ebook img

Dissertation Andreas Hiltebrand PDF

171 Pages·2015·11.32 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Dissertation Andreas Hiltebrand

ETH Library Entropy-stable discontinuous Galerkin finite element methods with streamline diffusion and shock-capturing for hyperbolic systems of conservation laws Doctoral Thesis Author(s): Hiltebrand, Andreas Publication date: 2014 Permanent link: https://doi.org/10.3929/ethz-a-010337672 Rights / license: In Copyright - Non-Commercial Use Permitted This page was generated automatically upon download from the ETH Zurich Research Collection. For more information, please consult the Terms of use. DISS. ETH NO. 22279 ENTROPY-STABLE DISCONTINUOUS GALERKIN FINITE ELEMENT METHODS WITH STREAMLINE DIFFUSION AND SHOCK-CAPTURING FOR HYPERBOLIC SYSTEMS OF CONSERVATION LAWS A thesis submitted to attain the degree of DOCTOR OF SCIENCES of ETH ZURICH (Dr. sc. ETH Zurich) presented by ANDREAS EDUARD HILTEBRAND MSc ETH CSE, ETH Zurich born on 17.03.1987 citizen of Bachenbu¨lach ZH, Switzerland accepted on the recommendation of Prof. Dr. Siddhartha Mishra, ETH Zurich, examiner Prof. Dr. R´emi Abgrall, University of Zurich, co-examiner Prof. Dr. Charalambos Makridakis, University of Sussex, co-examiner 2014 ISBN 978-3-906031-85-9 DOI 10.3929/ethz-a-010337672 Abstract We propose and analyse a space-time discontinuous Galerkin (DG) finite element method for hyperbolic conservation laws. Entropy stability is obtained by discretising the entropy variablesandusingentropy-stablenumericalfluxes. Assolutionsofhyperbolicconservation lawscandevelopdiscontinuities(shocks)infinitetime,weincludeastreamlinediffusionand a shock-capturing term to suppress spurious oscillations in the vicinity of shocks. We show that the approximate solutions converge to an entropy measure-valued solution for systems of conservation laws. Algorithmically, the discretisation leads to a big nonlinear system of algebraic equations in each time step. A Newton-Krylov method is applied to solve it. However, for effi- ciencyreasons, apreconditionermustbeused. WedesignandapplyblockJacobiandblock Gauss-Seidel type preconditioners and investigate their performance both analytically (in a simplified setting) and experimentally. The space-time formulation allows an easy local adaptation of the mesh in space and time. Weconsiderresidual-basedadaptivityaswellasgoal-orientedadaptivityusingduality estimates. Due to the implicitness of the scheme, there is no CFL condition that must be satisfied. Therefore, the scheme can also be used in problems with multiple time scales, such as flows near the incompressible limit. The high resolution and robustness properties of the method are demonstrated in several experiments. We consider, among others, the linear advection equation, Burgers’ equation, thewaveequation,andtheEulerequationsinoneortwospatialdimensions. Theseexamples show that also problems with complicated boundaries can be solved, thanks to the finite element formulation. iii Zusammenfassung Wir entwickeln und analysieren eine unstetige Galerkin-Finite-Element-Methode in Raum- zeit fu¨r hyperbolische Erhaltungsgleichungen. Entropie-Stabilit¨at erlangen wir durch das DiskretisierenderEntropievariablenunddasVerwendenEntropie-stabilernumerischerFlu¨s- se. Da L¨osungen hyperbolischer Differentialgleichungen in endlicher Zeit Unstetigkeiten (St¨osse) entwickeln k¨onnen, binden wir einen Stromliniendiffusions- und einen Stossaufl¨o- sungstermmitein,umst¨orendeOszillationenbeiSt¨ossenzuunterdru¨cken.Wirzeigen,dass dieapproximativenL¨osungenfu¨rSystemehyperbolischerDifferentialgleichungengegeneine masswertige Entropiel¨osung konvergieren. Aus algorithmischer Sicht fu¨hrt die Diskretisierung in jedem Zeitschritt auf ein grosses nichtlineares System algebraischer Gleichungen. Um es zu l¨osen, benutzen wir eine Newton- Krylov-Methode. Aus Effizienzgru¨nden muss jedoch ein Vorkonditionierer angewandt wer- den. Wir entwerfen und verwenden Vorkonditionierer vom Block-Jacobi- und Block-Gauss- Seidel-TypunduntersuchenihreEffizienzsowohlanalytisch(untervereinfachtenBedingun- gen) als auch experimentell. Die Raumzeit-Formulierung erlaubt eine einfache lokale