The approximation in the sense of Γ-convergence of nonisotropic Griffith-type functionals, with p−growth (p>1) in the symmetrized gradient, by means of a suitable sequence of non-local convolution type functionals defined on Sobolev spaces, is analysed.
Citation: Fernando Farroni, Giovanni Scilla, Francesco Solombrino. On some non-local approximation of nonisotropic Griffith-type functionals[J]. Mathematics in Engineering, 2022, 4(4): 1-22. doi: 10.3934/mine.2022031
[1] | Manuel Friedrich . Griffith energies as small strain limit of nonlinear models for nonsimple brittle materials. Mathematics in Engineering, 2020, 2(1): 75-100. doi: 10.3934/mine.2020005 |
[2] | Caterina Ida Zeppieri . Homogenisation of high-contrast brittle materials. Mathematics in Engineering, 2020, 2(1): 174-202. doi: 10.3934/mine.2020009 |
[3] | Luca Formaggia, Alessio Fumagalli, Anna Scotti . A multi-layer reactive transport model for fractured porous media. Mathematics in Engineering, 2022, 4(1): 1-32. doi: 10.3934/mine.2022008 |
[4] | Alessandra De Luca, Veronica Felli . Unique continuation from the edge of a crack. Mathematics in Engineering, 2021, 3(3): 1-40. doi: 10.3934/mine.2021023 |
[5] | Alberto Farina . Some results about semilinear elliptic problems on half-spaces. Mathematics in Engineering, 2020, 2(4): 709-721. doi: 10.3934/mine.2020033 |
[6] | Daniela De Silva, Ovidiu Savin . Uniform density estimates and $ \Gamma $-convergence for the Alt-Phillips functional of negative powers. Mathematics in Engineering, 2023, 5(5): 1-27. doi: 10.3934/mine.2023086 |
[7] | Stefano Almi, Giuliano Lazzaroni, Ilaria Lucardesi . Crack growth by vanishing viscosity in planar elasticity. Mathematics in Engineering, 2020, 2(1): 141-173. doi: 10.3934/mine.2020008 |
[8] | Matteo Novaga, Marco Pozzetta . Connected surfaces with boundary minimizing the Willmore energy. Mathematics in Engineering, 2020, 2(3): 527-556. doi: 10.3934/mine.2020024 |
[9] | Petteri Harjulehto, Peter Hästö, Jonne Juusti . Bloch estimates in non-doubling generalized Orlicz spaces. Mathematics in Engineering, 2023, 5(3): 1-21. doi: 10.3934/mine.2023052 |
[10] | Paola F. Antonietti, Chiara Facciolà, Marco Verani . Unified analysis of discontinuous Galerkin approximations of flows in fractured porous media on polygonal and polyhedral grids. Mathematics in Engineering, 2020, 2(2): 340-385. doi: 10.3934/mine.2020017 |
The approximation in the sense of Γ-convergence of nonisotropic Griffith-type functionals, with p−growth (p>1) in the symmetrized gradient, by means of a suitable sequence of non-local convolution type functionals defined on Sobolev spaces, is analysed.
The scope of this paper is to provide a generalization of recent results, obtained in [23], concerning the approximation of brittle fracture energies for linearly elastic materials, by means of nonlocal functionals defined on Sobolev spaces, which are easier to handle also from a computational point of view.
In [23] an approach originally devised by Braides and Dal Maso [6] for the approximation of the Mumford-Shah functional has been generalized to the linearly elastic context. Namely, it was shown that, for a given bounded increasing function f:R+→R+ the energies
Fε(u):=1ε∫Ωf(−∫Bε(x)∩ΩW(Eu(y))dy)dx |
Γ-converge to the functional
α∫ΩW(Eu(x))dx+2βHn−1(Ju), |
with α=f′(0) and β=limt→+∞f(t), in the L1(Ω)-topology. Above, W(Eu(y)) is a convex elastic energy depending on the linearized strain Eu, given by the symmetrized gradient of a vector-valued displacement u, whose jump set Ju represents the cracked part of a material. The energy space of the limit functional is the one of generalized functions with bounded deformation, introduced in [17].
It is noteworthy that the above result allowed one for a general (convex) bulk energy W having p-growth for p>1. The proof strategy must then avoid, at least when estimating the bulk part, any slicing procedure. This latter is instead successful in the special case* W(ξ)=|ξ|p, considered for instance in [22]. There, non-local convolution-type energies of the form
* We remark that this particular case is however not the most relevant one from a mechanical point of view, as even for an isotropic material additional terms in the bulk energy are expected to appear.
1ε∫Ωf(ε∫Rn|Eu(y)|pρε(x−y)dy)dx | (1.1) |
are considered, where ρ is a convolution kernel whose support is a convex bounded domain and ρε(z) is the usual sequence of convolution kernels ρ(z/ε)/εn. The Γ-limit of (1.1) with respect to the L1 convergence is given by
∫Ω|Eu(x)|pdx+∫Juϕρ(ν)dHn−1, |
where the anisotropic surface density ϕρ depends on the geometry and on the size of suppρ. A similar effort of generalizing the results of [6] to Mumford-Shah type energies with non-isotropic surface part has been previously performed in [13].
In this paper, we extend the focus of [22,23] by showing that general Griffith-type functionals of the form
α∫ΩW(Eu(x))dx+2β∫Juϕ(ν)dHn−1, | (1.2) |
where ϕ is any norm on Rn, can be obtained as variational limit of non-local convolution-type functionals
1ε∫Ωf(ε∫RnW(Eu(y))ρε(x−y)dy)dx. |
Above, f is again a bounded nondecreasing function with α=f′(0) and β=limt→+∞f(t), and the unscaled kernel ρ has the bounded convex symmetric domain ¯S:={ξ∈Rn:ϕ(ξ)≤1} as its support. This is the analogue, in the linear elastic setting, of the results in [13].
The proof strategy we devise is based on a localization method and involves nontrivial adaptions to the method used in [23], in particular when estimating the bulk term in the Γ-liminf inequality (Proposition 4.2). There, we have to impose (and this is the only point in the paper) an additional restriction on the convolution kernel ρ, namely of being nonincreasing with respect to the given norm ϕ (see Assumption (N2) below). This is namely needed in order to be able to estimate from below the size of the nonlocal approximations of the bulk term in an anisotropic tubular neighborhood of the set where they exceed the threshold βα, which heuristically corresponds to the breaking of the elastic bonds. With this, a set K′ε with small area and bounded perimeter, where the fracture energy concentrates can be explicitly constructed. This yields an estimate of the Γ-liminf which has an optimal constant in front of the bulk term, although being non-optimal for the surface energy.
Another non-optimal estimate for the Γ-liminf, but with an optimal constant for the surface energy can be instead obtained by a slicing procedure, involving a comparison argument and the convexity of the open set S (Proposition 4.3). As bulk and surface energy in (1.2) are mutually singular as measures, a localization procedure entails then the Γ-liminf inequality (Proposition 4.6). Finally, the Γ-limsup inequality (Proposition 5.1) can be obtained by a direct construction for a regular class of competitors having a "nice" jump set, and which are dense in energy. Notice indeed that such an approximation (see Theorem 2.3 for a precise statement) is possibile also with respect to an anisotropic norm ϕ, combining the recent results in [8] with the ones in [12].
