Rolling bearing fault diagnosis method based on 1-DCNN (1-Dimensional Convolutional Neutral Network) and LSTM (Long Short-Term Memory) fusion

Abstract

The invention discloses a rolling bearing fault diagnosis method based on 1-DCNN (1-Dimensional Convolutional Neutral Network) and LSTM (Long Short-Term Memory) fusion. The method comprises the following steps that a, data processing and expansion are carried out on original one-dimensional vibration signals through utilization of a sliding window overlapped sampling method, thereby obtaining expanded signals; b, the expanded signals are input into two channels of the 1-DCNN and the LSTM for training and analysis, and feature information is extracted, extraction is carried out in the 1-DCNN channel to obtain space feature information, and the extraction is carried out in the LSTM channel to obtain time feature information; c, the space feature information and the time feature information are concatenated through utilization of a concatenate layer, thereby obtaining concatenated features; and d, the concatenated features are connected with a full concatenate layer, and fault types are classified through utilization of a Softmax classifier. The method is relatively high in accuracy rate, relatively rapid in convergence rate and relatively low in loss and has high generalization and robustness.

Description

technical field [0001] The invention relates to a bearing fault diagnosis method, in particular to a rolling bearing fault diagnosis method based on fusion of 1-DCNN and LSTM. Background technique [0002] Rolling bearings are one of the core components of mechanical equipment, and their health is the key to ensuring the normal operation of equipment. According to incomplete statistics, about 30% of the faults in rotating machinery are caused by bearing damage. Therefore, the research on bearing fault diagnosis technology It has important economic value and social benefits. [0003] With the rise of artificial intelligence technology driven by big data, deep learning has been widely used in feature learning, pattern recognition, data mining and other fields with its powerful learning ability. Fault diagnosis has also made new research progress with the application of deep learning technology. As a typical technology of deep learning, the research of convolutional neural ne...

AI Technical Summary

Problems solved by technology

[0007] Although the above studies have achieved good results in bearing fault diagnosis, there are still some problems: the model input needs to convert the one-dimensional bearing signal into two-dimensional data input, which to a certain extent loses the integrity and spatial correlation of the original signal; Increased the artificial interference of data preprocessing, and did not take advantage of the powerful feature self-learning and data processing capabilities of the deep learning network

Images