Multiharmonic analysis for nonlinear acoustics with different scales Anastasia Thöns-Zuevaa,b, Kersten Schmidta,b,c, Adrien Semina,b a: Research center Matheon, 10623 Berlin, Germany b: Institut für Mathematik, Technische Universität Berlin, Straße des 17. Juni 136, 10623 Berlin, Germany c: Brandenburgische Technische Universität Cottbus-Senftenberg, Institut für Mathematik, Platz der deutschen Einheit 1, 03046 Cottbus, Germany Corresponding author: AnastasiaThöns-Zueva,InstitutfürMathematik,TechnischeUniversitätBerlin,Berlin,Ger- many Address: Technische Universität Berlin, Sekretariat MA 6-4, Straße des 17. Juni 136, D-10623 Berlin E-mail: [email protected] Tel: +49 (0)30 314 - 25192 7 1 0 Abstract 2 The acoustic wave-propagation without mean flow and heat flux can be described in terms of velocity and n pressure by the compressible nonlinear Navier-Stokes equations, where boundary layers appear at walls due to a the viscosity and a frequency interaction appears, i.e. sound at higher harmonics of the excited frequency ω is J generated due to nonlinear advection. We use the multiharmonic analysis to derive asymptotic expansions for 9 smallsoundamplitudesandsmallviscositiesbothoforderε2 inwhichvelocityandpressurefieldsareseparated into far field and correcting near field close to walls and into contributions to the multiples of ω. Based on the ] P asymptotic expansion we present approximate models for either the pressure or the velocity for order 0, 1 and A 2, in which impedance boundary conditions include the effect of viscous boundary layers and contributions at . frequencies 0 and 2·ω depend nonlinearly on the approximation at frequency ω. In difference to the Navier- h t Stokes equations in time domain, which has to be resolved numerically with meshes adaptively refined towards a thewallboundariesandexplicitschemesrequiretheuseofverysmalltimesteps,theapproximativemodelscan m be solved in frequency domain on macroscopic meshes. We studied the accuracy of the approximated models [ of different orders in numerical experiments comparing with reference solutions in time-domain. 1 Keywords v Acoustic wave propagation, Singularly perturbed PDE, Impedance Boundary Conditions, Asymptotic Expan- 7 sions. 9 AMS subject classification 0 35C20, 41A60, 42A16, 35Q30, 76D05 2 0 . 1 1 Introduction 0 7 InthisarticlewecontinueinvestigatingtheacousticequationsintheframeworkofLandauandLifschitz[13]as 1 : aperturbationoftheNavier-Stokesequationsaroundastagnantuniformfluidwhereheatfluxisnottakeninto v account. The aim of this study is to take into account nonlinear advection behaviour as well as viscous effects i X in the boundary layer near rigid walls. The governing equations in time domain similar to the works of Tam r et al. [24, 25], but for the case of isothermal process, i.e. pressure over density is constant over space, may be a written as ∂ v+(v·∇)v+∇p−ν∆v=f, in Ω, (1.1a) t ∂ p+c2 divv+div(pv)=0, in Ω, (1.1b) t v=0, on ∂Ω. (1.1c) where v is the acoustic velocity, p = p(cid:48)/ρ with p(cid:48) being the acoustic pressure and ρ > 0 being mean density, 0 0 c is the speed of sound, and ν > 0 is the kinematic viscosity. The introduction of p instead of p(cid:48) is only to simplify the equations by removing the constant ρ and we will regard p as pressure and approximations to 0 p as pressure approximations. In the momentum equation (1.1a) with some known source term f the viscous dissipationinthemomentumaswellastheadvectionnonlineartermarenotneglectedasweconsidernearwall regionswherethederivativesoftheacousticvelocityarerapidlyincreasingandmightbecrucial. Thecontinuity 1 equation (1.1b)relatestheacousticpressuretothedivergenceoftheacousticvelocity. Thesystemiscompleted by no-slip boundary conditions. For gases the viscosity ν is very small and leads to viscosity boundary