Modeling of a planar SOFC performances using artificial neural network

Abstract

The Planar Solid Oxide Fuel Cell (PSOFC) is one of the renewable energy technologies that is important as the main source for distributed generation and can play a significant role in the conventional electrical power generation. PSOFC stack modeling is performed in order to provide a platform for the optimal design of fuel cell systems. It is explained by the structure and operating principle of the PSOFC for the modeling purposes. PSOFC model can be developed using Artificial Neural Network approach. The data required to train the neural net-work model is generated by simulating the existing PSOFC model in the MATLAB/ Simulink software. The Radial Basis Function (RBF) and Multilayer Perceptron (MLP) neural networks are the most useful techniques in many applications and will be applied in developing the PSOFC model. A detailed analysis is presented on the best ANN network that gives the greatest results on the performances of the PSOFC. The simulation results show that Multilayer Perceptron (MLP) gives the best outcomes of the PSOFC performance based on the smallest errors and good regression analysis.