PRODIAG: A process-independent transient diagnostic system - II: Validation tests

The unique capabilities of the first-principles-based PRODIAG diagnostic sy stem to identify unanticipated process component faults and to be ported ac ross different processes/plants through modification of only input data fil es are demonstrated in two validation tests. The Braidwood Nuclear Power Pl ant full-scope operator training simulator is used to generate transient da ta for two plant systems used in the validation tests. The first rest consi sts of a blind test performed with 39 simulated transients of 20 distinct t ypes in the Braidwood chemical and volume control system. Of the 39 transie nts, 37 are correctly identified with varying precision within the first 40 s into the transient while the remaining two transients are nor identified . The second validation test consists of a double-blind test performed with 14 simulated transients in the Braidwood component coolant water system. I n addition to having no prior knowledge of the identity of the transients, in the double-blind test we also had no prior information regarding the ide ntity of the component faults that the simulator was capable of modeling. A ll 14 transient events are correctly identified with varying precision with in the first 30 s into the transient. The test results provide enough evide nce to successfully confirm the unique capabilities of the plant-level PROD IAG diagnostic system.