RBF neural network based adaptive constrained PID control of a solid oxide fuel cell

Abstract

This paper proposes a radial basis function (RBF) neural network based adaptive constrained PID control scheme for a solid oxide fuel cell (SOFC). First, a RBF neural network is designed for identification the dynamic model of SOFC. The Jacobian information can be obtained through the identification RBF model. Then, an on-line PID parameters tuning algorithm is designed by gradient descent method. At same time, in order to solve the input saturation problem, we design an anti-windup compensator for accommodate the reference. Finally, the simulation results on the dynamic model of SOFC are provided to demonstrate the effectiveness of the proposed constrained control approach.