Bernis estimates for higher-dimensional doubly-degenerate non-Newtonian thin-film equations

Christina Lienstromberg, Katerina Nik · 2024

Read Bernis estimates for higher-dimensional doubly-degenerate non-Newtonian thin-film equations by Christina Lienstromberg, Katerina Nik (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-15883.txt.

Introduction

Christina Lienstromberg and Katerina Nik proved key bounds for a thin film flow equation. Their paper, from twenty twenty-four, deals with a doubly degenerate, non-Newtonian thin film model. Such models track how a thin layer of fluid spreads when its flow rules are not the plain ones. The proof gives local Bernis estimates, a type of bound that ties high order terms to a base energy term.

The bounds hold for a set range of the two key powers in the model. One sharp result bounds the spread of a power of the film height by that base term. The pair adapted a known path from the plain case to this harder one. Readers in fluid math get tools to show these films stay smooth.

Catalog record

CREATORChristina Lienstromberg, Katerina Nik
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE81 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-15883

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Bernis estimates for higher-dimensional doubly-degenerate non-Newtonian thin-film equations Christina Lienstromberg, Katerina Nik Abstract . For the doubly-degenerate parabolic non-Newtonian thin-film equation u t + div ( u n | ∇ Δ u | p − 2 ∇ Δ u ) = 0 , u_{t}+\diverg\bigl(u^{n}|\nabla\Delta u|^{p-2}\nabla\Delta u\bigr)=0, we derive (local versions) of Bernis estimates of the form ∫ Ω u n − 2 ​ p | ∇ u | 3 ​ p d x + ∫ Ω u n − p 2 | Δ u | 3 ​ p 2 d x ≤ c ( n , p , d ) ∫ Ω u n | ∇ Δ u | p d x , \int_{\Omega}u^{n-2p}|\nabla u|^{3p}\;\mathrm{d}x+\int_{\Omega}u^{n-\frac{p}{2}}|\Delta u|^{\frac{3p}{2}}\;\mathrm{d}x\leq c(n,p,d)\int_{\Omega}u^{n}|\nabla\Delta u|^{p}\;\mathrm{d}x, for functions u ∈ W p 2 ​ ( Ω ) u\in W^{2}_{p}(\Omega) with Neumann boundary condition, where 2 ≤ p < 19 3 2\leq p<\frac{19}{3} and n n lies in a certain range. Here, Ω ⊂ ℝ d \Omega\subset\mathbb{R}^{d} is a smooth convex domain with d < 3 ​ p d<3p . A particularly important consequence is the estimate ∫ Ω | ∇ Δ ( u n + p p ) | p d x ≤ c ( n , p , d ) ∫ Ω u n | ∇ Δ u | p d x . \int_{\Omega}|\nabla\Delta(u^{\frac{n+p}{p}})|^{p}\;\mathrm{d}x\leq c(n,p,d)\int_{\Omega}u^{n}|\nabla\Delta u|^{p}\;\mathrm{d}x. The methods used in this article follow the approach of [ Grü01 ] for the Newtonian case, while addressing the specific challenges posed by the nonlinear higher-order term | ∇ Δ u | p − 2 ∇ Δ u |\nabla\Delta u|^{p-2}\nabla\Delta u and the additional degeneracy. The derived estimates are key to establishing further qualitative results, such as the existence of weak solutions, finite propagation of support, and the appearance of a waiting-time phenomenon. MSC (2020): 76A05, 76A20, 35A23, 46B70,35Q35, 35K35, 35K65 Keywords: non-Newtonian fluids, power-law fluids, fourth-order degenerate parabolic equations, thin-film equations, Bernis estimates 1. Introduction and main results 1.1. Background In this article, we derive so-called Bernis estimates for fourth-order doubly-degenerate parabolic problems of the form (1.1) \begin{cases}u_{t}+\diverg\bigl(u^{n}|\nabla\Delta u|^{p-2}\nabla\Delta u\bigr)=0&\quad\text{on }(0,\infty)\times\Omega\subset\mathbb{R}^{d+1},\\ \nabla u\cdot\nu=\nabla\Delta u\cdot\nu=0&\quad\text{in }(0,\infty)\times\partial\Omega,\\ u(0,\cdot)=u_{0}(\cdot)&\quad\text{in }\Omega,\end{cases} as they arise in the modeling of non-Newtonian thin-film flows with a strain rate-dependent power-law rheology. Problem ( 1.1 ) describes the evolution of the height