Case study of performance evaluation of ground source heat pump system based on ANN and ANFIS models

Abstract

This paper presents case studies with a method to predict the coefficient of performance (COP) of the heat pump and the COPs of ground source heat pump (GSHP) system with limited parameters. The method was based on an artificial neural network (ANN) model and an adaptive neuro-fuzzy inference system (ANFIS) model. Two GSHP systems were monitored respectively to get the training and test data. ANN models with different neurons in the hidden layer were compared according to the statistical validation results, and the models with five neurons in the hidden layer appears to be the best optimal topology for the prediction of COP of the heat pump and COPs of the system. ANFIS with different membership functions (MFs) and various numbers of MFs were compared. Gaussmf with three functions appeared to be the most optimal membership function for the ANFIS model calculating the COP of the heat pump. The optimal ANFIS was the model with two Gaussmf as its member function to predict the COPs of the system. It was found that the models provided high accuracy and reliability for calculating performance indexes of GSHP system.