Prediction of Heat Transfer Performance Using Polynomial Regression

Abstract

This paper is concentrated on the application of polynomial regression which is used to find the relation between two or more variables. In the present work polynomial regression model was used to find the contribution of the input parameters during spray impingement cooling on microchannel. The model was designed and trained using a total 120 experimental data points collected from the experiment. Experiment was carried out to find heat transfer coefficient and mass flux density during spray on microchannel which is used to cool electronic chips. The mass flux density calculated from the experiment, air and water pressures and nozzle tip to microchannel surface distance were used as input variables for the regression model, while the corresponding heat transfer coefficient was selected as its output variable. By comparing polynomial regression model with linear, ridge regression and ANN, it was evident that heat transfer coefficient with high accuracy was achieved for polynomial regression. Various degrees of model were also compared upon which it was found that regression model with degree three outperforms the others. Among four independent variables, Water Pressure, Nozzle height and Mass flux density turns out to be most significant variable in calculating the regression model for Heat Transfer Coefficient. © 2022 IEEE.