Design of actively-cooled microvascular materials: a genetic algorithm inspired network optimization
The design of a microvascular flow network embedded in an actively-cooled polymeric material is presented. A multi-objective Genetic Algorithm (GA) combined with the finite element method is first used to determine the quasi-optimized network configurations and provide insight into the behavior of the actively-cooled material. The objective functions and constraints involve improving the flow efficiency and minimizing the void volume fraction of the material, while maintaining an allowable temperature in the system. A periodic configuration is adopted for the embedded network based on the results of this study. We then solve an optimization problem at a considerably lower computational cost to improve the optimized network configuration. To determine the final design, we implement the information obtained from the GA optimization to describe the geometry of the embedded network in a simple mathematical form and conduct a parameter study to evaluate its optimized shape.