Ensemble Learning based Molecule Embedding to improve nanofluids' Thermal Conductivity prediction

Abstract

The thermal conductivity of nanofluids plays a critical role in numerous industrial applications, including lid-driven cavities and metallurgical lubrication technology. However, accurately predicting this property remains a persistent challenge, necessitating continued innovation. In this study, we propose a novel methodology that leverages ensemble learning with molecular embeddings to enhance the prediction of nanofluid thermal conductivity. By integrating existing correlations with experimental data, our approach generates robust predictive models that outperform state-of-the-art methods. Experiments conducted on a real-world dataset demonstrate the superior performance of the proposed framework, highlighting its potential to advance research and industrial applications in nanofluid heat transfer.