考虑邻近效应的变压器线圈多故障诊断方法研究

Research on multi fault diagnosis method of transformer coil considering proximity effect

传统离线检测方法,无法对变压器线圈情况进行在线检测,为此,提出一种新的考虑邻近效应的变压器线圈多故障诊断方法。在考虑邻近效应的情况下,对变压器线圈进行谐波分析,构建特征包络线,以体现谐波振幅改变情况,针对变压器谐波波形含有的时域信号与频域信号两层特征,利用傅里叶变化把信号置于纯频域进行分析,在此基础上,依据能量熵对故障特征进行提取,添加幅值一致,但相位角差180°的辅助噪声,将傅里叶转换中的剩余辅助噪声进行滤除,得到变压器线圈多故障特征。依据多故障特征构建考虑邻近效应的变压器线圈故障诊断模型,实现故障诊断。进行实验时,计算故障能量熵,和正常运行情况下能量熵进行比较,发现熵值可作为诊断变压器线圈故障的依据。对三种故障特征进行提取,通过支持向量机分类模型对变压器线圈多故障进行诊断,和文献[5]方法与文献[8]方法相比,所提方法诊断准确率最高,可靠性最强。

the traditional off-line detection method can not detect the situation of transformer coil on-line.Therefore,a new multi fault diagnosis method of transformer coil considering proximity effect is proposed.Considering the proximity effect,the harmonic analysis is carried out on the transformer coil,and the characteristic envelope is constructed to reflect the change of harmonic amplitude.According to the two-layer characteristics of time-domain signal and frequency-domain signal contained in the harmonic waveform of the transformer,the signal is analyzed in the pure frequency domain by using the Fourier change.On this basis,the fault characteristics are extracted according to the energy entropy,and the amplitude is added the residual auxiliary noise in Fourier transform is filtered out,and the multi fault characteristics of transformer coil are obtained.According to the multi fault characteristics,a transformer coil fault diagnosis model considering proximity effect is constructed to realize fault diagnosis.In the experiment,the energy entropy of fault is calculated and compared with that of normal operation.It is found that the entropy value can be used as the basis for diagnosing the fault of transformer coil.Three kinds of fault features are extracted,and the multi fault diagnosis of transformer coil is carried out by support vector machine classification model.Compared with literature [5] and literature [8],the proposed method has the highest diagnostic accuracy and the strongest reliability.