Particle Flow Behavior in a Shear Cell Device

This presentation describes ongoing research performed on a simple shear cell apparatus, previously described [1]. As a complement, discrete particle simulations and continuum models have been used to predict normal and shear forces in the ongoing experiments. The trends and orders-of-magnitude of the models and experiment are in basic agreement. Theoretical models used are constructed with basic principles, rather than curve fitting, to obtain effective properties of the mixture such as viscosity or conductivity. The experiment itself serves to determine the effects of shear rate, packing fraction, particle size and film thickness on the load carrying normal stress. Additionally, the frictional shear stress can be investigated. The working particulate medium within the apparatus consists of glass of aluminum spheres, poly-dispersed over four size increments, all less than 1.00 mm diameter. The upper annular disk is held stationary in a rotational sense by a force transducer, and applies predetermined normal stress values which vary according to a system of interchangeable counterweights. The lower transfer surface and the sidewalls of the annular ring are rotated by applied mechanical torque. Experimental trials consist of shear initiation, after which the trough velocity, film thickness, supporting load, and frictional torque are measured. From these measurements one can calculate the average shear rate, the average load-carrying normal stress, the average frictional shear stress, and the solids volume fraction. Such third body granular flow may apply to some solid lubrication mechanisms, and to applications such as smart clutches and dampers. The continuum theory presented is unique in that it addresses solid-like behavior and its transition to fluidized behavior. The discrete particle dynamics rely on the conceptual models of Iordanoff and colleagues [2]. Our findings are that the two theoretical predictions agree reasonably with the experimental results, suggesting validity of the approach. These results are promising, and may be used to further develop high level predictive models. Furthermore, similar methods of small scale experimental particle simulation can be used to develop simpler more usable continuum approaches.