Partial Differential Equations - Modelling and ... - ResearchGate
Partial Differential Equations - Modelling and ... - ResearchGate Partial Differential Equations - Modelling and ... - ResearchGate
32 Yu.A. Kuznetsov 3 Mixed Finite Element Method Let ∂ 0 E k,s be the part of the boundary ∂E k,s of a ⋂ polyhedral subcell E k,s belonging to the interior of E k , i.e. ∂ 0 E k,s = ∂E k,s Ek , s = 1,n k , k = 1,n. On ∂ 0 E k,s we define a set of t k,s non-overlapping flat polygons γ k,s,j which satisfies the following two conditions: 1. ∂ 0 E k,s = ⋃ t k,s j=1 ¯γ k,s,j, 2. each γ k,s,j belongs to ∂ 0 E k,s ′ for some s ′ ≠ s, s ′ ≤ n k , where t k,s is a positive integer, s = 1,n k , k = 1,n. Examples of the partitionings of ∂ 0 E k,s into polygons γ k,s,j are given in Figures 3 and 4. In Figure 3, the interface ∂ 0 E k,1 = ∂ 0 E k,2 between E k,1 and E k,2 consists of γ k,1 and γ k,2 . In Figure 4, ∂ 0 E k,1 consists of γ k,1,1 = γ 1 , γ k,1,2 = γ 2 ,andγ k,1,3 = γ 3 ,and∂ 0 E k,2 consists of γ k,2,1 = γ 3 , γ k,2,2 = γ 4 ,and γ k,2,3 = γ 5 . Finally, ∂ 0 E k,3 consists of γ k,3,1 = γ 1 , γ k,3,2 = γ 2 , γ k,3,3 = γ 4 ,and γ k,3,4 = γ 5 . Let T h,k,s = {e k,s,i } be conforming tetrahedral partitionings of E k,s , s = 1,n k , k = 1,n. The conformity of a tetrahedral partitioning (tetrahedral mesh) means that any two different intersecting closed tetrahedrons in T h,k,s have either a common vertex, or a common edge, or a common face. The boundaries ∂E k,s of E k,s are unions of polygons in {Γ k,i } and in {γ k,s,j }, s = 1,n k , k = 1,n. We assume that each of the tetrahedral meshes T h,k,s is also conforming with respect to the boundaries of polygons in {Γ k,i } and in {γ k,s,j } belonging to ∂E k,s , i.e. these boundaries belong to the union of E k, 1 γ 1 γ 2 E k, 3 γ 3 γ 4 E k, 2 γ 5 Fig. 4. An example of partitionings ∂ 0E k,s into segments γ k,j,s , j = 1,t k , s = 1, 3
Mixed FE Methods on Polyhedral Meshes 33 edges of tetrahedrons in T h,k,s , s = 1,n k , k = 1,n. We do not assume that the tetrahedral meshes T h,k,s and T h,k′ ,s′ are conforming on the interfaces between neighboring cells E k and E k ′ when k ′ ≠ k as well as on the interfaces between neighboring subcells E k,s and E k,s ′ when k ′ = k. Let T h be a tetrahedral partitioning of Ω such that its restrictions onto E k,s coincide with the tetrahedral meshes T h,k,s , and let RT 0 (E k,s )bethe lowest order Raviart–Thomas finite element spaces on T h,k,s , s = 1,n k , k = 1,n. We define the finite element spaces V h,k,s consisting of vector functions w ∈ RT 0 (E k,s ) which have constant normal fluxes w · n k,s on each of the flat polygons Γ k,i and γ k,j belonging to ∂E k,s ,wheren k,s are the outward unit normals to ∂E k,s , s = 1,n k , k = 1,n. Then, we define the spaces V h,k on E k assuming that the restrictions w k,s of any vector function w k ∈ V h,k onto E k,s belong to the spaces V h,k,s , s = 1,n k , and the normal components of w k are continuous through γ k,s,j , j = 1,t k . To satisfy the latter condition we assume that on each polygon γ k,s,j belonging to ∂E k,s ∩ ∂E k,s ′, s ′ ≠ s, the outward normal components of vector functions w k,s and w k,s ′ satisfy the equalities w k,s · n k,s + w k,s ′ · n k,s ′ = 0 (we recall that n k,s + n k,s ′ =0), j = 1,t k,s , k = 1,n. Finally, we define the finite element space V h assuming that the restrictions w k of any vector function w ∈ V h onto E k belong to the spaces V h,k and the normal components of w are continuous on the interfaces ∂E k ∩ ∂E l between E k and E l . To satisfy the latter condition we assume that on each polygon Γ k,i belonging to ∂E k ∩∂E l the outward normal components of vector functions w k and w l satisfy the condition w k · n k + w l · n l =0,1≤ i ≤ s k , l ≠ k, k, l = 1,n. We define the finite element space Q h for the solution function p by setting that functions in Q h are constant in each of the tetrahedrons in the partitionings T h,k,s , s = 1,n k , k = 1,n. With the defined FE spaces V h and Q h , the mixed finite element discretization to (4) is as follows: Find u h ∈ V h , u h · n =0on∂Ω, andp h ∈ Q h , such that ∫ ∫ ( a −1 ) u h · v dx − p h div v dx = 0, Ω Ω ∫ ∫ ∫ (14) − div u h q dx − cp h q dx = − fqdx Ω Ω for all v ∈ V h , v · n =0on∂Ω, andq ∈ Q h . Finite element problem (14) results in the system of linear algebraic equations Mū + B T ¯p + C T ¯λ =0, Bū − Σ ¯p = F, (15) Cū =0. Here, M ∈ Rˆn׈n is a symmetric positive definite matrix, Σ ∈ R N×N is either a symmetric positive definite or a symmetric positive semidefinite matrix, B ∈ R N׈n ,andC ∈ Rñ׈n ,whereˆn =dimV h , N is the total number of Ω
- Page 1 and 2: Partial Differential Equations
- Page 3 and 4: Partial Differential Equations Mode
- Page 5 and 6: Dedicated to Olivier Pironneau
- Page 7 and 8: VIII Preface computers has been at
- Page 9 and 10: Contents List of Contributors .....
- Page 11 and 12: List of Contributors Yves Achdou UF
- Page 13 and 14: List of Contributors XV Claude Le B
- Page 15 and 16: Discontinuous Galerkin Methods Vive
- Page 17 and 18: Discontinuous Galerkin Methods 5 2
- Page 19 and 20: Discontinuous Galerkin Methods 7 g
- Page 21 and 22: Discontinuous Galerkin Methods 9 (
- Page 23 and 24: Discontinuous Galerkin Methods 11 3
- Page 25 and 26: Discontinuous Galerkin Methods 13 3
- Page 27 and 28: Discontinuous Galerkin Methods 15 W
- Page 29 and 30: Let a h and b h denote the bilinear
- Page 31 and 32: Discontinuous Galerkin Methods 19 t
- Page 33 and 34: Discontinuous Galerkin Methods 21 l
- Page 35 and 36: Table 1. Primal DG for transport Di
- Page 37 and 38: Discontinuous Galerkin Methods 25 [
- Page 39 and 40: Mixed Finite Element Methods on Pol
- Page 41 and 42: Mixed FE Methods on Polyhedral Mesh
- Page 43: Mixed FE Methods on Polyhedral Mesh
- Page 47 and 48: 4 Hybridization and Condensation Mi
- Page 49 and 50: Mixed FE Methods on Polyhedral Mesh
- Page 51 and 52: is symmetric and positive definite,
- Page 53 and 54: with some coefficient α ∈ R wher
- Page 55 and 56: 44 E.J. Dean and R. Glowinski so fa
- Page 57 and 58: 46 E.J. Dean and R. Glowinski 2 A L
- Page 59 and 60: 48 E.J. Dean and R. Glowinski S:T=
- Page 61 and 62: 50 E.J. Dean and R. Glowinski minim
- Page 63 and 64: 52 E.J. Dean and R. Glowinski 6 On
- Page 65 and 66: 54 E.J. Dean and R. Glowinski Fig.
- Page 67 and 68: 56 E.J. Dean and R. Glowinski and
- Page 69 and 70: 58 E.J. Dean and R. Glowinski 7 Num
- Page 71 and 72: 60 E.J. Dean and R. Glowinski Fig.
