Use of rule-based methodology to develop and quantify event trees

The development of event trees during probabilistic risk assessment particularly when using the large-event-tree/small-fault-tree methodology, represents a large amount of effort. While it is desirable to accurately and comprehensively address plant conditions within the plant model, it is also advantageous to use efficient calculational tools to develop the plant model itself. Previous efforts to improve the quantification of large event trees have centered on condensing the event tree itself into a series of transfer, which may eventually require the review of literally thousands of scenarios when using the full event tree. Also, the plant model logic can reside either in the actual structure of the event tree, when the analyst removes an event tree node, or in the use of conditional split fractions, such that a given node branch has a zero frequency. The use of a rules-based methodology allows the development of a very compact, yet comprehensive framework for risk model development and quantification. Due to the compact nature of this methodology, several event trees can be linked and quantified in a single pass, removing the need to manually bin support system impact vectors. A rules-based methodology also has interest for practitioners of the small-event-tree or linked-fault-tree methodology.