• Aucun résultat trouvé

General decay for a porous thermoelastic system with thermo-viscoelastic damping

N/A
N/A
Protected

Academic year: 2021

Partager "General decay for a porous thermoelastic system with thermo-viscoelastic damping"

Copied!
18
0
0

Texte intégral

(1)

HAL Id: hal-01432632

https://hal.archives-ouvertes.fr/hal-01432632

Preprint submitted on 11 Jan 2017

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

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

General decay for a porous thermoelastic system with

thermo-viscoelastic damping

Kilani Brahim, Djebabla Abdelhak

To cite this version:

(2)

General decay for a porous thermoelastic

system with thermo-viscoelastic damping

Kilani Brahim

(1)

and Djebabla Abdelhak

(2)

(1)

Mathematics, Dynamics and Modelization Laboratory,

University Badji Mokhtar

Department of Mathematics

PO BOX 12 Annaba 23000, Algeria

E-mail: kilbra2000@yahoo.fr

and

(2)

Laboratory of Applied Mathematics

University Badji Mokhtar

Department of Mathematics

PO BOX 12 Annaba 23000, Algeria

E-mail: abdelhak.djebabla@univ-annaba.dz

Abstract

In this article, we consider the one-dimensional viscoelastic porous thermoelastic system. We introduce two new numbers χ0 and χ1 that

characterizes the decay. We establish a general decay result for the case χ0 = χ1 = 0. To better observe the effect of the weak

dissipa-tion (thermal effect or thermo-viscoelastic damping) on changing the nature of the stability numbers, we refer the reader to [1, 2, 3] .

Keywords: General decay; Porous; Relaxation function; Vis-coelastic damping; Thermal effect.

(3)

1

Introduction

The one-dimensional porous-elastic model has the form 

ρ0utt= µuxx+ βϕx

ρ0κϕtt = αϕxx− βux− τ ϕt− ξϕ

where u is the longitudinal displacement, ϕ is the volume fraction, ρ0 > 0 is

the mass density, κ > 0 is the equilibrated inertia, and µ, α, β, τ , and ξ are the constitutive constants which are positive and satisfy µξ > β2. The first

contribution in this direction has been investigated in 2003 by Quintanilla [6]. The authors analyzed this model in a bounded domain with initial conditions and mixed boundary conditions and showed that the damping in the porous equation (−τ ϕt) is not strong enough to obtain an exponential decay. Only

the slow decay has been proved. The case τ = 0 and with viscoelastic damping term of the form γutxx acting on the right- hand side of the first

equation has been investigated by Rivera et al. [1], they proved that the decay rate of the solution is polynomial and cannot be exponential. Casas and Quintanilla [7] considered a system of the form

               ρ0utt− µuxx− bϕx+ βθx = 0 in (0, ∞) × (0, π) ρ0κϕtt− αϕxx+ bux+ ζϕt− mθ = 0 in (0, ∞) × (0, π) cθt− κθxx+ βutx+ mϕt = 0 in (0, ∞) × (0, π), u (x, 0) = u0(x), ϕ(x, 0) = ϕ0(x), θ(x, 0) = θ0(x) in (0, π) ut(x, 0) = u1(x), ϕt(x, 0) = ϕ1(x), in (0, π) u (x, t) = ϕx(x, t) = θx(x, t) = 0, x = 0, π, t ≥ 0

where θ denotes the temperature difference, they showed that the presence of the macrotemperature and the porous dissipations acting together stabilize the system exponentially. Soufyane [15] introduced a viscoelastic damping in the porous equation together with a macrotemperature effect and showed that the decay is exponential (resp. polynomial) when the relaxation function is of exponential (resp.polynomial) decay. A similar result was also obtained by Soufyane et al. [16,17], for the same system, with the frictional damping −τ ϕt replaced by two boundary viscoelastic dissipations of the form

u(L, t) = −R0tg1(t − s) [µux(L, s) + bϕ(L, s)] ds,

v(L, t) = −αR0tg2(t − s) ϕ (L, s) ds.

