Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation
The approach to the solving of gradient problems of parameters identification of quasiideal fields with using applied quasipotential tomographic data based on numerical complex analysis methods is transferred to cases of anisotropic media. We, similar to the existing works of world scientists, some...
Gespeichert in:
| Veröffentlicht in: | Математичне та комп'ютерне моделювання. Серія: Фізико-математичні науки |
|---|---|
| Datum: | 2018 |
| Hauptverfasser: | , |
| Format: | Artikel |
| Sprache: | English |
| Veröffentlicht: |
Інститут кібернетики ім. В.М. Глушкова НАН України
2018
|
| Online Zugang: | https://nasplib.isofts.kiev.ua/handle/123456789/162214 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Zitieren: | Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation / A.Ya. Bomba, M.V. Boichura // Математичне та комп'ютерне моделювання. Серія: Фізико-математичні науки: зб. наук. пр. — Кам’янець-Подільський: Кам'янець-Подільськ. нац. ун-т, 2018. — Вип. 18. — С. 14-24. — Бібліогр.: 16 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
nasplib_isofts_kiev_ua-123456789-162214 |
|---|---|
| record_format |
dspace |
| spelling |
Bomba, A.Ya. Boichura, M.V. 2020-01-04T19:38:06Z 2020-01-04T19:38:06Z 2018 Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation / A.Ya. Bomba, M.V. Boichura // Математичне та комп'ютерне моделювання. Серія: Фізико-математичні науки: зб. наук. пр. — Кам’янець-Подільський: Кам'янець-Подільськ. нац. ун-т, 2018. — Вип. 18. — С. 14-24. — Бібліогр.: 16 назв. — англ. DOI: 10.32626/2308-5878.2018-18.14-24 https://nasplib.isofts.kiev.ua/handle/123456789/162214 519.6 The approach to the solving of gradient problems of parameters identification of quasiideal fields with using applied quasipotential tomographic data based on numerical complex analysis methods is transferred to cases of anisotropic media. We, similar to the existing works of world scientists, some additional information about the nature of the distribution of conductivity inside the domain (research object) is considered a priori known. However, in opposite to the traditional approaches to the statement and solving the problems of electrical impedance tomography, we set the local velocities distribution of a substance (liquid, current) in addition to the averaged potential at the contact sections of plate and body and at other sections (stream lines), we set the potential distribution (according to experimental data, which we approximate using splines, Bezier curves, etc.). Generation of initial data at the boundary of the investigated object is carried out in accordance with the polar model of current injection and a given sum of eigenvalues of the conductivity tensor of the medium. The presence of this kind of data greatly accelerates the process of further solving the problem, which is convenient, in particular, when verifying the method that developed by authors. The corresponding problem is reduced to the iterative solving of a series of problems for the Laplace type equations, where instead of «boundary numerical analogues of the Cauchy-Riemann type equations » appear the ratio of quasiorthogonality with using special types of optimization conditions. In particular: the minimizing functional is constructed by taking into account the Cauchy-Riemann type conditions, the relation between eigenvalues of corresponding anisotropy tensor and also regularizing term; the condition-restriction is built based on ellipticity conditions. Підхід до розв’язання градієнтних задач ідентифікації параметрів квазіідеальних полів за даними томографії прикладених квазіпотенціалів на основі числових методів комплексного аналізу перенесено на випадки анізотропних середовищ. При цьому, аналогічно до існуючих робіт світових вчених, апріорно відомими вважаються деякі додаткові відомості про характер розподілу провідності всередині області (об’єкта дослідження). Проте, на відміну від традиційних підходів до постановки та розв’язання задач електроімпедансної томографії, на ділянках контакту пластинки і тіла окрім усередненого потенціалу тут задається ще й розподіл локальних швидкостей речовини (рідини, струму), а на інших ділянках (лініях течії) — розподіл потенціалу (за експериментальними даними, які апроксимуємо із застосуванням сплайнів, кривих Безьє тощо). Генерація вихідних даних на межі досліджуваного об’єкта здійснюється відповідно до полярної моделі (схеми) інжекції струму при заданій сумі власних значень тензора провідності середовища. Наявність такого виду даних значно пришвидшує процес подальшого розв’язання поставленої задачі, що зручно, зокрема, при верифікації розробленого авторами методу. Відповідна задача зводиться до ітераційного розв’язання серії задач для рівнянь типу Лапласа, де замість «приграничних числових аналогів рівнянь типу Коші-Рімана» фігурують співвідношення квазіортогональності за спеціальних типів умов оптимізації. А саме: мінімізуючий функціонал побудований з урахуванням умов типу Коші-Рімана, співвідношення між відповідними тензору анізотропії власними значеннями, а також регуляризуючого доданку; умови-обмеження сформовані на основі умов еліптичності. en Інститут кібернетики ім. В.М. Глушкова НАН України Математичне та комп'ютерне моделювання. Серія: Фізико-математичні науки Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation Числовий метод комплексного аналізу ідентифікації параметрів анізотропних середовищ за даними томографії прикладених квазіпотенціалів. Частина 1: Постановка та наближення задачі Article published earlier |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| title |
Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation |
| spellingShingle |
Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation Bomba, A.Ya. Boichura, M.V. |
| title_short |
Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation |
| title_full |
Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation |
| title_fullStr |
Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation |
| title_full_unstemmed |
Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation |
| title_sort |
numerical complex analysis method for parameters identification of anisotropic media using applied quasipotential tomographic data. part 1: problem statement and its approximation |
| author |
Bomba, A.Ya. Boichura, M.V. |
| author_facet |
Bomba, A.Ya. Boichura, M.V. |
| publishDate |
2018 |
| language |
