Abstract
We study global regularity of nonlinear systems of partial differential equations depending on the symmetric part of the gradient with Dirichlet boundary conditions. These systems arise from variational problems in plasticity with power growth. We cover the full range of exponents \(p \in (1,\infty )\). As a novelty the degenerate case for \(p>2\) is included. We present a unified approach for all exponents by showing the regularity for general systems of Orlicz growth.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
In this paper we investigate the global regularity of solutions of nonlinear elliptic systems arising from problems in plasticity, see e.g. [25]. In particular, we are interested in systems of the form
Here, \(\Omega \subset \mathbb {R}^n\) is a bounded domain, \(p\in (1,\infty )\), \(\delta \ge 0\), \(u:\Omega \rightarrow \mathbb {R}^n\) is a vector field and \({\varepsilon }u{:}{=}\tfrac{1}{2} (\nabla u + (\nabla u)^T)\) is the symmetric part of \(\nabla u\), also denoted as symmetric gradient.
Similar systems arise in the context of non-Newtonian fluids (power-law fluids), where additionally a pressure and a incompressibility constraint appears, see e.g. [31]. These are then called p-Stokes systems.
For \(\delta = 0\) our system is also known as the symmetric p-Laplace system. Sometimes in the literature the case \(\delta =0\) and \(1<p<2\) is called singular and the case \(\delta =0\) and \(p>2\) is called degenerate.
The system (1.1) arises as the Euler-Lagrange equation of the functional
where \(\varphi \) is the Young function
For \(\delta =0\) we obtain \(\varphi (t) = \frac{1}{p} t^p\). In fact, although the system (1.1) is our main interest, we also consider systems arising from more general Young functions \(\varphi \). Our technique has the advantage that we do not have to distinguish in our proofs the cases \(p<2\) and \(p>2\) as is often the case in the literature.
In the study of our system (1.1) two quantities are of special interest
The former \(A(\varepsilon u)\) is sometimes called the stress tensor. The latter \(V(\varepsilon u)\) appears naturally in the study of regularity of nonlinear systems with power growth. Note that A and V fulfill the relation \(A(\varepsilon u): \varepsilon u = {|V(\varepsilon u) |}^2\). Our goal is to establish global regularity properties for solutions u of (1.1). This is best expressed in terms of \(V(\varepsilon u)\). In particular, we will show that \(V(\varepsilon u) \in W^{1,2}(\Omega )\) up to the boundary. This is the natural regularity when applying the technique of difference quotients which formally corresponds to testing the equation with \(\Delta u\).
We start by giving an overview of already existing results. Let us begin with the simpler case of the standard p-Laplace system, which depends on the gradient \(\nabla u\) rather than the symmetric gradient \({\varepsilon }u\). There the proof of \(V(\nabla u) \in W^{1,2}_{{\textrm{loc}}}(\Omega )\) is rather classical and can be found e.g. in [32, Theorem 1.11.1], [37, Lemma 3.1] and [1, Proposition 2.4] or in [17, Theorem 11] for Orlicz growth. Global regularity \(V(\nabla u) \in W^{1,2}(\Omega )\) has been established in [14] for parabolic systems. The elliptic case is a direct consequence of this by considering the limit \(t\rightarrow \infty \).
Let us now discuss the known results on regularity for our system (1.1) with symmetric gradients. Interior regularity \(V(\varepsilon u) \in W^{1,2}_{{\textrm{loc}}}(\Omega )\) for \(A(\varepsilon u)\) roughly of the form \(A(\varepsilon u) = (1+{|\varepsilon u |})^{p-2} \varepsilon u\) with \(p \in (1,2)\) has been shown in [34]. The technique using difference quotients is similar to the one for the case of dependence on \(\nabla u\). Only in the case of global regularity \(V(\varepsilon u) \in W^{1,2}(\Omega )\) the question becomes significantly more difficult. The first result on global regularity is due to Seregin and Shilkin [36] who proved the global regularity \(V({\varepsilon }u)\in W^{1,2}(\Omega )\) for the case of a flat boundary and for operators of the form \(A({\varepsilon }u){:}{=}(\delta +|{\varepsilon }u| )^{p-2}{\varepsilon }u\), where \(1<p\le 2\) and \(\delta >0\). This result was extended by Berselli and Růžička [11] to non-flat boundaries \(\partial \Omega \) of class \(C^{2,1}\) with \(1 < p \le 2\) and \(\delta \ge 0\) (see Sect. 4.1 for the precise definition of \(C^{2,1}\)-regular boundaries). The case \(p \ge 2\) but with \(\delta >0\) (non-degenerate case) has been proved in [9, 12], where also the parabolic situation is investigated. The estimates depend strongly on \(\delta >0\) and deteriorate for \(\delta \rightarrow 0\). In this paper we will prove \(V({\varepsilon }u)\in W^{1,2}(\Omega )\) for all \(1<p<\infty \) and \(\delta \ge 0\). To include all the possible variants of growth conditions we work in the even more general setting of Orlicz growth. For the special case of p-growth our main result is the following Theorem 1.1. The general case can be found in Theorem 2.4.
Theorem 1.1
Let \(\Omega \) be a bounded \(C^{2,1}\)-domain, \(1<p<\infty \), \(\delta \ge 0\) and \(f\in W_0^{1,p^\prime }(\Omega )\). Then the system
has a unique weak solution u fulfilling \(V({\varepsilon }u){:}{=}(\delta +|{\varepsilon }u|)^{\tfrac{p-2}{2}}{\varepsilon }u\in W^{1,2}(\Omega )\) with the estimate
If \(p\le 2\), this implies \(u\in W^{2,\tfrac{np}{n+p-2}}(\Omega )\), and if \(p\ge 2\), it implies \(u\in W^{1+\tfrac{2}{p},p}(\Omega )\).
Our assumption of a \(C^{2,1}\)-regular boundary is the same that is made in [11]. It is a technical assumption that we need because of the way we locally flatten the boundary. In particular, we straighten the boundary locally only in one direction. A more flexible method would allow slightly less regular boundaries.
Also our assumptions on f are in some cases more stringent than necessary. This is due to the fact that \(\varphi \) is a general N-function and does not need to have p growth for one specific p. If it was, e.g., of p growth with \(1<p\le 2\), then \(f\in L^{p'}(\Omega )\) would suffice. However, for \(p>2\) it is not sufficient to assume \(f\in L^{p'}\), see [13, Section 5]. It might be possible to reduce \(f \in W^{1,p'}\) for \(p > 2\) to \(W^{s,p'}\) for some \(s \in (\frac{p-2}{p},1)\) as done in [13]. The condition that f vanishes at the boundary is needed in order to perform integration by parts without getting any boundary terms and is important for \(p \ge 2\).
Let us make a few additional remarks where the regularity theory of the symmetric p-Laplace system lacks behind the one for the p-Laplace system involving \(\nabla u\) instead of \(\varepsilon u\). Higher Hölder regularity \(u \in C^{1,\alpha }\), in other words \(V(\nabla u) \in C^{0,\beta }\), has been shown in [1, 37] for the p-Laplace system and in [23] for the corresponding Orlicz growth. Such regularity up to the boundary has been proved in [14] even for parabolic problems. On the other hand, \(C^{1,\alpha }\)-regularity results are completely open for the symmetric p-Laplacian for \(n \ge 3\). The case \(n=2\) has been solved in [21, 29], where even the p-Stokes system is covered.
