Energy management strategy for electrically-powered hydraulic vehicle based on driving mode recognition

The effectiveness of the energy management strategy directly impacts the overall system performance of a vehicle, particularly under various driving modes. This paper proposes a novel electrically-powered hydraulic vehicle that integrates a hydraulic transmission system with an electric powertrain. A rule-based energy management strategy is developed to validate the feasibility of the system through steady-state simulation. To enhance system performance, Random Forest and gradient boosting tree algorithms are employed for velocity feature dimensionality reduction, while K-means clustering is used to segment driving modes. Subsequently, a genetic algorithm-optimized backpropagation neural network enables precise online driving mode recognition, and a fuzzy controller actively regulates energy flow in real time. Experimental results indicate that GBF-EMS achieves a final state of charge of 78.77%, reducing battery energy consumption by 16.26% compared to RB-EMS, 8.92% compared to RF-EMS and 2.52% compared to PMP-EMS. This study provides new insights into the further development and optimization of electro-hydraulic power systems.