where g1 and g2 are positive nonincreasing functions. Recently, Messaoudi

(4)

system of type III with the presence of a viscoelastic damping                ρ1ϕtt− k (ϕx+ ψ)x+ θx= 0, ρ2ψtt− αψxx+ k (ϕx+ ψ) − θ + Rt 0 g (t − s) ψxx(x, s) ds = 0, ρ3θtt− κθxx− δθtxx+ βϕttx+ βψtt= 0, ϕ (x, 0) = ϕ0(x) , ϕt(x, 0) = ϕ1(x) , ψ (x, 0) = ψ0(x) , ψt(x, 0) = ψ1(x) , θ (x, 0) = θ0(x) , θt(x, 0) = θ1(x) , ϕ (0, t) = ϕ (1, t) = ψ (0, t) = ψ (1, t) = θ (0, t) = θ (1, t) = 0

where ϕ is the longitudinal displacement, ψ is the volume fraction, θ is the difference in temperature and the relaxation function g : R+ → R is a

non-increasing function. They established a general decay result for the case of equal speeds ρk

1 =

α ρ2



. Notice that the term δθtxx in the third equation

represents a strong damping. The term βθtprovides a weaker damping

(fric-tional damping) and so we consider with initial data and boundary conditions the following system

   ρ1ϕtt− k (ϕx+ ψ)x+ γθx = 0, ρ2ψtt− bψxx + k (ϕx+ ψ) − γθ + Rt 0 g (t − s) ψxx(x, s) ds = 0, ρ3θtt− lθxx+ βθt+ γϕttx+ γψtt = 0,

in Ω × (0, +∞), where Ω = [0, L] and the coefficients ρ1, ρ2, ρ3, k, b, γ, β

and l are positive constants.

Our aim in this work is to prove the general decay of solutions in the energy norm. If we make the change of functions

 ϕt = χ ∈ H01(Ω) , ψt = ξ ∈ H01(Ω) , we find    ρ1χtt− k (χx+ ξ)x+ γθtx = 0, ρ2ξtt− bψxx+ k (χx+ ψ) − γθt+ Rt 0 g (t − s) ξxx(x, s) ds = 0, ρ3θtt− lθxx+ βθt+ γχtx+ γξt= 0, (1)

to be considered with the following initial data and boundary conditions

χ (0, .) = χ0, χt(0, .) = χ1, ψ (0, .) = ψ0,

ψt(0, .) = ψ1, θ (0, .) = θ0, θt(0, .) = θ1,

(5)

The plan of the paper is as follows: In the next section we prepare some material needed to prove our result. We present the different functionals by which we modify the classical energy to get an equivalent one. Section 3 is devoted mainly to the statement and proof the general decay result.

2

Preliminaries

In this section, we present our hypotheses, and state without proof a global existence result. First, we denote by ∗ the usual convolution term

(f ∗ h) (t) = R0tf (t − s) h (s) ds (3) and the binary operators ♦ and o respectively, by

(f ♦h) (t) = R0tf (t − s) (h (t) − h (s)) ds (4)

and

(f oh) (t) = R0tf (t − s) (h (t) − h (s))2ds. (5) For the relaxation function g we assume:

(H1) g : R+→ R+ is differentiable function such that

g (0) > 0, λ := b −R0∞g (s) ds := b − ¯g > 0 (6) (H2) There exists a nonincreasing differentiable function ζ : R+→ R+

g0(t) ≤ −ζ (t) g (t) , ∀ t ≥ 0. (7) Remark. There are many functions satisfying (H1) and (H2) (see [21]).

Proposition 1 Let ((χ0, χ1) , (ξ0, ξ1) , (θ0, θ1)) ∈ (H01(0, L) × L2(0, L)) 3

be given and assume that g satisfied (H1) and (H2). Then, problem (1.2) has a unique global solution:

(6)

then the solution satisfies

(χ, ξ, θ) ∈ (C (R+; H2(0, L)) ∩ C1(R+; H01(0, L)) ∩ C2(R+; L2(0, L))) 3

.

Remark 1 Proposition 1 can be established using standard methods such as the Galerkin method (see [4] for example).

The first-order energy associated with problem (1-2) is given as

E(t) = 1 2 R Ωρ 2 1χ2t + ρ22ξt2+ ρ3θt2+ k (χx+ ξ)2+ lθx2 +  b −R0tg (s) ds  ξx2+ goξx o dx, t ≥ 0. (8)

Lemma 1 Under assumptions (H1) and (H2), we have

R Ωξt Rt 0g(t − s)ξxx(s) dsdx = 1 2 d dt h (goξx) −  Rt 0 g (s) ds  R Ωξ 2 xdx i −1 2(g 0 x) + 12g(t) R Ωξ 2 xdx. t ≥ 0

Proof. Integrating by parts and using the boundary conditions, we get

R Ωξt Rt 0 g(t − s)ξxx(s) dsdx = −  Rt 0 g(s)ds  R Ωξ 2 x(t) dsdx −R Ω Rt 0 g(t − s)ξtx(t) [ξx(s) − ξx(t)] dsdx = 12dtd hR(goξx) dx −  Rt 0 g(s)ds  R Ωξ 2 x(t) dx i −1 2 R Ω(g 0 x) dx +12g(t) R Ωξ 2 x(t) dx.

