Research on the RBF control strategy based on dual-layer particle swarm optimization in a novel electro-mechanical-hydraulic power coupling vehicle

Electro-Hydromechanical Power Coupling Vehicle (EHPCV) integrates electrical, mechanical, and hydraulic energy through an Induction Asynchronous Mechanical-Electrical-Hydraulic Power Coupling (IA-MEHPC), significantly enhancing energy efficiency and reducing the motor's peak load. However, the existing Rule-Based energy management strategies exhibit limited adaptability under complex driving conditions, thereby restricting the performance and energy-saving potential of EHPCV. To address this limitation, this paper proposes a Radial Basis Function (RBF) neural network control strategy combined with Particle Swarm Optimization (PSO) for dual-layer optimization (DL-PSO-RBF). This approach utilizes the nonlinear mapping and data-driven characteristics of the RBF neural network to optimize energy management and mode switching. In this study, a model for the EHPCV system is established, and the RBF network is used to optimize the selection of operating modes, automatically adjusting control parameters to reduce the peak torque required by the motor and improve the efficiency of battery SOC management. The simulation results demonstrate that under the newly constructed Feature-Extracted Driving Cycle, the DL-PSO-RBF-EMS achieves improvements in energy efficiency of 11.79 % and 1.96 % compared to the EV and Rule-Based-EMS, respectively. Furthermore, this strategy significantly enhances torque smoothness and improves vehicle dynamic performance. This research provides valuable insights for the development of efficient and sustainable hybrid vehicle technologies.