基于边缘计算的电网边缘侧设备缺陷智能识别模型研究

Research on Intelligent Recognition Model of Grid Edge-side Equipment Defects Based on Edge Computing

目前,在电力设备的日常巡检和试验中,积累了大量关于设备故障情况的记录,缺乏相应的故障处理措施。传统集中式数据处理模型无法支持当前的强大电网系统。针对这一缺陷,文章提出了一种基于热点数据的电网边缘侧设备缺陷智能识别模型,模型分为云中心层、边缘层和现场层。在现场层,使用动态时间规整(Dynamic Time Warping,DTW)补充算法补充传感器数据;采用树突神经元模型(Dendritic Neuron Model,DNM)在边缘层进行故障初等分类,并将分类结果上传至云中心层;在云中心层利用数据之间的相关性实现故障分类。最后在公开数据集上进行设备缺陷识别模型验证,验证了模型的有效性和可行性。

At present,in the daily inspection and testing of power equipment,a large number of records of equipment failures have been accumulated,and there is no corresponding fault treatment measures.The traditional centralized data processing model centered on the cloud computing model has been unable to support the current power IoT system.Aiming at this defect,an intelligent identification model of grid edge-side equipment defects based on hot spot data is proposed.The system model is divided into cloud center layer,edge layer,and field layer.In the field layer,the DTW complement algorithm is used to complement the sensor data,the DNN algorithm is used to implement the fault primary classification in the edge layer,and the classification result is uploaded to the cloud center layer,the correlation between the data is used in the cloud center layer to realize the fault classification.After that,an experiment is designed to verify the efficacy of the modified m odel.