Optimization Control of Energy Management for Diesel-Electric Hybrid Trains Based on MPC-PMP

Appropriate energy management strategies can enhance the efficiency of train hybrid power systems and reduce fuel consumption costs. This paper addresses the problem of requiring global operating conditions for the Pontryagin Minimum Principle (PMP) strategy and proposes an energy management strategy based on model predictive control (MPC), with the goal of minimizing fuel consumption. This strategy can be implemented online. Firstly, a speed prediction model is established using Long Short-Term Memory (LSTM) neural network. Then, the PMP algorithm is applied for rolling optimization within the time domain of speed prediction, solving the optimal power control sequence at each moment. Finally, based on actual operating conditions, the energy management results of the MPC-PMP strategy, offline PMP strategy, and conventional threshold method are compared. Simulation results demonstrate that although the fuel economy of the MPC-PMP strategy is slightly lower than that of the offline PMP strategy, it improves fuel economy by 14.45% compared to the traditional threshold method, while meeting real-time requirements.