Research on fault diagnosis and fault location of nuclear power plant equipment

Nuclear energy, as a clean energy source, has advantages such as environmental friendliness and low consumption. However, it also carries the drawback of potential radioactive leaks. The impact of radioactive leaks from nuclear power plants on the environment and humans is far greater than that of other types of power plants. Therefore, there are higher requirements for safety in nuclear power plants. Prompt and accurate identification of equipment failures, as well as precise fault localization, can assist operators and maintenance personnel in taking appropriate actions to prevent further deterioration and improve the economic and safety aspects of nuclear power plants. To address these issues, this study has developed a nuclear power plant equipment fault diagnosis and fault localization system. Firstly, considering the large number of sensors in nuclear power plants and the high dimensionality of the collected data, which can affect the effectiveness of subsequent fault diagnosis and localization models, the use of Principal Component Analysis (PCA) for dimensionality reduction of highdimensional feature parameters is proposed. Then, a Residual Network (ResNet) is employed for fault classification. For fault localization, a Long-Short Term Memory Auto-Encoder (LSTM-AE) is used. The experimental results demonstrate that the developed system achieves high accuracy in nuclear power plant equipment fault diagnosis and fault localization.