Inverse problem for viscoelastic system in a vertically layered medium
Автор: Boltaev Asliddin A., Durdiev Durdimurod K.
Журнал: Владикавказский математический журнал @vmj-ru
Статья в выпуске: 4 т.24, 2022 года.
Бесплатный доступ
In this paper, we consider a three-dimensional system of first-order viscoelasticity equations written with respect to displacement and stress tensor. This system contains convolution integrals of relaxation kernels with the solution of the direct problem. The direct problem is an initial-boundary value problem for the given system of integro-differential equations. In the inverse problem, it is required to determine the relaxation kernels if some components of the Fourier transform with respect to the variables x1 and x2 of the solution of the direct problem on the lateral boundaries of the region under consideration are given. At the beginning, the method of reduction to integral equations and the subsequent application of the method of successive approximations are used to study the properties of the solution of the direct problem. To ensure a continuous solution, conditions for smoothness and consistency of initial and boundary data at the corner points of the domain are obtained. To solve the inverse problem by the method of characteristics, it is reduced to an equivalent closed system of integral equations of the Volterra type of the second kind with respect to the Fourier transform in the first two spatial variables x1, x2, for solution to direct problem and the unknowns of inverse problem. Further, to this system, written in the form of an operator equation, the method of contraction mappings in the space of continuous functions with a weighted exponential norm is applied. It is shown that with an appropriate choice of the parameter in the exponent, this operator is contractive in some ball, which is a subset of the class of continuous functions. Thus, we prove the global existence and uniqueness theorem for the solution of the stated problem.
Viscoelasticity, resolvent, inverse problem, hyperbolic system, fourier transform
Короткий адрес: https://sciup.org/143179315
IDR: 143179315 | DOI: 10.46698/i8323-0212-4407-h
Текст научной статьи Inverse problem for viscoelastic system in a vertically layered medium
A perfectly elastic material does not exist in nature; in fact, inelasticity is always present. This inelasticity results in energy dissipation or damping. Therefore, for a wide class of materials, it is not enough to use an elastic model to study their mechanical behavior. Therefore, viscoelastic foundational models have often been used to model the behavior of polymeric materials with respect to time variable.
-
# The research is supported by the Ministry of Education and Science of Russia, agreement № 075-022022-896.
Let be x = (х 1 ,х 2 ,х з ) € R 3 - Let us denote by O ij the projection onto the x i axis of the stress acting on the area with the normal parallel to the x j axis, and u i are the projection onto the x i axis of the vector particle displacement. According to Hooke’s law for viscoelastic media, stresses and deformations are related by the formulas [1, pp. 449–455], [2, ch. 3]:
∂u i ∂u j
^ij(x,t) = ^dj + dx-J + ^ij A div u
t
+ j Kij(t - т) |a (dX + dX) + 5ij A div uj (х,т) dT, i,j = 1, 2, 3, (1)
0 ji here ^ = ^(хз), A = А(хз) are Lame coefficients, 6ij is Kronecker symbol, Kij(t) are functions responsible for the viscosity of the medium and Kij = Kji, i, j = 1, 2, 3.
The equations of motion of a viscoelastic body particles in the absence of external forces have the form
∂ 2 u i ρ ∂t 2
3 ∂σ
= У ,, i = 1,2,3, ∂xj j=1
where p = р(х з ) is medium density, U(x,t) = (U i (x,t),U 2 (x,t),U 3 (x,t)) is displacement vector. Throughout this work, µ , λ , ρ are considered to be given functions.
Note that (1) can be considered as integral Volterra equations of the second kind with respect to the expression ^ (ddl i + d j) + ^ ij A div U, i, j = 1, 2, 3. For each fixed pair (i, j) solving these equations, we get
t
∂ui ∂uj oij (x,t)= ^(dX" + aX )+ 5ij A div u + rij (t - T )ст. (х,т) dT, i,j = 1, 2, 3, (3)
ji 0
where r ij are the resolvents of the kernels K ij and they are related by the following integral relations [3, 4]:
t rij(t) = -Kij(t) - j Kij(t - т)rij(t) dT, i,j = 1, 2, 3.
From the condition K ij = K ji implies the r ij = r ji .
Differentiating (3) with respect to t and introducing the notation u i = dt U i , we get
t
∂
∂t σ ij
∂u i ∂u j
(x,t) = ^(dX" + dX/+ 5 ij A div u + r ij (°)a ij (x,t)+ rij (t - т )ст. (х,т) dT. (5)
ji
Then the system of equations (1) and (2) for the velocity u i and strain ct . ( ct . = j ) in view of (3) — (5) can be written as a system of first-order integro-differential equations.
t
(A-X + B-X— + C"X— + D-X— + f) U(X,t)= [ R(t - т)U(х,т) dT,
∂t ∂x1 ∂x2
where U = (u i ,U 2 ,U 3 , ^ 11 , ^ 12 , ^ 13 , ^ 22 , ^ 23 , ^ 33 ) * , * is the transposition sign,
C2 =
D2 =
A =
(pI) 3 x 3 |
( O ) 3 x 3 |
( O ) 3 x 3 |
( O ) 3 x 3 |
(I) 3 x 3 |
( O ) 3 x 3 |
( O ) 3 x 3 |
( O ) 3 x 3 |
(I) 3 x 3 . |
/ — (A + 2ц) 0 |
||
B 1 = |
0 |
- µ |
0 |
0 |
|
( O ) 3 x 3 |
(C 1 ) 3 x 3 |
|
C = ( |
(C 3 ) 3 x 3 |
( O ) 3 x 3 |
(C 4 ) 3 x 3 |
(O) 3 x 3 |
,
(C 2 ) 3 x 3
( O )
( O )
- µ
-
,
C3 =
- µ
B=
3 x 3
3 x 3
-
λ
,
( O ) 3 x 3
(B 1 ) 3 x 3
(B 2 ) 3 x 3
B2 =
,
,
C1 =
-
D=
(O) 3 x 3
(D 3 ) 3 x 3
(D 4 ) 3 x 3
-
,
-
F =
(O)
(O)
3 x 3
6 x 3
(D 1 ) 3 x 3
( O ) 3 x 3
(O) 3 x 3
D3 =
(D 2 ) 3 x 3
( O ) 3 x 3
( O ) 3 x 3
- µ
,
-
λ
,
( — I) 3 x 3
(O)
(O)
3 x 3
3 x 3
- λ
- λ
C4 =
D1 =
D4 =
( O ) 3 x 3
( O ) 3 x 3
( O ) 3 x 3
,
,
-
,
-
(A + 2ц)
-
-
- µ
,
λ
- µ
-
,
-
λ
(A + 2ц)
,
( O ) 3 x 6
diag (rn(0), Г22(0), Г33(0), Г12(0), Г13(0), Г23
(0)) ,
p(+\ = ( (O) 3 x 3 (O) 3 x 6 A
(t) dbrnY'
\ ( O) 6 x 3 diag (r 11 , r 22 , r 33 , r 12 , r 13 , r 23) /
The system (6) can be reduced to a symmetric hyperbolic system [5].
