Dynamic performance improvement of proton exchange membrane fuel cell system by robust loop shaping and artificial intelligence optimized fractional order PI controllers

Due to inherent nonlinear behavior, the Proton Exchange Membrane (PEM) Fuel Cell system yields in poor power quality. The sudden change in stack current causes uncertainty in PEMFC behavior resulting in degraded output. This necessitates the control of supply manifold pressure on the cathode side to prevent oxygen starvation. In this work, three techniques, namely, robust loop shaping, smith predictor, and Fractional Order Proportional Integral (FOPI) controllers, are designed after incorporating parametric uncertainty in system description. Further, the Particle Swarm Optimization (PSO) artificial intelligence technique is implemented to optimize the parameters of smith predictor and Fractional Order PI controllers. By stabilizing the system across the entire parameterized uncertainty range, the proposed control strategies, namely, loop shaping, PSO-Smith Predictor, and PSO-FOPI controller effectively compensate uncertainty effects. The simulation result shows that the PSO-tuned FOPI controller has best overshoot at operating points 3, 4, and 5 having a value of 0.505% in comparison with PSO-Smith Predictor where overshoot (ranging from 19.88% to 40.141%) is too high; and loop shaping controller having high values of overshoot (more than 40% at every operating point) Thus, PSO-FOPI controller exhibits the best performance with far superior overshoot and steady-state stability.