2018 6th International Conference on Mechanical Engineering, Materials Science and Civil Engineering | |
Research on Rolling Bearing Fault Identification Method Based on LSTM Neural Network | |
材料科学;机械制造;土木建筑工程 | |
Luo, Pan^2 ; Hu, Yumei^1 | |
State Key Laboratory of Mechanical Transmission, Chongqing University, Chongqing | |
400024, China^1 | |
School of Automotive Engineering, Chongqing University, Chongqing | |
400024, China^2 | |
关键词: Fault data; Fault recognition; Fault-detection process; Feature learning; Neural network model; Original signal; Recognition accuracy; Rolling bearings; | |
Others : https://iopscience.iop.org/article/10.1088/1757-899X/542/1/012048/pdf DOI : 10.1088/1757-899X/542/1/012048 |
|
来源: IOP | |
【 摘 要 】
In order to simplify the fault detection process, improve the efficiency of fault detection and recognition accuracy, a rolling bearing fault recognition based on LSTM neural network is proposed. In this model, there is no need to perform any preprocessing on the original data. As long as the neural network model training is completed, the original signal can be detected and identified automatically by the model. In order to verify the performance of the model, the test results of the same fault data set are compared with the fault recognition model based on traditional machine learning. The results show that the fault recognition model based on LSTM neural network has obvious superior performance and higher recognition reliability. Its recognition accuracy rate reaches 98.00%, and the recognition accuracy of the fault recognition model based on traditional machine learning is only 94.20%.
【 预 览 】
Files | Size | Format | View |
---|---|---|---|
Research on Rolling Bearing Fault Identification Method Based on LSTM Neural Network | 574KB | download |