English |
| container_title |
Математичне та комп'ютерне моделювання. Серія: Фізико-математичні науки |
| publisher |
Інститут кібернетики ім. В.М. Глушкова НАН України |
| format |
Article |
| title_alt |
Числовий метод комплексного аналізу ідентифікації параметрів анізотропних середовищ за даними томографії прикладених квазіпотенціалів. Частина 1: Постановка та наближення задачі |
| description |
The approach to the solving of gradient problems of parameters identification of quasiideal fields with using applied quasipotential tomographic data based on numerical complex analysis methods is transferred to cases of anisotropic media. We, similar to the existing works of world scientists, some additional information about the nature of the distribution of conductivity inside the domain (research object) is considered a priori known. However, in opposite to the traditional approaches to the statement and solving the problems of electrical impedance tomography, we set the local velocities distribution of a substance (liquid, current) in addition to the averaged potential at the contact sections of plate and body and at other sections (stream lines), we set the potential distribution (according to experimental data, which we approximate using splines, Bezier curves, etc.). Generation of initial data at the boundary of the investigated object is carried out in accordance with the polar model of current injection and a given sum of eigenvalues of the conductivity tensor of the medium. The presence of this kind of data greatly accelerates the process of further solving the problem, which is convenient, in particular, when verifying the method that developed by authors. The corresponding problem is reduced to the iterative solving of a series of problems for the Laplace type equations, where instead of «boundary numerical analogues of the Cauchy-Riemann type equations » appear the ratio of quasiorthogonality with using special types of optimization conditions. In particular: the minimizing functional is constructed by taking into account the Cauchy-Riemann type conditions, the relation between eigenvalues of corresponding anisotropy tensor and also regularizing term; the condition-restriction is built based on ellipticity conditions.
Підхід до розв’язання градієнтних задач ідентифікації параметрів квазіідеальних полів за даними томографії прикладених квазіпотенціалів на основі числових методів комплексного аналізу перенесено на випадки анізотропних середовищ. При цьому, аналогічно до існуючих робіт світових вчених, апріорно відомими вважаються деякі додаткові відомості про характер розподілу провідності всередині області (об’єкта дослідження). Проте, на відміну від традиційних підходів до постановки та розв’язання задач електроімпедансної томографії, на ділянках контакту пластинки і тіла окрім усередненого потенціалу тут задається ще й розподіл локальних швидкостей речовини (рідини, струму), а на інших ділянках (лініях течії) — розподіл потенціалу (за експериментальними даними, які апроксимуємо із застосуванням сплайнів, кривих Безьє тощо). Генерація вихідних даних на межі досліджуваного об’єкта здійснюється відповідно до полярної моделі (схеми) інжекції струму при заданій сумі власних значень тензора провідності середовища. Наявність такого виду даних значно пришвидшує процес подальшого розв’язання поставленої задачі, що зручно, зокрема, при верифікації розробленого авторами методу. Відповідна задача зводиться до ітераційного розв’язання серії задач для рівнянь типу Лапласа, де замість «приграничних числових аналогів рівнянь типу Коші-Рімана» фігурують співвідношення квазіортогональності за спеціальних типів умов оптимізації. А саме: мінімізуючий функціонал побудований з урахуванням умов типу Коші-Рімана, співвідношення між відповідними тензору анізотропії власними значеннями, а також регуляризуючого доданку; умови-обмеження сформовані на основі умов еліптичності.
|
| url |
https://nasplib.isofts.kiev.ua/handle/123456789/162214 |
| citation_txt |
Numerical Complex Analysis Method for Parameters Identification of Anisotropic Media Using Applied Quasipotential Tomographic Data. Part 1: Problem Statement and its Approximation / A.Ya. Bomba, M.V. Boichura // Математичне та комп'ютерне моделювання. Серія: Фізико-математичні науки: зб. наук. пр. — Кам’янець-Подільський: Кам'янець-Подільськ. нац. ун-т, 2018. — Вип. 18. — С. 14-24. — Бібліогр.: 16 назв. — англ. |
| work_keys_str_mv |
AT bombaaya numericalcomplexanalysismethodforparametersidentificationofanisotropicmediausingappliedquasipotentialtomographicdatapart1problemstatementanditsapproximation AT boichuramv numericalcomplexanalysismethodforparametersidentificationofanisotropicmediausingappliedquasipotentialtomographicdatapart1problemstatementanditsapproximation AT bombaaya čisloviimetodkompleksnogoanalízuídentifíkacííparametrívanízotropnihseredoviŝzadanimitomografííprikladenihkvazípotencíalívčastina1postanovkatanabližennâzadačí AT boichuramv čisloviimetodkompleksnogoanalízuídentifíkacííparametrívanízotropnihseredoviŝzadanimitomografííprikladenihkvazípotencíalívčastina1postanovkatanabližennâzadačí |
| first_indexed |
2025-11-25T22:13:39Z |
| last_indexed |
2025-11-25T22:13:39Z |
| _version_ |
1850560934782697472 |
| fulltext |
Математичне та комп’ютерне моделювання
14
bounded and self-adjoint in the space of all k-periodic sequences. Then a spe-
cial functional was constructed, the critical points of which in this space are so-
lutions of the original equation. A Gateaux derivative of this functional is
found. Next we consider the Nehari manifold for a given variational problem,
which is a set of nontrivial critical points of a constructed functional in the
space of k-periodic sequences. It is shown that this manifold is a non-empty
and closed subset of a given space. In addition, the corresponding minimization
problem is considered and it is shown that this problem has a solution in the
Nehari manifold. Consequently, under these conditions the original equation
has nontrivial periodic solutions. Finally, due to the fact that saturable nonline-
arity satisfies these conditions, the existence of two nontrivial standing waves
with k-periodic amplitude for a discrete nonlinear Schrödinger equation with
saturable nonlinearity on a two-dimensional lattice is established. The results of
this paper are the distribution of already known results for discrete nonlinear
Schrödinger equations on 1D and 2D-lattices.
