XGBOOST-BASED MODEL FOR PREDICTION OF HEAT TRANSFER COEFFICIENTS IN LIQUID COLD PLATES

Extreme gradient boosting (XGBoost) algorithm is a newly developed machine learning (ML) technique with demonstrated excellent accuracy and performance in modelling complex processes in science and engineering. In the present study, an XGBoost-based model is developed to predict heat transfer coefficients in liquid cold plates (CPs) subjected to surface roughness. The CPs operate in turbulent flow over a wide range of Reynolds numbers. Roughness sizes range from zero (smooth surface) to 0.5 mm. The input dataset for training the XGBoost model is prepared using a computational fluid dynamics (CFD) approach and by solving three-dimensional fluid flow and heat transfer inside the CPs. It was found that the model exhibits excellent accuracy such that 63% and 90% of heat transfer coefficients are predicted within +/- 10% and +/- 20% of true values, respectively. The present finding suggests XGBoost as an effective modelling tool for performance analysis of thermal management solutions, specifically when there is limited performance data available in literature.