Dr. Don McGlinchey
Pneumatic conveying systems have been seen by many as a low-technology plant item with mechanically simple components and unsophisticated instrumentation and control. In reality, the physics and mathematical description of pneumatic transport is complex and challenging. Failure to recognize the underlying complexity can lead to inappropriate choices in system selection, specification, and operation.
The majority of systems in the industry are designed and operated in a dilute- or lean-phase where gas velocities are relatively high and the loading of solids is low. There are several options to make these systems work reliably. Larger-bore pipelines and conveying gas velocities well above the minimum make for safe, if inefficient, choices. The current economic climate and the cost of energy are driving a move to closely examine conveying processes. This results in attempts to make pneumatic conveying more energy efficient, either by
optimizing dilute-phase systems or by moving to dense-phase conveying. However, optimizing a dilute-phase pneumatic conveying system is trivial compared with the danger of making the system unstable and liable to blockages. The margin of error in designing dense-phase systems is considerably less. Therefore, there is a need to establish a reliable design practice based on a fundamental understanding of the physical mechanisms in a conveying system.
Understanding and describing the behaviors across the length scales requires innovation in experimentation, measurement techniques, mathematical modeling, and numerical simulation. It is convenient to discuss in general terms these length scales as micro-, meso-, and macro-level. Micro-level indicates the processes occurring at the individual particle level, like particle-particle contact mechanics. The macro-level takes more of a system, or global view, like system pressure drop. The meso-level lies somewhere in between.
At the micro-level, much of the industrial and academic research effort is being directed into modeling the behavior of assemblies of particles by the discrete element method (DEM). In this method each particle is treated as an element in the model and the motion of the particle can be described by Newtonian mechanics. Commercial DEM codes that couple through drag terms to computational fluid dynamics (CFD) software are now available. Fluid-particle interaction in a micro-scale has been studied using CFD adaptive meshing, Lattice Boltzmann simulation, and smoothed-particle hydrodynamics. In these cases the drag and lift are derived from the simulation itself rather than by imposing empirical coefficients. Experimental work with this scale includes such techniques as atomic force microscopy in the investigation of contact mechanics and laser Doppler velocimetry in particle and gas motion. The challenges here are the number of particles, irregular particle shape, and modeling turbulence.
To move up the length scale means sacrificing detail and turning to statistical methods to describe the particle system behavior at the meso-scale. A popular approach is to treat the particulate assembly as a pseudo-fluid or continuum. Here statistical physics and the formulation of a granular kinetic theory have lead to simulations of pneumatic transport. Versions of granular kinetic theory have been implemented in several commercial codes. The model requires continuum properties that can be established experimentally by a bench-scale rheometry tester or from data found from simulations at the micro-scale.
Lastly, there is a macro-scale understanding. Typically this is based on an extension of traditional single-phase fluid mechanics, where details such as momentum loss due to collision or fluid drag are absorbed into some general friction term analogous to a D’Arcy treatment. This can be a whole system term or broken down into pipeline components where they can be compared and provide feedback for meso-scale simulations. However, the terms are valid for one particular material only and are determined experimentally in large test facilities.
If the goal is to achieve a reliable design without needing to continually perform full-scale trials for every new material that comes along, it seems clear that there needs to be models that will take familiar bench-scale data as input, and output the performance characteristics of the material and pipeline system.
Don McGlinchey, PhD, is a reader within the School of Engineering and Computing at Glasgow Caledonian University (Glasgow, UK), where he is active in education, research, and consultancy in bulk solids handling. His research interests are in the areas of characterization of bulk solids, gas-solids flow, and multiphase flow modeling.