WCSE

Imbalanced Fault Classification of Industrial Bearings Based on Generative Adversarial Network with an Improved Structure

Abstract— With the development of modern industry, data collection is becoming heterogeneous with very huge amount, thus data-driven technology has become increasingly important in process monitoring. At present, the pattern recognition method based on Artificial Neural Network (ANN) has been widely used in fault classification of rotating machinery. However, they require a large amount of sample data to participate in training models to ensure the accuracy, while sample data is extremely lacking in engineering practice. Therefore, a suitable method is needed for fault classification of rotating machinery in the case of small samples and imbalanced data. This paper proposed a fault classification framework for small sample data, which can be used to generate fault data through GAN for fault classification. First, GAN generates different types of fault data with the same distribution as the original data, and then compares the generated time series data and the real time series data with various degree of difference evaluation indicators to obtain the fault classification results, which can verify the effectiveness of GAN in fault data classification.

Index Terms— GAN, Fault classification, Difference comparison, Imbalanced sample.

Zhao An
Taiyuan University of Technology, China
Lan Cheng
Taiyuan University of Technology, China
Yuanjun Guo
Shenzhen Institute of Advanced Technology Chinese Academy of Sciences, China
Mifeng Ren
Taiyuan University of Technology, China
Zhile Yang
Shenzhen Institute of Advanced Technology Chinese Academy of Sciences, China
Yanhui Zhang
Shenzhen Institute of Advanced Technology Chinese Academy of Sciences, China
Hong Qian
3Department of Automation Engineering Shanghai University of Electric Power Shanghai, China

[Download]

Cite: Zhao An, Lan Cheng, Yuanjun Guo, Mifeng Ren, Zhile Yang, Yanhui Zhang, Hong Qian, " Imbalanced Fault Classification of Industrial Bearings Based on Generative Adversarial Network with an Improved Structure, " WCSE 2022 Spring Event: 2022 9th International Conference on Industrial Engineering and Applications, pp. 1656-1663, Sanya, China, April 15-18, 2022.