Application of Machine Learning to the Analysis of Thermal Storage System

Abstract

Thermal energy storage is one of the methods to reduce irreversibilities in the thermal power generation process. Phase change material (PCM) is a material that has been investigated by researchers worldwide in that direction. Literature reveals an increase in thermal storage performance with the addition of metallic nanoparticles to PCM. Hence, in the present study, the influence of alumina nanoparticles on the thermal storage performance of paraffin wax is investigated. Experimental data was obtained by varying the nanoparticle concentration from 0% to 1.5% (by vol.). Further computations were carried out by subjecting the data to analysis of variance (ANOVA) (at 95% CI) to ascertain the impact of nanoparticles on thermal storage performance like temperature, heat absorbed/desorbed, and so on. Further, a regression equation was developed having a coefficient of determination (R²) more than 0.95. The equation is then used to generate more than 50 data sets by varying the nanoparticle concentration and a surface response plot is generated for each output against time. The data set is further used to arrive at an optimal nanoparticle concentration that maximizes output performance using particle swarm optimization (PSO). The optimization study revealed that a nanoparticle concentration of 0.72 within an initial period of 5 s would harness the maximum amount of energy absorbed or released from the PCM.