专用变压器异常用电检测方法研究

Research on Abnormal Power Consumption Detection Method of Special Transformer

专用变压器是电力系统中连接不同电压等级的关键设备,为了保证变压器的安全稳定运行,设计专用变压器异常用电检测系统,实时监测和采集变压器运行参数。使用STM32F103RBT6芯片作为主控单元作为专用变压器监测终端,使用红外温度传感器采集环境参数和变压器温度。系统基于图卷积神经网络构建异常检测模型,并使用有向水平可视图算法转换专用变压器的时序数据,通过特征池化提取有向水平图的特征向量,经过卷积层和池化层操作后得出异常检测结果。实验结果表明,系统模型的异常检测准确率最大为97.5%,模型运行时间最短为812 ms,检测时间最快为1224 ms。

The special transformer is the key equipment connecting different voltage levels in the power system.In order to ensure the safe and stable operation of the transformer,this paper designs a special transformer abnormal power consumption detection system to monitor and collect transformer operation parameters in real time.The STM32F103RBT6 chip is used as the main control unit as the special transformer monitoring terminal,and the infrared temperature sensor is used to collect environmental parameters and transformer temperature.The system builds an anomaly detection model based on graph convolutional neural network,and uses the directed horizontal visualizable algorithm to transform the time series data.Feature vectors of directed level graph are extracted through feature pooling,and anomaly detection results are obtained after operation of convolution layer and pooling layer.Experimental results show that the maximum accuracy of anomaly detection is 97.5%,the shortest running time is 812 ms,and the fastest detection time is 1224 ms.