Another aspect of regularity is the one of the stress \(A(\nabla u)\) in the case of the p-Laplace system. If \(f \in L^{n,1}(\Omega )\) (Lorentz space), then \(A(\nabla u) \in L^\infty (\Omega )\), resp \(\nabla u \in L^\infty (\Omega )\), has been proved in [15, 28]. If \(f = {\text {div}}F\) and \(F \in C^{0,\alpha }(\overline{\Omega })\), then \(A(\nabla u) \in C^{0,\beta }(\overline{\Omega })\) has been proved in [4]. The maximal regularity \(A(\nabla u) \in W^{1,2}(\Omega )\) for the p-Laplace system with \(f \in L^2(\Omega )\) and \(p > 2(2-\sqrt{2}) = 1.171532\dots \) has been shown in [3, 16]. The corresponding result \(A(\varepsilon u) \in W^{1,2}(\Omega )\) or even \(A(\varepsilon u) \in W^{1,2}_{{\textrm{loc}}}(\Omega )\) is unsolved for the symmetric p-Laplace system. The only known result on regularity transfer from the data to the stress can be found in [19]. There it is shown that if \(F \in L^q(\Omega )\) with \(p \le q < 3p\) and \(n=3\), then \(A(\varepsilon u) \in L^q_{{\textrm{loc}}}(\Omega )\). This is even shown for the p-Stokes system.
Let us make a few final remarks on the related p-Stokes system, where an additional pressure and incompressibility constraint appear. Here \(V(\varepsilon u)\in W^{1,2}(\Omega )\) is only known for systems with quadratic growth. In the language of N-functions this means \(\varphi ''(t) \eqsim 1\), see e.g. [25]. For \(1<p\le 2\) and \(\delta >0\) one has \(V(\varepsilon u) \in W^{1,q}(\Omega )\) for some \(q <2\), see [10].
The rest of the paper is organized as follows. In Sect. 2 we introduce uniformly convex N-functions and formulate our main result in the language of Orlicz spaces. In Sect. 3 we derive Korn’s and Caccioppoli’s inequalities in our setting. In Sect. 4 we prove the main result. We do this by showing it first for N-functions with quadratic growth (Sect. 4.2) and then use an approximation argument to prove the general case (Sect. 4.3). In the “Appendix A” we collect some elementary properties of uniformly convex N-functions that are needed throughout the proofs.
2 Objective and main result
Our main focus is the study of system (1.1). This fits nicely in the more general framework of systems with Orlicz growth. In this section we introduce the corresponding partial differential system and the necessary background on Orlicz spaces and N-functions.
2.1 Uniformly convex Orlicz spaces
We begin with a bit of standard notation. With \(a\lesssim b\) we mean that there exists a constant \(c>0\) independent of all important quantities such that \(a\le cb\). If \(a\lesssim b\) and \(b\lesssim a\), then we write \(a\eqsim b\). We use the notation \(a\vee b{:}{=}\max (a,b)\) and \(a\wedge b{:}{=}\min (a,b)\). For a measurable set \(U\subset \mathbb {R}^n\) and \(f\in L^1 (\mathbb {R}^n)\) we denote by |U| the Lebesgue measure \({\mathcal {L}}^n(U)\) and define
For a ball \(B \subset \mathbb {R}^n\) let 2B denote the ball with the same center but twice the diameter. Usually, we denote the radius of a ball B by r.
An N-function is a function \(\varphi : \mathbb {R}^{\ge 0} \rightarrow \mathbb {R}^{\ge 0}\) which has a right-continuous, non-decreasing derivative \(\varphi ^\prime : \mathbb {R}^{\ge 0} \rightarrow \mathbb {R}^{\ge 0}\), such that \(\varphi ^\prime (0)=0\), \(\varphi ^\prime (t)>0\) for \(t>0\), \(\lim \limits _{t\rightarrow \infty }\varphi ^\prime (t)=\infty \) and \(\varphi (t)=\int _0^t\varphi ^\prime (s)\,ds\) for all \(t\ge 0\).
For every N-function \(\varphi \) we define its conjugate N-function \(\varphi ^*\) via
With this definition, \(\varphi ^*\) is again an N-function and we have \(\varphi ^{**}=\varphi \). If \(\varphi ^\prime \) is increasing we also have \((\varphi ^*)^\prime (s)=(\varphi ^\prime )^{-1}(s)\) for all \(s\ge 0\).
For an N-function \(\varphi \) and open set \(\Omega \subset \mathbb {R}^n\) we define the Orlicz space \(L^\varphi (\Omega )\) as the set of all measurable functions \(v:\Omega \rightarrow \mathbb {R}\) such that \(\Vert v\Vert _{L^\varphi (\Omega )}<\infty \), where
We define the Sobolev-Orlicz space \(W^{1,\varphi }(\Omega ){:}{=}\lbrace v \in L^\varphi (\Omega ): |\nabla v|\in L^{\varphi }(\Omega ) \rbrace \) equipped with the norm \(\Vert v\Vert _{W^{1,\varphi }(\Omega )}{:}{=}\Vert v\Vert _{L^\varphi (\Omega )}+\Vert \nabla v\Vert _{L^\varphi (\Omega )}\).
The systems that we study require that \(\varphi \) fulfills the following condition of uniform convexity, see [18].
Definition 2.1
(Uniform convexity) An N-function \(\varphi \) is called uniformly convex, if \(\varphi \in C^1([0,\infty ))\cap C^2((0,\infty ))\) and we have
We call \(p^-\) the lower and \(p^+\) the upper index (of uniform convexity) of \(\varphi \).
Example 2.2
Standard examples of uniformly convex N-functions are
-
(a)
\(\varphi (t)=\tfrac{1}{p}t^p\) with \(p\in (1,\infty )\). Then \(p^-=p^+=p\). This example motivates the definition of \(p^-\) and \(p^+\).
-
(b)
\(\varphi (t)=\int _0^t (\delta +s)^{p-2} s\,ds \eqsim (\delta +t)^{p-2}t^2\) with \(p \in (1,\infty )\) and \(\delta > 0\). Then \(p^-= p \wedge 2\) and \(p^+=p \vee 2\).
-
(c)
\(\varphi (t)=t^p + t^q\) with \(p,q \in (1,\infty )\). Then \(p^- = p \wedge q\) and \(p^+ = p \vee q\).
Remark 2.3
An example for a non uniformly convex \(\varphi \) is given by
where \(\mu \in (1,\infty )\). In this case we have \(p^+=2\) but \(p^-=1\). The interior regularity of the corresponding Euler-Lagrange equation with the symmetric gradient (and \(f=0\)) is studied in [26, 27].
Uniform convexity can also be defined for slightly less regular N-functions with \(\varphi ' \in W^{1,\infty }_{{\textrm{loc}}}((0,\infty ))\) instead of \(C^2((0,\infty ))\), see [18, Definition 28]. We collected some properties of uniformly convex N-functions in the “Appendix”, Section A.
2.2 Main result
In this section we introduce our system and present our main result on global regularity.
We denote by \(\partial _i\) the partial derivative and by \(\nabla = (\partial _1, \dots , \partial _n)\) the gradient of either a scalar function, a vector field or a tensor field. If \(A:\mathbb {R}^n\rightarrow \mathbb {R}^{n \times n}\) is a function then by \({\text {div}}A:\mathbb {R}^n\rightarrow \mathbb {R}^n\) we mean the row-wise divergence of A, i.e. \({\text {div}}A = \sum _{k=1}^n \partial _k A_{j,k}\). For matrices \(M,N\in \mathbb {R}^{n \times n}\) we define the Euclidean inner product \(A:B{:}{=}\sum _{i,j=1}^n A_{ij}B_{ij}\) and for vectors u, v we set \(u\otimes v{:}{=}(u_iv_j)_{ij}\).
For a vector field \(u:\mathbb {R}^n\rightarrow \mathbb {R}^n\) we define the symmetric gradient \({\varepsilon }u\) via
In the following let \(\varphi \) be a fixed uniformly convex N-function with indices of uniform convexity \(p^-\) and \(p^+\). We define the two important quantities A and V as
with \(A(P) = V(P)=0\) for \(P=0\). The quantities A, V and the N-function \(\varphi \) are closely related by the following formula
Moreover, we have
In (2.2) and (2.3) all implicit constants depend only on \(p^-\) and \(p^+\). These and other useful equivalences can be found in the “Appendix”, Section A.
