基于分层网络的电能表运行状态评价

吴 宇; 赵莉; 冉光宇; 李竹

吴宇,赵莉,冉光宇,李竹.基于分层网络的电能表运行状态评价[J].电测与仪表,2024,61(5):218-224.
Wu Yu,Zhao Li,Ran Guangyu,Li Zhu.Evaluation of operation status of electricity meter based on hierarchical network[J].Electrical Measurement & Instrumentation,2024,61(5):218-224.

基于分层网络的电能表运行状态评价

Evaluation of operation status of electricity meter based on hierarchical network

DOI：10.19753/j.issn1001-1390.2024.05.030

中文关键词: 分层模型状态评价 DNDF 运行风险

英文关键词:hierarchical model, status evaluation, DNDF, operational risk

基金项目:国家重点研发计划2017YFB1401702

作者中文名	作者英文名	单位
吴宇	Wu Yu	国网重庆市电力公司营销服务中心
赵莉	Zhao Li	国网重庆市电力公司营销服务中心
冉光宇	Ran Guangyu	重庆广汇供电服务有限责任公司客户服务分公司
李竹	Li Zhu	国网重庆市电力公司营销服务中心

摘要点击次数: 517

中文摘要:

现有的电能表状态评价大都从孤立微观的角度对电能表运行状态进行评估，没有从更大的视角对台区电能表进行整体评价。孤立地对单个电能表进行状态评价，有可能造成对整个台区运行风险的误判。为了更加全面细致掌握电能表及台区的运行状态，文章提出了一种分层网络模型。方法是基于用电采集系统和计量生产调度平台的历史数据，根据各指标的状态对电能表质量，预期寿命和故障率进行评分，依据评分将电能表状态划分为稳定、正常、关注、严重和预警五个等级。基于台区终端记录的电能表异常事件数据，利用深度决策森林（Deep Neural Decision Forests）模型将台区中的异常事件等级程度划分为严重、重要、较重要和一般四个等级，模型分类正确率达到86.58%。实验结果表明，方法能够准确判断电能表的运行状态和整个台区的运行风险情况，为实现电网可靠稳定运行创造了条件。

英文摘要:

The current state evaluation of electricity meters mostly evaluates the operation status of electricity meters from an isolated and micro perspective, and does not conduct overall evaluation of electricity meters in the station area from a larger perspective. Evaluating the status of a single electricity meter in isolation may cause a misjudgment of the operational risk of the entire station area. In order to have a more comprehensive and detailed grasp of the operation status of the electricity meter and the station area, a hierarchical network model is proposed in this paper. The method is based on the historical data of the electricity acquisition system and the metering production scheduling platform. According to the status of each indicator, the quality, expected life and failure rate of the electricity meter are scored. According to the score, the status of the electricity meter is divided into five levels, including stable, normal, concerned, serious and early warning. Based on the abnormal event data of the electricity meter recorded by the terminal in the station area, the deep neural decision forests (DNDF) model is used to classify the abnormal event levels in the station area into four levels: serious, important, more important and general, and the model classification accuracy rate reaches 86.58%. Experimental results show that the proposed method can accurately determine the operating status of the electricity meter and the operational risk of the entire station area, creating conditions for the reliable and stable operation of the power grid.

查看全文查看/发表评论下载PDF阅读器

关闭