Bulletin of the American Physical Society
70th Annual Meeting of the APS Division of Fluid Dynamics
Volume 62, Number 14
Sunday–Tuesday, November 19–21, 2017; Denver, Colorado
Session Q30: Computational Fluid Dynamics: Immersed Boundary MethodsCFD FSI
|
Hide Abstracts |
Chair: Gianmarco Mengaldo, California Institute of Technology Room: 110 |
Tuesday, November 21, 2017 12:50PM - 1:03PM |
Q30.00001: Immersed Boundary Method for Shallow-Water Flow Solvers Ning Zhang The immersed boundary method (IBM) has been widely applied with Navier-Stoke equation solvers for flows over moving objects or objects with complex shapes. However, the IBM has not been often used with shallow-water flow solvers for estuary modeling applications. In regional scale hydrodynamic simulations, man-made or natural land structures such as levees, floodgates and small rivers/streams often have smaller scales than the grid resolutions in the simulations. Therefore, IBM could be a good candidate to realize the small shapes/forms of those structures on coarser simulation grids. In this study, IBM formulations have been developed to realize the floodgates and small rivers for several 2D depth-averaged shallow-water equation solvers. The research targets coastal areas in southwest Louisiana, particularly, the Calcasieu Lake and the surrounding coastal wetlands. The wetlands are protected by levees to avoid direct floods through the lake shore. The wetland water comes from the frequent floods through many small streams connecting the wetlands with the lake. It is very expensive to have grid resolutions smaller than the sizes of the streams. It is thus a good candidate for an IBM approach. [Preview Abstract] |
Tuesday, November 21, 2017 1:03PM - 1:16PM |
Q30.00002: A highly efficient sharp-interface immersed boundary method with adaptive mesh refinement for bio-inspired flow simulations. Xiaolong Deng, Haibo Dong Developing a high-fidelity, high-efficiency numerical method for bio-inspired flow problems with flow-structure interaction is important for understanding related physics and developing many bio-inspired technologies. To simulate a fast-swimming big fish with multiple finlets or fish schooling, we need fine grids and/or a big computational domain, which are big challenges for 3-D simulations. In current work, based on the 3-D finite-difference sharp-interface immersed boundary method for incompressible flows (Mittal et al., JCP 2008), we developed an octree-like Adaptive Mesh Refinement (AMR) technique to enhance the computational ability and increase the computational efficiency. The AMR is coupled with a multigrid acceleration technique and a MPI$+$OpenMP hybrid parallelization. In this work, different AMR layers are treated separately and the synchronization is performed in the buffer regions and iterations are performed for the convergence of solution. Each big region is calculated by a MPI process which then uses multiple OpenMP threads for further acceleration, so that the communication cost is reduced. With these acceleration techniques, various canonical and bio-inspired flow problems with complex boundaries can be simulated accurately and efficiently. [Preview Abstract] |
Tuesday, November 21, 2017 1:16PM - 1:29PM |
Q30.00003: An immersed boundary method for modeling a dirty geometry data Keiji Onishi, Makoto Tsubokura We present a robust, fast, and low preparation cost immersed boundary method (IBM) for simulating an incompressible high Re flow around highly complex geometries. The method is achieved by the dispersion of the momentum by the axial linear projection and the approximate domain assumption satisfying the mass conservation around the wall including cells. This methodology has been verified against an analytical theory and wind tunnel experiment data. Next, we simulate the problem of flow around a rotating object and demonstrate the ability of this methodology to the moving geometry problem. This methodology provides the possibility as a method for obtaining a quick solution at a next large scale supercomputer. [Preview Abstract] |
Tuesday, November 21, 2017 1:29PM - 1:42PM |
Q30.00004: Some recent developments of the immersed interface method for flow simulation Sheng Xu The immersed interface method is a general methodology for solving PDEs subject to interfaces. In this talk, I will give an overview of some recent developments of the method toward the enhancement of its robustness for flow simulation. In particular, I will present with numerical results how to capture boundary conditions on immersed rigid objects, how to adopt interface triangulation in the method, and how to parallelize the method for flow with moving objects. With these developments, the immersed interface method can achieve accurate and efficient simulation of a flow involving multiple moving complex objects. [Preview Abstract] |
Tuesday, November 21, 2017 1:42PM - 1:55PM |
Q30.00005: A pseudospectra-based approach to non-normal stability of embedded boundary methods Narsimha Rapaka, Ravi Samtaney We present non-normal linear stability of embedded boundary (EB) methods employing pseudospectra and resolvent norms. Stability of the discrete linear wave equation is characterized in terms of the normalized distance of the EB to the nearest ghost node ($\alpha$) in one and two dimensions. An important objective is that the CFL condition based on the Cartesian grid spacing remains unaffected by the EB. We consider various discretization methods including both central and upwind-biased schemes. Stability is guaranteed when $\alpha\leq \alpha_{\max}$ where $\alpha_{\max}$ ranges between 0.5 and 0.77 depending on the discretization scheme. Also, the stability characteristics remain the same in both one and two dimensions. Sharper limits on the sufficient conditions for stability are obtained based on the pseudospectral radius (the Kreiss constant) than the restrictive limits based on the usual singular value decomposition analysis. We present a simple and robust reclassification scheme for the ghost cells (``hybrid ghost cells") to ensure Lax stability of the discrete systems. This has been tested successfully for both low and high order discretization schemes with transient growth of at most $\mathcal{O}$(1). Moreover, we present a stable, fourth order EB reconstruction scheme. [Preview Abstract] |
Tuesday, November 21, 2017 1:55PM - 2:08PM |
Q30.00006: Numerical Simulation of h-Adaptive Immersed Boundary-Lattice Boltzmann Flux Solver for Freely Falling Disks. Pan Zhang, Zhenhua Xia, Qingdong Cai The flow field of a freely falling body contains very complicated unsteady characteristics. In general, there are four types of motion for a freely falling body in a viscous fluid medium, including flutter, tumble, steady or chaos fall, determined by its dimensionless moment of inertia and Reynolds number. In this work, direct numerical simulation (DNS) method is used to simulate the flow field with a parallel computation framework JASMIN. In order to efficiently simulate flows with moving boundaries, an adaptive numerical model is established combining the h-adaptive mesh refinement technique and the immersed boundary-lattice Boltzmann flux solver (IB-LBFS). Our numerical results agree well with the experimental results in all of the six degrees of freedom of the disk. Furthermore, very similar vortex structures observed in the experiment are also obtained. If there is a holes in the disk, an additional counter-rotating vortex ring was found at the disk's inner edge. We use the present method to study the effect of the central hole on the disk's falling mode. [Preview Abstract] |
Tuesday, November 21, 2017 2:08PM - 2:21PM |
Q30.00007: A moving control volume method for smooth computation of hydrodynamic forces and torques on immersed bodies Nishant Nangia, Neelesh A. Patankar, Amneet P. S. Bhalla Fictitious domain methods for simulating fluid-structure interaction (FSI) have been gaining popularity in the past few decades because of their robustness in handling arbitrarily moving bodies. Often the transient net hydrodynamic forces and torques on the body are desired quantities for these types of simulations. In past studies using immersed boundary (IB) methods, force measurements are contaminated with spurious oscillations due to evaluation of possibly discontinuous spatial velocity of pressure gradients within or on the surface of the body. Based on an application of the Reynolds transport theorem, we present a moving control volume (CV) approach to computing the net forces and torques on a moving body immersed in a fluid. The approach is shown to be accurate for a wide array of FSI problems, including flow past stationary and moving objects, Stokes flow, and high Reynolds number free-swimming. The approach only requires far-field (smooth) velocity and pressure information, thereby suppressing spurious force oscillations and eliminating the need for any filtering. The proposed moving CV method is not limited to a specific IB method and is straightforward to implement within an existing parallel FSI simulation software. [Preview Abstract] |
Tuesday, November 21, 2017 2:21PM - 2:34PM |
Q30.00008: Adaptive Mesh Refinement for the Immersed Boundary Lattice Green's Function method Gianmarco Mengaldo, Tim Colonius The immersed boundary lattice Green's function (IBLGF) method, recently developed by Liska and Colonius (JCP, vol. 331, pp. 257-279, 2017), is a recent scalable numerical framework to solve incompressible flows on unbounded domains. It uses an immersed boundary method, based on a 2$^{nd}$-order mimetic finite volume scheme that is used in conjunction with an adaptive block refinement approach, achieved via lattice Green's functions, whose scope is to limit the computational domain to vortical regions that dominate the flow evolution --- e.g. regions in proximity to the immersed body surface and in its wake. The method, as it stands, is competitive for low Reynolds number flows, as the staggered Cartesian mesh employed cannot be stretched or refined locally. In this talk we address this issue by presenting the development of adaptive mesh refinement (AMR) capabilities in the IBLFG method. As we shall see, this new feature and the adaptive block refinement already present in the code help overcoming the limitation of simulating high Reynolds number flows, issue that is endemic to the vast majority of immersed boundary-based methods. [Preview Abstract] |
