We study the homogenization of the steady Stokes equations with no-slip conditions in a domain perforated by ε-periodic obstacles, whose limit is Darcy's law for the homogenized pressure p∗. We prove the quantitative error estimate with rate ε, natural in the presence of a boundary layer, only assuming that f−∇p∗∈H1. Such estimates were previously available only under strong smoothness assumptions, such as f∈C1,1/2. Our geometrical assumptions are also less stringent than in the previous analyses. The outer domain and the reference fluid cell need only be Lipschitz, the perforations may deviate from periodicity in a layer of width O(ε) near the outer boundary, and the solid part may be connected (in three dimensions) and attached to the outer wall. In the absence of a boundary layer, the same argument yields the rate O(ε) assuming only f∈H1.
Introduction
We consider the Stokes problem in the perforated domain Ωε:=Ω∖Bε, Ω⊂Rd, d=2 or 3: find uε:Ωε→Rd and pε:Ωε→R solution of
−Δuε+∇pεdivuεuε=f=0=0in Ωε,in Ωε,on ∂Ωε.(1)
with a given f∈L2(Ω)d. Bε represents periodic perforations placed on a Cartesian grid of period ε. To define them, we start from the unit reference cell Y=(0,1)d, introduce an open set B⊂Y (the reference obstacle) and F=Y∖B (the reference fluid domain); the precise geometric hypotheses are collected in Assumptions 1 and 2 below. For ε>0 and k∈Zd let Ykε=ε(Y+k), Bkε=ε(B+k), Fkε=ε(F+k). A typical choice is
I={k∈Zd:Ykε∩Ω=∅},Bε=(k∈I⋃Bkε∩Ω)∘,Ωε=Ω∖Bε.(2)
The actual definition of Ωε may differ from (4) in a band of width ∼ε near the outer boundary ∂Ω.
To study the behavior of the solution for small ε, one introduces the cell Stokes problems: for i=1,…,d, find wi:F→Rd and πi:F→R, Zd-periodic, such that
The formal two-scale asymptotic expansion, cf. e.g. [10, 17], suggests that (uε,pε) is close to (u∗,p∗) with
u∗=ε2wi(ε⋅)(fi−∂ip∗),(4)
where p∗ satisfies the Darcy problem on the whole (non-perforated) domain
divA∗(f−∇p∗)=0in Ω,(5)
n⋅A∗(f−∇p∗)=0on ∂Ω.(6)
The constant matrix A∗ is defined by A∗ei=⟨wi⟩, i=1,…,d, where ⟨⋅⟩ stands for the average on F:
⟨v⟩=∣F∣1∫Fv(y)dy.(7)
The homogenization of the Stokes equations was first rigorously studied in [23], where the weak L2 convergence of the velocity and the strong L2 convergence of the pressure were established. The strong L2 convergence of the velocity was later proven in [2], where the case of a connected Bε was also considered (as opposed to the case of Bε composed of isolated obstacles strictly inside each periodic cell Ykε in the earlier literature). More recently, convergence rates of the type
have been obtained. The first such estimate is available in [15], with the sub-optimal rate 6ε instead of ε, for d=2. The convergence with the correct order ε has been proven in [12] for d=2 and in [19, 4, 3] for any d. Apart from their methods of proof (boundary layer correctors in [19] vs. cut-off techniques elsewhere), these articles differ in their assumptions on the perforations: in [12] and [19], Bε is composed of isolated obstacles, whereas [4, 3] cover the technically more involved and physically more realistic case of a connected structure Bε in dimension d=3. All of these papers make rather strong smoothness assumptions on the data f and on the boundary ∂Ω. In [19], f is assumed to be in C1,1/2(Ω)d and the boundary ∂B (hence ∂Bε) of class C1,α. In [4], f−∇p∗ is assumed to be in H2(Ω)d∩C1,α(Ω)d and ∂Bε of class Cm+2 with m>d/2. Obtaining an O(ε) estimate with less regularity on f, than C1,1/2 is singled out as “an interesting and challenging problem” in [19]. Lowering the regularity assumptions on ∂Bε is arguably even more important for applications in engineering and in numerical analysis, where Bε typically has corners and edges.
