Artificial neural network approach for irreversibility performance analysis of domestic refrigerator by utilizing LPG with TiO2-lubricant as replacement of R134a

This paper experimentally investigated irreversibility performance analysis of domestic refrigerator using 40 g, 50 g, 60 g and 70 g of LPG refrigerant with varying concentration of TiO2 nanoparticles dispersed in a mineral oil lubricant as replacement of R134a. The irreversibility performance analysis of the domestic refrigerator investigated with test parameters including irreversibility in the components, total irreversibility, and second law efficiency. Findings showed that under similar operating conditions irreversibilities in condenser, capillary tube, evaporator and compressor of the domestic refrigerator using LPG refrigerant with lubricant TiO2-mineral oil found lower than R134a by 6.05-48.10%, 17.83-46.98%, 2.51-33.17% and 2.21-23.29% respectively. Additionally, total irreversibility and second law efficiency in the domestic refrigerator using LPG refrigerant with TiO2, nanoparticles dispersed in a mineral oil lubricant found lower and higher than R134a by around 0.65-35.97% and 6.23-54.74% respectively. All in all 0.4 g L-1 of TiO2 and 40 g of LPG charge were the attained optimal values of nanoparticle concentration and LPG charge respectively in the domestic refrigerator utilizing LPG as replacement of R134a. Artificial Neural Networks (ANN) approach to delineate total irreversibility and second law efficiency of the domestic refrigerator was additionally applied to develop the models. It found that ANN models predictions concurred well with experimental results and brought out an absolute fraction of variance of (0.989-0.990), root mean square error of (0.831-1.061) and mean absolute percentage error of (1.734-2.056%) respectively with experimental results. (C) 2018 Elsevier Ltd and IIR. All rights reserved.