A New Generalization of the P1 Non-Conforming FEM to Higher Polynomial Degrees

Mira Schedensack

doi:10.1515/cmam-2016-0031

Publicly Available Published by De Gruyter October 18, 2016

A New Generalization of the P₁ Non-Conforming FEM to Higher Polynomial Degrees

Mira Schedensack

From the journal Computational Methods in Applied Mathematics

https://doi.org/10.1515/cmam-2016-0031

Abstract

This paper generalizes the non-conforming FEM of Crouzeix and Raviart and its fundamental projection property by a novel mixed formulation for the Poisson problem based on the Helmholtz decomposition. The new formulation allows for ansatz spaces of arbitrary polynomial degree and its discretization coincides with the mentioned non-conforming FEM for the lowest polynomial degree. The discretization directly approximates the gradient of the solution instead of the solution itself. Besides the a priori and medius analysis, this paper proves optimal convergence rates for an adaptive algorithm for the new discretization. These are also demonstrated in numerical experiments. Furthermore, this paper focuses on extensions of this new scheme to quadrilateral meshes, mixed FEMs, and three space dimensions.

Keywords: Non-Conforming FEM; Helmholtz Decomposition; Mixed FEM; Adaptive FEM; Optimality; Crouzeix–Raviart FEM; Rannacher–Turek FEM

MSC 2010: 65N30; 65N12; 65N15

1 Introduction

Non-conforming finite element methods (FEMs) play an important role in computational mechanics. They allow the discretization of partial differential equations (PDEs) for incompressible fluid flows, for almost incompressible materials in linear elasticity, and for low polynomial degrees in the ansatz spaces for higher-order problems. The projection property of the interpolation operator of the P1 non-conforming FEM, also named after Crouzeix and Raviart [21], states that the L2 projection of ∇⁡H01⁢(Ω) onto the space of piecewise constant functions equals the space of piecewise gradients of the non-conforming interpolation of H01⁢(Ω) functions in the P1 non-conforming finite element space. This property is the basis for the proof of the discrete inf-sup condition for the Stokes equations [21] as well as for the analysis of adaptive algorithms [6].

Many possible generalizations of the P1 non-conforming FEM to higher polynomial degrees have been proposed. All those generalizations are either based on a modification of the classical concept of degrees of freedom [23, 22, 41], are restricted to odd polynomial degrees [20, 3], or employ an enrichment by additional bubble functions [29, 28]. However, none of those generalizations possesses a corresponding projection property of the interpolation operator for higher moments (see Remark 3.15 below). This paper introduces a novel formulation of the Poisson equation (in (3.3) below) based on the Helmholtz decomposition along with its discretization of arbitrary (globally fixed) polynomial degree. This new discretization approximates directly the gradient of the solution, which is often the quantity of interest, instead of the solution itself. For the lowest-order polynomial degree, the discrete Helmholtz decomposition of [4] proves equivalence of the novel discretization with the known non-conforming Crouzeix–Raviart FEM [21] and therefore they appear in a natural hierarchy. In the context of the novel (mixed) formulation, these discretizations turn out to be conforming.

Although the complexity of the new discretization itself is competitive with that of a standard FEM, the method requires the pre-computation of some function φ such that its divergence equals the right-hand side. If this is not computable analytically, this results in an additional integration (see also Remark 3.6 below). However, this paper focuses on the Poisson problem as a model problem to introduce the idea of the new approach and to give a broad impression over possible extensions as quadrilateral discretizations (including a discrete Helmholtz decomposition on quadrilateral meshes for the non-conforming Rannacher–Turek FEM [33] as a further highlight of this paper), the generalization to three dimensions, or inhomogeneous mixed boundary conditions. The advantages of the new approach in the applications of linear elasticity, the Stokes equation, and higher-order problems (including the Kirchhoff plate) are the topic of the papers [36, 37].

The presence of singularities for non-convex domains usually yields the same sub-optimal convergence rate for any polynomial degree. This motivates adaptive mesh-generation strategies, which recover the optimal convergence rates. This paper presents an adaptive algorithm and proves its optimal convergence. The proof essentially follows ideas from the context of the non-conforming Crouzeix–Raviart FEM [6, 32]. This illustrates that the novel discretization generalizes it in a natural way. Since the efficient and reliable error estimator involves a data approximation term without a multiplicative power of the mesh-size, the adaptive algorithm is based on separate marking.

A possible drawback of the new FEMs is that the gradient of the solution ∇⁡u is approximated, but not the solution u itself. This excludes obvious generalizations to partial differential equations where u appears in lower-order terms.

The remaining parts of this paper are organized as follows. Section 2 defines some notation. Section 3 introduces the novel formulation based on the Helmholtz decomposition and its discretization together with an a priori error estimate. The equivalence with the P1 non-conforming FEM for the lowest-order case is proved in Section 3.3. Section 4 summarizes some generalizations. Section 5 is devoted to a medius analysis of the FEM, which uses a posteriori techniques to derive a priori error estimates. Section 6 proves quasi-optimality of an adaptive algorithm, while Section 7 outlines the generalization to 3D. Section 8 concludes this paper with numerical experiments.

2 Notation

Throughout this paper Ω⊆ℝ2 is a simply connected, bounded, polygonal Lipschitz domain. Standard notation on Lebesgue and Sobolev spaces and their norms is employed with L2 scalar product (⋅,⋅)L2⁢(Ω). Given a Hilbert space X, let L2⁢(Ω;X) resp. Hk⁢(Ω;X) denote the space of functions with values in X whose components are in L2⁢(Ω) resp. Hk⁢(Ω), and let L02⁢(Ω) denote the subset of L2⁢(Ω) of functions with vanishing integral mean. The space of L2 functions whose weak divergence exists and is in L2 is denoted with H⁢(div,Ω). The space of infinitely differentiable functions reads C∞⁢(Ω) and the subspace of functions with compact support in Ω is denoted with Cc∞⁢(Ω). The piecewise action of differential operators is denoted with a subscript NC. The formula A≲B represents an inequality A≤C⁢B for some mesh-size independent, positive generic constant C; A≈B abbreviates A≲B≲A. By convention, all generic constants C≈1 do neither depend on the mesh-size nor on the level of a triangulation but may depend on the fixed coarse triangulation 𝒯0 and its interior angles. The Curl operator in two dimensions is defined by Curl⁡β:=(∂⁡β/∂⁡x2,-∂⁡β/∂⁡x1) for sufficiently smooth β.

A shape-regular triangulation 𝒯 of a bounded, polygonal, open Lipschitz domain Ω⊆ℝ2 is a set of closed triangles T∈𝒯 such that Ω¯=⋃𝒯 and any two distinct triangles are either disjoint or share exactly one common edge or one vertex. Let ℰ⁢(T) denote the edges of a triangle T and ℰ:=ℰ⁢(𝒯):=⋃T∈𝒯ℰ⁢(T) the set of edges in 𝒯. Any edge E∈ℰ is associated with a fixed orientation of the unit normal νE on E (and τE=(0,-1;1,0)⁢νE denotes the unit tangent on E). On the boundary, νE is the outer unit normal of Ω, while for interior edges E⊈∂⁡Ω, the orientation is fixed through the choice of the triangles T+∈𝒯 and T-∈𝒯 with E=T+∩T- and νE:=νT+|E is then the outer normal of T+ on E. In this situation, [v]E:=v|T+-v|T- denotes the jump across E. For an edge E⊆∂⁡Ω on the boundary, the jump across E reads [v]E:=v. For T∈𝒯 and X⊆ℝn, let Pk⁢(T;X) denote the set of polynomials on a triangle T, and Pk⁢(𝒯;X) the set of piecewise polynomials, i.e.,

Pk(T;X):={v:T→X∣each component of v is a polynomial of total degree≤k},

Pk(𝒯;X):={v:Ω→X∣∀T∈𝒯:v|T∈Pk(T;X)},

and let Pk⁢(𝒯):=Pk⁢(𝒯;ℝ). Given a subspace X⊆L2⁢(Ω;ℝn), let ΠX:L2⁢(Ω;ℝn)→X denote the L2 projection onto X and let Πk abbreviate ΠPk⁢(𝒯;ℝn). Given a triangle T∈𝒯, let hT:=(meas2⁢(T))1/2 denote the square root of the area of T and let h𝒯∈P0⁢(𝒯) denote the piecewise constant mesh-size with h𝒯|T:=hT for all T∈𝒯. For a set of triangles ℳ⊆𝒯, let ∥⋅∥ℳ abbreviate

∥⋅∥ℳ:=∑T∈ℳ∥⋅∥L2⁢(T)2.

Given an initial triangulation 𝒯0, an admissible triangulation is a regular triangulation which can be created from 𝒯0 by newest-vertex bisection [40]. The set of admissible triangulations is denoted by 𝕋.

3 Problem Formulation and Discretization

This section introduces the new formulation based on the Helmholtz decomposition in Section 3.1 and its discretization in Section 3.2. Section 3.3 discusses the equivalence with the P1 non-conforming Crouzeix–Raviart FEM [21].

3.1 New Mixed Formulation of the Poisson Problem

Given the simply connected, bounded, polygonal Lipschitz domain Ω⊆ℝ2 and f∈L2⁢(Ω), the Poisson model problem seeks u∈H01⁢(Ω) with

(3.1)-Δ⁢u=f⁢ in ⁢Ω and u=0⁢ on ⁢∂⁡Ω.

The novel weak formulation is based on the classical Helmholtz decomposition [35]

(3.2)L2⁢(Ω;ℝ2)=∇⁡H01⁢(Ω)⊕Curl⁡(H1⁢(Ω)∩L02⁢(Ω))

for any simply connected domain Ω⊆ℝ2, where the sum is orthogonal with respect to the L2 scalar product.

Remark 3.1

Note that for Ω⊆ℝ2, the definition of the Curl implies

H⁢(Curl,Ω):={β∈L2⁢(Ω)∣Curl⁡β∈L2⁢(Ω)}=H1⁢(Ω).

Define X:=L2⁢(Ω;ℝ2) and Y:=H1⁢(Ω)∩L02⁢(Ω) and let φ∈H⁢(div,Ω) satisfy -div⁡φ=f. The novel weak formulation of the Poisson problem (3.1) seeks (p,α)∈X×Y with

(3.3)

(3.3a)(p,q)L2⁢(Ω)+(q,Curl⁡α)L2⁢(Ω)=(φ,q)L2⁢(Ω) ⁢for all ⁢q∈X,

(3.3b)(p,Curl⁡β)L2⁢(Ω)=0 ⁢for all ⁢β∈Y.

This formulation is the point of departure for the numerical approximation of ∇⁡u in Section 3.2.

Remark 3.2

Remark 3.2 (Existence of Solutions)

Since Curl⁡Y⊆X, any β∈Y satisfies the inf-sup condition

∥Curl⁡β∥L2⁢(Ω)≤supq∈X∖{0}⁡(q,Curl⁡β)L2⁢(Ω)∥q∥L2⁢(Ω).

This and Brezzi’s splitting lemma [11] imply the unique existence of a solution (p,α)∈X×Y to (3.3). The L2 orthogonality of p and Curl⁡α implies

∥p∥L2⁢(Ω)2+∥Curl⁡α∥L2⁢(Ω)2=∥φ∥L2⁢(Ω)2.

Remark 3.3

Remark 3.3 (Equivalence of (3.1) and (3.3))

The second equation of (3.3) and the Helmholtz decomposition (3.2) imply the existence of u~∈H01⁢(Ω) with p=∇⁡u~. Since φ∈H⁢(div,Ω) satisfies -div⁡φ=f, the L2 orthogonality in (3.2) implies that any v∈H01⁢(Ω) satisfies

(p,∇⁡v)L2⁢(Ω)=(φ,∇⁡v)L2⁢(Ω)=(f,v)L2⁢(Ω)

and, hence, u~ solves (3.1).

Remark 3.4

Remark 3.4 (Mixed Boundary Conditions)

Let ∂⁡Ω=ΓD∪ΓN with ΓD closed, ΓD∩ΓN=∅, and each connectivity component of ΓD has positive length. Assume that the triangulation resolves ΓD. Let H-1/2⁢(ΓN) denote the space of generalized normal traces of H⁢(div,Ω) functions and let uD∈H1⁢(Ω) and g∈H-1/2⁢(ΓN) in the sense that there holds g=q⋅ν on ΓN in the sense of distributions for some q∈H⁢(div,Ω). Consider the mixed boundary value problem -Δ⁢u=f in Ω with u|ΓD=uD on ΓD and (∇u⋅ν)|ΓN=g on ΓN. Let HD1⁢(Ω) denote the subspace of H1⁢(Ω) of functions with vanishing trace on ΓD. For ΓD=∅, define HD1⁢(Ω):=H1⁢(Ω)∩L02⁢(Ω). Define

H⋆1⁢(Ω):={β∈Y∣β⁢ is constant on each connectivity component of ⁢ΓN}.

The Helmholtz decomposition

L2⁢(Ω;ℝ2)=∇⁡HD1⁢(Ω)⊕Curl⁡H⋆1⁢(Ω)

for mixed boundary conditions [24, Corollary 3.1] then leads to the following formulation. Let φ∈H⁢(div,Ω) with -div⁡φ=f additionally fulfil the boundary condition φν|ΓN=g and seek (p,α)∈L2⁢(Ω;ℝ2)×H⋆1⁢(Ω) with

(p,q)L2⁢(Ω)+(q,Curl⁡α)L2⁢(Ω)=(φ,q)L2⁢(Ω) ⁢for all ⁢q∈L2⁢(Ω;ℝ2),

(p,Curl⁡β)L2⁢(Ω)=(∇⁡uD,Curl⁡β)L2⁢(Ω) ⁢for all ⁢β∈H⋆1⁢(Ω).