layers close to walls. These boundary layers are difficult to resolve in direct numerical simulations. Nevertheless, they have an essential influence on the absorption properties. Mainly based on experiments the physical community has introduced slip boundary conditions for the tangential component of the velocity, also known as wall laws, see for example [11, 16, 17]. For gases with small viscosity the Helmholtz equation can be completed by viscosity dependent boundary conditions [1] to obtain an approximation of high accuracy for which the boundary layers do not have to be resolved by finite element meshes [12]. In the earlier works we studied the linear acoustic equations taking into account viscous effects in the boundary layer near rigid walls. In [21] we derived a complete asymptotic expansion for the problem based on √ the technique of multiscale expansion in powers of η, where η is the dynamic viscosity and ν = η/ρ. This asymptotic expansion was rigorously justified with optimal error estimates. In [20] we proposed and justified (effective)impedanceboundaryconditionsforthevelocityaswellasthepressureforpossiblycurvedboundaries. In case of stable periodic oscillations in nonlinear dynamical systems the harmonic balance principal is used[15,23,28]. Itisdescribedasalinearcombinationofawaveoftheexcitationfrequencyanditsharmonics. Thereferredmethodwaspresentedasmultiharmonicanalysisforthemodellingofnonlinearmagneticmaterials [3, 4]. Its stability has been demonstrated within the eddy current model. For nonlinear Hamiltonian systems withasimpleoscillatoraspecialcaseofmultiharmonicanalysis,theso-calledmodulatedFourierexpansion[10], has been well developed. The multiharmonic analysis as a method in the frequency domain is especially attractive for nonlinear acoustics. This application has not previously been investigated, either numerically or with asymptotic expansions. In the current work we restrict ourselves to the case of small sound amplitudes that are of order O(ν). The article is ordered as follows. In Sec. 2 we introduce a frequency domain system for the quasi-stationary solution of the nonlinear acoustic wave propagation problem using the multiharmonic analysis. Moreover, the main ideas of the multiscale expansions separating far field and boundary layer contributions are introduced, that lead to the effective systems with impedance boundary conditions, both for the velocity and the pressure, thatarefinallyintroducedasmainresultsofthepaper. Thefarfieldandboundarylayertermsofthemultiscale expansion and the effective systems are derived in Sec. 3. Finally, in Sec. 4 we verify the effective systems by numerical computations using high-oder finite elements. 2 Multiharmonic analysis, multiscale expansion and approximative models 2.1 Multiharmonic analysis for the nonlinear system In many acoustic applications the source is of one single frequency ω >0 and so of the form 1(cid:16) (cid:17) f(t,x)= f(x) exp(−iωt)+f(x)exp(iωt) , 2 with f : Ω → C2 being complex valued. Then, we assume that the solution (v,p) of (1.1) tends to a quasi- stationary solution that is periodic in time with a period T = 2π which we denote by (v,p) again, i.e. ω v(t+ 2π,x)=v(t,x), p(t+ 2π,x)=p(t,x). (2.1) ω ω This does not mean in general that one obtains a mono-frequency solution of the same form as the source, but as the problem involves only linear and quadratic terms its solution can be written as combination of all the harmonics cos(kωt),k =0,1,... and sin(kωt),k =1,2,.... ∞ 1(cid:88) v(t,x)=v (x)+ v (x)exp(−ikωt)+v (x)exp(ikωt), 0 2 k k k=1 (2.2) ∞ 1(cid:88) p(t,x)=p (x)+ p (x)exp(−ikωt)+p (x)exp(ikωt) , 0 2 k k k=1 which is called multiharmonic ansatz [3, 4, 2]. Inserting expansion (2.2) into the time-dependent problem (1.1) and identifying the terms corresponding to