Investigation of thermal properties of TiN/MWCNT-OH hybrid nanofluids and GWO-BP neural network model

Abstract

Hybrid nanofluids have garnered significant attention due to excellent heat transfer performance and potential applications. Conducting comprehensive research on hybrid nanofluids holds paramount importance. This study investigates the effects of surfactants, particle concentrations, mixing ratio and storage time of TiN/MWCNT-OH hybrid nanofluids on stability, thermal conductivity, and viscosity. It proposes a Grey Wolf Optimizer-Backpropagation neural network model for predicting thermal properties. The results indicate that the inclusion of PVP-K30 surfactant leads to remarkable stability of hybrid nanofluids at a concentration of 50 ppm over a period of two weeks. An increase in the proportion of MWCNT-OH results in a slight increase in thermal conductivity, which exhibits a maximum increase of 46 % with elevated temperature and particle concentrations. The viscosity of hybrid nanofluids gradually decreases as temperature rises, although demonstrates a non-linear correlation with concentrations. The neural network model exhibits a high predictive accuracy of 99.3507 % for thermal conductivity and 98.8924 % for viscosity.