Abstract
We consider a fluid–structure interaction system composed by a three-dimensional viscous incompressible fluid and an elastic plate located on the upper part of the fluid boundary. The fluid motion is governed by the Navier–Stokes system whereas we add a damping in the plate equation. We use here Navier-slip boundary conditions instead of the standard no-slip boundary conditions. The main results are the local in time existence and uniqueness of strong solutions of the corresponding system and the global in time existence and uniqueness of strong solutions for small data and if we assume the presence of frictions in the boundary conditions.
We’re sorry, something doesn't seem to be working properly.
Please try refreshing the page. If that doesn't work, please contact support so we can address the problem.
Avoid common mistakes on your manuscript.
1 Introduction
The aim of this work is to analyze the interaction between a viscous incompressible fluid and a viscous elastic plate. Let us start by presenting the corresponding model. We denote by \(\omega \) the rectangular torus
For any function \(\eta : \omega \rightarrow (-1,\infty )\), we define (see Fig. 1)
In particular
We consider the following system describing the evolution of the fluid governed by the incompressible Navier–Stokes equations, and the movement of the elastic plate
In the above system, we have denoted by U the fluid velocity, P the fluid pressure and \(\eta \) the transversal plate displacement.
The Cauchy stress tensor \(\mathbb {T}(U,P)\) is defined by
The function \(\widetilde{\mathbb {H}}_{\eta }\) is the fluid strain on the structure and is defined by
We assume
These constants correspond respectively to the rigidity (\(\alpha \)), the stretching (\(\kappa \)), the damping on the structure (\(\delta \)) and the viscosity (\(\nu \)).
We have denoted by n the unitary exterior normal of \(\partial \Omega (\eta )\):
and on \(\Gamma (\eta )\):
Here and in what follows, \(|\cdot |\) denotes the Euclidian norm of \(\mathbb {R}^k\), \(k\geqslant 1\).
We complete (1.3) by the Navier slip boundary conditions. In order to write these boundary conditions, we need to introduce some notations. We denote by \(a_n\) and \(a_\tau \) the normal and the tangential parts of \(a\in \mathbb {R}^3\):
Then, our boundary conditions write as follows
In what follows, we write the above equations in the following more compact way
We assume that the friction coefficients \(\beta _1\) and \(\beta _2\) are constants satisfying
These boundary conditions can be compared with the standard no-slip boundary conditions usually considered with the Navier–Stokes system. In our case, these conditions would write as
The Navier slip boundary condition was proposed by Navier in 1823 [28] and is relevant in several physical contexts, see for instance [22, 24, 35].
To complete the system (1.3), (1.8), we add the following initial conditions
Let us remark that we don’t need to consider boundary conditions on the “lateral” boundaries since we work with the torus \(\omega \) [see (1.1) and (1.2)]. This means that we are considering periodic boundary conditions for U, P and \(\eta \):
and a similar relations for P.
Several works have been devoted to the study of the system (1.3), (1.10) with the Dirichlet boundary conditions (1.9): existence of strong solutions [3, 23], feedback stabilization [2, 30], global existence of strong solutions [15]. Let us point out that in this latter work, the authors manage to obtain in particular that there is no contact between the plate and the bottom of the domain in finite time for the system (1.3), (1.9), (1.10). This result, as previous works on fluid–structure interaction systems, shows that the standard no-slip boundary conditions may lead to some paradoxal results as the distance between two structures is going to 0: in the case of rigid bodies immersed into a viscous incompressible fluid, it is shown that in particular geometries there is no contact in finite time of two structures [18, 19] and in general, if there is contact, then it occurs with null relative velocity and null relative acceleration [31]. In [9, 10], the author considered boundary conditions involving the pressure. Here, our aim is to analyze the same system (1.3) with the Navier-slip boundary conditions (1.8) instead of the Dirichlet boundary conditions. Such a system was already considered in [17, 27] where the existence of weak solutions is proved in dimension 2 (global existence as long as the deformable structure does not touch the fixed bottom). The uniqueness of weak solutions for this system has been obtained in [16].
Our objective is to prove the existence and uniqueness of strong solutions for small time or for small data. This is the first work on strong solutions for such a system in the case of Navier-slip boundary conditions and to our knowledge, it is also the first work on strong solutions for this kind of systems in the 3D case.
In the case where the structures are rigid bodies immersed into a viscous incompressible fluid, several authors have already considered the Navier-slip boundary conditions: existence of weak solutions [12, 29], existence of contact in finite time [13], existence of strong solutions and study of contacts in finite time [36], uniqueness of weak solutions [7]. Let us also mention the work of [8] where they consider a nonlinear boundary condition of Tresca’s type.
The main result of this article is
Theorem 1.1
-
1.
Assume \(\beta _i\geqslant 0\) for \(i=1,2\) and (1.4). Suppose \(\eta ^0\in H^{3}(\omega )\), \(\eta ^1\in H^1(\omega )\) and \(U^0\in [H^1(\Omega (\eta ^0))]^3\) such that
$$\begin{aligned} 1+\eta ^0>0, \quad \nabla \cdot U^0=0 \quad \text {in} \ \Omega (\eta ^0), \quad (U^0-\eta ^1 e_3)_{n}=0 \quad \text {on} \ \Gamma (\eta ^0), \quad U^0_{n}=0 \quad \text {on} \ \Gamma _0. \end{aligned}$$There exists a time \(T_0\) such that the system (1.3), (1.8), (1.10) admits a unique strong solution \((U,P,\eta )\) on \((0,T_0)\):
$$\begin{aligned}&\displaystyle \eta \in L^2(0,T_0;H^4(\omega ))\cap C^0([0,T_0];H^3(\omega )) \cap H^1(0,T_0;H^2(\omega ))\cap C^{1}([0,T_0];H^1(\omega ))\cap H^2(0,T_0;L^2(\omega )),&\\&\displaystyle U\in L^2(0,T_0;[H^2(\Omega (\eta (t))]^3) \cap C^0([0,T_0];[H^1(\Omega (\eta (t)))]^3)\cap H^1(0,T_0;[L^2(\Omega (\eta (t)))]^3),&\\&\displaystyle \nabla P \in L^2(0,T_0;[L^2(\Omega (\eta (t)))]^3).&\end{aligned}$$ -
2.
