In recognition of advances in simulation and modeling of granular flows, the International Fine Particle Research Institute (IFPRI) is currently developing a collaborative program in which IFPRI will make available specific datasets from well controlled experiments wherein flow fields are well defined and stresses are measured over scales ranging from individual particles, chains and clusters of particles, and bulk particulate flows. The datasets include both model and industrially-relevant materials with distributed characteristics (e.g., size and shape) and a range of material properties (stiffness, restitution, etc). The challenge of the program is to develop theoretical approaches and models to describe the range of flows measured in the IFPRI experiments and validation thereof.
Broader Impact - in the mutual interest of Industry, Academia and Sustainability:
Industry requires useful methods to predict stress and flow behavior of granular and powder materials across industrially relevant flow regimes, as a function of operating conditions, material properties and particulate characteristics. In the trend toward sustainable processing and energy efficiency, industry strives toward the minimization of specific energy input in granular and powder processing, i.e., minimal power consumption at maximum production rate. The fundamental objectives of this project address this goal by providing a robust description of the physics of dense flows on particle and bulk scales, including energy dissipative interactions of realistic materials.
As a foundation for predictive methods, the broader community of industrial and academic researchers, modelers and engineers need to define and understand the relevant regimes of dense particulate flow, the underlying physics therein, and the effect of boundary conditions, material properties and particle characteristics.
This project was a collaborative effort involving NSF, the International Fine Particle Research Institute (IFPRI), and individual researchers. The short-term goal of the project was to test the reliability of computer models for granular flows by comparing calculations with current models to detailed, quantitative experimental data. The longer term goal was to improve models and to better understand how to accurately and cheaply describe granular flows of all types. The rationale for this kind of test is the following. Industrial users of powders and other granular-like materials lack models that are reliable, predictive, and inexpensive. Consequently, they must frequently rely on estimates and 'rules of thumb' when they design, implement, and manage industrial granular facilities. The financial consequences of the lack of good modeling tools are extraordinary. The value associated with granular materials and their handling has been estimated to be close to a trillion dollars annually in the US alone. Failures of industrial facilities are all too frequent and decidedly both costly and dangerous. Designers and operators of granular devices must deal as best they can with such failures, and the costs of failures are passed on to consumers. The basic approach of this project was as follows. A group of IFPRI PI's and representatives served as coordinators who provided experimental data developed by IFPRI PI's Behringer at Duke University and Tardos at City College, to modelers. In general, participation was open to any modeler, but the NSF-IFPRI collaborator specifically identified modeling experts who each received a modest amount of funding with the goal of modeling one or both of the Duke/City College experiments. (Both experiments involved flow out of a hopper-like container, which is a kind of flow that is particularly germane to industrial needs.) At the end of roughly one year, the collaboratory assembled in Chapel Hill, NC, to discuss and assess the results of the modeling. Several conclusions from this process are of particular importance. First, it was found that good agreement between models and actual physical devices is possible. In order to achieve such agreement, the model must provide an appropriate component to describe the frictional interaction between grains. With the continuing development of computational power, it seems possible that well chosen models should be able to make predictions for at least some classes of flows. However, models that are completely general are still out of reach. A detailed paper describing the process and results of the collaborator is in preparation. This paper will describe successes and shortcomings of the current state of models. Second, the collaboratory approach developed here may also be a useful funding model that can bring together a broader base of a academics and industrial partners. Finally, this project provided support and training to seven graduate students, of whom a number were women.
