基于深度强化学习的分布式能源智能体调控方法及系统

By employing deep reinforcement learning and perturbation immunity mechanisms, the problem of joint changes in operational data and network constraint data in distributed energy systems is solved. This enables the identification and management of unknown and high-risk perturbations, improves the stability and adaptability of the system, and ensures the safety and optimization effect of the control process.

CN122001026BActive Publication Date: 2026-07-17北京集联软件科技有限公司
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
北京集联软件科技有限公司
Filing Date
2026-04-08
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

Existing technologies struggle to simultaneously characterize the joint changes in operational data and network constraint data within distributed energy systems. They also lack effective mechanisms for identifying unknown and high-risk disturbances, leading to excessive control actions and insufficient stability in strategy updates, making it difficult to meet the requirements for safe and coordinated control in complex operating scenarios.

Method used

By employing deep reinforcement learning and perturbation immunity mechanisms, a perturbation fingerprint vector and perturbation level identifier are generated by constructing a reinforcement learning environment state representation vector. Combined with the immune affinity score sequence, this enables the coordinated control of distributed energy intelligent agents, which has the advantages of high security, strong adaptability and good stability.

Benefits of technology

It enables a fine characterization of the type and intensity of disturbances in distributed energy systems, improves the system's perception accuracy and adaptability to complex operating environments, ensures stable system operation, and avoids out-of-bounds control actions by exploring trigger judgment and contract verification mechanisms, thereby reducing the impact of operational disturbances and achieving continuous optimization decision-making.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122001026B_ABST
    Figure CN122001026B_ABST
Patent Text Reader

Abstract

本发明公开了基于深度强化学习的分布式能源智能体调控方法及系统，包括如下步骤：采集微网或园区配电网络中的分布式能源运行数据与网络约束数据，并执行预处理；构建强化学习环境状态表示向量；配置分布式能源智能体集合；获取扰动抗原指纹条目集合，生成免疫亲和度分值序列；生成常态调控动作指令集合并下发执行或进入微扰动采样流程；生成可行动作域和微扰动动作候选集合，并执行探索控制契约验证；选取目标微扰动动作并下发执行或者生成修正微扰动动作并下发执行；对深度强化学习策略网络执行更新。本发明采用深度强化学习与扰动免疫机制，实现分布式能源智能体协同调控，具备安全性高、适应性强与稳定性好的优点。
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Comprehensive energy system dynamic scheduling method based on multi-agent deep reinforcement learning

    CN119904057A

  • Adaptive distributed new energy power generation equipment cooperative control method and system

    CN119965994A