As a final remark, it would be desirable to get rid of the structural assumption (N2) on the convolution kernels, which is used only in Proposition 4.2. It is our opinion that this is going to require quite a delicate abstract analysis of the Γ-limit of nonlocal functionals which approximate free-discontinuity problems in GSBD, possibly including also finite-difference models which are well suited to numerical approximations (see [15] for a recent discrete finite-difference approximation of some Griffith-type functionals in GSBD). A similar analysis for the SBV setting has been performed in [11], where integral representation formulas for the limit energy have been provided. Furthermore, nontrivial sufficient conditions have been given under which the bulk part of the energy can be recovered by only considering weakly compact sequences in Sobolev spaces. We plan to defer this abstract analysis to a forthcoming contribution. For the asymptotic analysis via Γ-convergence of local free-discontinuity functionals in linear elasticity and the related issues, we refer the reader to the very recent papers [7,14,20].
Outline of the paper: The paper is structured as follows. In Section 2 we fix the basic notation and results on the function spaces we will deal with (Section 2.2), together with some technical lemmas (Section 2.3) which will be useful throughout the paper. In Section 3 we list the main assumptions, introduce our model (Eq (3.4)), and state the main results of the paper, provided {in} Theorem 3.1 and Theorem 3.2. Section 4 is devoted to the proof of the compactness statements in the main Theorems (Proposition 4.1), and to the Γ-liminf inequality, which is proved in Section 4.3 combining the estimates in Sections 4.1 and 4.2. The proof of the upper bound is given in Section 5.
The symbol |⋅| denotes the Euclidean norm in any dimension, while ⟨⋅,⋅⟩ stands for the scalar product in Rn. We will always denote by Ω an open, bounded subset of Rn {with Lipschitz boundary}, and by Sn−1 the (n−1)-dimensional unit sphere. The Lebesgue measure in Rn and the s-dimensional Hausdorff measure are written as Ln and Hs, respectively. A(Ω) stands for the family of the open subsets of Ω.
Let S be a bounded, open, convex and symmetrical set, i.e., S=−S. For η>0, we denote by ηS the η-dilation of S and we will often use the shorthand {S(x,η)} in place of x+ηS. We consider |⋅|S the norm induced by S, defined as
|x|S:=inf{η>0:x∈ηS}, | (2.1) |
whose unit ball {|x|S<1} coincides with S, and, correspondingly, we introduce the distance to a closed bounded set K⊂Rn; namely,
distS(x,K):=miny∈K|x−y|S,x∈Rn. | (2.2) |
In this section we recall some basic definitions and results on generalized functions with bounded deformation, as introduced in [17]. Throughout the paper we will use standard notations for the spaces (G)SBV and (G)SBD, referring the reader to [2] and [1,3,24], respectively, for a detailed treatment on the topics.
Let ξ∈Rn∖{0} and Πξ={y∈Rn:⟨ξ,y⟩=0}. If Ω⊂Rn and y∈Πξ we set Ωξ,y:={t∈R:y+tξ∈Ω} and Ωξ:={y∈Πξ:Ωξ,y≠∅}. Given u:Ω→Rn, n≥2, we define uξ,y:Ωξ,y→R by
uξ,y(t):=⟨u(y+tξ),ξ⟩, | (2.3) |
while if v:Ω→R, the symbol vξ,y will denote the restriction of v to the set Ωξ,y; namely,
vξ,y(t):=v(y+tξ). | (2.4) |
Let ξ∈Sn−1. For any x∈Rn we denote by xξ and yξ the projections onto the subspaces Ξ:={tξ:t∈R} and Πξ, respectively. For σ,r>0 and x∈Rn we define the cylinders
Cξσ,r(0):={x∈Rn:|xξ|<σ,|yξ|<r},Cξσ,r(x):=x+Cξσ,r(0). |
Note that Cξσ,r(x)=(xξ−σ,xξ+σ)×Bn−1r(yξ), where Bn−1 denotes a ball in the (n−1)-dimensional space Πξ.
Definition 2.1. An Ln-measurable function u:Ω→Rn belongs to GBD(Ω) if there exists a positive bounded Radon measure λu such that, for all τ∈C1(Rn) with −12≤τ≤12 and 0≤τ′≤1, and all ξ∈Sn−1, the distributional derivative Dξ(τ(⟨u,ξ⟩)) is a bounded Radon measure on Ω whose total variation satisfies
|Dξ(τ(⟨u,ξ⟩))|(B)≤λu(B) |
for every Borel subset B of Ω.
If u∈GBD(Ω) and ξ∈Rn∖{0} then, in view of [17,Theorem 9.1,Theorem 8.1], the following properties hold:
(a) ˙uξ,y(t)=⟨Eu(y+tξ)ξ,ξ⟩ for a.e. t∈Ωξ,y;
(b) Juξ,y=(Jξu)ξ,y for Hn−1-a.e. y∈Πξ, where
Jξu:={x∈Ju:⟨u+(x)−u−(x),ξ⟩≠0}. |
Definition 2.2. A function u∈GBD(Ω) belongs to the subset GSBD(Ω) of special functions of bounded deformation if, in addition, for every ξ∈Sn−1 and Hn−1-a.e. y∈Πξ, it holds that uξ,y∈SBVloc(Ωξ,y).
The inclusions BD(Ω)⊂GBD(Ω) and SBD(Ω)⊂GSBD(Ω) hold (see [17,Remark 4.5]). Although they are, in general, strict, relevant properties of BD functions are retained also in this weak setting. In particular, GBD-functions have an approximate symmetric differential Eu(x) at Ln-a.e. x∈Ω. Furthermore the jump set Ju of a GBD-function is Hn−1-rectifiable (this is proven in [17,Theorem 6.2 and Theorem 9.1], but it has been recently shown that this property is actually a general one for measurable functions [18]).
Let p>1. The space GSBDp(Ω) is defined as
GSBDp(Ω):={u∈GSBD(Ω):Eu∈Lp(Ω;Rn×nsym),Hn−1(Ju)<+∞}. |
Every function in GSBDp(Ω) can be approximated with the so-called "piecewise smooth" SBV-functions, denoted W(Ω;Rn), characterized by the three properties
{u∈SBV(Ω;Rn)∩Wm,∞(Ω∖Ju;Rn) for every m∈N,Hn−1(ˉJu∖Ju)=0,ˉJu is the intersection of Ω with a finite union of (n−1)-dimensional simplexes . | (2.5) |
This is stated by the following result, which combines [8,Theorem 1.1] with [12,Theorem 3.9].
