Signal validation techniques and power plant applications

Abstract

Computerized signal validation is a software technique for integrating information from redundant and from functionally diverse sensors to provide highly reliable information to operating crews and to auto matic controllers. High reliability results from greater and more comprehensive use of existing sensor redundancy, and from inherent fault detection and isolation (FDI) of faulty sensors. Operating crews at power plants have always performed some measure of manual signal validation. With the current thrust to introduce digital computers to support plant monitoring and operator aid functions# signal validation has become generally recognized as a key element in the interface between plant sensors and algorithms using sensor data. For any computerized system to be robust and to yield a false alarm rate that is credibly low, high information reliability must be assured at the front end. Progress of research in computerized signal validation techniques has benefited the U.S. utilities by improving the reliability of safety parameter display systems (SPDS) implemented in the nuclear power plants. Conventional signal validation techniques have been combined with more robust parity space and analytic redundancy methods to provide reliable and real-time SPDS monitoring information to PWR and BWR operators. The technical details of the methods, application to PWR and BWR SPDS, and the extension of the techniques to plant control functions are reviewed in this paper.