
基于边缘计算的电网主设备状态实时监测方法
Real-time monitoring method of power grid main equipment status based on edge computing
针对电网主设备状态实时监测性能和异常检测准确率低的问题,构建了一个基于边缘计算的电网主设备状态实时监测系统。该系统通过传感器模块对电网主设备状态进行实时数据采集;基于即时定位与地图创建(Simultaneous Localization and Mapping,SLAM)对电网主设备进行监测;边缘节点模块利用小波去噪对采集的数据进行处理,采用局部异常因子(Local Outlier Factor,LOF)算法对数据进行异常值检测;网络节点模块利用消息队列遥测传输协议(Message Queuing Telemetry Transport,MQTT)将处理后的结果值传输到管理平台;通过管理平台模块的分析展示、告警处理和权限设定实现对电网主设备状态的实时检测和预警。实验结果表明,该系统采用LOF算法对数据进行异常值检测,其误报率为1.5%,检测率为98.5%,准确率能达到98.6%。采用MQTT协议传输数据的平均时延为31.52 ms,到报率能达到99.78%,具有较好的实用性。
Aiming at the problems of low real-time monitoring performance and anomaly detection accuracy of power grid main equipment,a real-time monitoring system of power grid main equipment based on edge computing is constructed.The system collects real-time data of the state of power grid main equipment through the sensor module.Simultaneous Localization and Mapping (SLAM) is used to monitor the main equipment of power grid.The edge node module uses wavelet denoising to denoise the collected data,and uses the Local Outlier Factor (LOF) algorithm to detect the data outliers.The network node module uses Message Queuing Telemetry Transport (MQTT) to transmit the processed result value to the management platform.Through the analysis and display,alarm processing and permission setting of the management platform module,the real-time detection and early warning of the power grid main equipment status are realized.The experimental results show that the system uses LOF algorithm to detect outliers.The false positive rate is 1.5%,the detection rate is 98.5%,and the accuracy rate can reach 98.6%.The average time delay of data transmission using MQTT protocol is 31.52 ms, and the transmission rate can reach 99.78%,which is highly practical.
边缘计算 / 电网主设备 / 实时检测 / 小波去噪 / LOF算法 / MQTT协议 {{custom_keyword}} /
edge computing / power grid main equipment / real-time detection / wavelet denoising / LOF algorithm / MQTT protocol {{custom_keyword}} /
薛士然 {{custom_editor}},
表1 部分真实数据集Table 1 Part of the real data set |
时间 | 温度转换电压/V | 振动电压/mV | 输出信号电压/V |
---|---|---|---|
8:20 | 2.80 | 80.12 | 1.26 |
8:30 | 2.90 | 102.33 | 0.96 |
8:40 | 2.96 | 136.57 | 1.13 |
8:50 | 2.92 | 128.32 | 1.05 |
9:00 | 2.87 | 122.71 | 1.02 |
9:10 | 4.03 | 188.55 | 2.55 |
9:20 | 4.05 | 224.64 | 2.78 |
表2 3种协议实验结果Table 2 Experimental results of the three protocols |
协议 | 最大时延/ms | 最小时延/ms | 平均时延/ms | 到报率/% |
---|---|---|---|---|
MQTT | 355.76 | 0.36 | 31.52 | 99.78 |
CoAP | 662.52 | 2.48 | 52.88 | 96.15 |
DDS | 1250.31 | 8.52 | 115.27 | 94.23 |
表3 3种方法的传输效果Table 3 Transmission effects of the three methods |
序号 | 时间/ min | 期望传 输量/GB | 参考文献 [3]方法/GB | 参考文献 [4]方法/GB | 本文 方法/GB |
---|---|---|---|---|---|
1 | 10 | 356 786 | 256 745 | 306 701 | 356 732 |
2 | 10 | 245 663 | 198 601 | 215 601 | 245 645 |
3 | 10 | 432 352 | 322 345 | 398 311 | 432 321 |
4 | 10 | 356 773 | 236 701 | 307 121 | 356 723 |
5 | 10 | 457 854 | 307 801 | 397 812 | 457 834 |
6 | 10 | 457 745 | 323 701 | 407 745 | 457 712 |
7 | 10 | 534 663 | 302 631 | 494 643 | 534 634 |
8 | 10 | 453 455 | 253 401 | 393 451 | 453 419 |
9 | 10 | 346 645 | 142 145 | 306 611 | 346 601 |
10 | 10 | 346 645 | 246 601 | 306 632 | 346 621 |
[1] |
徐昊, 肖勇, 马雪菲. 基于灰色关联和层次分析的电网故障诊断[J]. 中国科技信息, 2023(14):101-104.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[2] |
叶飞, 王来善, 张静鑫, 等. 基于关联规则挖掘的电网设备差异化状态检测[J]. 信息技术, 2023, 47(4):157-160,166.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[3] |
高源, 张振宇, 罗翔, 等. 电网设备状态在线监控系统的研究[J]. 电子器件, 2021, 44(5):1183-1189.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[4] |
李岩. 基于多源数据融合感知的电网设备供应链管控与预警算法设计[J]. 电子设计工程, 2023, 31(12):169-172.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[5] |
高骞, 杨俊义, 刘凯, 等. 基于数据驱动技术的配电网拓扑结构及线路参数识别方法[J]. 电力电容器与无功补偿, 2023, 44(2):77-87.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[6] |
张亚芳. 基于四线制PT100的高精度温度采集系统[J]. 佳木斯大学学报(自然科学版), 2023, 41(2):162-165.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[7] |
王玲. 用冲击法校准压电式力传感器的灵敏度[J]. 上海计量测试, 2023, 50(2):43-46.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[8] |
刘伟, 赵玄玉, 查明虎. 基于霍尔传感器的工业机器人轴电流监测研究[J]. 湖北民族大学学报(自然科学版), 2022, 40(3):289-292.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[9] |
姬鑫. 基于信号分析的三电平逆变器开路故障诊断方法研究[D]. 西安: 西安工业大学, 2022.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[10] |
尤向前, 高盼. 基于Hadoop与局部异常因子的用电行为检测方法[J]. 科技创新与应用, 2023, 13(21):73-76.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[11] |
孙友林. 基于MQTT协议的物联网云平台设计与实现[J]. 物联网技术, 2023, 13(6):71-72.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[12] |
任雪峰, 武志伟. 一种电源监控管理模块的实现方法[J]. 雷达与对抗, 2022, 42(4):65-68.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_ref.label}} |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
/
〈 |
|
〉 |