mathematics

Article
Fictitious Point Technique Based on Finite-Difference Method
for 2.5D Direct-Current Resistivity Forward Problem
Xiaozhong Tong 1,2,3 and Ya Sun 1,2, *

1 School of Geosciences and Info-Physics, Central South University, Changsha 410083, China;
[email protected]
2 Key Laboratory of Metallogenic Prediction of Nonferrous Metals and Geological Environment Monitoring,
Ministry of Education, Central South University, Changsha 410083, China
3 Key Laboratory of Non-Ferrous and Geological Hazard Detection, Central South University,
Changsha 410083, China
* Correspondence: [email protected]

Abstract: With the widespread application of the direct-current resistivity method, searching for
accurate and fast-forward algorithms has become the focus of research for geophysicists and engineers.
Three-dimensional forward modeling can be the best way to identify geo-electrical anomalies but are
hampered by computational limitations because of the large amount of data. A practical compromise,
or even alternative, is represented by 2.5D modeling characterized using a 3D source in a 2D medium.
Thus, we develop a 2.5D direct-current resistivity forward modeling algorithm. The algorithm
incorporates the finite-difference approximation and fictitious point technique that can improve the
efficiency and accuracy of numerical simulation. Firstly, from the boundary value problem of the
Citation: Tong, X.; Sun, Y. Fictitious
electric potential generated by the point source, the discrete expressions of the governing equation are
Point Technique Based on Finite- derived from the finite-difference approach. The numerical solutions of the discrete electric potential
Difference Method for 2.5D Direct- are calculated after the approximate treatment of the boundary conditions with a finite-difference
Current Resistivity Forward Problem. method based on a fictitious point scheme. Secondly, through the simulation of a homogeneous half-
Mathematics 2024, 12, 269. https:// space model and a one-dimensional model, and compared with the analytical results, the correctness
doi.org/10.3390/math12020269 and stability of the finite-difference forward algorithm are verified. Lastly, through the numerical
Academic Editors: Valentina
simulation for a two-dimensional model, 2.5D direct-current sounding responses are summarized,
Yakovleva and Roman Parovik which can provide a qualitative interpretation of field data.

Received: 28 November 2023 Keywords: direct-current resistivity; forward modeling; finite-difference method; fictitious point
Revised: 3 January 2024
technique; 2.5D
Accepted: 12 January 2024
Published: 14 January 2024
MSC: 86A25
Correction Statement: This article
has been republished with a minor
change. The change does not affect
the scientific content of the article and 1. Introduction
further details are available within the Direct-current resistivity is a surface geophysical method that can provide essential
backmatter of the website version of
information for investigating subsurface geological structures by injecting electric current
this article.
into the Earth and measuring the corresponding voltage. It can find applications in min-
eral resource exploration [1–3], environmental and urban geological surveys [4–7], and
humanitarian geophysics [8]. Regular two-dimensional (2.5D) and three-dimensional (3D)
Copyright: © 2024 by the authors.
surveys are now conducted. In these multidimensional cases, the interpretation of apparent
Licensee MDPI, Basel, Switzerland.
resistivity data requires an accurate modeling approach. The computer resources required
This article is an open access article for 3D direct-current resistivity simulation severely restrict the practical applicability of any
distributed under the terms and automated interpretation technology. The 3D direct-current resistivity forward modeling
conditions of the Creative Commons can be expected not only from the direct development of 3D algorithms but also from a
Attribution (CC BY) license (https:// reasonable compromise between the degree of required forward model complexity and
creativecommons.org/licenses/by/ the level of computer resources consumed. A practical compromise, or even alternative, is
4.0/).

Mathematics 2024, 12, 269. https://doi.org/10.3390/math12020269 https://www.mdpi.com/journal/mathematics

Mathematics 2024, 12, 269 2 of 22

represented by 2.5D direct-current resistivity modeling characterized by the use of a 3D

point source in a 2D geological structure.
Numerical modeling approaches have been developed and applied for direct-current
resistivity forward modeling, such as finite-difference and finite-element methods, as
the approximation of partial differential equations (Poisson equation or Helmholtz equa-
tion) [9]. The finite-difference method is an effective approach for solving partial differential
equations by approximating second-order derivatives with a different scheme in the gov-
erning equation [10–13]. Several studies have explored the efficiency and accuracy of
finite-difference methods for direct-current resistivity modeling [14–18]. However, it is not
easy to calculate the exact space-domain electrical potential. The finite-element method
is another numerical approach for the direct-current resistivity modeling, which solves
the variational problem in integral form for the electrical potential, and then carries out a
numerical integration [19–22]. The finite-element methods use complex real-world infor-
mation, such as topography and bathymetry, when building density models, and improve
the flexibility of 2D, 2.5D, or 3D discretization domains [23]. Unfortunately, high-accuracy
direct-current resistivity modeling needs fine grids, which will lead to high computational
costs [24]. Additionally, there are some other numerical schemes for the 2.5D direct-current
resistivity modeling, such as the finite-volume method [25], the element-free Galerkin
method [26], and the boundary-element method [27]. These methods provide a good
research basis for both numerical simulations of 2D geo-electrical models and numerical
simulations in related geophysical disciplines.
In this work, the finite-difference method to incorporate a fictitious point technique
was developed for 2.5D direct-current resistivity modeling, which is different to the tra-
ditional finite-difference numerical method without a fictitious point scheme. The time
consumption for solving the finite-difference linear equation system without the fictitious
point technique would be longer than that of the finite-difference method with the fictitious
point approach. In addition, the finite-difference forward algorithm incorporated with
the fictitious point technique will give more accurate approximate results than the tradi-
tional finite-difference approach, especially in dealing with the 3D direct-current resistivity
forward problem.
The rest of this paper is arranged as follows. The boundary value problem for the
2.5D wavenumber domain potential is introduced in Section 2. Section 3 is devoted to
presenting the high-performance implementation of the finite-difference algorithm with a
fictitious point technique for 2.5D direct-current resistivity modeling. In Section 4, several
numerical experiments are used to confirm the accuracy and efficiency of the forward
algorithm. The main discussions and conclusions are summarized in Sections 5 and 6,
respectively. Appendix A contains the MATLAB code for 2.5D direct-current resistivity
modeling. Appendices B and C present the mathematical expressions of analytical solutions
in homogenous half-space and two-layer resistivity models, respectively.

2. Statement of 2.5D Direct-Current Resistivity Forward Problem

The 3D space-domain electrical potential u( x, y, z) generated by a point source located
at (xs , 0, 0) on the ground surface can be described as a Poisson equation [28]:

∇ · [σ( x, y, z)∇u( x, y, z)] = − Q( x, y, z)δ( xs )δ(ys )δ(zs ), in Ω (1)

where δ is the unit Dirac (impulse) function, Q( x, y, z) denotes the constant steady-state
current density specified at a point, Ω presents the computational domain, and ( xs , ys , zs )
indicate the coordinates of the point source of charge injected in the x-y-z space.
At the air–earth interface ΓS , shown in Figure 1, the electrical potential satisfies the
following Neumann boundary condition:

∂u
= 0. (2)
∂n ΓS
 also applied. It can be written as
u
=0 , (3)
n 
Mathematics 2024, 12, 269 3 of 22
where n is the outward-pointing normal vector on the truncated boundary   .

S

2
1



Figure 1.
Figure 1. Model
Model of
of 2D
2D geo-electrical
geo-electrical structure.
structure.

