Rotor fault diagnosis of centrifugal pumps in nuclear power plants based on CWGAN-GP-CNN for imbalanced dataset

Abstract

As a crucial device in nuclear power plants, centrifugal pumps undertake the critical role of cooling water circulation. Centrifugal pump rotor misalignment and unbalanced faults cause pump performance degradation, vibration increase, and equipment damage, thus seriously affecting the safety and reliability of nuclear power plants. In the process of centrifugal pump rotor fault, the difficulty in obtaining data samples and the limited amount of data can lead to an imbalance problem between the quantity of normal state and fault state samples in the dataset. In order to solve the problem, this paper proposed a CWGAN-GP model for generating rotor fault data based on CGAN and WGAN-GP models, and combined it with a two-stream CNN model to realize the rotor fault diagnosis with an imbalanced dataset. The quality and performance of the data generated by the proposed method were evaluated and validated in terms of visualization analysis, statistical indicators, and comparison with different data generation models. The results show that the CWGAN-GP model can generate high-quality data. Meanwhile, compared with other models on datasets with different degrees of imbalance, the two-stream CNN model is more effective in fault diagnosis on the expanded dataset by the CWGAN-GP model, and the improvement of fault diagnosis accuracy ranges from 1.40% to 13.33%.