Convergence of two-scale expansions for elastic heterogeneous plates

Adrien Lesage; Arthur Leb\'ee; Fr\'ed\'eric Legoll; Virginie Ehrlacher

Convergence of two-scale expansions for elastic heterogeneous plates

T0 review · 0 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper establishes order-$\sqrt{\varepsilon}$ strong convergence of two-scale expansions for oscillatory plate problems, including the previously resistant bending case.

desk verdict A rigorous, well-disclosed proof of new H^1_epsilon strong-convergence estimates for two-scale expansions of heterogeneous plates; the bending case is genuinely new and the main limitation (W^{1,∞} corrector regularity, d=2) is openly stated. read the letter →

arxiv 2507.20874 v1 pith:SZRSVHBL submitted 2025-07-28 math.AP

Virginie Ehrlacher , Arthur Lebée , Frédéric Legoll , Adrien Lesage This is my paper · ORCID

classification math.AP MSC 35B2774K2074Q0574B05

keywords two-scaleexpansionhomogenizationthinplateslinearelasticitymembraneproblembendingstrongconvergenceperiodiccoefficients

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper proves that, for heterogeneous plates whose in-plane microstructure varies on the same scale $\varepsilon$ as the thickness, the true solution is approximated by an explicit two-scale expansion with an error of order $\sqrt{\varepsilon}$ in the natural scaled energy norm. The first main result covers linear diffusion in arbitrary dimension; the constant is tracked explicitly in the size of the plate, which is what a numerical error analysis needs. For linear elasticity, under a classical mid-plane symmetry of the elasticity tensor, the problem decouples into a membrane and a bending problem, and the paper proves the membrane estimate in dimension two by adapting classical arguments. The bending case is the real obstacle: the classical proof scheme fails because no admissible $H^2_0(\omega)$ test function can be built from an arbitrary $H^1$ test field, and the authors replace it by a new strategy that identifies the weak limit of the stress tensor and then uses an exact identity that holds before passing to the limit.

What carries the argument

The load-bearing object is the two-scale expansion itself, built as $u^\star(x')+\varepsilon\,w_\alpha(x'/\varepsilon,x_d)\,\partial_\alpha(u^\star+g)$ for diffusion, $u^\star+\varepsilon\,w_{\alpha\beta}\,e_{\alpha\beta}(u^\star+g)$, with the transverse component scaled by $\varepsilon^2$, for the membrane case, and $u^\star+\varepsilon\,W^{\alpha\beta}\,\partial_{\alpha\beta}(u^\star_d+g_d)$ for the bending case. The rate is carried by three mechanisms: the scaled differential operators $\nabla_\varepsilon$ and $e_\varepsilon$, which encode the plate scaling in the $H^1_\varepsilon$ norm; the oscillation lemmas (Lemma 2.9 and its matrix counterpart Lemma 3.9) that kill the resonant in-plane oscillation of the corrector residual $Z(x'/\varepsilon,x_d)$ by representing it as the divergence of a periodic skew-symmetric potential and exploiting $\varepsilon\nabla_\varepsilon$, and which, for $d=2$, use a two-scale Taylor expansion of the test displacement; and, in the bending case, the enriched correctors $W^{\alpha\beta,\xi}$ together with the exact identity $(\Sigma^\varepsilon-\Sigma^\star):e(v)=0$, which replaces the unavailable homogenized test function by a stress-limit characterization.

What would settle it

Numerically compute the two-scale error $\|u^\varepsilon-u^{\varepsilon,1}\|_{H^1_\varepsilon(\Omega)}$ for a two-dimensional periodic bending problem whose cell corrector $W^{11}$ is only in $H^1$ (for instance a coefficient field with strong periodic discontinuities or corner-like oscillations), and check whether the rate is still $O(\sqrt{\varepsilon})$; if the rate worsens, the $W^{1,\infty}$ regularity assumptions are load-bearing rather than technical.

Watch

Extended reading notes

Core claim

The paper's central claim is that the two-scale ansatz—homogenized displacement plus cell corrector applied to the in-plane strain in the membrane case, or to the Hessian of the transverse displacement in the bending case—is accurate at order $\sqrt{\varepsilon}$ in the $H^1_\varepsilon$ norm for solutions of elliptic problems on plates of thickness $\varepsilon$ with $\varepsilon$-periodic in-plane heterogeneities. In the diffusion case (Theorem 2.5) this holds in any dimension with a constant explicit in $|\omega|$. In elasticity, under the symmetry assumption (80) and in dimension two, the membrane case (Theorem 3.12) and, by a different and new argument, the bending case (Theorem 3.21) both deliver the same $\sqrt{\varepsilon}$ rate, with the bending bound containing the data norms $N^{\mathrm{bend}}$ and $N^{\mathrm{bend}}_\Omega$. The genuinely new ingredient is the bending proof: instead of constructing a homogenized test function from the error, it passes to the limit in the stress through enriched bending correctors $W^{\alpha\beta,\xi}$, identifies the in-plane stress limit $\Sigma^\star_{\alpha\beta}=S^\star_{\alpha\beta\gamma\delta}\partial_{\gamma\delta}(u^\star_d+g_d)$, upgrades the regularity of $\Sigma^\star$, and then uses the exact identity $\int_\Omega (\Sigma^\varepsilon-\Sigma^\star):e(v)=0$, valid for every $v$ without taking $\varepsilon\to0$.

Load-bearing premise

The proofs require the cell correctors and the homogenized displacement to be Lipschitz, and in the bending case the transverse displacement to be smooth up to fourth order, regularity that is not automatic for general bounded elliptic coefficients.

Editorial extensions

If this is right

The estimates make two-scale expansions a legitimate tool for numerical analysis of plate multiscale methods: local elements of size $|\omega|^{1/(d-1)}$ inherit explicit error bounds, the application the paper identifies as pivotal for the companion MsFEM analysis.
The diffusion theorem upgrades the known weak homogenization limit for thin domains to strong convergence in the scaled $H^1$ norm in arbitrary dimension, with no restriction to $d=2$.
Under the symmetry assumption (80), membrane and bending problems are each approximated at the same $\sqrt{\varepsilon}$ order in dimension two, so the decoupled plate model is justified beyond weak convergence.
The bending theorem supplies a strong-convergence result for heterogeneous plates in the bending regime, and its proof route—identify the stress limit through enriched correctors, then freeze the identity before $\varepsilon\to0$—is new to this setting.
If the two-dimensional technical Lemma 3.9 is ever proven in dimension $d\ge3$, the same statements follow in all dimensions, as the authors explicitly note.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

As an extension the authors leave implicit: because the bending proof produces an explicit tensor $S^\star_{\alpha\beta\gamma\delta}$ and an exact identity $(\Sigma^\varepsilon-\Sigma^\star):e(v)=0$ for every admissible $v$, the same machinery should yield a computable a posteriori error indicator proportional to $\|\Sigma^\varepsilon-\Sigma^\star\|_{L^2}$ for plate multiscale methods; this is our
A testable extension is whether the rate sharpens to $O(\varepsilon)$ when the data are smoother: the displayed bound already contains an $\varepsilon^{3/2}N^{\mathrm{bend}}$ term, but removing the leading $\sqrt{\varepsilon}\,|\omega|^{(d-2)/(2(d-1))}$ term would require a boundary-layer correction of higher order than the paper constructs.
For random stationary coefficients the periodic divergence-potential argument used in Lemmas 2.9 and 3.9 has no direct analogue, so an open and natural test is whether the same $\sqrt{\varepsilon}$ two-scale rate survives in expectation for random microstructures.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

This paper delivers what its title promises: H^1_epsilon error estimates for two-scale expansions in heterogeneous plates, in the diffusion, membrane, and bending cases. The genuinely new part is the bending case (Theorem 3.21), where the authors replace the standard test-function route with an enriched family of correctors (166) and a stress-identification lemma (Lemma 3.17). That strategy is not in the prior strong-convergence literature, which only handled homogeneous plates [8,10,11] or weak convergence [4,5]. The proofs are detailed and, as far as I can tell, coherent. The stress-tensor identification (170) is a real div-curl-type argument, and Lemma 3.20 gives the exact identity (no limit) needed to close the estimate. The paper also tracks the dependence on |ω| and on the loads, which is the right thing for the MsFEM application in their companion paper.