We reduce the system (6) to canonical form with respect to the variables t and X 3 . To do this, multiply (6) on the left by A - 1 and compose the equation
| A - 1 D - vI I =0,
where I is the identity matrix of dimension 9. The last equation with respect to ν has following solutions:
ν 1
-V9 = -Vp =
IA + 2ц ρ
v 2,3 = — v 7,8 = —v s
-
v 4,5,6 = 0,
here v s and v p define velocities of the transverse and longitudinal seismic wave, respectively. Now we choose a nondegenerate matrix Y(x 3 ,t) so that the equality
Y-1A-1DY = Л
is hold, where Л is a diagonal matrix, the diagonal of which contains the eigenvalues (for each fixed X 3 ) (8) of the matrix A - 1 D that is Л = diag ( — v p , —v s , —v s , 0, 0, 0, v s , v s , v p ).
From the formula (9) implies the equality
A-1 DY = YЛ, which means that the column with the number i of the matrix Υ is an eigenvector of the matrix A-1DΥ, corresponding to the eigenvalue λi. Direct calculations show that the matrix Υ, satisfying the above conditions, can be chosen as (not uniquely)
0 0 |
0 1 |
1 0 |
0 0 |
0 0 |
0 0 |
0 1 |
1 0 |
0 0 |
||
1 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
1 |
||
λ |
0 |
0 |
1 |
0 |
1 |
0 |
0 |
λ |
||
ν p |
ν p |
|||||||||
Y(X 3 ) = |
0 |
0 |
0 |
0 |
1 |
0 |
0 |
0 |
0 |
|
0 |
0 |
ρν s |
0 |
0 |
0 |
0 |
-PV s |
0 |
||
λ |
0 |
0 |
0 |
0 |
1 |
0 |
0 |
λ |
||
ν p |
v p |
|||||||||
0 |
ρν s |
0 |
0 |
0 |
0 |
-PV s |
0 |
0 |
||
ρν p |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
-Pvp / |
We introduce the vector function ϑ by the equality
U = Υϑ.
Making this change in the equation (6) and then multiplying it on the left by Υ - 1 A - 1 , then we get
t
( 4 +Л-^- + Bvl— + C1^— + F1) ^(x,t)= [ Ri(t - т,хз)^(Х,т) dr,
∂t ∂x 3 ∂x 1 ∂x 2
where
B 1 (x 3 ) |
= Υ - 1 A - 1 BΥ = (b ij ), F 1 (x 3 ) = Υ - 1 A - 1 D ∂Υ ∂x 3 R 1 (x 3 , t) = Υ - 1 A |
C 1 (x 3 ) = Υ - 1 A - 1 CΥ |
= (c ij ) , |
||||||||
+ Υ - 1 A - 1 F Υ = |
(p ij ) , |
||||||||||
- 1 RΥ = |
^ ij ) |
||||||||||
/ |
r 3′ 3 2 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
r 3′3 2 |
\ |
|
0 |
r ′23 2 |
0 |
0 |
0 |
0 |
r 2′ 3 2 |
0 |
0 |
|||
0 |
0 |
r 1′3 2 |
0 |
0 |
0 |
0 |
r 1′3 2 |
0 |
|||
λ(r 1 ′ 1 - r 2 ′ 2 ) ν p |
0 |
0 |
r 1 ′ 1 |
0r |
′′ 11 r 22 |
0 |
0 |
λ(r 2 ′ 2 - r 1 ′ 1 ) ν p |
(11) |
||
0 |
0 |
0 |
0 |
r 1 ′ 2 |
0 |
0 |
0 |
0 |
. |
||
λ(r 2 ′ 2 - r 3 ′ 3 ) ν p |
0 |
0 |
0 |
0 |
r 2 ′ 2 |
0 |
0 |
λ(r 3 ′ 3 - r 2 ′ 2 ) ν p |
|||
0 |
r 2′3 2 |
0 |
0 |
0 |
0 |
r 2′3 2 |
0 |
0 |
|||
0 |
0 |
r 1′3 2 |
0 |
0 |
0 |
0 |
r 1′3 2 |
0 |
|||
1 |
r 3′3 2 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
r 3′ 3 2 |
The purpose of this article is to study the direct and inverse problems for the system (11). Moreover, the direct problem is an initial-boundary value problem for this system in domain D = { (x 1 , x 2 , X 3 , t) : (x 1 ,x 2 ) E R 2 , X 3 E (0, H), t > 0 } , H = const, and in the inverse problem, the elements of the matrix R are assumed to be unknown, which are included in the definition of the matrix R 1 (12).
The is organized as follows. Section 1 presents the formulations of the direct and inverse problems and investigates the direct problem. In Section 2, the inverse problem is reduced to solving of an equivalent closed system of integral equations. In Section 3, we present the formulation and proof of the main result, which consists in the unique global solvability of the inverse problem. At the end there is a list of literatures used in the article.
