Using CFD and a genetic algorithm, the baffle design of a parallel heat exchanger header is numerically optimized.

Abstract

One of the primary causes of decreased thermal performance in parallel heat exchangers is
improper refrigerant flow distribution. In order to improve the flow distribution of two-phase flow refrigerant
inside a heat exchanger header, this work presents a numerical approach to automatically estimate the height of
the baffles. This technique uses a genetic algorithm and the computational fluid dynamics (CFD) model to
simultaneously produce the best dimensions for several baffles. The numerical results demonstrate that, in
comparison to the original design, the flow maldistribution might be reduced by up to 9 times. This framework
also has the potential to increase the design area, making it a valuable tool for improving header designs that
incorporate parallel heat exchangers.