Fault Tree Analysis for Large Systems
John Andrews · Encyclopedia of Statistics in Quality and Reliability · 2007
Abstract This article describes the binary decision diagram (BDD) technique of fault tree analysis. The BDD provides an alternative to the traditional fault tree analysis method of kinetic tree theory which is progressed in two phases. The first phase is a qualitative stage to determine the minimal cut sets. The second stage then uses the minimal cut sets together with the component failure probabilities to conduct the system quantification. For large fault trees, this process is inefficient. While fault trees express the system failure logic in a convenient way from the engineering viewpoint, they do not lend themselves to the mathematical manipulation. BDDs express the failure logic in a disjoint form, which makes quantification very fast. The calculations require just one pass through the BDD structure. Minimal cut sets are not required as an intermediate step, although there are methods by which they can be obtained if they are of interest. Calculations are exact and do not need any form of approximation. The trade‐off for the gains in speed and accuracy is that the fault tree has to be converted to a BDD. The advantages of this are more evident on large, complex fault tree structures.