Computer Science · Artificial Intelligence
The Synthesis of Fault Trees
Steven A. Lapp, Gary J. Powers
See on the mapAbstract
Fault tree synthesis models, in order to be useful in assessment and control of risk, must be modular in the modeling of behavior; all possible modes of behavior must be describable; loops should be modeled naturally; range of variables must be greater than 0.1; and the fault tree or Boolean equation must be deduced from the models. Multiply-edged directed graphs models meet the needs of a fault tree synthesis. The Lapp-Powers fault tree synthesis algorithm is based on logical combinations of digraph variables which could cause a particular deviation in an output variable. Given the correct models, the algorithm will generate the correct fault tree for the system. A flow control system and a heat exchanger system are analyzed, as examples of using the models and the algorithm.