The purpose of the present article is to establish an estimate of type (10) under weaker regularity assumptions on f and ∂Bε. We only assume that f−∇p∗∈H1(Ω)d. We make no smoothness assumptions on Ω or on the periodic perforations beyond Lipschitz conditions plus some mild restrictions on the behavior of the perforations near ∂Ω (where they are allowed to deviate from the periodic pattern). In particular, Bε can be connected (in dimension d=3) and touch the outer boundary ∂Ω. We note in passing that the preceding works, even those allowing for a connected structure, assume that the structure is separated from the wall ∂Ω by a free space (filled by fluid) of width ∼ε, which is a non-physical situation of a structure floating freely in the fluid. The only exception is Appendix B of [3], where a connected Bε touching ∂Ω is allowed, but Ω is restricted to be a box ∏i=1d(0,niε) with integers ni.
Under the assumptions of the present article, the solutions (wi,πi) of the cell problems have only the regularity H1(F)d×L2(F), and one cannot guarantee that u∗, as defined by (6), belongs to H1(Ωε)d. We circumvent this problem by replacing f−∇p∗ in the definition of u∗ with its interpolant by continuous piecewise polynomial functions on the mesh naturally associated with the periodic perforations, i.e. the uniform Cartesian mesh of size ε. We thus introduce the regularized homogenized velocity
u^∗=ε2wi(ε⋅)Iε0(fi−∂ip∗),(9)
where Iε0 is a Clément-type interpolation [7] onto the Q1 finite element space on the ε-cells Ykε. Iε0 produces continuous functions vanishing on ∂Ω and, if necessary, in a band of width ∼ε near ∂Ω. By construction, u^∗∈H01(Ωε)d, so that the estimate
makes sense. This regularization by finite elements is similar to a device introduced in the theory of operator error estimates for elliptic homogenization [5, 24, 25]. There, the classical first-order approximation u0+εχ(⋅/ε)∇u0 is replaced by u0+εχ(⋅/ε)Sε∇u0, where Sε is the Steklov average over cubes of size ε. The purpose is the same as here: to make sense of the product χ(⋅/ε)Sε∇u0 under minimal regularity assumptions. The estimate O(ε) in the presence of boundary layers is also proved by these methods in [25, 16, 22], see also [18]. Using Q1 interpolant in place of Sε has certain advantages in the present setting: the cut-off near ∂Ω is built into the interpolant, which vanishes outside the interior cells. With Sε, on the contrary, one needs an extension of u0 beyond Ω and a separate cut-off function, as in [18].
Besides the differences in the regularity settings and geometrical assumptions, our contribution differs from the papers on periodic homogenization of Stokes equations on perforated domains mentioned above in its method of proof. It borrows ideas from the finite element analysis, introducing an inf-sup condition between velocity and pressure with respect to a mesh-dependent pressure norm, proved by a local argument in the spirit of the macroelement technique [21, 14]. This norm includes the jumps of the cell means of the pressure across the faces of the ε-mesh, which is particularly reminiscent of the stabilized Q1–P0 finite element methods [11, 20, 14], cf. also the textbook [8].
Assumptions and main result
Assumption 1 (periodic pattern).B⊂Y=(0,1)d is open and non-empty, F=Y∖B is connected and Lipschitz. Moreover, for j=1,…,d, the fluid part of two cells adjacent in the direction ej, i.e. F(j)=(F∪(F+ej)), is connected and Lipschitz.
To formulate the second assumption, we recall the notations Ykε=ε(Y+k), Fkε=ε(F+k), for k∈Zd, and introduce the definition of a special class of domains, in which one can construct the inverse of operator div, cf. Lemma 5 below:
Definition 1 (John domains). A bounded domain ω⊂Rd is a John domain if there exist a point x0∈ω (the John center) and a constant J>0 such that, for every x∈ω, there is a rectifiable curve γ:[0,L]⟶ω parametrized by the arc length and satisfying γ(0)=x, γ(L)=x0, and dist(γ(t),∂ω)≥t/J for every t∈[0,L].