Assume \(\beta _i\geqslant 0\) for \(i=1,2\) with \(\beta _{1}+\beta _{2}>0\) and (1.4). There exist \(\gamma _0>0\) and \(R_0>0\) such that if \(\eta ^0\in H^{3}(\omega )\), \(\eta ^1\in H^1(\omega )\) and \(U^0\in [H^1(\Omega (\eta ^0))]^3\) satisfy
$$\begin{aligned} 1+\eta ^0>0, \quad \nabla \cdot U^0=0 \quad \text {in} \ \Omega (\eta ^0), \quad (U^0-\eta ^1 e_3)_{n}=0 \quad \text {on} \ \Gamma (\eta ^0), \quad U^0_{n}=0 \quad \text {on} \ \Gamma _0. \end{aligned}$$and
$$\begin{aligned} \left\| U^0\right\| _{[H^1(\Omega )]^3} +\left\| \eta ^0\right\| _{H^3(\omega )}+\left\| \eta ^1\right\| _{H^1(\omega )}\leqslant R_0, \end{aligned}$$then the system (1.3), (1.8), (1.10) admits a unique strong solution \((U,P,\eta )\) on \((0,\infty )\):
$$\begin{aligned}&\eta \in L^2_\gamma (0,\infty ;H^4(\omega ))\cap BC^0_\gamma ([0,\infty ];H^3(\omega )) \cap H^1_\gamma (0,\infty ;H^2(\omega ))\cap BC^{1}_\gamma ([0,\infty ];H^1(\omega ))\cap H^2_\gamma (0,\infty ;L^2(\omega )), \\&U\in L^2_\gamma (0,\infty ;[H^2(\Omega (\eta (t))]^3) \cap BC^0_\gamma ([0,\infty ];[H^1(\Omega (\eta (t)))]^3)\cap H^1_\gamma (0,\infty ;[L^2(\Omega (\eta (t)))]^3), \\&\nabla P \in L^2_\gamma (0,\infty ;[L^2(\Omega (\eta (t)))]^3), \end{aligned}$$for \(\gamma \in [0,\gamma _0]\).
In the above statement, the spaces \(L^p\), \(H^s\) are the classical Lebesgue, Sobolev spaces. We use the notation \(BC^0=C^0\cap L^\infty \) and \(BC^1=C^1\cap W^{1,\infty }\). The notation \(\cdot _\gamma \) is explained below in (2.2), (2.3) and corresponds to an exponential decay of order \(\gamma \). Finally, the notation \(L^2(0,T;H^1(\Omega (\eta (t))))\) corresponds to the fact that the fluid velocity and pressure are written in a moving domain depending on \(\eta \). To obtain our result, we thus need to use a change of variables for U and P and the fluid velocity and pressure after change of variables are obtained in spaces of the form \(L^2(0,T;H^1(\Omega ))\) with a fixed \(\Omega \). The precise definition of strong solutions is given in Sect. 3 (Definition 3.1) and we reformulate the above result in a more precise way in Theorem 6.1.
Remark 1.2
We can write a bi-dimensional version of the system (1.3), (1.8), (1.10) and for such a system, one can prove a similar result as Theorem 1.1. In fact, in that case, one could obtain a global in time existence of strong solutions up to a possible contact between the beam and the bottom of the domain by following the arguments in [15].
Remark 1.3
For the sake of simplicity in the proof of Theorem 1.1 and in the remaining part of this article, we assume \(\kappa =\sigma =0\) since these constants do not play any role in the analysis.
The plan of this paper is as follows: In Sect. 2, we give some notation. In Sect. 3, we remap the problem into a fixed domain using a change of variables like it was introduced in [21], and we restate Theorem 1.1. We obtain some regularity properties of the Stokes system in domains of class \(H^3\) in Sect. 4. In Sect. 5, we study the linearized problem by writing it as an evolution equation. We prove in particular that the associated semigroup is analytic and in Sect. 6, we prove the main result using a fixed-point argument.
2 Notation
During the course of our analysis, we will use some functional spaces that we introduce in this section.
First, let us note that due to the incompressibility of the fluid and to the boundary conditions (1.8)\(_{1}\) and (1.8)\(_{3}\), we have
For simplicity, we assume throughout the paper that
so that
It yields to consider the following space
and the orthogonal projection \(M : L^2(\omega ) \rightarrow L^2_0(\omega )\). Applying M on the plate Eq. (1.3)\(_3\), we find
where
and
The projection of (1.3)\(_3\) onto \(L^2_0(\omega )^\perp \) leads to impose the choice of the constant normalizing the pressure, see for instance [15].
We denote by \(H^s(0,T;\mathfrak {X})\) the usual Sobolev spaces with values in a Banach space \(\mathfrak {X}\). For \(s>0\), \(s\notin \mathbb {N}\), the norm of these spaces can be defined by using
More precisely, the norm \( \left\| \cdot \right\| _{H^{s}(0,T;\mathfrak {X})}\) for \(s\in (0,1)\) is given by
We recall (see [6]) that if \(s\in \left( \dfrac{1}{2},1\right) \), then the norm \(\lfloor \cdot \rfloor _{s,2,(0,T),\mathfrak {X}}\) is equivalent to the norm defined in (2.1) in the space \(\left\{ \xi \in H^{s}(0,T;\mathfrak {X})\;|\;\xi (0)=0\right\} \).
Let \(\mathfrak {X}_1\), \(\mathfrak {X}_2\) be two Banach spaces endowed with the norm \(\left\| \cdot \right\| _{\mathfrak {X}_1}\) respectively \(\left\| \cdot \right\| _{\mathfrak {X}_2} \). For \(s\geqslant 0\), we define the following space
endowed with norm
For \(s=1\), we will denote \(W^1(0,T;\mathfrak {X}_1,\mathfrak {X}_2)\) by \(W(0,T;\mathfrak {X}_1,\mathfrak {X}_2)\).
For \(\gamma >0\), we also consider the spaces
and
For these spaces, we use the norms defined by
In what follows, we set
for the local existence and
for the global existence.
In order to differentiate the normal or the normal and tangential component of a vector v in \(\Omega \) and in \(\Omega (t)\), we use the notation \(n_0\), \(v_{n_0}\) and \(v_{\tau _0}\) for the configuration \(\Omega \).
We denote by
the space of infinitely differentiable functions with free divergence in \(\Omega \) with compact support .
Let us also define the following space
endowed with the norm
If \(T=+\infty \) and \(\gamma \geqslant 0\), we will write
endowed with the norm
To write the boundary conditions, we also introduce the operator \(\mathcal {T}\) defined as follows (see [2]):
We can verify that \(\mathcal {T}_{\eta ^0}\in \mathcal {L}(L^2(\omega );[L^2(\partial \Omega )]^3)\) and that
We set
We also define
3 Change of variables
For \(\eta ^1, \eta ^2 \in H^{3}(\omega )\) with
we can consider the change of variables \(X_{\eta ^1,\eta ^2}\) defined below
The mapping \(X_{\eta ^1,\eta ^2}\) is invertible of inverse \( X_{\eta ^2,\eta ^1}\). Moreover, using the Sobolev embedding \(H^3(\omega ) \hookrightarrow C^1(\overline{\omega })\) and that
we deduce that \(X_{\eta ^1,\eta ^2}\) is a \(C^1\)-diffeomorphism from \(\Omega (\eta ^1)\) onto \(\Omega (\eta ^2)\).
In the case \(\Omega =\Omega (\eta ^0)\) [see (2.4)], we set
and in the case \(\Omega =\Omega (0)\) [see (2.5)], we set
We have in both cases that \(Y(t,\cdot )=\left[ X(t,\cdot )\right] ^{-1}\).
We consider the following transformation of u and p:
Here, \(({\text {Cof}}\nabla X(t,y))^*\) denotes the transpose of \(({\text {Cof}}\nabla X(t,y))\). After some standard calculations (see, for instance, [21]), the system (1.3), (1.8), (1.10) can be written as
with the boundary conditions
and with the initial conditions
In order to write the nonlinearities F, H, G, we first set
Then
and
To define G, we introduce the following notations.
and
Then \(G(u,\eta )\) is given by
More precisely, let us note that
writes as
The formula (3.14) for G is such that
so that (3.16) is equivalent to the second condition of (3.6), with G tangential.
