Anomaly Detection in a Reactor Coolant Pump Flywheel System via Pulse Shape Analysis

In this study, a method that discriminates between anomalies present or absent in the vibration signal of a flywheel system is developed. First by means of MATLAB and Simulink, a simple flywheel system under different feasible conditions is simulated using equations of motion to capture the dynamic behaviors of the components of the system along their lines of action. The resulting vibration signals obtained from the simulations are combined with varying levels of noise and then subjected to pulse shape analysis (PSA). PSA is a tool that has been mostly used in the field of nuclear engineering, and it is explored and used differently here with the objective of developing a suitable PSA algorithm that can differentiate between vibration signals based on the presence or absence of an anomaly. The algorithm is a time-domain technique with minimal computational time that can be very easily applied. At the end, it is shown that the developed PSA algorithm can identify an anomaly in a vibration signal on the basis of a defined pattern under certain attainable conditions.