This report describes the methodology and advances reached for the project “Micro-Inspired Continuum Modeling Using Virtual Experiments.” The objective of this project was to develop new hardening rules for the most common plasticity constitutive models by using accurate micromechanical simulations that can capture the main features of granular behavior under cyclic loading — a crucial feature for earthquake engineering applications. To achieve this objective, laboratory experiments were performed and scanned with X-ray computed tomography (CT). The sample processed with CT was used as input for simulations using the Level Set Discrete Element Method (LS-DEM), developed by our group. This allowed a one-to-one comparison between the individual grains of the sample and the digital twins used in the simulation. Based on our results, it was possible to observe the same dilatancy, shear banding, and cyclic changes also found in experiments. Moreover, the simulation provides the convenience of tracking the evolution of force chains and fabric tensor descriptors under loading, which is not observable from an ordinary triaxial test. The evolution of this mesoscale structure was studied to determine how it can be linked to constitutive models. After the experiments and simulations were compared, additional analyses were done to examine fabric changes as a result of cyclic loading. Different types of samples under varied configurations were tested with LS-DEM and multiple cycles of shear and compressional loading were applied to them. As a result, evolution in coordination number, void ratio, contact fabric tensor, and force weight contact tensor were found to develop within the grains and, in some cases, cumulatively leaned to a densification of the sample over time (change in macroscopic properties from micro-scale alterations). Further testing and calibration can then be used to develop continuum formulations that relate the fabric tensor and its descriptors to cyclic constitutive models.
Two-page summary: click here.
Full List of PEER Reports: click here.