Key words: discrete nonlinear Schrödinger equation, 2D-lattice,
standing waves, critical points, Nehari manifold, saturable nonlinearity.
Отримано: 14.11.2018
UDC 519.6
DOI: 10.32626/2308-5878.2018-18.14-24
A. Ya. Bomba, Doct. of Techn. Sciences,
M. V. Boichura
Rivne State Humanitarian University, Rivne
NUMERICAL COMPLEX ANALYSIS METHOD FOR
PARAMETERS IDENTIFICATION OF ANISOTROPIC MEDIA
USING APPLIED QUASIPOTENTIAL TOMOGRAPHIC DATA.
PART 1: PROBLEM STATEMENT AND ITS APPROXIMATION
The approach to the solving of gradient problems of parameters
identification of quasiideal fields with using applied quasipotential
tomographic data based on numerical complex analysis methods is
transferred to cases of anisotropic media. We, similar to the existing
works of world scientists, some additional information about the na-
ture of the distribution of conductivity inside the domain (research
object) is considered a priori known. However, in opposite to the tra-
ditional approaches to the statement and solving the problems of
electrical impedance tomography, we set the local velocities distribu-
tion of a substance (liquid, current) in addition to the averaged poten-
tial at the contact sections of plate and body and at other sections
(stream lines), we set the potential distribution (according to experi-
mental data, which we approximate using splines, Bezier curves,
etc.). Generation of initial data at the boundary of the investigated
object is carried out in accordance with the polar model of current in-
jection and a given sum of eigenvalues of the conductivity tensor of
© A. Ya. Bomba, M.V. Boichura, 2018
Серія: Фізико-математичні науки. Випуск 18
15
the medium. The presence of this kind of data greatly accelerates the
process of further solving the problem, which is convenient, in par-
ticular, when verifying the method that developed by authors. The
corresponding problem is reduced to the iterative solving of a series
of problems for the Laplace type equations, where instead of
«boundary numerical analogues of the Cauchy-Riemann type equa-
tions» appear the ratio of quasiorthogonality with using special types
of optimization conditions. In particular: the minimizing functional is
constructed by taking into account the Cauchy-Riemann type condi-
tions, the relation between eigenvalues of corresponding anisotropy
tensor and also regularizing term; the condition-restriction is built
based on ellipticity conditions.
Key words: applied quasipotential tomography, quasiconformal
mappings, anisotropy, identification, nonlinear problems.
Introduction. As it is known [1–5], the image reconstruction of an iso-
tropic conductivity coefficient that based on the applied quasipotential tomog-
raphy (AQT) requires the imposition of a large number of conditions at the
domain bound, and also the structure of the corresponding medium. It turns
out that in the general case (in contrast to some specific [6]) of anisotropy, it is
necessary to set much more information about the conductivity distribution for
its parameter identification [1, 7–12]. This, obviously, weakens the correctness
of the problem in comparison with the isotropy. And, consequently, it requires
(in comparison with, for example, [5]) the necessity of using a regularizing
functional, in particular the Tikhonov type [1, 3, 4, 9, 11]. The ways to apply
additional data about the conductivity tensor (CT), depending on the infor-
mation type about it are proposed in a number of papers [6–9, 12]. In particu-
lar, several options for specifying information about eigenvectors offer in [8].
The regularization procedure was adapted in such a way that the additional
data were taken into account as much as possible [9]. A new visual CT repre-
sentation, in which it is easy to understand its distribution, is given in [6].
On the other hand, today a promising methodology for identifying the
conductivity coefficient using AQT data, according to which the solving of the
sequence of so-called analysis and synthesis problems is reduced to the alter-
nate application of numerical quasiconformal mappings methods and the pa-
rameters identification of the conductivity of the medium, respectively [4, 5] is
developed. In this paper we discuss the transfer of this methodology to the
case of the parameter identification of anisotropic media. At the same time, the
additional information is the dependency between eigenvalues of CT.
In this case, the practical application of the results of this research can
take place in a number of branches of science and technology (see, for exam-
ple [1, 9–13]). In particular, in medicine, the object of this kind of research
may be the medium with fibrous or layered areas (which includes muscles,
bones, etc.), in which there is a stream of non-spherical particles (for example,
red blood cells), and in geology, with areas of soil layer compression.