Our main result is the following.
Theorem 2.4
Let \(\Omega \subset \mathbb {R}^n\) be an open and bounded domain with \(C^{2,1}\)-regular boundary. Let \(\varphi \) be a uniformly convex N-function and A and V defined as in (2.1). Then for every \(f\in W^{1,\varphi ^*}_0(\Omega )\) the weak solution \(u\in W^{1,\varphi }(\Omega )\) to the system of equations
satisfies \(V({\varepsilon }u)\in W^{1,2}(\Omega )\) together with the estimate
where \(c>0\) is a constant depending only on n, \(\Omega \) and the upper and lower indices of uniform convexity of \(\varphi \).
The system (2.4) is the Euler-Lagrange system to the energy (1.2). Note that Theorem 1.1 is a special case of Theorem 2.4 with \(\varphi (t) = \int _0^t (\delta +s)^{p-2}s\,ds\), \(p \in (1,\infty )\) and \(\delta \ge 0\).
3 Basic properties of weak solutions
In this section we begin with the basic properties of weak solutions from existence to Caccioppoli estimates. We also present Korn’s inequalities that are needed in the context of the symmetric gradient.
3.1 Korn’s inequality
We say that an N-function \(\varphi \) fulfills the \(\Delta _2\) condition (\(\varphi \in \Delta _2\)) if there exists a constant \(c>0\) such that for all \(t\ge 0\)
We denote by \(\Delta _2(\varphi )\) the smallest constant c which verifies inequality (3.1). If an N-function \(\varphi \) is uniformly convex then we have \(\varphi \in \Delta _2\) and \(\varphi ^*\in \Delta _2\).
We need the following Korn-type inequality in Orlicz spaces that can be found for example in [22, Theorem 6.13].
Proposition 3.1
(Korn inequality) Let \(\varphi \) be an N-function with \(\varphi , \varphi ^*\in \Delta _2\). Then for all balls \(B\subset \mathbb {R}^n\) and \(u\in W^{1,1}(B;\mathbb {R}^n)\) we have
The constants \(c_1\) depends only on \(\Delta _2(\varphi )\), \(\Delta _2(\varphi ^*)\) and n.
The condition \(\varphi ,\varphi ^*\in \Delta _2\) is actually necessary, see [5, Theorem 1.1]. As a simple consequence we will obtain the following version for partial zero boundary values.
Proposition 3.2
(Korn inequality; boundary) Let \(\varphi \) be an N-function with \(\varphi , \varphi ^*\in \Delta _2\). Let \(\Omega \) be a bounded Lipschitz domain. Let \(B \subset \mathbb {R}^n\) be a ball with center in \(\overline{\Omega }\) such that \(\Omega \cap B\) is again a Lipschitz domain. Then for all \(u\in W^{1,1}(B \cap \Omega ;\mathbb {R}^n)\) such that u vanishes on \(\partial \Omega \cap \overline{B}\) we have
The constant \(c_2\) depends only on \(\Delta _2(\varphi )\), \(\Delta _2(\varphi ^*)\), n and on the ratio \(|B \cap \Omega |/ |B|\).
Proof
We will reduce the claim to an application of Theorem 3.1. For this we extend u by zero to a \(W^{1,1}(B)\) function on the full ball B. Then with \(B^+ {:}{=}\Omega \cap B\)
We have
Combining this with (3.2), Theorem 3.1 and Jensen’s inequality we get
This proves the claim. \(\square \)
In similar fashion we have the following Poincaré type inequality.
Proposition 3.3
(Poincaré inequality) Let \(\varphi \) be an N-function with \(\varphi ,\varphi ^*\in \Delta _2\).
-
(a)
For all balls \(B\subset \mathbb {R}^n\) with radius r and \(u\in W^{1,1}(B;\mathbb {R}^n)\) we have
$$\begin{aligned} \int _B \varphi \left( |u-\langle {u}\rangle _B|\right) \,dx \le c_1 \int _B \varphi (r|\nabla u|)\,dx. \end{aligned}$$ -
(b)
Let \(B \subset \mathbb {R}^n\) be a ball with radius r and center in \(\overline{\Omega }\) such that \(\Omega \cap B\) is again a Lipschitz domain. Then for all \(u\in W^{1,1}(B \cap \Omega ;\mathbb {R}^n)\) such that u vanishes on \(\partial \Omega \cap \overline{B}\) we have
$$\begin{aligned} \int _{B \cap \Omega }\varphi (|u|)\,dx\le c_2 \int _{B \cap \Omega }\varphi (r|\nabla u|)\,dx. \end{aligned}$$
The constants \(c_1\), \(c_2\) depend only on \(\Delta _2(\varphi )\), \(\Delta _2(\varphi ^*)\), n and on the ratio \(|B \cap \Omega |/ |B|\).
The Korn and Poincaré inequality enable us to work in spaces depending on the full gradient instead of the symmetric gradient. In case of non-uniformly convex \(\varphi \) optimal embeddings for the spaces with the symmetric gradient can be found in [2].
Note that Propositions 3.1, 3.2 and 3.3 also hold for cylinders instead of balls. We will use them in this form later.
3.2 Weak solutions
For every \(f \in L^{\varphi ^*}(\Omega )\) there exists a unique weak solution \(u \in W_0^{1,\varphi }(\Omega )\) of system (2.4) in the sense that for all \(\psi \in W_0^{1,\varphi }(\Omega )\) we have
Moreover, testing with \(\psi =u\) and using the Poincaré inequality we obtain the standard energy estimate
where \(c>0\) depends only on n, \(\Omega \), \(p^+\) and \(p^-\).
3.3 Caccioppoli estimates
In this subsection we derive a Caccioppoli-type inequality. Before that we recall some formulas for the symmetric gradient.
If \(u\in C^1(\mathbb {R}^n,\mathbb {R}^n)\) is a vector field and \(\eta \in C^1(\mathbb {R}^n,\mathbb {R})\) is a scalar function then
where for \(u,v\in \mathbb {R}^n\) we set
The set of smooth functions u with \({\varepsilon }u=0\) is given by the space of infinitesimal rigid motions
Lemma 3.4
(Caccioppoli inequality) Let u be the weak solution of system (2.4). We then have for every ball B with radius r and \(2B\subset \Omega \) the estimate
If \(B \subset \mathbb {R}^n\) is a ball with center in \(\overline{\Omega }\) such that \(\Omega \cap B\) is again a Lipschitz domain, then for all \(u\in W^{1,1}(B \cap \Omega ;\mathbb {R}^n)\) such that u vanishes on \(\partial \Omega \cap \overline{B}\) we have
The constants \(c_1,c_2>0\) depend only on \(p^-\), \(p^+\), n and \({|\Omega \cap B |}/{|B |}\).
Proof
We begin with the case \(2B \subset \Omega \). Let \(\eta \) be a cut-off function with \(\mathbb {1}_B \le \eta \le \mathbb {1}_{2B}\) and \({\Vert \nabla \eta \Vert }_\infty \lesssim 1/r\). For arbitrary \(\pi \in {\mathcal {R}}\) we use the test function \(\psi = \eta ^{p^+} (u-\pi )\). This gives us
Thus, using the properties of \(\eta \) we get
The first integral on the right hand side of (3.7) can be estimated using Young’s inequality (A.3) and Lemma A.2, giving us for every \(\lambda >0\)
Choosing \(\lambda >0\) small enough, we can absorb the first term into the left hand side. The second integral on the right hand side of (3.7) can directly be estimated using Young’s inequality (A.3). This shows (3.5).
