A genetic algorithm-based topology optimization method for convective cooling of a heating surface with multiple-peak heat flux

The heat-generating surface with a multiple peak heat flux is commonly encountered in many thermal devices, including electronic components, PV/T panels, Lithium-ion battery packs of electric vehicles, etc. If not properly cooled, the uneven heating condition would cause the existence of local hot spots in these devices, resulting in lowered working performance and shortened service time. The convective cooling of the heating surface using a single-phase fluid-based heat sink is often considered as an effective thermal management method to handle this issue due to its simplicity of design, high cooling performance, and compact size. In particular, the cooling performance improvement by optimizing the fluid flow configuration in the heat sink has become an important area of research.

This study presents the development of a genetic algorithm based topology optimization (GATO) method for convective cooling of a heating surface under multiple-peak heat flux. In more detail, the middle area of the heat sink receiving heat flux is treated as the design area and represented as a 50 x 50 binary matrix. Each element in the matrix is considered either as fluid or as solid, and their allocation is optimized with the objective of minimizing the peak temperature at the heating surface of the heat sink under the constraint of constant void volume for the fully-connected fluid domain. For each optimization step, the fluid flow and temperature characteristics are obtained by CFD simulation using OpenFoam and the GA operations (selection, crossover, mutation, etc.) are applied.

The numerical results obtained show that the proposed GATO method could successfully determine the optimal distribution of the fluid/solid elements in the design area, leading to the minimized peak temperature of the heating surface at the cost of a reasonable pressure drop increase. Compared to a conventional parallel straight mini-channel heat sink, better cooling performance (lower peak temperature, higher temperature uniformity) can be achieved by the topologically optimized heat sink.

A design parametric study has also been performed and the effects of some influencing factors (porosity, fluid inlet velocity, matrix dimension, etc.) on the performance of the topologically optimized heat sink are evaluated. The testing and performance comparison of different heat sink prototypes are our on-going work, in order to experimantally validate the developed GATO method. This study could be a useful contribution to the optimization of heat exchangers/heat sinks in general.

Work In Progress

Contributeurs
Yijun Li
Stephane Roux
Cathy Castelain
Lingai Luo
Yilin Fan
Contact
yijun.li@univ-nantes.fr
Groupe thématique
Mots-clés
Topology optimization
Genetic algorithm
Convective cooling
Heat Exchanger