The class of John domains is bigger than that of Lipschitz, but it excludes certain geometrical features, like external cusps or narrow channels connecting two parts of the domain (assuming a moderate value for the John constant J).
Assumption 2 (perforated domain and mesh).Ω⊂Rd, d=2 or 3, is a bounded connected Lipschitz domain and Ωε⊂Ω is open. Moreover, Ω is covered by a mesh Tε of closed cells with pairwise disjoint interiors, Ω=⋃T∈TεT, divided into two disjoint families Tε=Tεin∪Tεbl such that:
(a) (interior cells) Tεin={Ykε:k∈I0} for a set I0⊂Zd, and Ωε∩Ykε=Fkε for all k∈I0;
(b) (boundary layer) every T∈Tεbl satisfies T⊂Ω and dist(x,∂Ω)≤c0ε for all x∈T;
(c) (shape of the layer cells) every T∈Tεbl is of diameter ≤c2ε; ε−1T is Lipschitz with constants independent of ε; the fluid part of any such cell, i.e. T∩Ωε, is a John domain of constant J
and of T has measure ≥c1εd; moreover, for any other cell T′∈Tε adjacent to T, i.e. such that the intersection T∩T′ is of positive (d−1)-measure, the fluid part (T∪T′)∘∩Ωε is also a John domain of constant J. There are at most Ni such cells T′ for any T∈Tεbl.
The parameters c0,c1,c2,J,Ni are independent of ε.
Remark 1 (consequences of the assumptions)
(i) Assumption 1 implies that Fper=⋃k∈Zd(F+k) is connected, hence A∗ is symmetric positive definite. Indeed, ξ⋅A∗ξ=∣F∣1∫F∣∇wξ∣2 with wξ=ξiwi, πξ=ξiπi. If this vanishes, then wξ=0 and ∇πξ=ξ on the connected set Fper, so πξ(y)=ξ⋅y+c, which is periodic only if ξ=0. Assumption 1 covers isolated Lipschitz inclusions B⊂Y in any dimension and, for d=3, connected solid structure crossing the faces of Y, provided the fluid remains connected. It excludes tangential contacts between solid parts, which create cusps in the fluid.
(ii) Assumption 2(c) allows for a solid attached to the wall ∂Ω but forbids tangential contacts or similar situations creating small features of length scale ≪ε. The configuration where a strip of width ∼ε without any solid in it, assumed in previous studies [12, 19, 4], is also allowed. Assumption 2(c) is essentially void in this case, except for the careful choice of the mesh cells Tεbl in the boundary layer, which is always possible.
Interpolation operator. Let Ωεin=(⋃k∈I0Ykε)∘. For a node a∈εZd, let φa be the continuous Q1 hat function of a, i.e. φa(x)=∏s=1dϕ((xs−as)/ε) with ϕ:R→[0,1] the piecewise affine function on the grid Z such that ϕ(t)=δt0 at all nodes t∈Z. Let ωa be support of φa, i.e. the union of the 2d cells of the grid {Ykε}k∈Zd having a as a vertex. If a∈Ωεin, then ωa⊂Ωεin. For v∈L1(Ω) we define the Clément-type interpolant
Iε0v=a∈εZd∩Ωεin∑(∣ωa∣1∫ωav)φa.(11)
Iε0v is continuous on Ω, a Q1 polynomial on every grid cell in Tεbl, and vanishes outside Ωεin, so Iε0v∈W1,∞(Rd).
Theorem 1. Let Assumptions 1 and 2 hold and ε≤1. Let (uε,pε)∈H01(Ωε)d×L2(Ωε) solve (1)–(3), and let p∗∈H1(Ω) be the weak solution of (5)–(6), normalized by ∫Ωεp∗=0. Assume f−∇p∗∈H1(Ω)d. Then, with u∗ defined in (9),
with a constant C independent of ε. C depends only on Ω, F, and the constants in Assumption 2.
