(310f) Predicting Segregation In Handling Processes Using Hybrid Discrete and Continuum Approaches

One of the main causes for unscheduled plant shut-down is segregation of the mixture within the system.  Individual components can segregate due to sifting of fines through a matrix of course particles.  Differences in repose angle can induce material separation during pile formation.  Fines may be carried by air currents and deposit in the process equipment where gas velocities decrease sufficiently to drop the particles out of the gas stream, causing separation as material fills a process vessel.  Each of these mechanisms has a segregation pattern that is characteristic of the segregation profile.  Many theories describing segregation are limited to two components.  Multiple components often complicate the interpretation of segregation mechanism.  In addition, prediction of segregation in process vessels is a complicated process especially when dealing with multi-component mixture or multiple mechanisms.  To understand the influence of segregation in process vessels one must be able to compute the velocity profiles in process equipment during flow coupled with a mathematical model describing the cascade action of the free pile surface during both filling and emptying cycles.  The effect of multiple processes in series can complicate the segregation issue, making prediction using simple rules ineffective.  This paper presents a methodology to predict segregation in real processes using results from a novel segregation test method, standard flow properties, process geometry, and operation parameters.  The net result of this approach is the ability to predict the segregation expected as material passes a prescribed location in the process.  This approach should be valuable to process engineers to aid reliable process design.