Experimental and numerical investigations of statistical size effect in S235JR steel structural elements: Fid-Bau Portal

Experimental and numerical investigations of statistical size effect in S235JR steel structural elements

Li, Zheng / Pasternak, Hartmut

Highlights • Directly input uncertainty of material properties based on the discrete random field. • Statistical size effect exist in the steel structures. • Numerical simulation of statistical size effect for tensile and flexural components. • Material properties obtained with small specimen are unsuitable for large structures.

Abstract This paper presents the analysis of the statistical size effect in steel structural elements using the stochastic finite element (FE) method. The three-dimensional random field theory is applied to simulate the uncertainty of a nonlinear material. The proposed approach, which combines the discrete random field describing the randomness of the material properties via Monte Carlo simulations, is used for computing the bearing capacity variability of the steel structures. In addition, the mapping interpolation method makes it possible to separate the random field mesh and FE model mesh. The three-dimensional random field for the material properties is established according to the yield strength with a lognormal distribution based on reasonable assumptions, and the corresponding random field parameters are studied. Finally, the results of experimental investigation and simulations are compared based on the tensile and bending tests, and they indicate that the statistical size effect cannot be overlooked in S235JR steel structural elements and that the use of stochastic FE in the analysis of the statistical size effect has tremendous potential.