Soft spots, in proportion:

1. The W^{1,∞} regularity assumptions on the correctors and on the homogenized solution are load-bearing. For merely L^∞-elliptic period coefficients (e.g., piecewise-constant composites), correctors are only H^1, so the rates in Theorems 2.5, 3.12, and 3.21 are not established in that common physical setting. This is a genuine scope limitation, but it is stated plainly in each theorem, and it is standard for two-scale-expansion rate results. I don't see it as a hidden flaw.

2. The d=2 restriction for elasticity is explicit: Lemma 3.9 is proved only for d=2, as the authors disclose at length. That is a real restriction for plate applications (d=3), but not an internal gap.

3. The regularity assumptions on f, g, h are somewhat strong (e.g., gd ∈ H^4(ω) in the bending case), but they are compatible with elliptic regularity on u⋆ and are not hidden.

I checked the stress-test concern: it correctly identifies that the W^{1,∞} corrector regularity is not automatic and is needed for the pointwise cutoff estimates. That is a fair concern, but it is already in the paper as a stated assumption, not an unacknowledged gap. The stress-test note itself makes clear the proofs appear internally gap-free. I agree with that.

There is no circularity or fitted constants. The estimates are derived from the original PDE and corrector problems; the stress limit is characterized, not calibrated. The citation pattern is appropriate; the companion-paper reference [13] is in preparation, but the present results stand alone.

Who should read it: anyone working on rigorous numerical analysis of MsFEM for plates, or on strong-convergence homogenization in thin domains. It deserves a serious referee; the main theorems are tightly stated and proven with unusual care about the local dependence on ω. My own verdict is accept, with the standard caveat that the W^{1,∞} regularity assumption should be highlighted in the abstract or introduction so users see it immediately.

Recommendation: engage with it. Send to a referee who knows homogenization and plate theory, and expect the referee to verify Lemma 3.17 and the Step 1 boundary estimates in Theorem 3.21, but not to find a fatal gap.

Referee Report

0 major / 4 minor

Summary. The manuscript proves strong convergence results, in the scaled H^1_ε norm, between solutions of highly oscillatory problems on thin plates and their two-scale expansions. Section 2 treats scalar diffusion: under the shape regularity condition (20) and regularity assumptions on the homogenized solution and the cell correctors, Theorem 2.5 gives an explicit √ε bound with the dependence on |ω| tracked. Section 3 treats linear elasticity under the mirror symmetry assumption (80). The membrane case is handled in Theorem 3.12 by a careful adaptation of the diffusion argument, while the bending case, Theorem 3.21, requires a genuinely different strategy: one introduces rescaled stress fields Σε, proves weak compactness and regularity (Lemmas 3.15 and 3.19), identifies the in-plane stress limit through enriched correctors (Lemma 3.17), and then uses the exact identity ∫(Σε−Σ⋆):e(v)=0 for all v∈V (Lemma 3.20) to close the error estimate. Both elasticity theorems are restricted to d=2 because the auxiliary Lemma 3.9 is established only in that dimension. Appendices A–D collect the rescaling identities, H^div trace results, Korn inequalities, and self-contained proofs of the weak homogenization theorems.

