基于多业务融合分析的配变量测数据完整性异常辨识技术  

Abnormal Identification Technology for Measurement Data Integrity of Distribution Transformer Based on Multi-business Fusion Analysis

作　　者：陈锦铭 陈烨 韦磊 赵新冬 车伟 焦昊 凌绍伟 周祉君 CHEN Jinming;CHEN Ye;WEI Lei;ZHAO Xindong;CHE Wei;JIAO Hao;LING Shaowei;ZHOU Zhijun(State Grid Jiangsu Electric Power Co.,Ltd.,Research Institute,Nanjing 211103,Jiangsu Province,China;State Grid Jiangsu Electric Power Co.,Ltd.,Nanjing 210024,Jiangsu Province,China;School of Electrical Engineering,Southeast University,Nanjing 210096,Jiangsu Province,China)

机构地区：[1]国网江苏省电力有限公司电力科学研究院,江苏省南京市211103 [2]国网江苏省电力有限公司,江苏省南京市210024 [3]东南大学电气工程学院,江苏省南京市210096

出　　处：《电力信息与通信技术》2023年第10期41-47,共7页Electric Power Information and Communication Technology

基　　金：国家重点研发计划项目“面向超大规模电网设备节点的图计算分析与优化软件”(2022YFB2404200)。

摘　　要：配变量测数据具有较高的业务价值,然而受多因素影响,其完整性存在少量瑕疵。传统的状态估计、插值等补全技术忽略了配变停电所导致的数据缺失,将影响业务分析的正确性。文章在营、配、调多源数据融合基础上,提出了配变量测完整性异常辨识框架及方法。首先,基于配变数据缺失的片段生成断点碎片,在抖动异常过滤后结合相似日电量增长率趋势进行初步研判。其次,根据时空关联特征将断点碎片聚合为断点事件。最后,结合停电、线损等配电网核心业务视角对断电事件进行校核,完成由于停电所导致配变数据缺失的标注。实际案例证明了上述方法的有效性。该技术已在江苏工程化应用,显著提升了配变数据质量,同时形成了以用促治的良好互动模式,推动了线损、可靠性管理等存量业务系统的深化应用。The measurement data of distribution transformers has high business value,but due to multiple factors,there are a few flaws in their integrity.Traditional data completion techniques such as state estimation and interpolation ignore the data loss caused by power outage in distribution transformers,which will affect the accuracy of business analysis.On the basis of multi-source data fusion in operation,distribution,and dispatch,this paper proposes a framework and method for identifying integrity anomalies in distribution transformer data.Firstly,breakpoint fragments are generated based on the missing fragments of distribution transformer data,and preliminary analysis is conducted after filtering for jitter anomalies,based on the trend of electricity growth rate on the similar day.Secondly,breakpoint fragments are aggregated into breakpoint events according to the spatiotemporal correlation characteristics.Finally,based on the core business perspectives of distribution network such as power outage and line loss,the power outage event is verified and the missing distribution transformer data caused by power outage is annotated.Practical cases have proven the effectiveness of the above methods.This technology has been applied in engineering in Jiangsu,significantly improving the quality of distribution transformer data.Furthermore,a good interactive model of promoting governance through use has been formed,promoting the deepening application of existing business systems such as line loss and reliability management.

关 键 词：配变量测 数据完整性 多业务融合 

分 类 号：TN915.853[电子电信—通信与信息系统]