• 検索結果がありません。

WAVE FRONT SETS OF SOLUTIONS TO ELASTIC WAVE PROPAGATION PROBLEMS IN STRATIFIED MEDIA (Analytical Studies for Singularities to the Nonlinear Evolution Equation Appearing in Mathematical Physics)

N/A
N/A
Protected

Academic year: 2021

シェア "WAVE FRONT SETS OF SOLUTIONS TO ELASTIC WAVE PROPAGATION PROBLEMS IN STRATIFIED MEDIA (Analytical Studies for Singularities to the Nonlinear Evolution Equation Appearing in Mathematical Physics)"

Copied!
9
0
0

読み込み中.... (全文を見る)

全文

(1)

WAVE FRONT SETS OF

SOLUTIONS

TO ELASTIC

WAVE

PROPAGATION PROBLEMS

IN

STRATIFIED MEDIA

SENJO

SHIMIZU

(清水 扇丈)

Faculty of

Engineering, Shizuoka

University,

Hamamatsu 432-8561

1. Introduction

We considerelasticwavepropagationproblems in plane-stratified media$\mathrm{R}^{3}$with

the planer interface $x_{3}=0$. This problem is formulated as

an

elastic mixed or

initial-interfece problem in a stratified media.

An elastic equation has two speeds. Pressure or Primary wave (for simplicity

called $\mathrm{P}$ wave) and Share or Secondary wave

($\mathrm{S}$ wave). $\mathrm{P}$ wave is a longitudinal

wave and $\mathrm{S}$ wave is a transversal

wave. In general the speed of $\mathrm{P}$ wave is

grater

than that of $\mathrm{S}$

wave. In plane-stratified media problem, a lower half-space $\mathrm{R}_{-}^{3}$

called Medium $I$has $P_{1}$ and $S_{1}$ waves and an upper half-space $\mathrm{R}_{+}^{3}$ called Medium

II has $P_{2}$ and $S_{2}$ waves. The speed of $P_{1}$ (resp. $P_{2}$) wave is grater than that of$S_{1}$

(resp. $S_{2}$ ) wave. So the order relation of the speeds of

$P_{1},$ $P_{2},$ $S_{1}$, and $S_{2}$

waves

are six

cases.

Here

we

assume

$P_{2},$ $S_{2,1}P,$ $S_{1}$ waves in order of speedsince it is the

most complex

case.

We put unit impulse Dirac’s delta in the lower half-space Medium I. Then $P_{1}$

incident wave which speed is faster than $S_{1}$ incident wave bumps against the

in-terface andcauses $P_{1}$ and $S_{1}$ reflected waves in Medium I and $P_{2}$ and $S_{2}$ refracted

waves in Medium II as in Figure 1.

$X\mathrm{a}$

Figure 1 Reflected waves and refracted waves

Moreover when time goes on, lateral waves in other words glancing waves or

total

reflected

(or refracted)

waves

arise. In Figure 2, dotted

arrows

show $P_{2^{-}}P_{1}$

and $P_{2^{-}}S_{1}$ lateral waves in Medium I, and $P_{2^{-}}S_{2}$ lateral wave in

Medium

II for

1991 Mathematics Subject Classification. Primary $35\mathrm{L}67$; Secondary $73\mathrm{C}3535\mathrm{E}9935\mathrm{L}20$

.

This work was supported in part by $\mathrm{G}\mathrm{r}\mathrm{a}\mathrm{n}\mathrm{t}\mathrm{S}^{- \mathrm{i}}\mathrm{n}$ Aid for Encouragement

of Young Scientists

(grantA-09740098) from the Ministry of Education, Science and Culture of Japan.

(2)

$P_{1}$ incident

wave.

$P_{2^{-}}P_{1}$ lateral wave means that the wave originally should have

been $P_{2}$ reflected

wave

tends to total reflection, then becomes

source

and

causes

$P_{1}$

reflected wave.

Figure 2 Lateral waves

We have 11th kind of lateral waves in all. It is a characteristic of our elastic

wave propagation problems in stratified media. If half-space problem which has

two speeds, there exist only one kind of lateral wave. If plane-stratified media

problem that each medium has one speed, there exist onlyonekind oflateral wave.

Thus this elastic wave propagation problems in plane-stratified media has many

lateral waves.

In this paper we prove the above physical situation mathematically by using a

expression of an inner estimate of singularities. The main technical tool of our

analysis is alocalization method.

We

gave

an inner estimate of the location ofsingularities of the reflected and

re-fracted Riemann functions by making use of the localization method $[7,8]$. This

method is first studied by M. F. Atiyah, R. Bott, and L. $\mathrm{G}^{\mathrm{Q}}\mathrm{a}\mathrm{r}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{g}[1]$ and L.

