Abstract:

In order to process inexact failure probability efficiently, a novel interval analysis method of fault trees is proposed. The plausibility function and belief function in evidence theory are used to calculate the interval probability of basic events. A hyper-ellipsoidal model is introduced to structure the interval operators, by which the uncertainty propagation is implemented, and the Monte-Carlo simulation method is used to calculate conditional extremum. The proposed method can be applied when statistical data are incomplete. The good performance of the proposed method is demonstrated by theoretical analysis and numerical results.