Quantification of stochastically stable representative volumes for random heterogeneous materials (Q2505802): Difference between revisions

The paper investigates the problem of determination of the representative volume element (RVE) size for random heterogeneous materials. The RVE size is found as the minimum size of microstructure cell that fulfills the requirement of statistical homogeneity (i. e. is lower bound). The criterion for determining the RVE size takes into account sizes and volume fractions of the inclusions. First, a procedure is proposed to determine the RVE size based on the statistical analysis of numerical experimental data. To find the size of RVE with a given value of volume fraction of inclusions, the chi-square criterion is used. Then, the finite-element method is used to simulate the response of the unit cell and three-nodal triangular elements are applied. The investigation is carried out by using three-phase material consisting of matrix, inclusions and an interfacial transition zone surrounding each inclusion. The RVE size sensitivity is studied in relation to test parameters (i. e. to loading schemes -- tension and shear) and material parameters. The results show that the RVE size hardly depends on the loading scheme, and the dependence on the parameter of interest (stress based or stiffness based) is not significant. At the same time, there is a strong dependence of the RVE size on the material properties, e. g. changing stiffness ratio. Due to the stochastic deformation process is considered, the stability of the RVE size is redefined in stochastic terms. To distinguish deterministic and stochastic types of stability, it is introduced a new concept of stochastic stability (direct hit stability, DH-stability) coming from the idea that considered process `hit' the initially defined region of permissible values. This region is determined as a result of physical, mechanical and geometrical properties of the process and prescribed error tolerance. The strict definition of DH-stability is considered together with known definitions of Lyapunov stability and Lyapunov stability in probability.