Математичне та комп’ютерне моделювання
16
Parameters identification problem of anisotropic mediums using
AQT data. Consider quasiideal processes of particles movement (in particu-
lar, fluids, electrical currents) in a single-connected curvilinear domain (hori-
zontal anisotropic formation, anisotropic plate — the tomographical cross-
section) zG (Fig. 1, a), bounded by a smooth closed curve ( ) :zG x, y
( ),x x % ( ),y y % 0(0) (2 ) ,x x x % % % 0(0) (2 ) ,y y y % % % 0 2 ,
where ( ),x % ( )y % are defined continuosly differentiated functions,
0 0( , )O x y% % is given an initial starting point}, generated by the existence of
potential differences ( )
* ,p *( )p ( )*( )
*( ),pp which are given on select-
ed equipotential lines p pA B and ,p pC D where ,pA ,pB ,pC pD are
marked points on the ;zG 1, 2,...p is injection number (see for example
[3–5, 8, 11]); p pB C and p pA D are impermeable boundary flow lines. We
model a current injection through tomographic cross-section, similarly to [4,
5], by using sets of values ( ) ( ) ( ) ( ){ , , , },p p p p
B DA C corresponding to which
( ) ( )( ), ( ) ,p p
p A AA x y % % ( ) ( )( ), ( ) ,p p
p B BB x y % %
( ) ( )( ), ( ) ,p p
p C CC x y % % ( ) ( )( ), ( ) .p p
p D DD x y % %
We denote the corresponding for current injection bound of domain zG
with given four marked points by ( )p
zG ( ) ( ) ( )( ).p p pz x iy
a) b)
Fig. 1. Tomographical cross-section zG (a)
and corresponding complex quasipotential domains ( )pG (b)
As a mathematical model of AQT [1] we use, similarly to [4, 5, 14],
systems of differential equations in partial derivatives which connect mu-
tually quasicomplex conjugate quasipotentials ( ) ( , )p x y of density
( ) ( , )pgrad x y and stream functions ( ) ( ) ( , )p p x y when have place
Серія: Фізико-математичні науки. Випуск 18
17
CT , 1,2
( , ,...)x y
identification under corresponding boundary
conditions:
( ) ( ) ( ) ( )
11 12 21 22( ) ( ) 0;p p p p
x y x x y y (1)
( ) ( )
* ,
p p
p p
A B ( ) *( ) ,
p p
p p
C D ( ) 0;
p p p p
p
B C A Dj
ρ
(2)
( )( ) ( ) ( ),
p p
pp
B CM M ( )( ) ( ) ( ),
p p
pp
A DM M
( )( )
*( ) ( ),
p p
pp
A Bj M M
ρ
*( )( ) ( ) ( );
p p
pp
C Dj M M
ρ
(3)
11 22 1 2 . (4)
Here (1) is a consequence of the linear law of motion such as Ohm, Darcy
etc. ( ) ( )p pj grad
ρ
and the equation of continuity ( ) 0pdiv j
ρ
[1–8,
10, 13]; ( , ,...)x y are limited continuously differentiated in do-
main ( )p
zG functions, which characterize conductivity and anisotropy of
medium; ( , )x y is given distribution of the sum of eigenvalues 1,
2 of corresponding to (1) matrix, that equivalent to the sum of diagonal
elements of given tensor [14]; nρ is unit vector of outer normal; M is a
running point of the corresponding curve. The functions
( ) ( )( ) ( ,...)p pM ( ) ( )( ),p p
BC ( ) ( )( ) ( ,...)p pM
( ) ( )( ),p p
DA ( ) ( )
* *( ) ( ,...)p pM ( ) ( )( ),p p
B A
*( ) *( )( ) ( ,...)p pM ( ) ( )( )p p
D C
can be built by the interpolation of the experimentally received their values
( )
( ) ,p
p
i ( )
( ) ,p
p
i ( )
*
( )
* ,p
p
j *( )
*( )
p
p
j having some ( )
( ) ,p
p
i ( )
( ) ,p
p
i ( )
*
( )
* ,p
p
j
*( )
*( )
p
p
j on sections ,p pB C ,p pA D ,p pA B ,p pC D respectively
( )
( ) ( ) *( )
*( ,p
p p p
i ( )
( ) ( ) *( )
* ,p
p p p
i ( )
*
( )
* 0,p
p
j
*( )
*( ) 0,p
p
j ( ) *( )0 1,p pi m ( )( )
*0 1,ppi m
( ) ( )
* *0 1,p pj n *( ) *( )0 1)p pj n
as in [5]. We search the CT components as:
,
( ) ( )( ) ( )
11 11 ,0 0, ,
, 0
( , , , ..., ) ,
a a
a a a
a a a a a
a a
s k
p k r p rp p
s s k r r
k r
x y a a a x y
,
( ) ( )( ) ( )
22 22 ,0 0, ,
, 0
( , , , ..., ) ,
b b
b b b
b b b b b
b b
s k
p k r p rp p
s s k r r
k r
x y b b b x y
Математичне та комп’ютерне моделювання
18
,
( ) ( )( ) ( )
12 12 ,0 0, ,
, 0
( , , ,..., ) ,
c c
c c c
c c c c c
c c
s k
p k r p rp p
s s k r r
k r
x y c c c x y
21 12 , (5)
where , ,
a a ak r ra , ,
b b bk r rb ,c c ck r rc ( 0,..., ,a ak s 0,..., ,a ar k
0,..., ,b bk s 0,..., ,b br k 0,..., ,c ck s 0,..., )c cr k are the parameters
that are being found during the problem solving.
The problem lies in image reconstructing of CT. In this, the accompany-
ing is calculating of corresponding electrodynamic meshes and velocity fields.