Significance. If the results are correct, this is a valuable contribution: strong convergence rates for two-scale expansions in heterogeneous elastic plates were previously available in very few settings, and the bending-case argument is genuinely new. In particular, the exact stress identity of Lemma 3.20, and the use of enriched bending correctors to identify the in-plane stress limit in Lemma 3.17, are substantial technical advances. The explicit dependence of the constants on |ω| is important for the companion MsFEM error analysis. The proofs are detailed and the main theorems are conditional in a transparent way: they assume W^{1,∞} regularity of the cell correctors and W^{2,∞}-type regularity of the homogenized quantities, which is standard in two-scale expansion rate results but does not follow from mere L∞ ellipticity; the d=2 restriction is explicitly traced to Lemma 3.9. I found no internal derivation gap, and the stress-limit identification is not circular, as Remark 3.18 correctly notes.

minor comments (4)

[§3.5.1, proof of Lemma 3.9, around (88)–(95)] The symbol B is first used for J^i_{·,j}·∇_εφ and then for the symmetrized matrix eB+eB^T; this makes the manipulations in (95) and the following paragraph hard to follow. Rename the second object, for example B^{sym}, or use a different font.
[Theorem 3.21, Step 1, bounds for E_0 and E_1] The displayed exponents for |ω| in the bounds on E_0 and E_1 appear scrambled (e.g. |ω|^{d−1 d−2}); they should read |ω|^{(d−2)/(d−1)}. Since the theorem is stated for d=2 the issue is harmless, but the typography should be corrected for readability.
[§3.5, Theorem 3.21 proof and Lemmas 3.15–3.20] The proof of Theorem 3.21 constructs Σ⋆ along a weakly convergent subsequence, and the statement of the theorem concerns the whole sequence. Uniqueness of the full matrix Σ⋆ follows from Lemma 3.17 for the in-plane block and from the formulas in Lemma 3.19 for the remaining components; it would help to state this explicitly so that the reader sees immediately that the estimate holds without a final subsequence argument.
[Appendix D.3] When Lemma 3.17 is invoked to establish the bending homogenized equation before (84) has been proved, it is not immediately obvious that the argument is non-circular. A sentence referring the reader to Remark 3.18, and noting that u⋆∈V^B_KL follows from the symmetry of uε before the homogenized equation is identified, would remove potential confusion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the strong-convergence theorems are derived from the original PDE, explicitly stated corrector problems, and the homogenized equation; the recalled weak limits are re-proved in Appendix D, and no fitted parameter is promoted to a prediction.

full rationale

The derivation chain is self-contained. The weak homogenization limits (Theorem 2.4 and Theorem 3.1) are quoted from Caillerie [4,5], but full proofs are supplied in Appendix D by the oscillating-test-function method, so the later strong-convergence results do not rest on an unverified self-citation. The two-scale expansions are built from the homogenized solution u* and the cell correctors, and the error estimates are obtained by coercivity of the original problem together with explicit remainder bounds that use the corrector equations and the variational formulation of the homogenized problem. No constant is fitted to data, and no quantity that is defined in terms of the target error is later called a prediction. In the bending case, the identification of the stress limit in Lemma 3.17 is explicitly noted in Remark 3.18 not to use the homogenized equation for u*, only weak convergence of u_epsilon^d; the homogenized equation is then derived from this identification, which is the correct logical order. Lemma 3.19 uses elliptic regularity of the homogenized problem only to obtain regularity of the limiting stress, under assumptions that are stated in Theorem 3.21. The only load-bearing caveat is the explicit W^{1,infinity}-type regularity assumed for the correctors and for nabla^2(u*_d+g_d); this is disclosed in the theorem statements and is a standard hypothesis in two-scale-expansion rate results, not a circular input. Self-references such as [13] are motivational and not used in any proof. Therefore no circular step is present.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The central claims rest on standard functional-analytic tools (Korn, Poincare, elliptic regularity) and on regularity and symmetry assumptions on the material and data that are stated but not proved. No parameters are fitted to data; all constants in the estimates are explicit in terms of c+-, eta, and norms of the data.

