Comprehensive analysis and design method of ground heat exchangers under the influence of groundwater seepage

Traditional semi-empirical methods are difficult to accurately determine the length of ground heat exchangers (GHEs) in ground source heat pump system, particularly under seepage conditions, which predominantly employ equivalent thermal conduction to approximate the convective heat transfer. This work develops a dynamic model integrating building loads, heat pump units, and pipe groups to simulate annual system operation considering the coupled effects between heat conduction and groundwater advection, and innovatively introduces data-driven modeling to propose an intelligent design method. The optimal length of GHEs under underground parameters is first explored, the sensitivity analysis results indicate that soil thermal conductivity has greatest impact, followed by seepage velocity and then aquifer height. The load characteristic parameters are followed discussed, finding that unlike accumulation parameters and standard deviation, the influence of intensity parameters enhances with increase of seepage strength. When the maximum cooling load increased by 51.26 %, the length increased by 25.93 % and 42.86 % at seepage velocities of 0.05 m/d and 0.50 m/d, respectively. Finally, a database is established by extracting the numerical results, and the length design model is developed using deep neural network to learn from the database. The results show that the design deviation ranged from −6.88 % to 5.76 % across multiple samples. This work can provide a new way for the design method of GHEs.