基于批量匹配算法模型的配电网故障诊断方法
2022,30(8):19-24
摘要：配电网发生故障后线路电流增加、电压迅速下降,对配电网的电能质量和供电可靠性产生影响,本研究建立配电网故障诊断系统,利用采集到的线路故障信息迅速发现并定位故障位置。采集断路器跳闸前两个周波的变压器低压侧相电压波形,并进行傅里叶加窗变换计算基波的有效值,分析发生故障后低压侧相电压的变化规律。系统应用基于多语义交互的批量匹配算法,利用改善后的编码器进行故障文本信息的编码。实验结果显示本研究系统的故障诊断模型的训练速度较快,故障诊断准确率最高达到100%,语言推送任务测试中,SNL数据集的准确率为100%。
关键词：配电网故障;故障定位;变压器低压侧;傅里叶加窗变换;多语义交互;批量匹配算法
Distribution Network Fault Diagnosis Method based on Batch Matching Algorithm Model
Abstract：After the fault occurs in the distribution network, the line current increases and the voltage drops rapidly, which affects the power quality and power supply reliability of the distribution network. In this study, the fault diagnosis system of the distribution network is established, and the fault location is quickly found and located by using the fault information collected. The transformer low-voltage side phase voltage waveform of two cycles before circuit breaker trip is collected, and the RMS value of fundamental wave is calculated by Fourier plus window transform, and the variation rule of low-voltage side phase voltage after failure is analyzed. The system uses a batch matching algorithm based on multi-semantic interaction and encoders are used to encode fault text information. Experimental results show that the training speed of the fault diagnosis model of the system in this study is fast, and the fault diagnosis accuracy is up to 100%. In the language push task test, the accuracy of SNL data set is 100%.
Key words：distribution network fault; Fault location; Transformer low voltage side; Fourier windowed transform; Multi-semantic interaction; Batch matching algorithm
收稿日期：2021-12-09
基金项目：