Lemma 2 For any function g ∈ C ([0 ∞) , R) and for any h ∈ L2(Ω) we

have that

[(g♦h) (t)]2 ≤Rt

0 g(τ )dτ (goh) (t), t ≥ 0.

Proof. By Cauchy-Schwarz inequality, we have

[(g♦h) (t)]2 ≤hRt 0 g (t − s) (h (t) − h (s)) ds i2 = h Rt 0g 1 2 (t − s) g 1 2 (t − s) (h (t) − h (s)) ds i2 ≤   Rt 0g (t − s) ds 12  Rt 0 g (t − s) (h (t) − h (s)) 2 ds 1 2 2 ,

(7)

Lemma 3 There exists a positive constant C0 such that R Ω(bξx− (g ∗ ξx)) 2 dx ≤ C0 R Ωξ 2 xdx + 4g R Ω(goξx) dx, t ≥ 0, with C0 = (2b2+ 4g2) .

Proof. Using the fact that (a2+ b2) ≤ 2a2+ 2b2 and lemma 3, we get

R Ω(bξx− (g ∗ ξx)) 2 dx ≤ 2b2Rξx2dx + 2R(g ∗ ξx) 2 dx ≤ 2b2R Ωξ 2 xdx + 2 R Ω Rt 0g(t − s) (ξx(t) − ξx(s) − ξx(t)) 2 dsdx ≤  2b2+ 4R0tg(t − s)ds 2 R Ωξ 2 xdx + 4 R Ω(g♦ξx) 2 dx ≤ (2b2+ 4g2)R Ωξ 2 xdx + 4g R Ω(goξx) dx.

Our result reads as follows.

Theorem 1 Let ((χ0, χ1) , (ξ0, ξ1) , (u0, u1)) ∈ (H01(0, 1) × L2(0, 1)) 3

be given, assume that g satisfies (H1) and (H2) and the coefficients of the system sat-isfy the condition

κ0 = γ + bρk1 − ρ2 = 0 and κ1 = ρρ32b + bγk − l = 0. (9)

Then, there exist two positive constants c0 and c1such that

E (t) ≤ c0e−c1 Rt1

t0 ζ(s)ds, ∀t ≥ t0.

3

Proof of the main result

The proof of our main result will be established through several lemmas.

Lemma 4 Let (ϕ, ψ, θ) be the solution of (1-2). Then the energy functional E, defined by (8) satisfies d dtE(t) = −β R Ωθ 2 tdx − 1 2g(t) R Ωξ 2 xdx + 1 2 R Ω(g 0 x) dx ≤ 0, t ≥ 0. (10)

Proof : Multiplying the first equation of (1-2) by χt, the second by ξt

and the third by θt then we integrate over (0, L) and using lemma 2, we get

(10) for any regular solution.

Next, we introduce the multiplier w given by the solution of the Dirichlet problem

(8)

Lemma 5 Let (ϕ, ψ, θ) be the solution of (1-2). Then the functional I1(t) = ρ1 R Ωχtwdx + ρ2 R Ωξtξdx, t ≥ 0. (11)

satisfies, for any positive constant ε1, the estimate d dtI1(t) ≤ − λ 2 R Ωξ 2 xdx + ε1 R Ωχ 2 tdx +  ρ2+ ρ2 1Cp 4ε1  R Ωξ 2 tdx 3γ2C p λb R Ωθ 2 tdx + 3g 2λ R Ω(goξx) (t)dx, t ≥ 0. . (12)

where Cp is the Poincar´e constant and λ is defined in (H1).

