1. grid generation and refinement€¦ · 1. grid generation and refinement • numerical...

102
Grid Generation and . . . Basic Types of Grids Grid Manipulations Structured Grids – . . . Composite Structured . . . Block-Structured Grids Grid Transformation PDE Grid Generators 1 PDE Grid Generators 2 Algebraic Grid Generators Unstructured Grids – . . . Point Generation Advancing front methods Further Grid . . . Adaptive Refinement Varying Geometries 1 Varying Geometries 2 Page 1 of 17 Introduction to Scientific Computing 13. Grids Miriam Mehl 1. Grid Generation and Refinement

Upload: others

Post on 25-Apr-2020

16 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 1 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

1. Grid Generation and Refinement

Page 2: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 1 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

1. Grid Generation and Refinement

• numerical treatment of PDE requires approximate descriptionand discretization of the resp. domain

Page 3: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 1 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

1. Grid Generation and Refinement

• numerical treatment of PDE requires approximate descriptionand discretization of the resp. domain

• main tasks: generatingand refininggrids or meshes; closely re-lated with discretization of PDE:

Page 4: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 1 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

1. Grid Generation and Refinement

• numerical treatment of PDE requires approximate descriptionand discretization of the resp. domain

• main tasks: generatingand refininggrids or meshes; closely re-lated with discretization of PDE:

– point discretization, cf. finite differences

Page 5: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 1 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

1. Grid Generation and Refinement

• numerical treatment of PDE requires approximate descriptionand discretization of the resp. domain

• main tasks: generatingand refininggrids or meshes; closely re-lated with discretization of PDE:

– point discretization, cf. finite differences

– cell discretization, cf. finite elements or volumes

Page 6: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 1 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

1. Grid Generation and Refinement

• numerical treatment of PDE requires approximate descriptionand discretization of the resp. domain

• main tasks: generatingand refininggrids or meshes; closely re-lated with discretization of PDE:

– point discretization, cf. finite differences

– cell discretization, cf. finite elements or volumes

• objectives:

Page 7: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 1 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

1. Grid Generation and Refinement

• numerical treatment of PDE requires approximate descriptionand discretization of the resp. domain

• main tasks: generatingand refininggrids or meshes; closely re-lated with discretization of PDE:

– point discretization, cf. finite differences

– cell discretization, cf. finite elements or volumes

• objectives:

– accuracy: accurate (and dense) enough to catch the essen-tial physical phenomena

Page 8: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 1 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

1. Grid Generation and Refinement

• numerical treatment of PDE requires approximate descriptionand discretization of the resp. domain

• main tasks: generatingand refininggrids or meshes; closely re-lated with discretization of PDE:

– point discretization, cf. finite differences

– cell discretization, cf. finite elements or volumes

• objectives:

– accuracy: accurate (and dense) enough to catch the essen-tial physical phenomena

– boundary approximation: sufficiently detailed to representboundaries and boundary conditions

Page 9: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 1 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

1. Grid Generation and Refinement

• numerical treatment of PDE requires approximate descriptionand discretization of the resp. domain

• main tasks: generatingand refininggrids or meshes; closely re-lated with discretization of PDE:

– point discretization, cf. finite differences

– cell discretization, cf. finite elements or volumes

• objectives:

– accuracy: accurate (and dense) enough to catch the essen-tial physical phenomena

– boundary approximation: sufficiently detailed to representboundaries and boundary conditions

– computational efficiency: small overhead for handling of datastructures, no loss of performance on supercomputers

Page 10: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 1 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

1. Grid Generation and Refinement

• numerical treatment of PDE requires approximate descriptionand discretization of the resp. domain

• main tasks: generatingand refininggrids or meshes; closely re-lated with discretization of PDE:

– point discretization, cf. finite differences

– cell discretization, cf. finite elements or volumes

• objectives:

– accuracy: accurate (and dense) enough to catch the essen-tial physical phenomena

– boundary approximation: sufficiently detailed to representboundaries and boundary conditions

– computational efficiency: small overhead for handling of datastructures, no loss of performance on supercomputers

– numerical adequacy: features with a negative impact on nu-merical efficiency should be avoided (angles, distortions)

Page 11: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 2 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

2. Basic Types of Grids

Page 12: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 2 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

