Rotating machine fault diagnosis method based on an optimized structure convolution neural network

Abstract

The invention discloses a rotating machine fault diagnosis method based on an optimized structure convolution neural network. Firstly, working signals are collected in the normal state and the fault state of a rotating machine, then are converted into gray scale images, and the gray scale images and the corresponding fault labels are used as training samples to train the constructed convolution neural network; and in the working process of the rotating machine, a working signal is collected and is converted into a gray scale image, and the trained convolution neural network is input for faultdiagnosis. The invention adopts the method of converting the collected working signals of the rotating machine into gray-scale images, and completes the task of multi-classification of faults of the rotating machine through a convolution neural network.

Description

technical field [0001] The invention belongs to the technical field of fault diagnosis of construction machinery systems, and more specifically relates to a fault diagnosis method for rotating machinery based on an optimized structure convolutional neural network. Background technique [0002] Rotating machinery is the most widely used machinery in industry. With the development of modern industry and the improvement of mechanical automation, its reliability, maintainability and safety have attracted more and more attention. Rolling bearings are one of the core components of rotating machinery. According to statistics, about 30% of mechanical failures in rotating machinery using rolling bearings are related to bearing damage. Moreover, compared with other mechanical parts and components, rolling bearings have the characteristics of large dispersion of service life. Therefore, in actual work, some bearings have exceeded their design life but can still work normally, and some ...

AI Technical Summary

Problems solved by technology

These methods can effectively deal with non-stationary time series, but because the original time series is hidden and the method of relying on feature classification is used, it is difficult to fully express the information of the original time series, and it is impossible to achieve high-precision classification.

The final recognition accuracy of traditional machine learning models depends heavily on the extracted features, and there are two problems in feature extraction: one is whether the processed data has a good ability to express signal features, and the other is that the preprocessing process is extremely time-consuming

Therefore, in the actual use of engineering, the traditional machine learning model has the problems of slow speed and low accuracy, which need to be solved by research

Images