HAL Id: hal-01143558
https://hal.archives-ouvertes.fr/hal-01143558
Preprint submitted on 18 Apr 2015
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
Spectral element discretization of the heat equation with variable diffusion coefficient
Y Daikh, W Chikouche
To cite this version:
Y Daikh, W Chikouche. Spectral element discretization of the heat equation with variable diffusion coefficient. 2015. �hal-01143558�
Spectral element discretization of the heat equation with variable diffusion coefficient
Y. Daikh, W. Chikouche Université de Jijel
Laboratoire de Mathématiques Pures et Appliquées B.P. 98, Ouled Aissa
18000 Jijel, Algeria
yasmina-daikh@univ-jijel.dz, wided-chikouche@univ-jijel.dz.
Abstract
We are interested in the discretization of the heat equation with a diffusion coefficient depending on the space and time variables. The discretization relies on a spectral element method with respect to the space variables and Euler’s implicit scheme with respect to the time variable. A detailed numerical analysis leads to optimal a priori error estimates.
Keywords: Heat equation, diffusion coefficient, spectral element methods, a priori esti- mates.
2010 Mathematics Subject Classification: 35K05, 65N35, 35B45.
1 Introduction
An impressive amount of work has been done concerning a priori and a posteriori analysis of parabolic type problems for finite element methods, see [6] and [1] for instance. An exten- sion in spectral element method of some results obtained by Bergam et al. in [1] has been performed recently by N. Chorfi et al. in [4]. They were interested in the a posteriori analysis of the spectral element discretization of the one-dimensional heat equation with constant diffu- sion coefficient. The spectral element method consists on approaching the solution of a partial differential equation by polynomial functions of high degree on each element of a decomposition.
In this paper, we are interested in the discretization of the heat equation with a diffusion coefficient depending on the space and time variables by an implicit Euler’s scheme with respect to the time variable and spectral element method with respect to the space variables in a two- or three-dimensional bounded domain. For the space discretization, we consider a partition of the domain into rectangles in dimension 2 or rectangular parallelepipeds in dimension 3 which is conforming and without overlap. The discrete spaces are constructed from tensorized spaces of polynomials of the same high degree on each subdomain. The full discrete problem is then obtained by Galerkin method with numerical integration.
An outline of the paper is as follows:
• In Section 2, we present the linear heat equation and we study the continuous problem ant its stability.
• In Section 3, we describe its time semi-discretization and the corresponding stability property.
• Section 4 is devoted to the description of the space discretization of the problem by using spectral element method.
The well-posedness of the corresponding problem in each section is proved.
• Optimal error estimates are proved in Section 5.
2 Position of the problem
LetΩ be a connected and bounded open set in Rd (d = 1,2,or 3)with a Lipschitz-continuous boundary. Also let T be a fixed positif integer. We consider the heat equation
(1)
∂tu−div(λ∇u) = f inΩ×]0, T[
u = 0 on∂Ω×]0, T[
u|t=0 = u0 in Ω,
whereλ is a given continuous function on Ω×[0, T] satisfying for some positive constants λmin and λmax,
(2) ∀x∈Ω,∀t∈[0, T], λmin ≤λ(x, t)≤λmax.
The data are the distribution f and the function u0; the unknown is the function u.
As usual, we denote by Lp(Ω),1 ≤ p ≤ ∞, the Lebesgue spaces and by Hs(Ω), s > 0, the standard Sobolev spaces. The usual norm and seminorm of Hs(Ω) are denoted by k · ks,Ω and
| · |s,Ω respectively. The space H01(Ω) stands for the closure in H1(Ω) of the space of infinitely differentiable functions with a compact support in Ω, and H−1(Ω) stands for its dual space.
For simplicity, we denote by (·,·) the scalar product on L2(Ω) and by k · k0,Ω the associated norm. By extension, the duality pairing between H−1(Ω) and H01(Ω), is also denoted by (·,·).
We define C0(0, T;L2(Ω)), as the space of continuous functions, with values in L2(Ω), and also L2(0, T;H01(Ω)),respectivelyL2(0, T;H−1(Ω)),as the space of square-integrable functions with values in H01(Ω), respectively in H−1(Ω).
