Mixed refrigerants suitability analysis using artificial neural networks

This paper presents an artificial neural network approach with back propagation algorithm (BPA) to find an alternative to Chloroflurocarbon (CFC) by considering the mixture of Hydro fluorocarbon (HFC) and Hydrocarbon (HC). The thermodynamic properties of refrigerants are obtained usinf REFPROP 9. Correspondingly, the Coefficients of Performances (COPs) of the mixed refrigerants have been obtained. The testing of the ANN shows high performance in estimating the closest COP.