1. Statement of the Direct and Inverse Problems
Consider the system of equations (10) in the domain D with a bounded Г = Г о U Г U Г 2 :
Г 0 = { (x,t) : (x i ,x 2 ) G R 2 , 0 < x 3 ^ H, t = 0 } ,
Г 1 = { (x, t) : (x 1 ,x 2 ) G R 2 , x 3 = 0, t > 0 } ,
Г 2 = { (x, t) : (x i ,x 2 ) G R 2 , x 3 = H, t > 0 } .
For this system the direct problem we pose as follows: determine the solution of the system of equations (10) at the following initial and boundary conditions:
^ i \ t =0 = ^ i (x), i = 1,..., 9, (12)
^ i \ x 3 = H = g i (x i ,x 2 ,t), i = 1 2, 3 ^ i | x 3 =0 = g i (x i ,x 2 ,t), i = 7 8, 9. (13)
Here ^ i (x), g i (x i ,x 2 ,t) are given functions. It is known that [5, 6] the problem (10), (12), (13) is posed well.
The inverse problem is to determine the nonzero components of the matrix kernel R, that is r j (t) , i,j = 1, 2, 3 ( R is included in R i according to the formula (12)) in (10) if the following conditions are known:
^ iL = h i (x i ,x 2 ,t), i = 1,..., 6, (14)
x3= where hi (xi , x2, t), i = 1, . . . , 6, are the given functions.
In the inverse problem, the numbers r ij (0) , i, j = 1, 2, 3 , are also considered to be given.
Currently, the problems of determining kernels from one hyperbolic integro-differential equations of the second order [7–22] have been widely studied. One- and multidimensional inverse problems are investigated and unique solvability theorems are obtained. Typically, second-order equations are derived from systems of first-order partial differential equations under some additional assumptions.
The inverse problem of determining the convolution kernels of integral terms from a system of first-order integro-differential equations of general form with two independent variables was studied in [23]. The theorem of local existence and global uniqueness is obtained. In the work of [24], the method for studying the work of [23] was applied to the investigating of the inverse problem of determining the diagonal relaxation matrix from the system of Maxwell’s integro-differential equations.
It seems completely natural to study inverse problems on the determination of the kernels of integral terms of a system of integro-differential equations directly in terms of the system itself. This article is a natural continuation of this circle of problems and to a certain extent generalizes the results of [23] to the case of a three-dimensional system of viscoelasticity equations (1), (2).
Let functions ^(x), g i (x i ,x 2 , t) included in the right-hand side of (10) and the data (12) , (13) are compact support in x i , x 2 for each fixed x 3 , t . From the existence for the system (10) of a compact support domain of dependence and compact support with respect to x 1 , x 2 of the right-hand side (10) and data (12) , (13) implies the compact support in x i , x 2 solutions to the problem (10) - (13) .
Let us study the property of solution to this problem. More precisely, we restrict ourselves to studying the Fourier transform in the variables x i , X 2 of the solution. In what follows, for convenience, we put X 3 = z and introduce the notation
----
^ (n i ,n 2 ,z,
t)=/
R 2
^(x 1 , x 2 , z, t)e i[n 1 x 1 +n 2 x 2 1 dx 1 dx 2 ,
where n i , П 2 are transformation parameters. We fix n i , П 2 and for convenience, we introduce the notation ^(n i , П 2 , z, t) = ^(z, t) •
In terms of the function ϑ we write the equations (10) as
/ d d \ x"- - /A"-
— т) dr,
I dt + vjdzpj(z,t) = Z^P'jk(z^k(z,t) + ZVi^-k)^k(z,t z k=1 0 k=1
j = 1,..., 9, where pjk = —inibjk — in2Cjk - Pjk-
We will use a similar notations for the Fourier images of functions included in the initial, boundary and additional conditions (12)–(14):
