Fault detection in boiling water reactors by noise analysis

This paper describes the principles and main ideas in a research project concerning adaptive fault detection and malfunction diagnostics, applied to the surveillance of a Boiling Water Reactor using Noise Analysis. Different methods for fault detection are described, like pattern recognition and innovation methods based on Kalman Filters. An adaptive fault detection scheme is presented. For diagnostics, the use of ''Knowledge based expert systems'' are discussed and some examples are given. The implementation aspects of a small test system under development are presented