Prediction of thermophysical properties of hybrid nanofluids using machine learning algorithms

The current research focuses on identifying machine learning algorithms that provide results with high accuracy. The present work is conducted in three phases: conduction of heat transfer experiments, development of correlation, implementation, and comparison of machine learning algorithms with the correlation. Experiments were conducted using hybrid nanofluids with graphene platelets, and carbon nanotubes dispersed in Ethylene glycol-water mixtures. Ethylene glycol percentage in the base fluid varied from 0 to 100%. The nanoparticles are dispersed in concentrations of 0.5, 0.25, 0.125, and 0.0625 weight fractions. The results achieved a 15 to 24% enhancement in thermal conductivity. Results showed viscosity increased in temperatures ranging from 50 to 70 °C but less in higher temperatures. Correlation formulas were developed, and they predicted the thermal conductivity and viscosity values with a maximum deviation of 10%. Machine learning (ML) models have been implemented, and a comparative analysis with correlation results has been conducted. These ML models provided results with a maximum deviation of 4% for viscosity and 3% for thermal conductivity.