Machine learning specific heat capacities of nanofluids containing CuO and Al2O3

Abstract

AbstractEmpirical work and modeling approaches have shown that fundamental thermophysical parameters of constituents, nanoparticles, and base liquids have complex but synergistic effects on the specific heat capacity of nanofluids. In this work, we develop the Gaussian process regression model to investigate the statistical relationship among temperature, specific heat capacities of nanoparticles and base liquids, nanoparticle volume concentrations, and specific heat capacities of nanofluids. The model is developed with a dataset containing nanofluids with CuO and Al2O3 nanoparticles, and water and ethylene glycol (EG) as base liquids. The model also applies well to nanofluids containing mixtures of water and EG, and Al2O3 nanoparticles with different particle size distributions. The model is highly accurate and stable that contributes to fast and low‐cost estimations of specific heat capacities of nanofluids over a wide range of compositions and temperature.