assumptions (6)

standard math Korn's inequality in H^1(Omega) and H^1_0(Omega) (Lemmas C.1 and C.2 of the paper, quoted from [6])
Used to obtain uniform a priori bounds on the scaled strain e^epsilon(u^epsilon) in (65) and the Poincare-Korn estimate (78); the paper proves only the V-version (Lemma C.3).
standard math Lemma 2.13 (Jikov-Kozlov-Oleinik, p.6): each zero-mean divergence-free periodic vector field is the divergence of a skew-symmetric periodic matrix
Basis for the representation (29) and the proof of the oscillatory-integral estimates in Lemmas 2.9 and 3.9.
standard math Standard elliptic regularity for the fourth-order homogenized plate equation (84), giving u*_d in H^4(omega) from L^2 data on a smooth domain omega
Used in Lemma 3.19 to obtain H^2(omega)-regularity of Sigma*_{alpha beta} via (170).
domain assumption Regularity assumptions on the correctors and homogenized solution: w_alpha, w_{alpha beta}, W^{alpha beta} in (W^{1,infinity})^d, nabla(u*+g) in W^{1,infinity}, nabla^2(u*_d+g_d) in W^{2,infinity}
Stated explicitly in Theorems 2.5, 3.12 and 3.21; used to control the boundary-layer and lower-order terms in the two-scale expansion. Not automatic for general L^infinity periodic coefficients.
domain assumption Symmetry of the elasticity tensor with respect to the plate mid-plane (Assumption (80)), which decouples membrane and bending problems and yields K*_{12}=0
Used in Lemma 3.7 to split the problem and in Lemma 3.8 to characterize the homogenized limit; classical and satisfied by isotropic plates symmetric about the mid-plane.
domain assumption Shape regularity of the in-plane domain: diam(omega)/rho_0 <= eta (Assumption (20))
Required for the Poincare-type estimates (Lemmas 2.7 and 3.4) with constants independent of |omega|, which are needed for the dependence of the error on the cell size.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Convergence of two-scale expansions for elastic heterogeneous plates." pith.science (2026). https://pith.science/paper/SZRSVHBL

@misc{pith2026250720874,
  author       = {Pith},
  title        = {Pith review of: Convergence of two-scale expansions for elastic heterogeneous plates},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SZRSVHBL}},
  note         = {Machine review of arXiv:2507.20874}
}

read the original abstract

The aim of this article is to prove strong convergence results on the difference between the solution to highly oscillatory problems posed in thin domains and its two-scale expansion. We first consider the case of the linear diffusion equation and establish such results in arbitrary dimensions, by using a straightforward adaptation of the classical arguments used for the homogenization of highly oscillatory problems posed on fixed (non-thin) domains. We next consider the linear elasticity problem, which raises challenging difficulties in its full generality. Under some classical assumptions on the symmetries of the elasticity tensor, the problem can be split into two independent problems, the membrane problem and the bending problem. Focusing on two-dimensional problems, we show that the membrane case can actually be addressed using a careful adaptation of classical arguments. In the bending case, the scheme of the proof used in the membrane and diffusion cases can however not be straightforwardly adapted. In that bending case, we establish the desired strong convergence results by using a different strategy of proof, which seems, up to our knowledge, to be new.

Figures

Figures reproduced from arXiv: 2507.20874 by the authors.

**Figure 1.** The plate and its microstructure for d = 3. main result in that case) can indeed be obtained, in arbitrary dimensions, by using standard arguments. The situation turns out to be different in the case of linear elasticity, which we address in Section 3. We assume there that the mechanical composition of the heterogeneous material is symmetric with respect to its medium plane, which corresponds to assuming that the co… view at source ↗

**Figure 2.** Schematic representation of the plate Ωε . Let (ei)1≤i≤d be the canonical basis of R d . For any x = (xi)1≤i≤d ∈ R d , we set x ′ := (xi)1≤i≤d−1 ∈ R d−1 . For any M := (Mij )1≤i,j≤d ∈ R d×d , we set M′ := (Mij )1≤i,j≤d−1 ∈ R (d−1)×(d−1). The set of d × d symmetric matrices is denoted by R d×d s and c−, c+ > 0 are some fixed positive constants. We also define the periodic cells Y := (0, 1)d−1 and Y := Y × − 1 2 , 1… view at source ↗

