• Aucun résultat trouvé

On the scattering of electromagnetic waves by small bodies

N/A
N/A
Protected

Academic year: 2021

Partager "On the scattering of electromagnetic waves by small bodies"

Copied!
52
0
0

Texte intégral

(1)

HAL Id: hal-01677560

https://hal.inria.fr/hal-01677560

Submitted on 8 Jan 2018

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

On the scattering of electromagnetic waves by small bodies

Justine Labat, Victor Péron, Sébastien Tordeux

To cite this version:

Justine Labat, Victor Péron, Sébastien Tordeux. On the scattering of electromagnetic waves by small bodies. 2017. �hal-01677560�

(2)

On the scattering of electromagnetic waves by small bodies

Justine LABAT

PhD student in Applied Mathematics at the University of Pau

EPC Magique 3D, INRIA Bordeaux Sud-Ouest, LMAP UMR CNRS 5142, Universit ´e de Pau et des Pays de l’Adour Advisors : Victor P ´ERONet S ´ebastien TORDEUX

Instituto de Matematicas, Pontificia Universidad Cat ´olica de Valpara´ıso, Chile May 5, 2017

(3)

Scattering problems of electromagnetic waves by small inclusionsin 3D

incident field

δ δR3

δ: small parameter O(1)δ , . . . ,O(N)δ : inclusions δ=R3\SO(j)δ Γ(j)δ =O(j)δ

Γδ=S Γ(j)δ

Reference shapeOb(j)

1

O(j)δ =cj+δbO(j)

Ob(j)Lipschitz-continuous Asymptotic assumption

δλinc

(4)

Motivation−Applications and numerical difficulties

I Applications

I Medical imaging: detecting small tumours

I Non-destructive testing civil engineering: detecting impurities into concrete I Numerical approximation difficulties

I Infinite domain

I Mesh refinement near bodies (pictures from G. Vial)

I Our approach

I Reduced models derived thanks to asymptotic analysis: the radiusδis the small parameter

I Integration in a Boundary Element library (unbounded domain)

Figure:

https://people.eecs.ku.edu/∼

shontz/nsf career project.html

(5)

Non-exhaustive references−Scattering wave problems by small bodies

I Electromagnetic waves

H. Ammari, M. S. Vogelius, D. Volkov (2001)

Asymptotic formulas for perturbations in the electromagnetic fields due to the presence of inhomogeneities of small diameter II. The full Maxwell equations

H. Ammari, H. Kang (2004)

Reconstruction of small inhomogeneities from boundary measurements

I Linear elasticity problems

V. Bonnaillie-No ¨el, D. Brancherie, M. Dambrine, F. H ´erau, S. Tordeux, G. Vial (2011) Multiscale expansion and numerical approximation for surface defects

I Acoustic waves

V. Mattesi (2014)

Propagation des ondes dans un domaine comportant des petites h ´et ´erog ´en ´eit ´es : mod ´elisation asymptotique et calcul num ´erique

A. Bendali, P-H. Cocquet, S. Tordeux (2014)

Approximation by multipoles of the multiple acoustic scattering by small obstacles and application to the Foldy theory of isotropic scattering

(6)

Scattering of electromagnetic waves byonesmall perfect condutor in3D

(Einc,Hinc) δ

n Oδ

Oδ=δOb: perfect conductor δ=R3\Oδ: exterior domain Γδ=∂Ωδ

κ: wave number

Einc,Hinc: incident fields (given) Escaδ ,Hscaδ : scattered fields

Time-harmonic Maxwell equations

curl Escaδ Hscaδ =0 inδ

curl Hscaδ +Escaδ =0 inδ

n×Escaδ =−n×Einc onΓδ

n·Hscaδ =−n·Hinc onΓδ

|x|→+∞lim |x|(Hscaδ ×erEscaδ ) =0

κ2=ω2µ

ε+iσ ω

Im(κ)0

ω >0: frequency

µ >0: magnetic permeability ε >0: electric permittivity σ0: electric conductivity

(7)