Using the above transformation, we can now introduce our definition of strong solutions for system (1.3), (1.8), (1.10)
Definition 3.1
The triplet \((U,P,\eta )\) is a strong solution of (1.3), (1.8), (1.10) if the following conditions are satisfied
Following this definition, in order to prove Theorem 1.1, we have to prove the existence and uniqueness of
solution of the system (3.5), (3.6), (3.7) and satisfying (D2).
4 Regularity properties of the Stokes system
In this section, we obtain some results on the stationary system in \(\Omega (\eta )\) for \(\eta =\eta ^0\) [see (2.4)] or for \(\eta =0\) [see (2.5)]:
Let define the following space
We give the definition of a weak solution of the system (4.1).
Definition 4.1
We say that \((\overline{u},\overline{p})\) is a weak solution of (4.1) if \((\overline{u},\overline{p})\in [ H^1(\Omega (\eta ))]^3\times L^2(\Omega (\eta ))/\mathbb {R}\), if \(\nabla \cdot \overline{u}=\overline{g}\) in \(\Omega (\eta )\), \(\overline{u}_{n}=\overline{a}\) on \(\partial \Omega (\eta )\) and if the following variational equation is satisfied:
for all \(\phi \in H^1_{\tau }\).
We have the following result
Theorem 4.2
Assume \(\beta \geqslant 0\) and \(\alpha \geqslant 0\) with \(\beta _1+\beta _2>0\) or \(\alpha >0\). Let \(\eta \in H^3(\Omega (\eta ))\) and \(\delta _0>0\) such that \(1+\eta >\delta _0\) on \(\omega \). For any
such that
there exists a unique weak solution \((\overline{u},\overline{p})\in [H^1(\Omega (\eta ))]^3\times L^2_0(\Omega (\eta ))\) to the Stokes system (4.1). Moreover, we have the following estimates:
where C is a constant which depends on \(\left\| \eta \right\| _{H^3(\omega )} \) and \(\delta _0\).
Moreover, if
such that (4.2) holds, then \((\overline{u},\overline{p})\in [H^2(\Omega (\eta ))]^3\times (H^1(\Omega (\eta ))\cap L^2_0(\Omega (\eta )))\) and we have the following estimates:
where C is a constant which depends on \(\left\| \eta \right\| _{H^3(\omega )} \) and \(\delta _0\).
In the case where \(\eta \in C^{1,1}(\omega )\) such a result is already known, see [1] (see also [4]). Here, we manage to obtain the result for \(\eta \in H^3(\omega )\) by following an idea of [14, 15].
Proof of Theorem 4.2
The proof follows closely the proof of Lemma 6 in [15]. We assume here \(\beta _1+\beta _2>0\) and \(\alpha =0\), the proof is similar with \(\alpha >0\).
First, we write the system (4.1) in the domain
by using the change of variables \(X_{0,\eta }\) defined by (3.1). Then we set
and we define
Then system (4.1) is transformed into the following system
where N is defined by (1.5) and \(n_0\) is the unit exterior normal to \(\Omega \) (that is \(\pm e_3\)).
Since \(\eta \in H^3(\omega )\), we deduce that
for all \(s\geqslant 0\) and the corresponding norms depend on \(\left\| \eta \right\| _{H^3(\omega )} \) and \(\delta _0\). Moreover, using the embeddings \(H^1(\omega )\hookrightarrow L^p(\omega )\) for all \(p\geqslant 1\) and \(H^2(\omega )\hookrightarrow L^\infty (\omega )\), we deduce that it is sufficient to prove that the solution of (4.6) satisfies
Step 1: Weak solutions. Let note that the solution of (4.6) verifies
Let \(\lambda >0\) and consider the following system
with
To simplify the notations, we set
We define
We look for weak solutions to the system (4.8). Let \(f\in V'\), \(g\in L^2(\Omega )\), \(a\in [H^{1/2}(\partial \Omega )]^3\) and \(b\in [H^{-1/2}(\partial \Omega )]^3\). We have \(B_\eta \nabla \left( \frac{g}{\det (\nabla X_{0,\eta })}\right) \in V'\):
Therefore \(\widetilde{f}\in V'\) and we multiply the first equation of (4.8) by \(v \in V\) and the second equation of (4.8) by \(\psi \in L^2(\Omega )\) to obtain
We consider a lifting w satisfying
In order to this, we use [4, Corollary 8.2] and (4.2) to deduce the existence of \(\overline{w}\in [H^1(\Omega )]^3\) such that
Then \(w=(B_{\eta }^*)^{-1}\overline{w}\) satisfies (4.10) and the estimate
We set \(u=\widehat{u}+w\). Then, a couple (u, p) is a weak solution of the system (4.8) if and only if \((\widehat{u},p)\) verifies the following variational formulation
We have that
and writing
we deduce
with \(\overline{v}\cdot n=0\) on \(\partial \Omega (\eta )\). Applying a Korn inequality (see Proposition 4.5 below):
Hence, we can apply the Lax–Milgram theorem and using (4.11), we deduce the existence of a unique solution of \((u,p)=(u_\lambda ,p_\lambda )\in [H^1(\Omega )]^3\times L^2(\Omega )\) for (4.8) which verifies the estimates
Taking \(\psi =0\) and \(v\in [H^1_0(\Omega )]^3\) in (4.9), we obtain
This shows that \(\nabla p\in [H^{-1}(\Omega )]^3\) and using standard result (see, for instance [4, Proposition 1.1]), we deduce
Then, combining (4.15), (4.16) and (4.11), we obtain the estimate independent of \(\lambda \):
We can thus pass to the limit as \(\lambda \rightarrow 0\) in (4.8) to obtain a weak solution (u, p) of (4.6). To prove uniqueness, let us consider \((\overline{u}^*,\overline{p}^*)\) another weak solution corresponding to the same data. It follows that \(\overline{u}-\overline{u}^*\in H^1_\tau \), \(\nabla \cdot (\overline{u}-\overline{u}^*)\). Then, from Definition 4.1, we obtain
Thus, using Proposition 4.5, we get \(\overline{u}=\overline{u}^*\) in \(\Omega (\eta )\).
Step 2: Strong solutions. We use an argument developed in [14, 15]: if we approximate \(\eta \) by \(\eta _\varepsilon \in C^{1,1}(\omega )\), and the corresponding \(u_\varepsilon \), \(p_\varepsilon \) are \(H^2\) and \(H^1\). We show below that their norms depend only on the \(H^3\) norm of \(\eta _\varepsilon \) so that we can pass to the limit. To simplify, we do not write any \(\varepsilon \) below.
