Basava, Seshadri (2023)
Gradient robust methods for nearly incompressible materials.
Technische Universität Darmstadt
doi: 10.26083/tuprints-00026463
Dissertation, Erstveröffentlichung, Verlagsversion
Kurzbeschreibung (Abstract)
Mixed finite elements for incompressible Navier-Stokes equations have seen a great success in mathematical fluid dynamics [16, 50, 35, 24, 21], to name a few. However, the dependency on pressure causes numerical instability. Linke [39], proposed a cure for this by introducing the gradient-robust interpolation operator π div.
We construct necessary assumptions and conditions needed to choose the suitable finite dimensional subspace of H_div (Ω; Rd ), given a stable inf-sup finite element pair solving the linear elasticity problem. We use Raviart-Thomas (RT_1) and Brezzi-Douglas-Marini (BDM_2) elements for Q_2 × DGQ_1 and Q_2 × DGP_1 finite element pairs respectively.
For computation, we use C++ based open source finite element libraries deal.II [5] and DOpElib [25]. We develop the FEValuesInterpolated class, which is derived from FEValues class of deal.II. Our class gives the value of π_div v_h , while the latter gives v_h. In case of linear elasticity, under the influence of external thermal force, we show that the gradient-robust method gives a well represented solution with fewer elements, compared to the non-gradient robust techniques for both incompressible and nearly incompressible materials. As an extension of our work [9], we show that, for phase-field fracture models under the effect of external thermal force, a well represented solution of displacement and fracture propagation for gradient-robust methods can be obtained with fewer elements, compared to non-gradient robust techniques.
Typ des Eintrags: | Dissertation | ||||
---|---|---|---|---|---|
Erschienen: | 2023 | ||||
Autor(en): | Basava, Seshadri | ||||
Art des Eintrags: | Erstveröffentlichung | ||||
Titel: | Gradient robust methods for nearly incompressible materials | ||||
Sprache: | Englisch | ||||
Referenten: | Wollner, Prof. Winnifried ; Tscherpel, Prof. Tabea ; Giesselman, Prof. Jan ; Wick, Prof. Thomas | ||||
Publikationsjahr: | 19 Dezember 2023 | ||||
Ort: | Darmstadt | ||||
Kollation: | xii, 82 Seiten | ||||
Datum der mündlichen Prüfung: | 5 Oktober 2023 | ||||
DOI: | 10.26083/tuprints-00026463 | ||||
URL / URN: | https://tuprints.ulb.tu-darmstadt.de/26463 | ||||
Kurzbeschreibung (Abstract): | Mixed finite elements for incompressible Navier-Stokes equations have seen a great success in mathematical fluid dynamics [16, 50, 35, 24, 21], to name a few. However, the dependency on pressure causes numerical instability. Linke [39], proposed a cure for this by introducing the gradient-robust interpolation operator π div. We construct necessary assumptions and conditions needed to choose the suitable finite dimensional subspace of H_div (Ω; Rd ), given a stable inf-sup finite element pair solving the linear elasticity problem. We use Raviart-Thomas (RT_1) and Brezzi-Douglas-Marini (BDM_2) elements for Q_2 × DGQ_1 and Q_2 × DGP_1 finite element pairs respectively. For computation, we use C++ based open source finite element libraries deal.II [5] and DOpElib [25]. We develop the FEValuesInterpolated class, which is derived from FEValues class of deal.II. Our class gives the value of π_div v_h , while the latter gives v_h. In case of linear elasticity, under the influence of external thermal force, we show that the gradient-robust method gives a well represented solution with fewer elements, compared to the non-gradient robust techniques for both incompressible and nearly incompressible materials. As an extension of our work [9], we show that, for phase-field fracture models under the effect of external thermal force, a well represented solution of displacement and fracture propagation for gradient-robust methods can be obtained with fewer elements, compared to non-gradient robust techniques. |
||||
Alternatives oder übersetztes Abstract: |
|
||||
Status: | Verlagsversion | ||||
URN: | urn:nbn:de:tuda-tuprints-264632 | ||||
Sachgruppe der Dewey Dezimalklassifikatin (DDC): | 500 Naturwissenschaften und Mathematik > 510 Mathematik | ||||
Fachbereich(e)/-gebiet(e): | 04 Fachbereich Mathematik 04 Fachbereich Mathematik > Optimierung 04 Fachbereich Mathematik > Optimierung > Nonlinear Optimization |
||||
Hinterlegungsdatum: | 19 Dez 2023 13:27 | ||||
Letzte Änderung: | 20 Dez 2023 11:34 | ||||
PPN: | |||||
Referenten: | Wollner, Prof. Winnifried ; Tscherpel, Prof. Tabea ; Giesselman, Prof. Jan ; Wick, Prof. Thomas | ||||
Datum der mündlichen Prüfung / Verteidigung / mdl. Prüfung: | 5 Oktober 2023 | ||||
Export: | |||||
Suche nach Titel in: | TUfind oder in Google |
Frage zum Eintrag |
Optionen (nur für Redakteure)
Redaktionelle Details anzeigen |