Abstract:Multi-dimensional hybrid feature extraction is difficult and the classification and identification accuracy is low in main grid circuit breaker operating mechanism abnormality monitoring. In this regard, the study proposes a multi-source signal feature extraction method by means of variational modal decomposition algorithm with integrated Mayer inverse spectral coefficients. On the basis of this method, a new type of abnormal monitoring and diagnosis technology for main power grid circuit breaker operating mechanism is proposed by using kernel principal component cluster analysis and optimised support vector machine. The experimental results show that under normal working conditions, the standard deviation and root mean square value of the signal feature extraction of the research method are 0.10 and 1.12, respectively, which verifies the effectiveness of the research method in multi-source signal feature extraction and noise suppression. In addition, the highest classification recognition accuracy of the research method reaches 97.7%. It can be seen that the proposed method of the research has significant advantages in improving the accuracy of anomaly monitoring and diagnostic efficiency. It can not only provide a new technical path for the anomaly monitoring and diagnosis of the main grid circuit breaker operating mechanism, but also provide an effective solution for the analysis of multi-source signals in practical applications.