Compression Law and Settlement Calculation Method of Over-Wet Soil Based on Large Samples

Previous studies have shown that over-wet soil is challenging to compact and exhibits large creep deformation. The consolidation test of small specimens cannot accurately reflect the compression law, and creep is underestimated owing to size effects, which affects the engineering quality. In order to accurately analyze the compression process of over-wet soil and establish its settlement calculation method, this study focuses on over-wet soil in Anhui Province, China, and uses a large-sized tester to load and analyze its compression law. The thermogravimetric analysis method was employed to investigate the changes in water with different binding forces during the compression process, and the settlement calculation method for over-wet soil was explored. The results show that the creep of over-wet soil is large and long-lasting, and the three-stage consolidation division method based on the d−t curve is more effective in analyzing its regularity. The creep of over-wet soil is directly proportional to its water content. When the load exceeds the pre-consolidation pressure, the creep deformation becomes more significant, accounting for about 60% of the deformation under a single level load. It is recommended to use the creep coefficient (λ) for calculation. The results of the thermogravimetric analysis indicate that during the primary consolidation stage, free water is discharged, and weakly bound water is mainly discharged during the third consolidation stage, which is the main cause of creep. Finally, based on the relationship between the creep strain and water content of large samples, a calculation method for the settlement of over-wet soil foundations based on the layered summation method was established, which had a higher prediction accuracy than the conventional layered summation method. The results of this study will help clarify the deformation process and principle of over-wet soil and improve the quality of engineering.