Rolling bearing fault diagnosis method and system based on ISA-SDAE

Abstract

The invention relates to a rolling bearing fault diagnosis method based on ISCA-SDAE. The rolling bearing fault diagnosis method comprises the steps of performing normalization processing on motor rolling bearing fault data; initializing parameters; updating the particle value according to an ISCA particle value updating formula; checking whether particle values in the population are in a search range or not, if the particle values are in the search range, calculating classification error rates, screening out the minimum classification error rate, and taking the particle value corresponding to the minimum classification error rate as a current solution to output optimized SDAE network hyper-parameters; and substituting the particle value with the minimum classification error rate into corresponding parameters of a stacked noise reduction automatic encoder (SDAE), and performing classification by using a Soft-max classifier to obtain a test set fault classification result. The invention further discloses a rolling bearing fault diagnosis system based on the ISA-SDAE. The method has better feature learning ability, stronger robustness and generalization ability, and ultimately improves the accuracy of fault classification by balancing the hyper-parameters of the SDAE deep network structure optimized by the ISCA with strong ergodicity and high convergence efficiency.

Description

technical field [0001] The invention relates to the technical field of motor rolling bearing fault diagnosis, in particular to an ISCA-SDAE-based rolling bearing fault diagnosis method and system. Background technique [0002] As an important part of motors, rolling bearings are widely used in industrial production. During the long-term operation of the motor, the inner ring, rolling elements and outer ring of the rolling bearing are prone to failures, which will directly affect the performance of the motor, and even cause serious consequences such as motor function failure, economic losses or casualties. Therefore, it is of great significance to deeply study the fault diagnosis method of rolling bearings for the safety and reliability of the motor. [0003] The fault diagnosis method based on artificial intelligence has been widely used in the fault diagnosis of rolling bearings and achieved certain results. At present, most of the fault diagnosis of rolling bearings is t...

AI Technical Summary

Problems solved by technology

[0007] In the existing rolling bearing fault diagnosis algorithm, the improved sine-cosine algorithm is not used to introduce the stacked noise reduction autoencoder

Images