Proof : Differentiating I1(t) in (11) and using the first and second

equa-tions of (1-2), we obtain d dtI1(t) = ρ1 R Ωχtwtdx + k R Ωw 2 xdx + γ R Ωθtwxdx + ρ2 R Ωξ 2 tdx −bR Ωξ 2 xdx − k R Ωξ 2dx + γR Ωθtξdx + R Ω(g ∗ ξx) ξx(t) dx. Hence, thanks to R Ωw 2 tdx ≤ Cp R Ωw 2 txdx ≤ Cp R Ωψ 2 tdx,

and the Young inequality, we obtain, for all δ > 0,

d dtI1(t) ≤ ε1 R Ωχ 2 tdx + ρ2 1Cp 4ε1 R Ωξ 2 tdx + δ R Ωξ 2 xdx + γ2Cp 4δ R Ωθ 2 tdx +ρ2 R Ωξ 2 tdx − b R Ωξ 2 xdx + δ R Ωξ 2 xdx + γ2C p 4δ R Ωθ 2 tdx +R0tg (τ ) dτRξ2 xdx + R Ω(g♦ξx) ξx(t)dx. (13)

Next, exploiting the Young’s inequality for the last term in (13), and lemma 3, we obtain, for all δ > 0,

R Ω(g♦ξx) ξx(t)dx ≤ δ R Ωξ 2 xdx +4δ1 Rt 0 g(τ )dτ R Ω(goξx) (t)dx.

The choice of δ = λ6 gives the result.

Lemma 6 Let (ϕ, ψ, θ) be the solution of (1-2). Then the functional

I2(t) := −ρ1

RL

0 χtχdx − ρ2

RL

0 ξtξdx, t ≥ 0, (14)

satisfies, the estimate

(9)

Proof. A Straightforward computation, using the .rst and the second equations in (1-2), yields d dtI2(t) = −ρ1 R Ωχ 2 tdx − ρ2 R Ωξ 2 tdx + k R Ω(χx+ ξ) 2 dx +bRξx2dx − γRθt(χx+ ξ) dx − Rt 0 g(τ )dτ R Ωξ 2 xdx −R Ω(g♦ξx) ξx(t) dx.

Thanks to Young.s inequality, we get

−γR Ωθt(χx+ ξ) dx ≤ γ2 2k R Ωθ 2 tdx + k2 R Ω(χx+ ξ) 2 dx −R Ω(g♦ξx) ξx(t) dx ≤ b 2 R Ωξ 2 xdx +2b1 R Ω(g♦ξx) 2 dx.

Finally, it remains to conclude by lemma 3.

Lemma 7 Let (ϕ, ψ, θ) be the solution of (1-2). Then the functional

I3(t) := −ρ2

R

Ωξt(g♦ξ) (t) dx, t ≥ 0. (16)

satisfies, for any positive constant ε2 and δ1 > 0, the estimate

I30 (t) ≤ −ρ2 Rt 0 g (τ ) dτ − δ1  RL 0 ξ 2 tdx + ε2(1 + g) R Ωξ 2 xdx +γ2 Rθ2 xdx + kε2 R Ω(χx+ ξ) 2 dx + C (ε2) R Ω(goξx) dx −ρ22 4δ1g(0)Cp R Ω(g 0 x) (t)dx, t ≥ 0. (17) where C(ε2) = h ε2+12 + b 2 4ε2 + kCp 4ε2 + γCp 2  gi.

Proof. Differentiating the functional I3(t) and using the second equation

(10)

We now estimate the terms in this last identity, using Young’s inequality and lemma 3, we obtain, for all δ1 > 0,

−ρ2 R Ωξt(g 0 ♦ξ) (t)dx ≤ δ1 R Ωξ 2 tdx + ρ2 2 4δ1 Rt 0(−g 0(τ ) dτ )R Ω(−g 0oξ) (t)dx ≤ δ1 R Ωξ 2 tdx + ρ2 2 4δ1g(0)Cp R Ω(−g 0 x) (t)dx

Similary, we have for ε2 > 0,

bRξx(g♦ξx) (t)dx ≤ ε2 R Ωξ 2 xdx + b2g 4ε2 R Ω(goξx) (t)dx, kR Ω(χx+ ξ) (g♦ξ) (t)dx ≤ kε2 R Ω(χx+ ξ) 2 dx +kCpg 4ε2 R Ω(goξx) (t)dx, −γR Ωθt(g♦ξ) (t)dx ≤ γ 2 R Ωθ 2 xdx + γCpg 2 R Ω(goξx) (t)dx, finally −R Ω  Rt 0 g (t − s) (ξx(s) ds)  (g♦ξx) (t)dx ≤ 12 R Ω(g♦ξx) (t)dx +ε2 2 R Ω  Rt 0 g (t − s) (ξx(t) − ξx(s) − ξx(t) ds) 2 dx ≤ ε2  Rt 0 g(τ )dτ  R Ωξ 2 x(t) dx +  ε2 +12  R Ω(g♦ξx) (t)dx ≤ ε2  Rt 0 g(τ )dτ  R Ωξ 2 x(t) dx +  ε2+1 2  gR Ω(goξx) (t)dx.