The boundary estimate (3.6) follows in the same way using the test function \(\psi = \eta ^{p^+} u\). Note that \(\psi =0\) on \(\partial \Omega \), since \(u=0\) on \(\partial \Omega \). \(\square \)
4 Global regularity
In this section we prove our main result. Since the proof for interior regularity is similar but easier than the proof of boundary regularity, we focus in our presentation on the regularity at the boundary. First, we explain how we treat the boundary by flattening. Secondly, we show our main result for a subclass of Orlicz functions, which we call N-functions with quadratic growth, see Definition 4.1. Afterwards, we use an approximation argument to prove the general case.
4.1 Flattening of the boundary
We flatten the boundary by using local isomorphisms which change \(\Omega \) only in one direction. Similar to [10] we introduce the following notation. First we fix a point \(P\in \partial \Omega \) and define what it means for \(\partial \Omega \) to be locally \(C^{2,1}\)-regular at P. For this we fix a local coordinate system in which \(P=0\) holds. In this coordinate system we use for every \(x\in \mathbb {R}^n\) the notation \(x^\prime {:}{=}(x_1,...,x_{n-1})\). Furthermore, we denote by \(B'_r\) the \((n-1)\)-dimensional ball with radius r centered at 0.
We now say that \(\partial \Omega \) is locally \(C^{2,1}\)-regular at P if there exists \(0<r<1\) and a \(C^{2,1}\)-function \(a:B^\prime _r\rightarrow (-r,r)\) with \(a(0)=0\), such that \(\partial \Omega \) is locally the graph of a, i.e.
-
(a)
\(\Omega _P {:}{=}\lbrace x\in \mathbb {R}^n: x^\prime \in B^\prime _r,~ a(x^\prime )<x_n<a(x^\prime )+r\rbrace \subset \Omega \),
-
(b)
\(x\in \partial \Omega \cap \big (B^\prime _r\times (-r,r)\big ) \iff x_n=a(x^\prime )\).
Furthermore, we assume that the local coordinate system is chosen such that \(\nabla a (0)=0\) holds. This also implies that we can make \(\Vert a\Vert _{C^1(B^\prime _r)}\) arbitrarily small by choosing r small enough. Finally, we say \(\partial \Omega \) is \(C^{2,1}\)-regular, if it is locally \(C^{2,1}\)-regular at every point \(P\in \partial \Omega \).
For \(0<\lambda \le 1\) we define the scaled sets
We use the notation \(B^+{:}{=}\tfrac{1}{2}\Omega _P\) and \(2B^+{:}{=}\Omega _P\). We call them cylinders at the boundary. We use the convention that we call the first \((n-1)\) directions the tangential directions and denote them with Greek letters. The n-th direction we call normal direction. Note however, that while the n-th unit vector \(e_n\) is indeed orthogonal to \(\partial \Omega \) at P, it is only approximately orthogonal to \(\partial \Omega \) at points near P. This is due to the curvature of the boundary.
Next we define curved partial derivatives along the boundary. For this we fix a tangential direction \(e_\alpha \) and define for a function \(f:\Omega _P\rightarrow \mathbb {R}\) and \(h>0\) the tangential translation via
For vector- or matrix-valued functions we apply this definition component wise. If \(f\in W^{1,1}(\Omega )\) with \({\text {supp}}f\subset \Omega _P\) we have for almost every \(x\in \Omega _P\)
In case of a (locally) flat boundary this coincides with the partial derivative in the tangential direction \(e_\alpha \). Otherwise, the second term accounts for the fact that we take the limit of small steps parallel to the boundary and not in the direction of any fixed unit vector. We now collect some formulas for the curved derivatives \(\widetilde{\partial }\).
Firstly, the curved derivatives \(\widetilde{\partial }_\alpha \) do not commute with the standard partial derivatives. A straight-forward calculation reveals that for \(f\in W^{1,1}(\Omega ,\mathbb {R}^n)\) we have in \(\Omega _P\)
If \(f,g \in W^{1,1}(\Omega )\) with \({\text {supp}}f \cap {\text {supp}}g \subset \Omega _P\) then we have the integration by parts formula
Later on we will frequently use cut-off functions to localize either on a ball in the interior of \(\Omega \) or on a cylinder at the boundary of \(\Omega \).
Let B be a ball such that \(2B\subset \Omega \). Then we choose \(\eta =\eta _B\in C_c^\infty (2B)\) such that
where \(C>0\) depends only on the dimension n.
For cylinders at the boundary we choose a cut-off function \(\eta \in C^\infty (2B^+,[0,1])\) such that \(\mathbb {1}_{B^+}\le \eta \le \mathbb {1}_{2B^+}\) and the estimate in (4.3) holds. Note that with this definition \(\eta \) does not vanish on \(\partial \Omega \).
4.2 Boundary regularity for systems with quadratic growth
Our main result is proved for general uniformly convex N-functions. While it is in principle possible to work directly in this generality, it is difficult to justify some steps. It is more convenient to work with N-functions that still have some kind of quadratic growth in the sense that \(\delta \le \varphi ''(t) \le \delta ^{-1}\) for small \(\delta >0\). The corresponding partial differential equation then has linear growth. This allows us to work in the framework of \(u \in W^{2,2}(\Omega )\) simplifying the calculations. The estimates that we will derive will be independent of the choice of \(\delta \) and depend only on \(p^-\) and \(p^+\). This allows us later by some approximation process of the N-function \(\varphi \) to cover the general case of uniformly convex N-functions.
Definition 4.1
(N-functions with quadratic growth) We say that a uniformly convex N-function \(\varphi \) has quadratic growth if there exists a \(\delta >0\) such that for all \(t\ge 0\) we have
If \(\varphi \) has quadratic growth, then \(\frac{1}{2} \delta t^2 \le \varphi (t) \le \frac{1}{2}\delta ^{-1}t^2\) and hence \(W^{1,\varphi }(\Omega )=W^{1,2}(\Omega )\) with equivalent norms.
Note that in (2.2) and (2.3) all constants depend on \(\varphi \) only through \(p^-\) and \(p^+\). Using a limiting process we get for every \(u\in W^{2,2}(\Omega )\) and almost every \(x\in \Omega \) the equivalences
where the implicit constants again only depend on \(p^-\), \(p^+\) and n. This also holds for the curved derivatives in the sense that for every cylinder \(B^+\) at the boundary and \(u\in W^{2,2}(B^+,\mathbb {R}^n)\)
holds pointwise almost everywhere in \(B^+\) with implicit constants depending only on \(p^-\), \(p^+\) and n.
The goal is to show the main result for N-function with quadratic growth.
Proposition 4.2
Let \(\varphi \) be a uniformly convex N-function with quadratic growth, see Definition 4.1. Let \(\Omega \subset \mathbb {R}^n\) be a bounded \(C^{2,1}\)-regular domain and \(f\in W^{1,2}(\Omega ) \cap W^{1,\varphi ^*}_0(\Omega )\). Then there exists a unique weak solution \(u\in W_0^{1,2}(\Omega )\cap W^{2,2}(\Omega )\) to system (2.4) which fulfills \(V({\varepsilon }u) \in W^{1,2}(\Omega )\) with the estimate
where \(c>0\) depends only on \(\Omega \), n and the indices \(p^-\) and \(p^+\) of \(\varphi \).
We already know that (2.4) admits a unique weak solution. Since \(\varphi \) has quadratic growth our system is of 2-structure in the sense of [11, Definition 2.1]. Thus, using [11, Theorem 1.1] we get even more regularity, namely
Note that this implies
In order to prove Proposition 4.2, we need to control \(\partial _i V({\varepsilon }u)\) for every direction \(1\le i \le n\). In Lemma 4.3 we do this for all directions in the interior and for all tangential derivatives \(\partial _\alpha \) at the boundary. The main idea is to test the weak equation (3.3) with \(\psi =\partial _{\alpha \alpha }u\). Since this is not a test function, we have to multiply it with a cut-off function.
