Long-term performance prediction of ground source heat pump system based on co-simulation and artificial neural network

Abstract

Ground source heat pump (GSHP) systems are efficient and widely used form of shallow geothermal energy utilization. In practical application, GSHP systems face the problem in predicting the long-term performance and energy efficiency variations in GSHP systems in regions with thermal imbalances. In this study, a coupling heat transfer model of vertical ground heat exchanger (GHE), heat pump, and building load was established to achieve refined simulation calculations of the long-term operation of GSHP systems. Parametric modeling was conducted by employing the co-simulation of COMSOL and MATLAB, enabling dynamic simulations of 600 sets of GSHP system models throughout the entire year. A database was established by extracting the numerical results, and a prediction model was developed using artificial neural network (ANN) methodology to learn from the database. The model predicted the ground temperature, seasonal energy efficiency ratio (SEER), seasonal coefficient of performance (SCOP), annual electricity consumption, and annual energy savings of the GSHP system and the prediction model was valid. The results were used to estimate the ground temperature, coefficient of performance (COP), payback period for static investment, and carbon dioxide emission reduction. Finally, the prediction model was applied to a GSHP system engineering project in Yancheng, China, forecasting the operation performance of the system for 10 years. The accuracy of ANN in predicting 10 years of long-term performance is verified.