Fault detectability analysis in PCA method during condition monitoring of sensors in a nuclear power plant

Principal component analysis (PCA) is applied in this paper for sensor condition monitoring in a nuclear power plant (NPP). In order to estimate the fault detectability of the proposed PCA model, non-detection zones of sensors in the model are calculated. Then the diagnosis ability of the PCA model is more specific, and the diagnosis results are more credible and reliable. Meanwhile an improved statistics-based method is applied to directly reduce the false alarms during monitoring which makes contribution to the improvement of model performance. Finally, the credibility of the non-detection zones of sensors are evaluated with sensor measurements from a real NPP. According to simulation results, the non-detection zones of sensors are proved to be credible and reliable, and the false alarm reducing method also makes contribution to the condition monitoring performance of the developed PCA model.