Performance optimization of a finned-curved tubes compact heat Exchanger: Numerical simulation and machine learning method

Abstract

This study aims to develop efficient heat-transferring systems by optimizing the performance of compact heat exchangers. The research proposes novel tube shapes and examines the placement of vortex generators to enhance efficiency. Various tube configurations, including simple oval, oval-segmented, Cassini oval, parabolic-segmented, converging oval-segmented, and diverging oval-segmented shapes, are evaluated. The study considers Reynolds numbers ranging from 500 to 1500, and vortex generator angles at three distinct values. Results indicate that diverging ovalsegmented tubes with horizontal vortex generators achieve the highest Nusselt number, while Cassini-shaped tubes exhibit the highest friction factor. The maximum performance evaluation criterion exceeds 1.5. Using the simultaneous proposed geometries and vortex generators, the goodness factor can increase more than 60 %. Additionally, a more than 30 % increment in the Colburn factor is expected. The vortex generator with a 90◦ angle showed the best improvement in Nusselt number, outperforming other types by 10–15 % and cases without vortex generators by less than 10 %. Furthermore, to reach the maximum performance evaluation criterion, it incorporates a combination of artificial neural networks and genetic algorithms to determine optimal Reynolds numbers and vortex generator angles, contributing to the innovation in the design and performance of compact heat exchangers