Nonlinear model predictive control for efficient and robust airpath management in fuel cell vehicles.

The fuel cell airpath multivariable control problem of optimally coordinating the electric compressor motor and the back-pressure valve to achieve efficient and safe conditions, for both steady state and transient operation, has not been completely addressed in the literature. This paper proposes a nonlinear model predictive control strategy, implemented via the Garrett Motion proprietary NMPC toolbox, to regulate the oxygen stoichiometry and the cathode pressure of an automotive fuel cell airpath system, while avoiding compressor surge and air starvation. The controller set-points are optimized, using the nonlinear model, to achieve the maximum system power as a function of the operating stack condition. The effectiveness and robustness of the proposed control strategy have been validated by means of a simulated World harmonized Light-duty vehicles Test Cycle (WLTC), under both state feedback and model parameters uncertainties. [Display omitted] • MPC ensures compressor surge and oxygen starvation avoidance under severe dynamic scenarios. • Controller robustness assessed via a WLTC regulatory cycle under noisy and uncertain conditions. • Fuel efficiency is not significantly sensitive with respect to different controller calibrations. [ABSTRACT FROM AUTHOR]