exp(−ikωt), k = 0,1,... leads to the infinite system of non-linear 2 equations in space and frequency domain (cid:18) (cid:19) f(x) L (V,P)(x)+N (V,P)(x)= δ in Ω, k =0,1,... , (2.3a) k k 0 k=1 V=0 on ∂Ω , (2.3b) (cid:0) (cid:1)(cid:62) (cid:0) (cid:1)(cid:62) withthevectorsV= v ,v ,... andP = p ,p ,... ,collectingthecoefficientsoftheFourieransatz(2.2), 0 1 0 1 the linear differential operators (cid:32) (cid:33) −ikωv −ν∆v +∇p k k k L (V,P)= , k −ikωp +c2divv k k and the nonlinear differential operators 1(cid:32)(v0·∇)v0(cid:33) 1 (cid:88)∞ (cid:32)(vm·∇)vm+(vm·∇)vm(cid:33) N (V,P)= + 0 2 div(v p ) 4 div(v p +v p ) 0 0 m=0 m m m m 1 (cid:88)k (cid:32)(vm·∇)vk−m(cid:33) 1 (cid:88)∞ (cid:32)(vm·∇)vm−k+(vm−k·∇)vm(cid:33) N (V,P)= + , k >0. k 2 div(v p ) 2 div(v p +v p ) m=0 m k−m m=k m m−k m−k m Theorem 2.1. The existence and uniqueness of a quasi-stationary solution (v,p) of (1.1) satisfying (2.1) is equivalent to the existence and uniqueness of a solution (v ,v ,...,p ,p ,...) of (2.3). 0 1 0 1 Proof. First, let (v,p) be the unique solution of (1.1). Defining the vector (v ,v ,...,p ,p ,...) by 0 1 0 1 ω (cid:90) 2π/ω ω (cid:90) 2π/ω v (x)= v(t,x)dt, v (x)= v(t,x)exp(ikωt)dt, k (cid:62)1, (2.4a) 0 2π k π 0 0 ω (cid:90) 2π/ω ω (cid:90) 2π/ω p (x)= p(t,x)dt, p (x)= p(t,x)exp(ikωt)dt, k (cid:62)1. (2.4b) 0 2π k π 0 0 the decomposition (2.2) holds and by construction (v ,v ,...,p ,p ,...) is solution of (2.3). Moreover, [2, 0 1 0 1 Lemma 3.3] implies that the vector is unique. Now, let (v ,v ,...,p ,p ,...) be the unique solution of (2.3). It is easy to see that (v,p) defined by (2.2) 0 1 0 1 satisfies (1.1). If (1.1) would admit another solution (w,q), then its coefficients w , q defined similarly as k k in (2.4) would fulfil (1.1) as well, which is a contradiction to the assumption of unicity. 2.2 Asymptotic ansatz for small sound amplitude and viscosity Toinvestigateacousticvelocityvandacousticpressurepforsmallacousticexcitationandviscosityweintroduce asmallparameterε∈R+ andreplacetheacousticsourcef byε2 (cid:80)∞ εjf , whereeachtermf isindependent j=0 j j of ε, and the viscosity ν by ε2ν with ν ∈R+. Moreover, we consider the leading order source term ε2f to be 0 0 0 curl -free, i.e. curl f = 0, having in mind that f = ∇p where the pressure p corresponds to a solution 2D 2D 0 0 0 0 of a linear and inviscid wave equation. In addition, we assume for simplicity the source to disappear on the boundary. Theimpedanceboundaryconditionswithadditionaltermsduetomoregeneralsourcefunctionswill be given in the Appendix A.3. For these small acoustic excitations the leading part of the solution satisfies a linear equation in the whole domainasconsideredin[21]andthenonlinearitywillcomeintoplayonahigherorder. Thesmallviscositieson theotherhandleadstoboundarylayerswhosethicknessbecomesproportionaltoε. Indicatingtheirdependency onεwewilllabeltheacousticvelocityvε andtheacousticpressurepε withasuperscriptε. Theyaredescribed by the system (cid:18)f (x)+εf (x)+ε2f (x)(cid:19) Lε(Vε,Pε)(x)+N (Vε,Pε)(x)=ε2 0 1 2 δ , (2.5) k k 0 k=1 with the vectors Vε = (cid:0)vε,vε,...(cid:1)(cid:62), Pε = (cid:0)pε,pε,...(cid:1)(cid:62) of velocity and pressure coefficients and the linear 0 1 0 1 differential operators (cid:32) (cid:33) −ikωvε −ε2ν ∆vε +∇pε Lε(Vε,Pε)= k 0 k k . k −ikωpε +c2divvε k k 3 Ω ∂Ω ∂Ω f f n(τ) Ω s ∂Ω τ (a) (b) Figure1: (a)Definitionofageneraldomainwithalocalcoordinatesystem(τ,s)closetothewall;(b)Definition of an annulus domain for numerical simulations. In the following we specify first the domain and its boundary before we introduce the ansatz for an asymptotic expansion with far field terms and near field correctors and their coupling conditions. The geometrical setting LetΩ⊂R2 beaboundeddomainwithsmoothboundary∂Ω. Theboundaryshall be described by a mapping x :τ ∈Γ→R2 from a