Combining all the above estimates, we prove the assertion of the lemma.

Lemma 8 Let (ϕ, ψ, θ) be the solution of (1-2). Then the functional

I4(t) := ρ3 R Ωθtθdx + γ R Ω(χx+ ξ) θdx + β 2 R Ωθ 2dx, (18)

satisfies, for any positive constant ε3, the estimate

I40 (t) ≤ −lR0Lθ2 xdx +  γ2 4kε3 + ρ3  RL 0 θ 2 tdx + kε3 RL 0 (χx+ ξ) 2 dx. (19)

Proof. Differentiating I4(t) and using the third equation in (1-2), we

(11)

By using Young’s inequality, we obtain for any ε3 > 0, I40 (t) ≤ −lRθ2xdx +  γ2 4kε3 + ρ3  R Ωθ 2 tdx + kε3 R Ω(χx+ ξ) 2 dx, which is exactly (19).

Lemma 9 Assume that the coefficients ρ1, ρ2, ρ3, b, k and γ satisfy the

rela-tion (9). Then the funcrela-tional I5(t) := ρ2 R Ωξt(χx+ ξ) dx + (ρ2− γ) R Ωχtξxdx −γk +ρ3 ρ2  R Ωθx(g ∗ ξx) (t) dx − ρ1 k R Ωχt(g ∗ ξx) (t) dx +ρ3 R Ωξtθtdx + l R Ωθxξxdx, (20)

satisfies, for any positive constant ε3, the estimate

I50 (t) ≤ −k2R(χx+ ξ) 2 dx + ε3 R Ωχ 2 tdx + ε3 R Ωθ 2 xdx + ρ2 R Ωξ 2 tdx + [χx(bξx− (g ∗ ξx) (t))]x=Lx=0 + C1(ε3) R Ωθ 2 tdx − g 2ε3C2 R Ω(g 0 x) (t) dx +g 2(0) 2ε3 C2 R Ωξ 2 xdx, ∀t ≥ 0, (21) where C1(ε3) = γ 2 k + β2 4γ + ρ3γ ρ2 + ρ23k ρ2 2 and C2 =   γ k + ρ3 ρ2 2 + ρ1 k 2  .

Proof. A straifhtfoward computation, using the second, the third equa-tions in (1-2) and integrating by parts, yields

I50(t) = −bRξx(χx+ ξ)xdx − k R Ω(χx+ ξ) 2 dx +γR Ωθt(χx+ ξ) dx + R Ω(g ∗ ξx) (t) (χx+ ξ)xdx + [χx(bξx− (g ∗ ξx) (t))] x=L x=0 + ρ2 R Ωξ 2 tdx +ρ2 R Ωξtχtxdx + (ρ2− γ) R Ωχttξxdx + (ρ2− γ) R Ωχtξtxdx −γk +ρ3 ρ2  R Ωθtx(g ∗ ξx) (t) dx −  γ k + ρ3 ρ2  R Ωθx(g ∗ ξx)t(t) dx −ρ1 k R Ωχtt(g ∗ ξx) (t) dx − ρ1 k R Ωχt(g ∗ ξx)t(t) dx −ρ3b ρ2 R Ωξxθtxdx − ρ3k ρ2 R Ωθt(χx+ ξ) dx + ρ3γ ρ3 R Ωθ 2 tdx +ρ3 ρ2 R Ωθtx(g ∗ ξx) (t) dx − l R Ωθxξtxdx − β R Ωξtθtdx −γR Ωξtχtxdx − γ R Ωξ 2 tdx + l R Ωθtxξxdx + l R Ωθxξtxdx,

(12)

By using Young.s inequality, relation (9) and the properties of g, the result follows.

