Reinforcement learning driven equivalent fuel consumption minimization energy management method
By building a framework for equivalent fuel consumption minimization strategy driven by reinforcement learning, combined with deep reinforcement learning SAC algorithm and integrated strategy network, the reliability problem of fuel cell vehicle energy management strategies under complex operating conditions is solved, and intelligent and precise control of fuel cell hybrid vehicles is achieved.
CN120337765APending Publication Date: 2025-07-18BEIJING INST OF TECH
Patent Information
- Application Number
- CN202510455151.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-11
- Publication Date
- 2025-07-18
Smart Images
Figure CN120337765A_ABST
Abstract
The invention provides a reinforcement learning-driven equivalent fuel consumption minimization energy management method, designs a reinforcement learning-guided equivalent fuel consumption minimization method, realizes deep fusion of data driving and model optimization, and overcomes the technical bottlenecks of low reliability and limited model adaptability of a traditional energy management strategy; by constructing an integrated strategy network architecture and an uncertainty perception mechanism and combining an online decision credibility evaluation system, the reliability of uncertainty control under a dynamic working condition is effectively ensured; according to the real-time energy management strategy with working condition adaptability established by the method, the strategy reliability of the vehicle in a complex operation scene can be improved, and intelligent and accurate control over the fuel cell hybrid electric vehicle is facilitated.
Need to check novelty before this filing date? Find Prior Art
Citation Information
Cited By
Method and device for maintaining electric quantity of plug-in hybrid electric vehicle and medium
CN121947452A