one-dimensional reference domain Γ⊂R. We assume the ∂Ω boundary to be C∞ such that points in some neighbourhood Ω of ∂Ω can be uniquely written as Γ x(τ,s)=x (τ)−sn(x (τ)) (2.6) ∂Ω ∂Ω where n is the outer normalised normal vector and s the distance from the boundary. Without loss of generality we assume |x(cid:48) (τ)| = 1 for all τ ∈ Γ. The orthogonal unit vectors in these ∂Ω tangential and normal coordinate directions are e = −n⊥, where we use the notation u⊥ = (u ,−u )(cid:62) for a τ 2 1 turned vector clockwise by 90◦, and e =−n. This allows us to write the tangential derivative ∇u(x)·e of a s τ function u∈C1(Ω) with abuse of notation as (∂ u)(x):=∂ u(x (τ)). (2.7) τ τ ∂Ω Moreover, the curvature κ on the boundary ∂Ω is given by x(cid:48) (τ)x(cid:48)(cid:48) (τ)−x(cid:48) (τ)x(cid:48)(cid:48) (τ) κ(x (τ)):= ∂Ω,1 ∂Ω,2 ∂Ω,2 ∂Ω,2 . ∂Ω (x(cid:48) (τ)2+x(cid:48) (τ)2)3/2 ∂Ω,1 ∂Ω,2 Asymptotic ansatz. Within this article we consider the acoustic source of the same order as the boundary layer,i.e.f =fε =O(ε2). Inthelinearmodeltheresultingacousticvelocityandpressureareofthesameorder and for the considered nonlinear model the same is true. The solution Vε, Pε of (2.5) should be approximated by a two-scale asymptotic expansion in the framework of Vishik and Lyusternik [27] and for each coefficient we take the ansatz ∞ ∞ (cid:88) (cid:16) (cid:17) (cid:88) (cid:16) (cid:17) vε ∼ εj+2 vj(x)+vj (x) , pε ∼ εj+2 pj(x)+pj (x) , for ε→0 , (2.8) k k BL,ε,k k k BL,ε,k j=0 j=0 wherevj andpj arethefarfield velocityandpressureoforderj andvj andpj representtherespective k k BL,ε,k BL,ε,k nearfield velocityandpressure. TheyareseekedinscaledcoordinateS(s)= s ofthelocalnormalisedcoordinate ε 4 system (2.6) in the form vj (x)=Φj (τ,s) e (τ)+Φj (τ,s) e (τ) (2.9a) BL,ε,k k,τ ε τ k,s ε s pj (x)=Πj(τ,s) (2.9b) BL,ε,k k ε taking into account the fact that the boundary layer thickness scales linearly with ε. For the desired decay properties we require the near field terms Φj (τ,S), Φj (τ,S) and Πj(τ,S) as well as their higher derivatives k,τ k,s k to vanish with S →∞. The subscript · stands for “boundary layer” expressing the nature of the near field BL,ε terms that they are essentially defined in a small layer close the boundary. Indeed the equality (2.9) can be assumedtobetrueonlyinanO(1)neighbourhoodoftheboundaryinwhichthelocalcoordinatesystem(2.6)is defined. Outsidethisneighbourhoodtheexpressionontherighthandsidesof (2.9),thatdecayingexponentially in an O(ε) distance, is multiplied with a smooth cut-off function such that the product is exactly zero where the local coordinate system is not defined. In the linear case [21] the near field velocity turned out tobe divergence free such thatthere is no boundary layerforthepressure. Duetothecouplingofthevelocityandpressurebythenonlineartermsthispropertycan notbeassumedingeneral. However,wewillseeinouranalysisthatthenearfieldpressuretermsΠj vanish BL,ε,k at least up to order 2 and up to this order the resulting near field terms for the frequency ω of the excitation remain exactly the same as for the linear system. Coupling of far and near field by the no-slip boundary conditions By the homogeneous Dirichlet boundary condition the tangential trace (cid:0)vε +vBL,ε(cid:1)·e and normal trace (cid:0)vε +vBL,ε(cid:1)·n vanish for any k k k τ k k and separately in the orders in ε, cf (1.1c), therefore the traces of the far field have to fulfil the conditions Φj (τ,0)=−vj (τ):=−vj(x(τ,0))·e , (2.10a) k,τ k,τ k τ Φj (τ,0)=−vj (τ):= vj(x(τ,0))·n . (2.10b) k,s k,s k The far and near field terms will be derived order by order up order 2 in Section 3 as well as the effective systems with impedance boundary conditions that we will present already in the following subsection. 