Optimization of Liquid Nitrogen Freezing in a Water-Rich Sand Stratum with High-Velocity Seepage Based on the PSO Algorithm

The quality of artificial ground freezing construction deteriorates significantly in tunneling projects involving water-rich sandy strata with high flow velocities. This study aims to optimize the design of freezing schemes by improving construction efficiency, while ensuring cost-effectiveness under these challenging hydrogeological conditions. Using the liquid nitrogen freezing project in a high-permeability sand stratum as a research objective, COMSOL Multiphysics (version 6.1)-based numerical simulations are conducted to analyze the evolution of the temperature field and the formation of the frozen arch under high-velocity seepage conditions. An enhanced freezing configuration is subsequently developed through integration with the particle swarm optimization (PSO) algorithm. The results indicate that rapid groundwater flow delays upstream frozen arch closure by 3.37 days and induces displacement and thickness irregularities, as reflected by a reduction in standard deviation from 0.71 to 0.27 following optimization. The optimized scheme achieves a reduction of at least 20.14 days in the total freezing duration compared with conventional uniformly spaced freezing pipe arrangements. Temperature uniformity within the frozen zone improves significantly, with the upstream–downstream thermal differential decreasing from 30.85°C to 1.12°C. Furthermore, the revised configuration effectively mitigates upstream thermal erosion through synchronized arch development, thereby enhancing mechanical stability, evidenced by a 15% increase in load-bearing capacity and operational safety, while also preventing excessive energy consumption.