Instrument failure monitoring in nuclear power systems
Methods of monitoring dynamic systems for instrument failures were developed and evaluated. In particular, application of these methods to nuclear power plant components is addressed. For a linear system, statistical tests on the innovations sequence of a Kalman filter driven by all system measurements provides a failure detection decision and identifies any failed sensor. This sequence (in an unfailed system) is zero-mean with calculable covariance; hence, any major deviation from these properties is assumed to be due to an instrument failure. Once a failure is identified, the failed instrument is replaced with an optimal estimate of the measured parameter. This failure accommodation is accomplished using optimally combined data from a bank of accommodation Kalman filters (one for each sensor), each driven by a single measurement. Using such a sensor replacement allows continued system operation under failed conditions and provides a system operator with information otherwise unavailable. To demonstrate monitor performance, a liner failure monitor was developed for the pressurizer in the Loss-of-Fluid Test (LOFT) reactor plant. LOFT is a small-scale pressurized water reactor (PWR) research facility located at the Idaho National Engineering Laboratory. A linear, third-order model of the pressurizer dynamics was developed from first principles and validated. Using data from the LOFT L6 test series, numerous actual and simulated water level, pressure, and temperature sensor failures were employed to illustrate monitor capabilities. Failure monitor design was applied to nonlinear dynamic systems by replacing all monitor linear Kalman filters with extended Kalman filters. A nonlinear failure monitor was derived for LOFT reactor instrumentation. A sixth-order reactor model, including descriptions of reactor kinetics, fuel rod heat transfer, and core coolant dynamics, was obtained and verified with test data.
- OSTI ID:
- 6427188
- Country of Publication:
- United States
- Language:
- English
Similar Records
Evaluation of a Kalman filter based power pressurizer instrument failure detection system implemented on a nuclear power plant training simulator
Low-order model of the Loss-of-Fluid Test (LOFT) reactor plant for use in Kalman filter-based optimal estimators
Related Subjects
210200 -- Power Reactors
Nonbreeding
Light-Water Moderated
Nonboiling Water Cooled
22 GENERAL STUDIES OF NUCLEAR REACTORS
220400* -- Nuclear Reactor Technology-- Control Systems
DETECTION
FAILURES
LOFT REACTOR
MEASURING INSTRUMENTS
PERFORMANCE
PRESSURIZERS
PWR TYPE REACTORS
REACTORS
RESEARCH AND TEST REACTORS
TANK TYPE REACTORS
TEST REACTORS
WATER COOLED REACTORS
WATER MODERATED REACTORS