(416a) Inertial Migration of Neutrally Buoyant Prolate and Oblate Spheroids in Plane Poiseuille Flow Using Dissipative Particle Dynamics Simulations | AIChE

(416a) Inertial Migration of Neutrally Buoyant Prolate and Oblate Spheroids in Plane Poiseuille Flow Using Dissipative Particle Dynamics Simulations

Authors 

Larson, R. - Presenter, University of Michigan
Marson, R. L., University of Michigan
Huang, Y., Harbin Institute of Technology
Inertial migration of single neutrally buoyant prolate and oblate rigid spheroids in a plane Poiseuille flow at Reynolds number is studied using dissipative particle dynamics (DPD) simulations. The particles have aspect ratios AR (ratio of the major axis to the minor axis) from unity to three for prolate and from unity to eight for oblate spheroids, and the ratio of the diameter of the major axis to the gap L/H ranges from 0.2 to around 0.4 for both types of the spheroids. The equilibrium positions move closer to the channel center as Re and L/H increase as seen in previous experimental studies, but we also find that differences for oblate vs. prolate shape at fixed L/H, deemed negligible by previous studies, are when L/H > 0.3 and which were not previously explored. The major axis of a prolate spheroid generally tumbles in the flow-gradient plane, while an oblate spheroid changes from log-rolling with the minor axis in the vorticity direction (AR = 1.5), to inclined rolling with the minor axis tilted at an angle with respect to the vorticity axis (AR=3), to a mixed state of inclined-rolling and inclined-tumbling (AR=5), and finally to an approximate steady state (AR=8). The utility of the DPD methods for exploring inertial migration phenomena of particles of various shapes, along with the limitations of the DPD method, especially its limitation to Re between around 50 and 500, are explored. In particular, we find that achieving low Re, while retaining high Peclet number Pe to reduce Brownian noise requires a high Schmidt number Sc, but conventional methods of achieving high Sc in DPD simulations also reduce the resolution of thin boundary layers, making migration predictions inaccurate by DPD. On the other hand, at high Re, the Mach number Ma becomes large, thus limiting the Reynolds number to Re < 500.