We can reduce (1)–(5) to the series of more general boundary value prob-
lems on quasiconformal mapping ( ) ( ) ( )( ) ( , ) ( , )p p pz x y i x y of the
physical domains ( )p
zG (Fig. 1, a) onto the corresponding domains of the com-
plex quasipotential ( )pG (Fig. 1, b) by the way, similarly to [4, 5, 14], of intro-
ducing the stream functions ( ) ( ) ( , )p p x y ( 1, ),p p % which are complex
conjugated to ( ) ( ) ( , )p p x y ( 1, ),p p % under (4) and (5) conditions:
( ) ( ) ( ) ( ) ( ) ( )
11 12 21 22, ;p p p p p p
x y y x x x (6)
( ) ( )
* ,
p p
p p
A B ( ) *( ) ,
p p
p p
C D
( ) 0,
p p
p
A D
( ) ( ) ;
p p
p p
B C
Q (7)
( ) ( ) , , ;p p
p p p p
MN
j dl Q M B C N A D
ρ
( )( ) ( ) ( ),
p p
pp
B CM M ( )( ) ( ) ( ),
p p
pp
A DM M
( )( )
*( ) ( ),
p p
pp
A BM M *( )( ) ( ) ( ),
p p
pp
C DM M (8)
where
( ) ( )
* *( ) ( ) ,
p
p p
A M
M M dl *( ) *( )( ) ( ) ;
p
p p
D M
M M dl
( )( ) *( ) ( )
*( , ): ,0 ;pp p pG Q
( )pQ is discharge of the vector fields (current) through the contact sec-
tions ( p pA B and );p pC D dl is arc element of corresponding curve.
The synthesis of the numerical quasiconformal mapping method
and an idea of the rotational block parameterization. The most common
approach for solving of the AQT forward problem is based on the using the
finite element method. But the continuity conditions and, thus, the conserva-
Серія: Фізико-математичні науки. Випуск 18
19
tion law for both a mesh cell and a whole domain doesn't satisfy [15]. So, we
apply the finite difference method for discretization of functions and parame-
ters of AQT mathematical model where the analogies of electrodynamic prin-
ciple are considered. Also we apply quit modified algorithm for numerical
solving of the boundary value problems on quasiconformal mapping of the
domains with the different geometric configuration, which are limited by the
stream and equiquasipotential lines [14] (including the advantages of the ap-
propriate method) for the searching the unknown function of the current’s
potential ( ) ( , ).p x y We search the functions ( ) ( ) ( , )p p x y and
( ) ( ) ( , )p p x y when value , 1,2
( , ,...)x y
is known, by the way of
solving of the boundary value problems, inversed to (4)–(8), on quasiconfor-
mal mapping ( ) ( )p p
zG G (relative to ( ) ( ) ( , )p px x and
( ) ( ) ( , ))p py y with statements which have the form [4, 5, 14]:
( ) ( ) ( )
21 12( )11 22 21 12
11 11 11 11
( ) ( ) ( )
12 21( )11 22 21 12
22 22 22 22
0,
0;
p p p
p
p p p
p
x x x
x
y y y
y
(9)
( ) ( )( )
* *( , ) ( ) ,p ppx x % ( ) ( )( )
* *( , ) ( ) ,p ppy y %
( ) ( )( ) ( , ) ( ) ,p ppx Q x % ( ) ( )( ) ( , ) ( ) ,p ppy Q y %
*( ) *( )( ) ( , ) ( ) ,p ppx x % *( ) *( )( ) ( , ) ( ) ,p ppy y %
( )( ) ( ,0) ( ) ,ppx x % ( )( ) ( ,0) ( ) ;ppy y % (10)
( ) ( )( )
* * ( ) ( ) ( ) ( )( ) ( )
11 12* * * *( ) ( )( )
* *
( , )
( , ) ( , )
( , )
p pp
p p p pp p
p pp
y
y x
x
( ) ( ) ( ) ( ) ( ) ( )( ) ( )
21 22* * * *( , ) ( , ), [ ; ],p p p p p pp p
B Ay x
( ) *( ) *( )
( ) *( ) *( ) ( ) *( ) *( )
11 12( ) *( ) *( )
( , )
( , ) ( ,
( , )
p p p
p p p p p p
p p p
y
y x
x
( ) ( )( ) *( ) *( ) ( ) *( ) *( )
21 22( , ) ( , ), [ ; ],p pp p p p p p
D Cy x
( ) ( ) ( ) ( ) ( ) ( )
11 12( ,0) ( ,0) ( ,0)p p p p p py y x
( ) ( )( ) ( ) ( ) ( ) ( ) ( )
21 22( ,0) ( ,0) ( ,0) , [ ; ],p pp p p p p p
DAx y x
Математичне та комп’ютерне моделювання
20
( ) ( ) ( )
( ) ( ) ( ) ( ) ( ) ( )
11 12( ) ( ) ( )
( ) ( )( ) ( ) ( ) ( ) ( ) ( )
21 22
( , )
( , ) ( , )
( , )
( , ) ( , ), [ ; ]
p p p
p p p p p p
p p p
p pp p p p p p
BC
y Q
y Q x Q
x Q
y Q x Q
(11)
under (4) and (5) conditions, where ( )
* ( ),p ( ) ( ),p
*( ) ( ),p ( ) ( )p ( ) *( )
*( ,p p ( )0 )pQ — functions,
which can be built by the way of interpolation of the experimentally received
their values
*
( )
* ,p
j ( ) ,p
i *
*( ) ,p
j ( )p
i having some arguments
*
( )
* ,p
j ( ) ,p
i
*
*( ) ,p
j ( )p
i on sections ,p pA B ,p pB C p pC D and ,p pA D respectively.
