摘要
电力感知数据是一种时间序列数据。电力系统在运行中产生大量的数据,导致在查询数据频繁项时的系统负载大、查询效率低。为了提高电网电能传输质量,对基于时域特征的电力感知数据频繁项查询方法进行了研究。构建电力系统拓扑结构,预测其运行状态。根据电网的纵向连续性和横向连续性,定向采集传输活跃的电力感知数据。以三轴加速度相同的两个滑动窗口为目标区域,提取电力感知数据的时域特征。将提取的时域特征输入到贝叶斯分类算法中,使用贝叶斯分类器进行分类。考虑由谐波震颤效应导致的逻辑疏密性下降问题,通过锁定频次占比较大的类别区间,实现电力感知数据频繁项查询。试验结果表明,所提方法的查全率高于97%、查询的电力感知数据在180个以上、时间开销为5 ms、内存开销为10 MB。该方法可有效提升电力感知数据频繁项的查询性能和查询效率,提高电网电能传输质量。
Power sensing data is a kind of time series data. The power system generates a large amount of data in operation, which leads to high system load and low query efficiency when querying the frequent items of data. To improve the quality of power transmission in the power grid, the query method for frequent items of power-aware data based on time-domain features is investigated. The power system topology is constructed to predict its operation state. According to the vertical continuity and horizontal continuity of the power grid, the transmission-active power sense data are collected directionally. Two sliding windows with the same three-axis acceleration are taken as the target area to extract the time-domain features of the power sensing data. The extracted time-domain features are input into a Bayesian classification algorithm and classified using a Bayesian classifier. Considering the logical sparsity degradation problem caused by the harmonic tremor effect, the frequent item query of power sensing data is realized by locking the category intervals with a larger frequency share. The experimental results show that the proposed methods’s query rate is higher than 97%, the query of power sensing data is more than 180 items, the time overhead is 5 ms, and the memory overhead is 10 MB.The proposed method can effectively improve the query performance and query efficiency of the frequent items of power sensing data and improve the quality of power transmission in the power grid.
作者
汪江
温炜
WANG Jiang;WEN Wei(Training Center,State Grid Ningxia Electric Power Co.,Ltd.,Yinchuan 750001,China;School of Computer Science and Technology,Huazhong University of Science and Technology,Wuhan 430074,China)
出处
《自动化仪表》
CAS
2023年第12期80-84,共5页
Process Automation Instrumentation
基金
宁夏回族自治区科技攻关(支撑)计划基金资助项目(2020BEA21B06)。
关键词
电力感知数据
能量管理系统
时域特征
贝叶斯分类算法
频繁项查询
谐波震颤效应
Power sensing data
Energy management system
Time-domain features
Bayesian classification algorithm
Frequent term query
Harmonic tremor effect