2.3 Effective systems with impedance boundary conditions In this section we present effective models of order N = 0, 1 and 2 for approximative far field solutions Vε,N = (vε,N,vε,N,...)(cid:62), Pε,N = (pε,N,pε,N,...)(cid:62), in which the nonlinear and viscous behaviour in the 0 1 0 1 layers close to the boundary are incorporated with impedance boundary conditions. The steps for deriving the systemsforapproximativevelocityandpressurewillfollowinSection3.4. Contrarytotheoriginalsystem(1.1) in time domain or its multiharmonic approximation (2.5), for which all modes of velocity and pressure couple, the approximative pressure and velocity coefficients decouple for all modes k > 0. Therefore, we introduce separately systems for pressure coefficients pε,N only, where associated velocity coefficients wε,N are defined k k afterwordsasafunctionofthepressure,andsystemsforvelocitycoefficientsvε,N only,whereassociatedpressure k coefficients qε,N follow directly. Only for the static mode k =0 we have coupled velocity and pressure systems. k In general, the directly defined pressure coefficients pε,N and the pressure coefficients qε,N computed from the k k velocity may differ as well as the two velocity approximations vε,N and wε,N. We also distinguish the two for k k thestaticmodek =0evensoherevelocityandpressurecoefficientsaredefinedinacoupledsystemastheright hand side of this system depends on the different approximations. Derived from the asymptotic expansion, both approximative far field solutions for velocity and pressure order N shall be close to the respective far field expansion of order N, i.e. for the frequency mode k we expect that (cid:18)vε,N(cid:19) (cid:18)wε,N(cid:19) (cid:88)N (cid:18)vj(cid:19) k , k = εj+2 k +O(εN+3) . (2.11) qε,N pε,N pj k k j=0 k Even equally important the asymptotic regime of small sound amplitudes leads to an iterative procedure to obtain the coefficients for different modes k (see Table 1). In general, the coefficients for k = 1 can be defined independently and the neighbouring modes for k = 0 and k = 2 follow. Moreover, up to order 2 there are no modes for k > 2 as indicating that the response at the higher harmonics 3ω,4ω,... are more than two orders in ε smaller than the excitation amplitude. In general, for approximation of order N we have only the modes k =0,1,...,(cid:100)N+1(cid:101). 2 5 order pressure velocity N =0 pε,0 vε,0 ⇐= vε,0 1 0 1 N =1 pε,1 vε,1 ⇐= vε,1 1 0 1 N =2 pε,2 ⇐= pε,2 =⇒ pε,2 vε,2 ⇐= vε,2 =⇒ vε,2 0 1 0 0 1 2 Table1: Pressureapproximationspε,N andvelocityapproximationsvε,N canbecomputedseparatelyfromeach k k other (except for static mode for k = 0) and sequentially in the mode index k. Then, velocity approximations wε,N are directly deduced from pε,N and pressure approximations qε,N from vε,N. The velocity and pressure k approximations for the excitation frequency (k = 1) coincides with the respective solutions for the linear case up to order 1 (see [20]). 2.3.1 Systems for the pressure Herewepresenttheapproximativemodelsforthefarfieldpressure. Thisisdifferenttotheoriginalequationsin which no boundary conditions for the pressure, but for both velocity components, are imposed for each order. Approximative velocities can be computed a-posteriori (see Sec. 2.3.2). Order O(ε2) Theapproximativemodelforthepressureinfrequencyoftheexcitation1·ω isgivenbyalinear system ω2 ∆pε,0+ pε,0 =divf, in Ω, (2.12a) 1 c2 1 ∇pε,0·n=0, on ∂Ω, (2.12b) 1 All the terms pε,0 for k (cid:54)=1 are zero, meaning that the limit acoustic pressure is exactly as in the linear case. k Order O(ε3) The approximative model in frequency 1·ω is given by the linear system ω2 ∆pε,1+ pε,1 =divf, in Ω, (2.13a) 1 c2 1 (cid:114) ν ∇pε,1·n+(1+i) ∂2pε,1 =0, on ∂Ω. (2.13b) 1 2ω τ 1 Again, all the terms pε,1 for k (cid:54)= 1 are zero and the resulting acoustic pressure approximation is exactly as in k the linear case. The