We reconstruct CT, similarly to [3–5], using the minimizing functional
of the sum of the squares of the residual of expressions, receiver from Cauchy-
Riemann type conditions, (4) condition, within using regularization ideas
(1) ( ) (1) ( )
,0 ,0 ,0 0, 0, 0,( ,..., , ,..., , , , ,..., , , )
a b c a b c
df
p p
s s s s s sx x y y a b c a b c % %
2( ) ( ) ( ) ( ) ( ) ( )
11 12 21 22
1
p
p p p p p p
p
y x x y x y
%
2( ) ( ) ( ) ( ) ( ) ( )
11 12 21 22
p p p p p py x x y x y (12)
2 2 2
, , ,
0 0 0 0 0 0
2
100 100 100
a a b b c c
a a a b b b c c c
a b c
a a b b c c
s k s k s k
k r r k r r k r r
k k k
k r k r k r
a b c
2
11 222 ( ) min
and elliptic conditions [14]
2
11 22 12 0, 11 22 0, (13)
where is multiplier, which affects the degree of implementation of
the (4) condition, is regularizing parameter.
Let’s write the difference analogues of problem (4), (5), (9)–(11) in
mesh domains ( ) ,p
zG similarly to [4, 5, 14], in such form:
( ) ( ) ( ) ( )( )2 ( ) ( ) ( )2 ( ) ( )
, , ,1, 1, , 1 , 1
( ) ( ) ( ) ( ) ( ) ( )( ) ( ) ( )
, ,1, 1 1, 1 1, 1 1, 1 , 1 , 1
( )( )
, 1, 1
2(1 ) ( )
( ) / 4 ( )
(
p p p pp p p p p p
i j i j i ji j i j i j i j
p p p p p pp p p
i j i ji j i j i j i j i j i j
pp
i j i j i
x x B x B x x
A x x x x x x D
C x x
( ) ( ) ( ) ( )2 ( ) ( )
, ,, 1, 1,
( ) ( ) ( ) ( ) ( ) ( )( )2 ( ) ( ) ( )
, ,, 1 , 1 1, 1 1, 1 1, 1 1, 1
( ) ( ) ( ) ( ) ( )
, ,, 1 , 1 1
) 0, 0 2(1 )
( ) ( ) /
/4 ( ) (
p p p p p p
i j i jj i j i j
p p p p p pp p p p
i j i ji j i j i j i j i j i j
p p p p p
i j i ji j i j i
y y B y
B y y y y y y A
y y F E y
( ) ( ) ( ) ( )
, 1, ), ( 1, , 1, );p p p p
j i jy i m j n
(14)
Серія: Фізико-математичні науки. Випуск 18
21
( ) ( )
0, * ( ) ,p p
jjx x % ( ) ( )
0, * ( ) ,p p
jjy y % ( )
( ) ( )
, 1 ( ) ,p
p p
ii nx x %
( )
( ) ( )
, 1 ( ) ,p
p p
ii ny y % ( )
( ) *( )
1, ( ) ,p
p p
jm jx x %
( )
( ) *( )
1, ( ) ,p
p p
jm jy y % ( ) ( )
,0 ( ) ,p p
iix x %
( ) ( )
,0 ( )p p
iiy y % ( )(0 1,pi m ( )0 1);pj n (15)
( ) ( ) ( ) ( ) ( ) ( ) ( )
11 120, 1, 2, 0, 1 0, 1 0, 1 0, 1(3 4 ) ( ) ( )p p p p p p p
j j j j j j jy y y y y x x
( ) ( ) ( ) ( ) ( ) ( ) ( )
21 220, 1 2 0 1 0 1 0 1 0 1(3 4 ) ( ) ( ) ,p p p p p p p
j , j , j , j , j , j , jx x x y y x x
( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )
111, , 1, 1, 1 1, 1(3 4 ) ( )p p p p p
p p p p p
m j m j m j m j m jy y y y y
( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
12 211, 1 1, 1 1, 1 1, 1( ) ( )p p p p
p p p p
m j m j m j m jx x y y
( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )
22 1, 1 1, 1 1, , 1,( ) (3 4 ),p p p p p
p p p p p
m j m j m j m j m jx x x x x
( ) ( ) ( ) ( ) ( ) ( )
11 12,0 ,1 ,2 ,0 ,1 ,2( ) ( )
1,0 1,0 ( ) ( )
1,0 1,0
(3 4 ) (3 4 )
( )
p p p p p p
i i i i i ip p
i i p p
i i
y y y x x x
y y
x x
( ) ( ) ( ) ( ) ( ) ( )
21 22,0 ,1 ,2 ,0 ,1 ,2(3 4 ) (3 4 ) 0,p p p p p p
i i i i i iy y y x x x
( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )
11 12, 1 , , 1 , 1 ,(3 4 ) (3 4p p p p p
p p p p p
i n i n i n i n i ny y y x x
( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )
21, 1 1, 1 1, 1 , 1 ,) ( ) (3 4p p p p p
p p p p p
i n i n i n i n i nx y y y y
( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )
22, 1 , 1 , , 1 1, 1) (3 4 ) (p p p p p
p p p p p
i n i n i n i n i ny x x x x
( )
( )
1, 1) 0,p
p
i nx ( )=0, +1,pi m ( )=0, +1;pj n (16)
, ,
( ) ( )( )
, 11 22 ,
i j i j
p pp
i j
(17)
where ( )p is quasiconformal invariant [14] for the corresponding domains
( ) ( ) ( )( , ) :p p p
i jG
( ) ( ) ( ) ,p p p
i i ( )=0, +1;pi m
( ) ( ) ,p p
j j ( )=0, +1;pj n ( ) ( ) ( ) ( )( ) / ( 1),p p p pm
( ) ( ) ( )/ ( 1),p p pQ n ( ) ( ) ( )= / ,p p p ( ) ,pm ( ) ;pn N
( ) ( ) ( ) ( )
, ( , ),p p p p
i j i jx x ( ) ( ) ( ) ( )
, ( , ),p p p p
i j i jy y
,
( ) ( ) ( )
, ,( , ),
i j
p p p
i j i jx y
( ) ( ) ( )
, , ,( , );p p p
i j i j i jx y ( )
, ,p
i jA ( )
, ,p
i jB ( )
, ,p
i jC ( )
, ,p
i jD ( )
, ,p
i jE ( )
,
p
i jF
are defined in the same way as [14].