Since p=φ-Curl⁡α∈H⁢(div,Ω), the equivalence follows as in Section 3.1 and with

(p⋅ν)|ΓN=(φ⋅ν)|ΓN-(Curlα⋅ν)|ΓN=g-(∇α⋅τ)|ΓN=g.

Remark 3.5

Remark 3.5 (Multiply Connected Domains)

If Ω⊆ℝ2 is a multiply connected polygonal bounded Lipschitz domain and ∂⁡Ω=ΓD∪ΓN, such that all parts of ΓD lie on the outer boundary of Ω (on the unbounded connectivity component of ℝ2∖Ω), then the Helmholtz decomposition of Remark 3.4 still holds and a discretization as above is then immediate. However, if the Dirichlet boundary ΓD also covers parts of interior boundary, that Helmholtz decomposition does no longer hold: There exist harmonic functions which are constant on different parts of ΓD and, hence, are neither in ∇⁡HΓD1⁢(Ω), nor in Curl⁡H⋆1⁢(Ω).

Remark 3.6

Remark 3.6 (Computation of φ)

The computation of φ appears as a practical difficulty because φ needs to be defined through an integration of f. If f has some simple structure, e.g., f is polynomial, this can be done manually, while for more complicated f, φ=(φ1,0) can be defined by a numerical integration of f, e.g., by approximating

φ1⁢(x,y):=∫Rxf~⁢(t,y)⁢𝑑t

for some R∈ℝ with Ω⊆[-R,∞)×ℝ and f~ defined by f~|Ω=f and f~|([-R,∞)×ℝ)∖Ω=0. This is possible in parallel.

3.2 Discretization

Let 𝒯 be a regular triangulation of Ω and k∈ℕ∪{0} and define

Xh⁢(𝒯):=Pk⁢(𝒯;ℝ2) and Yh⁢(𝒯):=Pk+1⁢(𝒯)∩Y.

The discretization of (3.3) seeks ph∈Xh⁢(𝒯) and αh∈Yh⁢(𝒯) with

(3.4)

(3.4a)(ph,qh)L2⁢(Ω)+(qh,Curl⁡αh)L2⁢(Ω)=(φ,qh)L2⁢(Ω) ⁢for all ⁢qh∈Xh⁢(𝒯),

(3.4b)(ph,Curl⁡βh)L2⁢(Ω)=0 ⁢for all ⁢βh∈Yh⁢(𝒯).

Remark 3.7

Since there are no continuity conditions on qh∈Xh⁢(𝒯) and since Curl⁡Yh⁢(𝒯)⊆Xh⁢(𝒯), equation (3.4a) is fulfilled in a strong form, i.e.,

ph+Curl⁡αh=Πk⁢φ.

In contrast to classical finite element methods, the approximation ph of ∇⁡u is a gradient only in a discrete orthogonal sense, namely (3.4b). For k=0, Section 3.3 below proves that this discrete orthogonal gradient property is equivalent to being a non-conforming gradient of a Crouzeix–Raviart finite element function. The main motivation of the novel formulation is the generalization of this scheme to any polynomial degree k.

Remark 3.8

Remark 3.8 (Existence of Discrete Solutions)

Since Curl⁡Yh⁢(𝒯)⊆Xh⁢(𝒯), the discrete inf-sup condition

∥Curl⁡βh∥L2⁢(Ω)≤supqh∈Xh⁢(𝒯)∖{0}⁡(qh,Curl⁡βh)L2⁢(Ω)∥qh∥L2⁢(Ω) for all ⁢βh∈Yh⁢(𝒯)

is fulfilled. This inequality together with Brezzi’s splitting lemma [11] implies the unique existence of a solution (ph,αh)∈Xh⁢(𝒯)×Yh⁢(𝒯) to (3.4). The equality in

∥ph∥L2⁢(Ω)2+∥Curl⁡αh∥L2⁢(Ω)2=∥Πk⁢φ∥L2⁢(Ω)2≤∥φ∥L2⁢(Ω)2.

follows from the L2 orthogonality of ph and Curl⁡αh.

The conformity of the method and the inf-sup conditions from Remarks 3.2 and 3.8 imply the following best-approximation result.

Theorem 3.9

Theorem 3.9 (Best-Approximation)

The solution (p,α)∈X×Y to (3.3) and the discrete solution (ph,αh)∈Xh⁢(T)×Yh⁢(T) to (3.4) satisfy

(3.5)∥p-ph∥L2⁢(Ω)+∥Curl(α-αh)∥L2⁢(Ω)≲minqh∈Xh⁢(𝒯)∥p-qh∥L2⁢(Ω)+minβh∈Yh⁢(𝒯)∥Curl(α-βh)∥L2⁢(Ω).

Remark 3.10

A direct analysis of the bilinear form ℬ:(X×Y)×(X×Y)→ℝ defined by

ℬ⁢((p,α),(q,β)):=(p,q)L2⁢(Ω)+(q,Curl⁡α)L2⁢(Ω)+(p,Curl⁡β)L2⁢(Ω)

for all p,q∈X and all α,β∈Y reveals that the inf-sup constant of ℬ equals 5 and, hence, the constant hidden in ≲ in (3.5) is 5.

Remark 3.11

The best-approximation of Theorem 3.9 contains the term

minβh∈Yh⁢(𝒯)∥Curl(α-βh)∥L2⁢(Ω)

on the right-hand side, which depends on the choice of φ. This seems to be worse than the best-approximation results for standard FEMs, which do not involve such a term. However, if φ is chosen smooth enough, then Curl⁡α=φ-∇⁡u has at least the same regularity as ∇⁡u, and therefore the convergence rate is not diminished. On the other hand, the approximation space for p does not have any continuity restriction and so the first approximation term

(3.6)minqh∈Xh⁢(𝒯)∥p-qh∥L2⁢(Ω)

is superior to the best-approximation of a standard FEM, where p=∇⁡u is approximated with gradients of finite element functions. However, [42, Theorem 3.2] and the comparison results of [15] prove equivalence of (3.6) and the best-approximation with gradients of a standard FEM up to some multiplicative constant.

The following lemma proves a projection property. This means that for any v∈H01⁢(Ω), the best-approximation of ∇⁡v in Xh⁢(𝒯) is a discrete orthogonal gradient in the sense that it is orthogonal to Curl⁡Yh⁢(𝒯) and so belongs to the set of discrete orthogonal gradientsWh⁢(𝒯) defined by

(3.7)Wh⁢(𝒯):={qh∈Xh⁢(𝒯)∣(qh,Curl⁡βh)L2⁢(Ω)=0⁢ for all ⁢βh∈Yh⁢(𝒯)}.

The projection property is the key ingredient in the optimality analysis of Section 6.

Lemma 3.12

Lemma 3.12 (Projection Property)

It holds that Wh⁢(T)=ΠXh⁢(T)⁢∇⁡H01⁢(Ω). Moreover, if T⋆ is an admissible refinement of T, then ΠXh⁢(T)⁢Wh⁢(T⋆)=Wh⁢(T).

Proof.

Let q∈∇⁡H01⁢(Ω). Since Curl⁡Yh⁢(𝒯)⊆Xh⁢(𝒯) and Yh⁢(𝒯)⊆Y, the orthogonality in the Helmholtz decomposition (3.2) implies for any βh∈Yh⁢(𝒯) that

(ΠXh⁢(𝒯)⁢q,Curl⁡βh)L2⁢(Ω)=(q,Curl⁡βh)L2⁢(Ω)=0.

This proves ΠXh⁢(𝒯)⁢∇⁡H01⁢(Ω)⊆Wh⁢(𝒯). For the converse direction, let ph∈Wh⁢(𝒯) and let u∈H01⁢(Ω) be a solution (possibly not unique) to

(ΠXh⁢(𝒯)⁢∇⁡u,ΠXh⁢(𝒯)⁢∇⁡v)L2⁢(Ω)=(ph,ΠXh⁢(𝒯)⁢∇⁡v)L2⁢(Ω) for all ⁢v∈H01⁢(Ω).

The orthogonality of ph-ΠXh⁢(𝒯)⁢∇⁡u to ∇⁡H01⁢(Ω) implies the existence of α∈Y such that ph-ΠXh⁢(𝒯)⁢∇⁡u=Curl⁡α. Therefore, Curl⁡α∈Xh⁢(𝒯) and, hence, α is a piecewise polynomial of degree ≤k+1 and therefore α∈Yh⁢(𝒯). But since ph∈Wh⁢(𝒯), it holds that

∥Curl⁡α∥L2⁢(Ω)2=(ph-ΠXh⁢(𝒯)⁢∇⁡u,Curl⁡α)L2⁢(Ω)=0

and, hence, α=0. This proves ΠXh⁢(𝒯)⁢∇⁡u=ph and, therefore, Wh⁢(𝒯)⊆ΠXh⁢(𝒯)⁢∇⁡H01⁢(Ω).

A similar proof applies in the discrete case and proves ΠXh⁢(𝒯)⁢Wh⁢(𝒯⋆)=Wh⁢(𝒯). ∎

Remark 3.13

Remark 3.13 (Computational Costs)

Problem (3.4) is equivalent to: Find (ph,αh)∈Xh⁢(𝒯)×Yh⁢(𝒯) such that

(Curl⁡βh,Curl⁡αh)L2⁢(Ω)=(φ,Curl⁡βh)L2⁢(Ω) for all ⁢βh∈Yh⁢(𝒯),

ph=ΠXh⁢(𝒯)⁢φ-Curl⁡αh.

Therefore, the system matrix is (in 2D) the same as that of a standard FEM (up to degrees of freedom on the boundary).

3.3 Equivalence with Crouzeix–Raviart FEM

The non-conforming Crouzeix–Raviart finite element space [21] reads

CR01(𝒯):={vCR∈P1(𝒯)∣vCR is continuous at midpoints of interior edges

and vanishes at midpoints of boundary edges}.

Since CR01⁢(𝒯)⊈H01⁢(Ω) (if the triangulation consists of more than one triangle), the weak gradient of a function vCR∈CR01⁢(𝒯) does not exist in general. However, the piecewise version ∇NC⁡vCR∈P0⁢(𝒯;ℝ2) defined by (∇NCvCR)|T:=∇(vCR|T) for all T∈𝒯 exists. The P1 non-conforming discretization of the Poisson problem seeks uCR∈CR01⁢(𝒯) with

(3.8)(∇NC⁡uCR,∇NC⁡vCR)L2⁢(Ω)=(f,vCR)L2⁢(Ω) for all ⁢vCR∈CR01⁢(𝒯).

The lowest-order space of Raviart–Thomas finite element functions [34] reads

RT0⁢(𝒯):={qRT∈H⁢(div,Ω)∣∀T∈𝒯⁢∃aT∈ℝ2,bT∈ℝ⁢ with ⁢qRT⁢(x)=aT+bT⁢x}.

The Raviart–Thomas functions have the property that the integration by parts formula holds for functions in H01⁢(Ω) as well as for functions in CR01⁢(𝒯).

The following proposition proves the equivalence of the P1 non-conforming discretization and the discretization (3.4) for k=0. Note that the discretization (3.8) is a non-conforming discretization, while the discretization (3.4) is a conforming one.

Proposition 3.14

Proposition 3.14 (Equivalence with CR-NCFEM)

Let f∈P0⁢(T) be piecewise constant and let φRT∈RT0⁢(T) satisfy -div⁡φRT=f. Then the discrete solution (ph,αh)∈P0⁢(T;R2)×(P1⁢(T)∩Y) to (3.4) for k=0 and the gradient of the discrete solution uCR∈CR01⁢(T) to (3.8) coincide,

(3.9)ph=∇NC⁡uCR.

Proof.

The crucial point is the discrete Helmholtz decomposition [4]

(3.10)P0⁢(𝒯;ℝ2)=∇NC⁡CR01⁢(𝒯)⊕Curl⁡(P1⁢(𝒯)∩Y).

Since ph is L2 orthogonal to Curl⁡(P1⁢(𝒯)∩Y), this implies ph=∇NC⁡u~CR for some u~CR∈CR01⁢(𝒯). Let qh=∇NC⁡vCR for some vCR∈CR01⁢(𝒯). Then qh is L2 orthogonal to Curl⁡(P1⁢(𝒯)∩Y) and a piecewise integration by parts and (3.4) imply

(∇NC⁡u~CR,∇NC⁡vCR)L2⁢(Ω)=(ph,qh)L2⁢(Ω)=(φRT,qh)L2⁢(Ω)

=(-div⁡φRT,vCR)L2⁢(Ω)=(f,vCR)L2⁢(Ω).

Hence, u~CR=uCR solves (3.8). ∎

The projection property from Lemma 3.12 generalizes the famous integral mean property

∇NC⁡INC⁢v=ΠP0⁢(𝒯;ℝ2)⁢∇⁡v for all ⁢v∈H01⁢(Ω)

of the non-conforming interpolation operator INC.