Tuesday, November 21, 2017 2:34PM - 2:47PM |
Q30.00009: Vortex Particle-Mesh method for flows past bodies: a comparison of the iterative penalization and immersed interface methods. Thomas Gillis, Gregoire Winckelmans, Philippe Chatelain The Vortex Particle-Mesh (VPM) method is well suited for solving advection dominated incompressible flows. However, the efficient and accurate handling of solid boundaries in this method still constitutes an active domain of research. The boundary enforcement conditions the vorticity production at the wall and is thus paramount to the accuracy of the global method. We here focus on two iterative algorithms: the iterative penalization method and the immersed interface method. The iterative penalization uses a mask function over the grid and is therefore quite straightforward to use as it simply adds a penalization sub-step to the time step, already implemented in a body-free method. On the contrary, the immersed interface method uses an interface detection and modification of the numerical scheme for points near the interface. This preserves the order of accuracy up to the boundary, yet it increases the implementation efforts. We compare these two techniques on the benchmark of the flow past an impulsively started cylinder at Re$=$9500. We analyze the solutions of both methods, their accuracy up to the boundary and the related cost. [Preview Abstract] |
Tuesday, November 21, 2017 2:47PM - 3:00PM |
Q30.00010: Accurate solution of the Poisson equation with discontinuities Jean-Christophe Nave, Alexandre Marques, Rodolfo Rosales Solving the Poisson equation in the presence of discontinuities is of great importance in many applications of science and engineering. In many cases, the discontinuities are caused by interfaces between different media, such as in multiphase flows. These interfaces are themselves solutions to differential equations, and can assume complex configurations. For this reason, it is convenient to embed the interface into a regular triangulation or Cartesian grid and solve the Poisson equation in this regular domain. We present an extension of the Correction Function Method (CFM), which was developed to solve the Poisson equation in the context of embedded interfaces. The distinctive feature of the CFM is that it uses partial differential equations to construct smooth extensions of the solution in the vicinity of interfaces. A consequence of this approach is that it can achieve high order of accuracy while maintaining compact discretizations. The extension we present removes the restrictions of the original CFM, and yields a method that can solve the Poisson equation when discontinuities are present in the solution, the coefficients of the equation (material properties), and the source term. We show results computed to fourth order of accuracy in two and three dimensions. [Preview Abstract] |
Tuesday, November 21, 2017 3:00PM - 3:13PM |
Q30.00011: A geometry-adaptive IB--LBM for FSI problems at moderate and high Reynolds numbers Fangbao Tian, Lincheng Xu, John Young, Joseph C. S. Lai An FSI framework combining the LBM and an improved IBM is introduced for FSI problems at moderate and high Reynolds numbers. In this framework, the fluid dynamics is obtained by the LBM. The FSI boundary conditions are handled by an improved IBM based on the feedback scheme where the feedback coefficient is mathematically derived and explicitly approximated. The Lagrangian force is divided into two parts: one is caused by the mismatching of the flow velocity and the boundary velocity at previous time step, and the other is caused by the boundary acceleration. Such treatment significantly enhances the numerical stability. A geometry-adaptive refinement is applied to provide fine resolution around the immersed geometries. The overlapping grids between two adjacent refinements consist of two layers. The movement of fluid-structure interfaces only causes adding or removing grids at the boundaries of refinements. Finally, the classic Smagorinsky large eddy simulation model is incorporated into the framework to model turbulent flows at relatively high Reynolds numbers. Several validation cases are conducted to verify the accuracy and fidelity of the present solver over a range of Reynolds numbers. [Preview Abstract] |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700