Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Civil and Environmental Engineering

First Advisor

George Z. Voyiadjis


Concrete exhibits a significant strain-softening behavior beyond the peak stress, and moreover, the mechanism of inelastic deformation in concrete consists of both plastic slip and microcracking. Hence, a combined plasticity and damage mechanics model is proposed for modeling concrete behavior under both multiaxial monotonic and cyclic loadings. The model adopts a bounding surface concept for both plasticity and damage. The introduced plasticity bounding surface is a function of the maximum compressive strain experience by the material, while the damage bounding surface is a function of the accumulated damage parameter. By this definition, both surfaces shrink in size in the stress space consistently with strain and damage. In this model, the material parameters are identified by fitting well-documented test data. The functional dependence of the material parameters on stress history, $\varepsilon\sb{\rm max}$, and damage parameter allow realistic modeling of the complex cyclic behavior of concrete. The proposed model combines plastic strain with strain due to damage, which account for softening behavior of concrete. Plastic strain components are calculated by using the plastic modulus which is a function of the distance from the current stress point to the bounding surface along the deviatoric stress direction S$\sb{\rm ij}$. Similarly, damage growth rate is obtained by the hardening modulus which is a function of the distance defined above. The hardening behavior of concrete is assumed herein to be controlled by both damage and plasticity, while the strain softening regime is controlled by damage processes only. The simultaneous use of the plasticity surface and the damage surface, leads to a constitutive model that accounts for the essential features of concrete such as pressure sensitivity, shear compaction-dilatancy, and stiffness degradation. Comparison of model predictions with the available experimental data has been made and the results show good agreement. The model is computationally efficient and appears promising for implementation in generalized finite element programs.