Channel shape optimization of solid oxide fuel cells using advanced numerical techniques

Shape optimization of fuel channels in solid oxide fuel cells (SOFC) is presented. A three-dimensional, implicit, multi-species fuel cell solver is developed to compute the baseline solution. Sensitivity derivatives required for the optimization are computed using the discrete adjoint method. A gradient-based optimizer is utilized to update the design variables. The geometry of the fuel channel is parameterized to obtain meaningful design variables using a recently developed technique that is applicable for multi-disciplinary design optimization. During the design process, the mesh is continually modified to reflect the changes in the underlying geometry. An automated environment to sequentially execute the aforementioned processes is developed. Cost functions representating the cell voltage and uniform distribution of fuel among all channels are optimized with respect to the shape of the fuel channels.