An integrated approach to sensor FDI and signal reconstruction in HTGRs. Part II. Case studies

Abstract

This second paper in a two part series presents the application and validation of an integrated approach for sensor fault detection and isolation (FDI) in high temperature gas-cooled reactors (HTGRs) in the form of two case studies. In the first case study the integrated approach is applied to actual data retrieved from the prototype pebble bed micro model (PBMM) plant. This case study is particularly important to gain valuable insight regarding the statistical parameters used for the statistical and fuzzy-rule based algorithms. In the second case study, the integrated approach is applied to a complete and detailed Flownex® model of the new generation single-shaft, Brayton cycle-based pebble bed modular reactor (PBMR). This case study provides further insight on the actual performance of the integrated approach applied to a more realistic representation of an actual HTGR plant. In part I of this two part series, the theoretical concepts and methodology of the integrated approach to sensor FDI and signal reconstruction are given. This paper attempts to investigate if the proposed approach can effectively detect and isolate sensor malfunctions such as sensor bias or offset, dead sensors, excessive noise, sensor drift and random spikes. In addition, the ability of the method to diagnose multiple simultaneous sensor failures is also evaluated. It is illustrated that this method is especially suited for safety critical systems such as nuclear power plants where sensors can only be removed or tested once the plant is open for maintenance. This is usually only once every fuel cycle. Therefore, by implementing the proposed method on-line, sensor readings can be validated in real-time. Another advantage of the method is the ability to reconstruct faulty sensors. This will enable the control of safety critical components despite sensor failure