We first differentiate system (4.6) with respect to \(y_1\) and \(y_2\) to obtain a similar problem as (4.6) with source and boundary terms corresponding to the differentiates of f, g, a, b, \(A_\eta \) and \(B_\eta \). We only need to estimate the terms coming from \(A_\eta \) and \(B_\eta \), that is
Here we use a nice idea proposed in [14, 15]: we estimate the above terms by using the \(H^2\) regularity of u and the \(H^1\) regularity of p. More precisely, using the embeddings \(H^{1/2}(\omega )\subset L^4(\omega )\) and \(H^{1/4}(\omega )\subset L^{8/3}(\omega )\), we deduce that the above terms are estimated by
Using the first part of this proof and in particular (4.17), we obtain for \(i=1,2\)
We differentiate (4.6)\(_2\) with respect to \(y_3\), we obtain
Then, going back to (4.6)\(_1\), we also obtain
Since \(A_{33}=\frac{1}{1+\eta }(1+(y_3\partial _{y_1}\eta )^2+(y_3\partial _{y_2}\eta )^2)>0\), we deduce
Combining this with (4.19), we deduce the result. \(\square \)
We also need the following theorem which is proved in [32].
Theorem 4.3
Assume \(\beta \geqslant 0\) with \(\beta _1+\beta _2>0\). Let \(\eta \in H^3(\omega )\) and \(\delta _0>0\) such that \(1+\eta >\delta _0\) on \(\omega \). Let us consider the following non stationary Stokes system:
There exists \(\gamma _0>0\) such that if
for some \(\gamma \in [0,\gamma _0]\). Then the problem (4.22) admits a unique solution which satisfies the estimate
where C is a positive constant.
We recall that the spaces \(W_{\gamma }^s(0,\infty ;X_1,X_2)\) and \(L^2_{\gamma }(0,\infty ;L^2(\Omega ))\) are defined by (2.3), (2.2).
Remark 4.4
In [32], the author assumes that \(\eta \) is more regular but such an assumption is only used to obtain a lift of the boundary condition by taking a stationary Stokes system of the form (4.1), see relation (75) in [32].
Note also that in [32], the condition (4.23) is replaced by the equivalent condition
Such an equivalence can be obtained by using the surjectivity of the trace operator (see [25, p.21, Theorem 2.3]).
We end this section by proving a Korn’s type inequality (that we used in the above proof).
Proposition 4.5
Assume \(\eta \in W^{1,\infty }(\omega )\). Assume that \(\beta _1+\beta _2\ne 0\). There exists a positive constant \(C>0\), such that
for all \(u\in [H^1(\Omega (\eta ))]^3\).
Proof
We first show by contradiction that
Assume \(u_k\in [H^1(\Omega (\eta ))]^3\) with
and
Using the classical Korn inequality (see, for instance, [20]), the above relations imply that \((u_k)_k\) converges weakly to some \(u\in [H^1(\Omega (\eta ))]^3\) with \(D(u)=0\) and \(\sqrt{\beta } u=0\) on \(\partial \Omega (\eta )\). In particular, see [34, Lemma 1.1 p.18], there exist \(a,b\in \mathbb {R}^3\), such that for any \(y\in \Omega (\eta )\), \(u(y)=a+b\wedge y\). Using that
we deduce that \(b=0\), then \(u=a\) in \(\Omega (\eta )\). Since \(\sqrt{\beta } u=0\) on \(\partial \Omega (\eta )\), we obtain that \(u=0\) in \(\Omega (\eta )\). Up to a subsequence \(u_k \rightarrow u\) strongly in \([L^2(\Omega (\eta ))]^3\) and thus from (4.27), we get \(\left\| u \right\| _{[L^2(\Omega (\eta ))]^3}=1\) which leads to a contradiction. In order to prove (4.25), we combine (4.26) and the classical Korn inequality (using that \(\Omega (\eta )\) is Lipschitz continuous). \(\square \)
5 Linear System
Let us consider a linearized system of (3.5), (3.6), (3.7):
with the boundary conditions
and with the initial conditions
Let us consider \((v,\pi )\) the solution of (4.22) associated with \(\widetilde{g}\). Then \(w=u-v\) and \(q=p-\pi \) satisfy
with the boundary conditions
and with the initial conditions
To solve (5.4)–(5.6), we use a semigroup approach. We endow the space \([L^2(\Omega )]^3 \times \mathcal {D}(A_1^{1/2})\times L_0^2(\omega )\) with the scalar product
We consider the following functional spaces
We also denote by \(\mathbb {P}\) the orthogonal projector
Finally, we define
and
Using the above definition, we can write (5.4)–(5.6) as
with
Proposition 5.1
Assume that \(\beta _1+\beta _2\ne 0\). The operator A defined by (5.8)–(5.10) is the infinitesimal generator of a strongly continuous semigroup of contraction on \(\mathbb {H}\).
Proof
First we show that the operator A is dissipative: assume \(W=\begin{pmatrix} w\\ \eta _1\\ \eta _2 \end{pmatrix}\in \mathcal {D}(A)\). Then, by integration by parts, we obtain:
We write
and we deduce
Second, we show that the operator A is m-dissipative: we prove that for some \(\lambda > 0\) the operator \(\lambda I+A\) is onto. Let \(F=\begin{pmatrix} f\\ g\\ h \end{pmatrix}\in \mathbb {H}\). The problem is to find \(\begin{pmatrix} w\\ \eta _1\\ \eta _2 \end{pmatrix} \in \mathcal {D}(A)\) solution of the equation
which is equivalent to the system
To solve the above system, we use that \(\eta _1=\frac{1}{\lambda }(g+\eta _2)\) to obtain a system in \((u,\eta _2)\) and we introduce the space
We can thus write the Eq. (5.12) in a variational form: find \((w,\eta _2)\in \mathcal {V}\) such that
with \(a: \mathcal {V} \times \mathcal {V}\longrightarrow \mathbb {R}\) given by
and \(L:\mathcal {V}\longrightarrow \mathbb {R}\) given by
The bilinear form a is continuous and coercive on \(\mathcal {V}\) thanks to the classical Korn inequality. We can also check that L is linear and continuous on \(\mathcal {V}\). By the Lax–Milgram theorem, there exists a unique \((u,\eta _2)\in \mathcal {V}\) solution of (5.14).
Now, taking \(\xi =0\) and \(\phi \in \mathcal {D}_\sigma (\Omega )\), the Eq. (5.14) becomes
which is equivalent to
Using the De Rham theorem [33, Proposition 1.2, p.14] , we deduce the existence of a unique \(q\in L^2(\Omega )/\mathbb {R}\) such that (5.13a) holds. In particular, we have \(\nabla \cdot \mathbb {T}(w,q)\in [L^2(\Omega )]^3\) and \(\mathbb {T}(w,q)\in [L^2(\Omega )]^9\). Therefore, we deduce that \(\mathbb {T}(w,q)n_0\in [H^{-1/2}(\partial \Omega )]^3\) and
for all \(\phi \in [H^1(\Omega )]^3,\;\nabla \cdot \phi =0,\;\phi _{n_0}=0\). On the other hand, taking \(\xi =0\) in (5.14) yields
for all \(\phi \in [H^1(\Omega )]^3,\;\nabla \cdot \phi =0,\;\phi _{n_0}=0\). Comparing (5.15) and (5.16) and taking into account that
we obtain
Let \(\phi \in [H^{1/2}(\partial \Omega )]^3\) such that \(\phi _{n_0}=0\), and let consider the system
The above system admits a unique solution \((\widehat{g},\widehat{q})\in [H^1(\Omega )]^3\times L^2_0(\Omega )\) such that \(\nabla \cdot \widehat{g}=0\) and \(\widehat{g}|_{\partial \Omega }=\phi \). This implies that (5.17) holds for all \(\phi \in [H^{1}(\Omega )]^3\), \(\phi _{n_0}=0\). Inserting (5.17) in (5.15) we get
for all \(\phi \in [H^1(\Omega )]^3\), \(\phi _{n_0}=0\).
