Condition monitoring and breakage assessment of steam generator heat transfer tubes in nuclear power plants

The steam generator (SG) is a critical component of the steam power conversion system in nuclear power plants. The heat transfer tubes of steam generators are susceptible to mechanical and chemical damage due to prolonged exposure to high-temperature, high-pressure environments, and high-radiation media. Timely detection of abnormal states and accurate assessment of the breach degree in the heat transfer tubes are crucial for enhancing the economic and operational safety of nuclear power plants. This study focuses on simulating the normal state of steam generator heat transfer tubes and different degrees of abnormal states using a simulator, while collecting characteristic parameters that can be monitored by sensors. In order to improve the fidelity of the simulated signals to real-world engineering signals, in cases where the breach degree is significant, the reactor undergoes an emergency shutdown, resulting in a smaller amount of effective signal data collected for larger breach degrees. To address these issues, this paper employs the Synthetic Minority Oversampling Technique (SMOTE) to expand the capacity of small sample data. Additionally, to mitigate the impact of high-dimensional feature parameters on subsequent condition monitoring and breach degree assessment, a Denoised AutoEncoder (DAE) is employed to reduce the dimensionality of the feature parameters. The One-Class Support Vector Machine (One-Class SVM) is then utilized to monitor the condition of the steam generator heat transfer tubes. When an abnormality is detected in the heat transfer tubes, a Bi-directional Long Short-Term Memory (Bi-LSTM) model is used to evaluate the magnitude of the tube leakage. The experimental results demonstrate that the developed system achieves a high monitoring accuracy and provides a good assessment of the fault degree.