The real-time performance bottleneck of energy management strategies (EMS) based on model predictive control (MPC) severely restricts their vehicle-grade deployment in series-parallel plug-in hybrid electric vehicles (SPPHEVs). This research develops a real-time adaptive-mode MPC (RTAM-MPC) designed to jointly minimize fuel consumption, electricity usage, and battery aging under strict vehicle-grade execution constraints. An adaptive framework is established by integrating driving pattern recognition (DPR) with MPC, which dynamically adjusts the prediction time grid, solver initialization, and speed prediction configurations. To ensure computational efficiency suitable for embedded systems, a fast numerical optimization method is proposed, alongside a DPR-guided speed prediction model based on a coyote optimization algorithm-optimized kernel extreme learning machine. The results show that RTAM-MPC achieved 98.54% dynamic programming (DP) performance. Compared to the equivalent consumption minimization strategy (ECMS), it demonstrated a 5.37% improvement in economic efficiency and a 24.67% reduction in battery aging. Compared to standard MPC, the average computation time is 11.07 ms, a decrease of 94.48%.
Pan et al. (Sat,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: