Model-Based Intelligent Simulation of Underground Heat Exchanger

摘要

Referring to load aggregation scheme and superposition of pulses scheme, combining the conventional mathematical method with artificial neural network (ANN) is used to improve the modeling of the underground heat exchanger. In order to study computation time and calculation precision of the new model based on ANN, the improved model and the cylindrical source model are employed to simulate GSHPS operation under same simulation conditions. The simulation results show that new model-based intelligent simulation of ground heat exchanger has a good precision. Mean error of heat absorption capacity between the new model based intelligent and the cylindrical source model is 2.6%. Mean error of outlet flow temperature results between the new model based intelligent and the cylindrical source model is 1.28%. The computing time of new model is less than that of underground heat exchanger model based on superposition principle and cylindrical source theory. The computing time of improved model is lower than it of the no improved model about an order of magnitude under same simulation conditions.