^{a}, Zhiqiang Wan

^{b}and Jie Li

^{c}

^{a}chenjb@tongji.edu.cn

^{b}wanzhiqiang@tongji.edu.cn

^{c}lijie@tongji.edu.cn

Reliability assessment of real-world structures is of paramount importance. When dealing with the problems involving uncertain parameters, e.g., the mechanical properties of materials or components, usually the independence between the parameters is assumed for the sake of simplicity. However, it is noticed that the mechanical properties are usually not independent, nor completely dependent, but partly dependent or correlated. Therefore, how to reasonably represent such partial dependence and whether such partial dependence real matters in stochastic response and reliability are crucial issues. In the present paper, the underlying physical mechanism of dependent random variables is stressed in the quantification of such dependence. Specifically, the modeling of correlation between the compressive strength and the initial modulus of elasticity of concrete along with its effect on structural responses are addressed based on the probability density evolution method. Numerical examples are illustrated as well.