Modeling and simulation of cohesive granular flow

Abstract

Granular materials are found in various industrial and daily life applications. Among the different behavior shown by these systems, mixing and segregation have caught the most attention. In contrast with the mixing of fluids, a significant complicating factor in granular systems is their tendency to segregate or demix as a result of differences in the properties of the individual grains. Segregation studies have been limited and a fundamental understanding of the mechanisms involved is still lacking. Most studies on granular segregation have focused on systems that are non-cohesive. However, powder cohesion is a norm rather than exception. Understanding mixing and segregation in cohesive systems will not only improve the design of blending equipment used in the pharmaceutical industry, but also facilitate controlling the product properties and process parameters for various gas-solid fluidization processes and polymer processing applications.

The objective of this work is to study the flow properties of granular systems, which lead to segregation or mixing. A two-dimensional Monte Carlo simulation as developed by Rosato (1985) has been used. According to this technique, particle movements are generated by means of random perturbations to their location, and the tendency of the system to segregate during this flow is investigated. An approach for incorporating different inter-particle forces such as electrostatic, frictional or mechanical interlocking, and liquid bridge forces has been successfully proposed. Suitable models for these forces have been selected from literature and their effects on particle movement.

Simulations were run with different magnitudes of cohesive forces and their effects are observed. Particle percolation seems to be the most dominating mechanism for granular segregation in non-cohesive systems. In the absence of inter-particle forces the bed segregates due to size differences. Differences in the density of particles does not cause appreciable segregation in the bed. When cohesive forces are acting, transitions in segregation pattern are seen and the overall model predictions are in line with experimental observations recently reported in the literature. The simulation offers the possibility to explain these behaviors by tracking motions of individual particles and evaluating the nature and magnitude of cohesive interactions at the particulate level. The evaluation of force profiles of inter-particle interactions within the region of particle movements was found to be useful in the analysis of the results and guidelines were proposed for predicting bed behaviours in various blending operations. While very few experimental studies have been conducted to study the effect of cohesive forces on the nature of granular flow, quantitative information about the development of segregation patterns in cohesive systems will also be very important for the determination of model parameters used in the simulation of granular flow and for further validation of the proposed simulation.