A concept of instrumentation failure detection for nuclear power plant monitoring system

The monitoring system is one of the most important parts of the instrumentation and control (I and C) of nuclear power plants (NPPs). This system has a function to monitor NPP variables and provide data to other I and C systems and to operator for controlling the operation of the NPP. It has to provide the operators with accurate and appropriate information during both normal and abnormal operations. For providing the accurate and appropriate information, the monitoring system is supported by hundreds of measurements (instruments). Therefore it is important that instrument failures be detected before significant performance degradation. In this paper, an application of nonlinear filter to instrumentation failure detection mechanism is presented. The mechanism uses a bank of nonlinear filters to generate optimal estimates of the states. The nonlinear filter can work well for highly nonlinear problem that are difficult for the conventional Kalman filter. (author)