Assuming
On the boundary
the distant strike of the
or geological structure isΓ∞
truncated boundary aligned with the boundary
, the Neumann y-axis, thecan
above
be
problem can be simplified and
also applied. It can be written as efficiently simulated in the Fourier domain. The forward
Fourier-cosine transform is defined as follows: ∂u
= 0, (3)
∂n+Γ∞
U ( x,  , z ) =  u ( x, y, z ) cos (  y ) dy (4)
where n is the outward-pointing normal vector 0
on the truncated, boundary Γ∞ .
Assuming the strike of the geological structure is aligned with the y-axis, the above
where  presents the wavenumber and U ( x,  , z ) denotes the wavenumber domain
problem can be simplified and efficiently simulated in the Fourier domain. The forward
electrical potential.
Fourier-cosine Applying
transform the forward
is defined as follows:Fourier-cosine transform to Equations (1)–(3),
after some calculation, the Helmholtz Zboundary value problem for the wavenumber do-
+∞
main electrical potential U ((x,  , z)) =can be expressed
x, λ, u( x, y, z) cosas(λy)dy, (4)
0

   U  and   U U( x, 
 λ, z2) denotes the wavenumber domain elec-
 x   x  + z   z   −   U = −Q   ( xs )  ( zs ) 
where λ presents the wavenumber
trical potential. Applying the forward
  Fourier-cosine
 transform to Equations (1)–(3), after

U
some calculation, the Helmholtz boundary value problem for the wavenumber  domain
=0  (5)
electrical potential U ( x,λ, z
n ) can
S
be expressed as 
h  U  
2 σU = − Q · δ ( x ) δ ( z ) 
 i
∂U =0 ∂
∂
∂x n∂x
σ + ∂z σ ∂U
∂z − λ s s  

,



∂U
= 0 , (5)
where  S and   denote
∂n Γ
S the corresponding 2D boundary of the computational 
 domain.
= 0 current density Q in ( x,  , z ) space can
∂U 
The constant steady-state  be related to the


∂n Γ∞
current density I injected at ( xs , zs ) by
where ΓS and Γ∞ denote the corresponding 2D boundary of the computational domain.
The constant steady-state current density IQ in ( x, λ, z) space can be related to the
Q= ,
current density I injected at ( xs , zs ) by 2A

where A denotes a representative area in the

I x-z plane related to the injection source
Q= ,
point ( xs , zs ) . 2∆A
where ∆A denotes a representative area in the x-z plane related to the injection source point
( x s , z s ).

3. Methodology
To solve the wavenumber domain electrical potential in Equation (5), we should
divide the computational domain Ω into (N − 1) × (M − 1) rectangular elements, shown
in Figure 2. The mesh is designed to be rectangular with an irregular grid spacing in the
x-direction and z-direction. The nodes in the x-axis are represented by i = 1, 2, 3, · · · , N,
and the nodes in the z-axis are represented by j = 1, 2, 3, · · · , M. The representative area
∆A for a source point in the interior can be written as
 To solve the wavenumber domain electrical potential in Equation (5), we should di-
vide the computational domain  into (N − 1) × (M − 1) rectangular elements, shown in
Figure 2. The mesh is designed to be rectangular with an irregular grid spacing in the x-
direction and z-direction. The nodes in the x-axis are represented by i = 1, 2,3, , N , and
Mathematics 2024, 12, 269 of 22 A
the nodes in the z-axis are represented by j = 1, 2,3, , M . The representative 4area
for a source point in the interior can be written as

( xi + xi −1 )  ( z j + z j −1 )
A = + ∆x ) · ∆z + ∆z
,
(∆x i i −1 4j j −1
∆A = ,
4
and in the limit, at the ground surface with z → 0 ,
and in the limit, at the ground surface with z → 0 ,

(∆xi (+
x∆x )  j z j
i + x)i −·1 ∆z
∆A =A =
.
. .
i −1
4 4

z1
z 2
z3

zM −1

x1 x2 x3 xN −1

Figure2.2.Discretization
Figure Discretizationfor
fortwo-dimensional
two-dimensionalgeo-electric model.
geo-electric model.

3.1. Discretization of the Helmholtz Governing Equation

3.1. Discretization of the Helmholtz Governing Equation
The 2D conductivity distribution σ( x, z) is discretized at each node (i, j) by σi,j and
The 2D conductivity
the wavenumber distribution
domain numerical solution  ( xU, i,j
z ) atiseach discretized
node (i, j)atneedseachtonode (i, j) by  i , j
be calculated.
At
andanytheinterior
wavenumber j) with annumerical
node (i, domain irregular grid solutionspacing Ui ,in the x-axis and z-axis, the finite-
j at each node (i, j) needs to be cal-
difference approximation of second-order
culated. At any interior node (i, j) with an irregular grid spacing derivatives in Equation (5) in
canthebex-axis
writtenandas z-axis,
the hfinite-difference
 i approximation  of second-order   derivatives
 in Equation (5) can be
1
written
∂
σ ∂U
∂x as ∂x ≈ ∆xi +∆xi −1 σ ∂U
∂x 1
− σ ∂U
∂x 1
i,j i + ,j i − 2 ,j
2
h σ +σ U2 − U i −1,j + σi,j Ui,j −Ui −1,j
i
2  i + 1,j i,j i + 1,j i,j σ

   U   ≈ U  −
∆x
1i + ∆x i −1 2  i U 
∆x 2 ∆x
 x   x    x + xσi−1,j+σi,j x  1 −  h x σi−1,j  i −1
, (6)
+σi,j σi+1,j +σi,j

i
  i, j
i
1 
=i (∆x +i −∆x
i−1 i−1 2
)∆x 
U i +i −, 1,j
j 
−
i

2 i −
−
(∆x +∆x )∆x
i
1
1
,
i −1
j +
i (∆x +∆x )∆x
i −1
U
i
i,j
2 σi+1,j +σi,j
+ (∆x +∆x )∆x Ui+1,j
2 i  i−i +11, j +i  i , j U i +1, j − U i , j  i −1, j +  i , j U i , j − U i −1, j 
  − 
and xi + xi −1  2 xi 2 xi −1  , (6)
 i −1, j +  i , j     i −1, j + i , j   i +1, j +  i , j 
∂Ui −1, j −  +  Ui, j
h  i
= 1 U
i,j ( xi + j xi −1j−)1xi −1 ∂z i,j+ 1 ( x +  ) ( )
∂
σ ∂U ≈ σ − σ ∂U
  +  
∂z ∂z ∆z +∆z
2

2 i
∂zxi −i,j
1 − 1x
2 i −1 x i x i −1 xi 
 i2+1, j +  ii,j, j+1 i,j i,j+1 i,j − i,j−1 i,j i,j i,j−1
h σ + σ U −U σ + σ U − U
i
≈
+ ∆z j +∆z j−1 2 U ∆z j 2 ∆z j−1
( xi σ+i,j−1x+i −σ1i,j) xi i +1, j  σi,j−1 +σi,j σi,j+1 +σi,j
 . (7)
= Ui,j−1 − + Ui,j
(∆z j +∆z j−1 )∆z j−1 (∆z j +∆z j−1 )∆z j−1 (∆z j +∆z j−1 )∆z j
σi,j+1 +σi,j
+ Ui,j+1
(∆z j +∆z j−1 )∆z j
Mathematics 2024, 12, 269 5 of 22

Substituting Equations (6) and (7) into Equation (5), the finite-difference equation in
the discretized form at any interior node can be written as follows:
σi−1,j +σi,j σ 1,j +σi,j
U
(∆xi +∆xi−1 )∆xi−1 i −1,j
+ (∆x i++∆x Ui+1,j
i i −1 ) ∆xi
σi,j−1 +σi,j σi,j+1 +σi,j
+ U + U
(∆z j +∆z j−1 )∆z j−1 i,j−1 (∆z j +∆z j−1 )∆z j i,j+1
+σ σ 1,j +σi,j
h σ
− (∆x +i−∆x1,j )i,j∆x + (∆x i++∆x
i i −1 i −1 i i −1 ) ∆xi . (8)

σi,j−1 +σi,j σi,j+1 +σi,j
+ + + λ2 σi,j Ui,j
(∆z j +∆z j−1 )∆z j−1 (∆z j +∆z j−1 )∆z j
= − Q · δ( xs )δ(zs )

Using scientific notation, Equation (8) can be written as

ij ij ij ij ij
CW Ui−1,j + CE Ui+1,j + CN Ui,j−1 + CS Ui,j+1 + CP Ui,j = − Q · δ( xs )δ(zs ). (9)

ij ij
where CW is the coupling coefficient related to the node (i − 1, j) and node (i, j), CE is the
ij
coupling coefficient related to the node (i, j) and node (i + 1, j), CS is the coupling coefficient
ij
related to the node (i, j − 1) and node (i, j), CS is the coupling coefficient related to the
ij
node (i, j) and node (i, j + 1), and CP is the self-coupling coefficient at node (i, j). All these
coupling coefficients are given by

ij σi−1,j + σi,j
CW = , (10)
(∆xi + ∆xi−1 )∆xi−1

ij σi+1,j + σi,j
CE = , (11)
(∆xi + ∆xi−1 )∆xi
ij σi,j−1 + σi,j
CN = , (12)
∆z j + ∆z j−1 ∆z j−1

ij σi,j+1 + σi,j
CS = , (13)
∆z j + ∆z j−1 ∆z j


h i
ij ij ij ij ij
CP = − CW + CE + CN + CS + λ2 σi,j . (14)

3.2. Fictitious Point Technique for Boundary Conditions

3.2.1. Boundary Nodes Located on the Top Edge
For all top boundary nodes (i, j) with i = 2, 3, · · · , N − 1 and j = 1, the boundary
∂Ui,j
condition is of the homogeneous Neumann type ∂n = 0, i.e.,

Ui,j−1 = Ui,j+1 .