Remark 2 (on the regularity hypothesis) If Ω is a domain of class C1,1, and f∈H1(Ω)d, then p∗∈H2(Ω)[9] and ∥f−∇p∗∥H1≤C∥f∥H1. Hence, the error bound (12) holds assuming only f∈H1(Ω)d. On a general Lipschitz domain, f−∇p∗∈H1(Ω)d is an additional hypothesis on (f,p∗): the Neumann problem (5)–(6) only gives p∗∈H3/2(Ω) in general.
Remark 3 (error without boundary layer) If the boundary layer is absent, i.e. if Ω is a box, consisting of whole ε-cells Ykε, with periodic conditions on ∂Ω, and assuming that u∗ is based on the interpolant Iε with the sum in (11) extended to all the nodes εZd, the same proof as below gives
It suffices to get rid of all the bounds related to the boundary layer, no longer needed, and to adapt the interpolation error estimates (21)–(24) gaining there an extra ε (this is seen immediately in the proof of Lemma 9).
Auxiliary results
This section collects several Lemmas necessary for the proof of the main theorem in the forthcoming Section 4. We shall systematically use the letter C to denote constants depending only on d, F, Ω and the constants in Assumption 2; they never depend on ε or on the data. We assume ε≤1.
Compact notation. For a Zd-periodic function ϕ, we systematically write (ϕ)ε(x)=ϕ(x/ε). For a measurable set D, ∥⋅∥D denotes the norm in L2(D). Summation over repeated indices i is implied.
Notations related to the ε-mesh and fluid parts. We recall the notion of the mesh Tε and its submeshes Tεin, Tεbl from Assumption 2. A face is a set F=T∩T′ with T=T′∈Tε such that the (d−1)-measure ∣F∣ is positive (∣F∣>0). We denote by Eε the set of faces and by Einε the set of faces between two interior cells T=T′∈Tεin. For every face we fix once and for all an orientation, i.e. we call one of the two cells, sharing F, T and the other T′, and denote by nF the unit normal to F pointing from T to T′.
For any cell T∈Tε we shall write FTε for its fluid part, i.e. FTε=T∩Ωε (in particular, for the interior cells T=Ykε, k∈I0, the fluid part is FTε=Fkε). Similarly, for a face F=T∩T′ between two cells T,T′∈Tε, we shall denote by FFε the fluid part of the union of these cells, i.e. FFε=(T∪T′)∘∩Ωε.
Poincaré and trace inequalities
Lemma 2 (strip estimates on a Lipschitz domain) For t>0, let Ωt={x∈Ω:dist(x,∂Ω)<t} be the boundary strip of width t.
(a) if w∈H1(Ω), then ∥w∥Ωt≤Ct∥w∥H1(Ω);
(b) if w∈H01(Ω), then ∥w∥Ωt≤Ct∥∇w∥ΩCt.
Proof. Part (a) is borrowed from [18 Proposition 3.1.7]. There, the statement is in fact stronger: ∥w∥Ωt≤Ct∥w∥W1,q(Ω) with q=d+12d. Our statement follows from this since q<2.
To prove part (b), we first consider a function v∈H1(Qr+), where Qr+={x=(x′,xd)∈Rd:∣x′∣<r,xd>0}, vanishing on {xd=0}. For s>0,
Part (b) follows by covering ∂Ω with coordinate patches in which Ω lies above the graph of a Lipschitz function γ, and applying the result above after the bi-Lipschitz change of variables (x′,xd)↦(x′,xd−γ(x′)), which maps Ωt into a strip of width Ct.
Lemma 3 (Poincaré inequality) For ϕ∈H01(Ωε), ∥ϕ∥Ωε≤Cε∣ϕ∣H1(Ωε).
Proof. Extend ϕ by 0 to Ω. By Assumption 2, Ω⊂Ωc0ε∪⋃k∈I0Ykε in the notations of Lemma 2. We can apply part (b) of this Lemma to bound ∥ϕ∥Ωc0ε≤Cε∣ϕ∣H1(Ωc0ε). On Ykε, k∈I0, ϕ vanishes on Bkε, and the scaled Poincaré inequality on Y for functions vanishing on the non-empty open set B gives ∥ϕ∥Ykε≤Cε∣ϕ∣H1(Ykε).