Scattering of electromagnetic waves byonesmall perfect condutor in3D

(Einc,Hinc) δ

n Oδ

Oδ=δOb: perfect conductor δ=R3\Oδ: exterior domain Γδ=∂Ωδ

κ: wave number

Einc,Hinc: incident fields (given) Escaδ ,Hscaδ : scattered fields

Time-harmonic Maxwell equations

curl Escaδ Hscaδ =0 inδ

curl Hscaδ +Escaδ =0 inδ

n×Escaδ =−n×Einc onΓδ

n·Hscaδ =−n·Hinc onΓδ

|x|→+∞lim |x|(Hscaδ ×erEscaδ ) =0

(curl EincHinc=0 inR3 curl Hinc+Einc=F inR3

I Fsmoothsource term I 0/suppF

(8)

Guideline

Approximation method: Matched Asymptotic Expansions

“Formal” part Analysis part

I Postulate asymptotic expansions I Derive problems independent ofδ I Define a global approximationENδ

I Modal representation of solutions I Validation of reduced models

ENδ Escaδ

X

=? O

δ→0N)

Numerical resolution method: Equivalent Source Problem

I Boundary obstacles considered as surface/ponctual source terms

I Implementation of Matched Asymptotic Expansions method

I Comparison with volumical methods

(9)

Outline

1 Method of matched asymptotic expansions Asymptotic domain decomposition Far field expansion

Near field expansion Matching conditions Global approximation

2 Modal decomposition of the asymptotics Modal decomposition of the far field Modal decomposition of the near field Application of the matching conditions Case of the sphere: analytical solutions

3 Conclusion and perspectives

(10)

Outline

1 Method of matched asymptotic expansions Asymptotic domain decomposition Far field expansion

Near field expansion Matching conditions Global approximation

2 Modal decomposition of the asymptotics Modal decomposition of the far field Modal decomposition of the near field Application of the matching conditions Case of the sphere: analytical solutions

3 Conclusion and perspectives

(11)

Asymptotic domain decomposition

Far field domain Near field domain

ηδF

ηδN

δ→0limηδF=0

δ→0lim ηδN

δ = +∞

(12)

Asymptotic domain decomposition

Far field domain Matching area

ηδF

ηNδ ηFδ

δ→0limηδF=0

δ→0lim ηδN

δ = +∞

Asymptotic assumption δηFδ< ηδNλinc

A.M. Il’in (1991)

Matching of Asymptotic Expansions of Solutions of Boundary Value Problems

(13)

Asymptotic domain: far field

Far field domain

ηFδ

δ−→0

Asymptoticfar field domain

R3\B(0, ηδF) ?=R3\ {0}

(14)

Asymptotic domain: near field

Near field domain Change of coordinates

Oδ

ηNδ

Oδ=δOb bx=x

δ

δ0

Ob

ηNδ δ

Ob

Asymptoticnear field domain Ω =b R3\bO

(15)

Far field expansion inΩ? =R3\ {0}

1. Postulateformal series

Escaδ (x)

δ→0

X

p=0

δpEep(x) Hscaδ (x)

δ→0

X

p=0

δpHep(x)

I Eep,Hepdefined in? I Eep,Hepindependent ofδ

I Eep,Heppotentially singular atx=0

2. Deriveformal problems: Forp0, findEep,HepHloc(curl,?)s.t.

(curlEepHep=0 in? curlHep+eEp=0 in? + Silver-M ¨uller radiation condition

|x|→+∞lim |x|

Hep×erEep

=0

I Ill-posed problems

I Missing information matching conditions I Singular behavior atx=0

(16)

Far field expansion inΩ? =R3\ {0}

1. Postulateformal series

Escaδ (x)

δ→0

X

p=0

δpEep(x) Hscaδ (x)

δ→0

X

p=0

δpHep(x)

I Eep,Hepdefined in? I Eep,Hepindependent ofδ

I Eep,Heppotentially singular atx=0

2. Deriveformal problems: Forp0, findEep,HepHloc(curl,?)s.t.

(curlEepHep=0 in? curlHep+eEp=0 in? + Silver-M ¨uller radiation condition

|x|→+∞lim |x|

Hep×erEep

=0

I Ill-posed problems

I Missing information matching conditions I Singular behavior atx=0

(17)