Theorem 2.3. Let ϕ be a norm on Rn. Let Ω⊂Rn be a bounded open Lipschitz set, and let u∈GSBDp(Ω;Rn). Then there exists a sequence (uj) such that uj∈W(Ω;Rn) and
uj→uin measure on Ω, | (2.6) |
Euj→Eu in Lp(Ω;Rn×nsym), | (2.7) |
∫Jujϕ(νuj)Hn−1→∫Juϕ(νu)Hn−1. | (2.8) |
Moreover, if ∫Ωψ(|u|)dx is finite for ψ:[0,+∞)→[0,+∞) continuous, increasing, with
ψ(0)=0,ψ(s+t)≤C(ψ(s)+ψ(t)),ψ(s)≤C(1+sp),lims→+∞ψ(s)=+∞ |
then
limj→+∞∫Ωψ(|uj−u|)dx=0. | (2.9) |
As observed in [9,Remark 4.3], we may even approximate through functions u such that, besides (2.5), have a closed jump set strictly contained in Ω made of pairwise disjoint (n−1)-dimensional simplexes, with Ju∩Πi∩Πl=∅ for any two different hyperplanes Πi, Πl.
We recall the following general GSBDp compactness result from [10], which generalizes [17,Theorem 11.3].
Theorem 2.4 (GSBDp compactness). Let Ω⊂Rn be an open, bounded set, and let (uj)j⊂GSBDp(Ω) be a sequence satisfying
supj∈N(‖Euj‖Lp(Ω)+Hn−1(Juj))<+∞. |
Then there exists a subsequence, still denoted by (uj), such that the set A∞:={x∈Ω:|uj(x)|→+∞} has finite perimeter, and there exists u∈GSBDp(Ω) such that
(i) uj→u in measure on Ω∖A∞,(ii) Euj⇀Eu in Lp(Ω∖A∞;Rn×nsym),(iii) lim infj→∞Hn−1(Juj)≥Hn−1(Ju∪(∂∗A∞∩Ω)), | (2.10) |
where ∂∗ denotes the essential boundary of a set with finite perimeter.
Remark 2.5. If in the statement above one additionally assumes that
supj∈N∫Ωψ(|uj|)dx<+∞ |
for a positive, continuous and increasing function ψ with lims→+∞ψ(s)=+∞, then A∞=∅, so that |u| is finite a.e., and (i) holds on Ω. Moreover, if ψ is superlinear at infinity, that is
lims→+∞ψ(s)s=+∞, |
by the Vitali dominated convergence theorem one gets that u∈L1(Ω) and (i) holds with respect to the L1-convergence in Ω.
We recall here (without adding the standard proofs) some properties of integral convolutions in the setting of Sobolev spaces.
Proposition 2.6. Let w∈W1,p(Ω;Rn) and ρ∈L∞(Rn) be a convolution kernel, with suppρ⊂¯S for some bounded, open and convex set S⊂Ω. Set ρθ(x):=1θnρ(xθ). Then the following holds:
(i) let Ω′⊂⊂Ω and 0≤θ≤distS(Ω′,∂Ω). The convolution
φθ(x):=∫Ωw(y)ρθ(y−x)dy |
belongs to W1,p(Ω′;Rn). Moreover, it holds that
∇φθ(x)=∫Ω∇w(y)ρθ(y−x)dya.e. on Ω′. | (2.11) |
(ii) assume that wε→w in L1(Ω;Rn) and let θε be any sequence with θε→0 when ε→0. Then the sequence
ˆwε(x):=∫Ωwε(y)ρθε(y−x)dy |
satisfies ˆwε→cw in L1(Ω;Rn), where c=∫Rnρ(x)dx.
We also recall the following convergence property of one-dimensional sections of averaged functions (see, e.g., [23,Lemma 2.7(ii)]).
Lemma 2.7. Assume that wε→w in L1(Ω;Rn) and let ηε be any sequence with ηε→0 when ε→0. Then for all ξ∈Sn−1 and a.e. y∈Πξ, the sequence
ˆwξ,yε(t):=−∫Bn−1ηε(y)wε(z+tξ)dz |
satisfies ˆwξ,yε→wξ,y in L1(Ωξ,y;Rn), where wξ,y(t):=w(y+tξ).
We will also make use of the following localization result, dealing with the supremum of a family of measures (see, e.g., [4,Proposition 1.16]).
Lemma 2.8. Let μ:A(Ω)⟶[0,+∞) be a superadditive function on disjoint open sets, let λ be a positive measure on Ω and let φh:Ω⟶[0,+∞] be a countable family of Borel functions such that μ(A)≥∫Aφhdλ for every A∈A(Ω). Then, setting φ:=suph∈Nφh, it holds that
μ(A)≥∫Aφdλ |
for every A∈A(Ω).
Lower semicontinuous increasing functions can be approximated from below with truncated affine functions. We refer the reader to [23,Lemma 2.10] for a proof of the following result.
Lemma 2.9. Consider a lower semicontinuous increasing function f:[0,+∞)→[0,+∞) such that there exist α,β>0 with
limt→0+f(t)t=α,limt→+∞f(t)=β. |
Then there exist two positive sequences (ai)i∈N, (bi)i∈N with
supiai=α,supibi=β |
and min{ait,bi}≤f(t) for all i∈N and t∈R.
Let (X,d) be a metric space. We recall here the definition of Γ-convergence for families of functionals Fε:X→[−∞,+∞] depending on a real parameter ε (see, e.g., [5,16]).
For all u∈X, we define the lower Γ-limit of (Fε) as ε→0+ by
F′(u):=inf{liminfj→+∞Fεj(uj):εj→0+,uj→u}, | (2.12) |
and the upper Γ-limit of (Fε) as ε→0+ by
F"(u):=inf{limsupj→+∞Fεj(uj):εj→0+,uj→u}. | (2.13) |
We then say that (Fε) Γ-converges to F:X→[−∞,+∞] as ε→0+ iff
F(u)=F′(u)=F"(u),for all u∈X. |
The following one-dimensional Γ-convergence result will be useful in the proof of the lower bound for the surface term. In the statement below, functions in L1(I) with I⊂R are extended by 0 outside I, so that the functionals Hε are well-defined (actually, the result is not affected by the considered extension).
Theorem 2.10. Let p>1, let I⊂R be a bounded interval and consider a lower semicontinuous, increasing function f:[0,+∞)→[0,+∞) complying with
limt→0+f(t)t=α,limt→+∞f(t)=β |
for some α,β>0. Let Hε:L1(I)→[0,+∞] be defined by
Hε(u):=1ε∫If(12∫x+εx−ε|u′(y)|pdy)dx, |
where it is understood that
f(12∫x+εx−ε|u′(y)|pdy)=β |
if u∉W1,p(x−ε,x+ε). Then the functionals (Hε) Γ-converge as ε→0+ to the functional
H(u):={α∫I|u′|pdt+2β#(Ju),if u∈SBV(I),+∞,otherwise |
in L1(I).
Proof. The proof can be found, e.g., in [4,Theorem 3.30].
In this section we list our assumptions and introduce the main results of the paper. Let Ω⊂Rn be an open set with Lipschitz boundary, let 1<p<+∞ and f:[0,+∞)→[0,+∞) a lower semicontinuous, increasing function satisfying
limt→0+f(t)t=α>0,limt→+∞f(t)=β>0. | (3.1) |
Let ρ∈L∞(Rn;[0,+∞)) be a convolution kernel. The minimal assumption is that
(N1) ρ is Riemann integrable with ‖ρ‖1=1 and S=Sρ:={x∈Rn:ρ(x)≠0} is a bounded, open, convex and symmetrical set.