**Figure 3.** ). This is of course a standard step when studying plate problems (see e.g. [4, eq. (3.1)]). Ω ε ε 1 Ω [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 29 canonical work pages

[13]

Ehrlacher, A

V. Ehrlacher, A. Leb´ ee, F. Legoll, and A. Lesage. Multiscale Finite Element Methods for elastic heterogeneous plates. in preparation

work page
[1]

G. Allaire. Shape optimization by the homogenization method , volume 146 of Applied Mathematical Sciences. Springer, New York, 2002

work page 2002
[2]

Bensoussan, J.-L

A. Bensoussan, J.-L. Lions, and G. Papanicolaou. Asymptotic analysis for periodic struc- tures, volume 374. American Mathematical Soc., 2011

work page 2011
[3]

Blanc and C

X. Blanc and C. Le Bris. Homogenization theory for multiscale problems: An introduction , volume 21 of Modeling, Simulation and Applications . Springer, 2023

work page 2023
[4]

Caillerie

D. Caillerie. Homog´ en´ eisation des ´ equations de la diffusion stationnaire dans les domaines cylindriques aplatis. RAIRO Analyse num´ erique, 15(4):295–319, 1981

work page 1981
[5]

Caillerie

D. Caillerie. Thin elastic and periodic plates. Mathematical Methods in the Applied Sci- ences, 6(1):159–191, 1984

work page 1984
[6]

P.G. Ciarlet. Mathematical Elasticity: Volume I: three-dimensional elasticity . North- Holland, 1988

work page 1988
[7]

P.G. Ciarlet. Mathematical Elasticity: Volume II: theory of plates . North-Holland, 1997

work page 1997

Show all 29 references

[8]

Ciarlet and P

P.G. Ciarlet and P. Destuynder. Justification of the two-dimensional linear plate model. J. Mec., 18(2):315–344, 1979

1979
[9]

Cioranescu and P

D. Cioranescu and P. Donato. An introduction to homogenization . Oxford University Press, New York, 1999

1999
[10]

Dauge and I

M. Dauge and I. Gruais. Asymptotics of arbitrary order for a thin elastic clamped plate, I. Optimal error estimates. Asymptotic Analysis, 13(2):167–197, 1996

1996
[11]

Destuynder

P. Destuynder. Comparaison entre les mod` eles tridimensionnels et bidimensionnels de plaques en ´ elasticit´ e.RAIRO Analyse num´ erique, 15(4):331–369, 1981

1981
[12]

Efendiev and T

Y. Efendiev and T. Hou. Multiscale Finite Element Methods: Theory and Applications , volume 4 of Surveys and Tutorials in the Applied Mathematical Sciences . Springer New York, 2009. 81

2009
[14]

Engquist and P.E

B. Engquist and P.E. Souganidis. Asymptotic and numerical homogenization. Acta Nu- merica, 17:147–190, 2008

2008
[15]

Griso and B

G. Griso and B. Miara. Homogenization of periodically heterogeneous thin beams. Chinese Annals of Mathematics, Series B , 39(3):397–426, 2018

2018
[16]

Gustafsson and J

B. Gustafsson and J. Mossino. Non-periodic explicit homogenization and reduction of dimension: the linear case. IMA Journal of Applied Mathematics , 68(3):269–298, 2003

2003
[17]

Gustafsson and J

B. Gustafsson and J. Mossino. Compensated compactness for homogenization and reduc- tion of dimension: the case of elastic laminates. Asymptotic Analysis , 47(1-2):139–169, 2006

2006
[18]

Hornung, S

P. Hornung, S. Neukamm, and I. Velˇ ci´ c. Derivation of a homogenized nonlinear plate theory from 3D elasticity. Calculus of Variations and Partial Differential Equations , 51(3- 4):677–699, 2014

2014
[19]

Hornung, M

P. Hornung, M. Pawelczyk, and I. Velˇ ci´ c. Stochastic homogenization of the bending plate model. Journal of Mathematical Analysis and Applications , 458(2):1236–1273, 2018

