No abstract

The purpose of this study was to develop an Expert System (ES) to evaluate Nuclear Power Plant (NPP) operators. The study included a review of the Artificial Intelligence (AI) literature relevant to the NPP and human resource management, knowledge acquisition, knowledge representation, knowledge encoding, and Inference engine. An expert system development tool FuzzyCLIPS6.1 was used to develop a fuzzy rule-based expert system called NPPOEX. NPP operator records obtained from a Chinese NPP were entered into NPPOEX and the system produces a crisp value for Evaluation in the range [0,1], a certainty factor and brief explanations for the result.

No abstract

Fault trees (FT) and event trees (ET) have been used thoroughly in risk analysis and there have been a few published articles outlining how to map FTs and ETs to Bayesian Belief Networks (BBN). There have been documented benefits of a BBN being able to consider Common Cause Failures (CCF) and conditional dependencies. With modelling CCFs in a BBN, there is a possibility to increase the level of analysis of a CCF by breaking down the analysis to the respective CCF Categories, such as Environment, Maintenance or Design. This allows a better understanding of the contributing events given a defined accident scenario. Also, in the decommissioning industry, there is no established database yet for CCF of components, as decommissioning projects are sparse and spread out across different operating conditions. Hence it may be practical to adjust generic CCFs to obtain facility-specific parameters for common cause failures. The paper thus highlights how to express CCFs with a Beta-Factor Model in a BBN and by extension, undertake an extended level of analysis according to CCF categories and adjust generic database common cause factors to a facility-specific factor based on a checklist. The technique is applied to a risk analysis of a well plugging and abandonment event.

No abstract