Experimental and Machine Learning Approach to Evaluate the Performance of Refrigerator and Air Conditioning Using TiO2 Nanoparticle

Abstract

160The refrigerants used in the vapour compression refrigeration system, have global warming potential and ozone depletion potential that is harmful to the natural environment. In this research, R134a is used which is eco-friendly and widely used refrigerant and it has zero ozone depletion potential. Polyolester oil is compressor oil that is of great suitability with R134a refrigerant in the context of functionality. In this paper, experimental as well as simulation work was carried out on TiO2-polyolester oil nano lubricant and R134a refrigerant used in a vapour compression refrigeration system. The nano lubricant is prepared with different concentrations of TiO2 in polyolester oil. In general, a circular shape capillary tube is used for the expansion of refrigerant but we used a cubic shape capillary tube as an expansion device in a system. The result shows that when added 1.5g/L of TiO2 nanoparticles, a better coefficient of performance is obtained and works efficiently and safely in the system. The simulation was done through machine learning to predict the coefficient of performance with different algorithms like support vector regression and Gaussian process regression. Machine Learning algorithms have different kernel functions but Pearson VII and Radial-based kernel functions were used for prediction purpose. The developed machine learning models are used to compare the predicted outcomes with experimental results. 75% of readings of experimental data were used for training the developed model and the remaining were used for testing models. It is found that support vector regression works better in the context of Gaussian process regression technique.