H\"ormander [2] for initial value problem, then studied by M. Matsumura [5], M.

Tsuji [9], and S. Wakabayashi $[10,11]$ for half-space mixed problem.

In this paper, we give an outer estimate of wave front sets of the incident,

reflected and refracted Riemann functionsbymaking use ofthelocalization method.

$\mathrm{A}\mathrm{t}\mathrm{i}\mathrm{y}\mathrm{a}\mathrm{h}- \mathrm{B}\mathrm{o}\mathrm{t}\mathrm{t}- \mathrm{G}^{\mathrm{o}}\mathrm{a}\mathrm{r}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{g}[1]$ studied theouter estimate ofwavefront sets of solutions to

initial value problem. Wakabayashi studied for half-space mixed problem [12], and

for more general case [13]. We analysis an outer estimate of wave front sets ofthe

Riemann functions to the elastic mixed problem based on Wakabayashi’s theorem

[12, Theorem 4.2]. Combiningthe innerestimate and the outer estimate, weobtain

the exact wave front sets of the elastic mixed problem in stratified media.

I would like to express my gratitude to Professor Seiichiro Wakabayashi for his

invaluable suggestions.

2. Formulation ofProblems

We consider elastic wave propagation problems in the following plane-stratified

media $\mathrm{R}^{3}$ with the planar interface $x_{3}=0$:

$(\lambda(x_{3}), \mu(X_{3}),$$\rho(_{X))=}3\{$

$(\lambda_{1}, \mu_{1}, \rho_{\mathrm{l}})$ for $x_{3}<0$,

($\lambda_{2,\mu_{2},\rho_{2})}$ for $x_{3}>0$

.

Here the constants $\lambda_{1},$ $\lambda_{2},$

(3)

$\rho_{1},$ $p_{2}$ are densities.

We

shalldenote the lower half-space $\mathrm{R}_{-}^{3}$ by

Medium

$I$and the upper half-space $\mathrm{R}_{+}^{3}$ by Medium II, respectively, as in Figure 3.

Figure 3 Stratified media

We

assume

that

(1.1) $\lambda_{i}+\mu_{i}>0$, $\mu_{i}>0$, $\rho_{i}>0$, $i=1,2$.

(1.1) is the natural assumption in practicalsituation. From the ro$o\mathrm{t}\mathrm{s}$ofthe

charac-teristic equationsof$P^{I}(D)$ and$P^{II}(D)$ which aredefined below $3\cross 3$ matrix valued

hyperbolic partial differential operators in Medium I and Medium II, respectively,

weobtain two speeds correspondto Pressure or Primary wave (for simplicity called

$\mathrm{P}$ wave) and Share or Secondary wave ($\mathrm{S}$ wave) on

each medium. $\mathrm{P}$

wave

is a

lon-gitudinal wave and $\mathrm{S}$ wave is a transversal wave.

$c_{\mathrm{P}1}$ denotes the speed of

$\mathrm{P}$ wave

in Medium I and $c_{s_{1}}$ denotes the speed of

$\mathrm{S}$ wave

in Medium I. $c_{p_{2}}$ and $c_{s_{2}}$ denote

the speed of$\mathrm{P}$ and $\mathrm{S}$

wave in Medium II, respectively. They are given by $c_{p_{i}}^{2}= \frac{\lambda_{i}+2\mu_{i}}{p_{i}}$, $c_{s_{i}}^{2}=$

.

$\frac{\mu_{i}}{p_{i}}$, $i=1,2$

.

By assumption $(1.\perp)$, the speed of $\mathrm{P}$ wave is greater than that of $\mathrm{S}$ wave

in each

medium. On account of this, these are six cases of the order relation of the speeds

of$\{cc, cc\}p_{1}’ s1p_{2’ S2}$

.

Herewe assume that

$c_{s_{1}}<c_{\mathrm{P}1}\leq c_{s_{2}}<c_{p_{2}}$,

since ifwe put an unit impulse Dirac’s delta in Medium I, it is the

case

that the

most number of lateral waves are appeared. The other cases can be treated in a

similar

manner

(cf. [6, Section 3]).

Let $x=(x_{0,1,2,3}xxx)=(x’, x\mathrm{s})=(x_{0}, x^{\prime/})=(x_{0,3}x^{\prime//}, x)$ in $\mathrm{R}^{4}$. The variable

$x_{0}$ will play a role of time, and $x^{\prime/}=(x_{1,}.x_{2}, X_{3})$ will play that of space. $\xi$ is

a

real

dual variable of$x$ andis equal to $(\xi_{0}, \xi_{1}, \xi 2, \xi_{3})=(\xi’, \xi_{3})=(\xi_{0}, \xi/’)=(\xi_{0}, \xi^{\prime/}/, \xi_{3})$ in

$\mathrm{R}_{\xi}^{4}$. We use the differential symbol $D_{j}=i^{-1}\partial/\partial X_{j}(j=0,1,2,3)$, where $i=\sqrt{-1}$

.