2. Basic Types of Grids

• structured grids:

– construction of points or elements follows some regular pro-cess

Page 13: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 2 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

2. Basic Types of Grids

• structured grids:

– construction of points or elements follows some regular pro-cess

– geometric(coordinates) and topological information (neigh-bour relations) can be derived

Page 14: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 2 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

2. Basic Types of Grids

• structured grids:

– construction of points or elements follows some regular pro-cess

– geometric(coordinates) and topological information (neigh-bour relations) can be derived

• unstructured grids:

– completely irregular generation, even random choice is pos-sible

Page 15: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 2 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

2. Basic Types of Grids

• structured grids:

– construction of points or elements follows some regular pro-cess

– geometric(coordinates) and topological information (neigh-bour relations) can be derived

• unstructured grids:

– completely irregular generation, even random choice is pos-sible

– explicit storage of basic geometric and topological informa-tion

Page 16: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 3 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

3. Grid Manipulations

Page 17: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 3 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

3. Grid Manipulations

• grid generation:

initial placement of grid points or elements

Page 18: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 3 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

3. Grid Manipulations

• grid generation:

initial placement of grid points or elements

• grid adaptation:

– need for grid points often becomes clear only during thecomputations

Page 19: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 3 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

3. Grid Manipulations

• grid generation:

initial placement of grid points or elements

• grid adaptation:

– need for grid points often becomes clear only during thecomputations

– requires possibilities of both refinement and coarsening

Page 20: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 3 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

3. Grid Manipulations

• grid generation:

initial placement of grid points or elements

• grid adaptation:

– need for grid points often becomes clear only during thecomputations

– requires possibilities of both refinement and coarsening

• grid partition :

standard parallelization techniques are based on some subdivi-sionor decompositionof the underlying domain

Page 21: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 4 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

4. Structured Grids – Prototypes

Page 22: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 4 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

4. Structured Grids – Prototypes

• rectangular meshes:

– rectangles (2D) or cuboids (3D)

Page 23: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 4 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

4. Structured Grids – Prototypes

• rectangular meshes:

– rectangles (2D) or cuboids (3D)

• triangular meshes:

– triangles (2D) or tetrahedra (3D)

• restricted with respect to complexity of domain

Page 24: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 5 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

5. Composite Structured Grids

• subdivide (complicated) domain into subdomains of simpler formand use regular meshs there

Page 25: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 5 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

5. Composite Structured Grids

• subdivide (complicated) domain into subdomains of simpler formand use regular meshs there

blockor patchedgrids:glue subregions togetheralong interfaces (with orwithout continuity)

Page 26: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 5 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

5. Composite Structured Grids

• subdivide (complicated) domain into subdomains of simpler formand use regular meshs there

blockor patchedgrids:glue subregions togetheralong interfaces (with orwithout continuity)

overlaidor chimeragrids:subdomain grids are com-pletely independent and do notfit together (overlap, interpola-tion)

Page 27: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 6 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

6. Block-Structured Grids

Page 28: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 6 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

6. Block-Structured Grids

• subdivision into logically rectangular subdomains (with logicallyrectangular local grids)

Page 29: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 6 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

6. Block-Structured Grids

• subdivision into logically rectangular subdomains (with logicallyrectangular local grids)

• subdomains fit together in an unstructured way, but continuity isensured (coinciding grid points)

Page 30: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 7 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

7. Grid Transformation

• idea:

transformation of the unit square to the computational domain

Page 31: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 7 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

7. Grid Transformation

• idea:

transformation of the unit square to the computational domain

Page 32: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 7 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

7. Grid Transformation

• idea:

transformation of the unit square to the computational domain

• types:

– elliptic

Page 33: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 7 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

7. Grid Transformation

• idea:

transformation of the unit square to the computational domain

• types:

– elliptic

– inverse elliptic

Page 34: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 7 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

7. Grid Transformation

• idea:

transformation of the unit square to the computational domain

• types:

– elliptic

– inverse elliptic

– hyperbolic

Page 35: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 7 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

7. Grid Transformation

• idea:

transformation of the unit square to the computational domain

• types:

– elliptic

– inverse elliptic

– hyperbolic

– algebraic

Page 36: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 8 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

8. PDE Grid Generators 1

• elliptic:

