The fast spectrum kurtosis (FSK) algorithm can adaptively identify and select the resonant frequency band and extract the fault feature by the envelope demodulation method. However, in practical applications, the fault source may be located in different resonant frequency bands; plus in noise interference, the weak side of the compound fault is not easy to be identified by the FSK. In order to improve the accuracy of fast spectral kurtosis analysis method, a modified method based on maximum correlation kurtosis deconvolution (MCKD) is proposed. According to the possible fault characteristic frequencies, the period of MCKD is calculated, and the appropriate filter length is selected to filter the original compound fault signal. In this way, the compound fault located in different resonance bands is separated. Then, the signal after MCKD filtering is analyzed by FSK. Through the simulation and experimental analysis, the MCKD can separate the compound fault information in different frequency band and eliminate the noise interference; the FSK can accurately identify the resonance frequency and identify the weak fault characteristics of compound fault.
As high-voltage circuit breakers (HVCBs) are directly related to the safety and the stability of a power grid, it is of great significance to carry out fault diagnoses of HVCBs. To accurately identify operating states of HVCBs, a novel mechanical fault diagnosis method of HVCBs based on multi-feature entropy fusion (MFEF) and a hybrid classifier is proposed. MFEF involves the decomposition of vibration signals of HVCBs into several intrinsic mode functions using variational mode decomposition (VMD) and the calculation of multi-feature entropy by the integration of three Shannon entropies. Principle component analysis (PCA) is then used to reduce the dimension of the multi-feature entropy to achieve an effective fusion of features for selecting the feature vector. The detection of an unknown fault in HVCBs is achieved using support vector data description (SVDD) trained by normal-state samples and specific fault samples. On this basis, the identification and classification of the known states are realized by the support vector machine (SVM). Three faults (i.e., closing spring force decrease fault, buffer spring invalid fault, opening spring force decrease fault) are simulated on a real SF6 HVCB to test the feasibility of the proposed method. The detection accuracies of the unknown fault are 100%, 87.5%, and 100% respectively when each of the three faults is assumed to be the unknown fault. The comparative experiments show that SVM has no ability to detect the unknown fault, and that one-class support vector machine (OCSVM) has a weaker ability to detect the unknown fault than SVDD. For known-state classification, the adoption of the MFEF method achieved an accuracy of 100%, while the use of a single-feature method only achieved an accuracy of 75%. These results indicate that the proposed method combining MFEF with hybrid classifier is thus more efficient and robust than traditional methods.
Mechanical fault diagnosis of a circuit breaker can help improve the reliability of power systems. Therefore, a new method based on multiscale entropy (MSE) and the support vector machine (SVM) is proposed to diagnose the fault in high voltage circuit breakers. First, Variational Mode Decomposition (VMD) is used to process the high voltage circuit breaker's vibration signals, and the reconstructed signal can eliminate the effect of noise. Second, the multiscale entropy of the reconstructed signal is calculated and selected as a feature vector. Finally, based on the feature vector, the fault identification and classification are realized by SVM. The feature vector constructed by multiscale entropy is compared with other feature vectors to illustrate the superiority of the proposed method. Through experimentation on a 35 kV SF 6 circuit breaker, the feasibility and applicability of the proposed method for fault diagnosis are verified.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.