孪生神经网络技术在电力设备状态分析中的应用

Application of Twin Neural Network Technology in State Analysis of Power Equipment

电力设备状态影响因素较多,难以提取电力设备状态特征数据,导致电力设备状态感知与故障预警难度增加,为此,本文研究了孪生神经网络技术在电力设备状态分析中的应用。首先,采集电力设备运行数据,将数据录入孪生神经网络模型中,训练电力设备状态数据,并提取训练后的电力设备状态特征数据。其次,挖掘数据集合,提取电力设备异常信号,根据空间中信号的密度与状态对数据进行聚类与故障映射。最后,引进物联网感知技术,结合电力设备的常态化运行条件设计设备运行安全阈值,实现对电力设备故障状态的诊断与预警。以某地区大型发电厂为例设计对比实验,实验证明,本文设计的电力设备状态分析方法不仅可以实时感知设备运行中的故障状态,还可以实现对故障的精准预警。

There are many factors that affect the status of power equipment,making it difficult to extract the characteristic data of power equipment status,which increases the difficulty of power equipment status perception and fault warning,therefore,this paper studies the application of twin neural network technology in power equipment status analysis.Firstly,collect power equipment operation data,input the data into a twin neural network model,train the power equipment status data,and extract the trained power equipment status feature data.Secondly,mining data sets,extracting abnormal signals from power equipment,clustering and fault mapping data based on the density and status of signals in space.Finally,introduce IoT sensing technology,design equipment operation safety thresholds based on the normalized operating conditions of power equipment,and achieve diagnosis and early warning of power equipment fault status.Taking a large power plant in a certain region as an example,a comparative experiment was designed to demonstrate that the power equipment status analysis method designed in this paper can not only provide real-time perception of fault status during equipment operation,but also achieve accurate warning of faults.