379947 Flow in Irregular Geometries Simulated By Dissipative Particle Dynamics Guided By Body Forces,

Tuesday, November 18, 2014: 5:30 PM
Marquis Ballroom A (Marriott Marquis Atlanta)
Bruce Caswell, School of Engineering, Brown University, Providence, RI, Mingge Deng, Division od Applied Mathematics, Brown University, Providence, RI, Alireza Razdani, Division of Applied Mathematics, Brown University, Providence, RI and George Em Karniadakis, Division ofApplied Mathematics, Brown university, Providence, RI

Fluid flows simulated by Dissipative Particle Dynamics (DPD) have been successfully demonstrated for several geometries. However, efficient convergence starting from rest is achievable mainly in regular geometries, such as rectilinear channel flow driven by a spatially uniform body force applied to each DPD particle. A continuum solution of such flows is achievable with equal ease whether with a body force driver or by specification of normal pressures and inflow/outflow across the channel section. For particle models random fluctuations are an essential feature, and a body force derived in the same domain can function as a local guide to drive the particles in the general direction defined by the boundaries. In This work we derive the guiding force from the pressure gradient of the Navier-Stokes (N-S) system in the same domain. The equations governing the flows of complex fluids in irregular geometries are generally troublesome to solve whether by a continuum constitutive model or by a particle model such as DPD, SPH, etc.. If N-S derived body forces drive a complex fluid then the calculated V-P fields will differ from their N-S counterparts because of the disturbances induced by the addition drops, cells, macromolecules, etc. . In this work we demonstrate this concept in the DPD simulation of two flows. The first is a square lattice of counter-rotating line vortices which can be viewed as a periodic square containing four vortices bounded by streamlines. This potential flow satisfies the N-S equation, and its periodicity avoids the need to model real-wall boundary conditions. For a Newtonian fluid of standard DPD particles excellent agreement is found between the analytical and the simulated stagnation-point velocity fields and stresses. Then a bead-spring molecular model with bending stiffness is introduced into the stagnation region. Its response is compared to experiments on actin molecules suspended in the same flow. The second example is flow in a plane channel with a rectangular indentation on one wall. The N-S pressure field is calculated from the Nektar code, and its gradient is the force-guide to study the deposition of platelets suspended in plasma flowing in the same channel.

Extended Abstract: File Not Uploaded