As in [24], to deal with the boundary terms appearing in (21), we define the function

q(x) = 2 − 4xL, x ∈ Ω

Lemma 10 Let (ϕ, Ψ, u) be the solution of (1-2). The functionals J1 and J2

defined by J1(t) := εk3  ρ1R Ωχtq (x) χxdx + ρ3 R Ωθtq (x) θxdx +γR Ωχxq (x) θxdx,  (22) J2(t) := ρ2 3 R Ωξtq (x)  bξx− Rt 0 g (t − s) ξx(s) ds  dx, (23) satisfy, for any ε3 > 0, the estimates

J10(t) ≤ −ε3[χ2x(L) + χ2x(0)] + 2ε3ρ1 kL R Ωχ 2 tdx +ε3 L2 + 1  R Ωχ 2 xdx + ε3 k 2l L + β + γ  R Ωθ 2 xdx ε3 R Ωξ 2 xdx + ε3 k 2ρ3 L + β  R Ωθ 2 tdx + ε3γ k R Ωξ 2 tdx (24) and J20 (t) ≤ −1 3   bξx(L) − Rt 0g (t − s) ξx(L, s) ds 2 − 1 4ε3   bξx(0) − Rt 0 g (t − s) ξx(0, s) ds 2 C3(ε3) R Ωξ 2 xdx + kε3 R Ω(χx+ ξ) 2 dx +ρ2 2ε3 1 + 1 L  R Ωξ 2 tdx + γ2L 4ε3 R Ωθ 2 tdx +3g 3 + kg 4ε3 3  R Ω(goξx) dx − ρ2g(0) 4ε3 R Ω(g 0 x) dx, (25) where C3(ε3) =  1 2Lε3 + 1 Lε3 + k 8ε3 3  (b2+ 2g2) + ρ2 4ε3g 2(0).

Proof. Differentiating J1(t) and using the first and third equations in

(13)

Thanks to Young’s inequality, we obtain the desired result.

Next, from the second equation in (1-2) and integrating by parts, we get

J20 (t) := 2Lε1 3 R Ω  bξx− Rt 0g (t − s) ξx(s) ds 2 dx − 1 4ε3   bξx(L) − Rt 0 g (t − s) ξx(L, s) ds 2 − 1 4ε3   bξx(0) − Rt 0 g (t − s) ξx(0, s) ds 2 + γ 4ε3 R Ωθtq (x)  bξx− Rt 0 g (t − s) ξx(s) ds  dx − k 4ε3 R Ω(ϕx+ ξ) q (x)  bξx− Rt 0 g (t − s) ξx(s) ds  dx +ρ2 4ε3 R Ωξtq (x)  bξx− Rt 0 g (t − s) ξx(s) ds  tdx. (26)

Next, by using Young’s inequality, lemma 3, lemma 4 and the fact that q2(x) ≤ 4, ∀x ∈ [0, L] , we estimate the terms in (26) as follows,

(14)

By combining all the above estimates, we estabilish the assertion of the lemma.

Proof of Theorem. For some positive constants; N , N1, N2, to be chosen

appropriately later. We define the Lyapunov functional by

£ (t) = N E + N1I1(t) +14I2(t) + N2I3(t)

+N3I4(t) + I5(t) + J1(t) + J2(t) , ∀t > 0.

(27)

Next, taking into account (10), (12), (15), (17), (19), (21), (24), (25) and the following relations RL 0 χ 2 xdx ≤ 2 RL 0 (χx+ ξ) 2 dx + 2Cp RL 0 ξ 2 xdx, h χx  bξx− Rt 0g (t − s) ξx(x, s) ds ix=L x=0 ≤ ε3[χ 2 x(L) + χ2x(0)] +1 3   bξx(L) − Rt 0g (t − s) ξx(L, s) ds 2 +1 3   bξx(0) − Rt 0 g (t − s) ξx(0, s) ds 2 , we obtain £ (t) ≤ −hλN1 2 − 3b 8 − N2ε2(1 + g) − g2(0) 2ε3 C2− ε3− C3(ε3) − 2ε3 2 L+ 1 Cp i R Ωξ 2 xdx −hN2  ρ2 Rt 0 g (τ ) dτ − δ1  − 5ρ2 4 − N1  ρ2+ ρ2 1Cp 4ε1  − ε3γ k − ρ2 2ε3 1 + 1 L iRL 0 ξ 2 tdx −hN β −γ8k2 − N3  γ2 4kε3 + ρ3  −3N1γ2Cp λb − C1(ε3) − ε3 k 2ρ3 L + β − γ2L 4ε3 i R Ωθ 2 tdx −k 8 − kε2N2− ε3 N3k + k + 2 2 L + 1  R Ω(χx+ ξ) 2 dx −ρ1 4 − N1ε1− ε3 1 + 2ρ1 kL  R Ωχ 2 tdx −N3l −γ2N2− εk3 2lL + β + γ + 1  RL 0 θ 2 xdx +h3gN1 2λ + g 8b + N2C (ε2) +  3g Lε3 + kg 4ε3 3 i R Ω(goξx) (t)dx +hN 2 − N2 ρ2 2 4δ1g(0)Cp− ρ2g(0) 4ε3 − g 2ε3C2 i R Ω(g 0 x) dx (28) Since g is continuous, positive and g (0) > 0, then for any t0 > 0, we