As every Riemann integrable function is continuous at almost every point, we may also suppose, up to a modification on a null set, that ρ is lower semicontinuous. Also notice that, by a simple scaling argument, one can always consider the case of kernels with unit mass, up to modifying the constant α in (3.1).
A sequence (ρε)ε>0 of convolution nuclei is then obtained by setting, for every x∈Rn and ε>0,
ρε(x):=1εnρ(xε). |
For every v∈Rn we define
ϕρ(v):=2sup{|⟨y,v⟩|:y∈S}. | (3.2) |
Under the previous assumptions on S, the function ϕρ turns out to be a norm on Rn.
To obtain our main result, we will have to couple (N1) with the additional assumption that the convolution kernel is a non-increasing function with respect to the norm |⋅|S, that is
(N2) |x|S≥|y|S⟹ρ(x)≤ρ(y) for all x, y∈Rn.
Equivalently, we require that it exists a non-increasing function ϱ:R+→R+ such that ρ(x)=ϱ(|x|S). Notice that, in the case S=B1, every non-increasing radial function ρ complies with (N2).
Let W:Rn×n→R be a convex positive function on the subspace Mn×nsym of symmetric matrices, such that
W(0)=0,c|M|p≤W(M)≤C(1+|M|p). | (3.3) |
For every ε>0 we consider the functional Fε:L1(Ω;Rn)→[0,+∞] defined as
Fε(u)={1ε∫Ωf(ε∫ΩW(Eu(y))ρε(x−y)dy)dx, if u∈W1,p(Ω;Rn),+∞, otherwise on L1(Ω;Rn). | (3.4) |
We will deal with a localized version of the energies (3.4). Namely, for every A∈A(Ω), we will denote by Fε(u,A) the same functional as in (3.4) with the set A in place of Ω. When A=Ω, we simply write Fε(u) in place of Fε(u,Ω).
The following theorem is the first main result of this paper. We notice that the additional assumption (N2) on the structure of the convolution kernel is required in (ii) below only to obtain the optimal lower bound for the bulk term of the energy.
Theorem 3.1. Let ρ∈L∞(Rn;[0,+∞)) be a convolution kernel as in (N1), and let Fε be defined as in (3.4). Under assumptions (3.1) and (3.3), it holds that
(i) there exists a constant c0 independent of ε such that, for all (uε)⊂Lp(Ω;Rn) satisfying Fε(uε)≤C for every ε>0, one can find a sequence ¯uε∈GSBVp(Ω;Rn) with
¯uε−uε→0 in measure on ΩFε(uε)≥c0(∫ΩW(E¯uε)dx+2Hn−1(J¯uε∩Ω)). |
(ii) If, in addition, ρ complies with (N2), then the functionals (Fε) Γ-converge, as ε→0, to the functional
F(u)={α∫ΩW(Eu)dx+β∫Juϕρ(ν)dHn−1,if u∈GSBDp(Ω)∩L1(Ω;Rn),+∞,otherwise on L1(Ω;Rn), | (3.5) |
with respect to the L1 convergence in Ω.
The L1-convergence on the whole Ω can be enforced with the addition of a lower order fidelity term, as we have discussed in Remark 2.5. This motivates the statement below, {where} we consider a continuous increasing function ψ:[0,+∞)→[0,+∞) such that
ψ(0)=0,ψ(s+t)≤C(ψ(s)+ψ(t)),ψ(s)≤C(1+sp),lims→+∞ψ(s)s=+∞ | (3.6) |
and we set for every A∈A(Ω)
Gε(u,A)={Fε(u,A)+∫Aψ(|u|)dx, if u∈W1,p(A;Rn),+∞, otherwise on L1(A;Rn). | (3.7) |
As before, we simply write Gε(u) in place of Gε(u,Ω).
Then we have the following result.
Theorem 3.2. Under assumptions (3.1), (N1), (3.3), and (3.6) it holds that
(i) If (uε)⊂Lp(Ω;Rn) is such that Gε(uε)≤C for every ε>0, then (uε) is compact in L1(Ω;Rn).
(ii) If, in addition, (N2) holds, the {functionals} (Gε) Γ-converge, as ε→0, to the functional
G(u)={F(u)+∫Ωψ(|u|)dx,if u∈GSBDp(Ω)∩L1(Ω;Rn),+∞,otherwise on L1(Ω;Rn), |
with respect to the L1 convergence in Ω.
With the following proposition, we prove the compactness statements in Theorem 3.1(i), and Theorem 3.2(i), respectively. These results can be easily inferred by a comparison with non-local integral energies whose densities are averages of the gradient on balls with small radii, for which a compactness result has been provided in [23,Proposition 4.1]. In order to do that, we will only require assumption (N1) on the convolution kernel ρ.
Proposition 4.1. Let A∈A(Ω), and let Fε, Gε be defined as in (3.4), and (3.7), respectively, where ρ∈L∞(Rn;[0,+∞)) satisfies (N1). Then:
(i) Assume (3.1), (3.3). If (uε)⊂Lp(Ω;Rn) is such that Fε(uε,A)≤C for every ε>0, one can find a sequence ¯uε∈GSBVp(A;Rn) with
¯uε−uε→0 in measure on AFε(uε,A)≥c0(∫AW(E¯uε)dx+2Hn−1(J¯uε∩A)) |
for some c0>0.
(ii) Assume (3.1), (3.3), and (3.6). If (uε)⊂Lp(Ω;Rn) is such that Gε(uε,A)≤C for every ε>0, then (uε) is compact in L1(A;Rn).
Proof. Let η∈(0,1) be fixed such that Bη(0)⊂⊂S, and denote by mη the minimum of ρ on ¯Bη, which is strictly positive as we are assuming that ρ>0 on S. Setting ˜f(t):=f(mηωnηnt) and for any ε>0, we consider the energies
˜Fε(u)={1ε∫Ω˜f(ε−∫Bηε(x)∩ΩW(Eu(y))dy)dx, if u∈W1,p(Ω;Rn),+∞, otherwise on L1(Ω;Rn). | (4.1) |
Since Bη(0)⊆S and ρ≥mη on Bη(0), a simple computation shows that
˜Fε(u,A)≤Fε(u,A) | (4.2) |
for every u∈W1,p(Ω;Rn) and A⊆Ω open set. By virtue of (4.2), to obtain (i) it will suffice to apply the argument of [23,Proposition 4.1] to the sequence ˜Fε in (4.1). We then omit the details.
We now come to (ii). If additionally Gε(uε,A)≤C, following the argument for [23,Proposition 4.1(ii)], it can be shown that the sequence (ˉuε) constructed in (i) complies with
∫Aψ(|¯uε(x)|)dx+∫A|E¯uε(x)|pdx+Hn−1(J¯uε∩A)≤C<+∞ |
for all ε. Therefore, in view of the growth assumption (3.6) on ψ, Theorem 2.4 and Remark 2.5 apply, and this provides the compactness of the sequence (¯uε) in L1(A;Rn). Then, since ¯uε−uε→0 in measure on A, with the Vitali dominated convergence Theorem we infer that (uε) is compact in L1(A;Rn) as well. This concludes the proof of (ii).
