Design and Development of Intelligent Visual Simulator for Fault Detection, Identification and Diagnosis in PWR Nuclear Power Plant

In this research, the AP600 Pressurized Water Reactor (PWR)-type Nuclear Power Plant (NPP) is studied due to its large number of components and complex, diversified systems. Operating a reliable and economical PWR NPP without malfunctions is desirable, with maximum safety as the primary goal. A Personal Computer Transient Analyzer (PCTRAN) is used as a data-driven source for AP600 PWR NPP, enabling simulations of both normal and abnormal operations. A state-of-the-art, fully automated, intelligent fault detection, identification, and diagnosis software (AI-FDID-PCTRAN) is designed and developed in Visual Basic to address various safety concerns and enhance the reliability and availability of AP600 PWR NPP systems. AI-FDID-PCTRAN is formulated, programmed, and configured based on unsupervised machine learning using Principal Component Analysis (PCA), a fully Automated Multivariate Statistical Process Control Technique (AMSPCT). The proposed PCA-based technique is a purely software-driven, systematically structured, and fully automated approach, developed specifically for the AP600 PWR nuclear industry. This specialized software offers capabilities not found in highly expensive, commercially available alternatives. FDD-PCTRAN has been tested against benchmark normal and abnormal transients available in AP600 PCTRAN and has proven to be highly reliable and accurate in fault detection and diagnosis.