Tag
This paper presents a hybrid analytical-PINN model for simulating geothermal heat exchangers in heterogeneous subsurface, removing singularities and enabling efficient neural network training via an analytical correction approach.