Electric vehicle fast charging station charging scheduling method and system based on security reinforcement learning
By introducing a state-by-state security value function and a security reinforcement learning method using Lagrange networks, the problems of grid stability and constraint violations in electric vehicle charging scheduling are solved, and efficient and safe charging scheduling of electric vehicle fast charging stations is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANDONG UNIV
- Filing Date
- 2026-01-27
- Publication Date
- 2026-06-09
AI Technical Summary
Traditional electric vehicle charging control methods lack flexibility and adaptability when the power grid operating conditions change dynamically. Reinforcement learning algorithms have failed to effectively guarantee power grid stability, and existing safety reinforcement learning methods cannot ensure zero-constraint violations.
A safety-based reinforcement learning approach is adopted, which utilizes the state-by-state safety value function of the control barrier function and the state-by-state Lagrange network to quantify the risk of constraint violation and dynamically adjust the penalty term to ensure the safety and efficiency of charging scheduling.
It achieves zero voltage constraint violation during high load periods, reduces peak demand by 7.3-17.5%, and ensures grid stability and maximizes charging benefits.
Smart Images
Figure CN122175192A_ABST