u=u(t,x)\geq 0 of a thin layer of a viscous, non-Newtonian, and incompressible fluid over an impermeable flat bottom, as sketched in Figure 1 . The partial differential equation \eqref{eq:PDE}_{1} is a fourth-order quasilinear equation that is doubly-degenerate parabolic, meaning that parabolicity gets lost when either the film height u or the third-order term \nabla\Delta u vanishes. The term u^{n} represents the mobility of the flow, with the exponent n>0 determined by the choice of the slip condition at the lower boundary (i.e., the interface between the solid bottom and the fluid film). For a no-slip condition, we have n=p+1 , while for a Navier-slip condition, we have n=p . Other values of n\in(p-2,p+1) have been proposed to model weaker or stronger slippage; see, for example, [ Gre78 ] . The term |\nabla\Delta u|^{p-2}\nabla\Delta u has a dual physical interpretation: the higher-order term \nabla\Delta u accounts for surface tension effects (we assume that the dynamics of the flow is driven solely by surface tension, thus neglecting gravitational effects), while its appearance in the nonlinear form |s|^{p-2}s is due to the non-Newtonian power-law rheology of the fluid. The evolution equation \eqref{eq:PDE}_{1} is supplemented with Neumann-type boundary conditions \nabla u\cdot\nu=0 and \nabla\Delta u\cdot\nu=0 , which represent a zero-contact angle condition and a no-flux condition at the lateral boundary (0,\infty)\times\partial\Omega , respectively. Finally, u_{0} denotes the initial film height. x z y Figure 1. Sketch of a fluid film on an impermeable solid bottom. The natural energy inequality for non-negative solutions to ( 1.1 ) is given by (1.2) \frac{1}{2}\int_{\Omega}|\nabla u(t)|^{2}\;\mathrm{d}x+\int_{0}^{t}\int_{\Omega}u^{n}|\nabla\Delta u|^{p}\;\mathrm{d}x\;\mathrm{d}s\leq\frac{1}{2}\int_{\Omega}|\nabla u_{0}|^{2}\;\mathrm{d}x,\quad t\geq 0. However, even in the one-dimensional case or for the classical Newtonian thin-film equation ( p=2 ), this estimate on the dissipation \int_{\Omega}u^{n}|\nabla\Delta u|^{p}\;\mathrm{d}x is not sufficient for many purposes. Further estimates on the integrals \int_{\Omega}|\nabla\big(u^{\frac{n+p}{3p}}\big)|^{3p}\;\mathrm{d}x , \int_{\Omega}|\Delta\big(u^{\frac{2(n+p)}{3p}}\big)|^{\frac{3p}{2}}\;\mathrm{d}x , and \int_{\Omega}|\nabla\Delta\big(u^{\frac{n+p}{p}}\big)|^{p}\;\mathrm{d}x are needed to obtain qualitative results such as finite propagation of support or the occurrence of a waiting-time phenomenon. See Section 1.3 for more details on related results. 1.2. Aim of the article and main results The main goal of this article is to derive inequalities of the form \begin{split}\int_{\Omega}\varphi^{3p}u^{n-2p}|\nabla u|^{3p}\;\mathrm{d}x+\int_{\Omega}\varphi^{3p}u^{n-\frac{p}{2}}|\Delta u|^{\frac{3p}{2}}\;\mathrm{d}x\leq c(n,p,d)\left(\int_{\Omega}\varphi^{3p}u^{n}|\nabla\Delta u|^{p}\;\mathrm{d}x+\int_{\Omega}|\nabla\varphi|^{3p}u^{n+p}\;\mathrm{d}x\right)\end{split} for a given non-negative test function \varphi , and for positive functions u\in W^{2}_{p}(\Omega),\ 2\leq p<\frac{19}{3} , satisfying the Neumann boundary condition \nabla u\cdot\nu=0 on \partial\Omega and having finite dissipation \int_{\Omega}u^{n}|\nabla\Delta u|^{p}\;\mathrm{d}x<\infty . These are established in Theorem 1.1 below. From these inequalities, we can deduce further estimates of the form \begin{split}&\int_{\Omega}\Big|\nabla\big(u^{\frac{n+p}{3p}}\big)\Big|^{3p}\;\mathrm{d}x+\int_{\Omega}\Big|\Delta\big(u^{\frac{2(n+p)}{3p}}\big)\Big|^{\frac{3p}{2}}\;\mathrm{d}x+\int_{\Omega}\Big|\nabla\Delta\big(u^{\frac{n+p}{p}}\big)\Big|^{p}\;\mathrm{d}x\leq c(n,p,d)\int_{\Omega}u^{n}|\nabla\Delta u|^{p}\;\mathrm{d}x\end{split} for u as above. This result is proven