摘要
采用FCM算法、改进模糊c均值聚类算法进行电力计量设备故障监测时缺乏约束规则,导致监测效果较差,为此提出基于图像深度学习的电力计量设备故障监测方法。构建CNN模型,确定深度神经网络权值最优解,避免出现过拟合现象。经过前向传播和反向传播网络训练,不断更新权值,经过图像预处理后识别故障。设置关联规则,结合抄读电能表相关数据,设计故障监测流程。以电能表为例,通过拓扑结构进行仿真实验分析,由实验结果可知,该方法电压和电流监测曲线与实际值曲线基本吻合,具有精准的监测效果,能够更好地保障电力计量装置发挥好其自身作用。
The FCM algorithm and the improved fuzzy c-means clustering algorithm are lack of constraint rules in the fault monitoring of power metering equipment,resulting in poor monitoring effect.Therefore,a fault monitoring method of power metering equipment based on image deep learning is proposed.In order to avoid over fitting phenomenon,CNN model is constructed to determine the optimal weight solution of depth neural network.After forward propagation and back propagation network training,the weights are updated continuously,and the faults are identified after image preprocessing.Set up the association rules,and design the fault monitoring process combining with the relevant data of reading electricity meter.Taking the electric energy meter as an example,through the simulation experiment analysis of the topological structure,it can be seen from the experimental results that the voltage and current monitoring curve of the method is basically consistent with the actual value curve,which has the precise monitoring effect and can better guarantee the power metering device to play its own role.
作者
张海永
ZHANG Haiyong(Nanjing TAISIDE Smart Electric Co.,Ltd.,Nanjing 211100,China)
出处
《电子设计工程》
2021年第9期103-106,111,共5页
Electronic Design Engineering
基金
贵州电网有限责任公司科技项目(066600KK52180018)。
关键词
图像深度学习
电力计量设备
故障监测
卷积神经网络
imagedeeplearning
powermeteringequipment
faultmonitoring
convolutionalneuralnetwork
作者简介
张海永(1985—),男,江苏南通人,中级工程师。研究方向:电力系统自动化及人工智能。