– grid coordinates ξ(x, y) and η(x, y) are obtained as solu-tions of a system of elliptic PDE

Page 37: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 8 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

8. PDE Grid Generators 1

• elliptic:

– grid coordinates ξ(x, y) and η(x, y) are obtained as solu-tions of a system of elliptic PDE

∆ξ(x, y) = 0 on ]0, 1[2,

∆η(x, y) = 0 on ]0, 1[2

Page 38: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 8 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

8. PDE Grid Generators 1

• elliptic:

– grid coordinates ξ(x, y) and η(x, y) are obtained as solu-tions of a system of elliptic PDE

∆ξ(x, y) = 0 on ]0, 1[2,

∆η(x, y) = 0 on ]0, 1[2

– boundary conditions:(ξ, η) shape of the computational domain’s boundary

Page 39: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 8 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

8. PDE Grid Generators 1

• elliptic:

– grid coordinates ξ(x, y) and η(x, y) are obtained as solu-tions of a system of elliptic PDE

∆ξ(x, y) = 0 on ]0, 1[2,

∆η(x, y) = 0 on ]0, 1[2

– boundary conditions:(ξ, η) shape of the computational domain’s boundary

– ensures very smooth grids, even if boundaries are not smooth

Page 40: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 8 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

8. PDE Grid Generators 1

• elliptic:

– grid coordinates ξ(x, y) and η(x, y) are obtained as solu-tions of a system of elliptic PDE

∆ξ(x, y) = 0 on ]0, 1[2,

∆η(x, y) = 0 on ]0, 1[2

– boundary conditions:(ξ, η) shape of the computational domain’s boundary

– ensures very smooth grids, even if boundaries are not smooth

– explicit grid control (exact position of points) is difficult

Page 41: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 8 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

8. PDE Grid Generators 1

• elliptic:

– grid coordinates ξ(x, y) and η(x, y) are obtained as solu-tions of a system of elliptic PDE

∆ξ(x, y) = 0 on ]0, 1[2,

∆η(x, y) = 0 on ]0, 1[2

– boundary conditions:(ξ, η) shape of the computational domain’s boundary

– ensures very smooth grids, even if boundaries are not smooth

– explicit grid control (exact position of points) is difficult

– in nonconvex case, lines may leave domain

Page 42: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 9 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

9. PDE Grid Generators 2

• inverse elliptic:

– Laplacians are defined on the computational (curvilinear)domain

∆x(ξ, η) = 0 on Ω,∆y(ξ, η) = 0 on Ω

Page 43: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 9 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

9. PDE Grid Generators 2

• inverse elliptic:

– Laplacians are defined on the computational (curvilinear)domain

∆x(ξ, η) = 0 on Ω,∆y(ξ, η) = 0 on Ω

– boundary conditions:(x, y) shape of the unit square’s boundary

Page 44: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 9 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

9. PDE Grid Generators 2

• inverse elliptic:

– Laplacians are defined on the computational (curvilinear)domain

∆x(ξ, η) = 0 on Ω,∆y(ξ, η) = 0 on Ω

– boundary conditions:(x, y) shape of the unit square’s boundary

– no external lines, but now a more complicated system to besolved

Page 45: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 9 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

9. PDE Grid Generators 2

• inverse elliptic:

– Laplacians are defined on the computational (curvilinear)domain

∆x(ξ, η) = 0 on Ω,∆y(ξ, η) = 0 on Ω

– boundary conditions:(x, y) shape of the unit square’s boundary

– no external lines, but now a more complicated system to besolved

• hyperbolic:

– solve hyperbolic system

Page 46: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 9 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

9. PDE Grid Generators 2

• inverse elliptic:

– Laplacians are defined on the computational (curvilinear)domain

∆x(ξ, η) = 0 on Ω,∆y(ξ, η) = 0 on Ω

– boundary conditions:(x, y) shape of the unit square’s boundary

– no external lines, but now a more complicated system to besolved

• hyperbolic:

– solve hyperbolic system

– for physically unbounded domains

Page 47: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 10 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

10. Algebraic Grid Generators

• interpolation-based, no extra PDE to be solved

Page 48: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 10 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

10. Algebraic Grid Generators

• interpolation-based, no extra PDE to be solved

• most famous representative:

transfinite interpolation or Coons patch

Page 49: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 10 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

10. Algebraic Grid Generators

• interpolation-based, no extra PDE to be solved

• most famous representative:

transfinite interpolation or Coons patch

F1(x, y) = (1 − x) · c(0, y) + x · c(1, y)

F2(x, y) = (1 − y) · c(x, 0) + y · c(x, 1)

F12(x, y) = (1 − x)(1 − y) · c(0, 0) + x(1 − y) · c(1, 0) +

(1 − x)y · c(0, 1) + xy · c(1, 1)

TF (x, y) = (F1 + F2 − F12)(x, y)

Page 50: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 10 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

10. Algebraic Grid Generators

• interpolation-based, no extra PDE to be solved

• most famous representative:

transfinite interpolation or Coons patch

F1(x, y) = (1 − x) · c(0, y) + x · c(1, y)

F2(x, y) = (1 − y) · c(x, 0) + y · c(x, 1)

F12(x, y) = (1 − x)(1 − y) · c(0, 0) + x(1 − y) · c(1, 0) +

(1 − x)y · c(0, 1) + xy · c(1, 1)

TF (x, y) = (F1 + F2 − F12)(x, y)

• interpolation of boundary curves into interior, same in 3D

Page 51: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 10 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

10. Algebraic Grid Generators

• interpolation-based, no extra PDE to be solved

• most famous representative:

transfinite interpolation or Coons patch

F1(x, y) = (1 − x) · c(0, y) + x · c(1, y)

F2(x, y) = (1 − y) · c(x, 0) + y · c(x, 1)

F12(x, y) = (1 − x)(1 − y) · c(0, 0) + x(1 − y) · c(1, 0) +

(1 − x)y · c(0, 1) + xy · c(1, 1)

TF (x, y) = (F1 + F2 − F12)(x, y)

• interpolation of boundary curves into interior, same in 3D

• cheap, easy to control

Page 52: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 10 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

10. Algebraic Grid Generators

• interpolation-based, no extra PDE to be solved

• most famous representative:

transfinite interpolation or Coons patch

F1(x, y) = (1 − x) · c(0, y) + x · c(1, y)

F2(x, y) = (1 − y) · c(x, 0) + y · c(x, 1)

F12(x, y) = (1 − x)(1 − y) · c(0, 0) + x(1 − y) · c(1, 0) +

(1 − x)y · c(0, 1) + xy · c(1, 1)

TF (x, y) = (F1 + F2 − F12)(x, y)

• interpolation of boundary curves into interior, same in 3D

• cheap, easy to control

• non-smooth grids

Page 53: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 10 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

10. Algebraic Grid Generators

• interpolation-based, no extra PDE to be solved

• most famous representative:

transfinite interpolation or Coons patch

F1(x, y) = (1 − x) · c(0, y) + x · c(1, y)

F2(x, y) = (1 − y) · c(x, 0) + y · c(x, 1)

F12(x, y) = (1 − x)(1 − y) · c(0, 0) + x(1 − y) · c(1, 0) +

(1 − x)y · c(0, 1) + xy · c(1, 1)

TF (x, y) = (F1 + F2 − F12)(x, y)

• interpolation of boundary curves into interior, same in 3D

• cheap, easy to control

• non-smooth grids

• leaving lines possible

Page 54: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 11 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

11. Unstructured Grids – Delauny Triangluation

• closely related to FEM, typically triangles/tetrahedra

Page 55: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 11 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

11. Unstructured Grids – Delauny Triangluation

• closely related to FEM, typically triangles/tetrahedra

• Delaunay triangulations:

– suppose you already have grid points – how to define ele-ments?

Page 56: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 11 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

11. Unstructured Grids – Delauny Triangluation

• closely related to FEM, typically triangles/tetrahedra

• Delaunay triangulations:

– suppose you already have grid points – how to define ele-ments?

– go back to Dirichlet and Voronoi and define Voronoi regionsaround each given grid point:

Page 57: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 11 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

11. Unstructured Grids – Delauny Triangluation

• closely related to FEM, typically triangles/tetrahedra

• Delaunay triangulations:

– suppose you already have grid points – how to define ele-ments?