in Theorem 1.3 below. These inequalities, established in Theorems 1.1 and 1.3 , provide a basis for applying appropriate interpolation results, which are in turn crucial for proving qualitative properties of solutions to ( 1.1 ), such as finite speed of propagation or waiting-time phenomena. Our results generalize the findings of [ Grü01 ] , where the Newtonian analogue of ( 1.1 ) is addressed, to non-Newtonian thin-film equations of the form ( 1.1 ). In the one-dimensional non-Newtonian case, similar Bernis inequalities were shown in [ AG04 ] in the parameter range p\geq\frac{4}{3} and \frac{p-1}{2}<n<2p-1 . In the following, we present the main results of this article, Theorems 1.1 and 1.3 , in detail. Theorem 1.1. Let \Omega\subset\mathbb{R}^{d} , d<3p , be a bounded, convex domain with smooth boundary \partial\Omega . Let 2\leq p<\frac{19}{3} and 2p-2\frac{2+\sqrt{(3p-2)^{2}+d(3p-4)}}{-3p+6+\sqrt{(3p-2)^{2}+d(3p-4)}}<n<2p-1. Assume u\in W^{2}_{p}(\Omega) is positive, satisfies \nabla u\cdot\nu=0 on \partial\Omega , and \int_{\Omega}u^{n}|\nabla\Delta u|^{p}\,dx<\infty. Then there exist positive constants c_{i}=c_{i}(n,p,d)>0 , i=1,2,3 , such that the estimates (1.3) \begin{split}\int_{\Omega}\varphi^{3p}u^{n-2p}|\nabla u|^{3p}\;\mathrm{d}x\leq c_{1}\left(\int_{\Omega}\varphi^{3p}u^{n}|\nabla\Delta u|^{p}\;\mathrm{d}x+\int_{\Omega}|\nabla\varphi|^{3p}u^{n+p}\;\mathrm{d}x\right)\end{split} and (1.4) \begin{split}&\int_{\Omega}\varphi^{3p}u^{n-2p+2}|\nabla u|^{3p-6}|D^{2}u\nabla u|^{2}\;\mathrm{d}x+\int_{\Omega}\varphi^{3p}u^{n-2p+2}|\nabla u|^{3p-4}|D^{2}u|^{2}\;\mathrm{d}x\\ &+\int_{\partial\Omega}\varphi^{3p}u^{n-2p+2}|\nabla u|^{3p-4}\mathrm{II}(\nabla u,\nabla u)\;\mathrm{d}\mathcal{H}^{d-1}\leq\ c_{2}\left(\int_{\Omega}\varphi^{3p}u^{n}|\nabla\Delta u|^{p}\;\mathrm{d}x+\int_{\Omega}|\nabla\varphi|^{3p}u^{n+p}\;\mathrm{d}x\right)\end{split} and (1.5) \begin{split}\int_{\Omega}\varphi^{3p}u^{n-\frac{p}{2}}|\Delta u|^{\frac{3p}{2}}\;\mathrm{d}x\leq c_{3}\left(\int_{\Omega}\varphi^{3p}u^{n}|\nabla\Delta u|^{p}\;\mathrm{d}x+\int_{\Omega}|\nabla\varphi|^{3p}u^{n+p}\;\mathrm{d}x\right)\end{split} hold true for all non-negative test functions \varphi\in W^{1}_{\infty}(\Omega) satisfying \int_{\Omega}|\nabla\varphi|^{3p}u^{n+p}\;\mathrm{d}x<\infty . Here, \mathrm{II} denotes the (positive semi-definite) second fundamental form of \partial\Omega . Remark 1.2. (i) For p=2 and \varphi\equiv 1 , the inequalities in ( 1.3 ) and ( 1.4 ) recover exactly the results in [ Grü01 , Theorem 1.1] . The localized version for p=2 with a non-negative localization function \varphi is mentioned in [ Grü05 ] , where a proof is sketched for the range 2-\sqrt{\tfrac{8}{8+d}}<n<3 . (ii) For d=1 , estimates ( 1.3 ) and ( 1.5 ) are derived in [ AG04 , Proposition 4.1] for p\geq\frac{4}{3} and \frac{p-1}{2}<n<2p-1 . The restriction p\geq 2 in our case (instead of p\geq\frac{4}{3} , as in d=1 ) is due to the term |\nabla u|^{3p-6} in the integral \int_{\Omega}\varphi^{3p}u^{n-2p+2}|\nabla u|^{3p-6}|D^{2}u\nabla u|^{2}\;\mathrm{d}x. In the case d=1 , this integral simplifies to \int_{\Omega}\varphi^{3p}u^{n-2p+2}|\nabla u|^{3p-4}|D^{2}u|^{2}\;\mathrm{d}x. Note, however, that we can estimate |D^{2}u\nabla u|^{2}\leq|D^{2}u|^{2}|\nabla u|^{2} , where |D^{2}u|=\sqrt{\sum_{i,j=1}^{d}(\partial_{ij}u)^{2}} denotes the Frobenius norm of the Hessian matrix D^{2}u=(\partial_{ij}u)_{i,j=1,\dots,d} . (iii) The restriction p<\frac{19}{3} is used only once, in ( 2.8 ) of the proof of Proposition 2.3 , to ensure the non-negativity of the integral \mathcal{I}_{4} . This integral does not appear for d=1 , as in this case D^{2}u=\Delta u . (iv) For smooth functions, Theorem 1.1 is valid for any dimension

Read the complete text