Modeling the collisional-plastic stress transition for bin discharge of granular material

We propose a heuristic model for the transition between collisional and frictional/plastic stresses in the flow of granular material. Our approach is based on a physically motivated, nonlinear 'blending' function that produces a weighted average of the limiting stresses, depending on the local void fraction in the flow field. Previously published stress models are utilized to describe the behavior in the collisional (Lun et al., 1984) and quasi-static limits (Schaeffer, 1987 and Syamlal et al., 1993). Sigmoidal and hyperbolic tangent functions are used to mimic the observed smooth yet rapid transition between the collisional and plastic stress zones. We implement our stress transition model in an opensource multiphase flow solver, MFIX (Multiphase Flow with Interphase eXchanges, www.mfix.org) and demonstrate its application to a standard bin discharge problem. The model's effectiveness is illustrated by comparing computational predictions to the experimentally derived Beverloo correlation. With the correct choice of function parameters, the model predicts bin discharge rates within the error margins of the Beverloo correlation and is more accurate than one of the alternative granular stress models proposed in the literature. Although a second granular stress model in the literature is also reasonably consistent with the Beverloo correlation, we propose that our alternative blending function is likely to be more adaptable to situations with more complex solids properties (e.g., 'sticky' solids).