Remark 3.15

Remark 3.15 (Higher Polynomial Degrees)

For higher polynomial degrees k≥1, the discretization (3.4) is not equivalent to known non-conforming schemes [23, 20, 21, 28], in the sense that Wh⁢(𝒯)≠∇NC⁡Vh⁢(𝒯) for those non-conforming finite element spaces Vh⁢(𝒯). This follows from ∇NC⁡Vh⁢(𝒯)⊈Wh⁢(𝒯) for non-conforming FEMs with enrichment. A dimension argument shows

dim⁢(Wh⁢(𝒯))>dim⁢Vh⁢(𝒯)

for the non-conforming FEMs of [23, 20] without enrichment and therefore Wh⁢(𝒯)≠∇NC⁡Vh⁢(𝒯). Moreover, this proves that the generalization of the projection property to higher moments from Lemma 3.12 cannot hold for those finite element spaces, in contrast to the discretization (3.4).

4 Extensions

Section 4.1 generalizes the novel FEM to quadrilateral meshes and proves a new discrete Helmholtz decomposition for the Q1 rotated non-conforming Rannacher–Turek FEM [33]. Section 4.2 discusses a discretization with Raviart–Thomas functions.

4.1 Quadrilateral Finite Elements

For this subsection, consider a regular partition 𝒯 of Ω in quadrilaterals. Define, for the reference rectangle T^=[0,1]2,

Qk⁢(T^):={vh∈P2⁢k⁢(T^)∣∃f,g∈Pk⁢([0,1]):vh⁢(x,y)=f⁢(x)⁢g⁢(y)}.

Given T∈𝒯, let ψT:T^→T denote the bilinear transformation from the reference rectangle to T. For consistency, let P-1⁢([0,1]):={0} and set

VQ,k(𝒯):={βh∈Y∣∀T∈𝒯:(βh∘ψT)|T^∈Qk(T^)},

Xkrect(T^):={τh∈L2(T^;ℝ2)∣∃a∈ℝ,b,c∈Pk-2([0,1]),d,e∈Qk-1(T^) such that ∀(x^,y^)∈T^:

τh(x^,y^)=a(-x^k⁢y^k-1x^k-1⁢y^k)+(x^k⁢b⁢(y^)+d⁢(x^,y^)y^k⁢c⁢(x^)+e⁢(x^,y^))},

Xkrect(𝒯):={τh∈L2(Ω;ℝ2)∣∀T∈𝒯:∃ρT∈Xkrect(T^) such that

(τh∘ψT)|T^=(01-10)D(ψT-1)⊤∘ψT(0-110)ρT}.

Then a discretization with respect to the quadrilateral partition seeks ph∈Xkrect⁢(𝒯) and αh∈VQ,k⁢(𝒯) with

(ph,qh)L2⁢(Ω)+(qh,Curl⁡αh)L2⁢(Ω)=(φ,qh)L2⁢(Ω) ⁢ for all ⁢qh∈Xkrect⁢(𝒯),

(ph,Curl⁡βh)L2⁢(Ω)=0 ⁢ for all ⁢βh∈VQ,k⁢(𝒯).

Let βh∈VQ,k⁢(𝒯), i.e., (βh∘ψT)|T^∈Qk(T^). A direct calculation reveals, for all T∈𝒯,

((Curlβh)⊤∘ψT)|T=(∇(βh∘ψT∘ψT-1))⊤∘ψT(01-10)=(∇(βh∘ψT))⊤D(ψT-1)∘ψT(01-10).

Let (βh∘ψT)⁢(x^,y^)=(𝔞⁢x^k+f⁢(x^))⁢(𝔟⁢y^k+g⁢(y^)) with 𝔞,𝔟∈ℝ and f,g∈Pk-1⁢([0,1]). Then it holds

∇⁡(βh∘ψT)=𝔞⁢𝔟⁢k⁢(x^k-1⁢y^ky^k-1⁢x^k)+(𝔟⁢y^k⁢∂⁡f⁢(x^)/∂⁡x^𝔞⁢x^k⁢∂⁡g⁢(y^)/∂⁡y^)+(𝔞⁢k⁢x^k-1⁢g⁢(y^)+g⁢(y^)⁢∂⁡f⁢(x^)/∂⁡x^𝔟⁢k⁢y^k-1⁢f⁢(x^)+f⁢(x^)⁢∂⁡g⁢(y^)/∂⁡y^)

and therefore

(∇⁡(βh∘ψT))⊤⁢(0-110)=(𝔞⁢𝔟⁢k⁢(y^k-1⁢x^k-x^k-1⁢y^k)+(𝔞⁢x^k⁢∂⁡g⁢(y^)/∂⁡y^-𝔟⁢y^k⁢∂⁡f⁢(x^)/∂⁡x^)+(𝔟⁢k⁢y^k-1⁢f⁢(x^)+f⁢(x^)⁢∂⁡g⁢(y^)/∂⁡y^-𝔞⁢k⁢x^k-1⁢g⁢(y^)-g⁢(y^)⁢∂⁡f⁢(x^)/∂⁡x^))⊤

=:(ρT(x^,y^))⊤.

This implies ρT∈Xkrect⁢(T^) and (∇⁡(βh∘ψT))=(0,1;-1,0)⁢ρT. The combination of the previous equalities leads to

((Curlβh)∘ψT)|T=(01-10)D(ψT-1)⊤∘ψT(0-110)ρT.

Consequently, Curl⁡βh∈Xkrect⁢(𝒯). This and the conformity of the method prove as in Section 3 the following statements:

unique existence of solutions,
the best-approximation result
∥p-ph∥L2⁢(Ω)+∥Curl(α-αh)∥L2⁢(Ω)≲(minqh∈Xkrect⁢(𝒯)∥p-qh∥L2⁢(Ω)+minβh∈VQ,k⁢(𝒯)∥Curl(α-βh)∥L2⁢(Ω)),
the projection property
ΠXkrect⁢(𝒯)⁢∇⁡H01⁢(Ω)⊆Whrect⁢(𝒯)
for
(4.1)Whrect⁢(𝒯)={qh∈Xkrect⁢(𝒯)∣∀βh∈VQ,k⁢(𝒯):(qh,Curl⁡βh)L2⁢(Ω)=0}.

Remark 4.1

The properties (i)–(iii) still hold for any X~h⁢(𝒯) with Xkrect⁢(𝒯)⊆X~h⁢(𝒯)⊆X.

The remaining part of this subsection proves the equivalence of the lowest-order rectangular discretization with the non-conforming Rannacher–Turek FEM [33]. To this end, define for the reference rectangle T^ and the bilinear transformation ψT:T^→T,

Qrot⁢(T^):=span⁢{1,x,y,x2-y2},

VNCrot(𝒯):={vh∈L2(Ω)∣∀T∈𝒯:(vh∘ψT)|T^∈Qrot(T^) and ∫Evhds is continuous

(4.2) for all interior edges E and vanishes at boundary edges E}.

The following lemma proves a relation between the cardinalities of the quadrilaterals, nodes, and interior edges of a quadrilateral partition similar to Euler’s formulae

(4.3)card⁢(ℰ)+card⁢(ℰ⁢(Ω))=3⁢card⁢(𝒯),card⁢(ℰ⁢(Ω))+card⁢(𝒩)=2⁢card⁢(𝒯)+1

on triangles. This enables a dimension argument in the proof of the discrete Helmholtz decomposition in Theorem 4.3 below.

Lemma 4.2

Lemma 4.2 (Euler Formula for Quadrilateral Partitions)

Let T be a regular partition of Ω in quadrilaterals with edges E, interior edges E⁢(Ω), and vertices N. Then it holds that 3⁢c⁢a⁢r⁢d⁢(T)+1=card⁢(E⁢(Ω))+card⁢(N).

Proof.

Define a triangulation 𝒯Δ of Ω in triangles by the division of each quadrilateral into two triangles by a diagonal cut. Let ℰΔ denote the edges of 𝒯Δ, ℰΔ⁢(Ω) the interior edges and 𝒩Δ the vertices. Then the following relations between the two partitions hold:

card⁢(𝒯Δ)=2⁢c⁢a⁢r⁢d⁢(𝒯),card⁢(ℰΔ)=card⁢(ℰ)+card⁢(𝒯),

card⁢(ℰΔ⁢(Ω))=card⁢(ℰ⁢(Ω))+card⁢(𝒯),card⁢(𝒩Δ)=card⁢(𝒩).

This and Euler’s formulae for triangles (4.3) prove

card⁢(ℰ⁢(Ω))+card⁢(𝒩)=card⁢(ℰΔ⁢(Ω))-card⁢(𝒯)+card⁢(𝒩Δ)

=2⁢c⁢a⁢r⁢d⁢(𝒯Δ)+1-card⁢(𝒯)=3⁢c⁢a⁢r⁢d⁢(𝒯)+1.∎

The following theorem proves that the solution space Whrect⁢(𝒯) from (4.1) equals the piecewise gradients of functions in VNCrot⁢(𝒯) on a partition in squares for k=1.

Theorem 4.3

Theorem 4.3 (Discrete Helmholtz Decomposition on Squares)

Let T be a regular partition of Ω in squares. Then,

(4.4)X1rect⁢(𝒯)=∇NC⁡VNCrot⁢(𝒯)⊕Curl⁡VQ,1⁢(𝒯)

and the decomposition is L2 orthogonal.

Remark 4.4

The L2 orthogonality in (4.4) still holds for a partition in parallelograms. However,

∇NC⁡VNCrot⁢(𝒯)⊈X1rect⁢(𝒯)

for general quadrilateral partitions.

Proof of Theorem 4.3.

Let vh∈VNCrot⁢(𝒯) and βh∈VQ,1⁢(𝒯). A piecewise integration by parts leads to

(∇NC⁡vh,Curl⁡βh)L2⁢(Ω)=∑E∈ℰ∫E[vh]E⁢∇⁡βh⋅τE⁢d⁢s.

Since 𝒯 consists of parallelograms, the bilinear transformation ψT:T^→T is affine and, hence, βh|E is affine on each edge E∈ℰ. This implies that ∇⁡βh⋅τE is constant. Since the integral mean of [vh]E vanishes, this proves the L2 orthogonality.

Let vh∈VNCrot⁢(𝒯). A computation reveals for all T∈𝒯 that there exist fT∈ℝ and gT∈ℝ2 such that

∇⁡vh⁢(x,y)=D⁢(ψT-1)⊤⁢(fT⁢(-xy)+gT).

For k=1, X1rect⁢(𝒯) reads

X1rect(𝒯)={τh∈L2(Ω;ℝ2)∣∀T∈𝒯:∃aT∈ℝ,dT∈ℝ2 such that

(τh∘ψT)|T^=(01-10)(D(ψT-1)⊤∘ψT)(0-110)(aT(-xy)+dT)}.

Since all T∈𝒯 are squares, D⁢ψT and (0,1;-1,0) commute, and, hence, ∇⁡vh∈X1rect⁢(𝒯). Thus,

∇NC⁡VNCrot⁢(𝒯)⊕Curl⁡VQ,1⁢(𝒯)⊆X1rect⁢(𝒯).

The dimension of ∇NC⁡VNCrot⁢(𝒯) equals card⁢(ℰ⁢(Ω)) and the dimension of Curl⁡VQ,1⁢(𝒯) equals card⁢(𝒩)-1, while the dimension of X1rect⁢(𝒯) equals 3⁢c⁢a⁢r⁢d⁢(𝒯). This and Lemma 4.2 prove the assertion. ∎

Remark 4.5

Remark 4.5 (Arbitrary Quadrilaterals)

The best-approximation (ii) from above proves quasi-optimal convergence even for arbitrary quadrilaterals. Standard interpolation error estimates for VQ,1⁢(𝒯) and for P0⁢(𝒯;ℝ2)⊆X1rect⁢(𝒯) (see [19]) lead to first-order convergence rates of h for sufficiently smooth solutions. This should be contrasted with [33], where quasi-optimal convergence is only obtained for a modification of (4.2) where VNCrot⁢(𝒯) is defined in terms of local coordinates.

4.2 Relation to Mixed Raviart–Thomas FEM

This subsection shows that the classical mixed Raviart–Thomas FEM [34] can be regarded as a particular choice of the ansatz spaces in the new mixed scheme.

Let 𝒯 denote a regular triangulation of Ω in triangles. Define the space of Raviart–Thomas functions [34]

XRT(𝒯)={qRT∈H(div,Ω)∣∀T∈𝒯:qRT|T(x)∈Pk(T;ℝ2)+Pk(T)x}

and

YRT⁢(𝒯):=Pk+1⁢(𝒯)∩Y.

Then the following problem is a discretization of (3.3): Find (pRT,αRT)∈XRT⁢(𝒯)×YRT⁢(𝒯) with

(4.5)

(4.5a)(pRT,qRT)L2⁢(Ω)+(qRT,Curl⁡αRT)L2⁢(Ω)=(φ,qRT) ⁢for all ⁢qRT∈XRT⁢(𝒯),

(4.5b)(pRT,Curl⁡βRT)L2⁢(Ω)=0 ⁢for all ⁢βRT∈YRT⁢(𝒯).