It can be implemented by assuming a fictitious row of nodes in the air at j = 0, such

that the wavenumber domain electrical potential Ui,2 and the conductivity σi,2 at nodes (i,
2) are reflected at the imaginary nodes (i, 0). This fictitious point technique leads to the
finite-difference form of Equation (5) given by
ij ij ij ij
CW Ui−1,j + CE Ui+1,j + CS Ui,j+1 + CP Ui,j = − Q · δ( xs )δ(zs ), (15)

where the coupling coefficients are given by

Mathematics 2024, 12, 269 6 of 22

ij σi−1,j + σi,j
CW = , (16)
(∆xi + ∆xi−1 )∆xi−1
ij σi+1,j + σi,j
CE = , (17)
(∆xi + ∆xi−1 )∆xi

ij σi,j+1 + σi,j
CS =
2 , (18)
∆z j
h i
ij ij ij ij
CP = − CW + CE + CS + λ2 σi,j . (19)

3.2.2. Boundary Nodes Located on the Bottom Edge

For all bottom boundary nodes (i, j) with i = 2, 3, · · · , N − 1 and j = M, the homoge-
∂Ui,j
neous Neumann boundary condition ∂n = 0 can be written as

Ui,j+1 = Ui,j−1 .

Using the fictitious point technique, it will lead to the finite-difference form of Equation (5)
given by
ij ij ij ij
CW Ui−1,j + CE Ui+1,j + CN Ui,j−1 + CP Ui,j = − Q · δ( xs )δ(zs ), (20)
where the coupling coefficients are given by

ij σi−1,j + σi,j
CW = , (21)
(∆xi + ∆xi−1 )∆xi−1

ij σi+1,j + σi,j
CE = , (22)
(∆xi + ∆xi−1 )∆xi
ij σi,j−1 + σi,j
CN =
2 , (23)
∆z j−1
h i
ij ij ij ij
CP = − CW + CE + CN + λ2 σi,j . (24)

3.2.3. Boundary Nodes Located on the Left Edge

Using the fictitious point technique, the difference equation for boundary nodes (i, j)
with j = 2, 3, · · · , M − 1 and i = 1, has the form given as follows:
ij ij ij ij
CE Ui+1,j + CN Ui,j−1 + CS Ui,j+1 + CP Ui,j = − Q · δ( xs )δ(zs ), (25)

where the coupling coefficients are given by

ij σi+1,j + σi,j
CE = , (26)
(∆xi )2

ij σi,j−1 + σi,j
CN = , (27)
∆z j + ∆z j−1 ∆z j−1

ij σi,j+1 + σi,j
CS = , (28)
∆z j + ∆z j−1 ∆z j


h i
ij ij ij ij
CP = − CE + CN + CS + λ2 σi,j , (29)
Mathematics 2024, 12, 269 7 of 22

3.2.4. Boundary Nodes Located on the Right Edge

Using the fictitious point technique, the finite-difference equation for boundary nodes
(i, j) with j = 2, 3, · · · , M − 1 and i = N, has the form given as follows:
ij ij ij ij
CW Ui+1,j + CN Ui,j−1 + CS Ui,j+1 + CP Ui,j = − Q · δ( xs )δ(zs ), (30)

where the coupling coefficients are given by

ij σi−1,j + σi,j
CW = , (31)
(∆xi−1 )2

ij σi,j−1 + σi,j
CN = , (32)
∆z j + ∆z j−1 ∆z j−1

ij σi,j+1 + σi,j
CS = , (33)
∆z j + ∆z j−1 ∆z j


h i
ij ij ij ij
CP = − CW + CN + CS + λ2 σi,j . (34)

3.2.5. Corner Nodes with Fictitious Point Technique

For the top-left corner node (1, 1), the finite-difference equation can be written as
ij ij ij
CE Ui+1,j + CS Ui,j+1 + CP Ui,j = − Q · δ( xs )δ(zs ), (35)

σi+1,j +σi,j σi,j+1 +σi,j

h i
ij ij ij ij ij
where CE = 2 , CS = 2 , and CP = − CE + CS + λ2 σi,j .
(∆xi ) (∆z j )
For the top-right corner node (N, 1), the finite-difference equation is
ij ij ij
CW Ui+1,j + CS Ui,j+1 + CP Ui,j = − Q · δ( xs )δ(zs ), (36)

σi−1,j +σi,j σi,j+1 +σi,j

h i
ij ij ij ij ij
where CW = , CS = , and CP = − CW + CS + λ2 σi,j .
(∆xi−1 )2 (∆z j )
2

For the bottom-left corner node (1, M), the finite-difference equation is
ij ij ij
CE Ui+1,j + CN Ui,j+1 + CP Ui,j = − Q · δ( xs )δ(zs ), (37)

σi+1,j +σi,j σi,j−1 +σi,j

h i
ij ij ij ij ij
where CE = 2 , CN = 2 , and CP = − CE + CN + λ2 σi,j .
(∆xi ) (∆z j−1 )
For the bottom-right node (N, M), the finite-difference equation is
ij ij ij
CW Ui+1,j + CN Ui,j+1 + CP Ui,j = − Q · δ( xs )δ(zs ), (38)

σi−1,j +σi,j σi,j−1 +σi,j

h i
ij ij ij ij ij
where CW = 2 , CN = 2 , and CP = − CW + CN + λ2 σi,j .
(∆xi−1 ) (∆z j−1 )

3.3. Solution of the Discrete Forward Equation System

With the governing boundary value problem of Equation (5) being approximated
using the finite-difference method in all N × M grid nodes, the linear algebraic equations
(Equations (9), (15), (20), (25), (30) and (36)–(38)) must be properly arranged into a linear
system as follows:
KU = p, (39)
where K is a matrix originating from the coupling coefficients, U is the unknown wavenum-
ber domain electrical potentials vector, and p is the source vector containing I/2∆A at the
current electrode positions. The matrix K generated by the finite-difference algorithm based
on the fictitious point technique in Equation (39) can be a sparse, symmetric, and positive-
Mathematics 2024, 12, 269 8 of 22

definite matrix. Figure 3 shows the non-zero element distribution for an 8 × 8 grid (just for
illustration purposes). Usually, Equation (39) can be considered an ill-conditioned problem.
Therefore, the linear equation system generated by the finite-difference approximation
for the wavenumber domain electrical potential can be solved by Krylov-type iterative
methods. Meanwhile, the appropriate pre-conditioners can also significantly improve the
speed of convergence. In forward modeling, the ILU-BICGSTAB iterative method is9 used,
Mathematics 2024, 12, x FOR PEER REVIEW of 24
which combines a bi-conjugate gradient stabilization algorithm [29,30] with an incomplete
LU decomposition [31].

Figure
Figure 3.
3. Nonzero
Nonzero element
element distribution
distribution of finite-difference coefficient matrix with an 8 ××88grid.
of finite-difference grid.

For the
For the numerical
numericalsimulations
simulationsofofa simple
a simple2D2Dmodel, the the
model, discrete elements
discrete are designed
elements are de-
as 100 × 100 in the whole computational domain and it takes 0.12 s by the
signed as 100 × 100 in the whole computational domain and it takes 0.12 s by the ILU- ILU-BICGSTAB
iterative method.
BICGSTAB iterativeFigure
method.4 shows
Figure the fast convergence
4 shows curve and
the fast convergence curvetotal
andnumber of the
total number
ILU-BICGSTAB iterative method. For multiple point sources, Equation
of the ILU-BICGSTAB iterative method. For multiple point sources, Equation (39) needs (39) needs to be
to be solved multiple times and the apparent resistivity needs to be calculated based the
solved multiple times and the apparent resistivity needs to be calculated based on on
corresponding
the correspondingobservation device.
observation device.
3.4. Calculation of Spatial-Domain Electrical Potential and Apparent Resistivity
After solving the discrete forward equation system, we can obtain the wavenumber
domain electrical potentials U ( x, λ, z) on each node, and then spatial-domain electrical
potentials u( x, y, z) can be calculated by the Fourier inverse transform. On the profile
(y = 0), the Fourier inverse transform formula can be expressed as
Z +∞
2
u( x, 0, z) = U ( x, λ, z)dλ (40)
π 0

Using numerical integration, Equation (40) can be rewritten as

N
u (r ) ≈ ∑ gi U (r, λi ) (41)
i =1
√
where N is the number of wavenumbers; r = x2 + z2 is the distance from the measured
point on the profile to the source point; λi are the discrete wavenumbers; and gi is the corre-
Figure 4. Convergence curve of the ILU-BICGSTAB iterative method with tolerance = 10−12.

