Hydrogen Energy paper 10

Abstract

Since the computing time is very large to extend the PID controller to the intelligent fuzzy-neural one directly, and in order to make the steady state error being zero, this paper proposes a MIMO fuel cell design using a hybrid controller with Ziegler-Nicholsbased integrator and fuzzy-neural-based PD-type controller. The first step is to find a Ziegler-Nichols-based PID-type controller. The second one is to apply a hybrid controller with conventional (i.e. Ziegler-Nichols-based) integrator and PD-type fuzzyneural controller. Besides, to save the computing time this paper also applied an algorithm with GD (gradient descent), SCG (scaled conjugate gradient) and LM (Levenberg-Marquardt) optimization methods alternatively to determine the neural network weighting factors and biases. Comparing with the other methods by Matlab simulation, the proposed system performances are better and robust even for the system with stack current disturbance variation and gas valve's hysteresis effect.