Artificial neural network assisted multi-objective optimization of a methane-fed DIR-SOFC system with waste heat recovery

Abstract

The main purpose of this study is to enhance the performance of solid oxide fuel cell systems. For this purpose, a mathematical model of a direct internal reforming (DIR) methane-fed solid oxide fuel cell system with waste heat recovery was designed in the engineering equation solver program. We optimised the performance of the solid oxide fuel cell using a genetic algorithm and TOPSIS technique considering exergy, power, and environmental analyzes. An ANN working with the Levenberg-Marquardt training function was designed in the MATLprogram to create the decision matrix to which the TOPSIS method will be applied. According to the power optimization, 786 kW net power was obtained from the system. In exergetic optimization, the exergy efficiency was found to be 57.6%. In environmental optimization, the environmental impact was determined as 330.6 kgCO2/MWh. According to the multi-objective optimization results, the exergy efficiency, the net power of the solid oxide fuel cell system, and the environmental impact were 504.1 kW, 40.08%, and 475.4 kgCO2/MWh.