Enhancing Efficiency and Longevity of PEMFCs Through Robust Non-Linear Control Strategies: A Model-Based Novel Approach

This research addresses the challenge of controlling the air delivery system of Proton Exchange Membrane Fuel Cells (PEMFC) to ensure smooth operation and prolong device lifespan. The study is based on PEMFC model developed by Pukrushpan et al. for the Ford P2000 fuel cell vehicle. The stack temperature has been incorporated in the state-space model to improve its accuracy. Robust non-linear control strategies are designed and compared, including first-order Sliding Mode Control (SMC), second-order SMC, Integral SMC (ISMC), and Fast Terminal Integral SMC (FTISMC) with Super Twisting Algorithm (STA). In addition, a Uniform Robust Exact Differentiator (URED) based observer is proposed for estimating supply manifold pressure, mitigating measurement noise effects by eliminating the need of its direct measurement. Moreover, a three-layered feedforward neural network (TLFFNN) using FTISMC has been employed to estimate the continuous part of the control, trained to enhance robustness against noise. Extensive software simulations are conducted to validate the effectiveness of the proposed controllers. The simulation results demonstrate superior performance in terms of closed-loop stability, tracking accuracy, and robustness against measurement noise and system uncertainties, confirming the capability of the proposed control strategies to achieve reliable and accurate PEMFC air supply control. The results demonstrate that the proposed observer and all the control approaches achieve stable operation. However, FTISMC exhibits fastest convergence, superior tracking accuracy, and greater robustness to parameter uncertainties, external disturbances, and measurement noise. The proposed framework, thus, contributes towards the development of efficient and environmentally sustainable fuel cell technologies. Future research avenues may explore ways to enhance PEMFC robustness using online NN methods.