Algorithmische Aspekte der Fluid-Struktur-Wechselwirkung auf kartesischen Gittern

In vielen physikalischen Systemen und technischen Anwendungen spielen Fluid-Struktur-Wechselwirkungen eine wesentliche Rolle. Die Wechselwirkung von Fluiden und flexiblen Strukturen stellt ein gekoppeltes Problem dar, bei dem die Bewegung eines Fluides und einer Struktur uber die so genannte nasse Oberflache (Kopplungsoberflache) der Struktur bidirektional gekoppelt sind. So sind Windlasten an Gebauden und Brucken, das Aufblasen von Airbags, das Offnen von Fallschirmen, der Blutfluss in einer Ader oder die Stromung zwischen den Lamellen eines Autoreifens Beispiele fur diese Art der Kopplung. Bei der Untersuchung von Fluid-Struktur-Wechselwirkungen ist die numerische Simulation ein unerlassliches Hilfsmittel. Oft werden diese Simulationen durch so genannte partitionierte Ansatzen realisiert. Diese sind dadurch gekennzeichnet, dass getrennte und fur die einzelnen Teilprobleme konzipierte und angepasste Programme zur Berechnung der Stromungen und der Strukturbewegungen bzw. -verformungen eingesetzt werden -- im Gegensatz zu so genannten monolithischen Ansatzen, bei denen alle Teilprobleme gemeinsam diskretisiert und in einem Programm behandelt werden. Bei partitionierten Ansatzen konnen Teile der Berechnungen mit bewahrten und fur den jeweiligen Teilaspekt am besten geeigneten Softwarelosungen erfolgen. Damit ist jedoch eine zusatzliche Programmkomponente erforderlich, die den Ablauf der gekoppelten Simulation steuert und den Austausch der Daten zwischen den Simulationsprogrammen ermoglicht und die somit einen integralen Bestandteil partitionierter Ansatze darstellt. Dies zeigt deutlich, dass sich bei der Simulation von Fluid-Struktur-Wechselwirkungen mit partitionierten Ansatzen, neben den ingenieurwissenschaftlichen Herausforderungen (wie bspw. dem Losen konkreter Problemstellungen) und den mathematischen Herausforderungen (wie bspw. dem Sicherstellen von Konvergenz und Robustheit), insbesondere auch softwaretechnische und damit informatische Herausforderungen ergeben. Die vorliegende Arbeit befasst sich schwerpunktmasig mit den resultierenden Fragestellungen zur Steuerung der Kopplung, zur Verknupfung der in den Programmen unterschiedlichen geometrischen Darstellungen der nassen Oberflache und zum Austausch der kopplungsrelevanten Daten. Die physikalische Beschreibung des Fluid-Struktur-Wechselwirkungsproblems fordert die Erfullung von Gleichgewichtsbedingungen auf der Kopplungsoberflache zu jedem Zeitpunkt. Fur partitionierte Ansatze existieren je nach Anwendungsfall unterschiedliche Strategien und Methoden zum Austausch der Kopplungsdaten und zur Steuerung der Kopplung in der Zeit, um diese Gleichgewichte zwischen den getrennten Simulationen sicherzustellen. Dies erfordert eine Softwarelosung zur Kopplung der Simulationsprogramme, die neben einer einfachen und mit geringem Aufwand durchzufuhrenden Anpassung der Programme und einer flexiblen Moglichkeit zur Steuerung der Kopplung, eine Losung zum Transfer der kopplungsrelevanten Daten -- zwischen den auf unterschiedlichen Diskretisierungen und Datenstrukturen aufbauenden Reprasentationen der Kopplungsober Flache in den Simulationsprogrammen -- bietet. Dieser Transfer von Kopplungsdaten zwischen den verschiedenen Oberflachenreprasentationen kann durch die Einbettung in eine ubergeordnete Geometriereprasentation wirkungsvoll unterstutzt werden. Hierfur bieten sich insbesondere hierarchisch-strukturierte Zerlegungen des Raumes durch kartesische Volumenelemente (z. B. Oktalbaume) als laufzeit- und speichereffiziente Losung an. Diesem Effizienz-Gedanken folgend, stellt sich die Frage, ob diese strukturierten kartesischen Zerlegungen des Raumes nicht direkt als Basis fur die Diskretisierung des Stromungsgebietes bei der Simulation von Fluid-Struktur-Wechselwirkungen genutzt werden konnen. Die Untersuchung kartesischer Diskretisierungen im Kontext der Fluid-Struktur-Wechselwirkung bildet, neben den Fragestellungen der Realisierung der Kopplung den zweiten Schwerpunkt dieser Arbeit. Es werden entsprechende Methoden vorgestellt, untersucht und insbesondere durch die dreidimensionale Simulation des Transportes von Partikeln in Mikrostromungen validiert. Fluid-structure interactions occur in many physical systems and engineering applications. The interaction between fluids and flexible structures represents a coupled problem, where the movement of a fluid and a structure is coupled bidirectionally through the so called wet surface (coupling surface) of the structure. Hence, the wind load on buildings and bridges, the flow in blood vessels, or the gas stream inflating an airbag are examples for this kind of coupling. Numerical simulations represent a key technology for investigations on fluid-structure interactions. Such simulations are often realised by partitioned approaches. These approaches differ from monolithic approaches in the usage of different simulation programs, typically using different discretisations. So both the fluid and the structure part can be simulated using reliable programs which are well suited to the specific aspects of each part. However, an additional software component is necessary which steers the coupled simulation and establishes the connection between the simulation programs and, hence, makes exchange of the data possible. Therefore, this additional component is a substantial element for partitioned approaches. This clearly shows that simulations of fluid-structure interactions with partitioned approaches are extremely challenging not only with respect to engineering aspects (e. g. solving concrete applications) or mathematical aspects (e. g. asking for convergence and robustness). Additionally, there are also challenges concerning aspects of the software design, and thus in computer science. This thesis primarily examines the questions resulting from the latter regarding the steering of the coupling, joining the different geometrical representations of the wet surface in the programs and the exchange of data that are relevant for the coupling. The physical description of fluid-structure interactions requires the fulfillment of equilibrium conditions on the coupling surface at each time. In oder to meet these balances between the separate simulations in partitioned approaches, different strategies and methods for the exchange of the coupling data and for the steering of the coupling in time were developed, depending on the concrete fluid-structure interaction application. The transfer of coupling data between the different surface representations can be supported in an effective way by embedding these representations into a superordinate geometrical representation. For this, in particular hierarchical structured decompositions of the threedimensional space by Cartesian volume elements for example octrees provide a very attractive approach. Octrees offer an efficient geometric interface regarding both memory efficiency and run time efficiency. Following this idea of efficiency leads in a straight way to the question whether these structured Cartesian decompositions of the space can be used directly as a basis for the discretisation of the flow domain in a fluid-structure interaction scenario. Hence, the principal objective is to apply efficiency and usability to both, the coupling software for a fluid-structure interaction simulation and a software to simulate fluid dynamics. Appropriate methods will be presented, examined and validated in particular by the three-dimensional simulation of the transport of particles in micro flows.

