Research On Dam Failure Risk Analysis Using Fault Tree And Bayesian Network

Dam failure is a kind of society disaster factor, which has low probability and high hazard. Our country is the big country of dam-maker. The dams are found everywhere and play on important role in our country development. But potential hazard is also high. Once dam failure happen, it will bring serious damage. Along with the old dams and the sick dams getting more and more, the dam security causes people's high attention. Meanwhile the dam risk annlysis is outspread rapidly through the world.This paper introduces comprehensively the basic theory and several kinds of calculation methods of dam failure risk analysis and their probability. Based on the fault tree principium, the relations of different events and burst of dam failure are analyzed. Thus, the models of earth-rock dam failure and tailing dam failure are established. To supply a theory platform for developing a software of dam failure risk analysis.In view of the absence of fuzzy analysis in the traditional risk analysis on dam failure, fuzzy risk analysis, which, together with randomness, constitutes the two inseparable uncertainty elements, deserves exposition. The theory of fuzzy set is applied to the fault tree analysis and the model of the fuzzy fault tree analysis is established based on the theory of risk analysis. And the fuzzy fault tree analysis based on triangular fuzzy number arithmetic is discussed. The methods to transform various kinds of fuzzy number into triangular ones are gived, which are adaptive to the need of practice. Finally, an example is gived to illustrate the method. Because of considering fuzzy set theory, the result is more reasonable.Finally, the fault tree analysis and Bayesian network are compared with each other. A procedure to transform a fault tree into a Bayesian network is presented. Then the Bayesian network approach based on fault tree is used to evaluate the reliability of earth-rock dam failure. The results indicate that the relationship between the causes and consequences for dam risk analysis can be expressed in an intuitive format using the Bayesian network. The common cause failure associated with reliability evaluation of dam can be handled using the Bayesian networks easily. The risk analysis model proposed in the paper is feasible and effective to evaluate the reliability of dam safety.