Far field expansion inΩ? =R3\ {0}

1. Postulateformal series

Escaδ (x)

δ→0

X

p=0

δpEep(x) Hscaδ (x)

δ→0

X

p=0

δpHep(x)

I Eep,Hepdefined in? I Eep,Hepindependent ofδ

I Eep,Heppotentially singular atx=0

2. Deriveformal problems: Forp0, findEep,HepHloc(curl,?)s.t.

(curlEepHep=0 in? curlHep+eEp=0 in? + Silver-M ¨uller radiation condition

|x|→+∞lim |x|

Hep×erEep

=0

I Ill-posed problems

I Missing information matching conditions I Singular behavior atx=0

(18)

Near field expansion inΩ =b R3\Ob 1. Postulateformal seriesin fast variablebx=xδ

Escaδ (x)

δ→0

X

p=0

δpEbp(bx) Hscaδ (x)

δ→0

X

p=0

δpHbp(bx)

I Ebp,Hbpdefined inb I Ebp,Hbpindependent ofδ

I Ebp,Hbppotentially increasing at

Einc(x) =

X

p=0

δp X

|α|=p

1

α!dαEinc(0)·bxα

| {z }

bEincp(bx)

2. Deriveformal problems: Forp0findEbp,HbpHloc(curl,Ω)b s.t.

curlbEpHbp−1=0 inb curlHbp+Ebp−1=0 inb divEbp=divHbp=0 inb (bE−1=Hb−1=0)+ Boundary condition onbΓ =b

n×Ebp=−n×bEincp onΓb n·Hbp=−n·Hbincp onΓb

I Ill-posed problems

I Missing information matching conditions I Increasing behavior at

(19)

Near field expansion inΩ =b R3\Ob 1. Postulateformal seriesin fast variablebx=xδ

Escaδ (x)

δ→0

X

p=0

δpEbp(bx) Hscaδ (x)

δ→0

X

p=0

δpHbp(bx)

I Ebp,Hbpdefined inb I Ebp,Hbpindependent ofδ

I Ebp,Hbppotentially increasing at

Einc(x)=

X

p=0

δp X

|α|=p

1

α!dαEinc(0)·bxα

| {z }

bEincp(bx)

2. Deriveformal problems: Forp0findEbp,HbpHloc(curl,Ω)b s.t.

curlbEpHbp−1=0 inb curlHbp+Ebp−1=0 inb divEbp=divHbp=0 inb (bE−1=Hb−1=0)+ Boundary condition onbΓ =b

n×Ebp=−n×bEincp onΓb n·Hbp=−n·Hbincp onΓb

I Ill-posed problems

I Missing information matching conditions I Increasing behavior at

(20)

Near field expansion inΩ =b R3\Ob 1. Postulateformal seriesin fast variablebx=xδ

Escaδ (x)

δ→0

X

p=0

δpEbp(bx) Hscaδ (x)

δ→0

X

p=0

δpHbp(bx)

I Ebp,Hbpdefined inb I Ebp,Hbpindependent ofδ

I Ebp,Hbppotentially increasing at

Einc(x) =

X

p=0

δp X

|α|=p

1

α!dαEinc(0)·bxα

| {z }

bEincp(bx)

2. Deriveformal problems: Forp0findEbp,HbpHloc(curl,Ω)b s.t.

curlbEpHbp−1=0 inb curlHbp+Ebp−1=0 inb divEbp=divHbp=0 inb (bE−1=Hb−1=0)+ Boundary condition onbΓ =b

n×Ebp=−n×bEincp onΓb n·Hbp=−n·Hbincp onΓb

I Ill-posed problems

I Missing information matching conditions I Increasing behavior at

(21)

Some notations

Far field decomposition

Eep(x) =Eeregp (x) +Eesingp (x) x? Hep(x) =Heregp (x) +Hesingp (x) x? I Eeregp ,Heregp : regular behavior atx=0(regular partof the far field) I Eesingp ,Hesingp : singular behavior atx=0(singular partof the far field) Near field decomposition

bEp(bx) =Ebregp (bx) +bEsingp (bx) bxb Hbp(bx) =Hbregp (bx) +Hbsingp (bx) bxb

I Ebregp ,Hbregp :“decreasing” behavior at(decreasing partof the near field) I Ebsingp ,Hbsingp : increasing behavior at(increasing partof the near field) Be careful. Weighted spaces X. Claeys, PhD thesis (2008)

