| Abstract |
Ground heat exchangers (GHEs), particularly compact and efficient helical borehole type (HBHEs), enable year-round renewable heating and cooling by utilizing stable subsurface temperatures, offering higher thermal performance and lower spatial impact than conventional systems—crucial for sustainable, space-limited urban energy solutions. Despite progress in GHE research, HBHE studies lack design optimization using comprehensive datasets. Existing works often omit combined effects of spiral aspect ratio (D/L), soil conductivity (k_soil), revolution number (N), and initial inlet temperature (T_ini).The present study therefore fills the gap by coupling computational fluid dynamics (CFD) simulation dataset with an Artificial Neural Network (ANN) based optimization algorithm to maximize the thermal efficiency of the HBHE. With this coupled approach, a comprehensive analysis of combined effects of such parameters on thermal efficiency is successfully made while identifying key constraints, enabling efficient, predictive heat extraction modelling. Conjugate heat transfer phenomenon in HBHE is simulated in transient manner using a multiphysics finite element fluid flow solver under varying conditions. Feature importance analysis identifies critical design constraints for optimization. The results show that D/L, k_soil, N, and T_ini significantly and nonlinearly affect HBHE thermal performance. Tighter spirals increase surface area and heat extraction rate but raise thermal interference at high N values. Soil conductivity dominates in low-flow conduction regimes. The initial inlet temperature captures inlet-ground interaction, indicating responsiveness. The machine learning model exceeds 95% accuracy, identifying D/L and T_ini as most sensitive. Findings reveal complex parameter interdependencies shaping thermal behavior of HBHE structure. |