^ilt=0 = A|z=H = 9г(t), i = 1, 2, 3, ^i|z=0 = /(t), i = 7, 8, 9, ^i|z=0 = hi(t), i = 1,..., 6. Where A(z), i = 1, • • •, 9, gi(t), i = 1, 2, 3, 7, 8, 9, hi(t), i = 1,..., 6, are the Fourier images of the corresponding functions from (12)-(14) for ni = 0, n2 = 0. We also denote by DH the projection of D onto the plane z, t. In what follows, we will consider the system of equations (15) in the domain Dh U Г under the conditions (16) and (17). Where Го= {(z,t) : 0 < z < H, t = 0}, Г1= {z,t) : z = 0, t > 0}, Г2= {(z,t) : z = H, t > 0}, Г = Го u ri U Г2. ^^ For the purpose of further research let us introduce the vector function w(z,t) = ^(z,t). To obtain a problem for a function w(z,t) similar to (15)-(18) differentiate the equations (15) and the boundary conditions (17) with respect to the variable t, and the condition for t = 0 is found using the equations (15) and the initial conditions (16). In this case, we get ∂∂ ∂t νj∂z "" j Uj (z,t) = ^pj'k (z)wk (z,t) + ^Tjk (z,t)$k (z) k=i k=i t" + / ^/jk(z,T)шк(z,t — т) dr, j = 1,..., 9, (19) 0 k=i U, t ,, = —vi d^d(z) + ^ Pij (z)iPj (z) =: фi(z), i = 1,..., 9, j = i Ui|z=H = -dg/i(t), i = 1, 2, 3, и,|z=0 = -dgi(t), i = 7, 8, 9. dtdt For functions ωi additional conditions (18) gets । d x Ш I = h (t), ilz=0 dt iv , i = 1,..., 6. Let us pass from the equalities (19)–(22) to the integral relations for the components of the vector ϑ with integration flux along the corresponding characteristics of the equations of the system (19). Recall that the characteristics corresponding to νpand νshave a positive slope, and the characteristics corresponding to -νpand -νshave a negative slope. We denote z №i(z) = I i = 1, 2, 3, 7, 8, 9, J Vi(e) №(z) = 0, i = 4, 5, 6. Inverse functions to №i(z) will be denoted by z = щ-10. Using the introduced functions, the equations of characteristics passing through the points (z, t) on the plane of variables £, т can be written in the form т = t + щ (£) - ^i(z), i = 1,..., 9. Consider an arbitrary point (z, t) G Dh on the plane of variables £, т and draw through it the characteristic of the i th of the system (15) equation tell to intersection in the domain т ^ t with boundary Г. The intersection point is denoted by (z0, t0) • Integrating the equations of the system (15) along the corresponding characteristics from the point (z0, t0) to the point (z, t) we find t 9 Wi(z,t) = Wi(z0,t0)+ / ^pikШк(^,т) t0 k=1 dτ Е=-1\т -t+»i(z)] Я9 T9 "I ^ Тгк,(£,т)$г(€)+ / ^ Tk(£,т - а)шк (£,а) da 0k=1 0 k=1 (24) dτ, Е=-1\т -t+»i(z)] i = 1,..., 9. We define in (24) tg. It depends on the coordinates of the point (z,t). It is not difficult to see that ti0(z, t) has the form ti0(z, t) = t - №i(z) + Щг(Н), t > ^i(z) - щг(Н), 0, 0 < t < ^i(z) - щг(Н), i = 1, 2, 3, t0(z,t) = 0, i = 4,5,6, t0(z,t) = ^ t - ^i(z), t > №i(z), 0, 0 i = 7, 8, 9. Then, from the condition that the pair (z0i , ti0) satisfies the equation (23) it follows x f H, t > щ(z) - №(H), z0(z,t) = < ’ P V h i = 1,2,3, (№i 1 (^i(z) - t), 0 < t < №i(z) - щг(Н), zi(z, t) = z, i = 4, 5, 6, zi(z, t) = < , , ^’ i = 7, 8, 9. 0 , 0 V-1Mz) -1) , 0 The free terms of the integral equations (24) are defined through the initial and boundary conditions (20) and (21) as follows: ^i(z0,to) = (55 Hi(t - №(z) + m(H)), I фi (r, ' (№(z) -t)), t > ^(z) - m(H), 0 < t < ^i(z) - ^i (H), i = 1, 2, 3, ^i(z0,t0) = Фi(z), ^i(z0,to) = ( d Иг(t- №(z)) , ( фi (г, ' (№(z) -t)) , i = 4, 5, 6, t ^ ^i(z), 0 i = 7,8,9. Let the following conditions hold Hi (H )= Hi (0) and ддг (t) = - d<3i (z) dtt=0 j3 dz + £Hij (h )Hj (h ), z=H j=i i = 1, 2, 3, (25) + Vp. (0)<Hj (0), i = 7, 8, 9. (26) z=0 j=i It is easy to see that the conditions for matching the initial and boundary data (16), (17) (20), (21) in corner points of the domain Dh coincide with the relations (25) and (26). Hence it is clear that at the fulfillment of the same equalities (25) and (26) equations (24) will have unique continuous solutions Wi(z,t), or the same 57 M^t), i = 1,... , 9 Suppose that all given functions included in (24) are continuous functions of their arguments in DH . Then this system of equations is a closed system of integral equations of the Volterra type of the second kind with continuous kernels and free terms. As usual, such a system has a unique solution in the bounded subdomain Dht = {(z,t) : 0 < z < H, 0 < t < T}, T > 0 are some fixed number. Theorem 1. Assume functions y(x), g(xi,X2,t) have compact support in xi, X2 for each fixed z, t. Let be p(z), ^(z), X(z),Cp(z) G C 1[0, H], g(t) G C1[0,T] , p(z) > 0, A(z) > 0, ^(z) > 0, rj (t) G C [0, T], i, j = 1, 2, 3 and conditions (25), (26) are satisfied. Then