Energy analysis of a domestic refrigerator system with ANN using LPG/TiO₂–lubricant as replacement for R134a

This paper experimentally investigated and also modeled using artificial neural networks (ANN) approach the energy analysis of a domestic refrigerator using selected charges of LPG refrigerants (40, 50, 60 and 70 g) and different concentrations (0, 0.2, 0.4 and 0.6 gL⁻¹) of TiO₂-based nanolubricants as replacement for R134a refrigerant. The parameters for energy analysis include: compressor power consumption, cooling capacity, COP, compressor discharge temperature and pressure ratio. The findings showed that cooling capacity and COP of the domestic refrigerator using LPG refrigerant with TiO₂ nanoparticles dispersed in a mineral oil lubricant was found to be higher than that of R134a by around 18.74–32.72 and 10.15–61.49%, respectively. Furthermore, compressor power consumption and pressure ratio of the domestic refrigerator using LPG refrigerant with TiO₂–mineral oil lubricant were also found to be lower than that of R134a by about 3.20–18.1 and 2.33–8.45%, respectively, under similar conditions. The compressor discharge temperature was also found to be lower using LPG refrigerant with lubricant TiO₂–mineral oil than R134a. The results further suggested that the best energetic performance of the domestic refrigerator was obtained using 40 g charge of LPG refrigerant with TiO₂–mineral oil lubricant at 0.4 gL⁻¹ of TiO₂ concentration under similar operating conditions. The predictions from ANN models had excellent alignment with experimental results giving a range of values for R² from 0.914 to 0.970, RMSE from 0.111 to 2.317, and MAPE from 0.865 to 3.148%.