Now, we turn to provide a first estimate of the Γ-liminf of the functionals Fε. This estimate is optimal, up to a small error, only for the bulk part of the energy, and this is the only very point where we need to require the additional assumptions (N2) on the convolution kernels (see Section 4.1). The proof of an optimal estimate for the surface term, instead, will be derived separately by means of a slicing argument (see Proposition 4.3 below) for more general kernels complying only with (N1) providing the comparison estimate (4.2). As the two parts of the energy are mutually singular, the localization method of Lemma 2.8 will eventually allow us to get the Γ-liminf inequality.
We begin by giving the announced estimate for the bulk term.
Proposition 4.2. Let A∈A(Ω) with A⊂⊂Ω, and consider a sequence uε∈W1,p(Ω;Rn) converging to u in L1(Ω;Rn). Assume (3.1) and (3.3), let η∈(0,1) be fixed and let ρ comply with (N1)–(N2). Suppose that
supε>0Fε(uε,A)≤C. | (4.3) |
Then, for every fixed 0<δ<1, there exist a constant Mδ,η only depending on f, δ and η, a constant ση depending on ρ,η such that ση→0 as η→0, and a sequence of functions (vδ,ηε)⊂GSBVp(A;Rn) such that
(i) α(1−ση)2(1−δ)2n+1∫AW(Evδ,ηε(x))dx≤Fε(uε,A);
(ii) Hn−1(Jvδ,ηε)≤Mδ,ηFε(uε,A);
(iii) vδ,ηε→u in L1(A;Rn) as ε→0.
Proof. We first consider the case f(t)=min{at,b}, with a,b>0. For given η, we introduce the truncated kernel ρη(x):=11−σηρ(x)(1−χηS(x)), where the constant ση is given by
ση:=∫ηSρ(x)dx, |
and ση→0 as η→0. Notice that with this choice of ση one has ∫Rnρη(x)dx=1.
For fixed δ∈(0,1), we then define
Cδ,η:=1(1−δ)n(1−ση), | (4.4) |
and the functions
ψη,δε(x):=ε∫ΩW(Euε(y))ρηε(1−δ)(y−x)dyψε(x):=ε∫ΩW(Euε(y))ρε(y−x)dy. |
Observe that, since W≥0, by the definition of ρη and assumption (N2), we get
ψη,δε(x)≤Cδ,ηψε(x) | (4.5) |
for all x∈A. Define now the following sets, depending on δ,η and S:
Kε:={x∈A:ψη,δε(x)≥Cδ,ηba}, | (4.6) |
K′ε:={x∈A:distS(x,Kε)≤δηε}. | (4.7) |
We prove the inclusion
K′ε⊆{x∈A:ψε(x)≥ba}. | (4.8) |
For this, if x∈K′ε then there exists z∈Kε such that |x−z|S≤δηε. Now, by the triangle inequality, for every y∈Ω it holds that
|x−y|Sε≤δη+|z−y|Sε, |
whence
|x−y|Sε≤|z−y|S(1−δ)ε |
if and only if |z−y|S≥(1−δ)ηε. In this case, since ρ is non-increasing with respect to |⋅|S, we have
ρ(y−xε)≥(1−ση)ρη(y−z(1−δ)ε). |
Notice that this inequality holds true also if |z−y|S<(1−δ)ηε. In this case, indeed, one has y−z(1−δ)ε∈ηS and hence ρη(y−z(1−δ)ε)=0 by definition of ρη. Rescaling the kernels and using (4.4) we get
ρηε(1−δ)(y−z)≤Cδ,ηρε(y−x), |
so that
ψε(x)≥ψη,δε(z)Cδ,η≥ba |
and the proof of (4.8) is concluded.
Now, from the inclusion (4.8) and the fact that f(t)=b for t≥ba, we deduce that
Ln(K′ε)≤εbFε(uε,A). | (4.9) |
Applying the coarea formula (see for instance [19,Theorem 3.14]) to the 1-Lipschitz function g(x):=distS(x,Kε) in the open set {0<g(x)<ηδε}⊂K′ε we get
εbFε(uε,A)≥Ln(K′ε)≥∫ηδε0Hn−1({g=t})dt. |
It follows that we can choose 0<δ′ε<ηδε such that, for
K′′ε:={x∈A:distS(x,Kε)≤δ′ε}, | (4.10) |
it holds
Hn−1(∂K′′ε)=Hn−1({x∈A:distS(x,Kε)=δ′ε})≤1ηδbFε(uε,A). | (4.11) |
We define a sequence (vδ,ηε) of functions in GSBVp(A;Rn) as
vδ,ηε(x):={∫Ωuε(y)ρη(1−δ)ε(y−x)dy if x∈A∖K′′ε,0 otherwise. | (4.12) |
Since ‖ρη‖1=1, by Proposition 2.6(ii) (applied for θε=ε(1−δ)) and the fact that, by construction and (4.9), it holds Ln(K′′ε)→0 when ε→0, we have that vδ,ηε→u in L1(A;Rn) as ε→0. We also have Hn−1(Jvδ,ηε)≤Hn−1(∂K′′ε), so that with (4.11) we deduce (ii) for Mδ,η=1ηδb.
Now, since Kε⊂K′′ε and A⊂⊂Ω, it holds ψη,δε(x)<Cδ,ηba for all x∈K′′ε. As f(t)=min{at,b}, this gives
f(ψη,δε(x))≥aCδ,ηψη,δε(x) | (4.13) |
for all x∈A∖K′′ε. Now, since the function f is concave and f(0)=0, it holds f(λt)≥λf(t) for all λ∈[0,1]. Combining with the monotonicity of f and (4.5), we have
f(ψε(x))≥1Cδ,ηf(ψη,δε(x)) | (4.14) |
for all x∈A. With this, using (4.13), the Jensen's inequality, (2.11), (4.12), and since W(0)=0, we get
Fε(uε,A)≥1ε∫A∖K′′εf(ψε(x))dx≥1εCδ,η∫A∖K′′εf(ψη,δε(x))dx≥aεC2δ,η∫A∖K′′εψη,δε(x)dx≥aC2δ,η∫A∖K′′εW(∫ΩEuε(y)ρη(1−δ)ε(y−x)dy)dx=aC2δ,η∫A∖K′′εW(Evδ,ηε(x))dx=a(1−ση)2(1−δ)2n∫AW(Evδ,ηε(x))dx. |
For a general f complying with (3.1), use Lemma 2.9 to find aδ,bδ>0 with aδ≥α(1−δ) and f(t)≥min{aδt,bδ} for all t∈R, and perform the same construction as in the previous step. This gives (iii), (ii) (with Mδ:=1δηbδ) and
Fε(uε,A)≥aδ(1−ση)2(1−δ)2n∫AW(Evδ,ηε(x))dx≥α(1−ση)2(1−δ)2n+1∫AW(Evδ,ηε(x))dx, |
that is (i).
