An Open-Source algorithm for automatic geometrical optimization of extruded liquid cold plates for enhanced thermal management in railway electronics

Abstract

This paper presents the development and application of an optimization algorithm for determining the geometric parameters of an extruded Liquid Cold Plate (LCP) with internally finned channels. The entire workflow operates within a fully open-source environment, offering a comprehensive and accessible solution for optimizing LCP geometric parameters for efficient thermal management in railway power electronics as well as other industrial applications. In particular, the aim is to minimize the maximum temperature and the temperature gradient at the interface between the LCP and an electronic device for electric trains that dissipates heat. The algorithm explores a defined range of geometric parameters and automatically generates combinations and performs Computational Fluid Dynamics (CFD) simulations, using the open-source C++ toolbox OpenFOAM. Implemented in a bash script, the algorithm not only automates the simulation process but also provides a geometry of the LCP that is easy to manufacture and cost-effective. The correct value of parameters, such as the distance between the fins bottom surface and the channel base (gap), along with others, has shown a significant impact, leading to a reduction in both the maximum interface temperature (8 K) and the temperature gradient (25 K/m) within the system.