Automated design of direct methanol fuel cell stacks: a quick optimization

Abstract

This work presents a computer-aided automated design system to quickly optimize the constructive parameters of direct methanol fuel cells (DMFCs) by minimizing a multicriteria design function that includes mass, volume and fuel consumption. In this regard, this work addresses the absence of procedures for DMFC stack design. Four main tasks were performed to develop the design system: stack parameterization, mathematical modelling, design function formulation and implementation of a genetic algorithm as optimization method. To evaluate the capabilities of the developed DMFC automated design system, a case study and a sensitivity analysis were carried out. The continuous decrease of the minimum design function value until it converges to that of the best solution demonstrates a reliable system convergence. In addition, a solution reproducibility of 77% was obtained for 8% mutation and 1500 generations. Thus, the proposed design system permits obtaining the optimum design in a very short period of time.