there is a unique solution to the problem (19)-(21) in the domain Dht■ The problem (15)–(17) in the domain DHT is equivalent to a linear integral equation of the second kind of Volterra type with respect to ϑ. As follows from the theory of linear integral equations, it has a unique solutions [3]. So we drop it. j dgi(t) ^i(0) = gi (0) and dt _ d<3i(z) j ∂z 2. Reduction of the Inverse Problem In this section, the inverse problem is reduced to solving of an equivalent closed system of integral equations. Consider an arbitrary point (z, 0) G Г and draw through it the characteristics (23) for i = 1, 2, 3, up to the intersection with the boundary of the domain Dht. Integrating the first six components of the equation (19), we obtain wi(z, 0) = dt Hi (ti) i Я9 9 "I ^ Pij(£)^3 (£,T) +^ Tij(£,T)Hj(£) j=i j=i i t1 t 9 -j ] Z1 rij(£,а)ш3 (£,T - a) da where ti = —^i(z), i = 1, 2, 3, ti = t, i = 4, 5, 6. dτ е=м-1 [t+^i(z)] dT, i = 1,..., 6, (27) ?=^i 1[t+^i(z)] For the purpose of further research we introduce the following notation for the unknowns: u1(z,t)= Wi(z,t), i = 1,..., 9, "2(t) = r11(t), "2(t) = r12(t), "3(t) = r‘3 (t)- (28) "4 (t) = r2 2(t), "5(t) = r2 з(t), "6(t) = r3 3(t), ∂ "i (Z,t) = Sy^i Cz-t)- i = 4,5-6- (29) ui3(z, t) = ∂ dz^i(z,t) - r332r- (^ - m=0 )) d^to- i = 1,9, (30) "i3(z, t) = ∂ dz '^z t - r23^ (ft(z0) - ?7(z0)) dzfo, i = 2, 7, (31) "i3(z, t) = ∂ дzШi(z,t) - r13^ (^CzO) - n8(=0)) дZt0- i = 3, 8. (32) Taking into account these notations and the explicit forms of the functions rij(z,t) in terms of rij (t) by the formula (11), we rewrite the equations (24) in the form "H^t) t9 = U01(z,t)+ / ^Pij(4)"j(4,T) ti0 j =1 - (n - ПО)] dτ ξ=µi-1[τ -t+µi(z)] - / j U22a) ("1 - "9) (4'T " “) Д“ ti0 0 dT, i = 1, 9, ξ=µi-1[τ -t+µi(z)] Ui(z,t) t9 = u01(z,t)+ / ^Pij(4)u1(4,T) ti0 j =1 - "2(т) (n - ^7) (4)] dτ ξ=µi-1[τ-t+µi(z)] - j j ^ ("2 - "7) (4,t - a) da ti0 0 t "H^t) 01 υi (z,t)+ / ^Pij(Ouj(C,T) i j=1 - i t0 - dT, i = 2, 7, ξ=µi-1[τ -t+µi(z)] ^ (й - ft) (4) j j ^ (u2 - u2) (4,t - a) da ti0 0 dτ ξ=µi-1[τ-t+µi(z)] dT, i = 3, 8, ξ=µi-1[τ-t+µi(z)] t9 tτ Ui(z,t) = j ^^ p4j(z')uj(z,T) dT + У У u2(a)u4 (z,T - a) da dT 0j=1 0 0 tτ + У j (u2 - U2) (a) VJ( (u1 - U9) (z, T - a) + u6(z, t - a)^ da dT t +/ 0 υ12 - "4)(т) (^ ($i - $9)(z) +^(z)) + "2(т)П4(z) dT, t9 tτt u1(z,t) = j ^ P5j (z)uj (z,T) dT + j j u2(a)u1 (z,t — a) da dT + ju2(t)hi (z) dT, 0 j=1 0 00 tτ u1(z,t) = j j ^7 ^u2 - Ue) (a) (u1 — u1) (z,t — a) + u2(a)u6(z,T — a)] dadT 00 p t9 + / ^P6j(z^Uj(z,T)dT + j [V- (U4 — Ue) (t) (hi — $9)(z) + U2(T) where u0i(z,t) = Wi(z0, t0), i = 1, 2, 3, 7, 8, 9. Consider (27) the initial conditions (20), we differentiate (27) with respect to z for i = 1,2, 3 and for t for i = 4, 5, 6. After simple calculations, taking into account (28)-(32), we pass to integral equations tt u2 (t) = U02(t) — —Mi [ U2(T) (^hi — d-дЛ (t — t) dT — Mi [ u.(t)-d^(t — t) dT νp dt dt dt —Mi / Г7" (U2— U4) (t) f ;dhi—;dh9)(t — T)—(U2— U4)(T) ;dhe(t — T)1dT νp dt dt dt t +AMi j u2 (t ) dz (hi—<$9) (€) ∂9 dT + 2AMi — $Г.Р1з(£ )Wj (^,t ) ξ=µ1-1[t-τ] ∂z j=1 tτ +AMi /1u2(a) [(U3— U9)(^ t — a) — U6(t) (<hi —^M^] da dτ ξ=µ1-1[t-τ] dτ ξ=µ1-1[t-τ] —Mi / [7" (u2(t) — u6(t)) f7+hi —779*^(t — T) + u4(t)7+h6(t — T)1 νp dt dt dt dτ, u2 (t) = .. (t) — M2 t U32(T) ddh5(t — T) dT, 0 - t U3(t) = U32(t) —M4 У u2(t) (ddth3 - d dtg8 (t — T) dT tτ +M31 /U3(a)f(U3— U8) (^,t — a) — U3(t) ( ξ=µ2-1[t-τ] dτ t M3 jU.1T)dZ ($3 — й)(5) ∂9 dT + 2M3 — V.P33 (£ )uj(7T) ξ=µ2-1[t-τ] ∂z j=1 dτ, ξ=µ2-1[t-τ] t u2(t) = u42(t) - —M5 [ и2(т) (-dhi - d-gA (t - т) dT+ νp dt dt tτ + XM5 / 1 U2(a) ^3 - ^ (^ T - a) - u2(t) (—. -£9) (£)] da dτ ξ=µ1-1[t-τ] +XM,Ju2(T)dZ (—-*)(£) ξ=µ1-1[t-τ] t∂9 dT +2AM5 / — У pij(£)u1(C,T) 0z j=1 dτ ξ=µ1-1[t-τ] -M5 / [vp (u4- 6 (T) (sh'(t - T)- 9 t -T)) + U2(T) dd h6 (t - t)j dT, u2(t) = и0 (t) + M6 У u2 (t ) dZ(—2 -£7) (£) dτ ξ=µ2-1[t-τ] tτ +M61 /u2(a) [(u3- u7^(^ T - a) - u2 (t) (<2-7) (€)] da dτ ξ=µ2-1[t-τ] tt - M7 [ u2(t) (hh- - У g7\ (t - T) dT + 2M6 [ ^yj p2j(p)uj (^,T) dt dt∂z 0 0 u2(t) = u62(t) +M8 ju6 (t) dZ (i<1 - (?9) (£) dτ ξ=µ1-1[t-τ] dτ, ξ=µ2-1[t-τ] - tτ +M8 1 / U2(a) [(U3 t - u3) (^,T - a) - u6(t) (i?1-<9)(£)] da ξ= dτ µ1-1[t-τ ] M9 / u6 (T) (dthi 0 - dй dtg9 ∂9 j (t-т)dT + 2M8 j y^Lpij(^)и(^,T) 0 j=1 dτ, ξ=µ1-1[t-τ] u02(t) = MiQ1(t), u02(t) = M2Q2(t), U302(t) = -2M3P3(t), u42(t) = M5Q4(t), u02(t) = -2M6P2 (t), u32(t) = -2M8 Pi (t), Pi (t) = 1Г2 ^i (t1) - d^Z фi(0) - ViB У ^ (°)^j (°’t)’ and Q11 d2 dt2 h4(t) .