A Relative Examination of CD — Re Relationships Used in Particle Trajectory Calculations

Abstract

For a large variety of gas-particle flows, realistic simulation of particle behavior in the flowfield is needed to calculate accurate particle trajectories. The accuracy depends largely on the flowfield (loading, compressibility, unsteadiness, turbulence level), the forces acting on the particles (drag, gravity, buoyancy, Basset, Saffam lift, apparent mass, thermophoresis), and the particle size, density, and shape. Previous investigators have employed simplified (and indeed justified) models in which spherical particles move in a steady, incompressible, single phase flow. In trajectory calculations, the flowfield is known in either an analytical or a discrete (mesh) form. The particle equations of motion are integrated numerically using appropriately small steps. The predominant force acting on a particle is the drag force represented by a drag coefficient, C[sub D], whose value depends mainly on the particle Reynolds number, Re[sub p]. At each step during the trajectory calculation, its value is input to the equations of motion that are being integrated. The dependence on Re[sub p] asks for accurate C[sub D]-Re[sub p] relationships. The present note examines the relative accuracy of 19 such relationships met in the literature. Two sample flowfields are considered: (1) a swirling flowfield similar to that found in cyclone separators and (2) amore » cylinder in crossflow. The accuracy of fit of the various formulae to experimental data is reflected on the trajectory results, which in turn may be critical in the design process and performance prediction analysis. It must be here emphasized that trajectory predictions compared to the mean prediction of 19 formulae do not necessarily define a measure of accuracy or confidence of a particular drag formula, but only gives a measure of standard deviation.« less