Integration of Genetic Programing With Genetic Algorithm for Correlating Heat Transfer Problems

Abstract

In the present paper, the genetic programing (GP) is integrated with the genetic algorithm (GA) for deriving heat transfer correlations. In the process of developing heat transfer correlations with the approach (GP with GA (GPA)), the GP is first employed to obtain some potential optimal forms. After that, the forms are further optimized with the global GA to reach minimum errors between the predicted values and experimental values. With the proposed approach, three typical different heat transfer problems are applied to the data reduction processes from published experimental data, which are heat transfer in a shell-and-tube heat exchanger (STHE) with continuous helical baffles, a single row heat exchanger with helically finned tubes and a finned oval-tube heat exchanger with double rows of tubes, respectively. The results indicate that the GPA approach could improve the performance of heat transfer correlations obtained with the GP. Compared with the power-law-based correlations, the heat transfer correlations obtained with the approach have higher predicted accuracies and more excellent robustness.