Research and Optimization of Predictive Energy Management Strategy Based on Front Powertrain Phase Characteristics

To enhance the energy management performance of series mild-hybrid tracked vehicles under complex working conditions and address issues such as delayed power response and low energy utilization efficiency, this paper proposes an energy management strategy based on front power chain phase characteristic prediction and Nonlinear Model Predictive Control (NMPC). The strategy employs a Long Short-Term Memory (LSTM) neural network to predict future power demand, introduces a phase-domain switching mechanism to improve the dynamic responsiveness of the engine-generator set, and utilizes an NMPC framework to achieve rolling optimization of power allocation and dynamic system coordination. Simulations under typical off-road cycle conditions demonstrate that, compared to traditional rule-based control and the Equivalent Consumption Minimization Strategy, the proposed strategy exhibits significant advantages in vehicle speed tracking, power response, fuel economy, battery utilization efficiency, and system stability.