The problem (1) admits the equivalent variational formulation:
Find u inC0(0, T;L2(Ω))∩L2(0, T;H01(Ω)) satisfying
(3) u|t=0 =u0 inΩ,
and such that, for a.e. t in]0, T[,
(4) ∀v ∈H01(Ω), (∂tu(t), v) + (λ(t)∇u(t),∇v) = (f(t), v).
It is well known [5, Ch. 3, § 4] that, for any f in L2(0, T;H−1(Ω)) and u0 in L2(Ω), problem (3)-(4) admits a unique solution. Moreover, let us introduce the norm
(5) [[v]](t) =
µ
kv(t)k20,Ω+ Z t
0
kλ12(s)∇v(s)k20,Ω
¶12 .
By taking v equal to u(t) in (4) and integrating on the interval ]0, t[, we easily derive the following estimate [1]: for all t∈[0, T]
(6) [[u]](t)≤
µ
ku0k20,Ω+ 1
λminkfk2L2(0,t;H−1(Ω))
¶1
2
.
3 The time semi-discrete problem
In order to describe the time discretization of equation (1), we introduce a partition of the interval [0, T] into subintervals[tk−1, tk], 1≤k ≤ K, such that 0 =t0 < t1 < ... < tK =T.We denote by τk := tk−tk−1, by τ the K− tuple (τ1, ..., τK) and by |τ| the maximum of the τk, 1≤k≤K. We also define the regularity parameter
στ = max
2≤k≤K
τk τk−1.
With each family (vk)0≤k≤K, we agree to associate the function vτ on [0, T] which is affine on each interval[tk−1, tk],1≤k ≤K, and equal tovk attk,0≤k ≤K.Equivalently, this function can be written, for 1≤k ≤K, as
(7) ∀t∈[tk−1, tk], vτ(t) = vk−tk−t
τk (vk−vk−1).
For simplicity, we introduce the notation λk =λ(tk) and fk =f(tk), which obviously requires the continuity of f with respect to t. The semi-discrete problem issued from Euler’s implicit scheme is now written as
(8)
uk−uk−1
τk −div(λk∇uk) = fk inΩ, 1≤k ≤K,
uk = 0 on∂Ω, 1≤k≤K,
u0 = u0 in Ω.
Equivalently, it admits the variational formulation:
Find (uk)0≤k≤K in L2(Ω)×H01(Ω)K satisfying
(9) u0 =u0 in Ω,
and such that, for1≤k ≤K,
(10) ∀v ∈H01(Ω), ak(uk, v) =Lk(v), where the bilinear forms ak,1≤k ≤K, are defined by
ak(u, v) = (u, v) +τk(λk∇u,∇v),
and the linear forms Lk,1≤k≤K, are defined as
Lk(v) = (uk−1, v) +τk(fk, v).
The existence and uniqueness of a solution (uk)0≤k≤K for any data f in C0(0, T;H−1(Ω)) and u0 inL2(Ω) is now a simple consequence of the Lax-Milgram lemma.
Moreover, by using the notation λkmin = inf
x∈Ωλ(x, tk) and taking v =uk in (10), we easily derive the following estimate
(11) kukk20,Ω+τk
°°(λk)12∇uk°
°2
0,Ω ≤ kuk−1k20,Ω+ τk
λkminkfkk2−1,Ω. We now define the norm on whole sequences v`, 0≤`≤k by
(12) [[(v`)]]k=
Ã
kvkk20,Ω+ Xk
`=1
τ`°
°(λ`)12∇v`°
°2
0,Ω
!12 .
Indeed, by summing up estimate (11) on k,we derive the semi-discrete analogue of (6):
(13) [[(u`)]]k≤
Ã
ku0k20,Ω+ Xk
`=1
τ`
λ`minkf`k2−1,Ω
!12 .
The norm [[(u`)]]k involved in this estimate is not equal to the norm [[uτ]](tk) (see (7) for the definition of the function uτ). However, when u0 is supposed to be in H1(Ω), they are equivalent, as proven in the next lemma [1].
Lemma 3.1 Assume that the function λ is continuously differentiable in time, with maximum value of∂tλ onΩ×[0, T]denoted byµmax.There exists a positive real numberα0,equal to 2µλminmax, such that the following equivalence property holds for |τ| ≤α0 and for any family (v`)0≤`≤K in H1(Ω)K+1
(14) 1
8[[(v`)]]2k ≤[[vτ]]2(tk)≤ 3 4
¡1 + 3 2στ¢
[[(v`)]]2k+3 4τ1°
°(λ1)12∇v0°
°2
0,Ω. Proof:
Owing to the definitions (5) and (12) of the norms, we have to compare the quantities X`=
Z t`
t`−1
°°λ12(s)∇vτ(s)°°2
0,Ωds and Y` =τ`°°(λ`)12∇v`°°2
0,Ω.