Thus, we deduce that (w, q) is a weak solution of (5.13a), (5.13b), (5.13e) and (5.13f) in the sense of Definition 4.1. Since \(\eta _2\in H^2(\omega )\), \(\mathcal {T}\eta _2\in [H^{2}(\partial \Omega )]^3\) we can apply Theorem 4.2 and obtain \((w,q) \in [H^2(\Omega )]^3\times H^1(\Omega )/\mathbb {R}\).
Going back to the variational formulation (5.14), we deduce
for any \(\xi \in \mathcal {D}(A_1^{1/2})\) and where \(\eta _1=\frac{1}{\lambda }(g+\eta _2)\). We have \(\mathbb {T}(w,q)n_0 \in [H^{1/2}(\partial \Omega )]^{3}\) and thus \(\mathcal {T}^*(\mathbb {T}(w,q)n_0) \in L^2_0(\omega )\). Moreover since \(\eta _2\in H^2(\omega )\), we deduce that \(\eta _2\in \mathcal {D}(A_2)\). Thus \(A_1\eta _1 \in L^2_0(\omega )\).
Applying Lumer-Phillips theorem, we conclude that \((e^{-tA})_{t\geqslant 0}\) is a semigroup of contractions on \(\mathbb {H}\).
\(\square \)
In order to prove that \((e^{-tA})_{t\geqslant 0}\) is an analytical semigroup, we use Lemma 3.10 in [2]. We first need to show that \((e^{-tA})_{t\geqslant 0}\) is exponentially stable.
Proposition 5.2
Assume that \(\beta _{1}+\beta _{2}\ne 0\). The semigroup \((e^{-tA})_{t\geqslant 0}\) is exponentially stable.
Proof
Since \((e^{-tA})_{t\geqslant 0}\) is a semigroup of contraction, we apply the classical result of Huang–Gearhart (see for instance [26, Theorem 1.3.2, p.4]). We have to show that
Using the proof of [2, Proposition 3.5], we only need to prove the existence of \(C>0\) such that
Let us consider \(\lambda \in \mathbb {C}\), with \({\text {Re}}\lambda \in (0,1)\), \(F=\begin{pmatrix} f\\ g\\ h \end{pmatrix}\in \mathbb {H}\) and \(\begin{pmatrix}w\\ \eta _1\\ \eta _2\end{pmatrix} \in \mathcal {D}(A)\) such that
We can write the above relation as the system (5.13). We multiply (5.13a) by \(\overline{w}\), (5.13d) by \(\overline{\eta }_2\) and we perfom integrations by parts to deduce
We have
On the other hand, we have
Using (4.25), (5.21) and the fact that \(\mathcal {T}\in \mathcal {L}(L^2(\omega ),[L^2(\partial \Omega )]^3)\), we obtain
Following the proof of Proposition 3.5 in [2], we have
Gathering the above inequality with (5.22) and (5.21), we obtain
for some positive constant C. This concludes the proof. \(\square \)
Proposition 5.3
Suppose that \(\beta _{1}+\beta _2\ne 0\). The operator A is the infinitesimal generator of an analytic semigroup on \(\mathbb {H}\).
Proof
We apply Lemma 3.10 in [2]: since \((e^{-tA})_{t\geqslant 0}\) is exponentially stable, it sufficient to show
Assume \(\lambda \in i\mathbb {R}^*\), \(F=\begin{pmatrix}f\\ g\\ h\end{pmatrix}\in \mathbb {H}\) and let us consider \(W=(\lambda I+{A})^{-1}F\). We write \(W=\begin{pmatrix}w\\ \eta _1\\ \eta _2\end{pmatrix}\) so that (5.13) holds. We now proceed as in [2, Proposition 3.11]: we multiply (5.13a) by \(\overline{u}\) and (5.13d) by \(\overline{\eta }_2\) and we integrate by parts
Multiplying by \(\overline{\lambda }\) and taking the real part, we find
Using the Cauchy-Schwarz inequality, we obtain
Since \(A_1\) and \(A_2\) are self-adjoint positive operators and \(\mathcal {D}(A_1^{1/4})=\mathcal {D}(A_2^{1/2})\), we apply [11, Theorem 1.1] to deduce that
is the infinitesimal generator of an analytical semigroup on \(\mathcal {D}(A_{1}^{1/2})\times L^2_0(\omega )\). We have in particular
Applying this estimate on (5.13c)–(5.13d), we deduce
We use the fact \(\mathcal {T}^*\in \mathcal {L}([L^2(\partial \Omega )]^3,L^2_0(\omega ))\) and we combine (5.26) and (5.25) to find
Combining Theorem 4.2 and an interpolation argument, we get for \(\varepsilon <1/4\)
The rest of the proof is similar to the proof of [2, Proposition 3.11]. \(\square \)
We recall that \(\mathcal {X}_{\infty ,\gamma }\) is the space given in (2.8). We are now in position to give the following theorem.
Theorem 5.4
Suppose that \(\beta _{1}+\beta _2\ne 0\). There exists \(\gamma _0>0\) such that if
and
for \(\gamma \in [0,\gamma _0]\), then there exists a unique solution \((u,p,\eta )\in \mathcal {X}_{\infty ,\gamma }\) on \((0,+\infty )\) of the system (5.1)–(5.3). Moreover there exists a positive constant C such that
Proof
Since A generates an analytical and exponentially stable semigroup, from [5, Theorem 3.1, p.143], the evolution Eq. (5.11) admits a unique strong solution and verifies the estimates
Applying the De Rham theorem [33, Proposition 1.2, p.14], we deduce the existence of \(q\in L^2_{\gamma }(0,\infty ;H^1(\Omega )/\mathbb {R})\) such that \((w,\eta ,q)\) is the solution of (5.4)–(5.6). Setting \(u=w+v\), \(p=q+\pi \) where \((v,\pi )\) is the solution of (4.22) associated with \(\widetilde{g}\), we obtain the result. \(\square \)
Corollary 5.5
Suppose that \(\beta _{1}+\beta _2\ne 0\). Assume \(T>0\) and
Then there exists a unique solution \((u,p,\eta )\in \mathcal {X}_T\) on (0, T) of the system (5.1)–(5.3). Moreover, there exists a positive constant independent of T such that
Proof
We extend f, \(\widetilde{g}\), h by 0 in \((T,\infty )\) and apply Theorem 5.4. \(\square \)
We can now deal with the case \(\beta _i=0\) for \(i=1,2\)
Theorem 5.6
Suppose that \(\beta _{1}=\beta _2= 0\). Assume \(T>0\) and
Then there exists a unique solution \((u,p,\eta )\in \mathcal {X}_T\) on (0, T) of the system (5.1)–(5.3). Moreover, there exists a positive constant (non decreasing with respect to T) such that
Proof
Let introduce the space
Let \((\widetilde{u},\widetilde{\eta _2})\in \mathbb {X}\). From Corollary 5.5 (with \(\beta _1=\beta _2=1\)), there exists a unique strong solution \((u,p,\eta )\in \mathcal {X}_T\) to the system (5.1), (5.3) with the boundary conditions
Using the trace theorems and the definition (2.6) of \(\mathcal {X}_T\) we can thus define the mapping
Let us prove that the mapping \(\mathbb {F}\) is a contraction for T small enough: assume \((\widetilde{u}^i,\widetilde{\eta }_2^i)\in \mathbb {X}\), \(i=1,2\) and let \((u^i,p^i,\eta ^i)\in \mathcal {X}_T\) \(i=1,2\) be the corresponding solutions of the system (5.1), (5.3), (5.33). We write
so that
From (5.31) and the boundedness of \(\mathcal {T}\), we obtain
From (2.6), (2.7), the trace theorem and Lemma A.5 in [6], there exists a constant C independent of T such that
From Corollary A.3 in [6] and (5.36), we deduce
and
Combining the estimates (5.38), (5.39), (5.40), we obtain
This shows that \(\mathbb {F}\) is a contraction for T small enough and using the Banach fixed-point theorem, we deduce the existence and the uniqueness of a strong solution for the system (5.1)–(5.3) (with \(\beta _1=\beta _2=0\)) and the estimate (5.32). To deduce the result fo any T, we simply reiterate the above procedure on small intervals \([kT_0,(k+1)T_0]\), where \(T_0\) is such that \(\mathbb {F}\) is a contraction. \(\square \)
6 Fixed point
In this section, we prove the main result Theorem 1.1. Using Definition 3.1, we first restate this result after change of variables.
