A machine learning–based surrogate model for estimating the air-side friction factor in plain finned-tube heat exchangers with staggered arrangement

Author

Faculty of engineering,ShohadayeHoveizeh Campus of Technology, Shahid Chamran University of Ahvaz, DashteAzadegan, Iran.

10.22044/jsfm.2026.17637.4058

Abstract

Accurate determination of the friction factor in finned-tube heat exchangers through experimental and numerical methods entails substantial costs and significant computational effort. Therefore, the development of an alternative approach for rapid and precise estimation is essential. In this study, the air-side friction factor of a plain finned-tube heat exchanger with a staggered arrangement is estimated with high accuracy using machine learning techniques, including Bagging, M5P, M5Rules and REPT. A large dataset comprising 1,149 observations, extracted from various reliable sources, is employed for training and testing the developed models. Fin pitch, transverse tube pitch, longitudinal tube pitch, and Reynolds number are considered as input parameters, while the air-side friction factor is defined as the target variable. The results indicate that the Bagging method achieves the highest accuracy among the investigated models, with correlation coefficients of 0.9983 and 0.9968 for the training and testing datasets, respectively. The mean absolute error (MAE) and root mean square error (RMSE) for this method are 0.0003 and 0.0004 for the training data, and 0.0004 and 0.0006 for the testing data, respectively, demonstrating the very low prediction error and high reliability of this approach in estimating the friction factor.

Keywords

Main Subjects

Journal of Solid and Fluid Mechanics
Volume 16, Issue 4
September and October 2026
Pages 1-14

Files

Share

How to cite

Statistics