3.4. Calculation of Spatial-Domain Electrical Potential and Apparent Resistivity

After solving the discrete forward equation system, we can obtain the wavenumber
domain electrical potentials U ( x,  , z ) on each node, and then spatial-domain electrical
 For the numerical simulations of a simple 2D model, the discrete elem
signed as 100 × 100 in the whole computational domain and it takes 0.12
Mathematics 2024, 12, 269 BICGSTAB iterative method. Figure 4 shows the fast convergence9curve of 22 and
of the ILU-BICGSTAB iterative method. For multiple point sources, Equati
to be solved multiple times and the apparent resistivity needs to be calcula
sponding weighting coefficient. We use the optimization scheme developed by Xu et al. [32]
the corresponding
to select observation
the wavenumber. device.
The wavenumber used in our 2.5D direct-current resistivity
forward modeling is shown in Table 1.

Convergence
Figure 4.4. curve of the ILU-BICGSTAB iterative method with tolerance −12
= 10 with
.
Figure Convergence curve of the ILU-BICGSTAB iterative method tolerance
Table 1. Discrete wavenumbers λi and corresponding coefficients gi of inverse Fourier transform.
3.4. Calculation of Spatial-Domain Electrical Potential and Apparent Resistivity
i λi gi
After solving
1 the discrete forward
0.0066954 equation system, we can obtain the
0.0099032
domain electrical potentials U ( x,  , z ) on each node, and then spatial-dom
2 0.5160113 0.1759048
3 0.2388812 0.0760419
4 ( ) 0.1295920 0.0578590 transform.
potentials u x, y, z can be calculated by the Fourier inverse
5 0.9918394 0.4483043
On
= 0), the Fourier
6 inverse transform formula can be expressed
0.0623708 0.0292485as
7 0.3515289 0.0678736
8 2.0863820 1.1451725
9 0.0282197 0.0170322
10 6.9549240 7.9501720

In direct-current resistivity exploration, a current source +I and a current sink −I

are used to energize the conductive Earth, and the apparent resistivity can be defined
as follows:
∆u
ρs = G (42)
I
where G is a geometric factor, which depends upon the locations of electrodes, and ∆u is
the electrical potential difference.
According to the above 2.5D finite-difference algorithm analysis, the complete work-
flow is shown in Figure 5, which can compute the approximate numerical electrical potential
and apparent resistivity.
Mathematics
Mathematics 2024, 12, 269
2024, 12, x FOR PEER REVIEW 10 of 11
22 of 24

Figure5.5.The
Figure Themain
mainsteps
steps
ofof 2.5D
2.5D direct-current
direct-current resistivity
resistivity finite-difference
finite-difference forward
forward algorithm
algorithm basedbased
on fictitious point technique.
on fictitious point technique.

4.4.Numerical
NumericalExperiments
Experiments
The
TheLenovo
LenovoWorkstation
WorkstationP520 with
P520 an an
with 8-Core Intel
8-Core Xeon
Intel W-2145
Xeon processor
W-2145 including
processor including
32 GB of RAM was applied to execute our numerical simulation. The forward algorithm
32 GB of RAM was applied to execute our numerical simulation. The forward algorithm
was developed in MATLAB (R2023a), and the subroutine is given in Appendix A.
was developed in MATLAB (R2023a), and the subroutine is given in Appendix A.
4.1. Benchmark with Homogeneous Half-Space
4.1. Benchmark with Homogeneous Half-Space
The benchmark of the accuracy of the finite-difference forward algorithm began with
The benchmark
a homogeneous of the
half-space accuracy
model. of the finite-difference
The computational forward as
size was designed algorithm
200 m × began
100 m.with
a homogeneous half-space model. The computational size was designed
The ground surface was considered flat and uniform, and homogeneous with a conductivity as 200 m × 100 m.
The
of 0.01ground
S/m. Thesurface wasand
positive considered flat andelectrode
negative current uniform,were andplaced
homogeneous
at (−20 m,with
0 m)a and
conduc-
tivity
(20 m, 0ofm),
0.01 S/m. The positive
respectively, and theand negative
current current
was set as I = electrode
1A. For thewere placedsimulation,
numerical at (−20 m, 0 m)
we
andsolved
(20 m,the problem
0 m), respectively, × 50the
on a 100 and grid.
current was set as I = 1A. For the numerical simula-
The numerical spatial-domain electrical
tion, we solved the problem on a 100 × 50 grid. potentials for the homogeneous half-space
modelTheare numerical
shown in Figure 6. It is evident that the spatial-domain
spatial-domain electrical potentials electrical potentials
for the homogeneous half-space
computed by the finite-difference method with the fictitious point
model are shown in Figure 6. It is evident that the spatial-domain electrical technique agree potentials
with
the analytical
computed bysolutions given in Appendix
the finite-difference method B. The
withfinite-difference numerical
the fictitious point results
technique indi-with
agree
cate that our fictitious point technique can provide high accuracy for 2.5D direct-current
the analytical solutions given in Appendix B. The finite-difference numerical results indi-
resistivity modeling.
cate that our fictitious point technique can provide high accuracy for 2.5D direct-current
resistivity modeling.
Mathematics 2024, 12, x FOR PEER REVIEW 12 o
Mathematics 2024, 12, 269 11 of 22

Figure 6. Comparison
Figure 6. Comparison ofofanalytical
analyticalandand finite-difference
finite-difference numerical
numerical solution
solution of the
of the point point source
source
potentials in the homogeneous half-space model.
tentials in the homogeneous half-space model.

The relative root-mean-square error was used to measure the overall accuracy of our
The relative root-mean-square error was used to measure the overall accuracy of
finite-difference scheme [33]:
finite-difference scheme [33]: q
2
∑iN=1 ∑ jM=1 ui,j − ũi,j
  ( )
N M 2
Error = q
N =1
iM

u ×
j =1 2 i , j
− u100
i, j
per cent, (43)
∑ ∑
Error = i=1 j=1 ũ i,j  100 per cent ,
  (u )
N M 2

where N and M are the numbers of observation i =1 j =1 i, j

nodes in the x- and y-direction, respectively,
and ui,j and uei,j represent the numerical and analytical solution, respectively. Based on
where
uniformNmeshes
and M in are the numbers
the x-direction ofhomogeneous
for the observationhalf-space
nodes inmodel,
the x-the
and y-direction, resp
convergence
tively, andlog–log
plot in the ui , j and ui , jtherepresent
scale for the numerical error
relative root-mean-square and isanalytical solution,
shown in Figure respectiv
7. The
error decreases as the grid step size becomes smaller, and numerical stability increases
Based on uniform meshes in the x-direction for the homogeneous half-space model,
with respect to grid size. This further illustrates the accuracy of our finite-difference
convergence plot in the log–log scale for the relative root-mean-square error is show
forward algorithm.
Figure 7. The error decreases as the grid step size becomes smaller, and numerical stab
4.2. Two-Layer
increases withConductivity
respect toModel
grid size. This further illustrates the accuracy of our finite-diff
ence forward algorithm.
A two-layer conductivity model was applied for another numerical experiment to
verify the accuracy of our forward modeling code. This two-layer model included a low-
conductivity medium in the first layer and a high-conductivity medium in the second layer,
as shown in Figure 8. The thickness of the low-conductivity medium was designed to be
4 m with a conductivity of 0.005 S/m, while the thickness of the high-conductivity medium
was designed to be 100 m with a bearing conductivity of 0.05 S/m.
The computational domain was designed as 200 m × 200 m. The direct-current
resistivity data were computed using the Wenner array by varying AB/2 from 1.5 m to 75 m,
and the apparent resistivity curve calculated by the finite-difference scheme is displayed in
Figure 9a. The finite-difference numerical results were compared to the analytical solutions
given in Appendix C from Telford et al. [34]. The comparison between finite-difference
numerical and analytical apparent resistivities shows a relative error below 2%, shown in
Figure 9, which indicates that the 2.5D forward algorithm can satisfy practical requirements
in terms of simulation accuracy. The maximum error occurs near the point source and AB/2
is approximately 70 m, but their relative errors are all within 2%. Further analysis shows
that the errors in the 2.5D finite-difference forward modeling of direct-current sounding
 where N and M are the numbers of observation nodes in the x- and y-dire
tively, and ui , j and ui , j represent the numerical and analytical solution
Mathematics 2024, 12, 269 Based on uniform meshes in the x-direction for the homogeneous half-spa
12 of 22

convergence plot in the log–log scale for the relative root-mean-square erro
Figure 7. The
mainly come fromerror decreases
three aspects: asthe
firstly, the gridarestep
errors size
caused bybecomes
singularity smaller, and num
problems near
increases withsecondly,
the point source; respect theto gridcan
errors size. This by
be caused further illustrates
truncating boundary the accuracy
conditions and of ou
finite-difference discretization;
ence forward algorithm. lastly, the errors are caused by numerical integration of the
discrete Fourier inverse transform.