Lemma 4 (trace inequality on the cells).For everyT∈Tεandw∈H1(T∘),
∥w∥∂T2≤C(ε−1∥w∥T2+ε∣w∣H1(T)2).
Consequently, forw∈H1(Ω),
F∈Eε∑∥w∥F2≤C(ε−1∥w∥Ω2+ε∣w∣H1(Ω)2).
Proof. Lemma 4 follows by scaling from the trace inequality ∥w∥∂T2≤C(∥w∥T2+∥w∥T∥∇w∥T) on the Lipschitz domain T=ε−1T, whose constant depends only on the Lipschitz character of T: it is obtained by integrating div(μw2) with a smooth vector field μ transversal to ∂T, constructed from the local graph representations of ∂T, see [9 Lemma 1.5.1.9 and Theorem 1.5.1.10]. □
Pressure means and the inf-sup condition
Lemma 5 (Bogovskiĭ operator).Letω⊂Rdbe a John domain. For everyq∈L02(ω)with∫ωq=0there isv∈H01(ω)dwithdivv=qand∣v∣H1(ω)≤CB∥q∥ω, where the constantCBdepends only on the John constantJofωas in Definition 1.
Proof. The operator defining v is constructed in [6]. The fact that this construction is valid on John domains and that CB depends only on J is proved in [1]. □
For q∈L2(Ωε) and T∈Tε let qT=∣FTε∣−1∫FTεq and Πεq be the piecewise constant function on the mesh Tε defined on every T∈Tε by Πεq∣T=qT. Πε maps L2(Ωε) into itself (after restricting Πεq to the fluid part FTε on every cell T). We introduce the jumps [[Πεq]]F=qT−qT′ across any face F∈Eε, F=T∩T′, using the conventions adopted in the beginning of this Section. Set
Proof. Let Q∈L2(Ω) be equal to Πεq on Ωε and to 0 in the solid. Then ∫ΩQ=∑T∈Tε∣FTε∣qT=∫Ωεq=0 and ∥Q∥Ω=∥Πεq∥Ωε. Let V∈H01(Ω)d satisfy divV=Q and ∥V∥Ω+∣V∣H1(Ω)≤C∥Q∥Ω (this V is given by Lemma 5 since the Lipschitz domain Ω is also John, and the constant C depends only on Ω through its John constant and its diameter, involved in the Poincaré–Friedrichs inequality). Observe
The last equality follows from the divergence theorem on every cell T, rearranging the sums over the cell boundaries, and the fact that V⋅n vanishes on ∂Ω, so that only the contributions on faces F∈Eε survive. We now use the Cauchy–Schwarz inequality and Lemma 4:
using ∣FTε∣≥cεd (Assumption 2(c)). Domain Fε is indeed John with constant J by Assumption 2(c), if one of the cells T,T′ is in Tb. For patches Fε composed of interior cells, the John constant is also independent of ε, cf. Assumption 1, since the John constant is scale invariant.
Set vp=ε∑F∈Eε∣F∣(pT−pT′)ΦF. Since (p,divΦF)Ωε=pT−pT′, we get
For every T∈Tε, (I−Πε)p is of zero mean on FTε. Thus, Lemma 5 gives vT∈H01(FTε)d with divvT=(I−Πε)p on FTε and ∣vT∣H1(FTε)≤CB∥(I−Πε)p∥FTε. The constant CB is ε-independent by the same considerations as above.
Let vp0=∑T∈TεvT (each extended by zero) and w=vp+vp0. Then
since divvp0 has zero mean on every FTε. Recalling ∣vp0∣H1(Ωε)≤CB∥(I−Πε)p∥Ωε and combining with (18) gives the announced bound on the norm of w. □
Remark 4 An alternative definition of the mesh-dependent norm on the pressure space could be
With this definition, Lemma 6 would be a triviality, and Lemma 7 would be simpler to prove. One can work with such definition as well, and obtain almost the same results as with the present definition (15), but the regularity requirement would be stronger (typically f−∇p∗∈H3/2) and the treatment of the boundary layer contributions would be more involved.
