Fault Detection and Identification for Sensor Channels in Steam Generator Level Control Loops

Faults in sensor channels as a result of calibration errors or a bias in transmitters can lead to various safety and operational issues in nuclear power plants, such as reduced trip margins, jeopardized performance, or creating a false sense of security regarding the health of the plant operation. Unfortunately, such faults are difficult to detect, specifically when the sensors are part of a closed-loop system, which has a tendency to reduce or to mask the effects of the fault through feedback control actions during a dynamic transient process. Detecting these types of faults within a steam generator (SG) level control system is investigated in this paper. It is shown that the tools based on analytical redundancy can be very effective in this case. Using a typical SG level control system as a reference plant, specific analytical redundancy formulations are derived. Three sensor fault scenarios have been considered in detail. It has been shown, by simulation and analysis, that the proposed technique can effectively detect the sensor faults that are otherwise difficult to detect.