Research on Energy Management Strategy for Series Hybrid Powertrain Based on Twin Delayed Deep Deterministic Policy Gradient

To improve the fuel economy and performance of hybrid power systems, this chapter proposes an energy management strategy for series hybrid power systems based on the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm, addressing the shortcomings of existing rule-based and optimization-based strategies. By developing mathematical models for key components of the hybrid system, including the engine, motor, and battery, the system’s operation under real-world conditions is simulated. The framework for the TD3-based energy management strategy and the training of the agent are completed to achieve multi-objective optimization of fuel consumption and performance. Simulation experiments are conducted to verify the effectiveness and superiority of the proposed strategy. The results show that, compared to the Deep Q-Network (DQN) algorithm, this strategy effectively improves the overall system efficiency, significantly reducing fuel consumption by 3.3% while meeting performance requirements, providing a feasible approach and method for optimizing energy management strategies in series hybrid power systems.