Study on sensor fault diagnosis in nuclear power plants based on PCA

To solve the fault diagnosis of the safety-related sensors in nuclear power plants, a monitoring model for several important sensors is established based on principal component analysis (PCA). Sensor fault detection, sensor identification and faulty sensor signals reconstruction can be achieved by calculating the squared prediction error, sensor validity index and reconstruction index. Employing the data from a real-time full-scope simulator of nuclear power plants, the monitoring model is proved effectively to detect the complete failure fault, fixed bias fault, drifting fault, and precision degradation fault. The simulation result illustrates the effectiveness of the model for the sensor fault diagnosis and recovery