We shall denote by $\mathrm{R}_{-}^{n}$ the half-space $\{x=(x_{1}, \cdots, x_{n})\in \mathrm{R}^{n}|x_{n}<0\}$ and by

$\mathrm{R}_{+}^{n}$ the half-space $\{x=(x_{1}, \cdots, x_{n})\in \mathrm{R}^{n}|x_{n}>0\}$, and also

use

the notation $|x|$

$=\sqrt{x_{1}^{2}+x_{n}^{2}}$

.

Let $u(x)={}^{t}(u_{1}(x), u2(x),$$u_{3}(x))\in \mathrm{R}^{3}$ be the displacement vector at time $x_{0}$

and position $x^{\prime/}$

.

The

propagation

(4)

is formulated as mixed (initial-interface value) problem:

Here

$P^{I}(D)u=-D^{2}Eu0+ \frac{\lambda_{1}+\mu_{1}}{p_{1}}\nabla_{x’’}(\nabla_{x^{;\prime}}\cdot u)+\frac{\mu_{1}}{\rho_{1}}\Delta_{x’’}u$,

is a $3\cross 3$ matrix valued secondorder hyperbolic differential operator with constant

coefficients where $E$ is a 3 $\mathrm{x}3$ identity matrix,

$(B^{I}(D)u)k=i\lambda_{1}(\nabla x’’\cdot u)\delta_{k3}+2\mu_{1}\epsilon k3(u)$, $k=1,2,3$ ,

are the k-th component of symmetric stress tensors of$B^{I}(D)u$ where

$\epsilon_{k3}(u)=i/2(D_{3}u_{k}+D_{k^{\prime u}}3)$ , $k=1,2,3$,

are strain tensors. The $P^{II}(D)u$ and $B^{II}(D)u$ are defined by replacing $\lambda_{1},$

$\mu_{1},$ $\rho_{1}$ by $\lambda_{2},$

$\mu_{2},$ $p_{2}$

,

respectively.

If

we

put unit impulse Dirac’s delta $\delta(x-y)$ with $x_{0}\geq y_{0}$ and $y_{3}<0$, that

is, put it in Medium I, then the Riemann function ofthis elastic mixed problem is

given by the following:

$G(x, y)=\{$

$E^{I}(x-y)-F^{I}(x, y)$ for $x_{3}<0$,

$F^{II}(x, y)$ for $x_{3}>0$.

Wecall $E^{I}(x-y),$ $FI(x, y)$, and $F^{II}(x, y)$ the incident, reflected, and refracted

Rie-mann functions, respectively, becausethesearecorresponding to incident, reflected, and refracted waves, respectively. $E^{I}(x)$ is the fundamental solution in Medium I

describing an incident wave defined by

$E^{I}(x)=(2 \pi)^{-4}\int_{\mathrm{R}_{\xi}^{4}}e^{ix}.P(\xi+i\eta)I(\xi+i\eta)-1d\xi$, $\eta\in-\gamma 0\theta-\Gamma(\det P^{I}, \theta)$,

where $\gamma_{0}$ is a positive real number,

$\theta$ and $\Gamma(\det P^{I}, \theta)$ are defined Definition 1.4

below, and $P^{I}(\xi+i\eta)^{-1}$ is a $3\cross 3$ inverse matrix. Taking partial Fourier-Laplace

transform with respect to $x’$ for the mixed problem, we obtain a interface value

problem for ordinary differential equation with parameters. Then taking partial

inverse Fourier-Laplace transform for the solution, we obtain explicit expressions

of reflected and refracted Riemann functions $F^{I}(x, y)$ and $F^{II}(x, y)$.

We define a wave front set $WF(u)$ and a analytic

wave

front set $WF_{A}(u)$ (cf.

(5)

Definition

1.1.

Let $u(x)\in D’(X)$. Then the wave front set $WF(u)$ is defined as

the complement in $X\cross(\mathrm{R}^{n}\backslash \{0\})$ of the collection of the points $(x^{0}, \xi 0)$ such that

there exist a conic neighborhood $\triangle$ of$\xi^{0}$ in $\mathrm{R}^{n}\backslash \{0\}$ and $\phi\in C_{0}^{\infty}(X)$ such that

$\phi(x^{0})\neq 0$ and

$|\mathcal{F}[\phi u](\xi)|\leq C_{N}(1+|\xi|)^{-N}$ when $\xi\in\triangle \mathrm{a}\mathrm{n}\mathrm{d}$ $N=0,1,2,$ $\cdots$

Here $\mathcal{F}$ denotes the Fourier transform.

