Scien&fic Data Model Han-Wei Shen The Ohio State University - PowerPoint PPT Presentation

Scien&fic Data Model Han-Wei Shen The Ohio State University

What is a Data Model? How do you describe the data represented by this image?

Data Model • Describe the objects represented by the data

Data Model v u • Describe the objects represented by the data – Structures of the objects

Data Model (u,v) Temperature v Pressure Cloud density … u • Describe the objects represented by the data – Structures of the objects – Proper&es of the objects

Data Model (u,v) Temperature v Pressure Cloud density … u • Describe the objects represented by the data – Structures of the objects – Proper&es of the objects – Rela&onships between the objects

Scien&fic Data Model Data Model m dependent variables x i (i=1..m) Data set – a single or • n independent variable v j (j = 1..n) mul&ple valued func&on y 1 = f 1 ( x 1 , x 2 , x 3 , ..., x n ) Temperature y 2 = f 2 ( x 1 , x 2 , x 3 , ..., x n ) Pressure … Cloud density … y m = f m ( x 1 , x 2 , x 3 , ..., x n ) Each dependent variable y i can have a tensor rank k k = 0 : scalar; k = 1: vector; k = 2; 2D matrix, etc. –

Scien&fic Data Model (u,v) Temperature v Pressure Cloud density … u Data set – a single or • Independent variables Dependent variables • • mul&ple valued (dimensions) – The func&on values of func&on – Spa&al coordinates independent variables (longitude, la&tude, height) – The number of values associated with each – Time dependent variable can be – Zone ID described by its tensor rank – … – 0: scalar Dimensionality - number of • – 1: vector independent variables – 2: n x n matrix …

Domain Discre&za&on v v Con&nuous Domain u u compute values v u

Scien&fic Data Set v v u u Scien&fic Data Set = Domain Structure A[ributes - Topology: property invariant One or mul&ple Domain Structure + under transforma&on values (scalars, vectors, - Geometry: instan&a&on of tensors) defined at A[ributes topology with specific points or cells posi&ons - Consists of Points and Cells , which define the Mesh

Domain Structure - Cell v v u u Cells are the fundamental building blocks of • scien&fic data sets Cells define how points are connected • together to form the basis for interpola&on Cells can be of different dimensionality • – 0 D: Ver&ces – 1 D: Line; Polylines; – 2 D: Triangle; Quadrilateral; Polygon – 3 D: Tetrahedron; Hexahedron; Voxel;

Cell Types 1D Ver&ces Line Polyline 2D Triangle Polygon Quad 3D Pyramid Hexahedron Cube Tetrahedron

A[ributes • Scalars (e.g. density), Vectors (e.g. momentum), , Tensors (e.g. stress tensor)

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structures) • Structured Grid • Unstructured Grid

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structures) – Structured Grid – Consis&ng of a collec&on of points and cells arranged on a regular labce

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structures) – Structured Grid – Consis&ng of a collec&on of points and cells arranged on a regular labce – Every point in the structured grid can be indexed by (i,j) in 2D, (i,j,k) in 3D, etc.

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structures) – Structured Grid – Consis&ng of a collec&on of points and cells arranged on a regular labce – Every point in the structured grid can be indexed by (i,j) in 2D, (i,j,k) in 3D, etc. – The posi&on of the points, and hence the geometry of the cells, can be either implicitly defined (Cartesian gird), or explicitly specified (rec&linear or curvilinear grid)

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structures) – Structured Grid • Cartesian mesh Cartesian Grid

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structures) – Structured Grid • Cartesian mesh • Rec&linear mesh Rec&linear Grid Cartesian Grid

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structures) – Structured Grid • Cartesian mesh • Rec&linear mesh Rec&linear Grid • Curvilinear mesh Cartesian Grid Curvilinear Grid

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structure) – Unstructured Grid • Also called irregular grid data

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structure) – Unstructured Grid • Also called irregular grid data • Unstructured grid points are irregularly distributed in space

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structure) – Unstructured Grid • Also called irregular grid data • Unstructured grid points are irregular located in space • It is ocen a result of space tessella&on with simple shapes

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structure) – Unstructured Grid • Also called irregular grid data • Unstructured grid points are irregular located in space • It is ocen a result of space tessella&on with simple shapes • Explicit connec&vity informa&on to form cells is necessary

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structure) – Unstructured Grid • Polygonal mesh Polygonal mesh

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structure) – Unstructured Grid • Polygonal mesh • Tetrahedral mesh Polygonal mesh Tetrahedral mesh

Scien&fic Dataset Types • Data sets are categorized into different types based on their underlying grid (domain structure) – Unstructured Grid • Polygonal mesh • Tetrahedral mesh Polygonal mesh Tetrahedral mesh • Hybrid Mesh Hybrid mesh