Since Curl⁡YRT⁢(𝒯)⊆Pk⁢(𝒯;ℝ2) and div⁡Curl⁡vRT=0 for all vRT∈YRT⁢(𝒯), it follows Curl⁡YRT⁢(𝒯)⊆XRT⁢(𝒯). This and the conformity of the method guarantee as in Sections 3 and 4.1 the unique existence of solutions, a best-approximation result, and the projection property

ΠXRT⁢(𝒯)⁢∇⁡H01⁢(Ω)⊆WRT⁢(𝒯):={qRT∈XRT⁢(𝒯)∣∀βRT∈YRT⁢(𝒯):(qRT,Curl⁡βRT)L2⁢(Ω)=0}.

The discrete Helmholtz decomposition of [26, 5, 12] proves

XRT⁢(𝒯)=∇RT⁡Pk⁢(𝒯)⊕Curl⁡YRT⁢(𝒯)

with the operator ∇RT:Pk⁢(𝒯)→XRT⁢(𝒯) defined for all vRT∈Pk⁢(𝒯) by

(∇RT⁡vRT,qRT)L2⁢(Ω)=-(vRT,div⁡qRT)L2⁢(Ω) for all ⁢qRT∈XRT⁢(𝒯).

This decomposition yields the equivalence of (4.5) with the problem: Find (pRT,u~RT)∈XRT⁢(𝒯)×Pk⁢(𝒯) with

pRT=∇RT⁡u~RT,

(wRT,div⁡pRT)L2⁢(Ω)=(div⁡ΠXRT⁢(𝒯)⁢φ,wRT)L2⁢(Ω) for all ⁢wRT∈Pk⁢(𝒯).

This is the classical Raviart–Thomas discretization with f replaced by div⁡ΠXRT⁢(𝒯)⁢φ.

Assume now that the right-hand side φ∈XRT⁢(𝒯) is a Raviart–Thomas function. Since by definition YRT⁢(𝒯)=Yh⁢(𝒯) with Yh⁢(𝒯) from Section 3.2 and since αRT is the solution of

(Curl⁡βRT,Curl⁡αRT)L2⁢(Ω)=(φ,Curl⁡βRT)L2⁢(Ω) for all ⁢βRT∈YRT⁢(𝒯),

it holds αRT=αh with αh from (3.4). Since φ=pRT+Curl⁡αRT and ΠXh⁢(𝒯)⁢φ=ph+Curl⁡αh, it follows

ph=ΠXh⁢(𝒯)⁢pRT.

For k=0, the equivalence with the Crouzeix–Raviart FEM (3.9) then proves the identity

∇NC⁡uCR=ΠXh⁢(𝒯)⁢pRT,

which is also known as Marini identity [3, 27].

5 Medius Analysis

The medius analysis of [25, 15] proves for the discrete solution uCR∈CR01⁢(𝒯) to (3.8) the best-approximation result

(5.1)∥∇NC(u-uCR)∥L2⁢(Ω)≲minvCR∈CR01⁢(𝒯)∥∇NC(u-vCR)∥L2⁢(Ω)+osc(f,𝒯).

The following theorem proves a generalization for the discretization (3.4) for the lowest order case k=0.

Theorem 5.1

Theorem 5.1 (Best-Approximation Property)

Let (p,α)∈X×Y be the solution to (3.3) and let (ph,αh)∈P0⁢(T;R2)×(P1⁢(T)∩Y) be the solution to (3.4). Then the following best-approximation result holds:

(5.2)∥p-ph∥L2⁢(Ω)≲∥p-Π0⁢p∥L2⁢(Ω)+osc⁢(f,𝒯)+supvCR∈CR01⁢(𝒯)∖{0}⁡(f,vCR)L2⁢(Ω)-(φ,∇NC⁡vCR)L2⁢(Ω)∥∇NC⁡vCR∥L2⁢(Ω).

Remark 5.2

If φ is a lowest-order Raviart–Thomas function, then it allows for an integration by parts formula also with Crouzeix–Raviart functions (see Section 3.3). Therefore, the third term on the right-hand side of (5.2) vanishes. This and the equivalence with the non-conforming FEM of Crouzeix and Raviart from Section 3.3 reveal the best-approximation result (5.1).

The remaining part of this section is devoted to the proof of Theorem 5.1. The following lemma from [17, 14] is the key ingredient of this proof. Recall the definition of CR01⁢(𝒯) from Section 3.3.

Lemma 5.3

Lemma 5.3 (Companion)

For any vCR∈CR01⁢(T) there exists v∈H01⁢(Ω) with the following properties:

Π0⁢∇NC⁡(v-vCR)=0.
Π0⁢(v-vCR)=0.
∥h𝒯-1⁢(vCR-v)∥L2⁢(Ω)+∥∇NC⁡(vCR-v)∥L2⁢(Ω)≲∥∇NC⁡vCR∥L2⁢(Ω).

Proof of Theorem 5.1.

Define qh:=Π0⁢p-ph∈P0⁢(𝒯;ℝ2). The projection property of Lemma 3.12 implies that qh∈Wh⁢(𝒯) and the discrete Helmholtz decomposition (3.10) guarantees the existence of vCR∈CR01⁢(𝒯) with qh=∇NC⁡vCR. Let v∈H01⁢(Ω) denote the companion of vCR from Lemma 5.3. Then

(5.3)(p-ph,qh)L2⁢(Ω)=(p,∇NC⁡(vCR-v))L2⁢(Ω)+(p,∇⁡v)L2⁢(Ω)-(ph,∇NC⁡vCR)L2⁢(Ω).

For the first term on the right-hand side, properties (i) and (iii) from Lemma 5.3 yield

(p,∇NC⁡(vCR-v))L2⁢(Ω)=(p-Π0⁢p,∇NC⁡(vCR-v))L2⁢(Ω)

(5.4)≲∥p-Π0⁢p∥L2⁢(Ω)⁢∥∇NC⁡vCR∥L2⁢(Ω).

For the second and third term on the right-hand side of (5.3), problems (3.3) and (3.4) lead to

(p,∇⁡v)L2⁢(Ω)-(ph,∇NC⁡vCR)L2⁢(Ω)=(φ,∇⁡v)L2⁢(Ω)-(φ,∇NC⁡vCR)L2⁢(Ω).

Since -div⁡φ=f, it follows

(φ,∇⁡v)L2⁢(Ω)-(φ,∇NC⁡vCR)L2⁢(Ω)=(f,v-vCR)L2⁢(Ω)+(f,vCR)L2⁢(Ω)-(φ,∇NC⁡vCR)L2⁢(Ω).

Properties (ii) and (iii) of Lemma 5.3 prove

(f,v-vCR)L2⁢(Ω)≲osc⁢(f,𝒯)⁢∥∇NC⁡vCR∥L2⁢(Ω).

The combination with (5.3) and (5.4) and a Cauchy inequality yield

(p-ph,qh)L2⁢(Ω)≲(∥p-Π0⁢p∥L2⁢(Ω)+osc⁢(f,𝒯)+supvCR∈CR01⁢(𝒯)∖{0}⁡(f,vCR)L2⁢(Ω)-(φ,∇NC⁡vCR)L2⁢(Ω)∥∇NC⁡vCR∥L2⁢(Ω))⁢∥qh∥L2⁢(Ω).

This and

∥p-ph∥L2⁢(Ω)2=∥p-Π0⁢p∥L2⁢(Ω)2+∥qh∥L2⁢(Ω)2=∥p-Π0⁢p∥L2⁢(Ω)2+(p-ph,qh)L2⁢(Ω)

prove the assertion. ∎

Remark 5.4

Remark 5.4 (Higher Polynomial Degrees)

For k≥1, Remark 3.15 implies that an analogue of Lemma 5.3 cannot be proved in the same way.

6 Adaptive Algorithm

This section defines an adaptive algorithm based on separate marking and proves its quasi-optimal convergence.

6.1 Adaptive Algorithm and Optimal Convergence Rates

Let 𝒯0 denote some initial shape-regular triangulation of Ω, such that each triangle T∈𝒯 is equipped with a refinement edge ET∈ℰ⁢(T). A proper choice of these refinement edges guarantees an overhead control [7].

Let 𝕋⁢(N) denote the subset of 𝕋 of all admissible triangulations with at most card⁢(𝒯0)+N triangles. The adaptive algorithm involves the overlay of two admissible triangulations 𝒯,𝒯⋆∈𝕋, which reads

𝒯⊗𝒯⋆:={T∈𝒯∪𝒯⋆∣∃K∈𝒯,K⋆∈𝒯⋆⁢ with ⁢T⊆K∩K⋆}.

Given a triangulation 𝒯ℓ, define for all T∈𝒯ℓ the local error estimator contributions by

(6.1)λ2⁢(𝒯ℓ,T):=∥h𝒯⁢curlNC⁡ph∥L2⁢(T)2+hT⁢∑E∈ℰ⁢(T)∥[ph]E⋅τE∥L2⁢(E)2 and μ2⁢(T):=∥φ-Πk⁢φ∥L2⁢(T)2

and the global error estimators by

(6.2)λℓ2:=λ2⁢(𝒯ℓ,𝒯ℓ) and μℓ2:=μ2⁢(𝒯ℓ)

with

(6.3)λ2⁢(𝒯ℓ,ℳ):=∑T∈ℳλ2⁢(𝒯ℓ,T) and μ2⁢(ℳ):=∑T∈ℳμ2⁢(T)

for any ℳ⊆𝒯ℓ. The adaptive algorithm is driven by these two error estimators and runs the following loop.

Algorithm 1

Algorithm 1 (AFEM.)

Remark 6.1

Remark 6.1 (Separate Versus Collective Marking)

The residual-based error estimator λ2+μ2 involves the term ∥φ-Πk⁢φ∥L2⁢(T) without a multiplicative positive power of the mesh-size. Therefore, the optimality of an adaptive algorithm based on collective marking (that is κ=∞ and λ replaced by λ2+μ2 in Algorithm 1) does not follow from the abstract framework from [13]. The reduction property (axiom (A2) from [13]), is not fulfilled. Algorithm 1 considered here is based on separate marking. In this context, the optimality of the adaptive algorithm (see Theorem 6.5) can be proved with a reduction property that only considers λ.

Remark 6.2

The step Mark in the second case (μℓ2>κ⁢λℓ2) can be realized by the algorithm Approx from [7, 16], i.e., the thresholding second algorithm [8] followed by a completion algorithm. For this algorithm, the assumption (B1) optimal data approximation, which is assumed to hold in the following, follows from the axioms (B2) and (SA) from Section 6.5, cf. [16]. For a discussion about other algorithms that realize Mark in the second case, see again [16].

For s>0 and (p,α,φ)∈X×Y×H⁢(div,Ω) define

|(p,α,φ)|𝒜s:=supN∈ℕ0⁡Ns⁢inf𝒯∈𝕋⁢(N)⁡(∥p-ΠXh⁢(𝒯)⁢p∥L2⁢(Ω)+infβ𝒯∈Yh⁢(𝒯)⁡∥Curl⁡(α-β𝒯)∥L2⁢(Ω)+∥φ-ΠXh⁢(𝒯)⁢φ∥L2⁢(Ω)).

Remark 6.3

Remark 6.3 (Pure Local Approximation Class)

Since Ω is assumed to be a Lipschitz domain, all patches in an admissible triangulation 𝒯∈𝕋 are edge-connected, i.e., for all vertices z∈𝒩 and triangles T,K∈𝒯 with z∈T∩K, there exists m∈ℕ0 and K0,…,Km∈𝒯 with K0=T, Km=K, z∈K0∩…∩Km and Kj-1∩Kj∈ℰ for all 1≤j≤m. Under this assumption, [42, Theorem 3.2] shows

minvh∈Pk+1⁢(𝒯)∩H1⁢(Ω)∥∇(v-vh)∥L2⁢(Ω)≈∥∇v-Πk∇v∥L2⁢(Ω) for all v∈H1(Ω).

Hence,

|(p,α,φ)|𝒜s≈|(p,α,φ)|𝒜s′

:=supN∈ℕ0⁡Ns⁢inf𝒯∈𝕋⁢(N)⁡(∥p-ΠXh⁢(𝒯)⁢p∥L2⁢(Ω)+∥Curl⁡α-ΠXh⁢(𝒯)⁢Curl⁡α∥L2⁢(Ω)+∥φ-ΠXh⁢(𝒯)⁢φ∥L2⁢(Ω)).

In the following, we assume that the following assumption (B1) holds for the algorithm used in the step Mark for μℓ2>κ⁢λℓ2 (see Remark 6.2).

Assumption 6.4

Assumption 6.4 ((B1) Optimal Data Approximation)

Assume that |(p,α,φ)|𝒜σ is finite. Given a tolerance Tol, the algorithm used in Mark in the second case (μℓ2>κ⁢λℓ2) in Algorithm 1 computes 𝒯⋆∈𝕋 with

card⁢(𝒯⋆)-card⁢(𝒯0)≲Tol-1/(2⁢σ) and μ2⁢(𝒯⋆)≤Tol.

The following theorem states optimal convergence rates of Algorithm 1.

Theorem 6.5

Theorem 6.5 (Optimal Convergence Rates of AFEM)

For 0<ρB<1 and sufficiently small 0<κ and 0<θ<1, Algorithm 1 computes sequences of triangulations (Tℓ)ℓ∈N and discrete solutions (pℓ,αℓ)ℓ∈N for the right-hand side φ of optimal rate of convergence in the sense that

(card⁢(𝒯ℓ)-card⁢(𝒯0))s⁢(∥p-pℓ∥L2⁢(Ω)+∥Curl⁡(α-αℓ)∥L2⁢(Ω))≲|(p,α,φ)|𝒜s.