Analyse asymptotique et num ´erique de la diffraction d’ondes

(22)

Matching conditions

Find missing information

I Singular part of the far field ←−Decreasing part of the near field I Increasing part of the near field←−Regular part of the far field

Idea: developeEpclose tox=0andEbpnear

eEp(x)

r→0

−1

X

`=−∞

r`eE(p)` (θ, ϕ) +

+∞

X

`=0

r`eE(p)` (θ, ϕ) x= (r, θ, ϕ)

bEp(bx)

R→∞

−1

X

`=−∞

R`Eb(p)` (θ, ϕ) +

+∞

X

`=0

R`bE(p)` (θ, ϕ) bx= (R, θ, ϕ) =r δ, θ, ϕ

(23)

Matching conditions

Find missing information

I Singular part of the far field ←−Decreasing part of the near field I Increasing part of the near field←−Regular part of the far field

Idea

I Far field expansion inV(0)

Escaδ (x)

δ→0 +∞

X

p=−∞

δp

−1

X

`=−∞

r`eE(p)` (θ, ϕ) +

+∞

X

`=0

r`Ee(p)` (θ, ϕ)

!

I Near field expansion inV(∞)withR=δr

Escaδ (x)

δ→0 +∞

X

p=−∞

δp

−1

X

`=−∞

R`bE(p)` (θ, ϕ) +

+∞

X

`=0

R`Eb(p)` (θ, ϕ)

!

(24)

Matching conditions

Find missing information

I Singular part of the far field ←−Decreasing part of the near field I Increasing part of the near field←−Regular part of the far field

Idea

I Far field expansion inV(0)

Escaδ (x)

δ→0 +∞

X

p=−∞

δp

+∞

X

`=−∞

r`Ee(p)` (θ, ϕ)

!

I Near field expansion inV(∞)withR=δr

Escaδ (x)

δ→0 +∞

X

p=−∞

δp

+∞

X

`=−∞

R`Eb(p)` (θ, ϕ)

!

I Identificationof the two series

eE(p)` =Eb(p+`)` p0, `=−p, . . . ,−1 bE(p)` =Ee(p−`)` p0, `=0, . . . ,p

eE(p)` =0 ` <−p bE(p)` =0 ` >p

(25)

Matching conditions

I Identificationof the two series

eE(p)` =Eb(p+`)` p0, `=−p, . . . ,−1 bE(p)` =Ee(p−`)` p0, `=0, . . . ,p

eE(p)` =0 ` <−p bE(p)` =0 ` >p

I Term by term Eep(x)

r→0

−1

X

`=−p

r`bE(p+`)` (θ, ϕ)+

+∞

X

`=0

r`eE(p)` (θ, ϕ) (far field)

Ebp(bx)

R→∞

−1

X

`=−∞

R`bE(p)` (θ, ϕ) +

p

X

`=0

R`Ee(p−`)` (θ, ϕ) (near field)

I Forp=0,eE0isonlyregular

(26)

Global approximation Truncated series

EeNδ(x) =

N

X

p=0

δpeEp(x) HePδ(x) =

N

X

p=0

δpHep(x) (far fields)

EbNδ(x) =

N

X

p=0

δpEbp(bx) HbNδ(x) =

N

X

p=0

δpHbp(bx) (near fields)

Global approximation

EPδ(x) =χδ(|x|)bEPδ(x) + 1χδ(|x|)

EePδ(x) xδ

HPδ(x) =χδ(|x|)HbPδ(x) + 1χδ(|x|)

HePδ(x) xδ

1

ηδF ηδN r

χδ(r)

Consistance and stabilityDetermineφs.t.

