Model parameter identification and efficiency-oriented cooperative control of a high-power proton exchange membrane fuel cell for heavy-duty commercial vehicles

To address the difficulties in constructing simulation models and developing control strategies for 130 kW proton exchange membrane fuel cells (PEMFCs) for heavy-duty commercial vehicles, this study conducts full-chain optimization encompassing “Modeling–Identification–Control.” A data-mechanism fusion model is constructed to achieve deep coupling between mechanistic equations and actual operating characteristics, validated by bench tests with stack voltage/cathode flow average relative errors of 0.68% ( R 2 = 0.952) and 1.96% ( R 2 = 0.996), overall error ≤2%. An improved adaptive snake optimization algorithm integrated with the Simulated Annealing Metropolis criterion is proposed, paired with a current-dependent fitness function matching parameter sensitivity differences between stack voltage and cathode inlet flow rate, which outperforms snake optimization/particle swarm optimization by 73.03%/88.32% in fitness and 32.5%/40% in convergence. A novel cross-subsystem efficiency-oriented cooperative control strategy with air-side feedforward proportional–integral + dynamic coupling compensation is presented, which suppresses oxygen excess ratio overshoot, minimizes air compressor speed fluctuation and energy redundancy, and significantly enhances stack voltage stability under dynamic loads. Validated via simulation-bench closed-loop tests, the model and algorithms provide an efficient engineering solution for high-power PEMFC industrial applications, supporting rapid control algorithm iteration and bench calibration with substantial practical value.