[1]  Carlos A. Felippa,et al.  Partitioned analysis for coupled mechanical systems , 1988 .

[2]  R. Rannacher,et al.  Benchmark Computations of Laminar Flow Around a Cylinder , 1996 .

[3]  Ivo Babuška,et al.  The p-Version of the Finite Element Method for Parabolic Equations. Part 1 , 1981 .

[4]  Steven Fortune,et al.  Voronoi Diagrams and Delaunay Triangulations , 2004, Handbook of Discrete and Computational Geometry, 2nd Ed..

[5]  Matthias Heil,et al.  oomph-lib — An Object-Oriented Multi-Physics Finite-Element Library , 2006 .

[6]  Sebastian Meynen,et al.  Numerical simulation of fluid-structure interaction for fluid damped oscillations , 1999 .

[7]  O. C. Zienkiewicz,et al.  Coupled Problems and Their Numerical Solution , 1989 .

[8]  Roland Wüchner,et al.  Algorithmic treatment of shells and free form-membranes in FSI , 2006 .

[9]  K. Wolf,et al.  Coupling fluids and structures codes on MPI , 1996, Proceedings. Second MPI Developer's Conference.

[10]  Ian Foster,et al.  The Grid 2 - Blueprint for a New Computing Infrastructure, Second Edition , 1998, The Grid 2, 2nd Edition.

[11]  Ekkehard Ramm,et al.  Large deformation fluid structure interaction - advances in ALE methods and new fixed grid approaches , 2006 .

