Vacuum switch mechanical fault diagnosis method and system, equipment and readable storage medium

Abstract

The invention discloses a vacuum switch mechanical fault diagnosis method and system, equipment and a readable storage medium, and the method comprises the steps: obtaining opening vibration signals of different vacuum fast switches in different mechanical fault states, carrying out the S transformation of the obtained opening vibration signals, obtaining an S transformation two-dimensional complex time-frequency matrix, then, carrying out modular operation on the S-transformation two-dimensional complex time-frequency matrix to obtain an S-transformation modular matrix, carrying out singular value decomposition on sub-matrixes after the modular matrix is divided, combining S transformation, singular value decomposition and an information entropy theory, calculating a maximum singular value energy entropy, and finally, adopting a random forest ensemble learning algorithm as a classifier, and inputting the feature vectors into a random forest model for fault classification and diagnosis. The results obtained after comparison with different feature quantities and classifiers show that the mechanical fault diagnosis method for the vacuum fast switch is good in feature consistency, high in classification accuracy and high in speed.

Description

technical field [0001] The invention belongs to the technical field of fault detection of high-voltage electrical appliances, and in particular relates to a vacuum switch mechanical fault diagnosis method, system, equipment and readable storage medium. Background technique [0002] In recent years, with the development of DC power grids, high-voltage vacuum circuit breakers have been more and more widely used in power grids. However, the traditional circuit breaker operating mechanism cannot meet the fast breaking requirements of the DC grid. Therefore, a new type of electromagnetic repulsion mechanism vacuum fast switch based on the principle of eddy current repulsion came into being. At present, research on this mechanism mainly focuses on characteristic analysis, structure optimization and energy conversion efficiency. However, the electromagnetic repulsion mechanism is easy to cause damage to its buffer device and circuit breaker due to its short action time, fast spee...

AI Technical Summary

Problems solved by technology

However, the training samples required by ANN are too large, and it is prone to over-learning and local optimum; the classification effect of SVM depends on the choice of kernel function, and different parameter optimization algorithms also have a great influence on the results; while the training time of RVM samples is relatively It is longer than SVM, and the choice of kernel function lacks theoretical guidance

Images