Theorem 6.1
-
1.
Let \(\beta _i\geqslant 0\), \(i=1,2\). Assume that \((u^0,\eta ^0,\eta ^1)\in \mathbb {V}\) with
$$\begin{aligned} 1+\eta ^0>0. \end{aligned}$$There exists a time \(T_0>0\) (depending only on \(\Vert (u^0,\eta ^0,\eta ^1)\Vert _{\mathbb {V}}\)) such that the system (3.5), (3.6) and (3.7) admits a unique strong solution \((u,p,\eta )\in \mathcal {X}_T\) for \(T<T_0\).
-
2.
Let \(\beta _i\geqslant 0\) with \(\beta _1+\beta _2>0\), \(i=1,2\). There exists \(R_0>0\) such that for any \((u^0,\eta ^0,\eta ^1)\in \mathbb {V}\) with
$$\begin{aligned} 1+\eta ^0>0 \quad \text {and with} \quad \Vert (u^0,\eta ^0,\eta ^1)\Vert _{\mathbb {V}}\leqslant R_0, \end{aligned}$$then the system (3.5), (3.6) and (3.7) admits a unique strong solution \((u,p,\eta )\in \mathcal {X}_{\infty ,\gamma }\) on \((0,\infty )\) for \(\gamma \in [0,\gamma _0]\).
We recall that \(\mathbb {V}\) is defined by (5.7). The above result is obtained by using a fixed-point argument.
First let us show the local in time existence. We define for all \(T>0\) the space
and for \(R>0\), we define the set
In the sequel, we denote by C a quantity which does not depend on R and T. We first start by assuming
Thus, applying Theorem 5.6, we know that for any \((f,\widetilde{g},h)\in \mathcal {B}_{T,R}\), there exists a unique solution \((u,p,\eta )\in \mathcal {X}_T\) of (5.1)–(5.3). Moreover, the estimate (5.29) yields
for some positive constant C. For the local existence, the constant R is fixed. In the next section, we show that for T small enough, we can define F, G, H by (3.9), (3.10) and (3.14) and thus consider the mapping \(\Phi \) defined as follows:
In what follows, we show that for T small enough, we have \(\Phi (\mathcal {B}_{T,R})\subset \mathcal {B}_{T,R}\) and that \(\Phi _{|\mathcal {B}_{T,R}}\) is a strict contraction.
First, we notice that (6.4) yields several other useful estimates. From (2.6), (2.7) and Lemma A.5 in [6], there exists a constant C independent of T such that
Here and in what follows, we use the following notation for the derivatives of the function \(\eta =\eta (t,s_1,s_2)\):
For simplicity, in all what follows, we assume
The above assumption simplifies the estimates in the sense that we only keep the smaller power of T. We also denote by \(C_R\) a constant that can depend on R in a nondecreasing way (typically the sum of \(C R^m\), \(m\in \mathbb {N}\), \(C>0\)). The value of these constants may change from one appearance to another.
6.1 Estimates on the change of variables
We first prove some useful estimates on \(\eta \)
Lemma 6.2
We have
In particular, there exists
such that if \(T\leqslant T_0\), then
We also have the following estimates
Proof
In order to prove (6.8), we write
and we combine it with (6.6) and with \(H^2(\omega )\hookrightarrow L^\infty (\omega )\).
Since
there exists \(\varepsilon >0\) such that \(1+\eta ^0>2\varepsilon \). Using (6.8), we obtain (6.9) if T is small enough.
We set \(\xi = \partial _{s_j}\eta -\partial _{s_j}\eta ^0\) and \(\xi ^*(z,\cdot )=\xi \left( z T,\cdot \right) \), \(z\in [0,1]\). Then we combine (A.1), the embedding \(H^{3/4}(0,1)\hookrightarrow L^\infty (0,1),\) Lemma A.1 in [6] and (6.6) to obtain
Then, we deduce (6.10) and (6.11) by using \(H^{3/2}(\omega ) \hookrightarrow L^\infty (\omega )\) and \(H^{1/2}(\omega ) \hookrightarrow L^4(\omega )\).
Finally, (6.12) is a consequence of (6.6) and (2.7). \(\square \)
Now, we show some estimates on the changes of variables X and Y defined by (3.2). We recall that \(a_{ik}\) is given by (3.8).
Lemma 6.3
Assume (6.7).
Proof
By definition [see (3.1) and (3.2)], we recall that
As a consequence, the estimate on \(\nabla Y(X)-I_3\) reduces to the estimate of the following terms
We have
By using (6.8) and (6.9), we deduce
On the other hand, for \(j=1,2\), we have
and thus, using (6.4), (6.3), (6.8) and (6.10),
Hence, we obtain (6.14) and thus (6.15).
We have for \(k,j\in \{1,2\}\),
Then, we obtain
Using (6.11), (6.10) and (6.8), we obtain (6.16). The other cases for k, j are easier to do and we skip them.
The third derivative \(\frac{\partial ^3Y}{\partial x_j\partial _k\partial x_l}\) involves the following terms
Thus, using (6.4), (6.10), (6.11), (6.8) and (2.7), we obtain (6.17).
We have
and thus
Thus, using (6.3) and (6.6), we obtain (6.18).
The terms appearing in \(\partial _t a_{ik}(X)\) are of the form
Consequently, using (6.8) and (6.10),
The above estimate and (6.12) yield (6.19). \(\square \)
Now, we need the following lemma to estimates the terms on the boundary.
Lemma 6.4
Assume (6.7). Then we have the following estimates
Proof
Relation (6.25) is a consequence of (6.21), (6.23), (1.5) and (3.11) combined with (6.11). We obtain (6.26) by using Lemma 6.2 with (3.8).
