Copula Simulation of Mechanical Structural Systems Failure Mode

Abstract

As a new and efficient tool of statistical analysis, Copula has the capability of putting multivariate random probability models into practice, and describing time-varying and nonlinear feature of statistcal dependence of random variables. Aiming at mechanical structural systems with failure-dependence, a new method is put forward for reliability modeling by introducing mixed copula, in which giving a reference to several copula functions and its applicable feature of the correlation. Copula model of joint probability distribution function is build between every functions, in which every functions value are chosen as the analytic variables of copula function and unknown parameters of copula are estimated by means of the random sampling data generated by Monte Carlo simulation in MATLAB software, thereby failure probability of the structural system is got by copula function. Finally, a practical case of copula demonstrates the feasibility of the method , which give a new path for the reliability analysis of mechanical structural system with correlated failure modes.