For the normal direction \(i=n\) we can not use the same approach, since \(\partial _{nn}u\) does not vanish on \(\partial \Omega \). Instead, we show in Lemma 4.5 that we can use the pointwise equation (4.7) to bound the normal derivative \(\partial _n V({\varepsilon }u)\) by the tangential derivatives. Lemma 4.4 is an auxiliary Lemma that we need for the proof of Lemma 4.5.
Lemma 4.3
Under the assumptions of Proposition 4.2 we have for every ball B with \(2B\subset \Omega \) and every \(1\le i \le n\)
where \(c>0\) depends only on n, \(p^-\) and \(p^+\). Furthermore, if \(B^+\) is a cylinder at the boundary of \(\Omega \), \(1\le \alpha \le n-1\), \(\eta \) a cut-off function supported on \(2B^+\) (see Sect. 4.1) and \(\kappa >0\), then
Here \(c>0\) depends only on n, \(p^-\) and \(p^+\).
Proof
Let us first fix a cylinder \(B^+\) at the boundary of \(\Omega \). Multiplying the pointwise equation (4.7) with \(\psi {:}{=}\widetilde{\partial }_\alpha (\eta ^2 \widetilde{\partial }_\alpha u)\) and integrating over \(\Omega \) we get
We estimate the right hand side in (4.10) using the integration by parts formula (4.2), Young’s inequality and Korn’s inequality:
The left hand side in (4.10) can be rewritten using (4.1) as
We have
and thus
In the second step we estimated the second summand using the pointwise estimate \(\varphi '(|{\varepsilon }u|)|\nabla u|\le \varphi '(|\nabla u|)|\nabla u|\lesssim \varphi (|\nabla u|)\) and Korn’s inequality. The remaining first summand can be estimated with the help of Young’s inequality: For every \(\lambda >0\) we have
We now turn to \(T_1\). By (4.1) we have
Using (4.5) we have
By the integration by parts formula (4.2) we get
and thus
These terms can be estimated the same way we estimated \(T_2\) in (4.11). Note that this is the only step where we need the full \(C^{2,1}\)-regularity of the boundary. Similarly, we have for \(T_{1,3}\)
and thus
Again, these terms can be estimated the same way as in (4.11). Summarizing our findings so far (and choosing \(\lambda \) small enough), we have shown
To conclude (4.9) from here we simply note that
Thus, (4.9) follows. The interior estimate (4.8) can be shown in the same way but with many simplifications, since we do not need to use the curved derivatives \(\widetilde{\partial }_\alpha \). \(\square \)
Lemma 4.4
Let \(1\le i,j \le n\) with \((i,j)\ne (n,n)\) and \(\kappa \in (0,1)\). Then
where c depends only on n, \(p^-\) and \(p^+\).
Proof
Assume \(i\ne n\). The case \(j\ne n\) can be treated in exactly the same way. By (4.4) we have for every \(\kappa >0\)
This finishes the proof. \(\square \)
Lemma 4.5
Let \(2B^+\) be a cylinder at the boundary and \(\eta \) a cut-off function supported on \(2B^+\). Then
where \(c>0\) depends on \(p^-\), \(p^+\), n, \(r\Vert \nabla ^2 a\Vert _{\infty }\) and \(r^2\Vert \nabla ^3 a\Vert _{\infty }\).
Proof
We adopt the method used in [36]. Recall that we use the convention that the greek indices \(\alpha \) and \(\beta \) run only over the tangential directions \(1,...,n-1\). We have
We start by estimating \(S_2\). Using the pointwise equation (4.7) we have
The first sum can be estimated using Lemmas 4.4 and 4.3. To estimate the second sum we use integration by parts. Because \(\eta ^2 f\) vanishes on all of \(\partial (2B^+)\), we get no boundary terms. Thus
Using Young’s inequality A.3, we can estimate both summands. Indeed, we have
and
We thus have shown
We can bound \(S_3\) using the same strategy. We are thus left with bounding \(S_1\). For this we use the identity
This identity allows us to split \(S_1\) into three terms, the first two of which can again be estimated using Lemmas 4.4 and 4.3. This leaves us only with the third term, which we can write as
where
Here we used integration by parts to move the \(\partial _n\) away from \(A({\varepsilon }u)\) and the \(\partial _\alpha \) away from \(u_n\), and the fact that \(\partial _n u_n={\varepsilon }_{nn}u\). We have
Therefore, by Young’s inequality, we have for all \(\lambda >0\)
The term \(A_4\) can be estimated in exactly the same way.
For \(A_2\) we have
Both terms can be estimated with the help of Lemma 4.3. Collecting our estimates for \(A_1\),...,\(A_4\), we get the claimed estimate for \(S_1\) which together with the estimates for \(S_2\) and \(S_3\) completes the proof. \(\square \)
We are now ready to prove the main Proposition of this section.
Proof of Proposition 4.2
For every cylinder \(B^+\) at the boundary and every cut-off function \(\eta \) supported on \(2B^+\) we know by Lemmas 4.3 and 4.5 that for every \(\kappa >0\) we have
If we choose \(\kappa \) and \(B^+\) sufficiently small we can assume that
With this assumption and using \(r<{\text {diam}}(\Omega )\) and Lemma A.2 we get
The same bound holds for all interior balls B with \(2B\subset \Omega \).
We now cover \(\Omega \) by choosing a finite collection of balls \((B_i)_{i\in I}\) resp. cylinders at the boundary \((B_j^+)_{j\in J}\) such that
-
(a)
The \(B_i\) together with the \(B_j^+\) cover \(\Omega \),
-
(b)
\(2B_i\subset \Omega \) resp. \(2B_j^+\subset \Omega \) holds for all \(i\in I\) resp. \(j\in J\),
-
(c)
There exists a natural number \(C=C(n)\) such that every \(x\in \Omega \) is contained in at most C of the \(B_i\) or \(B_j^+\).
Additionally we assume that (4.15) holds for all of the \(B_j^+\). We denote by \(\eta _i\) resp. \(\eta _j\) the cut-off function corresponding to \(B_i\) resp. \(B_j^+\). We have
Using (4.16) and the finite overlap of our covering we get
Together with the energy estimate (3.4) this implies the claimed bound. \(\square \)
4.3 Passage to the limit
Proposition 4.2 already contains the results of Theorem 2.4 but is restricted to N-functions with quadratic growth. In this section we will show how to approximate every uniformly convex N-function by those with quadratic growth and how to pass to the limit.
For this let \(\varphi \) be a uniformly convex N-function. For every \(0 \le \delta ^- \le \delta ^+ \le \infty \) and \(\delta {:}{=}(\delta ^-,\delta ^+)\) we define the truncated N-functions \(\varphi _\delta =\varphi _{(\delta ^-,\delta ^+)}\) via
and \(\varphi _\delta (t){:}{=}\int _0^t \varphi ^\prime _\delta (s)\,ds\). These functions were introduced in [18]. See the “Appendix” for more details. For example by Lemma A.6 we know that \(\varphi _\delta \) is a uniformly convex N-function with quadratic growth.
Accordingly, we define for \(Q\in \mathbb {R}^{n \times n}\)
We can apply Proposition 4.2 only to systems where f is contained in \(W^{1,2}(\Omega )\). Thus we also need to approximate f. For this we perform a Lipschitz truncation, i.e. we approximate f by functions \(f_\delta \in W^{1,\infty }(\Omega )\) which agree with f outside of a small set.
As an auxiliary tool we need the Hardy-Littlewood maximal operator
To approximate f we will use the following approximation result.
Lemma 4.6
(Lipschitz truncation) Let \(\Omega \subset \mathbb {R}^n\) be a Lipschitz domain. Let \(\varphi \) be a uniformly convex N-function. Let \(v \in W^{1,\varphi }_0(\Omega )\). Then for all \(\lambda >0\) there exists \(T_\lambda v \in W^{1,\infty }_0(\Omega )\) with the following properties
-
(a)
\({\{{v \ne T_\lambda v}\}} \subset {\{{M(\nabla v)>\lambda }\}}\).
