Rolling bearing fault diagnosis method based on SA-ACWGAN-GP

Abstract

The invention discloses a rolling bearing fault diagnosis method based on SA-ACWGAN-GP, and relates to the technical field of fault diagnosis. The method comprises the steps: firstly collecting one-dimensional time domain vibration data of a bearing, converting the one-dimensional time domain vibration data into a two-dimensional frequency domain feature gray-scale map through fast Fourier transform, and taking a convolutional neural network as a network structure of a model to avoid gradient disappearance; secondly, constructing a model with a proper number of layers and initializing parameters, and inputting the training set into the model for training until the number of iterations is reached; and finally, applying the trained model to rolling bearing fault diagnosis. According to the method, an original ACGAN framework is improved, a Wasserstein distance and gradient penalty are introduced, the characteristics of periodicity and time sequence of a rolling bearing vibration signal are considered, and the precision of bearing fault feature extraction and fault category recognition is improved by combining a self-attention mechanism and ACWGAN-GP.

Description

technical field [0001] The invention relates to the technical field of fault diagnosis, in particular to a rolling bearing fault diagnosis method based on SA-ACWGAN-GP. Background technique [0002] As an important mechanical part, rolling bearings are widely used in various types of industrial equipment, and have the advantages of small frictional resistance, convenient assembly, and high efficiency. As the core components of rotating machinery such as gearboxes and turbines, the health of rolling bearings has a great impact on the service life and stability of the machine. In the actual production and operation process, the working environment of high-speed rotating mechanical equipment is harsh, such as lubricating oil pollution or overload, etc., resulting in multiple effects of different loads on rolling bearings, which are prone to various forms of defective failures, mainly including deformation, wear, Corrosion, cracks, etc. Structural design problems, according to...

AI Technical Summary

Problems solved by technology

However, the disadvantage of this kind of network is that it can only use artificially provided samples, and cannot understand the spatial distribution and internal structure of samples from a deeper level, which affects the final classification effect.

[0004] In practical problems, traditional fault diagnosis methods and discriminative-based deep learning fault diagnosis methods rely on manual feature extraction, which requires rich expert experience, resulting in poor final diagnosis results, and cannot achieve effective distribution of original data characteristics in multi-classification scenarios. study

Images