impedance boundary conditions (2.13b) are of Wentzell type. See [5, 18] for the functional framework and variational formulation. Order O(ε4) For frequency 1·ω the pressure of order 2 is solution of (cid:16) iων(cid:17) ω2 1− ∆pε,2+ pε,2 =divf, in Ω, (2.14a) c2 1 c2 1 (cid:114) ν iν ∇pε,2·n+(1+i) ∂2pε,2+ ∂ (κ∂ pε,2)=0, on ∂Ω. (2.14b) 1 2ω τ 1 2ω τ τ 1 Even for N =2 the nonlinear terms do not affect the pressure approximation in frequency of excitation, which coincides with the approximations in the linear case and are also obtained via a system decoupled from the velocity. However, in this order of approximation the first time other frequency modes come into play, namely thatforthefrequency0·ω,asocalledacousticstreaming[14],andforthefrequency2·ω. Theacousticpressure at frequency 0·ω is explicitly defined by the algebraic equation pε,2 =− 1 (cid:12)(cid:12)f −∇pε,2(cid:12)(cid:12)2 (2.15) 0 4ω2 1 6 and the one at frequency 2·ω by the Helmholtz equation ∆pε,2+ 4ω2pε,2 = 1 ∆(cid:0)f −∇pε,2(cid:1)2+ 1 (cid:18)(cid:0)f −∇pε,2(cid:1)·∇pε,2+ω2(cid:0)pε,2(cid:1)2(cid:19), in Ω, (2.16a) 2 c2 2 4ω2 1 c2 1 1 c2 1 ∇pε,2·n=0, on ∂Ω. (2.16b) 2 Forawell-poseddefinitionofpε,2 thesourcefunctionf hastobecontinouslydifferentiableandalsothepressure 2 approximation pε,2 needs higher regularity. Note, that the right hand side of (2.16a) can be simplified using 1 thefactthatpε,2 issolutiontoanHelmholtzproblem(seeAppendixA.4). Foranumericalapproximationwith 1 C0-continuous finite elements, for which this regularity is only attained approximately, the right hand side can be evaluated as the projection of the pressure gradient ∇pε,2 to continuous vector fields. 1 2.3.2 Post-processing of velocity from systems for the pressure When the far field pressure is computed we may obtain a-posteriori approximations to the far field velocity to the respective order. The far field velocities at frequency 1·ω are defined at the different approximation orders by i (cid:16) (cid:17) i (cid:16) (cid:17) i (cid:16) (cid:17) ν wε,0 = f −∇pε,0 , wε,1 = f −∇pε,1 , wε,2 = f −∇pε,2 − ∇pε,2. (2.17) 1 ω 1 1 ω 1 1 ω 1 c2 1 and those at frequency 2·ω at order 2 by (cid:18) (cid:19) wε,2 =− i ∇pε,2+ 1 ∇(cid:12)(cid:12)f −∇pε,2(cid:12)(cid:12)2 . (2.18) 2 2ω 2 4ω 1 For the frequency 0·ω a far field velocity approximation wε,1 of order 1 can be obtained as solution of 0 linear Stokes system similarly to (2.21) in the following subsection that is directly for a velocity approximation, however, using wε,1 on its right hand side. Likewise, a far field velocity approximation wε,2 for order 2 can be 1 0 defined by a nonlinear Navier-Stokes like system as (2.23) that depends on wε,2. 1 The far field velocity can be used as approximation away from the boundary and has to be corrected by a near field velocity approximation (see Sec. 2.3.5). 2.3.3 Systems for the velocity Here we propose approximative models directly for the far field velocity. For each order an approximative pressurecanbecomputedafterwards(seeSec.2.3.4)aswellasanearfieldvelocityapproximation(seeSec.2.3.5). Order O(ε2) The limit model is given by a linear system in frequency of excitation ω2 iω ∇divvε,0+ vε,0 = f, in Ω, (2.19a) 1 c2 1 c2 vε,0·n=0, on ∂Ω, (2.19b) 1 and all other terms vε,0, k (cid:54)=1 are zero. So the limit acoustic velocity coincides with the one in the linear case. k Order O(ε3) In frequency of excitation the approximative model is given by ω2 iω ∇divvε,1+ vε,1 = f, in Ω, (2.20a) 1 c2 1 c2 c2(cid:114) ν vε,1·n−(1+i) ∂2divvε,1 =0, on ∂Ω, (2.20b) 1 ω2 2ω τ 1 and there is a non-zero acoustic streaming velocity at frequency 0·ω that satisfies the Stokes system 1(cid:16) (cid:17) −ν∆vε,1+∇qε,3 =− (vε,1·∇)vε,1+(vε,1·∇)vε,1 , in Ω , (2.21a) 0 0 4 1 1 1 1 divvε,1 =0, in Ω , (2.21b) 0 vε,1 =0, on ∂Ω . (2.21c) 0 The purely real right hand sideof (2.21) implies that its solution (vε,1,qε,3) is purely real. Notethat qε,3 is not 0 0 0 only a Lagrange multiplier but a higher order approximation of the pressure at zero frequency. 