– go back to Dirichlet and Voronoi and define Voronoi regionsaround each given grid point:

Vi = P : ‖P − Pi‖ < ‖P − Pj‖ ∀j 6= i

Page 58: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 11 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

11. Unstructured Grids – Delauny Triangluation

• closely related to FEM, typically triangles/tetrahedra

• Delaunay triangulations:

– suppose you already have grid points – how to define ele-ments?

– go back to Dirichlet and Voronoi and define Voronoi regionsaround each given grid point:

Vi = P : ‖P − Pi‖ < ‖P − Pj‖ ∀j 6= i– result is called a Voronoi diagram(subdivision of domain into

polygons or polyhedra, resp.)

Page 59: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 11 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

11. Unstructured Grids – Delauny Triangluation

• closely related to FEM, typically triangles/tetrahedra

• Delaunay triangulations:

– suppose you already have grid points – how to define ele-ments?

– go back to Dirichlet and Voronoi and define Voronoi regionsaround each given grid point:

Vi = P : ‖P − Pi‖ < ‖P − Pj‖ ∀j 6= i– result is called a Voronoi diagram(subdivision of domain into

polygons or polyhedra, resp.)

– draw mid-lines between all pairs of neighbouring points;leads to set of disjoint triangles or tetrahedra

Page 60: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 11 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

11. Unstructured Grids – Delauny Triangluation

• closely related to FEM, typically triangles/tetrahedra

• Delaunay triangulations:

– suppose you already have grid points – how to define ele-ments?

– go back to Dirichlet and Voronoi and define Voronoi regionsaround each given grid point:

Vi = P : ‖P − Pi‖ < ‖P − Pj‖ ∀j 6= i– result is called a Voronoi diagram(subdivision of domain into

polygons or polyhedra, resp.)

– draw mid-lines between all pairs of neighbouring points;leads to set of disjoint triangles or tetrahedra

– very widespread

Page 61: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 12 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

12. Point Generation

• point generation: how to get grid points?

Page 62: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 12 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

12. Point Generation

• point generation: how to get grid points?

– superimpose a regular grid and refine (quadtree, octree)

Page 63: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 12 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

12. Point Generation

• point generation: how to get grid points?

– superimpose a regular grid and refine (quadtree, octree)

– or:

* start with some boundary point distribution,

Page 64: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 12 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

12. Point Generation

• point generation: how to get grid points?

– superimpose a regular grid and refine (quadtree, octree)

– or:

* start with some boundary point distribution,

* generate Delaunay triangulation,

Page 65: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 12 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

12. Point Generation

• point generation: how to get grid points?

– superimpose a regular grid and refine (quadtree, octree)

– or:

* start with some boundary point distribution,

* generate Delaunay triangulation,

* continue with subdivision following suitable rules

Page 66: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 12 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

12. Point Generation

• point generation: how to get grid points?

– superimpose a regular grid and refine (quadtree, octree)

– or:

* start with some boundary point distribution,

* generate Delaunay triangulation,

* continue with subdivision following suitable rules

– if helpful, add point or lines sources (singularities, bound.layers)

Page 67: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 13 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

13. Advancing front methods

• start from the boundary (starting front) and advance step by stepinto the interior:

Page 68: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 13 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

13. Advancing front methods

• start from the boundary (starting front) and advance step by stepinto the interior:

– choose an edge on the current front, say PQ

Page 69: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 13 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

13. Advancing front methods

• start from the boundary (starting front) and advance step by stepinto the interior:

– choose an edge on the current front, say PQ

– create a new point R at equal distance d from P and Q

Page 70: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 13 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

13. Advancing front methods

• start from the boundary (starting front) and advance step by stepinto the interior:

– choose an edge on the current front, say PQ

– create a new point R at equal distance d from P and Q

– determine all grid points lying within a circle around R, ra-dius r

Page 71: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 13 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

13. Advancing front methods

• start from the boundary (starting front) and advance step by stepinto the interior:

– choose an edge on the current front, say PQ

– create a new point R at equal distance d from P and Q

– determine all grid points lying within a circle around R, ra-dius r

– order these points w.r.t. distance from R

Page 72: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 13 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

13. Advancing front methods

• start from the boundary (starting front) and advance step by stepinto the interior:

– choose an edge on the current front, say PQ

– create a new point R at equal distance d from P and Q

– determine all grid points lying within a circle around R, ra-dius r

– order these points w.r.t. distance from R

– for all these points, form triangles with P and Q, select one

Page 73: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 13 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

