A predictive model of solid oxide fuel cell stacks for thermal mangement

This paper presents a predictive model of solid oxide fuel cell (SOFC) stacks for thermal management by using a support vector machine (SVM). The operating temperature of the SOFC stack is the most important variable controlled for the generation system. To carry out the control research on the stack thermal management, the predictive model of the stack temperature must be established. The SOFC stack is a nonlinear, multi-variable system that is hard to model by conventional methods. A predictive model of the stack temperature based on the least squares support vector machine (LS-SVM) with the radial basis function (RBF) is presented, which is a powerful tool to predict how a SOFC stack will behave under different operating conditions. Checked by the experimental data, the model can be established fast and the predicting accuracy is high, which applies to the research on the online predictive control strategy.