Adaptition des Gitters in Raum und Zeit. Wir betrachten Adaptivit¨at basierend auf dem Residuum sowie Ziel-orientierte Adaptivit¨at unter Verwendung von Dualit¨atsabsch¨atzungen. DankderImplizit¨atdesVerfahrensmusskeineCFL-Bedingungerfu¨lltwerden.Daherkann dieMethodeauchfu¨rMulti-Zeitskalen-Problemegebrauchtwerden,wiez.B.fu¨rStr¨omungen nahe dem inkompressiblen Grenzfall. Die hohe Aufl¨osung und Robustheit des Verfahrens werden in diversen Experimenten demonstriert. Wir betrachten unter anderem die lineare Advektionsgleichung, die Burgers- gleichung,dieWellengleichungunddieEulergleichungineineroderzweir¨aumlichenDimen- sionen. Zus¨atzlich zeigen diese Beispiele auch, dass Probleme mit komplizierten R¨andern dank der Finite-Element-Formulierung gel¨ost werden k¨onnen. v Acknowledgements1 I would like to thank all the people that have made this thesis possible either directly or indirectly. First of all, I would like to thank Siddhartha Mishra for all the inspiring ideas, for the insights into the theory given to me, and for all the time spend in discussions and in assisting me in any respect. Furthermore, I would like to mention and thank the following people explicitly: Blanca AyusodeDiosforthehelpinpreconditioningandforpointingoutthatsuperapproximation canbeused,FranziskaWeberforherpatienceinseeminglyendlessdiscussionsonprojection errorsandforenablingmetounderstandcertainaspectsofthetheorymorethoroughly,and SandraMayforposingapparentlysimplequestionsthathighlightedfundamentalproperties of the schemes and allow to embed the scheme into the existing work. Moreover, I would like to thank her, Deep Ray, and Ulrik Fjordholm for critical reading of the thesis, which has improved it a lot. I would like to thank all the SAMies that have made the stay at the institute a pleasant time in my life. This includes the discussions on various topics, being scientific or non- scientific. Lastbutnotleast,Iwouldliketothankmyfamily. Withouttheirsupport,whichenabled me to concentrate my work, this thesis would just not have been possible. 1TheresearchwasfundedinpartbyERCSTG.NN306279SPARCCLE. vii Contents . Abstract iii . Zusammenfassung v . Acknowledgements vii . Introduction xiii . Notation xvii 1. Hyperbolic conservation laws 1 1.1. Basic notions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1.1. Integral formulation of conservation laws. . . . . . . . . . . . . . . . . 1 1.1.2. Differential formulation of conservation laws . . . . . . . . . . . . . . . 1 1.1.3. Hyperbolicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.1.4. Weak solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.1.5. Entropy conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1.6. Entropy measure-valued solutions. . . . . . . . . . . . . . . . . . . . . 4 1.2. Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.2.1. Scalar conservation laws . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.2.2. Linear symmetrisable systems . . . . . . . . . . . . . . . . . . . . . . . 6 1.2.3. Nonlinear systems of conservation laws. . . . . . . . . . . . . . . . . . 7 2. Numerical approximation of hyperbolic conservation laws 9 2.1. Finite volume schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.1.1. High-resolution schemes . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.1.2. Entropy-stable finite volume schemes . . . . . . . . . . . . . . . . . . . 11 2.2. Discontinuous Galerkin schemes . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3. The shock-capturing streamline diffusion DG formulation 15 3.1. The mesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3.2. Variational formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3.3. The DG quasilinear form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.3.1. Numerical fluxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.3.2. Entropy-conservative flux . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.3.3. Numerical diffusion operators . . . . . . . . . . . . . . . . . . . . . . . 18 ix

Description:
Entropy-stable discontinuous Galerkin finite element methods born on 17.03.1987 citizen of . 10. 2.1.2. Entropy-stable finite volume schemes .
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.