使用智能电表实现工业设备的大数据识别  

Big Data Recognition of Industrial Equipment Using Smart Meters

作　　者：张春萌 关艳 ZHANG Chunmeng;GUAN Yan(Liaoning Power Supply Bureau Co., Ltd., Shenyang 110006, China)

机构地区：[1]国网辽宁省电力有限公司,辽宁沈阳110006

出　　处：《微型电脑应用》2020年第11期114-117,共4页Microcomputer Applications

摘　　要：为了实现对工业设备负荷的大数据识别,采用低压智能电表和长短期记忆神经网络(LSTM)构建一个非侵入式工业设备负载监测系统,通过智能电表所采集的电力负荷大数据实现对工业设备的智能标识。具有实用性、价格合理优势的低压智能电表被用于采集负载数据。LSTM神经网络用于处理电力负载数据,标识设备类别。所构建的监测系统中采用了边缘计算技术用于实现并行计算以提高设备识别的效率。实验证明,结果参数调优后,该模型的单个电气设备连续数据的平均识别率可达到83.6%。In order to realize the big data identification of industrial equipment load,this paper uses low-voltage smart meter and long-term and short-term memory neural network(LSTM)to construct a non-intrusive industrial equipment load monitoring system,which realizes industrial equipment through the power load big data collected by smart meter.A low-voltage smart meter with practicality and reasonable price is used to collect load data.The LSTM neural network is used to process electrical load data and identify device categories.Edge computing technology is used in the monitoring system to achieve parallel computing to improve the efficiency of device identification.Experiments show that after the resulted parameters are tuned,the average recognition rate of continuous data of a single electrical device of the model can reach 83.6%.

关 键 词：大数据 设备识别 边缘计算 非侵入性负载监测 智能电表 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]