Computational Homogenization of Elastic-Viscoplastic Refractory Masonry with Dry Joints

Refractory masonry with dry joints is widely used in the steel-making industry for the linings of several high-temperature components (>1500 ∘ C) including steel ladles and furnaces. To properly optimize the design and performance of these linings, thorough numerical models that consider the presence of joints, joints closure and reopening and the nonlinear elastic-viscoplastic behaviour (creep and stress relaxation) of refractories at high temperature are required. The present study reports on the formulation, numerical implementation, and application of a homogenized multi-scale elastic-viscoplastic model for the simulation of refractory masonry linings with dry joints. Refractory bricks are considered to exhibit linear elasticity as well as rate-dependent plasticity. Four joint patterns are predefined based on the state of bed and head joints. The homogenized elastic-viscoplastic behaviour of each joint pattern is determined using finite element based nonlinear homogenization approach. The transition criteria between the four patterns are defined in terms of macroscopic stresses and strains. Verification of the developed homogenized constitutive laws is carried out by comparing the numerical results of the detailed micro models (brick and joints are considered) with the homogeneous equivalent material models. Furthermore, comparisons with experimental results of refractory masonry walls subjected to biaxial compression load at room and high temperature are carried out. Good agreements between the experimental and numerical results are obtained. Then, the validated models are employed to predict the mechanical behavior of refractory masonry structures subjected to different loading conditions. The present numerical model is able to simulate the orthotropic, compressible, rate-dependent homogenized behaviour of mortarless refractory masonry structures, and accounts for joints closure and reopening due to loading and unloading.