[12]  Christoph M. Hoffmann,et al.  Geometric and Solid Modeling , 1989 .

[13]  Rolf Rannacher,et al.  Adaptive Finite Element Approximation of Fluid-Structure Interaction Based on an Eulerian Variational Formulation , 2006 .

[14]  Carlos E. S. Cesnik,et al.  Evaluation of computational algorithms suitable for fluid-structure interactions , 2000 .

[15]  Rajeev K. Jaiman,et al.  Conservative load transfer along curved fluid-solid interface with non-matching meshes , 2006, J. Comput. Phys..

[16]  Hans-Joachim Bungartz,et al.  CARTESIAN DISCRETISATIONS FOR FLUID-STRUCTURE INTERACTION { CONSISTENT FORCES , 2006 .

[17]  Stefan Turek,et al.  A Monolithic FEM/Multigrid Solver for an ALE Formulation of Fluid-Structure Interaction with Applications in Biomechanics , 2006 .

[18]  Carlos A. Felippa,et al.  Architecture of a distributed analysis network for computational mechanics , 1981 .

[19]  Serge Piperno,et al.  Explicit/implicit fluid/structure staggered procedures with a structural predictor and fluid subcycling for 2D inviscid aeroelastic simulations , 1997 .

[20]  R. Sani,et al.  On pressure boundary conditions for the incompressible Navier‐Stokes equations , 1987 .

[21]  J. Happel,et al.  Low Reynolds number hydrodynamics , 1965 .

[22]  Hans-Joachim Bungartz,et al.  Introduction to computer graphics , 2004 .

[23]  Hester Bijl,et al.  Comparison of the conservative and a consistent approach for the coupling of non-matching meshes , 2006 .

[24]  M. F. Tomé,et al.  GENSMAC: a computational marker and cell method for free surface flows in general domains , 1994 .

[25]  Miriam Mehl,et al.  A cache‐oblivious self‐adaptive full multigrid method , 2006, Numer. Linear Algebra Appl..

[26]  Scott M. Murman,et al.  Implicit approaches for moving boundaries in a 3-D Cartesian method , 2003 .

[27]  Rainald Löhner,et al.  Conservative load projection and tracking for fluid-structure problems , 1996 .

[28]  R. Löhner,et al.  Adaptive embedded unstructured grid methods , 2004 .

[29]  P. Roache Fundamentals of computational fluid dynamics , 1998 .

[30]  J. Demmel,et al.  Sun Microsystems , 1996 .

[31]  Arthur Veldman,et al.  Symmetry-Preserving Discretization of Turbulent Channel Flow , 2002 .

[32]  J. Szmelter Incompressible flow and the finite element method , 2001 .

[33]  Derek M. Causon,et al.  The development of a Cartesian cut cell method for incompressible viscous flows , 2007 .

[34]  Ernst Rank,et al.  Thin Solids for Fluid-Structure Interaction , 2006 .

[35]  Craig A. Lindley,et al.  Practical ray tracing in C , 1992, Comput. Aided Des..

[36]  E. Ramm,et al.  Artificial added mass instabilities in sequential staggered coupling of nonlinear structures and incompressible viscous flows , 2007 .

[37]  I. Duff,et al.  Direct Methods for Sparse Matrices , 1987 .

[38]  Hans-Joachim Bungartz,et al.  Coupled Problems in Microsystem Technology , 1995 .

[39]  P. Tallec,et al.  Load and motion transfer algorithms for fluid/structure interaction problems with non-matching discrete interfaces: Momentum and energy conservation, optimal discretization and application to aeroelasticity , 1998 .

[40]  Ernst Rank,et al.  Computation of wind-induced vibrations of flexible shells and membranous structures , 2003 .

[41]  S. Turek,et al.  Proposal for Numerical Benchmarking of Fluid-Structure Interaction between an Elastic Object and Laminar Incompressible Flow , 2006 .

[42]  W. Joppich,et al.  MpCCI—a tool for the simulation of coupled applications , 2006 .

[43]  D. Dinkler,et al.  A monolithic approach to fluid–structure interaction using space–time finite elements , 2004 .

[44]  Michael Griebel,et al.  Numerical Simulation in Fluid Dynamics: A Practical Introduction , 1997 .

