06
3012405004818
Ciaco scrl
onixsuitesupport@onixsuite.com
20230608
eng
COM.ONIXSUITE.9782875581198
03
02
i6doc
01
SKU
87757
02
2875581198
03
9782875581198
15
9782875581198
10
BC
<TitleType>01</TitleType>
<TitleText/>
Numéro 422
Thèses de l'École polytechnique de Louvain
422
<TitleType>01</TitleType>
<TitleText textcase="01">Development of the discontinuous Galerkin method for high-resolution, large scale CFD and acoustics in industrial geometries</TitleText>
01
GCOI
28001100583200
1
A01
Koen Hillewaert
Hillewaert, Koen
Koen
Hillewaert
<p>Koen Hillewaert graduated as an electromechanical engineer from the university of<br />Ghent in 1995. Ever since he has worked on the simulation of turbomachinery flows<br />as well as on the development of CFD methods and grid generation strategies. After<br />passages at the von Karman institute and Numeca, he joined Cenaero in 2003 where<br />he currently heads the Argo team. In close collaboration with UCL, he has been working<br />on the industrialisation of DGM since 2006.</p>
1
01
eng
172
00
172
03
TEC000000
29
2012
3069
TECHNIQUES ET SCIENCES APPLIQUEES
01
06
01
<p> New unstructured high order methods hold the promise of transforming industrial?CFD practices. Their combination of high precision, high geometrical flexibility?and parallel efficiency, make them much better suited for unsteady scale resolving?simulations (DNS, LES, …) of turbulent industrial flows than current industrial codes.?As industrial practices for this type of simulations have not yet been established,?there is an opportunity to increase simulation reliability and resolution capacity by?introducing these novel high-potential methods.?The main objective of this work is the practical development of the discontinuous?Galerkin method, arguably the most mature high-order discretisation, for the scaleresolving?simulations of turbomachinery flows.?First optimal interior penalty parameters, adapted to hybrid and high aspect ratio?meshes, typical for CFD, have been developed. This includes the development of a?sharp trace inverse inequality for all types of elements, as well as the generalisation?of the coercivity analysis.?Based on algebraic primitives performance, a flexible yet extremely efficient blockedmatrix?structure is defined, as well as reorganised assembly routines that compensate?the increase of computational complexity of the high-order quadrature. Different?iterative strategies have implemented and compared, and a generic framework is?presented that provides high-quality transfer operators for h- and p-multigrid.?Some extensions of the DGM towards low Reynolds incompressible flows and?frequential acoustics are furthermore developed and analysed.?The capacities of the developed platform are demonstrated by the direct numerical?simulation of the transitional flow on a low-pressure turbine.</p>
03
<p> New unstructured high order methods hold the promise of transforming industrial?CFD practices. Their combination of high precision, high geometrical flexibility?and parallel efficiency, make them much better suited for unsteady scale resolving?simulations (DNS, LES, …) of turbulent industrial flows than current industrial codes.?As industrial practices for this type of simulations have not yet been established,?there is an opportunity to increase simulation reliability and resolution capacity by?introducing these novel high-potential methods.?The main objective of this work is the practical development of the discontinuous?Galerkin method, arguably the most mature high-order discretisation, for the scaleresolving?simulations of turbomachinery flows.?First optimal interior penalty parameters, adapted to hybrid and high aspect ratio?meshes, typical for CFD, have been developed. This includes the development of a?sharp trace inverse inequality for all types of elements, as well as the generalisation?of the coercivity analysis.?Based on algebraic primitives performance, a flexible yet extremely efficient blockedmatrix?structure is defined, as well as reorganised assembly routines that compensate?the increase of computational complexity of the high-order quadrature. Different?iterative strategies have implemented and compared, and a generic framework is?presented that provides high-quality transfer operators for h- and p-multigrid.?Some extensions of the DGM towards low Reynolds incompressible flows and?frequential acoustics are furthermore developed and analysed.?The capacities of the developed platform are demonstrated by the direct numerical?simulation of the transitional flow on a low-pressure turbine.</p>
02
The main objective of this work is the practical development of the discontinuous Galerkin method, arguably the most mature high-order discretisation, for the scale resolving simulations of turbomachinery...
