Investigation of Model-Based Anomaly Detection Method for the Full-Power Natural Circulation Small Modular Reactor NHR200-II

Abstract

The NHR200-II is a full-power range natural circulation small modular heating reactor characterized with inherent safety. The significant difference between NHR200-II and a general pressurized water reactor is that the load of NHR200-II is relatively smaller, and it could vary rapidly in a wider range under different operation conditions. The control system of NHR200-II is required to sustain the stability of reactor operating parameters and make sure the reactor follows the load change in time. Anomaly conditions may occur during the operation of NHR200-II. They may not induce safety-related accidents that gravely impair the reactor core or lead to the release of large-scale radioactivity, but they will fluctuate the power output. Therefore, it is practically valuable to detect abnormal working conditions in time. Furthermore, the detected information of anomaly is helpful for operators to conduct the reactor operation. A model-based anomaly detection method to diagnose abnormal conditions of circulation pump sensor in NHR200-II is investigated in this paper and verified with the NHR200-II model. The simulation results show that the circulation pump sensor fault will cause the variation of running parameters, including inlet and outlet temperature of the coolant inside the reactor core, nuclear power, and output steam flowrate from the steam generator. The variation of parameters will lead to system deviation from expected operating conditions and system stability degradation. When the proposed anomaly detection method is applied, the circulation pump sensor fault could be successfully diagnosed and detected, and a fault-tolerant control strategy is deployed to correct the deviations and stabilize the operation of the NHR200-II reactor.