Use of artificial neural network approach for depicting mass flow rate of R134a /LPG refrigerant through straight and helical coiled adiabatic capillary tubes of vapor compression refrigeration system

Abstract

In this work, an experimental investigation carried out with R134a and LPG refrigerant mixture for depicting mass flow rate through straight and helical coil adiabatic capillary tubes in a vapor compression refrigeration system.Various experiments conducted under steady-state conditions, by changing capillary tube length, inner diameter, coil diameter and degree of subcooling. The outcomes demonstrated that mass flow rate through helical coil capillary tube discovered lower than straight capillary tube by about 5−16%. Dimensionless correlation and Artificial Neural Network (ANN) models developed to predict the mass flow rate. It found that dimensionless correlation and ANN model predictions concurred well with experimental results and brought out an absolute fraction of variance of 0.961 and 0.988, root mean square error of 0.489 kg/h and 0.275 kg/h and mean absolute percentage error of 4.75% and 2.31%, respectively. The outcomes suggested that ANN model shows better statistical prediction than dimensionless correlation model.