Motor fault identification method and system based on neural network, and storage medium

Abstract

The invention discloses a motor fault identification method and system based on a neural network, and a storage medium. The method comprises the following steps: respectively collecting multiple groups of signals of a motor in normal operation and different fault states, performing wavelet packet transformation to obtain frequency band energy index values corresponding to different wavelet packet nodes, and constructing feature vectors of the motor in normal operation and different fault states by using the frequency band energy index values; inputting the feature vectors of the multiple groups of signals into a neural network for training test, taking the recognition rates of the motor in a normal operation state and different fault states as optimization targets, taking network structure parameters of the neural network as to-be-optimized parameters, obtaining multiple neural networks corresponding to an optimal solution set under multiple optimization targets, and obtaining the optimal solution set of the motor; and different neural networks can be flexibly selected and called according to different application scenes. The method can solve the problem of mutual exclusion between the recognition rates of different operation states of the motor, improves the applicability of the neural network, and achieves the intelligent recognition of normal and different fault states of the motor.

Description

technical field [0001] The present application relates to the technical field of fault identification, and more particularly, to a method, system and storage medium for motor fault identification based on a neural network. Background technique [0002] Asynchronous motors have obvious advantages such as simple structure, reliable operation, easy manufacture, low price, sturdiness and durability, high working efficiency and better working characteristics, and are widely used in various industrial production fields such as metallurgy, coal, mining, machinery and oil fields. . As an important core component of an industrial production system, the reliability of the motor will affect the performance of the entire system. Once a fault occurs, it is easy to produce a chain reaction and lead to the paralysis of the entire system. Therefore, the initial fault diagnosis of the motor is particularly important. Among them, broken rotor bars, air gap eccentricity, short circuit between...

AI Technical Summary

Problems solved by technology

Among them, rotor broken bars, air gap eccentricity, stator turn-to-turn short circuit, and bearing faults are common faults of asynchronous motors. In the existing motor fault diagnosis methods, current spectrum analysis is widely used due to the convenience of signal acquisition and rich fault information. , but in the traditional current spectrum analysis, the fault identification signal is single, which is easily affected by the leakage of the fundamental frequency component and the fluctuation of the motor load.

[0003] In addition, with the improvement of the intelligent operation and maintenance level of industrial production, a series of machine learning algorithms have been applied to the field of motor fault identification. Automatic training and learning are carried out through training samples to obtain a classifier model whose recognition accuracy meets the training error requirements. Automatic classification of operating status and intelligent identification of faults, but there are still many problems, and further research is needed to improve the performance of the model

Images