The proof follows from the abstract framework of [16], which employs the bounded overhead [7] of the newest-vertex bisection, under the assumptions (A1)–(A4) and (B2) and (SA) which are proved in Sections 6.2–6.5.

6.2 (A1) Stability and (A2) Reduction

The following two theorems follow from the structure of λ.

Theorem 6.6

Theorem 6.6 (Stability)

Let T⋆ be an admissible refinement of T and M⊆T∩T⋆. Let (pT⋆,αT⋆)∈Xh⁢(T⋆)×Yh⁢(T⋆) and (pT,αT)∈Xh⁢(T)×Yh⁢(T) be the respective discrete solutions to (3.4). Then,

|λ⁢(𝒯⋆,ℳ)-λ⁢(𝒯,ℳ)|≲∥p𝒯⋆-p𝒯∥L2⁢(Ω).

Proof.

This follows with triangle inequalities, inverse inequalities and the trace inequality from [10, p. 282] as in [18, Proposition 3.3]. ∎

Theorem 6.7

Theorem 6.7 (Reduction)

Let T⋆ be an admissible refinement of T. Then there exist 0<ρ2<1 and Λ2<∞ such that

λ2⁢(𝒯⋆,𝒯⋆∖𝒯)≤ρ2⁢λ2⁢(𝒯,𝒯∖𝒯⋆)+Λ2⁢∥p𝒯⋆-p𝒯∥L2⁢(Ω)2.

Proof.

This follows with a triangle inequality and the mesh-size reduction property h𝒯⋆2|T≤h𝒯2|T/2 for all T∈𝒯⋆∖𝒯 as in [18, Corollary 3.4]. ∎

6.3 (A4) Discrete Reliability

The following theorem proves discrete reliability, i.e., the difference between two discrete solutions is bounded by the error estimators on refined triangles only.

Theorem 6.8

Theorem 6.8 (Discrete Reliability)

Let T⋆ be an admissible refinement of T with respective discrete solutions (pT⋆,αT⋆)∈Xh⁢(T⋆)×Yh⁢(T⋆) and (pT,αT)∈Xh⁢(T)×Yh⁢(T). Then,

∥p𝒯-p𝒯⋆∥L2⁢(Ω)2+∥Curl⁡(α𝒯-α𝒯⋆)∥L2⁢(Ω)2≲λ2⁢(𝒯,𝒯∖𝒯⋆)+μ2⁢(𝒯,𝒯∖𝒯⋆).

Proof.

Recall the definition of Wh⁢(𝒯⋆) from (3.7). Since p𝒯-p𝒯⋆∈Xh⁢(𝒯⋆), there exist σ𝒯⋆∈Wh⁢(𝒯⋆) and r𝒯⋆∈Yh⁢(𝒯⋆) with p𝒯-p𝒯⋆=σ𝒯⋆+Curl⁡r𝒯⋆. Since Wh⁢(𝒯⋆)⊥L2⁢(Ω)Curl⁡Yh⁢(𝒯⋆), it holds

∥σ𝒯⋆∥L2⁢(Ω)2+∥Curl⁡r𝒯⋆∥L2⁢(Ω)2=∥p𝒯-p𝒯⋆∥L2⁢(Ω)2.

The orthogonality furthermore implies that the discrete error can be split as

∥p𝒯-p𝒯⋆∥L2⁢(Ω)2=(p𝒯-p𝒯⋆,σ𝒯⋆)L2⁢(Ω)+(p𝒯-p𝒯⋆,Curl⁡r𝒯⋆)L2⁢(Ω).

The projection property, Lemma 3.12, proves ΠXh⁢(𝒯)⁢σ𝒯⋆∈Wh⁢(𝒯). Hence, problem (3.4) implies that the first term of the right-hand side equals

(p𝒯-p𝒯⋆,σ𝒯⋆)L2⁢(Ω)=(ΠXh⁢(𝒯)⁢φ-φ,σ𝒯⋆)L2⁢(Ω)=(ΠXh⁢(𝒯)⁢φ-ΠXh⁢(𝒯⋆)⁢φ,σ𝒯⋆)L2⁢(Ω).

For any triangle T∈𝒯∩𝒯⋆, it holds (ΠXh⁢(𝒯)φ-ΠXh⁢(𝒯⋆)φ)|T=0. Therefore,

(ΠXh⁢(𝒯)⁢φ-ΠXh⁢(𝒯⋆)⁢φ,σ𝒯⋆)L2⁢(Ω)≤∥ΠXh⁢(𝒯)⁢φ-ΠXh⁢(𝒯⋆)⁢φ∥𝒯∖𝒯⋆⁢∥σ𝒯⋆∥L2⁢(Ω).

Since 𝒯⋆ is a refinement of 𝒯, it holds

∥ΠXh⁢(𝒯)⁢φ-ΠXh⁢(𝒯⋆)⁢φ∥𝒯∖𝒯⋆=∥ΠXh⁢(𝒯⋆)⁢(ΠXh⁢(𝒯)⁢φ-φ)∥𝒯∖𝒯⋆≤∥φ-ΠXh⁢(𝒯)⁢φ∥𝒯∖𝒯⋆.

Let r𝒯∈Yh⁢(𝒯) denote the quasi interpolant from [39] of r𝒯⋆ which satisfies the approximation and stability properties

∥h𝒯-1⁢(r𝒯⋆-r𝒯)∥L2⁢(Ω)+∥Curl⁡(r𝒯⋆-r𝒯)∥L2⁢(Ω)≲∥Curl⁡r𝒯⋆∥L2⁢(Ω)

and (r𝒯)|E=(r𝒯⋆)|E for all edges E∈ℰ⁢(𝒯)∩ℰ⁢(𝒯⋆). Since p𝒯∈Wh⁢(𝒯) and p𝒯⋆∈Wh⁢(𝒯⋆), it holds

(p𝒯-p𝒯⋆,Curl⁡r𝒯⋆)L2⁢(Ω)=(p𝒯,Curl⁡(r𝒯⋆-r𝒯))L2⁢(Ω).

An integration by parts leads to

(6.4)(p𝒯,Curl⁡(r𝒯⋆-r𝒯))L2⁢(Ω)=-(curlNC⁡p𝒯,r𝒯⋆-r𝒯)L2⁢(Ω)+∑E∈ℰ⁢(𝒯)∫E[p𝒯⋅τE]E⁢(r𝒯⋆-r𝒯)⁢𝑑s.

For a triangle T∈𝒯∩𝒯⋆, any edge E∈ℰ⁢(T) satisfies E∈ℰ⁢(𝒯)∩ℰ⁢(𝒯⋆). Hence, (r𝒯)|T=(r𝒯⋆)|T for all triangles T∈𝒯∩𝒯⋆. This, the Cauchy inequality and the approximation and stability properties of the quasi interpolant lead to

(6.5)-(curlNC⁡p𝒯,r𝒯⋆-r𝒯)L2⁢(Ω)≲∥h𝒯⁢curlNC⁡p𝒯∥𝒯∖𝒯⋆⁢∥Curl⁡r𝒯⋆∥L2⁢(Ω).

Since (r𝒯)|E=(r𝒯⋆)|E for all edges E∈ℰ⁢(𝒯)∩ℰ⁢(𝒯⋆), the approximation and stability properties of the quasi interpolant and the trace inequality [10, p. 282] lead to

(6.6)∑E∈ℰ∫E[p𝒯⋅τE]E⁢(r𝒯⋆-r𝒯)⁢𝑑s≲∑E∈ℰ⁢(𝒯)∖ℰ⁢(𝒯⋆)hT⁢∥[p𝒯⋅τE]E∥L2⁢(E)2⁢∥Curl⁡r𝒯⋆∥L2⁢(Ω).

The combination of the previous displayed inequalities yields

∥p𝒯-p𝒯⋆∥L2⁢(Ω)2≲λ2⁢(𝒯,𝒯∖𝒯⋆)+μ2⁢(𝒯,𝒯∖𝒯⋆).

Since Curl⁡α𝒯=ΠXh⁢(𝒯)⁢φ-p𝒯 and Curl⁡α𝒯⋆=ΠXh⁢(𝒯⋆)⁢φ-p𝒯⋆, the triangle inequality yields the assertion. ∎

The discrete reliability of Theorem 6.8 together with the convergence of the discretization proves reliability of the residual-based error estimator. This is summarized in the following proposition.

Proposition 6.9

Proposition 6.9 (Efficiency and Reliability of the Residual-Based Error Estimator)

Let (p,α)∈X×Y and (ph,αh)∈Xh⁢(T)×Yh⁢(T) be the solutions to (3.3) and (3.4) for some T∈T. There exist constants Ceff,Crel>0 with

Ceff-2⁢(λ2⁢(𝒯,𝒯)+μ2⁢(𝒯))≤∥p-ph∥L2⁢(Ω)2+∥Curl⁡(α-αh)∥L2⁢(Ω)2≤Crel2⁢(λ2⁢(𝒯,𝒯)+μ2⁢(𝒯)).

Proof.

The a priori error estimate from Theorem 3.9 implies the convergence of the discrete solutions. This and Theorem 6.8 prove the reliability. The efficiency follows from the standard bubble function technique [43]. ∎

6.4 (A3) Quasi-Orthogonality

The following theorem proves quasi-orthogonality of the discretization (3.4).

Theorem 6.10

Theorem 6.10 (General Quasi-Orthogonality)

Let (Tj∣j∈N) be some sequence of triangulations with discrete solutions (pj,αj)∈Xh⁢(Tj)×Yh⁢(Tj) to (3.4). Let ℓ∈N. Then,

∑j=ℓ∞(∥pj-pj-1∥L2⁢(Ω)2+∥Curl⁡(αj-αj-1)∥L2⁢(Ω)2)≲λℓ-12+μℓ-12.

Proof.

The projection property, Lemma 3.12, proves ΠXh⁢(𝒯j-1)⁢pj∈Wh⁢(𝒯j-1) with Wh⁢(𝒯j-1) from (3.7). Hence, problem (3.4) leads to

(pj-1,pj-pj-1)L2⁢(Ω)=(φ,ΠXh⁢(𝒯j-1)⁢pj-pj-1)L2⁢(Ω),

(pj,pj-pj-1)L2⁢(Ω)=(φ,pj)-(φ,ΠXh⁢(𝒯j-1)⁢pj)L2⁢(Ω).

The subtraction of these two equations and an index shift lead, for any M∈ℕ with M>ℓ, to

∑j=ℓM∥pj-pj-1∥L2⁢(Ω)2=∑j=ℓM(φ,pj-ΠXh⁢(𝒯j-1)⁢pj)L2⁢(Ω)-∑j=ℓM(φ,ΠXh⁢(𝒯j-1)⁢pj)L2⁢(Ω)+∑j=ℓ-1M-1(φ,pj)L2⁢(Ω)

(6.7)=(φ,pℓ-1-pM)L2⁢(Ω)+2⁢∑j=ℓM(φ,pj-ΠXh⁢(𝒯j-1)⁢pj)L2⁢(Ω).

Since pj-ΠXh⁢(𝒯j-1)⁢pj∈Xh⁢(𝒯j) is L2 orthogonal to Xh⁢(𝒯j-1), a Cauchy and a weighted Young inequality imply

2⁢∑j=ℓM(φ,pj-ΠXh⁢(𝒯j-1)⁢pj)L2⁢(Ω)=2⁢∑j=ℓM(ΠXh⁢(𝒯j)⁢φ-ΠXh⁢(𝒯j-1)⁢φ,pj-ΠXh⁢(𝒯j-1)⁢pj)L2⁢(Ω)

(6.8)≤2⁢∑j=ℓM∥ΠXh⁢(𝒯j)⁢φ-ΠXh⁢(𝒯j-1)⁢φ∥L2⁢(Ω)2+12⁢∑j=ℓM∥pj-ΠXh⁢(𝒯j-1)⁢pj∥L2⁢(Ω)2.

The orthogonality ΠXh⁢(𝒯j)⁢φ-ΠXh⁢(𝒯j-m)⁢φ⊥L2⁢(Ω)Xh⁢(𝒯j-m) for all 0≤m≤j proves

(6.9)∑j=ℓM∥ΠXh⁢(𝒯j)⁢φ-ΠXh⁢(𝒯j-1)⁢φ∥L2⁢(Ω)2=∥ΠXh⁢(𝒯M)⁢φ-ΠXh⁢(𝒯ℓ-1)⁢φ∥L2⁢(Ω)2.

The definition of μℓ yields

(6.10)∥ΠXh⁢(𝒯M)⁢φ-ΠXh⁢(𝒯ℓ-1)⁢φ∥L2⁢(Ω)=∥ΠXh⁢(𝒯M)⁢(φ-ΠXh⁢(𝒯ℓ-1)⁢φ)∥L2⁢(Ω)≤μℓ-1.

The combination of (6.7)–(6.10) and ∥pj-ΠXh⁢(𝒯j-1)⁢pj∥L2⁢(Ω)≤∥pj-pj-1∥L2⁢(Ω) leads to

(6.11)12⁢∑j=ℓM∥pj-pj-1∥L2⁢(Ω)2≤2⁢μℓ-12+(φ,pℓ-1-pM)L2⁢(Ω).

The combination of the arguments of (6.4), (6.5) and (6.6) proves

(6.12)(Curl⁡(αM-αℓ-1),pℓ-1)L2⁢(Ω)≲λℓ-1⁢∥Curl⁡(αM-αℓ-1)∥L2⁢(Ω).

