| 引用本文: | 任 佳,晋龙兴,郭乐欣,等.二次设备远程运维管控平台的研发与应用[J].电力系统保护与控制,2026,54(07):129-139. |
| REN Jia,JIN Longxing,GUO Lexin,et al.Development and application of a remote operation and management platform for secondary equipment[J].Power System Protection and Control,2026,54(07):129-139 |
|
| 摘要: |
| 针对电力二次设备远程运维中传统保信系统告警信息杂乱、二次回路故障定位误差大、人工监盘误判率高，且“感知-认知-决策”全链条存在明显短板的行业问题，构建了融合告警信息精确分级分类、设备运行场景智能辨识、告警处置措施智能生成的二次设备智能监盘体系，并研发对应的远程运维管控平台。智能监盘体系以自然语言处理与预训练模型实现告警标准化并建立动态迭代优化机制，依托多维约束场景知识库与规则推理引擎完成设备运行场景辨识，结合多策略检修信息滤除与“大模型 + 检索增强生成(retrieval-augmented generation, RAG)”技术生成告警处置方案。平台涵盖远方监视、巡视、控制、维护等功能，按安全分区架构设计。平台已部署于某省级电网，运行情况良好，推动了二次设备运维从“现场依赖”向“远程智能”转型。 |
| 关键词: 智能监盘 告警分级 场景辨识 远程运维管控 |
| DOI：10.19783/j.cnki.pspc.251188 |
| 分类号: |
| 基金项目:国家重点研发计划项目资助(2021YFB2401000)；南方电网公司科技项目资助(090000KK52210169) |
|
| Development and application of a remote operation and management platform for secondary equipment |
|
REN Jia, JIN Longxing, GUO Lexin, LU Zhengfei
|
|
Shenzhen Power Supply Bureau Co., Ltd., Shenzhen 518000, China
|
| Abstract: |
| To address industry challenges in the remote operation and maintenance (O&M) of power system secondary equipment, such as disordered alarm information in traditional protection and control systems, large errors in secondary circuit fault localization, high misjudgment rates in manual monitoring, and significant deficiencies across the entire “perception-cognition-decision” chain, an intelligent monitoring system for secondary equipment is developed. This system integrates precise alarm classification and grading, intelligent identification of equipment operation scenarios, and automated generation of alarm handling measures, along with the development of a corresponding remote O&M management platform. The intelligent monitoring system standardizes alarm information using natural language processing and pretrained models, while establishing a dynamic iterative optimization mechanism. It relies on a multidimensional, constraint-based scenario knowledge base and a rule-based reasoning engine to achieve accurate identification of equipment operation scenarios. By combining multi-strategy maintenance information filtering with large language models and retrieval-augmented generation (RAG) techniques, the system generates effective alarm handling solutions. The platform supports functions such as remote monitoring, inspection, control, and maintenance, and is designed based on a secure zoning architecture. It has been deployed in a provincial power grid and has demonstrated stable performance, promoting the transition of secondary equipment O&M from “on-site dependence” to “remote intelligence”. |
| Key words: intelligent monitoring alarm grading scenario recognition remote O&M management |