Mathematics 2024, 12, x FOR PEER REVIEW

4.2. Two-Layer Conductivity Model

A two-layer conductivity model was applied for another numerical e
verify the accuracy of our forward modeling code. This two-layer model in
conductivity medium in the first layer and a high-conductivity medium
layer, as shown in Figure 8. The thickness of the low-conductivity medium
to be 4 m with a conductivity of 0.005 S/m, while the thickness of the high
Figure
Figure 7. 7.
medium TheThe
was convergence
convergence plot
bein100
plot in log–log
designed to log–log
scale scalea using
using relative
m with relative
root-mean-square
bearing root-mean-square
error with step sizes
conductivity error
of 0.05 S/m.
in the
in theinterval [0.1,[0.1,
interval 1]. 1].

1 = 0.005 S/ m h1 = 4m

 2 = 0.05 S/ m

Figure 8. Schematic diagram of two-layered horizontal model.

Figure 8. Schematic diagram of two-layered horizontal model.

The computational domain was designed as 200 m × 200 m. The direct

tivity data were computed using the Wenner array by varying AB/2 from 1
and the apparent resistivity curve calculated by the finite-difference schem
in Figure 9a. The finite-difference numerical results were compared to the a
tions given in Appendix C from Telford et al. [34]. The comparison between
ence numerical and analytical apparent resistivities shows a relative err
Mathematics 2024, 12, x FOR PEER REVIEW 14 of 24
Mathematics 2024, 12, 269 13 of 22

4.3. A Rectangular Conductive Body in Half-Space

To illustrate the accuracy and the efficiency of our finite-difference forward algorithm,
the finite-element method with the structured rectangular mesh was adopted. In order
to show a comparison of accuracy and efficiency, the numerical experiment was applied
for a simple 2D model, which is shown in Figure 10. It was a symmetrical, rectangular,
conductive body with a resistivity of 10 Ω · m inside a uniform half-space. The background
resistivity of this 2D model was designed as 100 Ω · m. The width and the thickness14ofofthe
Mathematics 2024, 12, x FOR PEER REVIEW 24
rectangular body were 4 m and 2 m, respectively. The rectangular body was located at a
depth of 2 m from the ground surface.

Figure 9. Comparison of analytical and finite-difference solutions for the two-layer resistivity model:
(a) analytical and numerical values of apparent resistivity; (b) relative error.

4.3. A Rectangular Conductive Body in Half-Space

To illustrate the accuracy and the efficiency of our finite-difference forward algorithm,
the finite-element method with the structured rectangular mesh was adopted. In order to
show a comparison of accuracy and efficiency, the numerical experiment was applied for
a simple 2D model, which is shown in Figure 10. It was a symmetrical, rectangular, con-
ductive body with a resistivity of 10   m inside a uniform half-space. The background
resistivity of this 2D model was designed as 100   m . The width and the thickness of the
Figure 9. Comparison
Comparison
rectangular of analytical
body were 4 m andand
analytical 2 m,finite-difference
respectively. solutions
finite-difference for the two-layer
The rectangular body wasresistivity
located model:
at a
(a) analytical and numerical values
depth of 2 m from the ground surface.of apparent resistivity; (b) relative error.
error.

4.3. A Rectangular Conductive Body in Half-Space

To illustrate the accuracy and the efficiency of our finite-difference forward algorithm,
the finite-element method with the structured rectangular mesh was adopted. In order to
show a comparison of accuracy and efficiency, the numerical experiment was applied for
a simple 2D model, which is shown in Figure 10. It was a symmetrical, rectangular, con-
ductive body with a resistivity of 10   m inside a uniform half-space. The background
2 this
resistivity of = 10 2D
mmodel was designed as 100   m . The width and the thickness of the
rectangular body were 4 m and 2 m, respectively. The rectangular body was located at a
depth of 2 m from the ground surface.
 = 100   m
1

Figure 10. A two-dimensional

Figure two-dimensional geo-electrical
geo-electricalmodel
modelwith
withrectangular
rectangularbody
bodyburied
buriedinin
thethe
homogeneous
homogene-
ous half-space.
half-space.

The computational size was designed as 100 m × 100 m. The Wenner-alpha array was
selected for the simulation and the apparent resistivity pseudo-section calculated by
2 = 10   m

1 = 100   m
 finite-difference scheme is displayed in Figure 11a. In order to produce the closed contour
map, the resistivity distribution characteristics of the low-resistivity anomalous body (or
high-conductivity anomalous body) can be qualitatively distinguished, and the spatial
distribution
Mathematics 2024, 12, 269 characteristics of the anomalous body can be accurately distinguished. 14 of 22
Figure 11b shows the apparent resistivity pseudo-section of the Wenner-alpha array
calculated by the finite-element method from Res2dmod software (Windows version
The computational
3.03.06) [35]. The finite-difference resultssize wascompared
were designed asto100 × 100 m. The Wenner-alpha
themfinite-element solutions, array
was selected for the simulation and the apparent resistivity pseudo-section calculated by
and the absolute error is displayed in Figure 11c. The absolute errors are all within 4   m .
finite-difference scheme is displayed in Figure 11a. In order to produce the closed contour
By comparing the finite-difference results with the fictitious point technique and the finite-
map, the resistivity distribution characteristics of the low-resistivity anomalous body (or
element results, wehigh-conductivity
found that the accuracy
anomalousofbody)
the two ways
can be is almost
qualitatively the same, and
distinguished, and the
the spatial
results agree well. distribution characteristics of the anomalous body can be accurately distinguished.

s

Figure 11. NumericalFigure

results 11.for rectangular
Numerical resultsbody model bybody
for rectangular different
modelfinite-element methods methods
by different finite-element with with
Wenner-alpha electrode configuration:
Wenner-alpha (a)configuration:
electrode finite-difference solution based
(a) finite-difference onbased
solution fictitious pointpoint
on fictitious tech-technique;
nique; (b) finite-element solution based on Res2dmod software; (c) absolute error.
(b) finite-element solution based on Res2dmod software; (c) absolute error.

Figure 11b shows the apparent resistivity pseudo-section of the Wenner-alpha ar-
If the Schlumberger array was selected to simulate the vertical electrical sounding
ray calculated by the finite-element method from Res2dmod software (Windows version
data, the apparent resistivity
3.03.06) [35].pseudo-section calculated
The finite-difference by finite-difference
results were scheme is dis-
compared to the finite-element solutions,
played in Figure 12a. andAsthecan be seen
absolute inisFigure
error 12,inthe
displayed numerical
Figure 11c. Theapparent resistivities
absolute errors cal-4 Ω · m.
are all within
By comparing method
culated by the finite-difference the finite-difference
with the results withpoint
fictitious the fictitious point and
technique technique and the finite-
the finite-
element method with Res2dmod software are in close agreement with each other. Com-and the
element results, we found that the accuracy of the two ways is almost the same,
results agree well.
pared to the finite-element results (Figure 12b), the maximum absolute error for the ap-
If the Schlumberger array was selected to simulate the vertical electrical sounding data,
parent resistivity is the   m (Figure
3.4apparent resistivity12c).
pseudo-section calculated by finite-difference scheme is displayed
in Figure 12a. As can be seen in Figure 12, the numerical apparent resistivities calculated
by the finite-difference method with the fictitious point technique and the finite-element
method with Res2dmod software are in close agreement with each other. Compared to the
finite-element results (Figure 12b), the maximum absolute error for the apparent resistivity
is 3.4 Ω · m (Figure 12c).
, x FOR PEER REVIEW 16 of 24
Mathematics 2024, 12, 269 15 of 22

s

Figure 12. NumericalFigure

results 12.for rectangular
Numerical resultsbody model bybody
for rectangular different
modelfinite-element methods methods
by different finite-element with with
Schlumberger electrode configuration:
Schlumberger (a)configuration:
electrode finite-difference solution based
(a) finite-difference onbased
solution fictitious pointpoint
on fictitious tech-technique;
nique; (b) finite-element solution based
(b) finite-element on Res2dmod
solution software;
based on Res2dmod (c) absolute
software; error.
(c) absolute error.

