The vision of this research study is to exploit physical insights obtained through microscale simulations to develop better and accurate macroscale constitutive models in different regimes of granular flow. Development of these constitutive models at macroscale that incorporates microscale particle interactions, need tools such as, DEM (discrete element method) simulations, to probe microscale behavior. These DEM simulations are helpful in understanding the granular physics and mesoscale descriptors that link microscale particle interaction to macroscopic constitutive behavior.

In order to attain the primary goal of development of constitutive

models, DEM simulations are validated with the experiments in a

Couette shear device. It is found that DEM simulations are capable of capturing the regime transition from quasi-static to the intermediate behavior as observed in the experiments. Influence of microscale parameters on granular rheology is demonstrated using comprehensive regime map established using DEM data. Existence of a third stable granular phase is discovered that is neither completely solid-like nor completely fluid-like. A new modified form of the free energy density function is proposed to capture this third stable granular phase observed in the DEM simulations. Further, a constitutive model based on the order parameter (OP) framework is refined, and a linear model with new model coefficient extracted from data of 3D DEM simulations of homogeneously sheared granular flows is proposed, which is denoted as refined order parameter (ROP) model. Performance of this ROP model along with other existing constitutive models is assessed in the different regimes of granular flow. It is found that the intermediate regime poses significant challenge to predictive capability of the constitutive models. In order to capture this complex rheological behavior of the intermediate regime a constitutive model based on mesoscale descriptors (such as the coordination number and the fabric tensor) that links microscale particle interactions to the macroscale behavior is developed. It is shown that the proposed contact stress model is capable of capturing the correct scaling of the stress with the shear rate even in the intermediate and dense regime of granular flow.