φ(N) −→

N→∞+∞and for anyN>0,ρ > ρ0

Escaδ ENδ

H(curl,ΩδB(0,ρ))CEN,ρδφ(N)

Hscaδ HNδ

H(curl,ΩδB(0,ρ))CHN,ρδφ(N)

(27)

Global approximation Truncated series

EeNδ(x) =

N

X

p=0

δpeEp(x) HePδ(x) =

N

X

p=0

δpHep(x) (far fields)

EbNδ(x) =

N

X

p=0

δpEbp(bx) HbNδ(x) =

N

X

p=0

δpHbp(bx) (near fields)

Global approximation

EPδ(x) =χδ(|x|)bEPδ(x) + 1χδ(|x|)

EePδ(x) xδ

HPδ(x) =χδ(|x|)HbPδ(x) + 1χδ(|x|)

HePδ(x) xδ

1

ηF ηN r

χδ(r)

Consistance and stabilityDetermineφs.t.

φ(N) −→

N→∞+∞and for anyN>0,ρ > ρ0

Escaδ ENδ

H(curl,ΩδB(0,ρ))CEN,ρδφ(N)

Hscaδ HNδ

CHN,ρδφ(N)

(28)

Outline

1 Method of matched asymptotic expansions Asymptotic domain decomposition Far field expansion

Near field expansion Matching conditions Global approximation

2 Modal decomposition of the asymptotics Modal decomposition of the far field Modal decomposition of the near field Application of the matching conditions Case of the sphere: analytical solutions

3 Conclusion and perspectives

(29)

Modal decomposition of the far field of order 0

FindeE0andHe0Hloc(curl,R3)s.t.

curlEe0He0=0 inR3 curlHe0+eE0=0 inR3

|x|→∞lim |x|

He0×ereE0

=0

I Ee0,He0areregular(no singularity atx=0)

I Ee0,He0satisfy S.M.=eE0=0andHe0=0

I Far field expansion

Escaδ (x)

X

p=1

δpEep(x)

(30)

Modal decomposition of the far field of orderp>0

FindeEpandHepHloc(curl,?)s.t.

curlEepHep=0 in? curlHep+Eep=0 in? lim

|x|→∞|x|

Hep×erEep

=0

I Eep,Heparesingularatx=0

(31)

Modal decomposition of the far field of orderp>0

FindeEpandHepHloc(curl,?)s.t.

curlEepHep=0 in? curlHep+Eep=0 in? lim

|x|→∞|x|

Hep×erEep

=0

I Eep,Heparesingularatx=0 I Modal decomposition (singular at 0)

Eep(x) =

X

n=1 n

X

m=−n

(

βe(p),n,mEcurl

h(1)n (κ|x|)Yn,m(bx)x

+βe(p),n,mH

curl curl h

h(1)n (κ|x|)Yn,m(bx)xi )

I Finite numberofβe(p),n,mE,βe(p),n,mHnon-equal to 0 I βen,m(p),E,βen,m(p),Hare given bymatching conditions I h(1): spherical Hankel function of first kind

Références

Documents relatifs

We study the optical response of two coupled oriented dipoles with the dimer axis perpendicular to the wave vector of light by analyzing how their scattering matrix can be

As noted before, the main value of these decompositions is theoretical: demonstrating that the bound consistency propagator of [12] for the NV ALUE constraint can be sim- ulated using

To conclude, in addition to the resonant contributions due to the excitation of QNM of the scatterer by the ex- citation field, a non-resonant contribution must be taken into account

Regarding more complex rotor-structure coupling, Silverthorn [5] investigated an advancing whirl flutter mode on the main rotor of a YAH-64 helicopter using an

The result of the interference between the wave scattered by the specimen and the reference wave results in an intensity distribution, or holographic diffraction pattern,

(a) Reference eld; (b) measurement support with ran- dom acquisitions; (c) reconstructed eld using GPOD1; (d) reconstructed eld using GPOD2; (e) reconstructed eld using the

Organisation of the paper is as it follows: governing equations are derived as a dynamic perturbation of the stationary configuration of an inextensible catenary in section (2);

The subjunctive verb form in such cases carries an element of deontic modality rather than that of epistemic modality, as it contains the speaker’s will regarding the