This, the discrete problem (3.4), and the discrete reliability ∥Curl⁡(αM-αℓ-1)∥L2⁢(Ω)≲λℓ-1+μℓ-1 from Theorem 6.8 lead to

(pℓ-1-pM,ΠXh⁢(𝒯ℓ-1)⁢φ)L2⁢(Ω)=(pℓ-1-pM,pℓ-1+Curl⁡αℓ-1)L2⁢(Ω)

=(pℓ-1-pM,pℓ-1)L2⁢(Ω)

=(Curl⁡(αM-αℓ-1),pℓ-1)L2⁢(Ω)

≲λℓ-1⁢∥Curl⁡(αM-αℓ-1)∥L2⁢(Ω)

≲(λℓ-1+μℓ-1)2.

This and a further application of Theorem 6.8 lead to

(φ,pℓ-1-pM)L2⁢(Ω)=(φ-ΠXh⁢(𝒯ℓ-1)⁢φ,pℓ-1-pM)L2⁢(Ω)+(pℓ-1-pM,ΠXh⁢(𝒯ℓ-1)⁢φ)L2⁢(Ω)

≲∥φ-ΠXh⁢(𝒯ℓ-1)⁢φ∥L2⁢(Ω)⁢∥pℓ-1-pM∥L2⁢(Ω)+(λℓ-1+μℓ-1)L2⁢(Ω)2

(6.13)≲(λℓ-1+μℓ-1)2.

The combination of (6.11) with (6.13) implies

(6.14)∑j=ℓM∥pj-pj-1∥L2⁢(Ω)2≲λℓ-12+μℓ-12.

The Young inequality, the triangle inequality, and Curl⁡αj=ΠXh⁢(𝒯j)⁢φ-pj imply

∑j=ℓM∥Curl⁡(αj-αj-1)∥L2⁢(Ω)2≤2⁢∑j=ℓM∥pj-pj-1∥L2⁢(Ω)2+2⁢∑j=ℓM∥ΠXh⁢(𝒯j)⁢φ-ΠXh⁢(𝒯j-1)⁢φ∥L2⁢(Ω)2.

Since M>ℓ is arbitrary, the combination with (6.9), (6.10), and (6.14) yields the assertion. ∎

6.5 (B) Data Approximation

The following theorem together with Assumption 6.4 forms the axiom (B) from [16].

Theorem 6.11

Theorem 6.11 ((B2) Quasimonotonicity and (SA) Sub-Additivity)

Any admissible refinement T⋆ of T satisfies

μ2⁢(𝒯⋆)≤μ2⁢(𝒯) 𝑎𝑛𝑑 ∑T∈𝒯⋆,T⊆Kμ2⁢(T)≤μ2⁢(K) for all ⁢K∈𝒯.

Proof.

This follows directly from the definition of μ. ∎

7 Extension to 3D

This section is devoted to the generalization to 3D. Section 7.1 defines the novel discretization and comments on basic properties, while Section 7.2 is devoted to optimal convergence rates for the adaptive algorithm.

7.1 Weak Formulation and Discretization

For this section, let Ω⊆ℝ3 be a simply connected, bounded, polygonal Lipschitz domain in ℝ3. For the sake of simplicity, we also assume that ∂⁡Ω is connected (i.e., Ω is contractible). The Curl operator acts on a sufficiently smooth vector field β:Ω→ℝ3 as Curl⁡β=∇∧β with the cross product or vector product ∧. Let H⁢(Curl,Ω) denote the space of all β∈L2⁢(Ω;ℝ3) with Curl⁡β∈L2⁢(Ω;ℝ3) for the weak Curl, i.e.,

∫Ωv⋅Curl⁡β⁢d⁢x=∫Ωβ⋅Curl⁡v⁢d⁢x for all ⁢v∈Cc∞⁢(Ω;ℝ3).

In contrast to the two-dimensional case, H⁢(Curl,Ω)≠H1⁢(Ω;ℝ3). The Helmholtz decomposition in 3D reads

L2⁢(Ω;ℝ3)=∇⁡H01⁢(Ω)⊕Curl⁡H⁢(Curl,Ω)

and the sum is L2 orthogonal. It is a consequence of the identity

{r∈H⁢(div,Ω)∣div⁡r=0}=Curl⁡H⁢(Curl,Ω)

in the De Rham complex [9].

Let φ∈H⁢(div,Ω) with -div⁡φ=f. Then the Poisson problem (3.1) is equivalent to the problem: Find (p,α)∈L2⁢(Ω;ℝ3)×H⁢(Curl,Ω) with

(7.1)

(7.1a)(p,q)L2⁢(Ω)+(q,Curl⁡α)L2⁢(Ω)=(φ,q)L2⁢(Ω) ⁢for all ⁢q∈L2⁢(Ω;ℝ3),

(7.1b)(p,Curl⁡β)L2⁢(Ω)=0 ⁢for all ⁢β∈H⁢(Curl,Ω).

In contrast to the two-dimensional case, the operator Curl:H⁢(Curl,Ω)→L2⁢(Ω;ℝ3) has a non-trivial kernel. Classical results [35] characterize this kernel as ∇⁡H1⁢(Ω). To enforce uniqueness, we can reformulate (7.1) as follows: Find (p,α,w)∈L2⁢(Ω;ℝ3)×H⁢(Curl,Ω)×(H1⁢(Ω)∩L02⁢(Ω)) with

(p,q)L2⁢(Ω)+(q,Curl⁡α)L2⁢(Ω)=(φ,q)L2⁢(Ω) ⁢for all ⁢q∈L2⁢(Ω;ℝ3),

(p,Curl⁡β)L2⁢(Ω)+(β,∇⁡w)L2⁢(Ω)=0 ⁢for all ⁢β∈H⁢(Curl,Ω),

(α,∇⁡v)L2⁢(Ω)=0 ⁢for all ⁢v∈(H1⁢(Ω)∩L02⁢(Ω)).

Note that {β∈H⁢(Curl,Ω)∣Curl⁡β=0}=∇⁡H1⁢(Ω) implies w=0.

Standard finite element spaces to discretize H⁢(Curl,Ω) in 3D are the Nédélec finite element spaces [30, 31] (also called edge elements) which are known from the context of Maxwell’s equations. Let 𝒯 be a regular triangulation of Ω in tetrahedra in the sense of [19]. The spaces of first kind Nédélec finite elements read

YN,k⁢(T):=Pk⁢(T;ℝ3)+(x∧Pk⁢(T;ℝ3)),

YN,k(𝒯):={βh∈H(Curl,Ω)∣∀T∈𝒯:βh|T∈YN,k(T)}.

Let Xh⁢(𝒯):=Pk⁢(𝒯;ℝ3). Since Curl⁡YN,k⁢(𝒯)⊆Xh⁢(𝒯), a generalization of (3.4) to 3D seeks (ph,αh)∈Xh⁢(𝒯)×YN,k⁢(𝒯) with

(7.2)

(7.2a)(ph,qh)L2⁢(Ω)+(qh,Curl⁡αh)L2⁢(Ω)=(φ,qh)L2⁢(Ω) ⁢for all ⁢qh∈Xh⁢(𝒯),

(7.2b)(ph,Curl⁡βh)L2⁢(Ω)=0 ⁢for all ⁢βh∈YN,k⁢(𝒯).

The discrete exact sequence [9] implies that the elements in YN,k⁢(𝒯) with vanishing Curl are exactly the gradients of functions in Uh⁢(𝒯):=Pk+1⁢(𝒯)∩H1⁢(Ω)∩L02⁢(Ω). Therefore, the uniqueness in (7.2) can be obtained in the following formulation: Find (ph,αh,wh)∈Xh⁢(𝒯)×YN,k⁢(𝒯)×Uh⁢(𝒯) with

(7.3)

(7.3a)(ph,qh)L2⁢(Ω)+(qh,Curl⁡αh)L2⁢(Ω)=(φ,qh)L2⁢(Ω) ⁢for all ⁢qh∈Xh⁢(𝒯),

(7.3b)(ph,Curl⁡βh)L2⁢(Ω)+(βh,∇⁡wh)L2⁢(Ω)=0 ⁢for all ⁢βh∈YN,k⁢(𝒯),

(7.3c)(αh,∇⁡vh)L2⁢(Ω)=0 ⁢for all ⁢vh∈Uh⁢(𝒯).

Note that ∇⁡Uh⁢(𝒯) is the kernel of Curl:YN,k⁢(𝒯)→Pk⁢(𝒯;ℝ3) and so (7.3) implies wh=0. This variable is introduced in order that (7.3) has the form of a standard mixed system. The discrete Helmholtz decomposition of [1, Lemma 5.4] proves that for the lowest order discretization k=0, ph is a Crouzeix–Raviart function and so (7.3) can be seen as a generalization of the non-conforming Crouzeix–Raviart FEM to higher polynomial degrees.

The inf-sup condition follows from ∇⁡Uh⁢(𝒯)⊆YN,k⁢(𝒯) and Curl⁡YN,k⁢(𝒯)⊆Xh⁢(𝒯). This and the conformity of the method lead to the best-approximation result

∥p-ph∥L2⁢(Ω)+∥Curl⁡(α-αh)∥L2⁢(Ω)+∥∇⁡(w-wh)∥L2⁢(Ω)

≲minqh∈Xh⁢(𝒯)∥p-qh∥L2⁢(Ω)+minβh∈YN,k⁢(𝒯)∥Curl(α-βh)∥L2⁢(Ω)+minsh∈Uh⁢(𝒯)∥∇(w-sh)∥L2⁢(Ω).

Since w=wh=0, this is equivalent to

∥p-ph∥L2⁢(Ω)+∥Curl(α-αh)∥L2⁢(Ω)≲minqh∈Xh⁢(𝒯)∥p-qh∥L2⁢(Ω)+minβh∈YN,k⁢(𝒯)∥Curl(α-βh)∥L2⁢(Ω).

The following lemma states a projection property similar to Lemma 3.12 for the two-dimensional case. To this end, define

Zh⁢(𝒯):={βh∈YN,k⁢(𝒯)∣∀vh∈Uh⁢(𝒯):(βh,∇⁡vh)L2⁢(Ω)=0},

Wh⁢(𝒯):={qh∈Xh⁢(𝒯)∣∀βh∈Zh⁢(𝒯):(qh,Curl⁡βh)L2⁢(Ω)=0}.

Since ∇⁡Uh⁢(𝒯) is the kernel of Curl:YN,k⁢(𝒯)→Xh⁢(𝒯), it holds

Curl⁡YN,k⁢(𝒯)=Curl⁡Zh⁢(𝒯).

This implies

Wh⁢(𝒯)={qh∈Xh⁢(𝒯)∣∀βh∈YN,k⁢(𝒯):(qh,Curl⁡βh)L2⁢(Ω)=0}.

Lemma 7.1

Lemma 7.1 (Projection Property)

Let q∈L2⁢(Ω;R3) with (q,Curl⁡β)L2⁢(Ω)=0 for all β∈H⁢(Curl,Ω) (that means that q is a gradient of an H01⁢(Ω) function). Then ΠXh⁢(T)⁢q∈Wh⁢(T). If T⋆ is an admissible refinement of T, then ΠXh⁢(T)⁢Wh⁢(T⋆)⊆Wh⁢(T).

Proof.

Since Curl⁡YN,k⁢(𝒯)⊆Xh⁢(𝒯) and YN,k⁢(𝒯)⊆H⁢(Curl,Ω), the assertion follows with the arguments in the proof of Lemma 3.12. ∎

7.2 Adaptive Algorithm

This subsection outlines the proof of optimal convergence rates for Algorithm 1 in 3D driven by the error estimators λ and μ defined by the local contributions

λ2⁢(𝒯ℓ,T):=∥h𝒯⁢CurlNC⁡ph∥L2⁢(T)2+hT⁢∑E∈ℰ⁢(T)∥[ph∧νE]E∥L2⁢(E)2,

μ2⁢(T):=∥φ-ΠXh⁢(𝒯)⁢φ∥L2⁢(T)2

and (6.2), (6.3). Here, ℰ⁢(T) denotes the faces of a tetrahedron T∈𝒯 and h𝒯∈P0⁢(𝒯) denotes the piecewise constant mesh-size function defined by h𝒯|T:=hT:=meas3(T)1/3. The refinement of triangulations in Algorithm 1 is done by newest-vertex bisection [40]. Let 𝕋⁢(N) denote the space of admissible triangulations with at most N tetrahedra more than 𝒯0. As in Section 6.1, define the seminorm

|(p,α,φ)|𝒜s:=supN∈ℕ0⁡Ns⁢inf𝒯∈𝕋⁢(N)⁡(∥p-ΠXh⁢(𝒯)⁢p∥L2⁢(Ω)+infβ𝒯∈YN,k⁢(𝒯)⁡∥Curl⁡(α-β𝒯)∥L2⁢(Ω)+∥φ-ΠXh⁢(𝒯)⁢φ∥L2⁢(Ω)).

Assume that Assumption 6.4 holds. The following theorem states optimal convergence rates for Algorithm 1 for 3D.