For the definition of a analytic wave front set $WF_{A}(u)$, we prepare that there

exist a bounded sequence $\{\phi_{N}\}$ in $C_{0}^{\infty}$ such that $\phi_{N}=1$ on a fixed neighborhood

of$x^{0’}$ in $X$

, independent of$N$, and

$|D^{\alpha}\phi_{N}|\leq C(CN)^{1}\alpha|$ for $|\alpha|\leq N$.

Definition 1.2. Let $u(x)\in D^{/}(X)$. Then the analytic wave front set $WF_{A}(u)$ is defined as the complement in $X\cross(\mathrm{R}^{n}\backslash \{0\})$ of the collction of the points

$(x^{0}, \xi 0)$ such that for some sequence $\{\phi_{N}\}$ of the above type there exists a conic

neighborhood $\triangle$ of$\xi^{0}$ in $\mathrm{R}^{n}\backslash \{0\}$ with

$|\mathcal{F}[\phi_{N}u](\xi)|\leq C(CN)^{N}(1+|\xi|)^{-N}$ when $\xi\in\triangle$, $N=0,1,2,$$\cdots$

By Definition 1.1 and Definition 1.2, we obtain

$WF(u)\subset WF_{A}(u)\backslash$,

and $WF(u)$ and $WF_{A}(u)$ are closed subsets of$X\cross(\mathrm{R}^{n}\backslash \{0\})$.

We define a localization of polynomials according to $\mathrm{A}\mathrm{t}\mathrm{i}\mathrm{y}\mathrm{a}\mathrm{h}- \mathrm{B}\mathrm{o}\mathrm{t}\mathrm{t}_{- \mathrm{G}^{\circ}\mathrm{r}\mathrm{d}}\mathrm{a}\mathrm{i}\mathrm{n}\mathrm{g}$ (cf.

[1]$)$:

Definition 1.3. Let $P(\xi)$ be a polynomial ofdegree $m\geq 0$ and develop

$l\text{ノ^{}m}P(l\text{ノ}-1\xi+\eta)$ in ascending power of $\nu$:

$\nu^{m}P(\nu^{-1}\xi+\eta)=\nu^{p}P_{\xi}(\eta)+o(_{\mathcal{U}}p+1)$ as J ノ $arrow 0$,

where $P_{\xi}(\eta)$ is the first coefficient that does not vanish identically in

$\eta$

.

The

poly-nomial $P_{\xi}(\eta)$ is the localization of $P$ at $\xi$

,

the number

$p$ is the multiplicity of $\xi$

relative to $P$

.

Moreover we introduce the following:

Definition 1.4. $\Gamma=\Gamma(P, \theta)$ is the component of$\mathrm{R}_{\eta}^{n}\backslash \{\eta\in \mathrm{R}_{\eta}^{n}, P(\eta)=0\}$ which

contains $\theta=(1,0, \cdots, 0)\in \mathrm{R}^{n}$. Moreover $\Gamma’=\mathrm{r}^{l}(P, \theta)=\{x\in \mathrm{R}^{n}|x\cdot\eta\geq 0,$

$\eta\in$

$\Gamma\}$ is the dual cone of $\Gamma$ and is called the propagation cone.

3.

Results

We obtain the exact wave front sets of the elastic mixed problem in stratified

media by combining an inner estimate and

an

outer estimate. First we mention

about the results of the incident Riemann function, namely

fundamental

solution

ofMediumI $E^{I}(x)$

.

This proposition is aversionofthe theorem provedby

(6)

Proposition. For$\xi^{0}\in \mathrm{R}_{\xi}^{4}\backslash \{0\}$ satisfying $(\det P_{j}^{I})(\xi^{0})=0(j\in\{p_{1}, s_{1}\})$, that is, $(\det P_{p_{1}1}^{I})(\xi 0)=\xi_{0}^{0^{2}}-C_{\mathrm{p}}2|\xi 0//|^{2}=0$,