Using (6.6) and \(H^{5/4}(\omega )\hookrightarrow L^\infty (\omega )\), we obtain
For \((\alpha _1, \alpha _2, \alpha _3) \in \mathbb {N}^3\), we also deduce that
Nevertheless, one has to take care about the dependence in T of the corresponding norm. In order to do this, we notice that if
then
and
where \(\left\| \cdot \right\| _ {H^{7/8}(0,T;L^\infty (\omega )) \cap L^\infty (0,T;L^\infty (\omega ))}=\left\| \cdot \right\| _{H^{7/8}(0,T;L^\infty (\omega ))} + \left\| \cdot \right\| _{L^\infty (0,T;L^\infty (\omega ))}\).
The last estimate is obtained by writing the definition (2.1) of the norm in \(H^{7/8}(0,T;L^\infty (\omega ))\).
Then, combining (6.29) with (6.4), we obtain that
From this estimate and (6.21), (6.23), (1.5) and (3.11), we obtain (6.27).
To prove (6.28), we use that the terms appearing in \(\frac{\partial a_{mk}}{\partial x_j}(X)\) are of the form (6.24). Combining the above arguments with (6.6) and (6.4), we deduce the result. \(\square \)
6.2 Estimates of F, G, H
Proposition 6.5
Assume F, G, H are given by (3.9), (3.14), (3.10). Then we have
Proof
Using (6.14), (6.15), we obtain
and
Using (6.15) and (6.18), we obtain
Using (6.15) and (6.16), we get
From (6.19) and (6.6), it follows that
Using standard estimates on the nonlinear terms (see, for instance, [3, p.48]), we have
Combining this with (6.14) yields
Using (6.16), we have also
Hence, \(F(u,p,\eta )\) is \(L^2(0,T;[L^2(\Omega )]^3)\) and using (6.33), (6.34), (6.35), (6.39), (6.37) and (6.40), we get
We estimate now \(G(u,\eta )\) in \(W^{1/4}(0,T;[H^{1/2}(\partial \Omega )]^3,[L^2(\partial \Omega )]^3)\). We recall that the formula (3.14) for G involves \(\tau ^i\), \(\mathcal {W}\), \(\mathcal {V}^i\) [see (3.11), (3.12), (3.13)]. First we write for \(i=1,2\)
with
From (6.4) and trace results, we have
Combining this with (6.25) and (6.26), we deduce
and thus from (3.14), we finally obtain
For the estimate in \(H^{1/4}(0,T;L^2(\partial \Omega ))\), we use (A.5): for instance,
The last inequality is obtained by using both (6.25), (6.27) and (6.6).
The other kind of terms that has to be estimated are of the form
where we have used (A.5) and
All the other terms are estimated similarly so that we finally deduce (6.32). The estimate (6.31) on H can be done similarly as the estimate (6.32) for G. \(\square \)
6.3 Proof of Theorem 6.1
We are now in position to prove Theorem 6.1.
Proof of Theorem 6.1
First let us prove the local in time existence. We recall that \(\Phi \) is given by (6.5), with \(\mathcal {Y}_T\) given by (6.1). From (6.30), (6.32), (6.31), we obtain
Thus, for T small enough, we obtain that \(\Phi (\mathcal {B}_{T,R})\subset \mathcal {B}_{T,R}\), where \(\mathcal {B}_{T,R}\) is defined by (6.2). With computations similar as the ones done in the two previous subsections, we also obtain that for T small enough, \(\Phi |_{{\mathcal {B}_{T,R}}}\) is a contraction. Using the Banach fixed-point theorem, we deduce the existence and uniqueness of \((u,p,\eta )\) solution of the system (3.5), (3.6) and (3.7) provided that T is small enough.
For the second part of Theorem 6.1, the application \(\Phi \) is defined in a similar way as (6.5) but with \(T=\infty \) and
Here \(\gamma \in [0,\gamma _0]\), where \(\gamma _0\) is given by Theorem 5.4. In that case, we show that for R small enough \(\Phi (\mathcal {B}_{\infty ,R})\subset \mathcal {B}_{\infty ,R}\) and that \(\Phi _{|\mathcal {B}_{\infty ,R}}\) is a strict contraction. The estimates are similar to the previous case, but are simpler: for instance, Lemma 6.2 is replaced by the following estimates:
In particular, there exists \(R_0>0\) so that, if \(R\leqslant R_0\), then
We can then define the changes of variables X and Y by (3.3), and obtain similar estimates as in Lemma 6.3, Lemma 6.4 and Proposition 6.5.
This yields
and
for \((f,\widetilde{g},h), (f^{(i)},\widetilde{g}^{(i)},h^{(i)})\in \mathcal {B}_{\infty ,R}\). Then, we use the Banach fixed point by taking R small enough and we deduce the global existence and uniqueness of a strong solution \((u,p,\eta )\in \mathcal {X}_{\infty ,\gamma }\) for the system (3.5), (3.6) and (3.7) provided that R is small enough. \(\square \)
References
Acevedo, P., Amrouche, C., Conca, C., Amrita, G.: Stokes and Navier–Stokes Equations with Navier Boundary Condition. arXiv:1805.07760v1 (2018)
Badra, M., Takahashi, T.: Feedback boundary stabilization of 2D fluid–structure interaction systems. Discrete Contin. Dyn. Syst. 37(5), 2315–2373 (2017)
Beirão da Veiga, H.: On the existence of strong solutions to a coupled fluid–structure evolution problem. J. Math. Fluid Mech. 6(1), 21–52 (2004)
Beirão Da Veiga, H.: Regularity for Stokes and generalized Stokes systems under nonhomogeneous slip-type boundary conditions. Adv. Differ. Equ. 9(9–10), 1079–1114 (2004)
Bensoussan, A., Da Prato, G., Delfour, M.C., Mitter, S.K.: Representation and Control of Infinite Dimensional Systems. Systems & Control: Foundations & Applications, 2nd edn. Birkhäuser Boston Inc, Boston, MA (2007)
Boulakia, M., Guerrero, S., Takahashi, T.: Well-Posedness for the Coupling Between a Viscous Incompressible Fluid and an Elastic Structure. https://hal.inria.fr/hal-01939464 (2018) (preprint)
Bravin, M.: On the Weak Uniqueness of “Viscous Incompressible Fluid + Rigid Body” System with Navier Slip-With-Friction Conditions in a 2D Bounded Domain. https://hal.archives-ouvertes.fr/hal-01740859 (2018) (preprint)
Bălilescu, L., San Martín, J., Takahashi, T.: Fluid–rigid structure interaction system with Coulomb’s law. SIAM J. Math. Anal. 49(6), 4625–4657 (2017)
Casanova, J.-J.: Existence of Time-Periodic Strong Solutions to a Fluid–Structure System. https://hal.archives-ouvertes.fr/hal-01838262 (2018) (preprint)
Casanova, J.-J.: Fluid Structure System with Boundary Conditions Involving the Pressure. arXiv:1707.06382 (2017)
Chen, S.P., Triggiani, R.: Proof of extensions of two conjectures on structural damping for elastic systems. Pac. J. Math. 136(1), 15–55 (1989)
Gérard-Varet, D., Hillairet, M.: Existence of weak solutions up to collision for viscous fluid-solid systems with slip. Commun. Pure Appl. Math. 67(12), 2022–2075 (2014)