Theorem 7.2

Theorem 7.2 (Optimal Convergence Rates of AFEM for 3D)

Let s>0. For 0<ρB<1 and sufficiently small 0<κ and 0<θ<1, Algorithm 1 computes sequences of triangulations (Tℓ)ℓ∈N and discrete solutions (pℓ,αℓ)ℓ∈N for the right-hand side φ of optimal rate of convergence in the sense that

(card⁢(𝒯ℓ)-card⁢(𝒯0))s⁢(∥p-pℓ∥L2⁢(Ω)+∥Curl⁡(α-αℓ)∥L2⁢(Ω))≲|(p,α,φ)|𝒜s.

The proof follows as in Section 6 from (A1)–(A4) and (B) from [16] and the efficiency of λ and μ. The proof of efficiency follows with the standard bubble function technique [43]. The proofs of the axioms (A1)–(A4) and (B) are outlined in the following.

The axioms (A1) stability and (A2) reduction follow as in Section 6.2 with triangle inequalities, inverse inequalities, a trace inequality similar to [10, p. 282], and the mesh-size reduction property h𝒯⋆3|T≤h𝒯3|T/2 for all T∈𝒯⋆∖𝒯. However, for (A3) quasi-orthogonality and (A4) discrete reliability, the interpolation operator of [39] cannot be applied directly to r𝒯⋆∈YN,k⁢(𝒯⋆) as done in the proof of Theorem 6.8, because YN,k⁢(𝒯⋆)⊈H1⁢(Ω;ℝ3). This can be overcome by a quasi-interpolation based on a quasi-interpolation operator from [38] and a projection operator from [44]. Its properties are summarized in the following theorem.

Theorem 7.3

Theorem 7.3 (Quasi-Interpolation)

Let T⋆ be an admissible refinement of T and define

ℛ⁢(𝒯,𝒯⋆):={T∈𝒯∣∃K1∈𝒯∖𝒯⋆,K2∈𝒯⁢ with ⁢K1∩K2≠∅⁢ and ⁢T∩K2≠∅}.

Let γT⋆∈Zh⁢(T⋆). Then there exist γT∈YN,k⁢(T), ρ∈H1⁢(Ω), and Φ∈H1⁢(Ω;R3) with

γ𝒯⋆-γ𝒯=∇⁡ρ+Φ,

(γ𝒯⋆-γ𝒯)|T=0 for all T∈𝒯∖ℛ(𝒯,𝒯⋆),

∥h𝒯-1⁢Φ∥L2⁢(Ω)+∥∇⁡Φ∥L2⁢(Ω)≲∥Curl⁡γ𝒯⋆∥L2⁢(Ω).

Proof.

This follows as in the proof of [44, Theorem 5.3] and with the ellipticity on the discrete kernel from [2, Proposition 4.6]. ∎

The differences between the proof of (A4) discrete reliability and the proof of Theorem 6.8 are outlined in the following. Let (p𝒯⋆,α𝒯⋆)∈Xh⁢(𝒯⋆)×Zh⁢(𝒯⋆) and (p𝒯,α𝒯)∈Xh⁢(𝒯)×Zh⁢(𝒯) denote the discrete solutions to (7.2). As in the proof of Theorem 6.8, let σ𝒯⋆∈Wh⁢(𝒯⋆) and r𝒯⋆∈Zh⁢(𝒯⋆) such that p𝒯-p𝒯⋆=σ𝒯⋆+Curl⁡r𝒯⋆. The first term of the right-hand side of

∥p𝒯-p𝒯⋆∥2=(p𝒯-p𝒯⋆,σ𝒯⋆)L2⁢(Ω)+(p𝒯-p𝒯⋆,Curl⁡r𝒯⋆)L2⁢(Ω)

is estimated as in the proof of Theorem 6.8, while for the second term, the quasi-interpolant r𝒯∈YN,k⁢(𝒯) of r𝒯⋆ with r𝒯⋆-r𝒯=∇⁡ρ+Φ for ρ∈H1⁢(Ω) and Φ∈H1⁢(Ω;ℝ3) from Theorem 7.3 is employed. This yields

(p𝒯-p𝒯⋆,Curl⁡r𝒯⋆)L2⁢(Ω)=(p𝒯,Curl⁡(r𝒯⋆-r𝒯))L2⁢(Ω)=(p𝒯,Curl⁡Φ)L2⁢(Ω).

A piecewise integration by parts and the arguments of the proof of Theorem 6.8 conclude the proof. The crucial point is that Φ∈H1⁢(Ω;ℝ3) is smooth enough to allow for a trace inequality.

The proof of (A3) quasi-orthogonality follows as in the proof of Theorem 6.10 with the projection property of Lemma 7.1 and the following modifications in (6.12). Since (in the analogue notation as in (6.12)) αℓ-1∈Zh⁢(𝒯ℓ-1)⊆YN,k⁢(𝒯M), there exists γM∈Zh⁢(𝒯M) with Curl⁡γM=Curl⁡αℓ-1. Theorem 7.3 guarantees the existence of βℓ-1∈YN,k⁢(𝒯ℓ-1), ρ∈H1⁢(Ω) and Φ∈H1⁢(Ω;ℝ3) with αM-γM-βℓ-1=∇⁡ρ+Φ. This implies in (6.12) that

(Curl⁡(αM-αℓ-1),pℓ-1)L2⁢(Ω)=(Curl⁡(αM-γM-βℓ-1),pℓ-1)L2⁢(Ω)=(Curl⁡Φ,pℓ-1)L2⁢(Ω).

Since Φ∈H1⁢(Ω;ℝ3) is smooth enough, a piecewise integration by parts and the arguments of the proof of Theorem 6.8 then prove

(Curl⁡(αM-αℓ-1),pℓ-1)L2⁢(Ω)≲(λℓ-1+μℓ-1)⁢∥Curl⁡(αM-αℓ-1)∥L2⁢(Ω).

This and the arguments of Theorem 6.10 eventually prove the quasi-orthogonality.

8 Numerical Experiments

This section presents numerical experiments for the discretization (3.4) for k=0,1,2. Sections 8.1–8.3 compute the discrete solutions on sequences of uniformly red-refined triangulations (see Figure a for a red-refined triangle) as well as on sequences of triangulations created by the adaptive Algorithm 1 with bulk parameter θ=0.1 and κ=0.5 and ρ=0.75. The convergence history plots are logarithmically scaled and display the error ∥p-ph∥L2⁢(Ω) against the number of degrees of freedom (ndof) of the linear system resulting from the Schur complement. The underlying L-shaped domain Ω:=(-1,1)2∖([0,1]×[-1,0]) with its initial triangulation is depicted in Figure b.

Figure 1

Red-refined triangle and initial mesh for the L-shaped domain.

(a)

(b)

8.1 L-Shaped Domain, I

The function u given in polar coordinates by

u⁢(r,ϕ)=r2/3⁢sin⁡((2/3)⁢ϕ)

is harmonic. For the following experiment we choose φ≡0 and uD:=g⁢u with perturbation function g∈H2⁢(Ω),

g⁢(x):={0 if ⁢|x|≤1/2,16⁢|x|4-64⁢|x|3+88⁢|x|2-48⁢|x|+9 if ⁢1/2≤|x|≤1,1 if ⁢|x|≥1,

such that g|Γ=1 for Γ:=∂⁡Ω∖({0}×(-1,0)∪(0,1)×{0}). Since u|∂⁡Ω∖Γ=0, it holds uD|∂⁡Ω=u. Let

B1/2⁢(0):={x∈ℝ2∣|x|<1/2}

denote the ball with radius 1/2 and midpoint (0,0). Since g|B1/2⁢(0)=0 and u∈H2⁢(Ω∖B1/2⁢(0)), it holds uD∈H2⁢(Ω).

For non-homogeneous Dirichlet data, the jump [ph]E⋅τE is defined for boundary edges E∈ℰ, E⊆ΓD, with adjacent triangle T+ by

[ph]E⋅τE:=ph|T+⋅τE-∇uD⋅τE.

The error estimator λ is then defined by (6.1)–(6.3). The local data error estimator contributions read

μ2⁢(T):=∥(φ-∇⁡uD)-Πk⁢(φ-∇⁡uD)∥L2⁢(T)2.

The global error estimator μ is defined by (6.2) and (6.3).

The errors and error estimators for the approximation ph∈Pk⁢(𝒯;ℝ2) of ∇⁡u for k=0,1,2 are plotted in Figure 2 against the number of degrees of freedom. The errors and error estimators show an equivalent behavior with an overestimation of approximately 10. Uniform refinement leads to a suboptimal convergence rate of h2/3≈𝗇𝖽𝗈𝖿-1/3 for k=0,1,2. The adaptive refinement reproduces the optimal convergence rates of 𝗇𝖽𝗈𝖿-(k+1)/2 for k=0,1,2. Figure 3 depicts three meshes created by the adaptive algorithm for k=0, 1, and 2 with approximately 1000 degrees of freedom. The singularity at the re-entrant corner leads to a strong refinement towards (0,0), while the refinement for k=0,1 also reflects the behavior of the right-hand side, i.e., one also observes a moderate refinement on the circular ring {x∈Ω∣1/2≤|x|≤1}. The marking with respect to the data-approximation (μℓ2>κ⁢λℓ2 in Algorithm 1) is applied at the first 7 (resp. 5 and 10) levels for k=0 (resp. k=1 and k=2) and then at approximately every third level.

Figure 2

Errors and error estimators from Section 8.1.

Figure 3 Adaptively refined triangulations for the experiment from Section 8.1.

Figure 3

Adaptively refined triangulations for the experiment from Section 8.1.

8.2 L-Shaped Domain, II

For f≡-1 and uD≡0 define φ⁢(x,y):=(1/2)⁢(x,y) with -div⁡φ=f.

The error estimators are plotted against the degrees of freedom in Figure 4 for k=0,1,2. The error estimators show for k=0,1,2 a suboptimal convergence rate of h2/3≈𝗇𝖽𝗈𝖿-1/3 for uniform refinement. The adaptive Algorithm 1 recovers the optimal convergence rate of 𝗇𝖽𝗈𝖿-(k+1)/2. Adaptively refined meshes are depicted in Figure 5 for approximately 1000 degrees of freedom. The strong refinement towards the singularity at the re-entrant corner is clearly visible. The smoothness of φ∈P1⁢(Ω;ℝ2) implies that the data-approximation error estimator μℓ vanishes on all triangulations for k=1,2. For k=0, μℓ does not vanish, nevertheless, since μℓ2≤κ⁢λℓ2 for all ℓ, only the Dörfler marking is applied.

Figure 4

Error estimators for the experiment from Section 8.2.

Figure 5 Adaptively refined triangulations for the experiment from Section 8.2.

Figure 5

Adaptively refined triangulations for the experiment from Section 8.2.

Figure 6

Errors and error estimators for the experiment with singular α from Section 8.3.

Figure 7

Errors and error estimators for the experiment with singular α from Section 8.3 and uniform refinement.

Figure 8 Adaptively refined triangulations for the experiment from Section 8.3.

Figure 8

Adaptively refined triangulations for the experiment from Section 8.3.

8.3 Singular α

This subsection is devoted to a numerical investigation of the dependence of the error ∥p-ph∥L2⁢(Ω) on the regularity of α. The exact smooth solution u∈C∞⁢(Ω) of

-Δu=2sin(πx)sin(πy) in Ω and u|ΓD=0

reads u⁢(x,y)=sin⁡(π⁢x)⁢sin⁡(π⁢y). Define φ=∇⁡u+Curl⁡(α~) with α~∈H1⁢(Ω)∖H2⁢(Ω) defined by

α~⁢(r,ϕ)=r2/3⁢sin⁡(2⁢ϕ/3).

Then φ∈H⁢(div,Ω) with -div⁡φ=f.

The errors and error estimators are plotted in Figure 6 against the number of degrees of freedom. The convergence rate on uniform red-refined meshes for k=1,2 is h2/3≈𝗇𝖽𝗈𝖿-1/3 and, hence, the convergence rate seems to depend on the regularity of α. The errors and error estimators show the same convergence rate. Figure 7 focuses on the results for k=0 and uniform mesh-refinement. The error ∥p-ph∥L2⁢(Ω) and the error estimator λ2+μ2 show a convergence rate between h and h2/3, while ∥Curl⁡(α-αh)∥L2⁢(Ω) converges with a rate of h2/3≈𝗇𝖽𝗈𝖿-1/3 due to the singularity of α. This numerical experiment suggests that the error ∥p-ph∥L2⁢(Ω) does not depend on the regularity of α (at least in a preasymptotic regime). The triangle inequality implies ∥Curl⁡(α-αh)∥L2⁢(Ω)≤∥p-ph∥L2⁢(Ω)+μ. This upper bound is also plotted in Figure 7.

Figure 8 depicts adaptively refined meshes for k=0,1,2 with approximately 1000 degrees of freedom. The singularity of α leads to a strong refinement towards the re-entrant corner. The marking with respect to the data-approximation (μℓ2>κ⁢λℓ2 in Algorithm 1) is only applied at levels 1–5, 7, 12, and 18 for k=0. All other marking steps for k=0,1,2 use the Dörfler marking (μℓ2≤κ⁢λℓ2).

Acknowledgements

The author would like to thank Professor C. Carstensen for valuable discussions.