Thanks to the definition of µmax, we have the standard estimate
∀s∈[t`−1, t`], kλ(s)−λ`k∞,Ω ≤τ` µmax, so that, when |τ| ≤α0,
(15) ∀s∈[t`−1, t`], 1
2 ≤
°°
°λ(s) λ`
°°
°∞,Ω ≤ 3 2.
It can also be noted that, thanks to the definition of vτ, and for a.e. xin Ω,
(16)
Z t`
t`−1
|∇vτ(x, s)|2ds = τ` 3
¡|∇v`(x)|2+|∇v`−1(x)|2+∇v`(x)· ∇v`−1(x)¢ .
By combining (15) and (16), we obtain X` ≥ τ`
6
³°°(λ`)12∇v`°
°2+°
°(λ`)12∇v`−1°
°2+ ((λ`)12∇v`,(λ`)12∇v`−1)
´ .
So using the inequality ab≥ −14a2−b2 yields X` ≥ τ`
8
°°(λ`)12∇v`°°2
0,Ω = 1 8Y`, whence the first inequality in (14).
Similarly, by combining (15) and (16) and using the inequality ab≤ 12a2+12b2,we have X` ≤ 3τ`
4
¡°°(λ`)12∇v`°
°2
0,Ω+°
°(λ`)12∇v`−1°
°2
0,Ω
¢.
When `= 1, we keep this inequality without modification. When ` >1, we use an analogue of (15) to obtain
X` ≤ 3τ` 4
°°(λ`)12∇v`°
°2
0,Ω+3τ`−1 4
3 2στ°
°(λ`−1)12∇v`−1°
°2
0,Ω. By summing up the previous lines on `, we derive the second inequality in (14).
The a priori error estimate has been established in [1]: we observe that the family(ek)0≤k≤K, withek =u(tk)−uk,satisfiese0 = 0 and also, by integrating∂tubetweentk−1 andtkand using equation (10) and equation (4) at time t =tk,
(17) ∀v ∈H01(Ω), (ek, v) +τk(λk∇ek,∇v) = (ek−1, v) +τk(²k, v), where the consistency error ²k is given by
(²k, v) = µ1
τk Z tk
tk−1
(∂tu)(s)ds−(∂tu)(tk), v
¶ .
So, applying (13) to this new problem, we derive the estimate: If the solution u is such that
∂t2u belongs to L2(0, T;H−1(Ω)), and for 1≤k ≤K, [[(u(t`)−u`)]]k ≤ 2
3λmin12
³
1≤`≤kmaxτ`´°
°∂t2u°°2
L2(0,tk;H−1(Ω)).
Thanks to Lemma 3.1, this also induces a similar bound for the norm [[u−uτ]](tk).
4 The time and space discrete problem
From now on, we assume that Ω admits a partition without overlap into a finite number of subdomains
(18) Ω =∪Rr=1Ωr and Ωr∩Ωr0 =∅, 1≤r < r0 ≤R,
which satisfy the further conditions:
(i) Each Ωr, 1 ≤ r ≤ R, is a rectangle in dimension d = 2 or a rectangular parallelepiped in dimension d= 3.
(ii) The intersection of two subdomains Ωr and Ωr0,1 ≤ r < r0 ≤ R, if not empty, is either a vertex or a whole edge or a whole face of both Ωr and Ωr0.
We introduce the space PN(Ωr) of restrictions to Ωr of polynomials with d variables and degree≤N with respect to each variable. Relying on this definition, we introduce the discrete spaces, for an integer N ≥2,
YN ={w∈L2(Ω) | wr=w|Ωr ∈PN(Ωr), r = 1, ..., R}, X0N =YN ∩H01(Ω).
Setting ξ0 =−1 and ξN = 1, we introduce the N −1 nodes ξj,1≤j ≤N −1, and the N + 1 weights ρj,0 ≤ j ≤ N, of the Gauss-Lobatto quadrature formula on Λ := [−1,1]. We recall that the following equality holds
(19) ∀ϕN ∈P2N−1(Λ),
Z 1
−1
ϕ(ζ)dζ = XN
j=0
ϕ(ξj)ρj.
