Rolling bearing fault diagnosis method based on fft and fully connected layer-svm

Abstract

The invention discloses a rolling bearing fault diagnosis method based on FFT and fully connected layer-SVM, including the extraction, transformation, normalization and other processes of vibration signals of the rolling bearing under normal conditions, inner ring faults, outer ring faults, and rolling body faults; The rolling bearing fault detection method does not rely on the convolutional neural network when detecting rolling bearings, it only needs to train the SVM classifier and the fully connected layer to complete, and the parameters of the classifier do not need to be optimized by the algorithm, which can improve the training of the model Speed ​​and classification accuracy rate; Simultaneously the method of the present invention has only one layer of fully connected layer, does not need the convolution kernel operation of convolutional neural network, makes its parameter optimization time greatly reduce, experimental result proves, the method of the present invention measures The accuracy rate of rolling bearing faults reaches 100%, the detection time is shortened by more than 80%, and it has the characteristics of high classification accuracy and fast fault diagnosis speed.

Description

technical field [0001] The invention belongs to the technical field of fault diagnosis, and in particular relates to a fault diagnosis method of a rolling bearing based on FFT and fully connected layer-SVM. Background technique [0002] Modern industrial production requires very close production rhythm. Once mechanical equipment fails, it will disrupt the production process, causing huge economic losses and even casualties. Today, large-scale production relies more on highly reliable production and processing equipment, and mechanical fault diagnosis has become a A rapidly changing engineering technology; [0003] Rolling bearings are not only widely used in production, rolling bearings are also wearing parts in mechanical equipment. According to statistics, about 30% of the failures in rotating machinery are caused by rolling bearing failures; after using fault diagnosis technology, the probability of mechanical equipment failure can be extremely high. The maintenance cost...

AI Technical Summary

Problems solved by technology

[0003] Rolling bearings are not only widely used in production, rolling bearings are also vulnerable parts in mechanical equipment. According to statistics, about 30% of the faults in rotating machinery are caused by rolling bearing failures; after using fault diagnosis technology, the probability of mechanical equipment failure can be extremely high The reduction of maintenance cost is 25%-50% less than before; due to the long life of the rolling bearing, according to the national standard, there is a 90% reliability threshold as a standard part, even if the same material is used on the same machine with the same process The same batch of bearings is produced, and its service life is also different; usually in reality, regular inspection and repairs are carried out based on the designed fatigue life, which has a great chance of damaging the equipment and causing economic losses;

[0004] At present, in the practice of enterprises, if it is a rolling bearing in an unimportant occasion, it mainly depends on the experience of engineers and technicians to judge whether there is a fault in the operation of the bearing through touch and hearing; in more important occasions, it is necessary to regularly disassemble and check the operation of the bearing However, it is impossible to predict the fault state in a timely and effective manner; the current detection algorithms for bearing fault defects mainly focus on the traditional use of time domain and frequency domain signals as signal input features, and then use machine learning or deep learning methods for classification, with high accuracy , however, it does not meet the requirements of the industrial detection application of core components; at the same time, the current detection algorithm for bearing fault defects consumes a relatively long time, and the model convergence speed is too slow

Images