FAILURE-DEPENDENCE MECHANISM AND COPULA-BASED RELIABILITY MODEL OF STRUCTURAL SYSTEMS

For most engineering structural systems, reliability modeling becomes complicated because of failure-dependence of components. Aiming at this problem, a new method is put forward for reliability modeling of structural systems with failure-dependence by introducing copula, an powerful tool in multivariate statistical analysis. Considering the typical mechanism of failure-dependence, safety margin and fatigue life is chosen as the analytic variable for copula-based modeling of static-strength reliability and fatigue reliability respectively. In practical situation, unknown parameters of mixed copulas are determined by means of maximum likelihood method according to existing failure data. A simple example is given 10 illustrate basic steps of reliability modeling based on copulas. Advantages of the copula-based model are that most copulas possess flexible and concise expression, and there are specific relations between copulas and many dependence measures such as rank dependence coefficient, tail dependence coefficient by which nonlinear and time-varying dependence structure of components failures can be depicted finely. The new models give a new path for reliability assessment of engineering structural systems with failure-dependence.