We also recall [2, form. (13.20)] the following property, which is useful in what follows
(20) ∀ϕN ∈PN(Λ),kϕNk2L2(Λ) ≤ XN
j=0
ϕ2N(ξj)ρj ≤3kϕNk2L2(Λ).
Denoting by Fr the affine mapping that sends Λd onto Ωr, we introduce the local discrete products, defined on continuous functions u and v onΩr by
(21)
(u, v)rN =
meas(Ωr) 4
PN
i=0
PN
j=0u◦Fr(ξi, ξj)v◦Fr(ξi, ξj)ρiρj if d= 2,
meas(Ωr) 8
PN
i=0
PN
j=0
PN
p=0u◦Fr(ξi, ξj, ξp)v ◦Fr(ξi, ξj, ξp)ρiρjρp if d= 3.
The global product is then defined on continuous functions uand v onΩ by
(22) ((u, v))N =
XR
r=1
(u|Ωr, v|Ωr)rN.
We denote by iN the interpolation operator at the nodes ξj,0 ≤ j ≤ N. We need the lo- cal Lagrange interpolation operators INr : For each function ϕ continuous on Ωr,INrϕ be- longs to PN(Ωr) and is equal to ϕ at all nodes Fr(ξi, ξj),0 ≤ i, j ≤ N in dimension 2 and at Fr(ξi, ξj, ξp),0≤i, j, p≤ N in dimension 3. Finally, for each function ϕcontinuous on Ω,INϕ denotes the function equal to INrϕon each Ωr,1≤r ≤R.
The fully discrete problem is now constructed from (9)-(10) by using the Galerkin method combined with numerical integration. It reads:
Find ¡ ukN¢
0≤k≤K inYN ×(X0N)K, satisfying
(23) u0N =INu0 inΩ,
and such that, for1≤k ≤K,
(24) ∀vN ∈X0N(Ω), akN(ukN, vN) =LkN(vN), where the bilinear forms akN(·,·),1≤k ≤K, are defined by
akN(uN, vN) = ((uN, vN))N +τk((λk∇uN,∇vN))N, and the bilinear forms LkN are defined by
LkN(vN) = ((uk−1N , vN))N +τk((fk, vN))N.
It follows from (20) combined with Cauchy-Schwarz inequalities, that the forms akN and LkN are continuous on X0N ×X0N and X0N respectively, and akN are coercive with norms bounded independently of N.
In all that follows, cstands for a generic constant which can vary from one line to the next one but is always independent ofN. The proof of the next proposition is now standard.
Proposition 4.1 For any data f continuous on Ω×[0, T]and a continuous u0 on Ω,problem (23)-(24) has a unique solution (ukN)0≤k≤K in YN×(X0N)K. Moreover this solution satisfies for a constant c independent of N
(25) [[(u`N)]]k ≤c
³
1 + λmax λmin
´1
2
Ã
kINu0k20,Ω+
³
1 + λmax λmin
´Xk
`=1
τ` λ`min
°°INf`°
°2
0,Ω
!1
2
.
Proof: Taking vN equal to ukN in (24), we have thanks to Cauchy-Schwarz inequality ((ukN, ukN))N +τk((λk∇ukN,∇ukN))N ≤((uk−1N , uk−1N ))N12 ·((ukN, ukN))N12
+τk((INfk,INfk))N21 ·((ukN, ukN))N12, (26)
using (20), Poincaré-Friedrichs inequality and the inequality ab≤ 2ε1a2+ε2b2,∀ε >0,we obtain 1
2
°°ukN°
°2
0,Ω+τk((λk∇uNk,∇ukN))N ≤ 1 2
°°uk−1N °
°2
0,Ω+ 1 2ε
cτk λkmin
°°INfk°
°2
0,Ω+ ε 2τk°
°(λk)12∇ukN°
°2
0,Ω.