5. Discussion
5. Discussion
In the finite-difference algorithm for 2.5D direct-current resistivity forward modeling,
In the finite-difference algorithm
if the fictitious for 2.5D direct-current
point technique was not applied,resistivity
the coefficientforward
matrix modeling,
K in Equation (39)
must be unsymmetrical and non-positive. Figure 13 shows
if the fictitious point technique was not applied, the coefficient matrix K in Equation (39) the nonzero element distribution
must be unsymmetrical and non-positive. Figure 13 shows the nonzero element distribu-based on
of a finite-difference coefficient matrix formed without a fictitious point technique
an 8 × 8 grid (just for illustration purposes). The time consumption for solving the linear
tion of a finite-difference coefficient matrix formed without a fictitious point technique
equation system of the finite-difference method without the fictitious point technique could
based on an 8 × 8 grid (just than
be longer for illustration purposes). The
that of the finite-difference time
method consumption
with for solving
the fictitious point approach.
the linear equation system of the finite-difference
Comparing the efficiency of two method without
computing the fictitious
strategies, point tech-
the homogeneous half-space
model in Section 4.1 is adopted. The computational
nique could be longer than that of the finite-difference method with the fictitious pointsize is designed as 200 m × 100 m.
approach. The positive and negative current electrodes are placed at ( − 20 m, 0 m) and (20 m, 0 m),
respectively, and the current is set as I = 1 A. The time consumption of the finite-difference
Comparing themethodefficiency of two computing strategies, the homogeneous half-space
with the fictitious point technique is 0.12 s. However, the time consumption of the
model in Section 4.1finite-difference
is adopted. The computational
method size is designed
without the fictitious point techniqueas 200 m ×s.100
is 0.65 Them. The
computational
positive and negative costcurrent
is mainlyelectrodes
consumed for are placed
solving theat (−20equation
linear m, 0 m) and In
system. (20addition,
m, 0 m), the re-
maximum
spectively, and the relative
currenterror foras
is set theI =
electrical
1 A. Thepotential using the fictitious
time consumption point
of the technique is 0.39%, and
finite-difference
that without
method with the fictitious pointthe fictitious is
technique point
0.12scheme is 31.56%
s. However, the(Figure 14). Therefore,of
time consumption the accuracy
the
and efficiency of our 2.5D direct-current resistivity forward algorithm are superior to the
finite-difference method without the fictitious point technique is 0.65 s. The computational
finite-difference method without the fictitious point technique.
cost is mainly consumed for solving the linear equation system. In addition, the maximum
relative error for the electrical potential using the fictitious point technique is 0.39%, and
that without the fictitious point scheme is 31.56% (Figure 14). Therefore, the accuracy and
efficiency of our 2.5D direct-current resistivity forward algorithm are superior to the fi-
nite-difference method without the fictitious point technique.
Mathematics 2024,
Mathematics 2024, 12,
12, 269
x FOR PEER REVIEW 17 of 24
Mathematics 2024, 12, x FOR PEER REVIEW 17 16
of of
2422

Figure 13.
Figure Nonzero
13.Nonzero
13. Nonzero element
element
element distribution
distribution
distribution of finite-difference
ofoffinite-difference
finite-difference coefficient
coefficient
coefficient matrix
matrix
matrix formed
formed
formed without
without
without ficti-
ficti-
tious point
point
fictitious
tious technique
point based
technique
technique onan
based
based on an
on88an
××88grid.
grid.
× 8 grid.

Figure
Figure 14.
14. Comparison
Comparisonof
Comparison numerical
ofof
numerical solution
numerical without
solution
solution fictitious
without
without point
fictitious
fictitious technique
point
point and
technique
technique andnumerical
and so-so-
numerical
numerical
lution with
solution with
lution with fictitious
fictitious point
fictitious technique
point
point of
technique
technique the point
of point
of the source
the point potentials
source
source in the
potentials
potentials homogeneous
in theinhomogeneoushalf-space
the homogeneous half-
half-space
model.
model.model.
space

To
To discuss
To discuss the
discuss the stability
stabilityof
ofthe
theproposed
proposedfinite-difference
finite-differencescheme,
scheme,letletususalso
alsocompute
compute
the condition
the condition number
conditionnumber of
numberofofthe the matrix
matrix
the K
K in
matrix in Equation
K Equation (39),
(39),(39),
in Equation since
sincesince it
it turns turns out
out that
it turns that
outthis this is isa a
is athis
that critical
critical
quantity quantity in determining
in determining
critical quantity how
how rapidly
in determining rapidly certain
certaincertain
how rapidly iterative
iterative methods
iterative methods converge.
converge.
methods TheThe2-2-
The 2-norm
converge.
norm condition
condition numbernumber is defined
is defined by by
norm condition number is defined by
s
−1 λmax ( A T A)
cond(K )2 = ∥K ∥2 K = , (44)
2 λmin ( AA T )
 max ( AT A )
cond ( K )2 = K K −1
= , (44
Mathematics 2024, 12, 269
2 2
min ( AAT ) 17 of 22

where max and min are the maximum and minimum of the eigenvalues, respectively
where λmax and λmin are the maximum and minimum of the eigenvalues, respectively.
The 2-norm condition number values as a function of the number of nodes for differen
The 2-norm condition number values as a function of the number of nodes for different
meshesininthe
meshes the2D2D wavenumber
wavenumber domain
domain are shown
are shown in Figure
in Figure 15. The15. The
fact fact
that thethat the matrix
matrix
becomes more ill-conditioned as we refine the grid is responsible for the slow-down in the in the
becomes more ill-conditioned as we refine the grid is responsible for the slow-down
BICGSTABiterative
BICGSTAB iterative method.
method.

Figure15.
Figure 15.Condition
Condition number
number values
values as aas a function
function of theofnumber
the number of for
of nodes nodes for different
different meshes. meshes.

6. Conclusions
6. Conclusions
A high-accuracy 2.5D direct-current resistivity finite-difference forward algorithm
basedAonhigh-accuracy 2.5D
a fictitious point direct-current
technique resistivity
is developed. finite-difference
We present the calculation forward
formulasalgorithm
based on a fictitious point technique is developed. We present the calculation
of this approach and provide a successful implementation. All mathematical formulas of formulas o
thisforward
the approach and provide
modeling a successful
algorithm implementation.
are presented All mathematical
and implemented in MATLAB code. formulas
The of the
numerical results show
forward modeling that theare
algorithm proposed algorithm
presented has high accuracy
and implemented for 2.5D direct-
in MATLAB code. The nu
current
mericalresistivity
results showforward
thatmodeling,
the proposedwhichalgorithm
can provide
hasa high
qualitative interpretation
accuracy of
for 2.5D direct-cur
field data.
rent resistivity forward modeling, which can provide a qualitative interpretation of field
Unlike the conventional finite-difference method without a fictitious point approach
data.
in the spatial domain, which can lead to a linear equation with an unsymmetrical and non-
Unlike the conventional finite-difference method without a fictitious point approach
positive matrix, our scheme leads to a linear equation with a sparse, symmetrical positive
in the spatial
matrix. domain,
Therefore, which can lead
the computational time to
anda the
linear equation
memory usagewith
spentan onunsymmetrical
solving the linearand non
positive matrix, our scheme leads to a linear equation with a sparse,
system equations should be lower than that of the conventional finite-difference symmetrical
method. positive
matrix. Therefore, the computational time and the memory usage spent on solving the
Author
linear Contributions:
system equationsConceptualization,
should be X.T.lowerand than
Y.S.; formal analysis,
that of X.T.; funding acquisition,
the conventional finite-difference
Y.S.; methodology, X.T. and Y.S.; project administration, Y.S.; visualization, X.T.; supervision, Y.S. All
method.
authors have read and agreed to the published version of the manuscript.
Funding: This research work
Author Contributions: was partly supported
Conceptualization, by the
X.T. and National
Y.S.; formalNatural Science
analysis, Foundation
X.T.; funding acquisition
of China (grant nos. 42274083 and 41974049) and partly by the Hunan National Natural Science
Y.S.; methodology, X.T. and Y.S.; project administration, Y.S.; visualization, X.T.; supervision, Y.S
Foundation
All authors(grant
have no. 2023JJ30659).
read and agreed to the published version of the manuscript.
Data Availability Statement: Data associated with this research are available and can be obtained by
Funding: This research work was partly supported by the National Natural Science Foundation o
contacting the corresponding author.
China (grant nos. 42274083 and 41974049) and partly by the Hunan National Natural Science Foun
Acknowledgments: The authors would like to thank Dawei Gao, who modified this manuscript to
dation (grant no. 2023JJ30659).
improve the English writing quality and gave helpful discussions about the results of the models.
Data
We Availability
would also like toStatement: Data and
thank the editors associated with this
the reviewers research comments
for providing are available and can be obtained
that substantially
by contacting the
improved the paper. corresponding author.
Mathematics 2024, 12, 269 18 of 22

Conflicts of Interest: The authors declare no conflicts of interest.

