基于多源状态融合的智能变电站二次设备状态监测系统设计
集成电路应用
陈令球
国电南京自动化股份有限公司
摘要： 针对智能变电站二次设备状态信息分散、指标量纲不同及单阈值告警易受短时扰动影响的问题，该文设计了一种基于多源状态融合的监测系统。系统统一接入保护测控装置、合并单元、智能终端和过程层交换机的运行、通信、时钟及环境数据，经时间对齐和质量判断后完成异常度转换与健康评价，并通过持续判定、恢复回差和关键故障直通形成分级告警。110 kV半实物平台测试表明，系统状态分级准确率为96.3%，误报率和漏报率均为2.2%，平均告警时延为1.12 s；与单指标固定阈值法和多指标加权评价法相比，状态分级准确率分别提高8.5和3.3个百分点，短时扰动误报次数分别减少75%和50%，可用于二次设备集中监视与异常预警。
中图分类号：TM63文献标识码：ADOI:10.19339/j.issn.1674-2583.2026.010.007
中文引用格式：陈令球. 基于多源状态融合的智能变电站二次设备状态监测系统设计［J］.集成电路应用，2026，43（10）：40-43.
英文引用格式：Chen Lingqiu. Design of a secondary equipment condition monitoring system for smart substations based on multisource state fusion［J］.Application of IC，2026，43（10）：40-43.
中文引用格式：陈令球. 基于多源状态融合的智能变电站二次设备状态监测系统设计［J］.集成电路应用，2026，43（10）：40-43.
英文引用格式：Chen Lingqiu. Design of a secondary equipment condition monitoring system for smart substations based on multisource state fusion［J］.Application of IC，2026，43（10）：40-43.
Design of a secondary equipment condition monitoring system for smart substations based on multi-source state fusion
Chen Lingqiu
Guodian Nanjing Automation Co.
Abstract： A condition monitoring system based on multi-source state fusion is designed for secondary equipment in smart substations to address dispersed state information, heterogeneous indicator scales, and false alarms caused by transient disturbances. The system acquired operation, communication, time-synchronization, and environmental data from protection and control devices, merging units, intelligent terminals, and process-level switches. After time alignment and data-quality checking, anomaly score conversion and health assessment were completed, and graded alarms were generated by continuous confirmation, recovery hysteresis, and critical-fault bypass. Tests on a 110 kV hardware-in-the-loop platform showed a classification accuracy of 96.3%, false and missed alarm rates of 2.2% each, and an average alarm delay of 1.12 s. Compared with the single-indicator fixed-threshold method and the multi-indicator weighted evaluation method, the proposed method improved classification accuracy by 8.5 and 3.3 percentage points, respectively, while reducing transient-disturbance false alarms by 75% and 50%, which is available for centralized mornitoring and anomaly early warning for secondary equipment.
Key words :
引言
智能变电站的保护、测量和控制功能依赖二次设备与过程层网络协同运行，装置资源占用、通信质量及时钟同步状态均可能影响信息传输和保护动作。故障参数映射可建立设备状态与故障类型之间的联系，数字孪生模糊评价和区间型PCA评价为状态分级提供了不同路径［1-3］，云雾边协同运检及物联网远程监测进一步拓展了状态接入与远程运维方式［4-5］。由于多类设备的数据接口、更新周期和量纲不同，数据超时、短时抖动及关键故障仍需在统一监测链路中协调处理。为此，本文构建了多源状态接入、质量控制、健康评价和分级告警链路，并通过半实物平台检验状态识别与告警响应性能。
本文详细内容请下载：
http://www.chinaaet.com/resource/share/2000007337
投稿邮箱：ICAPP@belling.com.cn
作者信息：
陈令球
（国电南京自动化股份有限公司，江苏南京211106）
此内容为AET网站原创，未经授权禁止转载。