have Z t 0 g (s) ds ≥ Z t0 0 g (s) ds = g0, ∀t ≥ t0.

Now all the terms in the right-hand side of (28) become negative if we select carefully our constants. First, let us take δ1 = 4Nk

(15)

enough so that

ε1 ≤

ρ1

8N1

,

Second, by we select N1 large enough so that

λN1 4 − 3b 8 − g2(0) 2ε3 C2− C3(ε3) − 2ε3  2 L+ 1 + 1 2Cp  Cp > 0,

then select ε3 so small so that

ε3 ≤ min " ρ1 8  1 + 2ρ1 kL −1 , k 16  N3k + k + 2  2 L+ 1 −1# .

Next, we pick N3 large enough so that

N3l − γ 2N2− ε3 k  2l L + β + γ + 1  > 0,

and we choose N2 large enough so that

N2g0− 1 4 − 5ρ2 4 − N1  ρ2+ ρ2 1Cp 4ε1  − ε3γ k − ρ2 2ε3  1 + 1 L  > 0

and ε2 so small so that

ε2 < min  1 8N2 , λN1 4N2(1 + g) 

Finally, we choose N large enough so that

N β −γ 2 8k− N3  γ2 4kε3 + ρ3  −3N1γ 2C p λb − C1(ε3) − ε3 k  2ρ3 L + β  −γ 2L 4ε3 > 0, and N 2 − N2 ρ22 4δ1 g(0)Cp− ρ2g(0) 4ε3 − g 2ε3 C2 > 0

Therefore, (28) takes the form

£0(t) ≤ −k0E (t) + c (goξx) (t) , ∀t ≥ t0, (29)

for two positive constants k0 and c.

(16)

ζ (t) £0(t) ≤ −k0ζ (t) E (t) + cζ (t) (goψx) (t)

≤ −k0ζ (t) E (t) − c1(g0oξx) (t)

≤ −k0ζ (t) E (t) − c1E0(t)

[ζ (t) £ (t) + cE (t)]0− ζ0(t) E (t) ≤ −k0ζ (t) E (t) , ∀t ≥ t0.

Using the fact that ζ0(t) ≤ 0, we have

(ζ (t) £ (t) + cE (t))0 ≤ −k0ζ (t) E (t) , ∀t ≥ t0.

Again, by noting that

R (t) = ζ (t) £ (t) + cE (t) ∼ E (t) . we obtain for some positive constant α,

R0(t) ≤ −αζ (t) R (t) , ∀t ≥ t0. (30)

A simple integration of (30) over (t0, t) leads to

R (t) ≤ R (0) e−c1

Rt1

t0 ζ(s)ds, ∀t ≥ t0. (31)

Finally, the assertion of Theorem 11 is then obtained by virtue of the bound-edness of E and ζ and the fact that R ∼ E.

Acknowledgment

The authors are grateful to Prof. Said Mazouzi from the University of Annaba, Algeria, for his helpful comments.

References

[1] A. Djebabla and N. Tatar, Exponential stabilization of the Timoshenko system by a thermo-viscoelastic damping, J. Dyn. Control Syst. 16 (2010), no. 2, 189210.

[2] A. Djebabla and N. Tatar, Stabilization of the Timoshenko beam by thermal effect, Mediterr. J. Math. 7 (2010), no. 3, 373385.

(17)

[4] [3] L.H. Fatori, J.E. Munoz Rivera, Energy decay for hyperbolic ther-moelastic systems of memory type, Quart. Appl. Math. 59 (3) (2001) 441458.

