Abstract
In continuous strata engineering, such as foundations and underground caverns, the differences in shear strength between sandstone and mudstone rock mass–structural plane–structural block systems critically affect design and safety. However, the underlying mechanisms and controlling factors of these shear strength parameters remain poorly understood, leading to challenges in optimizing engineering strategies. This study investigates the differences in shear strength parameter correlations between sandstone and mudstone and develops an intelligent model for predicting rock mass displacement. We constructed multi-parameter correlation models using laboratory and field shear test data combined with a random forest algorithm. The results show that the model achieved high prediction accuracy (R2 = 0.997–0.998, RMSE = 1.649–3.898, MAE = 1.110–2.991). For instance, the peak shear strength of sandstone structural planes was approximately 54% higher than that of mudstone. Sensitivity analysis revealed that for sandstone, structural plane shear stress (27.80%) and structural block stress (25.50%) are the most sensitive factors, while for mudstone, structural plane shear displacement (35.20%) and structural block strain (34.20%) dominate. These correlations are model-predicted based on empirical data from shear tests. These findings provide a mechanistic understanding of plastic instability in sandstone and slip-strain-induced fissure extension in mudstone, and they can guide shear strength prediction and stability assessment in mixed sandstone–mudstone strata. The study contributes to the field by offering a quantitative basis for stratified adaptive design in continuous strata engineering, enhancing the efficiency and safety of foundation treatment and cavern support.