Математичне та комп’ютерне моделювання
22
We can rewrite the functional (12) and elliptic conditions (13) for all
internal nodal points ( )
,( ,p
i jx ( )
, )p
i jy ( 1, ,p p % ( )1, ,pi m ( )1, )pj n as
follows
( ) ( )
(1) ( )
,0 ,0 ,0 0, 0, 0,1,1 ,( ,..., , , , ,..., , , )p p
a b c a b c
p
s s s s s sm nz z a b c a b c % %
%
( ) ( )
, ,
( ) ( ) ( ) ( ) ( ) ( ), , , 1 , 1 , 1 , 111 12
( )
, , 1
( ) ( )p p
i j i j
p p p p p pp m n i j i j i j i j
p
p i j
y y x x
%
, ,
2 ( ) ( ) ( ) ( ) ( ) ( )( ) ( )
, 1 , 1 , 1 , 121 221, 1,
( ) ( )
( ) ( )
i j i j
p p p p p pp p
i j i j i j i ji j i j
p p
y y x xx x
(18)
2( ) ( ) 2 2
, ,1, 1,
( )
0 0 0 0100 100
a a b b
a a a b b b
a b
a a b b
p p s k s k
k r r k r ri j i j
p k k
k r k r
a by y
, ,
2
, ( ) ( )( ) 2
, 11 22
0 0
( ) min;
100
c c
c c c
i j i jc
c c
s k
k r r p pp
i jk
k r
c
, , ,
( ) ( ) ( )2
111 22 12 ,
i j i j i j
p p p
, ,
( ) ( )
211 22 ,
i j i j
p p
,
( )
311 ,
i j
p
,
( )
422 ,
i j
p (19)
where , ,
a a ak r ra , ,
b b bk r rb ,c c ck r rc ( 0, ,a ak s 0, ,a ar k 0, ,b bk s
0, ,b br k 0, ,c ck s 0, )c cr k are the desired parameters. The solving
of the nonlinear programming problem (18)–(19) is expedient to imple-
ment by one of the appropriate methods of local optimization (for exam-
ple, by the penalty functions method [16]).
Conclusions. The methodology [4, 5] of image reconstruction based
on the idea of the quasiconformal similarity in the small of building the
curvilinear quadrangles — dynamical mesh components in the physical
domain and the corresponding rectangles in the complex quasipotential
domain using AQT data and the rotational parameterization of internal
nodes of dynamic meshes (which are built for each injections) and
searched CT is generalized to the case of anisotropic media. We set the
liquid (current) distribution and the averaged potential at the contact sec-
tions of plate and body unlike the traditional approaches to the statement
and solving the problems of electrical impedance tomography [1–3, 6–13].
We plan to extend the proposed algorithm to the following cases: spatial
resolution, applying the quasipotential of the initial stream to several sections,
parameters identification of CT for piecewise-homogeneous and piecewise-
Серія: Фізико-математичні науки. Випуск 18
23
inhomogeneous media (which, in particular, have place in medical diagnos-
tics). Moreover, in medical practice there are media where the dependence of
conductivity from the direction is not necessarily determined by the corre-
sponding ellipse. In this case, it is expedient to represent the CT as a complex
function. This is one of the areas for our future researches too.
References:
1. Holder D. Electrical Impedance Tomography. Methods, History and Applica-
tions / D. Holder. — Bristol : IOP Publishing, 2005. — 456 p.
2. Horb M. S. Choice of the study object for mathematical model in electrical
impedance tomography / M. S. Horb, O. V. Husieva // Bulletin of NTUU
«KPI». Radiotechnique, Radioaparatus Building Series. — Kyiv : NTUU
«KPI», 2013. — Vol. 52. — P. 120–128. (Ukr).
3. Sherina E. S. Finite volume schemes for the electrical impedance tomography
problem / E. S. Sherina, A. V. Starchenko // Tomsk State University Journal of
Mathematics and Mechanics. — 2014. — Vol. 29 (3). — P. 25–38. (Rus).
4. Bomba A. Ya. Numerical Methods of Quasiconformal Mappings for Solving
Problems of Identifying of Electrical Conductivity Coefficient in an Applied
Potential Tomography / A. Ya. Bomba, L. L. Kroka // Volyn Mathematical
Bulletin. Applied Mathematics Series. — Rivne : Editorial and Publishing De-
partment of RSHU, 2014. — Vol. 11 (20). — P. 24–33. (Ukr).
