Rolling bearing fault diagnosis method based on parallel feature learning and multi-classifier

Abstract

The present invention proposes a rolling bearing intelligent fault diagnosis method based on multi-classifier integration and parallel feature learning, aiming at improving the classification accuracy of the model. The implementation steps are: obtaining a training sample set and a test sample set; establishing multiple stack autoencoders The model uses the training sample set as input to perform parallel training on the stacked autoencoder model, and extracts multiple features of the training sample set; evaluates the extracted features based on the softmax model, and screens the feature composition according to the corresponding threshold and evaluation index value Feature subsets; according to the feature subsets, multiple classifiers based on the softmax model are established, and the classification accuracy of each classifier is obtained with the feature subsets as input, and multiple classifiers are reselected according to the threshold to construct an integrated multi-classifier model. The voting method obtains the prediction label of the integrated multi-classifier model, and maps the prediction label with the rolling bearing fault type to realize the intelligent fault diagnosis of the rolling bearing.

Description

technical field [0001] The invention belongs to the technical field of intelligent fault diagnosis of rotating machinery, and relates to a rolling bearing fault diagnosis method, in particular to an intelligent rolling bearing fault diagnosis method based on parallel feature learning and integrated multi-classifiers, which can be used for automatic fault diagnosis of rolling bearings and other rotating machinery. Background technique [0002] Rotating machinery plays an important role in industrial equipment. Rolling bearings are one of the most important components in rotating machinery such as motors, wind turbines and gearboxes. It consists of rolling elements, outer rings, inner rings and cages. Rolling bearings usually work under complex working conditions, such as different working conditions, vibration, temperature, load, etc. These factors often lead to the decline of rolling bearing performance or even failure. The performance status of rolling bearings directly af...

AI Technical Summary

Problems solved by technology

However, although this method can extract deep features by integrating the deep autoencoder model in the feature extraction and selection stage, it does not effectively evaluate the extracted deep features, and cannot screen out the most relevant and more representative features that are highly relevant to the diagnostic target. The deep features of the model affect the classification and diagnosis accuracy of the model; in addition, this method only uses a softmax classifier for diagnosis in the fault classification and diagnosis stage, which has poor robustness and leads to low diagnostic accuracy of the model under random interference

Images