04
<p>
1 Introduction 1<br />
1.1 Motivation . 1<br />
1.2 Overview 8<br />
2 The discontinuous Galerkin method 11<br />
2.1 Model equations11<br />
2.2 Elements and functional spaces . 12<br />
2.3 Variational formulation 14<br />
2.4 Shape functions23<br />
3 Extending the DG variational formulation 27<br />
3.1 Stability of the interior penalty method on hybrid meshes27<br />
3.2 Non-conformal formulation . 35<br />
3.3 Frequential formulation of the homentropic LEE . 44<br />
4 Iterative methods 57<br />
4.1 Newton methods . 59<br />
4.2 Multigrid methods 63<br />
4.3 Concluding remarks . 71<br />
5 Efficient data structures 75<br />
5.1 Algebraic primitives on the computer . 76<br />
5.2 Data Structures84<br />
5.3 Efficient assembly . 89<br />
5.4 Conclusions 102<br />
6 noFUDGe: a first industrial application 105<br />
6.1 Description of the flow 106<br />
6.2 Computational setup . 106<br />
6.3 Comparison of computed flow fields108<br />
6.4 Validation . 114<br />
6.5 Comparison of computational cost. 116<br />
6.6 Scaling tests 116<br />
6.7 Conclusions 118<br />
i<br />
ii CONTENTS<br />
7 Current status and prospects 119<br />
7.1 Conclusions 119<br />
7.2 Current status of the Argo group 120<br />
7.3 Prospects 121<br />
A Elements of functional analysis A.3<br />
A.1 Hilbert spaces. A.3<br />
A.2 Solvability of variational problems. A.4<br />
A.3 The Lax-Milgram theoremA.5<br />
A.4 The most simple exampleA.5<br />
B Function spaces, reference elements and quadrature A.7<br />
B.1 Construction of Lagrange interpolants . A.7<br />
B.2 Interpolation on the boundary A.8<br />
B.3 Specific elements . A.9<br />
B.4 Quadrature rules . A.12<br />
C Sharp values for the trace inverse inequality A.15<br />
C.1 Simplices A.16<br />
C.2 Outline ofWarburton's method . A.16<br />
C.3 Tensor product elements. A.17<br />
C.4 Wedges . A.19<br />
C.5 Lagrange interpolation on pyramidsA.21<br />
C.6 The Pascal space on the pyramid A.25<br />
D Nonlinear instability of quadrature-free methods A.27<br />
D.1 Original formulation . A.27<br />
D.2 Extension to non-linear equations of state . A.30<br />
D.3 Spurious modesA.30<br />
Bibliography i</p>
04
01
http://www.i6doc.com/resources/titles/28001100583200/images/6459257ddab7b85bf4b57845e875e4d4/THUMBNAIL/9782875581198.jpg
17
03
01
https://i6doc.com/resources/publishers/35.jpg
18
09
01
https://i6doc.com/resources/publishers/73.png
38
https://i6doc.com/en/book/?GCOI=28001100583200
06
3052405007518
Presses universitaires de Louvain
01
06
3052405007518
Presses universitaires de Louvain
01
Presses universitaires de Louvain
http://pul.uclouvain.be/
Louvain-la-Neuve
BE
04
20130210
646
2013
02
WORLD
01
9.45
in
02
6.30
in
03
0.96
in
08
10.26
oz
01
24
cm
02
16
cm
03
0.96
cm
08
291
gr
06
3012405004818
CIACO - DUC
33
www.i6doc.com
http://www.i6doc.com
03
WORLD 01 2
20
1
02
00
02
02
STD
02
30.00
EUR
R
6.00
28.30
1.70