5. Bomba A. Ya. Applied Quasipotenrial Method for Solving Coefficient Prob-
lems of Parametric Identification / A. Ya. Bomba, M. V. Boichura // Bulletin
of NUWEE. Technical Sciences Series. — Rivne : Editorial and Publishing
Department of NUWEE, 2017. — Vol. 4 (76). — P. 163–177. (Ukr).
6. Martins T. C. Investigating Anisotropic EIT with Simulated Annealing / T. C.
Martins, M. S. G. Tsuzuki // IFAC-PapersOnLine. — 2017. — Vol. 50 (1). —
P. 9961–9966.
7. Lionheart W. R. B. Conformal uniqueness results in anisotropic electrical im-
pedance imaging Inverse Problems / W. R. B. Lionheart // Inverse
Problems. — 1997. — Vol. 13 (1). — P. 125–134.
8. Electrical impedance tomography in anisotropic media with known eigenvec-
tors / J.-F. P. J. Abascal, W. R. B. Lionheart, S. R. Arridge and other // Inverse
Problems. — 2011. — Vol. 27 (6). — P. 1–17.
9. Anisotropic resistivity tomography / J. V. Herwanger, C. C. Pain, A. Bin-
ley and other // Geophysical Journal International. — 2004. — Vol. 158 (2). —
P. 409–425.
10. Crabb M. EIT Reconstruction Algorithms for Respiratory Intensive Care. PhD
Thesis / M. Crabb // Manchester :University of Manchester. — 2014.
11. Tallman T. N. Conductivity-Based Nanocomposite Structural Health Monitor-
ing via Electrical Impedance Tomography. PhD Thesis / T. N. Tallman // Ann
Arbor : University of Michigan. — 2015.
12. Use of anisotropic modelling in electrical impedance tomography; Description
of method and preliminary assessment of utility in imaging brain function in
the adult human head / J.-F. P. J. Abascal, S. R. Arridge, D. Atkinson and oth-
er // NeuroImage. — 2008. — Vol. 43 (2). — P. 258–268.
Математичне та комп’ютерне моделювання
24
13. Vassilevski Yu. V. Finite element analysis in bioimpedance diagnostics /
Yu. V. Vassilevski, A. A. Danilov, D. V. Nikolaev, A. V. Smirnov // Computa-
tional Mathematics and Mathematical Physics. — 2012. — Vol. 52 (4). —
P. 733–745 (Rus).
14. Bomba A. Ya. Complex analysis methods / A. Ya. Bomba, S. S. Kashtan,
D. O. Pryhornytskyi, S. V. Yaroshchak. — Rivne : Editorial and Publishing
Department of NUWEE, 2013. — 415 p. (Ukr).
15. The comparison between FVM and FEM for EIT forward problem / G. Dong,
J. Zou, R. Bayford and other // IEEE Transactions on Magnetics. — 2005. —
Vol. 41 (5). — P. 1468–1471.
16. ALGLIB — C++/C# numerical analysis library. URL: http://www.alglib.net
(last accessed 2018/11/06).
ЧИСЛОВИЙ МЕТОД КОМПЛЕКСНОГО АНАЛІЗУ
ІДЕНТИФІКАЦІЇ ПАРАМЕТРІВ АНІЗОТРОПНИХ
СЕРЕДОВИЩ ЗА ДАНИМИ ТОМОГРАФІЇ
ПРИКЛАДЕНИХ КВАЗІПОТЕНЦІАЛІВ.
ЧАСТИНА 1: ПОСТАНОВКА ТА НАБЛИЖЕННЯ ЗАДАЧІ
Підхід до розв’язання градієнтних задач ідентифікації параметрів
квазіідеальних полів за даними томографії прикладених квазіпотенціа-
лів на основі числових методів комплексного аналізу перенесено на ви-
падки анізотропних середовищ. При цьому, аналогічно до існуючих ро-
біт світових вчених, апріорно відомими вважаються деякі додаткові ві-
домості про характер розподілу провідності всередині області (об’єкта
дослідження). Проте, на відміну від традиційних підходів до постанов-
ки та розв’язання задач електроімпедансної томографії, на ділянках ко-
нтакту пластинки і тіла окрім усередненого потенціалу тут задається ще
й розподіл локальних швидкостей речовини (рідини, струму), а на ін-
ших ділянках (лініях течії) — розподіл потенціалу (за експерименталь-
ними даними, які апроксимуємо із застосуванням сплайнів, кривих
Безьє тощо). Генерація вихідних даних на межі досліджуваного об’єкта
здійснюється відповідно до полярної моделі (схеми) інжекції струму
при заданій сумі власних значень тензора провідності середовища. Ная-
вність такого виду даних значно пришвидшує процес подальшого
розв’язання поставленої задачі, що зручно, зокрема, при верифікації ро-
зробленого авторами методу. Відповідна задача зводиться до ітерацій-
ного розв’язання серії задач для рівнянь типу Лапласа, де замість «при-
граничних числових аналогів рівнянь типу Коші-Рімана» фігурують
співвідношення квазіортогональності за спеціальних типів умов опти-
мізації. А саме: мінімізуючий функціонал побудований з урахуванням
умов типу Коші-Рімана, співвідношення між відповідними тензору ані-
зотропії власними значеннями, а також регуляризуючого доданку; умо-
ви-обмеження сформовані на основі умов еліптичності.
Ключові слова: томографія прикладених квазіпотенціалів, квазі-
конформні відображення, анізотропія, ідентифікація, нелінійні задачі.
Отримано: 28.11.2018
|