7 Order O(ε4) The approximative model in frequency 1·ω is defined by (cid:18) iων(cid:19) ω2 iω 1− ∇divvε,2+ vε,2 = f, in Ω, (2.22a) c2 1 c2 1 c2 c2(cid:16) (cid:114) ν iν (cid:17) vε,2·n− (1+i) ∂2divvε,2+ ∂ (κ∂ divvε,2) =0, on ∂Ω, (2.22b) 1 ω2 2ω τ 1 2ω τ τ 1 and that of frequency 0·ω by the nonlinear system −ν∆vε,2+(vε,2·∇)vε,2+∇qε,4 =−1(cid:0)(vε,2·∇)vε,2+(vε,2·∇)vε,2(cid:17), in Ω (2.23a) 0 0 0 0 4 1 1 1 1 divvε,2 =− 1 (cid:0)vε,2·f +vε,2·f(cid:1), in Ω (2.23b) 0 4c2 1 1 vε,2 =0, on ∂Ω (2.23c) 0 Again,thesolution(vε,2,qε,4)of (2.23)ispurelyrealandqε,4 isapressureapproximationofhigherorder,where 0 0 0 qε,4 =−1(cid:12)(cid:12)vε,2(cid:12)(cid:12)2+O(ε5). At frequency 2·ω a velocity approximation satisfies the Helmholtz equation 0 4 1 ∇divvε,2+ 4ω2vε,2 =−iω∇(cid:0)vε,2(cid:1)2− 1 ∇(vε,2·f)+ i ∇(cid:0)divvε,2(cid:1)2, in Ω , (2.24a) 2 c2 2 c2 1 2c2 1 2ω 1 vε,2·n=0, on ∂Ω . (2.24b) 2 Remark 2.2. Note, that (vε,2 ·∇)vε,2 = (vε,1 ·∇)vε,1 +O(ε5). Therefor instead of solving the nonlinear 0 0 0 0 system (2.23a)–(2.23c)onecould,first,findvε,1 bysolvingthelinearsystem (2.21a)–(2.21c),andthensubstitute 0 it in (2.23a)–(2.23c) for vε,2 in the advection term. That will lead again to a linear system. 0 2.3.4 Post-processing of pressure from systems for the velocity Whenthefarfieldvelocityapproximationiscomputedwemayobtaina-posteriorianassociatedfarfieldpressure approximation for the frequencies 1·ω and 2·ω. The approximations for frequency 1·ω are given by ic2 qε,N =− divvε,N, N =0,1,2 , (2.25) 1 ω 1 and the approximation of order 2 for frequency 2·ω by qε,2 =−ic2 divvε,2− i vε,2·f− c2 (cid:18)(cid:0)divvε,2(cid:1)2− ω2(cid:0)vε,2(cid:1)2(cid:19) . (2.26) 2 2ω 2 2ω 1 2ω2 1 c2 1 Moreover, a pressure approximation of order 2 at frequency 0·ω is given by qε,2 =−1(cid:12)(cid:12)vε,0(cid:12)(cid:12)2 . (2.27) 0 4 1 2.3.5 Post-processing of a near field velocity Close to the wall the far field velocity approximations Vε,N have to be corrected by boundary layer functions in tangential as well as normal direction N √ VBL,ε,N(x)=χ(x)(cid:88)H(cid:96)(Vτε,N)(x) e−(1−i) 2ωνs(x), (2.28) (cid:96)=0 where χ is an admissible cut-off function (see [21]) that takes the constant value 1 in some subset of Ω , s is Γ the distance function to the boundary, i.e. there exists for each point x∈Ω a base point x ∈∂Ω such that Γ ∂Ω x = x +s(x)n⊥(x ), and the operators H(cid:96) : (C∞(∂Ω))∞ → (C∞(Ω ))∞ with H(cid:96) = (0,h(cid:96),h(cid:96),...)(cid:62) that ∂Ω ∂Ω Γ 1 2 are acting on the tangential velocity traces Vε,N = (vε,N,vε,N,...)(cid:62) with vε,N(x ) := vε,N(x )·n⊥(x ) τ 0,τ 1,τ k,τ ∂Ω k ∂Ω ∂Ω 8 where we note that vε,N =0 for k >(cid:100)N+1(cid:101). Hence, to define VBL,ε,N for N =0,1,2 we state the operators k,τ 2 h0(Vε,N)(x)=−vε,Nn⊥ , (2.29a) 1 τ 1,τ (cid:113) h1(Vε,N)(x)=−1κs(x)vε,Nn⊥+(1+i) ν ∂ vε,Nn , (2.29b) 1 τ 2 1,τ 2ω τ 1,τ (cid:16) (cid:113) (cid:17) (cid:113) h2(Vε,N)(x)=−3κ2s(x) (1+i) 2ν −s(x) vn⊥−(1+i) ν s(x)∂2vn⊥ 1 τ 8 ω 2ω τ (cid:113) (cid:16) (cid:17) + 1(1+i) ν 2κ∂ vε,N +∂ κvε,N s(x)n+ iν∂ κvε,Nn , (2.29c) 2 2ω τ 1,τ τ 1,τ 2ω τ 1,τ h2(Vε,N)(x)=−vε,Nn⊥ . (2.29d) 2 τ 2,τ wherewenotethatκ,n⊥,nandvε,N arefunctionsofthebasepointx ofxand∂ isthetangentialderivative 1,τ ∂Ω τ defined in (2.7). 3 Derivation of terms of multiscale expansion and effective systems In Sec. 2.2 we have introduced the ansatz of the two-scale expansion (2.8), which expresses an approximation to the exact solution as a two-scale decomposition into far field terms, modelling the macroscopic picture of the solution, which are corrected in the neighbourhood of the boundary by near field terms. To separate the two scales we use the technique of multiscale expansion as described in Sec. 2.2, which defines the near field terms in a local normalised coordinate system (2.6) such that they decay rapidly away from the wall and are set to zero where the local coordinate system is not defined (using a cut-off function). In the following we define the terms of asymptotic expansion (2.8) order by order. 