$or$

$(\det P_{s_{1}}^{I})(\xi^{002})=\xi 0^{2}-C|s_{1}\xi 0//|^{2}=0$,

then we have

$\lim_{\nuarrow\infty}\nu e^{-}i\nu x\cdot\xi 0IE(X)=E_{j}I(_{X}\xi 0)$, $j\in\{p_{1}, s_{1}\}$,

in the distribution

sense

with respect to $x\in \mathrm{R}^{4}f$ where

$E_{j\xi^{0}}I(X)=(2 \pi)-4\int \mathrm{R}_{\xi}4eix\cdot(\xi+i\eta)\frac{(_{\mathrm{C}\mathrm{o}\mathrm{f}}P^{I})_{\xi^{0}}(\xi+i\eta)}{(\det P^{I}j)_{\xi}\mathrm{o}(\xi+i\eta)}d\xi$

for $\eta\in-\theta-\Gamma(\det PI\theta)j$

’ and $j\in\{p_{1}, s_{1}\}$

with a positive real $s$ large enough. Moreover we have

$WF(E^{I}(_{X)})= \bigcup_{\xi^{0}\neq 0}(\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{P}^{E}p_{1}\xi^{0}(_{X})\cup \mathrm{s}\mathrm{u}\mathrm{p}\mathrm{P}E0\xi(s1)X)II\cross\{\xi^{0}\}$,

and

$\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{p}E_{j\xi}^{I}0(x)=(\Gamma_{j\xi^{\mathrm{O}}})’=\{x\in \mathrm{R}^{4}$ : $x\cdot\eta\geq 0$ for any $\eta\in \mathrm{r}_{\mathrm{i}\xi^{0}}$

$u=\Gamma((\det P_{j}^{I})_{\xi}\mathrm{o}(\eta), \theta)\}$, $\theta=(1,0,0,0)$, $j\in\{p_{1}, s_{1}\}$.

In general, $\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{P}^{E_{j\xi}^{I}}\mathrm{o}(x)\subset(\Gamma_{j\xi^{0}})^{/}(j=\{p_{1\mathcal{J}}.s_{1}\})$, more precisely $ch[\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{p}E_{j\xi}^{I}0(x-$

$y)]=(\mathrm{r}_{j\xi^{0}})/$, where $ch$ denotes

convex

hull. However in our problem we obtain

$\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{p}E_{j\xi^{0}}^{I}(x)=(\Gamma\xi^{0})j/$ .

Secondly we mention about main result.

Since

we take a partial Fourier-Laplace

transform with respect to $x’$ of $\delta(x-y)$ regarding $y’$ as a parameter, $y’\mathrm{a}_{\mathrm{P}\mathrm{P}^{\mathrm{e}\mathrm{a}\mathrm{r}\mathrm{s}}}$

only in the form $x’-y’$. So we put

$\tilde{F}^{\iota}(x, y_{3})=F^{L}(_{X,0,)}y3,$ $\iota=\{I, II\}$.

The following Main Theorem shows the exact wave front sets of $\overline{F}^{I}(x, y_{3})$ and

$\tilde{F}^{II}(x, y_{3})$.

Main Theorem. For$\xi^{0}\in \mathrm{R}_{\xi}^{4}\backslash \{0\}$ satisfying $(detP_{j}^{I})(\xi^{0})=0(j\in\{p_{1}, S1\})$ we

have the following:

(1) For the

reflected

Riemann

function

$\tilde{F}^{I}(x, y\mathrm{s})$, we have

(3.1) $\lim_{\nuarrow\infty}\nu e^{-i\{}k)-y3\xi^{0}3\}\tilde{F}\nu x\xi^{0-(\xi}’\cdot\prime 0’(+x3\mathcal{T}l)X,$$y_{3}=\tilde{F}^{I}0k(j\xi X, y_{3})$,

(7)

and

if

$\xi^{0’}$ are

zeros

of

$\tau_{m}^{+}(\zeta’)$, that $i\mathit{8}_{y}\xi^{0}/sati_{\mathit{8}}fy|\xi^{0^{\prime//}}|=\xi_{0}^{0}/cm(m\in\{p_{1},p_{2,2}s\})$,

then we have

(3.2)

$\lim_{\nuarrow\infty}\{\nu^{\frac{3}{2}}e^{-}(-\xi^{0})-y3\xi_{3}0_{\}I}\tilde{F}x\cdot+x_{3}\tau_{k}(Xi\nu\{;_{\xi}0’;, y3)-\nu^{\frac{1}{2}}\tilde{F}^{I}0k(j\xi x, y3)\}=\tilde{F}^{I}j\xi 0km(x, y3)$,

$(j, k, m)=\{(p_{1},p1,p_{2}),$ $(p_{1},p\mathrm{l}, S2),$$(p1, S1,p_{2}),$ $(p_{1}, S_{1}, S2),$ $(s1,p1,p2)$,

$(s_{1},p_{1,2}s),$ $(s1, s_{1},p_{2}),$$(S1, s_{1,2}S),$ $(s_{1}, s1,p_{1})\}$

in the dist$7’ ibution$

sense

with respect to $(x, y_{3})\in \mathrm{R}_{-}^{4}\cross \mathrm{R}_{-}$

.

