Machine learning prediction of nanofluid thermal and flow characteristics

This research presents a study on the prediction of nanofluid thermal and flow characteristics using machine learning. The modelling and simulation of nanofluid is a useful tool in predicting nanofluid thermal and flow characteristics. It aids design of new products and helps researchers fully forecast how the fluid will behave under different conditions and settings.
The present study has been successful in improving the accuracy of nanofluid simulations and the thermal and flow characteristics prediction. Through a series of analyses using machine learning, new nanofluid models have been created, and tested, and new insights have been produced both on the nanofluids aspect and the machine learning aspect.
From the study, it was observed that three (3) modelling assumptions namely dispersion, Buongiorno, and discrete phase model (DPM) showed high accuracy in modelling nanofluids. They all have an accuracy within 5%, with the dispersion model being the best for both constant and temperature-dependent property assumptions. The discrete phase model (DPM) model seems to do better in turbulent flow than in laminar flow, while the Buongiorno models do better under laminar flow conditions. The single phase model and mixture models were observed to be the worst and only recommended where there are no other choices. The accuracy of the single phase model can be considerably improved by the substitution of the thermophysical properties of the nanofluid with an accurate machine learning model. The feature selection method based on the physics of the problem is a highly effective machine learning modelling strategy with the possibility of higher accuracy and generalization. Neural network models are considerably effective in modelling nanofluids' thermophysical properties.