—й - У P4j (°)^j (°,t) - M5[ j=1 d2 dt2 h6 (t) .—— - У P63(°b(°,t) - 2M8Pi(zA , j=1 29 29 Q2 = ^2Mt) - E—5j(°)^j(°,t), Q4 = ^2Mt) - E—6j(°)^j(°,t) - 2M8Pi(z), dt dt j=1 j=1 M2 ^5(0)’ M1 A (71(0) - £9(0))+^£б(0) + vp$4(0)’ 3 £3(0) - w(0)’ M4 V3(0) - 78(0) ’ M6 V2(0) - .(0)’ 5 νp A (vM0) — V9(0)) + vpV6(0) ’ £2(0) - £7(0) ’ M8 71(0) - V9(0) ’ M9 n(0) - .9(0) ■ The equation (39)-(44) contains unknown functions ^Z, j = 1’... ’ 9. For them we will receive integral equations from (24) by differentiating them with respect to the variable z. Using the notation (28)–(32), we obtain the integral equations for them t u3(z’t) = u03 (z’t)+ /ddZ ^ p'ij (4)U1 (^’T) ,i Lj=1 - ^ (71 - 79) (4) i t0 ti +^ / - 0 ’t0- T) dT tτ - / / U62a)dZ u -u9) (4’T - a) da to 0 u3(z’t) = u^Cz’t)+ / d|; [ ^ Pij(4 )u1 (4’t ) -+i Lj=i t0 dτ, ?=^i 1[t-t+Mi(z)] ЕэИ (7 - Vr)«)j i t0 + IZt0 / ^u1 - u7 ^ (z0 ’ t0- T)dT tτ - / / и52т) dz ^U1- u1) (4’T - a) da ti0 0 dτ, §=^-1[t-t+^i(z)] t ui3(z’t) = u03 (Z’t)+ fib ^ 7ij (4)uj «’T)-ei +i Lj-=1 ЕзИ (73 - 78) (4) i t0 ti +ddZt0 J ii 0’ t0 - T) dT dτ §=^-1 [t-t+W(z)] i = 1’ 9’ dτ e=^i1[t-t+^i(z)] i = 2’ 7’ dτ §=Ц-1[т-t+^i(z)] tτ - ti0 0 из(а) Э 2 dz (u3- u8) (4’t - a) da §=^i 1[t-t+^i(z)] dT’ i = 3’ 8’ t9 ∂ tτ ∂ u3(z,t) = ^ Ё d- [p4j (-)uj1 (z,T)] dT + j j U2(a) —u4(-,T - a) dadT 0 j=1 0 0 tτ + j j (u>2.(a) - u4(a)) d- 00 t d- [ip M - u9 ’ (-T - a) + u6(z,T — a) dadT + /[(u2(t) - u2(T)) d-^ (й 0 -<9)(-) +<6(-)) + u2(T)д|-<4(-) dT, (48) t9 ∂ u3 = / Ё d- ^T^ dT+ 0 j=1 tτ t j j u1(a) d-u1(-,t - a) dadT + j и2(т) ^-<1 (-) dT, (49) t9 ∂ tτ ∂ u6(-,t)= j^ d— [p6j (z)uj (-,t)] dT + j y‘u2(a) —u6(-,T - a) dadT 0 j=1 0 0 + tτ j j (u4(a) - u6(a)) d- vA (ui - u9) (z 00 p T- a) dα dτ t + / (u2 - u2) (t)— №(-) - (?9(-)) + u2(T)77" <6(-) dT, (50) ∂z νp ∂z where ∂ U03(-,t) = ^i(-° ’to) - d-to ^ pij(-9) шз(-o ,to)j, i = 1, 2, 3, 7, 8, 9. A [<6i(0) - <69(0)]+ vp<66(0) + vp<64(0) = 0, <65(0) =0, <6з(0) - <68(0) = 0, (51) A [<i(0) - <9(0)] + Vp<6(0) = 0, <2(0) - <7(0) = 0, <1 (0) - <9(0) = 0. (52) We require the fulfillment of the matching conditions d<6i(-) j ∂z + Ё P.j (0) , t=0 i = 1,..., 6. 3. Main Result The main result of this work is the following theorem: Theorem 2. Let the conditions of Theorem 1 are satisfied, besides function h(xi; X2; t) have compact support in xi, X2 for each fixed t, <6i(-) € C2[0, H] , i = 1,..., 9, gt(t) 6 C2[0, H] , i = 1, 2, 3, 7, 8, 9, hi(t) € C2[0, H] , i = 1,..., 6, equalities (51), (52) and matching conditions (25), (26), (53) hold. Then for any H > 0 on the segment [0, H] there is a unique solution to the inverse problems (15)-(18) from class rij(t) € C [0,H] ,i,j = 1, 2, 3. <1 Equations (33)-(50) supplemented by the initial and boundary conditions from the equalities (19) forms a closed system of equations for the unknowns wi(z,t), i = 1,..., 9, ri’j (t), i,j = 1, 2, 3, dZwi (z,t), i = 1,..., 9. Consider now a square Dq := {(z, t) : 0 C z C H, 0 C t C H} . Then, these equations show that the values of the functions Wi(z,t), i = 1,..., 9, r’j(t), i,j = 1, 2, 3, ddZWi(z,t), i = 1,..., 9 at (z,t) € Dq are expressed through integrals of some combinations of the same functions over segments lying in Dq. The system of equations (33)–(50) we rewrite in the operator form и = Au, where the operator A = (Ai^A?, A3), i = 1,..., 9, j = 1,..., 6, in accordance with the right-hand sides follow equations (33)–(50). Let Cs(Dq), (s ^ 0) be the Banach space of continuous functions induced by the family at weighted norms HJIs , ||u^s= max < max sup |u1(z,t)e (1CiC9(z,tWo st I , max sup H(t)e 1CiC6tE[Q,H] st1, max sup |u3(z,t)e st1>. ' Ki<9(z,t)GDo' ’* J Obviously, Cs with s = 0 is the usual space of continuous functions with the ordinary norm, denoted by ||-|| in what follows. Because e-sH|u| C |u|sC |u|, the norms |u|sand |u| are equivalent for any H € (0, to). We choose that number s later. Next, consider the set of functions S(uQ,r) C Cs(Dq), satisfying the inequality ||u— и0^ Cr, where r is a known number, the vector function uQ(z,t) = (uQ1(z,t), i = 1,...,9, uQ(t), i = 1,...,6, uQ3Cz,t), i = 1,...,9), defined by the free terms of the operator equation (54). It is easy to see that for u € S(uQ, r) the estimate |u|s C |uQ|s + r C |uQ|| + r := rQ. Thus, rQ is a known number. Let us introduce the following notation: hQ := iss Oh L -юн}1 ^Q := i^mlaxg {^ lc2tQ’H 1} , go := i=ima/7,8,9 {^ ^2[Q,H 1} , MQ= max{llMi(z)llc[Q,H]} , M0 = • .max ^WpijИ^н]} , 1,...,9x. i,i 1«...