Appendix A. MATLAB Program for 2.5D Direct-Current Resistivity Modeling

Functio n [U]=fdm_25DC_Neumann(x,dx,z,dz,t,I,sigm)
%Input arguments
%x: x-direction coordinate vector
%dx Mesh size in x-direction
%z: z-direction coordinate vector
%dz: Mesh size in z-direction
%t: The locattion of the surce
%I: The current value
%sigm: The conductivity matrix
%Output argument
%U: Space-domain electrical Potential
dA=(x(t+1)-x(t-1))*dz(1)/4;
M=length(x);
N=length(z); % Number of rows in the model matrix
q(M*N,1)=0;
xs=x(t);
zs=0;
q((t-1)*N+1)=2*I/(2*dA);
Ky = [0.00669536185444851, 0.516011331129247,...
0.23888121181836100, 0.129592031420649,...
0.99183940749273000, 0.062370794321097,...
0.35152894553111700, 2.086382045795400,...
0.02821971094872780, 6.95492402823451];
Gy = [0.00990316118235839, 0.175904762136477,...
0.07604192559029110, 0.057858976242153,...
0.44830432193231000, 0.029248474429237,...
0.06787363490235750, 1.145172473298260,...
0.01703221854812900, 7.950172047416480];
A=sparse(M*N,M*N);
for k=1:length(Ky)
for i=1:N
for j=1:M
m=(j-1)*N+i;
if i~=1 && i~=N && j~=1 && j~=M
aW=-2*(sigm(i,j-1)+sigm(i,j))/(dx(j)+dx(j-1))/dx(j-1);
aE=-2*(sigm(i,j+1)+sigm(i,j))/(dx(j)+dx(j-1))/dx(j);
aN=-2*(sigm(i-1,j)+sigm(i,j))/(dz(i)+dz(i-1))/dz(i-1);
aS=-2*(sigm(i+1,j)+sigm(i,j))/(dz(i)+dz(i-1))/dz(i);
AA= 2*Ky(k)ˆ2*sigm(i,j);
aP=-(aE+aW+aN+aS)+AA;
A(m,m)=aP;
A(m,m-1)=aN;
A(m,m+1)=aS;
A(m,m-N)=aW;
A(m,m+N)=aE;
elseif i==1 && j~=1 && j~=M
aW=-2*(sigm(i,j-1)+sigm(i,j))/(dx(j)+dx(j-1))/dx(j-1);
aE=-2*(sigm(i,j+1)+sigm(i,j))/(dx(j)+dx(j-1))/dx(j);
aS=-2*(sigm(i+1,j)+sigm(i,j))/dz(i)/dz(i);
AA= 2*Ky(k)ˆ2*sigm(i,j);
aP=-(aE+aW+aS)+AA;
Mathematics 2024, 12, 269 19 of 22

A(m,m)=aP;
A(m,m+1)=aS;
A(m,m-N)=aW;
A(m,m+N)=aE;
elseif i==N && j~=1 && j~=M
aW=-2*(sigm(i,j-1)+sigm(i,j))/(dx(j)+dx(j-1))/dx(j-1);
aE=-2*(sigm(i,j+1)+sigm(i,j))/(dx(j)+dx(j-1))/dx(j);
aN=-2*(sigm(i-1,j)+sigm(i,j))/dz(i-1)/dz(i-1);
AA= 2*Ky(k)ˆ2*sigm(i,j);
aP=-(aW+aN+aE)+AA;
A(m,m)=aP;
A(m,m-1)=aN;
A(m,m-N)=aW;
A(m,m+N)=aE;
elseif j==1 && i~=1 && i~=N
aE=-2*(sigm(i,j+1)+sigm(i,j))/dx(j)/dx(j);
aN=-2*(sigm(i-1,j)+sigm(i,j))/(dz(i)+dz(i-1))/dz(i-1);
aS=-2*(sigm(i+1,j)+sigm(i,j))/(dz(i)+dz(i-1))/dz(i);
AA= 2*Ky(k)ˆ2*sigm(i,j);
aP=-(aE+aN+aS)+AA;
A(m,m)=aP;
A(m,m-1)=aN;
A(m,m+1)=aS;
A(m,m+N)=aE;
elseif j==M && i~=1 && i~=N
aW=-2*(sigm(i,j-1)+sigm(i,j))/dx(j-1)/dx(j-1);
aN=-2*(sigm(i-1,j)+sigm(i,j))/(dz(i)+dz(i-1))/dz(i-1);
aS=-2*(sigm(i+1,j)+sigm(i,j))/(dz(i)+dz(i-1))/dz(i);
AA= 2*Ky(k)ˆ2*sigm(i,j);
aP=-(aW+aN+aS)+AA;
A(m,m)=aP;
A(m,m-1)=aN;
A(m,m+1)=aS;
A(m,m-N)=aW;
elseif i==N && j==1
aE=-2*(sigm(i,j+1)+sigm(i,j))/dx(j)/dx(j);
aN=-2*(sigm(i-1,j)+sigm(i,j))/dz(i-1)/dz(i-1);
AA= 2*Ky(k)ˆ2*sigm(i,j);
aP=-(aE+aN)+AA;
A(m,m)=aP;
A(m,m-1)=aN;
A(m,m+N)=aE;
elseif i==N && j==M
aW=-2*(sigm(i,j-1)+sigm(i,j))/dx(j-1)/dx(j-1);
aN=-2*(sigm(i-1,j)+sigm(i,j))/dz(i-1)/dz(i-1);
AA= 2*Ky(k)ˆ2*sigm(i,j);
aP=-(aW+aN)+AA;
A(m,m)=aP;
A(m,m-1)=aN;
A(m,m-N)=aW;
elseif i==1 && j==1
aE=-2*(sigm(i,j+1)+sigm(i,j))/dx(j)/dx(j);
aS=-2*(sigm(i+1,j)+sigm(i,j))/dz(i)/dz(i);
AA= 2*Ky(k)ˆ2*sigm(i,j);
Mathematics 2024, 12, 269 20 of 22

aP=-(aE+aS)+AA;
A(m,m)=aP;
A(m,m+1)=aS;
A(m,m+N)=aE;
elseif i==1 && j==M
aW=-2*(sigm(i,j-1)+sigm(i,j))/dx(j-1)/dx(j-1);
aS=-2*(sigm(i+1,j)+sigm(i,j))/dz(i)/dz(i);
AA= 2*Ky(k)ˆ2*sigm(i,j);
aP=-(aW+aS)+AA;
A(m,m)=aP;
A(m,m+1)=aS;
A(m,m-N)=aW;
end
end
end
[L,U] = ilu(A);
[VV(:,k),fl1,rr1,it1,rv1] = bicgstab(A,q,1e-10,20,L,U);
V(:,:,k) = reshape(VV(:,k),N,M);
end
U=zeros(N,M);
for k=1:length(Gy)
U=U+V(:,:,k)*Gy(k);
end

Appendix B. Mathematical Expressions of the Electrical Potential in Homogenous

Half-Space
In a homogeneous half-space, the electrical potential of a single current electrode at
the surface can be written as
Iρ
u= (A1)
2πr
Assuming the two current electrodes at the surface, the electrical potential can be
written as  
Iρ 1 1
u= − (A2)
2π r1 r2
where I is the electrical current, r is the distance between the current and the potential
electrode, and ρ is the resistivity value of the medium.