Moreover we have

$WF( \tilde{F}^{I}(x’, x_{3}, y3))=WFA(\tilde{F}^{I}(_{X}/, x_{3},y_{3}))=\xi\neq 0\bigcup_{0}$

$[_{(}j,k)=p1,’ \mathrm{P}1’ p1,s_{1}),/\cup(\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{p}\tilde{F}I(j\xi 0kx, X_{3},y_{3})\cross\{(\xi 0’, \mathcal{T}^{-(\xi}0’k), -\xi_{3}^{0})\})$

$\cup$

$(j,k^{\bigcup_{m}},)=$

$(\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{p}\tilde{F}_{j\xi km}^{I}0(X,$$x/3,$

$y_{3})\cross\{(\xi^{0^{//}}, \mathcal{T}_{k^{-}}(\xi \mathrm{o}), -\xi 30)\})]$

$\{(p_{1},p1,p2),(p_{1p_{12}),(},,Sp1,s1,p_{2})$,

$(p_{1,1S}s,2),(s_{1,p1},\mathrm{P}2),(s_{1},p1,s_{2})$,

$(s_{1},s_{1},p2),(_{S1s_{1},s_{2})},,(s1,s1,p1)\}$

and

$\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{p}\tilde{F}_{j\xi^{0}}^{I}(kX, y)=(\Gamma_{j\xi}0)k\equiv I\{(_{X}, y_{3})\in \mathrm{R}^{4}-\cross \mathrm{R}_{-}$ :

$(X’+X_{3\mathrm{g}\mathrm{a}}\mathrm{r}\mathrm{d}\xi\tau_{k}^{-}(\xi 0^{/}))\cdot\eta-/y3\eta_{3}\geq 0$ for any

$\eta\in\Gamma_{j\xi^{0}}\}$,

$(j, k)=\{(p1,p_{1}), (p1, S1), (S1,p1), (S_{1}, s_{1})\}$

for

$\xi^{0}$ satisfying

$\tilde{F}_{j\xi^{0}k}^{I}(x, y3)\neq 0$,

$\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{p}\overline{F}_{j\xi^{0}k}^{I}(mx, y)=(\Gamma_{j\xi^{0_{m}}})_{k}^{I}\equiv\{(x, y_{3})\in \mathrm{R}^{4}-\cross \mathrm{R}_{-}$:

$(x’+x3\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\xi\tau^{-}k(\xi 0’))\cdot\eta-/y_{3}\eta_{\mathrm{s}}\geq 0$ for any

$\eta\in\Gamma_{j\xi^{0}m}\}$,

$(j, k, m)=\{(p_{1}, p1,p_{2}),$ $(p1,p1, S2),$ $(p17S_{1},p2),$$(p1, s1, s2),$ $(_{S_{1,p_{1}}},p2)$,

$(s_{1},p_{1,2}S),$$(s_{1}, s1,p2),$ $(s_{1,1}s, s2),$ $(S1, S1,p_{1})\}$

for

$\xi^{0}$ satisfying $\tilde{F}_{j\xi^{0}km}^{I}(x, y3)\neq 0$.

(2) For the

refracted

Riemann

function

$\tilde{F}^{II}(x, y\mathrm{s})$, we have

(8)

$(j, k)=\{(p_{1},p_{2}), (p_{12}, s), (s_{1p2},), (S_{1}, S_{2})\}$

and

if

$\xi^{0’}$ are

zeros

of

$\tau_{m}^{+}(\zeta’)(m\in\{p_{2}\})$, then we have

(3.4)

$\lim_{\nuarrow\infty}\{\nu^{\frac{3}{2}-}ek\xi 0’)-y3\xi 3\}\overline{F}i\nu\{x’\cdot\xi 0’x+3\tau^{\dagger}(0II(x, y_{3})-\nu^{\frac{1}{2}}\tilde{F}_{j\xi k(}II0X, y_{3})\}=\tilde{F}^{I}Imj\xi 0k(x, y_{3})$ ,

$(j, k, m)=\{(p1_{\text{ノ}}.s_{2,p}2), (_{S_{1}}, S2,p2)\}$

in the distribution

sense

with respect to $(x, y_{3})\in \mathrm{R}_{+^{\mathrm{X}}}^{4}\mathrm{R}_{-}$.

