Application of an Artificial Neural Network (ANN) for predicting low-GWP refrigerant condensation heat transfer inside herringbone-type Brazed Plate Heat Exchangers (BPHE)

Abstract

This paper presents an Artificial Neural Network (ANN) model for predicting refrigerant condensation heat transfer coefficients inside herringbone-type Brazed Plate Heat Exchangers (BPHE). The model accounts for the effect of plate geometry, operating conditions and refrigerant properties both in the saturated and in the superheated vapour condensation regimes. The model predictions demonstrate good agreement with a database of 1884 data points comprising 12 plate geometries and 16 refrigerants (including 4 so-called natural refrigerants and 6 other low-GWP refrigerants). The Mean Absolute Percentage Error (MAPE) of the model predictions is 3.6%. The results demonstrates that the ANN model presented herein is capable of better predictive capability than most of the state-of-the-art BPHE analytical-computational models presented in the open literature. The characteristic parameters of the ANN model are reported in the paper.