Motor bearing fault diagnosis method based on single sample learning

Abstract

The invention discloses a motor bearing fault diagnosis method based on single sample learning, and belongs to the technical field of computers. The diagnosis method comprises the following steps: setting different sample numbers of each class and adding different noise intensities; carrying out data expansion on the existing one-dimensional vibration signal of the motor bearing; sending the processed data as input into a network for training and network parameter iterative updating; and performing fault diagnosis on the test sample so as to judge the model performance. According to the invention, learning is carried out based on a single sample of a twin network, so that the problem of low precision caused by data shortage during motor bearing fault diagnosis can be well solved; therefore, the phenomenon of industrial noise frequently occurring in a working condition environment can be solved, the feature information of the signal can be better extracted, and the improved multi-scale one-dimensional convolutional network fusion and LSTM combined method is adopted to extract the global features and local features of the signal, so that fault diagnosis is better carried out. The method has the advantages of high applicability and high stability.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a method for diagnosing faults of motor bearings based on single-sample learning. Background technique [0002] With the rapid development of technology and science, modern industrial machinery and equipment are becoming more and more complex. Rotating machinery is one of the most important pieces of equipment in modern industrial applications. Rolling element bearings, also known as rolling bearings, are common components of rotating machinery, and failure of rolling bearings can affect the normal operation of rotating machinery, causing severe equipment damage and economic costs, and sometimes even causing casualties. Many studies have shown that 40-50% of rotating machinery failures are directly caused by rolling bearing failures. Therefore, it is particularly important to diagnose the state of rolling bearings effectively, quickly and accurately, and this problem has attr...

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is: how to overcome the working conditions of the mechanical system in the prior art, which are very complicated and often change according to the production requirements, and the faulty motors in the industrial system cannot run for a long time, especially some critical Therefore, collecting and labeling enough training samples is difficult to achieve, and a single-sample learning-based motor bearing fault diagnosis method is provided, which uses multi-scale one-dimensional convolution according to the characteristics of the collected bearing raw signals How does the network perform feature fusion to extract local features at different levels, and the combination of multi-scale one-dimensional convolutional network and LSTM is used as a sub-network of twin network to extract global features, which has good classification performance, high reliability and strong adaptability features

Images