基于改进Hilbert-Huang变换和支持向量机的高压断路器触头超程状态识别

引入集合经验模态分解(EEMD)对Hilbert-Huang变换(HHT)方法进行改进,并将改进的HHT方法结合支持向量机(SVM)应用于高压断路器振动信号特征提取和触头超程状态识别中。采用EEMD提取反映振动信号局部特性的本征模态函数(IMF)分量,并计算IMF分量的Hilbert边际谱能量值,由此构造高压断路器触头超程状态特征量,利用得到的特征向量对SVM进行训练,实现高压断路器触头超程状态的自动识别。试验提取了高压断路器在不同触头超程下的振动信号并进行分析,结果表明所提方法能够有效识别高压断路器触头超程状态。

Over-travel detection of electrical contact for high-voltage circuit breaker based on improved HHT and SVM

The HHT(Hilbert-Huang Transform) is improved by EEMD(Ensemble Empirical Mode Decomposition). The improved HHT method combined with SVM(Support Vector Machine)is applied to feature extraction of vibration signal of HVCB(High-Voltage Circuit Breaker) and over-travel state recognition of HVCB''s electrical contacts. The IMF(Intrinsic Mode Function) components reflecting local characteristics of HVCB''s vibration signal are extracted by EEMD, then their Hilbert marginal spectrum energy is calculated to build characteristic vector of electrical contact over-travel of HVCB. The SVM is trained by the characteristic vector to realize the automatic detection of electrical contact over-travel. The vibration signals of HVCB under different over-travel states are extracted and analyzed, and the results show that the proposed method can effectively detect the electrical contact over-travel of HVCB.