An experimental study for evaluation of collapsible loess roadbed replacement method using lightweight soil

Water is one of the most significant variables that contribute to loess collapse, hence in collapsible loess areas, stringent waterproofing measures are frequently employed to limit loess roadbed collapse. During the construction of roads in sponge city, the roadbed must have natural-like permeability and water retention capacities to assure normal permeation and retention of rainwater. These capacities are contradictory with the collapse of the loess roadbed. This research proposes a collapsible loess roadbed replacement method based on optimized lightweight soil with cotton stalk fibers (LSCF) to overcome this contradiction. The optimization of the LSCF mixture ratio was based on density, unconfined compressive strength, and permeability tests. The optimized LSCF has a low density (1.05 g/cm³), sufficient strength (178.43 kPa), and similar permeability to natural loess (4.03 × 10⁻⁶ cm/s). In addition, a method for calculating replacement depth and the corresponding replacement construction processes were developed. Three methods, including theory calculation, centrifugal model test, and practical field demonstration, were used to further evaluate the applicability and effectiveness of the proposed replacement method. Theoretical and experimental results demonstrate that the replacement method effectively reduces the self-weight collapsibility settlement (more than 20%) and maintains good permeability and water retention capacity of the loess site. The research findings provide engineering recommendations for sponge city roadbed treatment in collapsible loess areas.