In this section we derive by slicing a lower bound for the surface term in the energy. It is worth mentioning that, by virtue of (4.17), the desired estimate could be probably also obtained by adapting to the GSBD-setting the semi-discrete approach of [22,Proposition 6.4]. Nonetheless, that argument is quite delicate for our purposes, and more complicated than we need. It indeed aimed to provide an optimal lower bound for both the bulk and the surface terms in a unique proof by means of a slicing procedure. In our case, the general form of the bulk energy we are considering does not comply with slicing arguments. Therefore, on the one hand, the two terms have to be estimated separately. On the other hand, an independent and simpler strategy can be followed to provide a lower bound with optimal constant in front of the surface energy.
We set
τξ:=H1({x∈S:x=tξ for t∈R}), | (4.15) |
for every ξ∈Sn−1.
Proposition 4.3. Let ρ∈L∞(Rn;[0,+∞)) be a convolution kernel complying with (N1), and let Fε be defined as in (3.4). Assume (3.1) and (3.3). Let δ∈(0,1) be fixed, and consider a sequence εj→0. Let A∈A(Ω) and uj∈W1,p(A;Rn) converging to u in L1(A;Rn). Assume that
liminfj→+∞Fεj(uj,A)<+∞. |
Then u∈GSBDp(A) and
liminfj→+∞Fεj(uj,A)≥β(1−δ)∫Jξu∩Aτξ|⟨ν,ξ⟩|dHn−1 | (4.16) |
for every ξ∈Sn−1.
Proof. It follows from Proposition 4.1 and Theorem 2.4 that u∈GSBDp(A). To prove (4.16), we first note that, by virtue of the growth assumption (3.3), we have
W(Eu)≥c|Eu|p≥c|⟨(Eu)ξ,ξ⟩|p, |
for every ξ∈Sn−1. Thus, for every fixed ξ, since f is non-decreasing, it will be sufficient to provide a lower estimate for the energies
Fξεj(uj,A):=1εj∫Af(cεj∫Sεj(x)|⟨(Euj(z))ξ,ξ⟩|pρεj(z−x)dz)dx. | (4.17) |
We proceed by a slicing argument, and for each x∈A we denote by xξ and yξ the projections of x onto Ξ and Πξ, respectively. Since S is open and convex, for every fixed ξ∈Sn−1 we can find a radius r=r(δ,S)>0 such that the cylinder
Cξ(1−δ),r=(−λξ,δ,λξ,δ)×Bn−1r(0)⊂⊂S, | (4.18) |
where λξ,δ:=τξ2(1−δ) and τξ is the length of the section Sξ. Indeed, since S is open, some η>0 can be found such that ¯Bη(0) is contained in S. Now, if t=(1−δ)s for some s∈Sξ and y∈ξ⊥ with |y|≤η, then tξ+δy∈S from the convexity of S. Thus, it will suffice to choose r:=δη.
If we denote by mC the minimum of ρ on ¯Cξ(1−δ),r, we then have
Fξεj(uj,A)=∫ΠξdHn−1(yξ)(1εj∫Aξ,yξf(cεj∫Sεj(x)|⟨(Euj(z))ξ,ξ⟩|pρεj(z−x)dz)dxξ)≥∫ΠξdHn−1(yξ)(1εj∫Aξ,yξ˜f(1εn−1j∫Cξ(1−δ)εj,rεj(x)|⟨(Euj(z))ξ,ξ⟩|pdz)dxξ), | (4.19) |
where ˜f(t):=f(cmCt). Note that ˜f(t)→β as t→+∞.
We now set
Fξ,yξεj(uj,Aξ,yξ):=1εj∫Aξ,yξ˜f(1εn−1j∫Cξ(1−δ)εj,rεj(x)|⟨(Euj(z))ξ,ξ⟩|pdz)dxξ. |
We denote (with a slight abuse of notation) still with z the (n−1)-dimensional variable in Bn−1rεj(yξ). Set wξ,yξj(t):=−∫Bn−1rεj(yξ)⟨uj(z+tξ)),ξ⟩dz.
By virtue of Lemma 2.7(ii), applied with θεj=rεj, we have that wξ,yξj converges to uξ,yξ in L1(Aξ,yξ) for a.e. yξ. Furthermore, setting g(t):=˜f(ωn−1rn−1t), Fubini's Theorem, Jensen's inequality and the monotonicity of ˜f entail that
Fξ,yξεj(uj,Aξ,yξ)=1εj∫Aξ,yξ˜f(1εn−1j∫Bn−1rεj(yξ)dz∫xξ+λξ,δεjxξ−λξ,δεj|⟨(Euj(z+tξ))ξ,ξ⟩|pdt)dxξ=1εj∫Aξ,yξ˜f(1εn−1j∫xξ+λξ,δεjxξ−λξ,δεj(∫Bn−1rεj(yξ)|⟨(Euj(z+tξ))ξ,ξ⟩|pdz)dt)dxξ≥1εj∫Aξ,yξ˜f(ωn−1rn−1∫xξ+λξ,δεjxξ−λξ,δεj(−∫Bn−1rεj(yξ)⟨(Euj(z+tξ))ξ,ξ⟩dz)pdt)dxξ=1εj∫Aξ,yξg(∫xξ+λξ,δεjxξ−λξ,δεj|˙wξ,yξj(t)|pdt)dxξ=λξ,δ1λξ,δεj∫Aξ,yξg(∫xξ+λξ,δεjxξ−λξ,δεj|˙wξ,yξj(t)|pdt)dxξ, | (4.20) |
Now, for the function t↦g(t) it still holds g(t)→β when t→+∞. Hence, applying Theorem 2.10 to the one-dimensional energies
˜Fξ,yξεj(wξ,yξj,Aξ,yξ):=1λξ,δεj∫Aξ,yξg(∫xξ+λξ,δεjxξ−λξ,δεj|˙wξ,yξj(t)|pdt)dxξ |
we obtain the lower bound
liminfj→+∞˜Fξ,yξεj(wξ,yξj,Aξ,yξ)≥2β#(Juξ,yξ∩Aξ,yξ). | (4.21) |
Therefore, using (4.20) and (4.21) we deduce
liminfj→+∞Fξ,yξεj(uj,Aξ,yξ)≥λξ,δliminfj→+∞˜Fξ,yξεj(wξ,yξj,Aξ,yξ)≥βτξ(1−δ)#(Juξ,yξ∩Aξ,yξ). |
With (4.19) and Fatou's Lemma we finally have
liminfj→+∞Fεj(uj,A)≥liminfj→+∞∫ΠξFξ,yξεj(uj,Aξ,yξ)dHn−1(yξ)≥∫Πξ(liminfj→+∞Fξ,yξεj(uj,Aξ,yξ))dHn−1(yξ)≥βτξ(1−δ)∫Πξ#(Juξ,yξ∩Aξ,yξ)dHn−1(yξ)=βτξ(1−δ)∫Jξu∩A|⟨νu,ξ⟩|dHn−1, |
where in the last equality we used the Area Formula. This concludes the proof of (4.16).
For any A∈A(Ω), we denote by F′(u,A) and G′(u,A) the lower Γ-limits of Fε(u,A) and Gε(u,A), respectively, as defined in (2.12). It holds that G′(u,A)≥F′(u,A) for each A∈A(Ω) and u∈L1(A;Rn) (see, e.g., [16,Proposition 6.7]). The results of the previous subsection lead to the following estimate.
