Trans-Scale Damage Evolution Modeling For Brittle Materials And Structures Containing Micro-Cracks

As a complex and heterogeneous material, concrete, as well as its rupture mechanism, is an old and new subject for scientific researches. In the thesis, a comprehensive study was performed on concrete rupture simulations, which covers single-aggregate ruptures, material-level ruptures, component-level ruptures, as well as fatigue ruptures. During this process, we developed a numerical model for trans-scale propagations of multiple cracks, and a numerical model for auto propagations of interface cracks. In the thesis, programs was developed on the platform of commercial finite element software ABAQUS via its Python language interface. And a relatively mature program consisting of several modules had been completed.To be specific, contents below are mainly included in the thesis.1. A numerical model for auto propagation of interface cracks was developed.During the research on the rupture behavior of a matrix containing one single aggregate and an initial crack on its surface, we developed a numerical model which could manipulate free propagation of interface cracks. Since current commercial softwares were unable to directly realize the simulation of interface crack propagation, we designed an algorithm and implemented it on ABAQUS through its Python script interface. This model could successfully simulate the growth of the interface crack, including both on-surface growth and the transition into the matrix.2. On rupture of a matrix containing an isolated aggregate.Base on the model above, the rupture simulation of a matrix containing an isolated aggregate (cell) was conducted. The influence of three factors, including the side-edge constraint, the aggregate direction and the fracture energy of the interface, was investigated. It is found that tensile constraint on the side edge, a smaller angle between tensile axis and the aggregate, as well as a higher fracture energy could lead to a higher rupture strength of the interface. It is also found that once the interface starts to grow, it immediately and unstably propagate to the two ends of the aggregate major axis, and then enters the matrix-probably the load may need to increase at this time. The three factor influences less on the character of above rupture path. With this in mind, the interface could be treated as normal crack by mapping it into the major axis so that the troublesome problem of modeling interface cracks within the complex meso-scale structure of concrete could be avoided. This conclusion is also one import basis of the model of trans-scale cracks propagation. The strong brittleness of matrix rupture indicates that the mechanical behavior of a cell could not be spread to concrete at the material level.3. A numerical model for trans-scale propagations of multiple cracks was developed.In the thesis, a numerical model for trans-scale propagations of multiple cracks was developed. With only several simple principles and directly based on the theory of linear-elastic fracture mechanics, the model could successfully simulate the micro crack coalescence, the formulation and growth of macro cracks, and the final concrete rupture. In this process, we solve a key problem, that is how to simulate the merging and intersecting of multiple cracks. Also, the program is developed on the platform of ABAQUS via Python scripts.4. Numerical experiments on tensile specimens and tensile rupture mechanisms.Above multiple-crack model was applied on tensile rupture modeling, and the influence of aggregate volume fraction on concrete tensile rupture, as well as the rupture mechanisms, were investigated. Results show that, tensile rupture exists two modes:single major-crack mode and multiple major-crack mode. Results also show that with the increase of aggregate volume fraction, the strength has a V curve-first decrease and then increase. Aggregate volume fraction could improve the concrete ductility. Generally speaking, increasing the volume fraction can improve the mechanical property of concrete. However, the fraction should be greater than 40%, or the concrete may have no advantage on the concrete with smaller aggregate volume fraction.5. A multi-scale model containing concurrently macro and micro-scale domains is developed.Inspired by the concurrent multi-scale modeling strategy, we developed a mutil-scale concrete model, in which macro meshes and micro meshes coexist. Here the micro meshes is the multiple-crack model mentioned above. Mirco and macro meshes were introduced in the hot and the non-hot regions, respectively, while the interface between the two has to hold displacement coordination. The multi-scale model is mainly for modeling concrete beam ruptures.6. Numerical experiments on concrete size effect.Using above concurrent multi-scale model, numerical experiments on concrete size effect was performed, and the mechanism with respect to damage trans-scale evolution was discussed. Results show that the increase of specimen size leads to smaller nominal strength, higher fracture energy and smaller data dispersion. Carpin-teri’s scale law seemly fits our results better than Bazant’s scale law. Damage trans-scale evolution may act as an import reason for the size effect.7. Fatigue rupture modeling of concrete beam.By introducing the crack fatigue propagation law, above multi-scale model were improved to accommo-date fatigue rupture simulations. The main character of the fatigue rupture model is that multiple cracks, rather than just the critical one, will propagate simultaneously in a computational step. Identical specimens applied with different fatigue stress levels were computed. Results show that the obtained S-N curve of the concrete beam fits well with actual situations. Also the computed failure mode agrees well with the experiments.