An extended two-scale random field model for stochastic response analysis and its application to RC Short-leg shear wall structure

Abstract

For the stochastic response analysis of concrete structures, it is necessary to model the heterogeneous concrete with significant spatial variability by random fields. In practical applications, multiple material parameters are typically required to characterize the nonlinear constitutive relationship of concrete. As a result, the material parameters of concrete will be represented by a multivariate random field in the nonlinear stochastic response analysis. Nevertheless, a vast number of random variables are frequently involved in representing the multivariate random field, and it is difficult to quantify the probabilistic dependence among the material parameters. To address the preceding issues, this paper proposes an extended two-scale random field model (ETSRF) to quantify the spatial variability of the constitutive relationship of concrete. The model can accurately depict the fluctuation of material properties at the micro-meso scale and the spatial correlation at the macro scale. Besides, it is proved that there is a linear relationship between the mesoscopic random field and ETSRF. Furthermore, a new framework of stochastic response analysis is proposed, where the probability density evolution method (PDEM) is applied to obtain the probabilistic information of the structural responses, and ETSRF is adopted to describe the spatial variability of the material properties. Finally, the proposed model is validated by two illustrative examples, one of which simulates the random fields of compressive and tensile strength of a plain concrete slab, and the other of which deals with the stochastic response analysis of a RC short-leg shear wall structure. The results indicate that the proposed model can depict the spatial variability of the nonlinear constitutive relationship of concrete effectively.