Summing up on k,and using (23), to obtain 1
2
°°ukN°
°2
0,Ω+ Xk
`=1
τ`((λ`∇u`N,∇u`N))N ≤ 1 2
°°INu0°
°2
0,Ω+ c 2ε
Xk
`=1
τ` λ`min
°°INf`°
°2
0,Ω
+ε 2
Xk
`=1
τ`
°°(λ`)12∇u`N°
°2
0,Ω. (27)
On the other hand and thanks to (20), we have
°°(λ`)12∇u`N°°2
0,Ω ≤λmax((∇u`N,∇u`N))N ≤ λmax
λmin((λ`∇u`N,∇u`N))N, 1≤` ≤k, so, (27) implies that
1 2
°°ukN°°2
0,Ω+ Xk
`=1
τ`°°(λ`)12∇u`N°°2
0,Ω ≤ 1
2(1 + λmax
λmin)°°INu0°°2
0,Ω+ c
2ε(1 + λmax
λmin) Xk
`=1
τ`
λ`min
°°INf`°°2
0,Ω
+ ε
2(1 + λmax λmin)
Xk
`=1
τ`°
°(λ`)12∇u`N°
°2
0,Ω
Finally, we choose ε= λ λmin
max+λmin to obtain (25).
5 Error estimate
We now wish to establish the error estimate between the solution(uk)0≤k≤K of problem (9)-(10) and the solution (ukN)0≤k≤K of problem (23)-(24).
LetΠ1,0N denote the orthogonal projection operator fromH01(Ω) ontoX0N for the scalar product associated with the norm | · |1,Ω. For 0≤` ≤k,Π1,0N−1u` will be denoted by v`N.
Proposition 5.1 Assume thatf andu0 are continuous onΩ×[0, T]and Ωrespectively. Then the following estimate holds for the error between the solution (uk)0≤k≤K of problem (9)-(10) and the solution (ukN)0≤k≤K of problem (23)-(24)
[[(u`−u`N)]]k≤c µ
[[(u`−vN` )]]k+³
1 + λmax λmin
´1
2³
ku0−vN0k0,Ω+ku0 − INu0k0,Ω
+
³
1 + λmax λmin
´1
2 Xk
`=1
³ τ` λ`min
´1
2¡
EN,`a,1 +EN,`a,2 +EN,`f ¢´¶ , where the quantities EN,`a,1, EN,`a,2 and EN,`f are defined by
EN,`a,1 = sup
vN∈X0N
(uk−uk−1
τk , vN)−((vkN −vNk−1
τk , vN))N
|vN|1,Ω ,
EN,`a,2 = sup
vN∈X0N
(λk∇uk,∇vN)−((λk∇vNk,∇vN))N
|vN|1,Ω
, EN,`f = sup
vN∈X0N
(f, vN)−((f, vN))N
|vN|1,Ω
.
Proof: We have
(28) [[(u`−u`N)]]k≤[[(u`−vN` )]]k+ [[(u`N −vN` )]]k,
so we have to estimate the term [[(u`N −vN` )]]k. It follows from (10) and (24) that
(29) ((ukN −vkN, vN))N +τk((λk∇(ukN −vNk),∇vN))N = ((uk−1N −vk−1N , vN))N +τkMNk(vN),
where MNk is the linear form onX0N defined by MNk(vN) = (uk−uk−1
τk , vN)−((vNk −vk−1N
τk , vN))N + (λk∇uk,∇vN)−((λk∇vNk,∇vN))N
+ ((fk, vN))N −(fk, vN).
Due to the Riesz’s theorem, there exists a unique polynomial FNk inX0N such that (30) ∀vN ∈X0N, MNk(vN) = ((FNk, vN))N.
Thus the family (ukN −vNk)0≤k≤K is a solution of the discrete problem (23)-(24) with INu0 − vN0 instead of INu0 and FNk instead of fk. Consequently, following the same arguments as in Proposition 4.1, we get
[[(u`N −v`N)]]k ≤c
³
1 + λmax λmin
´1
2³
kINu0−vN0 k20,Ω+
³
1 + λmax λmin
´Xk
`=1
τ` λ`min(ϕ`)2
´1
2,
where ϕ` = sup
vN∈X0N
((FN`, vN))N
|vN|1,Ω
.We conclude the proof thanks to (30).
In order to estimate the term EN,`a,1, we denote by w` the quantity u`−uτ`−1
` and we observe that vN` −vN`−1
τ` = Π1,0N−1w`, so as a consequence of the exactness property (19), the terms (Π1,0N−1w`, vN¢
0,Ω and ((Π1,0N−1w`, vN¢
)N coincide and thus,
(31) EN,`a,1 ≤°
°w`−Π1,0N−1w`°
°0,Ω.