«9 i 3S , lA(z)lc 1[Q^P l dl U’MQ=max { Vp(Z) c[q H] L J ^z L J I Note that the operator A maps the space Cs(Dq) into itself. Let us show that for a suitable choice of s (recall that H > 0 — is an arbitrary fixed number) it is on the set S(uQ, r) a contraction operator. First, let us make sure that the operator A takes the set S(uQ, r) into itself, that is, it follows from the condition u(z,t) € S(uQ, r) that Au € S(uQ, r), if s satisfies some constraints. In fact, for any (z,t) E Do and и E S(u0,r) the following inequalities hold: ||AiU - u01|s= sup | (Aiи — u01) e st| (z,t)∈D0 t V [ ^ Pij(^j (^,T) / 7 = 1 ij e ST+ EeHe i t0 t τ sτ - (V1 ^9)(€) e s(t τ) dτ ξ=µi-1[t-τ+µi(z)] + V^]! sα — (u1 — u9) (£,r — a)e s(t a)da i t0 dτ ξ=µi-1[t-τ +µi(z)] < (9M0M. + «,||* + M.M.) /■)dT < rs (9M0+ vo + ro) := 211 ro. t10 Using similar calculations for the remaining equations. Finally, we get ||Au - u0||s ^ 1 max J max {Yij} , max {Y2j} , max {Y3j} f := r070, s j =1,...,9 j=1,...,6 j=1,...,9 s where 71j := 9M0+ vo + ro, j = 2, 3, 7, 8, 9, 714 := 13M0+ 4M0ro + 3Wo + 3^0, 715 := 9M0+ vo + ro, 716 := 9M0+ 9Moro + 4M0vo + vo, 721 := 2ro + 2voro + 5 (M0)2 (go + ho) + 4ho + 9 (M0)3 , 722 := M%, 724 := (M0) (3go + 3ho + 2Vo + 18), Y2j := 2M0 (ro(1 + Vo) + Vo + ho + 9), j = 3,5,6, 73j := 11M0+ vo + ro, j = 1, 2, 3, 7, 8, 9, 734 := 13M0+ 4M0ro + 3vo(ro + 1), 735 := 9M0+ ro + vo, 736 := 9M0+ 9M0ro + Vo(4M0+ 1). Choosing s > (1/r)Y0, we get that the operator A maps the set S (u0, r) into itself. Now, let u and v be two arbitrary elements in S(u0,r). Using the obvious inequality |ukui — uku| |e-st ^ |uk — Vk||ul|e-st + |vk| |u| — V||e-st ^ 2ro^u — v^s, (z,t) E Do, after some easy estimations, we find that for (z,t) E Do, ^A1u — A1v||s = sup |(A1u — A1v) e-st| < ——— [9M0 + vo + 4ro] := 1741 ||u — v^s s (z,t)∈D0 s s and hence we have |Au — Au|s = |u — VL г r . r . r a 171B« — SUs , -------s max < max {74 Л , max {75 Л , max {76 Л > := s j=1,...,9 j j=1,...,6 j j=1,...,9 j where 74j := 9M0+ vo + 4ro, j = 2, 3, 7, 8, 9, 744 := 13M0+ 4Moro + 12Wo + 3vo, 745 := 9M0+ vo + 4ro, 746 := 9M0+ 9M% + 4M0vo + 4vo, Y51 := 2ro + 4^oro + 10 (M0)2 (go + he) + 10h + 9 (M0)3 , Y52 := MX, Y54 : = (M0) (12g0 + 3h0 + 10^0 + 18) , Y5j : = M0 (r0(1 + ^0) + ^0 + 4h0 + 9) , j = 3 5, 6 Y6j := 11M0+ ^0 + ro, j = 1, 2, 3, 7, 8, 9, Y64 := 13M0+ 4M0r, + 3^o(ro + 1), Y65 := 9M0 + Г0 + ^0, Y66 := 9M0 + 9M% + ^(4M0 + 1). Choosing now s > γ1, we get, that the operator A compresses the distance between the elements u,U to S (u0,r) . As follows from the performed estimates, if the number s is chosen from conditions s > s* := max{Y0, Y1}, then the operator A is contracting on S (u0, r) . In this case, according to the principle Banach the equation [25, pp. 87–97] (54) has the only solution in S υ0, r for any fixed H > 0. Theorem 2 is proved. > By the found functions r’j(t), i, j = 1, 2, 3, the functions rij(t), i, j = 1, 2, 3, are found by the formulas t rij (t) = rij (0)+ rij (t ) dT, i,j = 1, 2, 3. Note that by the functions rij(t), i,j = 1, 2, 3, the functions Kij(t), i,j = 1, 2, 3, are defined as solutions of integral equations (4).
Список литературы Inverse problem for viscoelastic system in a vertically layered medium
- Mura, T. Micromechanics of Defects in Solids, Second, Revised Edition, USA, IL, Evanston, Northwestern University, 1987.
- Galin, L. A. Contact Problems of the Theory of Elasticity and Viscoelasticity, Moscow, Nauka, 1980 (in Russian).
- Kilbas, A. А Integral Equations: Course of Lectures, Minsk, Belarusian State University, 2005 (in Russian).