Proposition 4.4. Assume (3.1), (3.3), and (3.6). Let Fε and Gε be defined as in (3.4) and (3.7), respectively, and let ρ comply with (N1) -(N2). Let u∈L1(Ω;Rn), A∈A(Ω), and define F′(u,A) and G′(u,A) by (2.12) in correspondence of Fε and Gε, respectively. If F′(u,A)<+∞, then u∈GSBDp(A) and
(i) F′(u,A)≥α∫AW(Eu)dx
(ii) G′(u,A)≥F′(u,A)≥β∫Jξu∩Aτξ|⟨νu,ξ⟩|dHn−1
for every ξ∈Sn−1. If in addition G′(u,A)<+∞ holds, then one also has
(iii) G′(u,A)≥α∫AW(Eu)dx+∫Aψ(|u|)dx.
Proof. With (2.12) and a diagonal argument, one may find (not relabeled) subsequences (uj) and (˜uj)} converging to u in L1(A;Rn) such that
F′(u,A)=liminfj→+∞Fεj(uj,A),G′(u,A)=liminfj→+∞Gεj(˜uj,A). |
With the first equality and Proposition 4.3 we have that, if F′(u,A)<+∞, then u∈GSBDp(A). By the second one, the superadditivity of the liminf, Fatou's lemma and (2.12), we have
G′(u,A)=liminfj→+∞Gεj(˜uj,A)≥liminfj→+∞Fεj(˜uj,A)+liminfj→+∞∫Aψ(|˜uj|)dx≥F′(u,A)+∫Aψ(|u|)dx. |
Hence, (iii) will follow once we have proved (i).
We therefore only have to check (i) and (ii). To this aim, let η,δ∈(0,1) be fixed. Then, by applying Proposition 4.2 to the sequence (uj), we can find a sequence (vδ,ηj)⊂GSBVp(A;Rn) such that vδ,ηj→u in L1(A) as εj→0 and
(a) α(1−ση)2(1−δ)2n+1∫AW(Evδ,ηj(x))dx≤Fεj(uj,A);
(b) Hn−1(Jvδ,ηj∩A)≤Mδ,ηFεj(uj,A).
Now, the equiboundedness of Fεj(uj,A) combined with the bounds (a) and (b) allows to apply the lower semicontinuity part of Theorem 2.4 to the sequence (vδ,ηj). Taking into account that A∞=∅ because u∈L1(A;Rn), by the convexity of W and (2.10), (ii), we have
α(1−ση)2(1−δ)2n+1∫AW(Eu(x))dx≤liminfj→+∞∫AW(Evδ,ηj(x))dx≤liminfj→+∞Fεj(uj,A)=F′(u,A). |
We then obtain (i) by letting δ→0 and η→0 above.
For what concerns (ii), from (4.16) of Proposition 4.3 we get
β(1−δ)∫Jξu∩Aτξ|⟨νu,ξ⟩|dHn−1≤liminfj→+∞Fεj(uj,A)=F′(u,A) |
for every ξ∈Sn−1, so that (ii) follows by taking the limit as δ→0 again.
For the proof of the Γ-liminf inequality, we need the following lemma, which can be found in [22,Lemma 4.5].
Lemma 4.5. Let S⊂Rn be a bounded, convex and symmetrical set, and let ϕρ and τξ be defined as in (3.2) and (4.15), respectively. Then
ϕρ(v)=supξ∈Sn−1τξ|⟨v,ξ⟩|. | (4.22) |
We are now in a position to prove the Γ-liminf inequality.
Proposition 4.6. Let ρ∈L∞(Rn;[0,+∞)) be a convolution kernel satisfying (N1)-(N2). Assume (3.1), (3.3), and (3.6). Consider Fε, and Gε given by (3.4), and (3.7), respectively. Let u∈L1(Ω;Rn) and let A∈A(Ω), and define F′(u,A) and G′(u,A) by (2.12) in correspondence of Fε and Gε, respectively. If F′(u,A)<+∞, then u∈GSBDp(A) and
F′(u,A)≥α∫AW(Eu)dx+β∫Ju∩Aϕρ(ν)dHn−1. |
If it additionally holds G′(u,A)<+∞, then
G′(u,A)≥α∫AW(Eu)dx+β∫Ju∩Aϕρ(ν)dHn−1+∫Aψ(|u|)dx. |
Proof. The proof can be obtained by a standard localization method based on Lemma 2.8. In order to prove, e.g., the second assertion containing an additional term, we can apply Lemma 2.8 to the set function μ(A):=G′(u,A), which is superadditive on disjoint open sets since Gε(u,⋅) is superadditive as a set function:
G′(u,A1∪A2)≥G′(u,A1)+G′(u,A2) whenever A1,A2∈A(Ω) with A1∩A2=∅. |
Then, we consider the positive measure λ(A):=Ln(A)+Hn−1(Ju∩A) and the sequence (φh)h≥0 of λ-measurable functions on A defined as
φ0(x):={αW(Eu(x))+ψ(|u(x)|), if x∈A∖Ju,0, if x∈A∩Ju, |
φh(x):={0, if x∈A∖Ju,βϕξh(x), if x∈A∩Ju, |
where
ϕξh(x)={τξh|⟨νu(x),ξh⟩|, if x∈Jξhu∩A,0, otherwise in Ju∩A, |
for (ξh)h≥1 a dense sequence in Sn−1.
Now, by virtue of Proposition 4.4 it holds that
μ(A)≥∫Aφhdλ |
for every h=0,1,…, so that all the assumptions of Lemma 2.8 are satisfied. The assertion then follows once we notice that, taking into account Lemma 4.5, it holds
suph≥0φh(x)=φ(x):={αW(Eu(x))+ψ(|u(x)|), if x∈A∖Ju,βϕρ(νu(x)), if x∈A∩Ju, |
for λ-a.e. x∈A.
We denote by F" and G" the upper Γ-limits of (Fε) and (Gε), respectively, as defined in (2.13).
Proposition 5.1. Let u∈GSBDp(Ω)∩L1(Ω;Rn). Then
F"(u)≤α∫ΩW(Eu)dx+β∫Juϕρ(ν)dHn−1. | (5.1) |
If, in addition, it holds that ∫Ωψ(|u|)dx<+∞, then
G"(u)≤α∫ΩW(Eu)dx+β∫Juϕρ(ν)dHn−1+∫Ωψ(|u|)dx. | (5.2) |
Proof. We only prove (5.1) by using the density result of Theorem 2.3, as (5.2) follows by an analogous construction with the additional property (2.9).
In view of Theorem 2.3 and remarks below, since we perform a local costruction and by a diagonal argument it is not restrictive to assume that u∈W(Ω;Rn) and that Ju is a closed simplex contained in any of the coordinate hyperplanes, that we denote by K.