We shall also use the combined velocity-pressure space X=H01(Ωε)d×L02(Ωε), equipped with the norm ∥(v,q)∥ε defined for any (v,q)∈X by
∥(v,q)∥ε2=∣v∣H1(Ωε)2+∥∥q∥∥ε2.(18)
Introduce the combined bilinear form c:X×X→R associated to the Stokes equations by
where the sum is taken over all the vertices of T inside Ωεin, and ωT denotes the union of node-associated patches ωa for all such vertices a. Bound (22) follows by summing the first inequality in (27) over all T∈Tεin, since the patches ωT overlap boundedly.
Let Tε,in,0 be the set of cells T∈Tεin whose 2d vertices all lie inside Ωεin, and Tε,in,bl=Tεin∖Tε,in,0. If T∈Tε,in,0, then ωT is a cube of side 3ε and Iε0c=c on T for any constant c. Taking c as the mean of v on ωT, (27) and the Poincaré inequality on ωT give, on T∈Tε,in,0,
If T∈Tε,in,bl, then T touches ∂Ωεin, which lies at distance at most c0ε from ∂Ω by Assumption 2(b). Hence T and ωT, as well as Ωεbl=Ω∖Ωεin, are contained in the strip Ωcε. Summing over all cells and recalling Iε0v=0 on Ωεbl, we obtain thanks to Lemma 2(a), and (27)–(28),
It remains to prove the estimates on faces. For (25), a scaling argument on a cube Ykε gives
∥∇Iε0v∥F2≤Cε−1∣Iε0v∣H1(Ykε)2
for a face F of Ykε, k∈I0. Thus, the left-hand side of (25) is bounded by Cε−1/2∣Iε0v∣H1(Ωin), and (25) follows from (24). For (26), Lemma 4 gives ∑F∈Eε∥v−Iε0v∥F2≤C(ε−1∥v−Iε0v∥2+ε∣v−Iε0v∣H1(Ω)), so that (26) follows from (23) and (24). □
We now recall the classical flux corrector of periodic homogenization [13, 18].
Lemma 10 (flux corrector). For i=1,…,d there is a Zd-periodic skew-symmetric matrix field ϕ(i)∈Hper2(Rd)d×d such that
We assume here that wi is extended by 0 in B=Y∖F.
Proof. Since wi∈H1(F), divwi=0 in Fper and wi=0 on ∂Fper, extending wi by 0 in B gives a field in Hper1(Rd)d and its divergence vanishes everywhere. Its mean over Y is ∫Fwi=∣F∣A∗ei.
Let ψ∈Hper3(Rd)d solve −Δψ=wi−∣F∣A∗ei and set ϕab(i)=∂aψb−∂bψa. Then (divϕ(i))a=∂adivψ−Δψa. Observe −Δdivψ=divwi=0. Thus, divψ is a periodic harmonic function, hence constant, and divϕ(i)=−Δψ. □
Corollary 11. Let h be a function of class C2 on a grid cell Ykε. Then
The last integral vanishes since ϕ is a skew-symmetric matrix and the matrix of second derivatives of h is symmetric. Writing, on ∂Ykε, ∇h=(∇h⋅n)n+∇τh, we conclude ∇h⋅(ϕ(i))εn=(ϕ(i))ε∇τh⋅n since (ϕ(i))εn⋅n=0 by skew-symmetry of ϕ(i). □
Remark 5 (products with oscillating functions). We shall frequently encounter the products of ε-periodic functions, eg. wi,∇wi,πi rescaled by ε, with Q1 functions on mesh Tεh. If one needs to calculate the norm of such a product, then multiplication by the periodic function will act as multiplication by a constant. Indeed, let ψ∈L2(F) be Zd-periodic and h be a polynomial of degree up to d on Ykε. Scaling estimates give then
In the last inequality we made the generic constant C absorb the norm ∥ψ∥F since, in all the cases encountered, it will depend only on F, and the constants C are allowed to hide this dependence.