References

[1] Alonso Rodríguez A., Hiptmair R. and Valli A., Mixed finite element approximation of eddy current problems, IMA J. Numer. Anal. 24 (2004), no. 2, 255–271. 10.1093/imanum/24.2.255Search in Google Scholar

[2] Amrouche C., Bernardi C., Dauge M. and Girault V., Vector potentials in three-dimensional non-smooth domains, Math. Methods Appl. Sci. 21 (1998), no. 9, 823–864. 10.1002/(SICI)1099-1476(199806)21:9<823::AID-MMA976>3.0.CO;2-BSearch in Google Scholar

[3] Arnold D. N. and Brezzi F., Mixed and nonconforming finite element methods: Implementation, postprocessing and error estimates, RAIRO Modél. Math. Anal. Numér. 19 (1985), no. 1, 7–32. 10.1051/m2an/1985190100071Search in Google Scholar

[4] Arnold D. N. and Falk R. S., A uniformly accurate finite element method for the Reissner–Mindlin plate, SIAM J. Numer. Anal. 26 (1989), no. 6, 1276–1290. 10.1137/0726074Search in Google Scholar

[5] Arnold D. N., Falk R. S. and Winther R., Preconditioning in H⁢(div) and applications, Math. Comp. 66 (1997), no. 219, 957–984. 10.1090/S0025-5718-97-00826-0Search in Google Scholar

[6] Becker R., Mao S. and Shi Z., A convergent nonconforming adaptive finite element method with quasi-optimal complexity, SIAM J. Numer. Anal. 47 (2010), no. 6, 4639–4659. 10.1137/070701479Search in Google Scholar

[7] Binev P., Dahmen W. and DeVore R., Adaptive finite element methods with convergence rates, Numer. Math. 97 (2004), no. 2, 219–268. 10.21236/ADA640658Search in Google Scholar

[8] Binev P. and DeVore R., Fast computation in adaptive tree approximation, Numer. Math. 97 (2004), no. 2, 193–217. 10.21236/ADA640671Search in Google Scholar

[9] Boffi D., Brezzi F. and Fortin M., Mixed Finite Element Methods and Applications, Springer Ser. Comput. Math. 44, Springer, Heidelberg, 2013. 10.1007/978-3-642-36519-5Search in Google Scholar

[10] Brenner S. C. and Scott L. R., The Mathematical Theory of Finite Element Methods, 3rd ed., Texts Appl. Math. 15, Springer, New York, 2008. 10.1007/978-0-387-75934-0Search in Google Scholar

[11] Brezzi F., On the existence, uniqueness and approximation of saddle-point problems arising from Lagrangian multipliers, Rev. Franc. Automat. Inform. Rech. Operat. Sér. Rouge 8 (1974), no. R-2, 129–151. 10.1051/m2an/197408R201291Search in Google Scholar

[12] Brezzi F., Fortin M. and Stenberg R., Error analysis of mixed-interpolated elements for Reissner–Mindlin plates, Math. Models Methods Appl. Sci. 1 (1991), no. 2, 125–151. 10.1142/S0218202591000083Search in Google Scholar

[13] Carstensen C., Feischl M., Page M. and Praetorius D., Axioms of adaptivity, Comput. Math. Appl. 67 (2014), no. 6, 1195–1253. 10.1016/j.camwa.2013.12.003Search in Google Scholar PubMed PubMed Central

[14] Carstensen C., Gallistl D. and Schedensack M., Adaptive nonconforming Crouzeix–Raviart FEM for eigenvalue problems, Math. Comp. 84 (2015), no. 293, 1061–1087. 10.1090/S0025-5718-2014-02894-9Search in Google Scholar

[15] Carstensen C., Peterseim D. and Schedensack M., Comparison results of finite element methods for the Poisson model problem, SIAM J. Numer. Anal. 50 (2012), no. 6, 2803–2823. 10.1137/110845707Search in Google Scholar

[16] Carstensen C. and Rabus H., Axioms of adaptivity for separate marking, preprint 2016, http://arxiv.org/abs/1606.02165. Search in Google Scholar

[17] Carstensen C. and Schedensack M., Medius analysis and comparison results for first-order finite element methods in linear elasticity, IMA J. Numer. Anal. 35 (2015), no. 4, 1591–1621. 10.1093/imanum/dru048Search in Google Scholar

[18] Cascon J. M., Kreuzer C., Nochetto R. H. and Siebert K. G., Quasi-optimal convergence rate for an adaptive finite element method, SIAM J. Numer. Anal. 46 (2008), no. 5, 2524–2550. 10.1137/07069047XSearch in Google Scholar

[19] Ciarlet P. G., The Finite Element Method for Elliptic Problems, Stud. Math. Appl. 4, North-Holland, Amsterdam, 1978. 10.1115/1.3424474Search in Google Scholar

[20] Crouzeix M. and Falk R. S., Nonconforming finite elements for the Stokes problem, Math. Comp. 52 (1989), no. 186, 437–456. 10.1090/S0025-5718-1989-0958870-8Search in Google Scholar

[21] Crouzeix M. and Raviart P.-A., Conforming and nonconforming finite element methods for solving the stationary Stokes equations. I, Rev. Franc. Automat. Inform. Rech. Operat. Sér. Rouge 7 (1973), no. R-3, 33–75. 10.1051/m2an/197307R300331Search in Google Scholar

[22] Fortin M., A three-dimensional quadratic nonconforming element, Numer. Math. 46 (1985), no. 2, 269–279. 10.1007/BF01390424Search in Google Scholar

[23] Fortin M. and Soulie M., A nonconforming piecewise quadratic finite element on triangles, Internat. J. Numer. Methods Engrg. 19 (1983), no. 4, 505–520. 10.1002/nme.1620190405Search in Google Scholar

[24] Girault V. and Raviart P.-A., Finite Element Methods for Navier–Stokes Equations, Springer Ser. Comput. Math. 5, Springer, Berlin, 1986. 10.1007/978-3-642-61623-5Search in Google Scholar

[25] Gudi T., A new error analysis for discontinuous finite element methods for linear elliptic problems, Math. Comp. 79 (2010), no. 272, 2169–2189. 10.1090/S0025-5718-10-02360-4Search in Google Scholar

[26] Huang J. and Xu Y., Convergence and complexity of arbitrary order adaptive mixed element methods for the Poisson equation, Sci. China Math. 55 (2012), no. 5, 1083–1098. 10.1007/s11425-012-4384-0Search in Google Scholar

[27] Marini L. D., An inexpensive method for the evaluation of the solution of the lowest order Raviart–Thomas mixed method, SIAM J. Numer. Anal. 22 (1985), no. 3, 493–496. 10.1137/0722029Search in Google Scholar

[28] Matthies G. and Tobiska L., Inf-sup stable non-conforming finite elements of arbitrary order on triangles, Numer. Math. 102 (2005), no. 2, 293–309. 10.1007/s00211-005-0648-8Search in Google Scholar

[29] Maubach J. M. L. and Rabier P. J., Nonconforming finite elements of arbitrary degree over triangles, RANA 0328, Technische Universiteit Eindhoven, 2003. Search in Google Scholar

[30] Nédélec J.-C., Mixed finite elements in ℝ3, Numer. Math. 35 (1980), no. 3, 315–341. 10.1007/BF01396415Search in Google Scholar

[31] Nédélec J.-C., A new family of mixed finite elements in ℝ3, Numer. Math. 50 (1986), no. 1, 57–81. 10.1007/BF01389668Search in Google Scholar

[32] Rabus H., A natural adaptive nonconforming FEM of quasi-optimal complexity, Comput. Methods Appl. Math. 10 (2010), no. 3, 315–325. 10.2478/cmam-2010-0018Search in Google Scholar

[33] Rannacher R. and Turek S., Simple nonconforming quadrilateral Stokes element, Numer. Methods Partial Differential Equations 8 (1992), no. 2, 97–111. 10.1002/num.1690080202Search in Google Scholar

[34] Raviart P.-A. and Thomas J. M., A mixed finite element method for 2nd order elliptic problems, Mathematical Aspects of Finite Element Methods (Rome 1975), Lecture Notes in Math. 606, Springer, Berlin (1977), 292–315. 10.1007/BFb0064470Search in Google Scholar

[35] Rudin W., Principles of Mathematical Analysis, 3rd ed., McGraw-Hill Book, New York, 1976. Search in Google Scholar

[36] Schedensack M., A new discretization for mth-Laplace equations with arbitrary polynomial degrees, SIAM J. Numer. Anal. 54 (2016), no. 4, 2138–2162. 10.1137/15M1013651Search in Google Scholar

[37] Schedensack M., Mixed finite element methods for linear elasticity and the Stokes equations based on the Helmholtz decomposition, ESAIM Math. Model. Numer. Anal. (2016), 10.1051/m2an/2016024. 10.1051/m2an/2016024Search in Google Scholar

[38] Schöberl J., A posteriori error estimates for Maxwell equations, Math. Comp. 77 (2008), no. 262, 633–649. 10.1090/S0025-5718-07-02030-3Search in Google Scholar

[39] Scott L. R. and Zhang S., Finite element interpolation of nonsmooth functions satisfying boundary conditions, Math. Comp. 54 (1990), no. 190, 483–493. 10.1090/S0025-5718-1990-1011446-7Search in Google Scholar

[40] Stevenson R., The completion of locally refined simplicial partitions created by bisection, Math. Comp. 77 (2008), no. 261, 227–241. 10.1090/S0025-5718-07-01959-XSearch in Google Scholar

[41] Stoyan G. and Baran Á., Crouzeix–Velte decompositions for higher-order finite elements, Comput. Math. Appl. 51 (2006), no. 6–7, 967–986. 10.1016/j.camwa.2005.10.011Search in Google Scholar

[42] Veeser A., Approximating gradients with continuous piecewise polynomial functions, Found. Comput. Math. 16 (2016), no. 3, 723–750. 10.1007/s10208-015-9262-zSearch in Google Scholar

[43] Verfürth R., A Review of a Posteriori Error Estimation and Adaptive Mesh-Refinement Techniques, Wiley-Teubner Ser. Adv. Numer. Math., Wiley, Chichester, 1996. Search in Google Scholar

[44] Zhong L., Chen L., Shu S., Wittum G. and Xu J., Convergence and optimality of adaptive edge finite element methods for time-harmonic Maxwell equations, Math. Comp. 81 (2012), no. 278, 623–642. 10.1090/S0025-5718-2011-02544-5Search in Google Scholar

Received: 2016-9-27

Accepted: 2016-9-29

Published Online: 2016-10-18

Published in Print: 2017-1-1

A New Generalization of the P1 Non-Conforming FEM to Higher Polynomial Degrees

Abstract

1 Introduction

2 Notation

3 Problem Formulation and Discretization

3.1 New Mixed Formulation of the Poisson Problem

Remark 3.2 (Existence of Solutions)

Remark 3.3 (Equivalence of (3.1) and (3.3))

Remark 3.4 (Mixed Boundary Conditions)

Remark 3.5 (Multiply Connected Domains)

Remark 3.6 (Computation of φ)

3.2 Discretization

Remark 3.8 (Existence of Discrete Solutions)

Theorem 3.9 (Best-Approximation)

Lemma 3.12 (Projection Property)

Proof.

Remark 3.13 (Computational Costs)

3.3 Equivalence with Crouzeix–Raviart FEM

Proposition 3.14 (Equivalence with CR-NCFEM)

Proof.

Remark 3.15 (Higher Polynomial Degrees)

4 Extensions

4.1 Quadrilateral Finite Elements

Lemma 4.2 (Euler Formula for Quadrilateral Partitions)

Proof.

Theorem 4.3 (Discrete Helmholtz Decomposition on Squares)

Proof of Theorem 4.3.

Remark 4.5 (Arbitrary Quadrilaterals)

4.2 Relation to Mixed Raviart–Thomas FEM

5 Medius Analysis

Theorem 5.1 (Best-Approximation Property)

Lemma 5.3 (Companion)

Proof of Theorem 5.1.

Remark 5.4 (Higher Polynomial Degrees)

6 Adaptive Algorithm

6.1 Adaptive Algorithm and Optimal Convergence Rates

Algorithm 1 (AFEM.)

Remark 6.1 (Separate Versus Collective Marking)

Remark 6.3 (Pure Local Approximation Class)

Assumption 6.4 ((B1) Optimal Data Approximation)

Theorem 6.5 (Optimal Convergence Rates of AFEM)

6.2 (A1) Stability and (A2) Reduction

Theorem 6.6 (Stability)

Proof.

Theorem 6.7 (Reduction)

Proof.

6.3 (A4) Discrete Reliability

Theorem 6.8 (Discrete Reliability)

Proof.

Proposition 6.9 (Efficiency and Reliability of the Residual-Based Error Estimator)

Proof.

6.4 (A3) Quasi-Orthogonality

Theorem 6.10 (General Quasi-Orthogonality)

Proof.

6.5 (B) Data Approximation

Theorem 6.11 ((B2) Quasimonotonicity and (SA) Sub-Additivity)

Proof.

7 Extension to 3D

7.1 Weak Formulation and Discretization

Lemma 7.1 (Projection Property)

Proof.

7.2 Adaptive Algorithm

Theorem 7.2 (Optimal Convergence Rates of AFEM for 3D)

Theorem 7.3 (Quasi-Interpolation)

Proof.

8 Numerical Experiments

8.1 L-Shaped Domain, I

8.2 L-Shaped Domain, II

8.3 Singular α

Acknowledgements

References

Journal and Issue

Articles in the same Issue

A New Generalization of the P₁ Non-Conforming FEM to Higher Polynomial Degrees