Now, in order to evaluate the term EN,`a,2, we introduce the orthogonal projection operator Πr,1N fromH1(Ωr)ontoPN(Ωr)for the scalar product associated with the normk · k1,Ωr. IfN˜ stands for the integer part of N−12 ,so as a consequence of the exactness property (19),we have for any vN ∈X0N
(λ`∇u`,∇vN)−((λ`∇vN` ,∇vN))N = (λ`(∇u`− ∇vN` ),∇vN) +
XR
r=1
³ Z
Ωr
(λ`∇vN` − INr˜λ`∇(Πr,1N˜ u`))(x)· ∇vN(x)dx +(INr˜λ`∇(Πr,1N˜ u`)−λ`∇vN` ,∇vN)rN
´ . (32)
Due to Cauchy-Schwarz inequality and by using the notationλ`max = sup
x∈Ω
λ(x, t`), the first term in the right hand side of (32) can be estimated as
(λ`(∇u`− ∇vN` ),∇vN)≤λmax` |u`−vN` |1,Ω|vN|1,Ω. Similar arguments also lead to
XR
r=1
Z
Ωr
(λ`∇vN` − INr˜λ`∇(Πr,1N˜ u`))(x)· ∇vN(x)dx ≤
³ λ`max¡
|u`−v`N|1,Ω+ XR
r=1
¯¯u`−Πr,1N˜ u`¯
¯1,Ωr
¢
+ XR
r=1
°°λ`− INr˜λ`°°
∞,Ωr
°°Πr,1N˜ u`°°
1,Ωr
´
|vN|1,Ω.
Now, thanks to Cauchy-Schwarz inequality and (20), the last term in the right hand side of (32) can be estimated as follows
XR
r=1
(INr˜λ`∇(Πr,1N˜ u`)−λ`∇v`N,∇vN)rN ≤ c³ λ`max¡
|u`−vN` |1,Ω+ XR
r=1
¯¯u`−Πr,1N˜ u`¯
¯1,Ωr
¢
+ XR
r=1
°°λ`− INr˜λ`°
°∞,Ωr
°°Πr,1N˜ u`°
°1,Ωr
´
|vN|1,Ω.
so that
EN,`a,2 ≤ c µ
λ`max
³
|u`−v`N|1,Ω+ XR
r=1
¯¯u`−Πr,1N˜ u`¯
¯1,Ωr
´
+ XR
r=1
°°λ`− INr˜λ`°
°∞,Ωrku`k1,Ωr
¶ . (33)
Finally, in order to estimate the term EN,`f , we introduce the orthogonal projection operator ΠrN from L2(Ωr) intoPN(Ωr). Indeed, using (19) leads to, for any vN inX0N,
Z
Ωr
f`(x)·vN(x)dx−(f`, vN)rN = Z
Ωr
(f`(x)−ΠrN−1f`)(x)·vN(x)dx−(INrf`−ΠrN−1f`, vN)rN, so that, owing to (20),
(34) EN,`f ≤c
XR
r=1
³°°f`−ΠrN−1f`°
°0,Ωr +°
°f`− INrf`°
°0,Ωr
´ .
Now, to make complete the evaluation of EN,`a,1, EN,`a,2 and EN,`f , we need the following results.
First, we recall from [2, Thm. 7.1, Thm. 7.3 and Thm. 14.2] the approximation properties of the operators ΠrN,Πr,1N and INr,1≤r≤R: for any function ϕin Hs(Ωr), s ≥0
°°ϕ−ΠrNϕ°
°0,Ωr ≤cN−skϕks,Ωr,
for any function ϕin Hs(Ωr), s≥1
°°ϕ−Πr,1N ϕ°
°1,Ωr ≤cN1−skϕks,Ωr,
and for any function ϕin Hs(Ωr), s > d2
°°ϕ− INrϕ°
°0,Ωr ≤cN−skϕks,Ωr.
The following result is derived from [3, Lem. VI.2.5] thanks to an interpolation argument, for any real number s ≥ 1, and any function ϕ in H01(Ω) such that each ϕ|Ωr,1 ≤ r ≤ R, belongs toHs(Ωr)
(35) °°ϕ−Π1,0N ϕ°°
1,Ω ≤cN1−s XR
r=1
kϕks,Ωr.