Online monitoring and diagnosis of high voltage circuit breaker faults: feature extraction analysis of vibration signals

Abstract

The development of power grid system not only increases voltage and capacity, but also increases power risk. This paper briefly introduces the feature extraction method of the vibration signal of high voltage circuit breaker and support vector machine (SVM) algorithm and then analyzed the high voltage circuit breaker in three states: normal operation, fixed screw loosening and falling of opening spring, using the SVM based on the above feature extraction method. The results showed that the accuracy and precision rates of fault identification of circuit breaker were the highest by using the wavelet packet energy entropy extraction features, the false alarm rate was the lowest, and the detection time was the shortest.

Condition Assessment of Power Circuit Breakers Based on Machine Learning Algorithms

“…The second stage -the decision stage -is directly dependent on the previous one. Thus, the authors choose machine learning classification methods based on Support Vector Machine (SVM), [20], [21], [28], [29] Self-Organizing Maps Network (SOM) [14], [30] Convolutional Neural Kerim Obarcanin, Senior Member IEEE, Dzenita Skulj, and Bakir Lacevic, Member, IEEE Network (CNN) [11], [31] Extreme Learning Machine (ELM) [32], Classification and regression tree (CART) [33] etc.…”

A novel circuit breaker fault diagnosis method based on dense residual and attention mechanism

“…The vibration signal produced by the HVCB during the opening and closing process can intuitively reflect the status information of each component of the operating mechanism. The vibration signal can be analyzed to extract mechanical fault characteristics of the operating mechanism and to determine the status of the HVCB [5][6][7].…”