基于ITD-ELM的变压器绕组微弱匝间短路故障诊断

Fault diagnosis of weak turn-to-turn short circuit in transformer windings based on ITD-ELM

为了解决油浸式变压器内部绕组绝缘故障检测难度大的问题,提出了一种基于本征时间尺度分解(ITD)和极限学习机(ELM)的变压器绕组微弱匝间短路故障诊断方法。首先,将采集到的变压器可听声信号利用ITD算法分解为若干旋转(PR)分量,并将峭度值较大的分量信号相加,对可听声信号进行重构;其次,将重构后的变压器可听声信号作为模型输入层,故障诊断结果作为模型输出层,构建了基于ITD-ELM的变压器绕组微弱匝间短路故障诊断模型;最后,以一台110 V变压器搭建实验模拟平台,对其微弱匝间短路故障进行训练并诊断。结果表明:基于ITD-ELM模型的微弱匝间短路故障诊断精确率为98%,较传统的ELM故障诊断精确度提升了3%,验证了所提变压器绕组微弱匝间短路故障诊断方法的准确性。研究所提出的故障诊断方法准确性较高,可应用于现场运行的不同电压等级的油浸式变压器。

In order to solve the problem of the difficulty in detecting insulation faults in internal windings of oil immersed transformers,a fault diagnosis method based on intrinsic time-scale decomposition(ITD)and extreme learning machine(ELM)was proposed.Firstly,the collected transformer audible signal was decomposed into proper rotation(PR)components by ITD algorithm,and the component signals with larger kurtosis value were added to reconstruct the audible signal.Secondly,taking the reconstructed transformer audible signal as the model input layer and the fault diagnosis result as the model output layer,a fault diagnosis model of transformer winding weak turn-to-turn short circuit based on ITD-ELM was constructed.Finally,a 110 V transformer was used to build an experimental simulation platform to train and diagnose the weak turn-to-turn short-circuit fault.The results show that the fault diagnosis accuracy of weak turn-to-turn short circuit based on ITD-ELM model is 98%,which is 3%higher than that of traditional ELM fault diagnosis.This verifies the accuracy of the fault diagnosis method of weak turn-to-turn short circuit in transformer windings the proposed fault diagnosis method has high accuracy and can be applied to oil-immersed transformers with different voltage levels in field operation.