利用窄带噪声辅助多元经验模态分解算法检测换流变压器用有载分接开关机械状态

Mechanical Condition Detection of On-Load Tap-Changer in Converter Transformer Based on Narrowband Noise Assisted Multivariate Empirical Mode Decomposition Algorithm

为有效检测换流变压器用有载分接开关的机械状态,提出一种窄带噪声辅助多元经验模态分解的方法对有载分接开关切换过程中采集到的多通道振动信号进行分析。具体应用时,在原始多通道振动信号中增加若干通道窄带噪声信号,并在统一高维超球面坐标系下进行分解计算,从而在有效抑制经验模态分解过程中模态混叠现象的同时,极大提高了多维振动信号分解的准确性并降低了运算复杂性。进而根据区间最大功率特征计算固有模态函数的功率矩阵,对有载分接开关的机械特征进行描述。对有载分接开关样机正常与典型机械故障时振动信号的计算结果表明,其不同工况下的功率特征有较大区别,所定义的矩阵相似度指标可较好地衡量有载分接开关典型故障时的振动差异程度。此外,触头松动时与正常工况的相似度指标低于绝缘板松动故障,说明触头松动故障对有载分接开关切换过程的影响更为明显。

In order to effectively detect the mechanical condition of on-load tap-changer(OLTC)in converter transformer,the narrowband noise assisted multivariate empirical mode decomposition(NNA-MEMD)method was proposed.This method can analyze the multichannel vibration signals during contacts switch.In detail,several narrowband noises were added to the multichannel vibration signals.Meanwhile the intrinsic mode functions(IMF)of each channel were uniformly calculated under the high dimensional hypersphere.The method will not only reduce the aliasing effect of EMD,but also improve the accurateness and reduce the complexity of signal decomposition.The power matrices of IMF based on interval maximum power feature under different conditions were also calculated to describe the mechanical characteristics.The results under normal and typical fault conditions of R type OLTC model show that,the power matrices of different conditions vary significantly,so that the matrix similarity index can measure the difference degree of vibration under typical faults.In addition,the index of contact looseness is lower than that of the insulated panel looseness.It indicates that the contact looseness fault has greater influence on switch-over process of OLTC.