Estimation of Properties of Liquid-Vapor Mixture of Some Refrigerants at High Pressure for Solar- Photovoltaic Refrigeration

Abstract

In this work, a hybrid method based on neural network and particle swarm optimization is applied to literature data to develop and validate a model that can predict with precision vapor–liquid equilibrium data for the binary systems (hexafluoroethane (R116(1)), 1,1,1,2-tetrafluoroethane (R134a) and R1234ze) . ANN was used for modelling the non-linear process. The PSO was used for two purposes: replacing the standard back propagation in training the ANN and optimizing the process. The training and validation strategy has been focused on the use of a validation agreement vector, determined from linear regression analysis of the predicted versus experimental outputs, as an indication of the predictive ability of the neural network model. Statistical analysis of the predictability of the optimized neural network model shows excellent agreement with experimental data (coefficient of correlation equal to 0.998). Furthermore, the comparison in terms of average relative deviation (AARD%) between, the predicted results for the whole temperature and pressure range shows that the ANN- PSO model can predict far better the mixture properties than cubic equations of state.