Answers are grounded in this paper and cite their sources.
Numerical modeling reveals ground freezing boosts conventional heat pump efficiency while delaying peak ground temperatures, highlighting the need to include phase change in thermal predictions.
Most existing numerical studies on ground source heat pump (GSHP) systems assume negligible ground freezing during operation. However, this assumption has not been systematically validated. The lack of quantitative evaluation of freezing effects introduces significant uncertainty in predicting ground thermal response and system performance. This study employs coupled heat transfer modeling implemented in a finite volume CFD tool and experimental validation to quantify the impact of ground freezing on the thermal and thermodynamic performance and prediction accuracy of GSHP and underground seasonal solar thermal energy storage systems in a cold climate. The findings show that neglecting ground freezing can cause the system's performance to be underestimated. This is because soil moisture and the phase change of water within the soil help regulate and stabilize temperature variations around the borehole. For conventional GSHP systems, ground freezing improved average heating and cooling COPs as well as ground thermal energy extraction rate by up to 2.6%, reduced heat pump energy consumption by 5%, and delayed peak ground temperature by approximately 2000 h — revealing the critical influence of phase-change dynamics on system longevity and efficiency. For SAGSHP systems with ground freezing, the average cooling COP reduced by 2.3% with no significant changes in the average heating COP, the total heat pump energy consumption, and the extracted ground thermal energy. The freezing radius for GSHP is within 0.3 m and 0.1 m, for GSHP and SAGSHP, respectively. Overall, this study contributes new insight into the coupled thermo-hydraulic behavior of GSHP systems under freezing conditions, offering a validated framework for integrating freezing effects into design and prediction models.
No takes yet. Share an insight, caveat, or question.
Adebayo et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: