Simulation of Particle Flow in Inertial Particle Separators with Eulerian VR-QMOM Method

Abstract

This paper presents research into practical simulations of particle flow in inertial particle separators typically used in helicopter and tilt-rotor aircraft propulsion systems. The flowfield of the carrier gas is predicted by a Reynolds-averaged Navier–Stokes computational-fluid-dynamics method with the Reynolds-Averaged turbulence model. An Eulerian methodology is used to trace the trajectories of foreign particles such as droplets, ice, and sand. To predict the characteristics of particle wall bouncing in dilute particle flow, the velocity-reassociated two-node quadrature-based method of moments is used. The particle distributions in the inertial particle separator are predicted for various particle sizes, and these are compared with results from a Lagrangian particle-tracking method. The particle–wall interactions and the separation efficiencies are studied for solid particles bouncing off perfectly elastic walls and an inertial particle separator shell coated with the M246 alloy, which changes the coefficients of restitution. The simulated separation efficiencies predicted by the Eulerian method are compared with the simulation using the Lagrangian method over a range of particle sizes. The velocity-reassociated two-node quadrature-based method of moments is seen to reproduce the particle bouncing and trajectory crossing behavior and to agree well with the Lagrangian method for predicted separation efficiencies. The new velocity-reassociated two-node quadrature-based method of moments is shown to be an accurate and convenient alternative to established Lagrangian approaches.