Tuesday, April 12, 2016: 8:30 AM
372 A & D (George R. Brown )
Domino effect has a significant characteristic with low frequency and high consequence, therefore how to prevent secondary accidents is really important especially in chemical park land use planning. However, High complexity and growing interdependences of chemical infrastructures make them increasingly vulnerable to domino effect. Based on problems existed, Bayesian network is used to calculate and model the most probable evolution of domino effect when defining each unit that may be involved in accidents as a node in model. Bayesian network helps reduce the uncertainty of domino effect given posterior probability. The obtained probability and propagation of domino effect is able to assist locate units themselves and neighboring area. The application of the developed methodology has been demonstrated via a fuel storage plant.