3.1 Correcting near field In this section we will give the near field equations and their solutions up to order 2. They are derived such thatthenearfieldvelocityandpressureexpansions(2.9)insertedinto(2.5)leavearesidualassmallaspossible in powers of ε and that the sum of tangential far and near field velocity vanishes at the boundary. The general form of the near field equations of any order and in any frequency can be found in the Appendix A.2. The near field terms of order O(ε2). The near field equation for j =0 in frequency 1·ω yields iωu0 +ν ∂2u0 =0, 1,τ 0 S 1,τ ∂ u0 =0, S 1,s ∂ q0 =0. S 1 It is easy to see that its unique solution together with the coupling condition for far and near fields (2.10) and decay condition for the near field is given by (cid:112) u0 (τ,S)=−v0 (τ)e−λ0S, with λ =(1−i) ω/2ν , (3.1a) 1,τ 1,τ 0 0 u0 (τ,S)=0, (3.1b) 1,s q0(τ,S)=0. (3.1c) 1 This is the dominating boundary layer term close to the wall. The near field terms of order O(ε3). The near field equations for j =1 in frequency 1·ω are given by iωu1 +ν ∂2u1 =κ(cid:0)3iωS+3ν S∂2 +ν ∂ (cid:1)u0 , 1,τ 0 S 1,τ 0 S 0 S 1,τ ∂ u1 =−∂ u0 , S 1,s τ 1,τ ∂ q1 =0, S 1 9 which unique solution, using the terms in (3.1) together with the coupling condition, is u1 (τ,S)=−(cid:0)v1 (τ)+ 1κSv0 (τ)(cid:1) e−λ0S, (3.2a) 1,τ 1,τ 2 1,τ 1 u1 (τ,S)=− ∂ v0 (τ)e−λ0S, (3.2b) 1,s λ τ 1,τ 0 q1(τ,S)=0 . (3.2c) 1 The near field terms of order O(ε4). The near field equations for j =2 in frequency 1·ω are given by iωu2 +ν ∂2u2 =κ(cid:0)3iωS+3ν S∂2 +ν ∂ (cid:1)u1 1,τ 0 S 1,τ 0 S 0 S 1,τ −ν ∂2u0 −κ2(cid:0)3iωS2+3ν S2∂2 +ν (2S∂ −1)(cid:1)u0 , 0 τ 1,τ 0 S 0 S 1,τ ∂ u2 =−∂ u1 +κ(S∂ u1 +u1 ), S 1,s τ 1,τ S 1,s 1,s ∂ q2 =0, S 1 which unique solution, using the terms in (3.1) and (3.1) together with the coupling condition, is (cid:18) 3κ2S (cid:18) 1 (cid:19) S (cid:19) u2 (τ,S)=− v2 (τ)+ 1κSv1 (τ)− −S v0 (τ)+ ∂2v0 (τ) e−λ0S, (3.3) 1,τ 1,τ 2 1,τ 8 λ 1,τ 2λ τ 1,τ 0 0 1 (cid:18) κ(cid:16) 1 (cid:17) κ(cid:48)(cid:16) 1 (cid:17) (cid:19) u2 (τ,S)=− ∂ v1 (τ)+ 3S+ ∂ v0 (τ)+ S+ v0 (τ) e−λ0S, (3.4) 1,s λ τ 1,τ 2 λ τ 1,τ 2 λ 1,τ 0 0 0 q2 =0. (3.5) 1 In frequency 2·ω the first non trivial terms appear for j =2 with the near field equations given by 2iωu2 +ν ∂2u2 =0, 2,τ 0 S 2,τ ∂ u2 =0, S 2,s ∂ q2 =0. S 2 Its unique solution together with the coupling condition is given by √ u2 (τ,S)=−v2 (τ)e− 2λ0S, (3.6a) 2,τ 2,τ u2 (τ,S)=0, (3.6b) 2,s q2(τ,S)=0. (3.6c) 2 For frequency 0·ω the unique solution is the trivial solution at least up to order j = 2, i.e. the boundary layer disappears. 3.2 Far field velocity terms In the following section we will derive the terms of asymptotic expansion for the far field velocity up to order 2. The resulting expressions in frequency 1·ω are exactly the expressions for the linear case which are derived and analysed in [21]. The expressions for frequencies 0·ω and 2·ω are only due to the nonlinear advection term and do not appear for the linear case. The general form of the far field equations of any order and in any frequency can be found in the Appendix A.1. Approximation of order O(ε2). The limit model for the far field velocity in frequency 1·ω is given by ω2 iω ∇divv0+ v0 = f , in Ω, (3.7a) 1 c2 1 c2 0 v0·n=0, on ∂Ω, (3.7b) 1 Thefarfieldapproximationforfrequency0·ω isgivenbythestationaryincompressibleNavier-Stokesequations 1 ∇p2+(v0·∇)v0−ν ∆v0 =− ∇|v0|2 in Ω (3.8a) 0 0 0 0 0 4 1 divv0 =0 in Ω (3.8b) 0 v0 =0, on ∂Ω, (3.8c) 0 10