on creep effects in soft soil. Using MATLAB to connect Abaqus with an Improved Whale Optimisation Algorithm (IWOA) allows researchers to effectively extract calculation parameters for soft soil layers. The simulation of settlement deformation in subgrade soil, with due consideration of the creep effect, is achieved through the use of the modified Drucker–Prager yield criterion model and the time hardening creep law coupling model. The innovative approach enhances the Whale Optimisation Algorithm’s ability to avoid local extrema stagnation and improves the overall optimisation process. The IWOA was evaluated on 16 benchmark functions and compared with six other AI algorithms. Its effectiveness in global function performance was analysed using the Friedman test on the mean values from 30 simulation results. The evidence suggests that IWOA performs exceptionally well. Thereafter, the proposed method’s effectiveness was confirmed by developing a finite element inversion model for a highway subgrade on soft soil. The model was used to simulate subgrade deformation throughout construction, allowing for a comparison with experimental data to validate its reliability. The outcomes demonstrate a clear correlation between subgrade deformation from inversion analysis and real-world data. It is projected that, for the next 10 years, the post-construction settlement will be limited to a maximum variation of 10 cm. These projections have the potential to offer valuable theoretical insights for improving the efficiency of highway management and maintenance.