Multi-integrated fault diagnosis method for rotating machinery based on deep autoencoder dae

Abstract

The invention provides a multi-integration fault diagnosis method of rotating machinery based on a deep auto-encoder DAE. The implementation steps are as follows: constructing training samples, verification samples and test samples by using vibration signals of the rotating machinery; establishing a parallel deep auto-encoder network, and training the parallel deep auto-encoder network; performingfeature extraction on the training samples, the verification samples and the test samples by using the trained parallel deep auto-encoder to construct three feature pools; optimizing the features inthe three feature pools by using a constructed softmax classifier; constructing and training the softmax classifier again by using the optimized features to obtain a finally optimized softmax classifier; and inputting all screened test samples in the test sample feature pool into the trained parallel deep auto-encoder to perform feature extraction, and then inputting the features into the finallyoptimized softmax classifier to obtain a final classification result.

Description

technical field [0001] The invention belongs to the technical field of machinery, and further relates to a multi-integrated fault diagnosis method for rotating machinery based on a deep auto-encoder (DAE) in the technical field of rotating machinery. The invention can be used for judging, identifying and detecting the faults of the rotating machinery, and provides a basis for fault diagnosis and maintenance of the rotating machinery equipment. Background technique [0002] With the great opportunities and challenges brought by the rapid development of modern mechanical equipment, mechanical fault diagnosis technology is also developing vigorously towards intelligence. In actual production, rotating machinery is a vulnerable equipment, and the complexity, variability, and uncertainty of its faults lead to a particularly prominent demand for fault diagnosis. The upsurge of artificial intelligence triggered by deep learning technology has swept many research fields including f...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art, to provide a rotating machinery fault diagnosis method based on deep self-encoder DAE, to solve the problem of difficult generalization ability of deep neural network parameter adjustment

Images