13. Advancing front methods

• start from the boundary (starting front) and advance step by stepinto the interior:

– choose an edge on the current front, say PQ

– create a new point R at equal distance d from P and Q

– determine all grid points lying within a circle around R, ra-dius r

– order these points w.r.t. distance from R

– for all these points, form triangles with P and Q, select one

– if accepted (no intersections etc.), add to list of points

Page 74: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 13 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

13. Advancing front methods

• start from the boundary (starting front) and advance step by stepinto the interior:

– choose an edge on the current front, say PQ

– create a new point R at equal distance d from P and Q

– determine all grid points lying within a circle around R, ra-dius r

– order these points w.r.t. distance from R

– for all these points, form triangles with P and Q, select one

– if accepted (no intersections etc.), add to list of points

– change triangulation: add new edges

Page 75: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 13 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

13. Advancing front methods

• start from the boundary (starting front) and advance step by stepinto the interior:

– choose an edge on the current front, say PQ

– create a new point R at equal distance d from P and Q

– determine all grid points lying within a circle around R, ra-dius r

– order these points w.r.t. distance from R

– for all these points, form triangles with P and Q, select one

– if accepted (no intersections etc.), add to list of points

– change triangulation: add new edges

– change front line: add new edges, remove old edges

Page 76: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 14 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

14. Further Grid Generation Methods

• spacetreewith boundary fitted closure:

– create a quadtree/octree of given accuracy, add triangles ortetrahedra, resp., at boundary

Page 77: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 14 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

14. Further Grid Generation Methods

• spacetreewith boundary fitted closure:

– create a quadtree/octree of given accuracy, add triangles ortetrahedra, resp., at boundary

• hybrid grids (structured-unstructured grids):

– many mixed forms to achieve optimum compromise betweenregularity and flexibility

Page 78: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 15 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

15. Adaptive Refinement

Page 79: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 15 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

15. Adaptive Refinement

• adaptivity with structured grids:

Page 80: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 15 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

15. Adaptive Refinement

• adaptivity with structured grids:

– block-wise refinement, interpolation at hanging nodes

Page 81: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 15 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

15. Adaptive Refinement

• adaptivity with structured grids:

– block-wise refinement, interpolation at hanging nodes

– hierarchical approach (quadtrees, octrees)

Page 82: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 15 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

15. Adaptive Refinement

• adaptivity with structured grids:

– block-wise refinement, interpolation at hanging nodes

– hierarchical approach (quadtrees, octrees)

• adaptivitiy with unstructured grids: more popular

Page 83: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 15 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

15. Adaptive Refinement

• adaptivity with structured grids:

– block-wise refinement, interpolation at hanging nodes

– hierarchical approach (quadtrees, octrees)

• adaptivitiy with unstructured grids: more popular

– global error estimator:tells us whether the result computed so far is sufficientlyaccurate

Page 84: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 15 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

15. Adaptive Refinement

• adaptivity with structured grids:

– block-wise refinement, interpolation at hanging nodes

– hierarchical approach (quadtrees, octrees)

• adaptivitiy with unstructured grids: more popular

– global error estimator:tells us whether the result computed so far is sufficientlyaccurate

– refinement criterion: determines the aim of refinement:

* balancing of error over grid points,

Page 85: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 15 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

15. Adaptive Refinement

• adaptivity with structured grids:

– block-wise refinement, interpolation at hanging nodes

– hierarchical approach (quadtrees, octrees)

• adaptivitiy with unstructured grids: more popular

– global error estimator:tells us whether the result computed so far is sufficientlyaccurate

– refinement criterion: determines the aim of refinement:

* balancing of error over grid points,

* keeping error below some threshold everywhere,

* . . .

Page 86: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 15 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

15. Adaptive Refinement

• adaptivity with structured grids:

– block-wise refinement, interpolation at hanging nodes

– hierarchical approach (quadtrees, octrees)

• adaptivitiy with unstructured grids: more popular

– global error estimator:tells us whether the result computed so far is sufficientlyaccurate

– refinement criterion: determines the aim of refinement:

* balancing of error over grid points,

* keeping error below some threshold everywhere,

* . . .

