Freezing performance of selective artificial ground freezing based on three-pipe structure and its influencing analysis

Abstract

Selective artificial ground freezing (S-AGF) enables directional cooling and mitigates unnecessary thermal loss through air-insulated casings, thereby exhibiting superior energy-saving and stratum-protection performance relative to conventional artificial ground freezing (C-AGF). Existing S-AGF investigations have primarily concentrated on single-pipe configurations, whereas the parametric mechanisms and matching principles governing multi-pipe layouts, which are extensively deployed in engineering practice, remain insufficiently elucidated, and quantitative criteria for refined design are still lacking. Therefore, the COMSOL-based thermo-hydraulic coupled numerical simulations were conducted to investigate the freezing performance and its influencing for three-pipe S-AGF. Comparative analyses were conducted between S-AGF and C-AGF regarding frozen-wall evolution, temperature-field distribution, and energy consumption; furthermore. And the quantitative influences of pipe spacing, coolant temperature, flow velocity, and air-insulation-layer thickness on directional freezing performance were elucidated. The results demonstrate that S-AGF effectively suppresses heat diffusion within non-target strata, yielding a 50.0% reduction in peak energy consumption and a 22.6% reduction in cycle-averaged energy consumption. Pipe spacing has the dominant influence on freezing behavior, when it increases from 222 mm to 372 mm, frozen-wall closure time increases by 158%, and 322 mm is identified as the optimal spacing that balances formation efficiency and temperature uniformity. Reducing the coolant temperature from −20 °C to −35 °C shortens closure time and qualified formation time by 28.72% and 30.33%, respectively. Coolant velocity demonstrates limited regulatory capacity, with 2.0 m/s identified as the effective threshold beyond which additional velocity increments yield negligible efficiency improvements. The air gap exhibits an optimal thickness of 3.8 mm; exceeding this thickness triggers internal natural convection and consequently deteriorates insulation performance. This study could provide the quantitative design references for layout optimization, coolant-system selection, and insulated-casing fabrication in practical directional-freezing engineering applications.