-
(b)
\({\Vert T_\lambda v \Vert }_{L^\varphi (\Omega )} \lesssim {\Vert v \Vert }_{L^\varphi (\Omega )}\) and \(\int _\Omega \varphi ({|T_\lambda v |})\,dx \lesssim \int _\Omega \varphi ({|v |})\,dx\).
-
(c)
\({\Vert \nabla T_\lambda v \Vert }_{L^\varphi (\Omega )} \lesssim {\Vert \nabla v \Vert }_{L^\varphi (\Omega )}\) and \(\int _\Omega \varphi ({|\nabla T_\lambda v |})\,dx \lesssim \int _\Omega \varphi ({|\nabla v |})\,dx\).
-
(d)
\({|\nabla T_\lambda v |} \lesssim \lambda \mathbb {1}_{{\{{M(\nabla v)>\lambda }\}}} + {|\nabla v |}_{\Omega {\setminus } {\{{M(\nabla v)> \lambda }\}}} \le \lambda \) almost everywhere.
-
(e)
\({\Vert \nabla (v-T_\lambda v) \Vert }_{L^\varphi (\Omega )} \lesssim {\Vert \nabla v \Vert }_{L^\varphi (\Omega )}\).
-
(f)
\(\int _\Omega \varphi ({|\nabla (v-T_\lambda v) |})\,dx \lesssim \int _\Omega \mathbb {1}_{{\{{M(\nabla v)>\lambda }\}}} \varphi ({|\nabla v |})\,dx\).
-
(g)
\(T_\lambda v \rightarrow v\) in \(W^{1,\varphi }_0(\Omega )\) as \(\lambda \rightarrow \infty \).
The hidden constants only depend on the indices \(p^-\) and \(p^+\) of \(\varphi \) and \(\Omega \).
Proof
The uniform convexity of \(\varphi \) ensures that \(1<p^- \le p^+ < \infty \). Now, the proof follows exactly as in [7, 20]. However, the boundedness of the maximal operator M in \(L^p\) has to be replaced by the corresponding results on Orlicz spaces, see [30]. \(\square \)
We are now ready to prove Theorem 2.4.
Proof of Theorem 2.4
In order to apply Proposition 4.2 we approximate f by applying the above Lipschitz truncation \(f_\delta {:}{=}T_\lambda f \in W^{1,\infty }_0(\Omega )\) on the level \(\lambda =c\,\varphi ^\prime (\delta ^+)\), so that \({|\nabla f_\delta |} \le \varphi '(\delta ^+)\). The level \(\lambda = c\varphi ^\prime (\delta ^+)\) is chosen in such a way that the truncation of \(\varphi \) at \(\delta ^{+}\) (near \(\infty \)) does not see the difference between \(\nabla f_\delta \) and \(\nabla f\).
By Proposition 4.2 we know that for every \(0<\delta ^- \le \delta ^+<\infty \) there exists a unique solution \(u_\delta \in W_0^{1,\varphi _\delta }(\Omega )\) such that
and (also using the Orlicz-Poincaré inequality Propositon 3.3)
where c is independent of \(\delta \). In a moment, it will turn out that \(K_\delta \) can be bounded independently of \(\delta \). Using \({|\nabla f_\delta |} \le \varphi '(\delta ^+)\), Lemmas A.6 (d) and 4.6 we have
The first term vanishes as \(\delta ^-\rightarrow 0\). Thus, combining this with (4.18), we have uniformly in \(\delta \)
In the following we assume \(\delta ^-\le 1\) and \(\delta ^+ \ge 1\) and are interesting in the limit \(\delta ^- \rightarrow 0\) and \(\delta ^+\rightarrow \infty \). Since \({\Vert V_\delta (\varepsilon u_\delta ) \Vert }_{L^2(\Omega )}\) is uniformly bounded, so is \({\Vert \varepsilon u_\delta \Vert }_{L^{\varphi _\delta }(\Omega )}\). Setting \(s{:}{=}p^- \wedge 2\), Lemma A.7 implies that the \(u_\delta \) are uniformly bounded in \(W^{1,s}(\Omega )\). Thus, after passing to a (non relabeled) subsequence we have for some \(u^*\in W^{1,s}(\Omega )\)
In the following we will show that \(u^*\) fulfills the regularity estimate (2.5) and then that it is indeed the weak solution to the system (2.4).
From (4.19) we can assume that there exists \(Q\in W^{1,2}(\Omega )\) such that
Using Lemma A.9, the pointwise convergence in (4.21) implies that \({\varepsilon }u_\delta =V_\delta ^{-1}\circ V_\delta ({\varepsilon }u_\delta )\) converges pointwise almost everywhere to \(V^{-1}(Q)\). Since we also know from (4.20) that \({\varepsilon }u_\delta \rightharpoonup {\varepsilon }u^*\) weakly in \(L^s\), we have \(V({\varepsilon }u^*)=Q\). We can now use the first convergence in (4.21) together with the uniform bound (4.19) to conclude
We are left to show that \(u^*\) is the weak solution of the system (2.4). Using Lemma A.9 we know that \(A_\delta ({\varepsilon }u_\delta ) \rightarrow A({\varepsilon }u^*)\) almost everywhere. Note that
Thus, by Lemma A.7, \(\varepsilon u_\delta \) is uniformly bounded in \(L^{p^- \wedge 2}(\Omega )\) and \(A_\delta ({\varepsilon }u_\delta )\) uniformly bounded in \(L^{(p^+)' \wedge 2}(\Omega )\). This together with the almost everywhere convergence imply convergence in \(L^1\) (for a subsequence). Thus, \(A_\delta ({\varepsilon }u _\delta ) \rightarrow A({\varepsilon }u^*)\) in \(L^1(\Omega )\). This implies for all \(\psi \in C_c^\infty (\Omega )\)
This shows that \(u^*\) indeed solves (2.4) in the weak sense. Therefore, (4.22) gives us the claimed bound (2.5), which finishes the proof. \(\square \)
Finally, we prove Theorem 1.1.
Proof of Theorem 1.1
We restrict our proof to the difficult case \(\delta =0\). The simpler case \(\delta >0\) can be treated in the same way. The claimed estimate follows immediately from Theorem 2.4. We are thus left to prove the claimed Sobolev regularity of u. In the case \(p\ge 2\) we note that \(V({\varepsilon }u) = |{\varepsilon }u|^{\tfrac{p-2}{2}}{\varepsilon }u\) implies
Using [8, Proposition 4.4]Footnote 1 we get that \({\varepsilon }u \in W^{2/p,p}\) and thus by Korn’s inequality \(u \in W^{1+2/p,p}\).
For \(p<2\) we first note that by Sobolevs embedding we have
Setting \(r{:}{=}\tfrac{np}{n+p-2}\) we get
We thus have \(\nabla {\varepsilon }u \in L^r(\Omega )\) which implies \(u \in W^{2,r}(\Omega )\). \(\square \)
Data availability
Data sharing is not applicable to this article as no new data were created or analyzed in this study.