Moreover we have

$WF( \tilde{F}^{II/}(X, x3, y3))=WFA(\tilde{F}^{II}(Xx\mathrm{s}, y/,3))=\bigcup_{\xi 0\neq 0}$

$[_{(j,k)\mathrm{t}(}(_{\mathit{8}1,p_{2}}),’(s1,s_{2}) \}x=\mathrm{P}1p\bigcup_{22),(p_{1},s)},$$(\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{P}^{\tilde{F}(}j\xi 0kx3,$$y_{3})\mathrm{X}\{(II/,\xi 0’.(J\tau k^{+}\xi^{00}/), -\xi_{3})\})$

$\cup(j,k,m)=\{(p1,s2,p_{2}),(\bigcup_{2s1,s,p_{2})\}}(I/,)\cross\{(\xi^{0}\tau_{k}(\xi), -\xi \mathrm{s})\})(^{\sup \mathrm{P}^{\tilde{F}_{j}^{I}(}}\xi 0kmXx_{3}, y3/,+0’0]$

and

$\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{p}\tilde{F}_{j\xi k}^{lI}0(_{X}, y_{3})=(\mathrm{r}_{j\epsilon^{0}})^{II}k\equiv\{(x, y3)\in \mathrm{R}_{+}^{4}\cross \mathrm{R}_{-:}$

$(x’+x\mathrm{s}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}_{\xi}\mathcal{T}(k\xi^{0}+/))\cdot\eta’-y3\eta 3\geq 0$ for any $\eta\in\Gamma_{j\xi^{0}}\}$,

$(j, k)=\{(p_{1},p2), (p1, s2), (s1,p_{2}), (s1, s_{2})\}$

for

$\xi^{0}$ satisfying $\tilde{F}_{j\xi^{0}k}^{II}(X, y_{3})\neq 0_{f}$

$\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{p}\tilde{F}_{j\xi km}^{lI}\mathrm{o}(_{X}, y_{3})=(\Gamma_{j\xi m}0)_{k}II\equiv\{(x, y_{3})\in \mathrm{R}_{-}^{4}\mathrm{x}\mathrm{R}_{-}$

:

$(x’+x_{3}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}_{\xi}\mathcal{T}_{k}+(\xi^{0}’))\cdot\eta-/y_{3}\eta 3\geq 0$for any $\eta\in\Gamma_{j\xi^{0}m}\}$,

$(j, k, m)=\{(p1, S2,p_{2}), (s1, S2,p_{2})\}$

for

$\xi^{0}$ satisfying $\overline{F}_{j\xi^{0}}^{II}(kmx, y3)\neq 0$

.

Here$F_{j\xi^{0}k}^{I}(X, y_{3}),$ $F_{j\xi^{0}km}^{I}(x, y_{3})_{\mathrm{z}}F_{j\xi^{0}k}^{II}(x, y_{3})$, and$F_{j\xi^{0}km}^{lI}(x, y\mathrm{s})$ are localizations

in the

sense

of

(3.1), $(\mathit{3}.\mathit{2})_{f}(\mathit{3}.\mathit{3})_{y}$ and $(\mathit{3},\mathit{4})_{f}re\mathit{8}pectively_{f}$ and

more

$preC?\mathit{8}e$

expres-sions are given in $([7]y[8])$. Moreover

$\Gamma_{j\xi^{0}}=\Gamma((\det P^{I})_{\xi^{0}}(\eta), \theta j)$, $\theta=(1,0,0, \mathrm{o})$, $j\in\{p_{1}, s_{1}\}$,

(9)

$\theta’=(1,0, \mathrm{o})$, $j\in \mathrm{f}^{p_{1}},$ $s_{1}\}$

,

$m\in\{p_{2}\}$,

$\tau_{p_{1}}^{\pm}(\xi’)=\mathrm{S}\mathrm{g}\mathrm{n}(\mp\xi_{0})\sqrt{\frac{\xi_{0}^{2}}{c_{p_{1}}^{2}}-|\xi^{0^{\prime//}}|^{2}}$, if

$\frac{\xi_{0}^{2}}{c_{p_{1}}^{2}}-|\xi 0^{\prime//}|2\geq 0$,

and $\tau_{p_{1}}^{\pm}(\xi’)$ is taken a branch

of

$\sqrt{\frac{\xi_{0}^{2}}{c_{\mathrm{p}_{1}}^{2}}-|\xi^{0^{\prime//}}|^{2}}\mathit{8}uchthat\pm{\rm Im} \mathcal{T}_{p_{1}}^{\pm}(\xi’)>0$

if

$\frac{\xi_{0}^{2}}{c_{p\mathrm{J}}^{2}}$

-$|\xi^{0^{;J/}2}|<0$ . $\tau_{s_{1}}^{\pm}(\xi’)_{f}\tau_{p_{2}}^{\pm}(\xi’)$, and $\tau_{s_{2}}^{\pm}(\xi’)$ are

defined

as the

same

$a\mathit{8}\tau_{p_{1}}^{\pm}(\xi^{l})$

substi-tuting$c_{p_{1}}$

for

$c_{s_{1}\mathrm{z}}c_{p_{2}}$, and$c_{s_{2})}$ respectively.