[45]  Ernst Rank,et al.  Computation of fluid-structure interaction on lightweight structures , 2001 .

[46]  Computation of Stokes Flow due to the Motion or Presence of a Particle in a Tube , 2005 .

[47]  Hans-Joachim Bungartz,et al.  Efficient Treatment of Complicated Geometries and Moving Interfaces for CFD Problems , 1999 .

[48]  W. F. Ames Solid modelling with DESIGNBASE : H. Chiyokura, Addison Wesley, Reading, Mass., 1988. 300 pp., US $39.75. ISBN 0-201-19245-4 , 1990 .

[49]  R. Courant,et al.  Über die partiellen Differenzengleichungen der mathematischen Physik , 1928 .

[50]  Michael T. Heath,et al.  Common‐refinement‐based data transfer between non‐matching meshes in multiphysics simulations , 2004 .

[51]  C. Fletcher Computational techniques for fluid dynamics , 1992 .

[52]  Michael Schäfer,et al.  An Implicit Partitioned Method for the Numerical Simulation of Fluid-Structure Interaction , 2006 .

[53]  Graham D. Riley,et al.  GCF : a general coupling framework , 2006 .

[54]  Z. Zhong Finite Element Procedures for Contact-Impact Problems , 1993 .

[55]  Hans-Joachim Bungartz,et al.  Fluid-Structure Interaction on Cartesian Grids: Flow Simulation and Coupling Environment , 2006 .

[56]  Michael T. Heath,et al.  A system integration framework for coupled multiphysics simulations , 2006, Engineering with Computers.

[57]  C. Farhat,et al.  Mixed explicit/implicit time integration of coupled aeroelastic problems: Three‐field formulation, geometric conservation and distributed solution , 1995 .

[58]  Leonid M. Brekhovskikh,et al.  Mechanics Of Continua And Wave Dynamics , 1993 .

[59]  Jay Walter Larson,et al.  The Model Coupling Toolkit: A New Fortran90 Toolkit for Building Multiphysics Parallel Coupled Models , 2005, Int. J. High Perform. Comput. Appl..

[60]  Jan Vierendeels,et al.  Implicit Coupling of Partitioned Fluid-Structure Interaction Solvers using Reduced-Order Models , 2005 .

[61]  Charbel Farhat,et al.  Geometric conservation laws for flow problems with moving boundaries and deformable meshes, and their impact on aeroelastic computations , 1996 .

[62]  Hans-Joachim Bungartz,et al.  A Parallel Adaptive Cartesian PDE Solver Using Space-Filling Curves , 2006, Euro-Par.

[63]  M. Iri,et al.  Practical use of Bucketing Techniques in Computational Geometry , 1985 .

[64]  C. W. Hirt,et al.  An Arbitrary Lagrangian-Eulerian Computing Method for All Flow Speeds , 1997 .

[65]  Manfred Krafczyk,et al.  Lattice-Boltzmann Method on Quadtree-Type Grids for Fluid-Structure Interaction , 2006 .

[66]  Jack Dongarra,et al.  PVM: Parallel virtual machine: a users' guide and tutorial for networked parallel computing , 1995 .

[67]  A. Chorin Numerical solution of the Navier-Stokes equations , 1968 .

[68]  W. Shyy,et al.  Regular Article: An Accurate Cartesian Grid Method for Viscous Incompressible Flows with Complex Immersed Boundaries , 1999 .

[69]  Haibo Dong,et al.  Hydrodynamic Performance of Deformable Fish Fins and Flapping Foils , 2006 .

[70]  Alan Sussman,et al.  Building complex coupled physical simulations on the grid with InterComm , 2006, Engineering with Computers.

[71]  F. Harlow,et al.  Numerical Calculation of Time‐Dependent Viscous Incompressible Flow of Fluid with Free Surface , 1965 .

[72]  Steven G. Parker,et al.  A component-based parallel infrastructure for the simulation of fluid–structure interaction , 2006, Engineering with Computers.

[73]  S. Armfield,et al.  A representation of curved boundaries for the solution of the Navier-Stokes equations on a staggered three-dimensional Cartesian grid , 2003 .

[74]  P. Tallec,et al.  Fluid structure interaction with large structural displacements , 2001 .

[75]  H. Wendland,et al.  Multivariate interpolation for fluid-structure-interaction problems using radial basis functions , 2001 .