References
Acerbi, E., Fusco, N.: Regularity for minimizers of nonquadratic functionals: the case \(1<p<2\). J. Math. Anal. Appl. 140(1), 115–135 (1989). https://doi.org/10.1016/0022-247X(89)90098-X
Breit, D., Cianchi, A.: Symmetric gradient Sobolev spaces endowed with rearrangement-invariant norms. Adv. Math. 391, 107954 (2021). https://doi.org/10.1016/j.aim.2021.107954
Balci, A.K., Cianchi, A., Diening, L., Maz’ya, V.: A pointwise differential inequality and second-order regularity for nonlinear elliptic systems. Math. Ann. (2021). https://doi.org/10.1007/s00208-021-02249-9
Breit, D., Cianchi, A., Diening, L., Schwarzacher, S.: Global Schauder estimates for the \(p\)-Laplace system. Arch. Ration. Mech. Anal. 243(1), 201–255 (2022). https://doi.org/10.1007/s00205-021-01712-w
Breit, D., Diening, L.: Sharp conditions for Korn inequalities in Orlicz spaces. J. Math. Fluid Mech. 14(3), 565–573 (2012). https://doi.org/10.1007/s00021-011-0082-x
Berselli, L.C., Diening, L., Růžička, M.: Existence of strong solutions for incompressible fluids with shear dependent viscosities. J. Math. Fluid Mech. 12(1), 101–132 (2010). https://doi.org/10.1007/s00021-008-0277-y
Bulíček, M., Diening, L., Schwarzacher, S.: Existence, uniqueness and optimal regularity results for very weak solutions to nonlinear elliptic systems. Anal. PDE 9(5), 1115–1151 (2016). https://doi.org/10.2140/apde.2016.9.1115
Balci, A.K., Diening, L., Weimar, M.: Higher order Calderón–Zygmund estimates for the \(p\)-Laplace equation. J. Differ. Equ. 268(2), 590–635 (2020). https://doi.org/10.1016/j.jde.2019.08.009
Beirão da Veiga, H., Kaplický, P., Růžička, M.: Boundary regularity of shear thickening flows. J. Math. Fluid Mech. 13(3), 387–404 (2011). https://doi.org/10.1007/s00021-010-0025-y
Berselli, L.C., Růžička, M.: Global regularity properties of steady shear thinning flows. J. Math. Anal. Appl. 450(2), 839–871 (2017). https://doi.org/10.1016/j.jmaa.2017.01.016
Berselli, L.C., Růžička, M.: Global regularity for systems with \(p\)-structure depending on the symmetric gradient. Adv. Nonlinear Anal. 9(1), 176–192 (2020). https://doi.org/10.1515/anona-2018-0090
Berselli, L.C., Růžička, M.: Natural second-order regularity for parabolic systems with operators having \((p,\delta )\)-structure and depending only on the symmetric gradient. Calc. Var. Partial Differ. Equ. 61(4), 137 (2022). https://doi.org/10.1007/s00526-022-02247-y
Brasco, L., Santambrogio, F.: A sharp estimate à la Calderón–Zygmund for the \(p\)-Laplacian. Commun. Contemp. Math. 20(3), 1750030 (2018). https://doi.org/10.1142/S0219199717500304
Chen, Y.Z., DiBenedetto, E.: Boundary estimates for solutions of nonlinear degenerate parabolic systems. J. Reine Angew. Math. 395, 102–131 (1989)
Cianchi, A., Maz’ya, V.G.: Global boundedness of the gradient for a class of nonlinear elliptic systems. Arch. Ration. Mech. Anal. 212(1), 129–177 (2014). https://doi.org/10.1007/s00205-013-0705-x
Cianchi, A., Maz’ya, V.G.: Optimal second-order regularity for the \(p\)-Laplace system. J. Math. Pures Appl. 132(9), 41–78 (2019). https://doi.org/10.1016/j.matpur.2019.02.015
Diening, L., Ettwein, F.: Fractional estimates for non-differentiable elliptic systems with general growth. Forum Math. 20(3), 523–556 (2008). https://doi.org/10.1515/FORUM.2008.027
Diening, L., Fornasier, M., Tomasi, R., Wank, M.: A relaxed Kačanov iteration for the \(p\)-Poisson problem. Numer. Math. 145(1), 1–34 (2020). https://doi.org/10.1007/s00211-020-01107-1
Diening, L., Kaplický, P.: \(L^q\) theory for a generalized Stokes system. Manuscr. Math. 141(1–2), 333–361 (2013). https://doi.org/10.1007/s00229-012-0574-x
Diening, L., Kreuzer, C., Süli, E.: Finite element approximation of steady flows of incompressible fluids with implicit power-law-like rheology. SIAM J. Numer. Anal. 51(2), 984–1015 (2013). https://doi.org/10.1137/120873133
Diening, L., Kaplicky, P., Schwarzacher, S.: Campanato estimates for the generalized Stokes system. Ann. Mat. Pura Appl. 193(6), 1779–1794 (2014). https://doi.org/10.1007/s10231-013-0355-5
Diening, L., Růžička, M., Schumacher, K.: A decomposition technique for John domains. Ann. Acad. Sci. Fenn. Math. 35(1), 87–114 (2010). https://doi.org/10.5186/aasfm.2010.3506
Diening, L., Stroffolini, B., Verde, A.: Everywhere regularity of functionals with \(\phi \)-growth. Manuscr. Math. 129(4), 449–481 (2009). https://doi.org/10.1007/s00229-009-0277-0
Fiorenza, A., Krbec, M.: Indices of Orlicz spaces and some applications. Comment. Math. Univ. Carolin. 38(3), 433–451 (1997)
Fuchs, M., Seregin, G.: Variational methods for problems from plasticity theory and for generalized Newtonian fluids. In: Lecture Notes in Mathematics, vol. 1749. Springer, Berlin, pp. vi+269 (2000). https://doi.org/10.1007/BFb0103751
Gmeineder, F., Kristensen, J.: Sobolev regularity for convex functionals on BD. Calc. Var. Partial Differ. Equ. 58(2), 56 (2019). https://doi.org/10.1007/s00526-019-1491-6
Gmeineder, F.: The regularity of minima for the Dirichlet problem on BD. Arch. Ration. Mech. Anal. 237(3), 1099–1171 (2020). https://doi.org/10.1007/s00205-020-01507-5
Kuusi, T., Mingione, G.: Vectorial nonlinear potential theory. J. Eur. Math. Soc. 20(4), 929–1004 (2018). https://doi.org/10.4171/JEMS/780
Kaplický, P., Málek, J., Stará, J.: On global existence of smooth two-dimensional steady flows for a class of non-Newtonian fluids under various boundary conditions. In: Applied Nonlinear Analysis, pp. 213–229. Kluwer/Plenum, New York (1999)
Krasnosel’skiĭ, M.A., Rutickiĭ, J.B.: Convex Functions and Orlicz Spaces, pp. xi+249. Russian. P. Noordhoff Ltd., Groningen (1961)
Málek, J., Nečas, J., Rokyta, M., Růžička, M.: Weak and measure-valued solutions to evolutionary PDEs. In: Applied Mathematics and Mathematical Computation, vol. 13, pp. xii+317. Chapman & Hall, London (1996). https://doi.org/10.1007/978-1-4899-6824-1
Morrey Jr, C.B.: Multiple integrals in the calculus of variations. In: Classics in Mathematics. Reprint of the 1966 edition [MR0202511], pp. x+506. Springer, Berlin (2008). https://doi.org/10.1007/978-3-540-69952-1
Růžička, M., Diening, L.: Non-Newtonian fluids and function spaces. In: Rákosník, J. (ed.) Nonlinear Analysis, Function Spaces and Applications. Proceedings of the Spring School held in Prague, May 30–June 6, vol. 8, pp. 95–143. Institute of Mathematics of the Academy of Sciences of the Czech Republic, 2007 (2006)
Serëgin, G.A.: On the regularity of minimizers of some variational problems in the theory of plasticity. Algebra Anal. 4(5), 181–218 (1992)
Shvartsman, P.: Local approximations and intrinsic characterization of spaces of smooth functions on regular subsets of \({\mathbb{R} }^n\). Math. Nachr. 279(11), 1212–1241 (2006). https://doi.org/10.1002/mana.200510418
Seregin, G.A., Shilkin, T.N.: Regularity of minimizers of some variational problems in plasticity theory. Zap. Nauchn. Sem. S.-Peterburg. Otdel. Mat. Inst. Steklov. 243, 270–298 (1997). https://doi.org/10.1007/BF02673602
Uhlenbeck, K.: Regularity for a class of non-linear elliptic systems. Acta Math. 138(3–4), 219–240 (1977). https://doi.org/10.1007/BF02392316
Funding
Open Access funding enabled and organized by Projekt DEAL.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by L. Szekelyhidi.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors thank the anonymous referee for pointing out a mistake in an earlier version of this paper. The authors gratefully acknowledge financial support by the German Research Foundation through the IRTG 2235 (Project 282638148).
Appendix A. Uniformly convex N-functions
Appendix A. Uniformly convex N-functions
In this “Appendix” we collect some properties of uniformly convex N-functions (also called uniformly convex Young functions), see Definition 2.1. We will carefully ensure that all constants (implicit and explicit) in this section depend on the Orlicz function \(\varphi \) only through the indices \(p^-\) and \(p^+\).
1.1 A.1 Basic properties
Let us collect a few basic properties of uniformly convex N-functions. In the following let \(p'=\frac{p}{p-1}\) denote the conjugate exponent of p.
Lemma A.1
If \(\varphi \) is a uniformly convex N-function with indices \(p^-\) and \(p^+\), then \(\varphi ^*\) is also uniformly convex with indices \((p^+)^\prime \) and \((p^-)^\prime \).
Proof
This is shown in [18, Lemma 29] and [33, Lemma 6.4]. \(\square \)
The following lemma shows that the indices of uniform convexity provide an upper and lower bound for the Simonenko indices. See [24] for an overview on indices on Orlicz spaces.
Lemma A.2
(Simonenko indices) Let \(\varphi \) be uniformly convex with indices \(p^-\) and \(p^+\). Then
In particular,
Proof
The upper bound with \(p^+\) follows from
The lower bound with \(p^-\) follows for analogously.
The rest of the claim follows by the theory on Simonenko indices, but let us include a short proof. If \(t\ge 1\), then
We thus have \(\varphi (st)\le t^{p^+}\varphi (s)\) for \(t \ge 1\). The other bounds follow similarly. \(\square \)
Corollary A.3
(\(\Delta _2\) and Young’s inequality) Let \(\varphi \) be a uniformly convex N-function with indices \(p^-\) and \(p^+\). Then \(\varphi \) and \(\varphi ^*\) satisfy the \(\Delta _2\)-condition, i.e. for all \(t \ge 0\)
Moreover, for all \(s,t\ge 0\) and \(\delta \in (0,1]\)
Proof
The first part follows from Lemma A.2. The second part follows by Young’s inequality and Lemma A.2, i.e.
The other estimate follows by replacing \(\varphi \) with \(\varphi ^*\). \(\square \)
For N-functions \(\varphi \) one has for \(t\ge 0\)
As a consequence one obtains for \(t \ge 0\)
Hence, if \(\varphi \) is uniformly convex with indices \(p^-\) and \(p^+\), then for \(t \ge 0\)
Recall that by (2.1)
Remark A.4
As A is induced by \(\varphi \), V is induced by \(\psi \), where \(\psi \) is the N-function defined by
One easily calculates
This proves that \(\psi \) is a uniformly convex N-function and the indices of \(\psi \) and \(\varphi \) are related by the formula
The quantities A, V and \(\varphi \) are linked by the following important equivalence.
Lemma A.5
Let \(\varphi \) be a uniformly convex N-function. Then for all \(P,Q \in \mathbb {R}^{n \times n}\)
where the implicit constants depend only on the indices \(p^-\) and \(p^+\).
Proof
This Lemma goes back to [17, Lemma 3]. \(\square \)
1.2 A.2 Truncated N-functions
For \(0 \le \delta ^-\le \delta ^+\le \infty \) and \(\delta = (\delta ^-,\delta ^+)\) we define the truncated N-function \(\varphi _\delta \) by
These functions were first introduced in [18]. If \(\delta ^+=\infty \), then they are related to the shifted N-functions introduced in [17].
Lemma A.6
Let \(\varphi \) be a uniformly convex N-function with indices \(p^-\) and \(p^+\) and \(0<\delta ^-\le \delta ^+<\infty \). Then
-
(a)
\(\varphi _\delta \) is a uniformly convex N-function with quadratic growth, see Definition 4.1.
-
(b)
The lower index of \(\varphi _\delta \) is \(p^- \wedge 2\), and the upper index of \(\varphi _\delta \) is \(p^+\vee 2\).
-
(c)
\(\varphi ^\prime (t)=\varphi ^\prime _\delta (t)\) and \(\varphi ^{\prime \prime }(t)=\varphi ^{\prime \prime }_\delta (t)\) for all \(t\in (\delta ^-,\delta ^+)\).
-
(d)
We have
$$\begin{aligned} (\varphi _{\delta })^*= (\varphi ^*)_{\varphi ^\prime (\delta ^-),\varphi ^\prime (\delta ^+)}. \end{aligned}$$ -
(e)
For \(t\in [0,\delta ^+]\) we have
$$\begin{aligned} {|\varphi (t) - \varphi _\delta (t) |} \le 2^{p^+-1}\varphi (\delta ^-). \end{aligned}$$
Proof
By the definition of \(\varphi _\delta \) we have
and thus
These observations readily imply (a)–(c).
To see (d) we note that
This proves (d).
It remains to prove (e). If \(t \in [0,\delta ^+]\), then
This proves the last claim. \(\square \)
Lemma A.7
Let \(\varphi \) be a uniformly convex N-function with indices \(p^-\) and \(p^+\) and \(0<\delta ^- \le 1 \le \delta ^+<\infty \). Then
with embedding constants uniformly bounded with respect to \(\delta \).
Proof
It follows from Lemmas A.2 and A.6 (b) that
Note that using \(1< 2 \wedge p^- \le 2 \vee p^+ < \infty \) we calculate
Hence,
This proves the claim. \(\square \)
For \(Q \in \mathbb {R}^{n \times n}\) define
Remark A.8
In the sense of Remark A.4\(A_\delta \) is induced by \(\varphi _\delta \) and \(V_\delta \) is induced by \(\psi _\delta \) (the truncation of \(\psi \)) and \(\psi '_\delta (t) = \sqrt{\varphi _\delta '(t) t}\). Note that also the \(\psi _\delta \) are uniformly convex N-functions by Lemma A.6.
The following continuity is inspired by [6].
Lemma A.9
Let \(\varphi \) be a uniformly convex N-function. Then \((\delta ^-,\delta ^+,t)\mapsto \varphi ^\prime _\delta (t)\) is continuous on \([0,1]\times [1,\infty ] \times \mathbb {R}^{\ge 0}\). Furthermore, the mapping \((\delta ^-,\delta ^+,Q)\mapsto A_{\delta }(Q)\) is continuous \([0,1]\times [1,\infty ] \times \mathbb {R}^{n \times n}\). The same holds true if we replace \(A_\delta \) with \(A^{-1}_\delta \), \(V_\delta \) or \(V_\delta ^{-1}\).
Proof
Recall that \(\varphi '\) is continuous and that
This proves that \((\delta ^-,\delta ^+,t) \mapsto \varphi _\delta '(t)\) is continuous on \([0,1] \times [1,\infty ] \times [0,\infty )\). Analogously
shows that \(A_\delta (Q)\) is continuous. Using \((\varphi _\delta )^* = (\varphi ^*)_{\varphi '(\delta )}\) it follows that \(A_\delta ^{-1}(Q)\) is continuous. The continuity of \(V_\delta (Q)\) and \(V_\delta ^{-1}(Q)\) follows by Remark A.8. \(\square \)
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Behn, L., Diening, L. Global regularity for nonlinear systems with symmetric gradients. Calc. Var. 63, 60 (2024). https://doi.org/10.1007/s00526-024-02666-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00526-024-02666-z