- Durdimurod, D., Shishkina, E. and Sitnik, S. The Explicit Formula for Solution of Anomalous Diffusion Equation in the Multi-Dimensional Space, Lobachevskii Journal of Mathematics, 2021, vol. 42, no. 6, pp. 1264–1273. DOI: 10.1134/S199508022106007X.
- Godunov, S. K. Equations of Mathematical Physics, Moscow, Nauka, Ch. Ed. Physical-Mat. Lit., 1979 (in Russian).
- Romanov, V. G. Inverse Problems of Mathematical Physics, Utrecht, The Netherlands, 1987.
- Lorenzi, A. An Identification Problem Related to a Nonlinear Hyperbolic Integro-Differential Equation, Nonlinear Analysis, Theory, Methods and Applications, 1994, vol. 22, no. 1, pp. 21–44. DOI: 10.1016/0362-546X(94)90003-5.
- Janno, J. and Von Wolfersdorf, L. Inverse Problems for Identification of Memory Kernels in Viscoelasticity, Mathematical Methods in the Applied Sciences, 1997, vol. 20, no. 4, pp. 291–314. DOI: 10.1002/(SICI)1099-1476(19970310)20:4<291::AID-MMA860-3.0.CO;2-W.
- Romanov, V. G. Stability Estimates for the Solution in the Problem of Determining the Kernel of the Viscoelasticity Equation, Journal of Applied and Industrial Mathematics, 2012, vol. 6, no. 3, pp. 360–370. DOI: 10.1134/S1990478912030118.
- Totieva, Zh. D. and Durdiev, D. Q. The Problem of Determining the Multidimensional Kernel of Viscoelasticity Equation, Vladikavkaz Mathematical Journal, 2015, vol. 17, no. 4, pp. 18–43. DOI: 10.23671/VNC.2015.4.5969.
- Durdiev, D. K. Some Multidimensional Inverse Problems of Memory Determination in Hyperbolic Equations, Zhurnal Matematicheskoi Fiziki, Analiza, Geometrii [Journal of Mathematical Physics, Analysis, Geometry], 2007, vol. 3, no. 4, pp. 411–423.
- Durdiev, D. K. and Safarov, Z. S. Inverse Problem of Determining the One-Dimensional Kernel of the Viscoelasticity Equation in a Bounded Domain, Mathematical Notes, 2015, vol. 97, no. 6, pp. 867–877. DOI: 10.1134/S0001434615050223.
- Romanov, V. G. On the Determination of the Coefficients in the Viscoelasticity Equations, Siberian Mathematical Journal, 2014, Vol. 55, no. 3, pp. 503–510. DOI: 10.1134/S0037446614030124.
- Romanov, V. G. The Problem of Determining the Kernel of the Viscoelasticity Equation, Doklady Akademii Nauk, 2012, vol. 446, no. 1, pp. 18–20 (in Russian).
- Durdiev, D. K. and Rakhmonov, A. A. Inverse Problem for the System Integro-Differential Equation SH Waves in a Visco-Elastic Porous Medium: Global Solvability, Theoretical and Mathematical Physics, 2018, vol. 195, no. 3, pp. 923–937, DOI: 10.1134/S0040577918060090.
- Durdiev, D. K. and Rakhmonov, A. A. The Problem of Determining Two-Dimensional Kernel in a System of Integro-Differential Equations of a Viscoelastic Porous Medium, Journal of Applied and Industrial Mathematics, 2020, vol. 14, no. 2, pp. 281–295. DOI: 10.1134/S1990478920020076.
- Durdiev, D. K. and Rahmonov, A. A. A 2D Kernel Determination Problem in a Viscoelastic Porous Medium with a Weakly Horizontally Inhomogeneity, Mathematical Methods in the Applied Sciences, 2020, vol. 43, no. 15, pp. 8776–8796. DOI: 10.1002/mma.6544.
- Durdiev, D. K. and Totieva, Z. D. The Problem of Determining the One-Dimensional Matrix Kernel of the System of Viscoelasticity Equations, Mathematical Methods in the Applied Sciences, 2018, vol. 41, no. 17, pp. 8019–8032. DOI: 10.1002/mma.5267.
- Totieva, Z. D. and Durdiev, D. K. The Problem of Finding the One-Dimensional Kernel of the Thermoviscoelasticity Equation, Mathematical Notes, 2018, vol. 103, no. 1–2, pp. 118–132. DOI: 10.1134/S0001434618010145.
- Safarov, J. SH. and Durdiev, D. K. Inverse Problem for an Integro-Differential Equation of Acoustics, Differential Equations, 2018, vol. 54, no. 1, pp. 134–142. DOI: 10.1134/S0012266118010111.
- Durdiev, D. K. and Totieva, Z. D. The Problem of Determining the One-Dimensional Kernel of Viscoelasticity Equation with a Source of Explosive Type, Journal of Inverse and Ill-Posed Problems, 2020, vol. 28, no. 1, pp. 43–52. DOI: 10.1515/jiip-2018-0024.
- Durdiev, U. D. An Inverse Problem for the System of Viscoelasticity Equations in Homogeneous Anisotropic Media, Journal of Applied and Industrial Mathematics, 2019, vol. 13, no. 4, pp. 623–628. DOI: 10.1134/S1990478919040057.
- Durdiev, D. K. and Turdiev, Kh. Kh. Inverse Problem for a First-Order Hyperbolic System with Memory, Differential Equations, 2020, vol. 56, no. 12, pp. 1634–1643. DOI: 10.1134/S00122661200120125.
- Durdiev, D. K. and Turdiev, Kh. Kh. The Problem of Finding the Kernels in the System of Integro-Differential Maxwell’s Equations, Journal of Applied and Industrial Mathematics, 2021, vol. 15, no. 2, pp. 190–211. DOI: 10.1134/S1990478921020022.
- Kolmogorov, A. N. and Fomin, S. V. Elements of Function Theory and Functional Analysis, Moskow, Nauka, 1989 (in Russian).