[5] M.A. Goodman, S.C. Cowin, A continuum theory for granular materials, Arch. Ration. Mech. Anal. 44 (1972) 249266.

[6] J.W. Nunziato, S.C. Cowin, A nonlinear theory of elastic materials with voids, Arch. Ration. Mech. Anal. 72 (1979) 175201.

[7] R. Quintanilla, Slow decay in one-dimensional porous dissipation elas-ticity, Appl. Math. Lett. 16 (2003) 487491.

[8] P.S. Casas, R. Quintanilla, Exponential decay in one-dimensional porous-thermo-elasticity, Mech. Res. Comm. 32 (2005) 652658.

[9] P.S. Casas, R. Quintanilla, Exponential stability in thermoelasticity with microtemperatures, Internat. J. Engrg. Sci. 43 (2005) 3347.

[10] M. Ciarletta and D. Iesan, Non-classical elastic solids. Pitman Research Notes in mathematics Series 293. New York (1993).

[11] A. Maga˜na, R. Quintanilla, On the time decay of solutions in one-dimensional theories of porous materials, Internat. J. Solids Structures 43 (2006)34143427.

[12] P.X. Pamplona, J.E. Mu˜noz Rivera, R. Quintanilla, Stabilization in elas-tic solids with voids, J. Math. Anal. Appl. 350 (2009) 3749.

[13] A. Maga˜na, R. Quintanilla, On the time decay of solutions in porous-elasticity with quasi-static microvoids, J. Math. Anal. Appl. 331 (#1) (2007) 617630.

[14] J.E. Mu˜noz Rivera, R. Quintanilla, On the time polynomial decay in elastic solids with voids, J. Math. Anal. Appl. 338 (2008) 12961309.

[15] A. Soufyane, Energy decay for porous-thermo-elasticity systems of mem-ory type, Appl. Anal. 87 (#4) (2008) 451464.

(18)

[17] A. Soufyane, M. Afilal, M. Aouam, M. Chacha, General decay of solu-tions of a linear one-dimensional porous-thermoelasticity system with a boundary control of memory type, Nonlinear Anal. 72 (2010) 39033910.

[18] P.X. Pamplona, J.E. Mu˜noz Rivera, R. Quintanilla, On the decay of solutions for porous-elastic systems with history, J. Math. Anal. Appl. 379 (2) (2011) 682705.

[19] S.C. Cowin, J.W. Nunziato, Linear elastic materials with voids, J. Elas-ticity 13 (1983) 125147.

[20] S.C. Cowin, The viscoelastic behavior of linear elastic materials with voids, J. Elasticity 15 (1985) 185191.

[21] S. A. Messaoudi, T. A. Aparala, General stability result in a memory-type porous thermoelasticity system of memory-type III. Arab J Math Sci 20(2) (2014), 213232.

[22] D. Iesan, A theories of thermoelastic materials with voids. Acta Me-chanica 60, (1986) 67-89

[23] D. Ieasan, R. Quintanilla, On thermoelastic bodies with inner structure and microtemperatures, J. Math. Anal. Appl. 354 (#1) (2009) 1223.

[24] A. Maga˜na, R. Quintanilla, On the exponential decay of solutions in one-dimensional generalized porous-thermo-elasticity, Asymptot. Anal. 49 (2006) 173187.

Références

Documents relatifs

Notre étude a montré que les femmes avec fractures vertébrales étaient plus âgées, plus maigres, avec une parité plus importante, une durée de ménopause plus longue

We show that if the damping vanishes as a H¨ older function |x| β , and in addition, the boundary of the damped region is strictly convex, the wave is stable at rate t −1+ 2β+7 2

In this paper, we consider global weak solutions to com- pressible Navier-Stokes-Korteweg equations with density dependent vis- cosities, in a periodic domain Ω = T 3 , with a

The first work to deal with model consistency for a large class of functions without the irrepresentable condition assumption is (Fadili et al., 2017), which in- troduces of the

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des

1. Fibich [7] noted that in the nonlinear optics context, the origin of the nonlinear damping is multiphoton absorption. Damped Nonlinear Schr¨ odinger Equation, Blow-up,

The choice of a viscoelastic material will be based on its properties (loss factor and Young modulus), in the frequency and temperature range of interest for the application.. But

On the other hand, in Section 4, we prove that under some restrictions on the initial data and if the interior source dominates the boundary damping then the L p -norm of the