REFERENCES

1. M. F. Atiyah, R.Bott, L. $\mathrm{G}[mathring]_{\mathrm{a}}\mathrm{r}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{g}$, Lacunas

forhyperbolic differentialoperators urith constant

coefficients I, Acta Math. 124 (1970), 109-189.

2. L. H\"ormander, On the singularities ofsolutionsofpamaldifferential equations, International

Conferenceof Functional Analysis and Related Topics, Tokyo (1969).

3. L. $\mathrm{H}_{\ddot{\mathrm{O}}\mathrm{r}\mathrm{m}\mathrm{a}}\mathrm{n}\mathrm{d}\mathrm{e}\mathrm{r}$, Uniqueness

theorems and wavefront sets for solutions of linear differential

equations with analytic coefficients, Comm. Pure Appl. Math. 24 (1971), 671-703.

4. L. H\"ormander, The $Analy\dot{\Re}s$ of Linear Partial Differential Operators I, Springer-verlag,

Berlin, Heidelberg, NewYork, Tokyo, 1983.

5. M. Matsumura, Localization theorem in hyperbolic mixed problems, Proc. Japan. Acad. 47

(1971), 115-119.

6. S. Shimizu, Eigenfunction expansionsforelastic wave$p\tau \mathrm{o}$pagation probiemsinstratifiedmedia

$R^{3}$, Tsukuba J. Math. 18 (1994),

283-350.

7. S. Shimizu, Singularities ofsolutions to elastic wavepropagahonproblems instmtifiedmedia

II, RIMS kokyuroku 1045 (1998), 73-102.

8. S. Shimizu, $Singula\dot{\mathcal{H}}tieS$ of the refiected and

refracted

Riemann

functions of elastic wave

propagation prvblems in 8tratified media, ISAAC’97 proceedingon wave and related topics,

KluwerAcademic Publishers, 1999 (to appear).

9. M. Tsuji, Propagation ofthe singularitiesforhyperbolic equations with constant $coeff_{\mathfrak{X}}ientS$,

Japan J. Math. 2 (1976), 369-373.

10. S. Wakabayashi, $Singula7\dot{\mathrm{v}}ties$ ofthe R\’iemann functions ofhyperbolic $m\dot{?}xed$

problemノs in a

quater-space, Proc. Japan. Acad. 50 (1974), 821-825.

11. S. Wakabayashi, $Singula\dot{\mathcal{H}}ties$ of the Riemann functions ofhyperbolic $m?\acute{x}ed$

problems in a

quarter-space, Publ. RIMS Kyoto Univ. 11 (1976), 417-440.

12. S. Wakabayashi, Analyticwaveffont sets ofthe Riemannfunctions ofhyperbolic mixed prob-lems in a quarter-space, Publ. RIMS Kyoto Univ. 11 (1976), 785-807.

13. S. Wakabayashi, Prvpagationof$singula\gamma\dot{\mathrm{v}}$ties ofthefundamentalsolutions

ofhyperbolic mixed

Figure 1 Reflected waves and refracted waves
Figure 2 Lateral waves
Figure 3 Stratified media We assume that

参照

関連したドキュメント

In recent years, several methods have been developed to obtain traveling wave solutions for many NLEEs, such as the theta function method 1, the Jacobi elliptic function

7, Fan subequation method 8, projective Riccati equation method 9, differential transform method 10, direct algebraic method 11, first integral method 12, Hirota’s bilinear method

In this paper, by using the integral bifurcation method 34–36, we mainly investigate some new exact solutions such as explicit solutions of Jacobian elliptic function type

A new method is suggested for obtaining the exact and numerical solutions of the initial-boundary value problem for a nonlinear parabolic type equation in the domain with the

Zhang, “The G /G-expansion method and travelling wave solutions of nonlinear evolution equations in mathematical physics,” Physics Letters A, vol. Li, “Application of the G

The damped eigen- functions are either whispering modes (see Figure 6(a)) or they are oriented towards the damping region as in Figure 6(c), whereas the undamped eigenfunctions

[18] , On nontrivial solutions of some homogeneous boundary value problems for the multidi- mensional hyperbolic Euler-Poisson-Darboux equation in an unbounded domain,

Interesting results were obtained in Lie group invariance of generalized functions [8, 31, 46, 48], nonlinear hyperbolic equations with generalized function data [7, 39, 40, 42, 45,