2018
[20]

Jikov, S.M

V.V. Jikov, S.M. Kozlov, and O.A. Ole ˘ ınik.Homogenization of differential operators and integral functionals. Springer-Verlag, Berlin, 1994

1994
[21]

Kohn and M

R.V. Kohn and M. Vogelius. A new model for thin plates with rapidly varying thickness. III: Comparison of different scalings. Quarterly of Applied Mathematics, 44(1):35–48, 1986

1986
[22]

C. Le Bris. Syst` emes multi-´ echelles: mod´ elisation et simulation, volume 47 of Math´ ematiques et applications. Springer, 2005

2005
[23]

Le Bris and F

C. Le Bris and F. Legoll. Examples of computational approaches for elliptic, possibly multiscale PDEs with random inputs. J. Comput. Physics , 328:455–473, 2017

2017
[24]

A. Lesage. Multi-scale approaches for the computation and the optimization of heterogeneous plates . PhD thesis, Universit´ e Paris-Est, 2020. (available at https://theses.hal.science/tel-03587182)

2020
[25]

Lewinski and J.J

T. Lewinski and J.J. Telega. Plates, laminates and shells: asymptotic analysis and homog- enization, volume 52 of Advances in Mathematics for Applied Sciences . World Scientific, 2000

2000
[26]

Marohni´ c and I

M. Marohni´ c and I. Velˇ ci´ c. Homogenization of bending theory for plates: the case of oscillations in the direction of thickness. Communications on Pure and Applied Analysis , 14(6):2151–2168, 2015

2015
[27]

Marohni´ c and I

M. Marohni´ c and I. Velˇ ci´ c. Non-periodic homogenization of bending-torsion theory for inextensible rods from 3D elasticity. Annali di Matematica Pura ed Applicata (1923-) , 195(4):1055–1079, 2016

1923
[28]

Murat and L

F. Murat and L. Tartar. H-convergence. In A. Cherkaev and R. V. Kohn, editors, Topics in the mathematical modelling of composite materials , volume 31 of Progress in Nonlinear Differential Equations and their Applications , pages 21–44. Birkhauser, 1997

1997
[29]

Velˇ ci´ c

I. Velˇ ci´ c. On the derivation of homogenized bending plate model.Calculus of Variations and Partial Differential Equations , 53(3-4):561–586, 2015. 82

2015

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [13]

Ehrlacher, A

V. Ehrlacher, A. Leb´ ee, F. Legoll, and A. Lesage. Multiscale Finite Element Methods for elastic heterogeneous plates. in preparation

work page

[2] [1]

G. Allaire. Shape optimization by the homogenization method , volume 146 of Applied Mathematical Sciences. Springer, New York, 2002

work page 2002

[3] [2]

Bensoussan, J.-L

A. Bensoussan, J.-L. Lions, and G. Papanicolaou. Asymptotic analysis for periodic struc- tures, volume 374. American Mathematical Soc., 2011

work page 2011

[4] [3]

Blanc and C

X. Blanc and C. Le Bris. Homogenization theory for multiscale problems: An introduction , volume 21 of Modeling, Simulation and Applications . Springer, 2023

work page 2023

[5] [4]

Caillerie

D. Caillerie. Homog´ en´ eisation des ´ equations de la diffusion stationnaire dans les domaines cylindriques aplatis. RAIRO Analyse num´ erique, 15(4):295–319, 1981

work page 1981

[6] [5]

Caillerie

D. Caillerie. Thin elastic and periodic plates. Mathematical Methods in the Applied Sci- ences, 6(1):159–191, 1984

work page 1984

[7] [6]

P.G. Ciarlet. Mathematical Elasticity: Volume I: three-dimensional elasticity . North- Holland, 1988

work page 1988

[8] [7]

P.G. Ciarlet. Mathematical Elasticity: Volume II: theory of plates . North-Holland, 1997

work page 1997