Proof of Theorem 1
Let us denote uˉ=f−∇p∗ and recall Definition (6) of u^∗, which we rewrite as u^∗=ε2(wi)εIε0uˉ and introduce the pressure reconstruction
Since Iε0uˉ vanishes outside Ωεin, u^∗∈H01(Ωε)d. By construction, pε,1∈L02(Ωε). Take some z∈H01(Ωε)d, to be chosen later in the proof. We have (uε−u^∗+z,pε−pε,1)∈X. By Lemma 8 there is (v,q)∈X with ∥(v,q)∥ε=1 and
The test function (Iε0uˉi)v is admissible since Iε0uˉi∈W1,∞ and its support is in Ωεin, where the perforations are periodic. Denoting σi=∇wi−πiI, we obtain
The portions of the boundaries ∂T lying on ∂Ω could be excluded from the last sum since Iε0uˉ=0 and n⋅A∗uˉ=0 on ∂Ω. We now rewrite the last sum over the boundaries of cells T∈Tε as the sum over the faces F∈Eε, the Cauchy–Schwarz inequality and (26) give
Choice of z. We now choose z so that to cancel exactly in (33) the term III1′′ introduced in (35), i.e. so that
III1′′=(q,divz)Ωε.(35)
We rewrite III1′′ as a sum over faces. For F∈Einε, F=T∩T′ with T=Ykε, T′=Ylε, set
JF=ε3∫F(ϕ(i))ε∇τIε0uˉi⋅nF.
Note that JF is unambiguously defined since the tangential derivatives of Iε0uˉi on F are the same from both sides T and T′. On a face of an interior cell which is not in Einε, Iε0uˉi and its tangential derivatives vanish. Passing from the sum over the boundaries of the interior cells to the sum over their faces, and recalling that nF is the outer normal of T and the inner normal of T′, we get
III1′′=F∈Einε∑(qT−qT′)JF.
On the other hand, the fields ΦF defined in the proof of Lemma 7 by (17) satisfy (q,divΦF)Ωε=qT−qT′. Hence the choice
z=F∈Einε∑JFΦF∈H01(Ωε)d(36)
gives (37). To bound z, note that ∥(ϕ(i))ε∥F≤Cε(d−1)/2, so that ∣JF∣≤Cε3+(d−1)/2∥∇τIε0uˉ∥F by the Cauchy–Schwarz inequality. Since ∣ΦF∣H1(FFε)≤Cε−d/2, we get with (25):
Conclusion. With z chosen above, (37) is satisfied, so that (33) and (35) give −III=III1′+III2. Thus, ∣III∣ is bounded by the sum of (34) and (36). Combining this with (32), (39), and inserting into (31) gives
Use of large language models. Claude (model Opus 5.5) was used to develop the central arguments in the proof of Theorem 1. The article grew out of a first draft from July 2024, and has greatly benefited from interactions with Claude AI in many aspects. The most outstanding contributions were:
From human part: The idea of using the finite element framework on the ε-grid and the mesh-dependent norm (17) on the pressure. As mentioned in Remark 4, this choice lead to technical complications and necessitated stronger regularity assumptions.
From AI part: Introducing the additional velocity correction z, as in (36).
From human part again: The idea of introducing gradient-like terms in the pressure norm (a priori inconsistent with the L2 framework for the pressure). The norm (17) becomes (13).
Claude was also used for several other refinements, proof checks, bibliographical search, and language editing. The author has checked all mathematical arguments and takes full responsibility for the content of the paper.
Acknowledgements
The author warmly thanks Profs. Grégoire Allaire and François Murat for their interest in the first version of this article and enlightening discussions. This work is part of the Franco-German project ”Numerical homogenization techniques for multiscale Stokes problems” (PHC PROCOPE). The exchanges with German partners, in particular Moritz Hauck, were very helpful.
References
[1]G. Acosta, R. G. Durán, and M. A. Muschietti, Solutions of the divergence operator on John domains, Advances in Mathematics, 206 (2006), pp. 373–401.DOI
[2]G. Allaire, Homogenization of the Stokes flow in a connected porous medium, Asymptotic Anal., 2 (1989), pp. 203–222.DOI
[3]L. Balazi, Multi-scale Finite Element Method for incompressible flows in heterogeneous media: Implementation and Convergence analysis, PhD thesis, 2024.DOI
[4]L. Balazi, G. Allaire, and P. Omnes, Sharp convergence rates for the homogenization of the Stokes equations in a perforated domain. working paper or preprint, May 2024.DOI
[5]M. Birman and T. Suslina, Second order periodic differential operators. threshold properties and homogenization, St. Petersburg Mathematical Journal, 15 (2004), pp. 639–714.DOI
[6]M. E. Bogovskii, Solution of the first boundary value problem for the equation of continuity of an incompressible medium, in Dokl. Akad. Nauk SSSR, vol. 248, 1979, pp. 1037–1040.