– local error estimator or indicator:tells us during computation where to refine the grid locally

Page 87: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 15 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

15. Adaptive Refinement

• adaptivity with structured grids:

– block-wise refinement, interpolation at hanging nodes

– hierarchical approach (quadtrees, octrees)

• adaptivitiy with unstructured grids: more popular

– global error estimator:tells us whether the result computed so far is sufficientlyaccurate

– refinement criterion: determines the aim of refinement:

* balancing of error over grid points,

* keeping error below some threshold everywhere,

* . . .

– local error estimator or indicator:tells us during computation where to refine the grid locally

– refinement procedure:determines the technical process of refinement (centroid,red, green)

Page 88: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 16 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

16. Varying Geometries 1

Page 89: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 16 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

16. Varying Geometries 1

• applications:

Page 90: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 16 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

16. Varying Geometries 1

• applications:

– free surface problems

* injection moulding,

* melting,

* freezing

Page 91: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 16 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

16. Varying Geometries 1

• applications:

– free surface problems

* injection moulding,

* melting,

* freezing

– multiphase flows

Page 92: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 16 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

16. Varying Geometries 1

• applications:

– free surface problems

* injection moulding,

* melting,

* freezing

– multiphase flows

– fluid-structure interactions

Page 93: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 16 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

16. Varying Geometries 1

• applications:

– free surface problems

* injection moulding,

* melting,

* freezing

– multiphase flows

– fluid-structure interactions

• common strategies:

Page 94: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 16 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

16. Varying Geometries 1

• applications:

– free surface problems

* injection moulding,

* melting,

* freezing

– multiphase flows

– fluid-structure interactions

• common strategies:

– front tracking methods:

* describe boundary or interface explicitly,

Page 95: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 16 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

16. Varying Geometries 1

• applications:

– free surface problems

* injection moulding,

* melting,

* freezing

– multiphase flows

– fluid-structure interactions

• common strategies:

– front tracking methods:

* describe boundary or interface explicitly,

* update of geometry due to movements,

Page 96: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 16 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

16. Varying Geometries 1

• applications:

– free surface problems

* injection moulding,

* melting,

* freezing

– multiphase flows

– fluid-structure interactions

• common strategies:

– front tracking methods:

* describe boundary or interface explicitly,

* update of geometry due to movements,

* accurate, but expensive,

Page 97: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 16 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

16. Varying Geometries 1

• applications:

– free surface problems

* injection moulding,

* melting,

* freezing

– multiphase flows

– fluid-structure interactions

• common strategies:

– front tracking methods:

* describe boundary or interface explicitly,

* update of geometry due to movements,

* accurate, but expensive,

* topology changes?

Page 98: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 17 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

17. Varying Geometries 2

• common strategies (continued):

Page 99: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 17 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

17. Varying Geometries 2

• common strategies (continued):

– front capturing methods:

* follow interface indirectly (some global quantity)

Page 100: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 17 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

17. Varying Geometries 2

• common strategies (continued):

– front capturing methods:

* follow interface indirectly (some global quantity)

* less precise, but more straightforward

Page 101: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 17 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

17. Varying Geometries 2

• common strategies (continued):

– front capturing methods:

* follow interface indirectly (some global quantity)

* less precise, but more straightforward

* examples: Volume-of-Fluid, Marker-and-Cell

Page 102: 1. Grid Generation and Refinement€¦ · 1. Grid Generation and Refinement • numerical treatment of PDE requires approximate description and discretization of the resp. domain

Grid Generation and . . .

Basic Types of Grids

Grid Manipulations

Structured Grids – . . .

Composite Structured . . .

Block-Structured Grids

Grid Transformation

PDE Grid Generators 1

PDE Grid Generators 2

Algebraic Grid Generators

Unstructured Grids – . . .

Point Generation

Advancing front methods

Further Grid . . .

Adaptive Refinement

Varying Geometries 1

Varying Geometries 2

Page 17 of 17

Introduction to Scientific Computing

13. GridsMiriam Mehl

17. Varying Geometries 2

• common strategies (continued):

– front capturing methods:

* follow interface indirectly (some global quantity)

* less precise, but more straightforward

* examples: Volume-of-Fluid, Marker-and-Cell

– sliding mesh techniques:

* part of the grid moves, the other part is fixed (cf. ALEapproach)