Gérard-Varet, D., Hillairet, M., Wang, C.: The influence of boundary conditions on the contact problem in a 3D Navier–Stokes flow. J. Math. Pures Appl. (9) 103(1), 1–38 (2015)
Grandmont, C.: On an Unsteady Fluid–Beam Interaction Problem. https://basepub.dauphine.fr/bitstream/handle/123456789/6848/2004-48.pdf (2004) (preprint)
Grandmont, C., Hillairet, M.: Existence of global strong solutions to a beam–fluid interaction system. Arch. Ration. Mech. Anal. 220(3), 1283–1333 (2016)
Guidoboni, G., Guidorzi, M., Padula, M.: Continuous dependence on initial data in fluid–structure motions. J. Math. Fluid Mech. 14(1), 1–32 (2012)
Guidorzi, M., Padula, M., Plotnikov, P.I.: Hopf solutions to a fluid–elastic interaction model. Math. Models Methods Appl. Sci. 18(2), 215–269 (2008)
Hillairet, M.: Lack of collision between solid bodies in a 2D incompressible viscous flow. Commun. Partial Differ. Equ. 32(7–9), 1345–1371 (2007)
Hillairet, M., Takahashi, T.: Collisions in three-dimensional fluid structure interaction problems. SIAM J. Math. Anal. 40(6), 2451–2477 (2009)
Horgan, C.O.: Korn’s inequalities and their applications in continuum mechanics. SIAM Rev. 37(4), 491–511 (1995)
Inoue, A., Wakimoto, M.: On existence of solutions of the Navier–Stokes equation in a time dependent domain. J. Fac. Sci. Univ. Tokyo Sect. IA Math. 24(2), 303–319 (1977)
Kistler, S.F., Scriven, L.E.: Coating flow theory by finite element and asymptotic analysis of the Navier–Stokes system. Int. J. Numer. Methods Fluids 4(3), 207–229 (1984)
Lequeurre, J.: Existence of strong solutions to a fluid–structure system. SIAM J. Math. Anal. 43(1), 389–410 (2011)
Liakos, A.: Finite-element approximation of viscoelastic fluid flow with slip boundary condition. Comput. Math. Appl. 49(2–3), 281–294 (2005)
Lions, J.-L., Magenes, E.: Problèmes aux Limites Non Homogènes et Applications. Travaux et Recherches Mathématiques, No. 18, vol. 2. Dunod, Paris (1968)
Liu, Z., Zheng, S.: Semigroups Associated with Dissipative Systems. Chapman & Hall/CRC Research Notes in Mathematics, vol. 398. CRC, Boca Raton, FL (1999)
Muha, B., Čanić, S.: Existence of a weak solution to a fluid–elastic structure interaction problem with the Navier slip boundary condition. J. Differ. Equ. 260(12), 8550–8589 (2016)
Navier, C.L.M.H.: Mémoire sur les lois du mouvement des fluides. Mémoires de l’Académie Royale des Sciences de l’Institut de France 6(1823), 389–440 (1823)
Planas, G., Sueur, F.: On the “viscous incompressible fluid + rigid body” system with Navier conditions. Ann. Inst. H. Poincaré Anal. Non Linéaire 31(1), 55–80 (2014)
Raymond, J.-P.: Feedback stabilization of a fluid-structure model. SIAM J. Control Optim. 48(8), 5398–5443 (2010)
San Martín, J.A., Starovoitov, V., Tucsnak, M.: Global weak solutions for the two-dimensional motion of several rigid bodies in an incompressible viscous fluid. Arch. Ration. Mech. Anal. 161(2), 113–147 (2002)
Shimada, R.: On the \(L_p\)-\(L_q\) maximal regularity for Stokes equations with Robin boundary condition in a bounded domain. Math. Methods Appl. Sci. 30(3), 257–289 (2007)
Temam, R.: Navier–Stokes Equations. Studies in Mathematics and its Applications. Theory and numerical analysis, With an appendix by F. Thomasset, vol. 2, revised edn. North-Holland Publishing Co., Amsterdam (1979)
Temam, R.: Problèmes Mathématiques en Plasticité, Méthodes Mathématiques de l’Informatique [Mathematical Methods of Information Science], vol. 12. Gauthier-Villars, Montrouge (1983)
Verfürth, R.: Finite element approximation of incompressible Navier–Stokes equations with slip boundary condition. Numer. Math. 50(6), 697–721 (1987)
Wang, C.: Strong solutions for the fluid–solid systems in a 2-D domain. Asymptot. Anal. 89(3–4), 263–306 (2014)
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that there is no conflict of interest.
Additional information
Communicated by G. P. Galdi
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Takéo Takahashi acknowledges the support of the Agence Nationale de la Recherche, Project IFSMACS, ANR-15-CE40-0010.
Appendix A: Technical results
Appendix A: Technical results
In this section, we give some technical estimates that have been elaborated in [6]. Given a function \(\xi \), we define for \(z\in [0,1]\), \(\xi ^*(z)=\xi (zT)\). Assume \(\mathfrak {X}\) is a Banach space. If \(\xi \in H^s(0,T; \mathfrak {X}) \), then \(\xi ^* \in H^s(0,1;\mathfrak {X}) \) and
Assume \(\sigma _2\in (1/2,1]\) and \(\sigma _1\in [0,\sigma _2]\). Using the above result, there exists a constant independent of T such that for any \(\xi \in H^{\sigma _2}(0,T;\mathfrak {X})\) and \(\xi (0)=0\), then
We also recall the following result on the interpolation estimates (with constants independent of T), see [6, Lemma A.5]: assume \(\sigma \in [0,1]\), \(\mu _1\geqslant 0\), \(\mu _2\geqslant 0\) and \(\mu =\sigma \mu _1+(1-\sigma )\mu _2\). Then there exists a constant C independent of T such that for any function \(u\in H^1(0,T;H^{\mu _1}(\Omega ))\cap L^2(0,T;H^{\mu _2}(\Omega ))\), we have
On the other hand, for \(p,\;q\in [1,+\infty ]\) and \(\dfrac{1}{r}=\dfrac{\sigma }{p}+\dfrac{(1-\sigma )}{q}\), we have
for \(u\in L^p(0,T;H^{\mu _1}(\Omega ))\cap L^q(0,T;H^{\mu _2}(\Omega ))\).
We give also a useful formula (see [6, Lemma A.7]) for the product of functions: assume that \(\mathfrak {X}_1\), \(\mathfrak {X}_2\) and \(\mathfrak {X}_3\) are Banach spaces such that
Let us assume \(\sigma \in (1/2,1]\), \(s\in [0,1/2]\), \(T_0>0\). Then there exists a constant C such that for any \(T\leqslant T_0\) we have
for all \(u_1\in H^s(0,T;\mathfrak {X}_1)\) and \(u_2\in H^\sigma (0,T;\mathfrak {X}_2)\).
Rights and permissions
About this article
Cite this article
Djebour, I.A., Takahashi, T. On the Existence of Strong Solutions to a Fluid Structure Interaction Problem with Navier Boundary Conditions. J. Math. Fluid Mech. 21, 36 (2019). https://doi.org/10.1007/s00021-019-0440-7
Accepted:
Published:
DOI: https://doi.org/10.1007/s00021-019-0440-7