- Page 73 and 74: 62 E.J. Dean and R. Glowinski Assum
- Page 75 and 76: Higher Order Time Stepping for Seco
- Page 77 and 78: u n+1 h Optimal Higher Order Time D
- Page 79 and 80: Optimal Higher Order Time Discretiz
- Page 81 and 82: Optimal Higher Order Time Discretiz
- Page 83 and 84: Optimal Higher Order Time Discretiz
- Page 85 and 86: Optimal Higher Order Time Discretiz
- Page 87 and 88: Optimal Higher Order Time Discretiz
- Page 89 and 90: Optimal Higher Order Time Discretiz
- Page 91 and 92: Optimal Higher Order Time Discretiz
- Page 93 and 94: Optimal Higher Order Time Discretiz
Mixed FE Methods on Polyhedral Meshes 33<br />
edges of tetrahedrons in T h,k,s , s = 1,n k , k = 1,n. We do not assume that the<br />
tetrahedral meshes T h,k,s <strong>and</strong> T h,k′ ,s′ are conforming on the interfaces between<br />
neighboring cells E k <strong>and</strong> E k ′ when k ′ ≠ k as well as on the interfaces between<br />
neighboring subcells E k,s <strong>and</strong> E k,s ′ when k ′ = k.<br />
Let T h be a tetrahedral partitioning of Ω such that its restrictions onto<br />
E k,s coincide with the tetrahedral meshes T h,k,s , <strong>and</strong> let RT 0 (E k,s )bethe<br />
lowest order Raviart–Thomas finite element spaces on T h,k,s , s = 1,n k , k =<br />
1,n. We define the finite element spaces V h,k,s consisting of vector functions<br />
w ∈ RT 0 (E k,s ) which have constant normal fluxes w · n k,s on each of the flat<br />
polygons Γ k,i <strong>and</strong> γ k,j belonging to ∂E k,s ,wheren k,s are the outward unit<br />
normals to ∂E k,s , s = 1,n k , k = 1,n. Then, we define the spaces V h,k on E k<br />
assuming that the restrictions w k,s of any vector function w k ∈ V h,k onto<br />
E k,s belong to the spaces V h,k,s , s = 1,n k , <strong>and</strong> the normal components of<br />
w k are continuous through γ k,s,j , j = 1,t k . To satisfy the latter condition<br />
we assume that on each polygon γ k,s,j belonging to ∂E k,s ∩ ∂E k,s ′, s ′ ≠ s,<br />
the outward normal components of vector functions w k,s <strong>and</strong> w k,s ′ satisfy<br />
the equalities w k,s · n k,s + w k,s ′ · n k,s ′ = 0 (we recall that n k,s + n k,s ′ =0),<br />
j = 1,t k,s , k = 1,n.<br />
Finally, we define the finite element space V h assuming that the restrictions<br />
w k of any vector function w ∈ V h onto E k belong to the spaces V h,k<br />
<strong>and</strong> the normal components of w are continuous on the interfaces ∂E k ∩ ∂E l<br />
between E k <strong>and</strong> E l . To satisfy the latter condition we assume that on each<br />
polygon Γ k,i belonging to ∂E k ∩∂E l the outward normal components of vector<br />
functions w k <strong>and</strong> w l satisfy the condition w k · n k + w l · n l =0,1≤ i ≤ s k ,<br />
l ≠ k, k, l = 1,n.<br />
We define the finite element space Q h for the solution function p by setting<br />
that functions in Q h are constant in each of the tetrahedrons in the partitionings<br />
T h,k,s , s = 1,n k , k = 1,n. With the defined FE spaces V h <strong>and</strong> Q h ,<br />
the mixed finite element discretization to (4) is as follows: Find u h ∈ V h ,<br />
u h · n =0on∂Ω, <strong>and</strong>p h ∈ Q h , such that<br />
∫<br />
∫<br />
(<br />
a −1 )<br />
u h · v dx − p h div v dx = 0,<br />
Ω<br />
Ω<br />
∫<br />
∫<br />
∫<br />
(14)<br />
− div u h q dx − cp h q dx = − fqdx<br />
Ω<br />
Ω<br />
for all v ∈ V h , v · n =0on∂Ω, <strong>and</strong>q ∈ Q h .<br />
Finite element problem (14) results in the system of linear algebraic equations<br />
Mū + B T ¯p + C T ¯λ =0,<br />
Bū − Σ ¯p = F,<br />
(15)<br />
Cū =0.<br />
Here, M ∈ Rˆn׈n is a symmetric positive definite matrix, Σ ∈ R N×N is either<br />
a symmetric positive definite or a symmetric positive semidefinite matrix,<br />
B ∈ R N׈n ,<strong>and</strong>C ∈ Rñ׈n ,whereˆn =dimV h , N is the total number of<br />
Ω