[7]P. Clément, Approximation by finite element functions using local regularization, Revue Française d’Automatique, Informatique, Recherche Opérationnelle. Analyse Numérique, 9 (1975), pp. 77–84.DOI
[8]H. C. Elman, D. J. Silvester, and A. J. Wathen, Finite Elements and Fast Iterative Solvers: with Applications in Incompressible Fluid Dynamics, Numerical Mathematics and Scientific Computation, Oxford University Press, Oxford, 2nd ed., 2014.
[9]P. Grisvard, Elliptic problems in nonsmooth domains, SIAM, 2011.DOI
[11]T. J. R. Hughes and L. P. Franca, A new finite element formulation for computational fluid dynamics: VII. The Stokes problem with various well-posed boundary conditions: symmetric formulations that converge for all velocity/pressure spaces, Computer Methods in Applied Mechanics and Engineering, 65 (1987), pp. 85–96.DOI
[12]G. Jankowiak and A. Lozinski, Non-Conforming Multiscale Finite Element Method for Stokes Flows in Heterogeneous Media. Part II: error estimates for periodic microstructure, Discrete and Continuous Dynamical Systems-B, 29 (2024), pp. 2298–2332; also available as preprint arXiv:1802.04389 (2018).
[13]V. Jikov, S. Kozlov, and O. Oleinik, Homogenization of differential operators and integral functionals, Springer-Verlag, 1994.DOI
[14]N. Kechkar and D. Silvester, Analysis of locally stabilized mixed finite element methods for the Stokes problem, Mathematics of Computation, 58 (1992), pp. 1–10.DOI
[15]E. Marušić-Paloka and A. Mikelić, An error estimate for correctors in the homogenization of the stokes and navier-stokes equations in a porous medium, Boll. Unione Mat. Ital., 7 (1996), pp. 661–671.
[16]M. A. Pakhinin and T. A. Suslina, Homogenization of the elliptic Dirichlet problem: error estimates in the (L₂ → H¹)-norm, Funktsional. Anal. i Prilozhen., 46 (2012), pp. 92–96.
[17]E. Sánchez-Palencia, Non-homogeneous media and vibration theory, vol. 127 of Lecture notes in physics, 1980.
[18]Z. Shen, Periodic Homogenization of Elliptic Systems, vol. 269 of Operator Theory: Advances and Applications, Birkhäuser, Cham, 2018.DOI
[19]———, Sharp convergence rates for Darcy’s law, Communications in Partial Differential Equations, 47 (2022), pp. 1098–1123.arxiv.org/abs/2011.14169
[20]D. J. Silvester and N. Kechkar, Stabilised bilinear-constant velocity-pressure finite elements for the conjugate gradient solution of the Stokes problem, Computer Methods in Applied Mechanics and Engineering, 79 (1990), pp. 71–86.
[21]R. Stenberg, Analysis of mixed finite element methods for the Stokes problem: a unified approach, Mathematics of Computation, 42 (1984), pp. 9–23.
[22]T. A. Suslina, Homogenization of the Dirichlet problem for elliptic systems: L₂-operator error estimates, Mathematika, 59 (2013), pp. 463–476.
[23]L. Tartar, Incompressible fluid flow in a porous medium-convergence of the homogenization process, Appendix of [17], (1980).
[24]V. Zhikov, On operator estimates in homogenization theory, in Doklady Mathematics, vol. 72, Pleiades Publishing, 2005, pp. 535–538.
[25]V. Zhikov and S. Pastukhova, On operator estimates for some problems in homogenization theory, Russian Journal of Mathematical Physics, 12 (2005), pp. 515–524.