For every h>0, let Kh:=∪x∈KS(x,h) be the anisotropic h-neighborhood of K. As K is compact and (n−1)-rectifiable, it holds (see for instance [21,Theorem 3.7])
limh→01hLn(Kh)=∫Kϕρ(ν)dHn−1 | (5.3) |
(observe that a factor 2 is already contained in our definition (3.2) of ϕρ). Let γε>0 be a sequence such that γε/ε→0 as ε→0. Notice that, for ε small,
K⊂Kγε⊂⊂Kγε+ε⊂⊂Ω, |
recalling that K⊂Ω. Let ϕε be a smooth cut-off function between Kγε and Kγε+ε, and set
uε(x):=u(x)(1−ϕε(x)). |
Since u∈W1,∞(Ω∖Ju;Rn) we have uε∈W1,∞(Ω;Rn). Note also that, by the Lebesgue Dominated Convergence Theorem, uε→u in L1(Ω;Rn). Moreover, since uε=u on S(x,ε)∩Ω if x∉Kγε+ε, we have
Fε(uε)≤1ε∫Ωf(ε∫S(x,ε)∩ΩW(Eu(y))ρε(y−x)dy)dx+βLn(Kγε+ε)ε. | (5.4) |
Setting
wε(x):=∫S(x,ε)∩ΩW(Eu(y))ρε(y−x)dy, |
we have that wε(x) converges to w(x):=W(Eu(x)) in L1loc(Ω) as ε→0. Since f complies with (3.1) and it is increasing, there exists ˜α>α such that f(t)≤˜αt for every t≥0.
This gives
1εf(εwε(x))≤˜αwε(x) for every x∈Ω and every ε>0, |
and, taking into account that limt→0+f(t)t=α, we also infer that
1εf(εwε(x))→αw(x) for a.e. x∈Ω. |
Thus, by Lebesgue's Dominated Convergence Theorem,
limε→01ε∫Ωf(ε∫S(x,ε)∩ΩW(Eu(y))ρε(y−x)dy)dx=α∫ΩW(Eu)dx. |
As γε+εε→1 as ε→0, from (5.3), (5.4), the subadditivity of the limsup and (2.13) we get (5.1).
Proof of Theorems 3.1 and 3.2. The two results follow by combining Propositions 4.1, 4.6, and 5.1
The authors are members of Gruppo Nazionale per l'Analisi Matematica, la Probabilità e le loro Applicazioni (GNAMPA) of INdAM. G. Scilla and F. Solombrino have been supported by the Italian Ministry of Education, University and Research through the Project "Variational methods for stationary and evolution problems with singularities and interfaces" (PRIN 2017).
The authors declare no conflict of interest.
[1] |
L. Ambrosio, A. Coscia, G. Dal Maso, Fine properties of functions with bounded deformation, Arch. Rational Mech. Anal., 139 (1997), 201–238. doi: 10.1007/s002050050051
![]() |
[2] | L. Ambrosio, N. Fusco, D. Pallara, Functions of bounded variation and free discontinuity problems, Clarendon Press, 2000. |
[3] |
G. Bellettini, A. Coscia, G. Dal Maso, Compactness and lower semicontinuity properties in SBD(Ω), Math. Z., 228 (1998), 337–351. doi: 10.1007/PL00004617
![]() |
[4] | A. Braides, Approximation of free-discontinuity problems, Springer Science & Business Media, 1998. |
[5] | A. Braides, Γ-convergence for beginners, Oxford: Oxford University Press, 2002. |
[6] |
A. Braides, G. Dal Maso, Non-local approximation of the Mumford-Shah functional, Calc. Var., 5 (1997), 293–322. doi: 10.1007/s005260050068
![]() |
[7] | F. Cagnetti, A. Chambolle, L. Scardia, Korn and Poincaré-Korn inequalities for functions with a small jump set, Math. Ann., 2021, https://doi.org/10.1007/s00208-021-02210-w. |
[8] |
A. Chambolle, V. Crismale, A density result in GSBDp with applications to the approximation of brittle fracture energies, Arch. Rational Mech. Anal., 232 (2019), 1329–1378. doi: 10.1007/s00205-018-01344-7
![]() |
[9] | A. Chambolle, V. Crismale, Phase-field approximation for a class of cohesive fracture energies with an activation threshold, Adv. Calc. Var., 2020, https://doi.org/10.1515/acv-2019-0018. |
[10] | A. Chambolle, V. Crismale, Compactness and lower semicontinuity in GSBD, J. Eur. Math. Soc., 23 (2021), 701–719. |
[11] |
G. Cortesani, Sequences of non-local functionals which approximate free-discontinuity problems, Arch. Rational Mech. Anal., 144 (1998), 357–402. doi: 10.1007/s002050050121
![]() |
[12] |
G. Cortesani, R. Toader, A density result in SBV with respect to non-isotropic energies, Nonlinear Anal., 38 (1999), 585–604. doi: 10.1016/S0362-546X(98)00132-1
![]() |
[13] |
G. Cortesani, R. Toader, Non-local approximation of non-isotropic free-discontinuity problems, SIAM J. Appl. Math., 59 (1999), 1507–1519. doi: 10.1137/S0036139997327691
![]() |
[14] | V. Crismale, M. Friedrich, F. Solombrino, Integral representation for energies in linear elasticity with surface discontinuities, Adv. Calc. Var., 2020, https://doi.org/10.1515/acv-2020-0047. |
[15] |
V. Crismale, G. Scilla, F. Solombrino, A derivation of Griffith functionals from discrete finite-difference models, Calc. Var., 59 (2020), 193. doi: 10.1007/s00526-020-01858-7
![]() |
[16] | G. Dal Maso, An introduction to Γ-convergence, Birkhäuser, 1993. |
[17] |
G. Dal Maso, Generalized functions of bounded deformation, J. Eur. Math. Soc., 15 (2013), 1943–1997. doi: 10.4171/JEMS/410
![]() |
[18] | G. Del Nin, Rectifiability of the jump set of locally integrable functions, Ann. Scuola Norm Sci., 2020, In press. |
[19] | L. Evans, R. Gariepy, Measure theory and fine properties of functions, revised edition, CRC Press, 2015. |
[20] | M. Friedrich, M. Perugini, F. Solombrino, Γ-convergence for free-discontinuity problems in linear elasticity: Homogenization and relaxation, arXiv: 2010.05461. |
[21] |
L. Lussardi, E. Villa, A general formula for the anisotropic outer Minkowski content of a set, P. Roy. Soc. Edinb. A, 146 (2016), 393–413. doi: 10.1017/S0308210515000542
![]() |
[22] | M. Negri, A non-local approximation of free discontinuity problems in SBV and SBD, Calc. Var., 25 (2005), 33–62. |
[23] |
G. Scilla, F. Solombrino, Non-local approximation of the Griffith functional, Nonlinear Differ. Equ. Appl., 28 (2021), 17. doi: 10.1007/s00030-021-00682-y
![]() |
[24] | R. Temam, Mathematical problems in plasticity, Gauthier-Villars, 1985. |
1. | Dario Mazzoleni, Benedetta Pellacci, Calculus of variations and nonlinear analysis: advances and applications, 2023, 5, 2640-3501, 1, 10.3934/mine.2023059 | |
2. | Roberta Marziani, Francesco Solombrino, Non-local approximation of free-discontinuity problems in linear elasticity and application to stochastic homogenisation, 2024, 154, 0308-2105, 1060, 10.1017/prm.2023.51 |