Appendix C. Apparent Resistivity Formula of Two-Layer Resistivity Model

For a two-layer resistivity model with the Wenner array, the measured potential
difference can be written as
 
∞ ∞
4km 4km
 
Iρ1  
∆u = 1+ ∑ h i1/2 − ∑ h i1/2  (A3)
2πa 
 m=1 1 + (2mz/a )2 m=1 4 + (2mz/a )2 

So the apparent resistivity is

 
∞ ∞
4km 4km

 

ρ a = ρ1 1 + ∑ h i1/2 − ∑ h i1/2  (A4)
m=1 1 + (2mz/a )2 m=1 4 + (2mz/a )2

 

ρ −ρ
where k = ρ22 +ρ1 . In this expression, k is a reflection coefficient whose value lies within ±1,
1
depending on the relative resistivities in the two media.
Mathematics 2024, 12, 269 21 of 22

References
1. Mosaad, A.H.; Farag, M.M.; Wei, Q.; Fahad, A.; Mohamed, S.A.; Hussein, A.S. Integration of electrical resistivity tomography and
induced polarization for characterization and mapping of (Pb-Zn-Ag) sulfide deposits. Minerals 2023, 13, 986. [CrossRef]
2. Mitchell, M.A.; Oldenburg, D.W. Using DC resistivity ring array surveys to resolve conductive structures around tunnels or
mine-workings. J. Appl. Geophys. 2022, 211, 104949. [CrossRef]
3. Oldenburg, D.W.; Li, Y.; Ellis, R.G. Inversion of geophysical data over a copper gold porphyry deposit: A case history for Mt.
Milligan. Geophysics 1997, 62, 1419–1431. [CrossRef]
4. Chambers, J.C.; Kuras, O.; Meldrum, P.I.; Ogilvy, R.D.; Hollands, J. Electrical resistivitytomography applied to geologic,
hydrogeologic, and engineering investigations at a former waste-disposal site. Geophysics 2006, 71, 231–239. [CrossRef]
5. Rucker, D.; Loke, M.H.; Levitt, M.T.; Noonan, G.E. Electrical resistivity characterization of an industrial site using long electrodes.
Geophysics 2010, 75, 95–104. [CrossRef]
6. Kim, J.H.; Tsourlos, P.; Karmis, P.; Vargemezis, G.; Yi, M.J. 3D inversion of irregular gridded 2D electrical resistivity tomography
lines: Application to sinkhole mapping at the Island of Corfu (West Greece). Near Surf. Geophys. 2018, 14, 275–285. [CrossRef]
7. Plank, Z.; Polgar, D. Application of the DC resistivity method in urban geological problems of karstic areas. Near Surf. Geophys.
2019, 17, 547–561. [CrossRef]
8. Sirota, D.; Shragge, J.; Krahenbuhl, R.; Swidinsky, A.; Yalo, N.; Bradford, J. Development and validation of a low-cost direct
current resistivity meter for humanitarian geophysics applications. Geophysics 2022, 87, 1–4. [CrossRef]
9. Zhou, B.; Greenhalgh, M.; Greenhalgh, S.A. 2.5-D/3-D resistivity modelling in anisotropic media using Gaussian quadrature
grids. Geophys. J. Int. 2009, 176, 63–80. [CrossRef]
10. Zhang, H.; Liu, Y.; Yang, X. An efficient ADI difference scheme for the nonlocal evolution problem in three-dimensional space.
J. Appl. Math. Comput. 2023, 69, 651–674. [CrossRef]
11. Zhou, Z.; Zhang, H.; Yang, X. H1-norm error analysis of a robust ADI method on graded mesh for three-dimensional subdiffusion
problems. Numer. Algorithms 2023, 94, 1–19. [CrossRef]
12. Yang, X.; Wu, L.; Zhang, H. A space-time spectral order sinc-collocation method for the fourth-order nonlocal heat model arising
in viscoelasticity. Appl. Math. Comput. 2023, 457, 128192. [CrossRef]
13. Tian, Q.; Yang, X.; Zhang, H.; Xu, D. An implicit robust numerical scheme with graded meshes for the modified Burgers model
with nonlocal dynamic properties. Comput. Appl. Math. 2023, 42, 246. [CrossRef]
14. Mufti, I.R. Finite-difference resistivity modeling for arbitrarily shaped two-dimensional structures. Geophysics 1976, 41, 62–78.
[CrossRef]
15. Vachiratienchai, C.; Boonchaisuk, S.; Siripunvaraporn, W. A hybrid finite difference-finite element method to incorporate
topography for 2D direct current (DC) resistivity modeling. Phys. Earth Planet. Interiors 2010, 183, 426–434. [CrossRef]
16. Gernez, S.; Bouchedda, A.; Gloaguen, E.; Paradis, D. AIM4RES, an open-source 2.5D finite difference MATLAB library for
anisotropic electrical resistivity modeling. Comput. Geosci. 2020, 135, 104401. [CrossRef]
17. Jahandari, H.; Lelièvre, P.; Farquharson, C.G. Forward modeling of direct-current resistivity data on unstructured grids using an
adaptive mimetic finite-difference method. Geophysics 2023, 88, 123–134. [CrossRef]
18. Suryavanshi, D.; Dehiya, R. A mimetic finite-difference method for two-dimensional DC resistivity modeling. Math. Geosci. 2023,
55, 1189–1216. [CrossRef]
19. Zhou, B.; Greenhalgh, S.A. Finite element three-dimensional direct current resistivity modelling: Accuracy and efficiency
considerations. Geophys. J. Int. 2001, 145, 679–688. [CrossRef]
20. Pan, K.J.; Tang, J. 2.5-D and 3-D DC resistivity modelling using an extrapolation cascadic multigrid method. Geophys. J. Int. 2014,
197, 1459–1470. [CrossRef]
21. Chou, T.K.; Chouteau, M.; Dubé, J.S. Intelligent meshing technique for 2D resistivity inverse problems. Geophysics 2016, 81, 45–56.
[CrossRef]
22. Yan, B.; Li, Y.; Liu, Y. Adaptive finite element modeling of direct current resistivity in 2-D generally anisotropic structures. J. Appl.
Geophys. 2016, 130, 169–176. [CrossRef]
23. Ren, Z.Y.; Qiu, L.; Tang, J. 3D direct current resistivity anisotropic modelling by goal-oriented adaptive finite element methods.
Geophys. J. Int. 2018, 212, 76–87. [CrossRef]
24. Doyoro, Y.G.; Chang, P.Y.; Puntu, J.M.; Puntu, J.M.; Lin, D.J.; Huu, T.V.; Rahmalia, D.A.; Shie, M.S. A review of open software
resources in python for electrical resistivity modelling. Geosci. Lett. 2022, 9, 3. [CrossRef]
25. Pidlisecky, A.; Knight, R. FW2_5D: A MATLAB 2.5-D electrical resistivity modelling code. Comput. Geosci. 2008, 34, 1645–1654.
[CrossRef]
26. Ma, C.; Liu, J.; Liu, H.; Guo, R.; Musa, B.; Cui, Y. 2.5D electric resistivity forward modeling with element-free Galerkin method.
J. Appl. Geophys. 2019, 162, 47–57. [CrossRef]
27. Xu, S.Z.; Zhou, H. Modelling the 2D terrain effect on MT by the boundary-element method. Geophys. Prospect. 1997, 45, 931–943.
[CrossRef]
28. Dey, A.; Morrisson, H.F. Resistivity modeling for arbitrarily shaped three-dimensional structures. Geophysics 1979, 44, 753–780.
[CrossRef]
29. Liu, J.; Liu, P.; Tong, X. Three-dimensional land FD-CSEM forward modeling using edge finite-element method. J. Cent. South
Univ. 2018, 25, 131–140. [CrossRef]
Mathematics 2024, 12, 269 22 of 22

30. Chen, J.; Haber, E.; Oldenburg, D.W. Three-dimensional numerical modelling and inversion of magnetometric resistivity data.
Geophys. J. Int. 2002, 149, 679–697. [CrossRef]
31. Pan, K.; Wang, J.; Hu, S.; Ren, Z.; Cui, T.; Guo, R.; Tang, J. An efficient cascadic multigrid solver for 3-D magnetotelluric forward
modelling problems using potentials. Geophys. J. Int. 2022, 230, 1834–1851. [CrossRef]
32. Xu, S.Z.; Duan, B.C.; Zhang, D.H. Selection of the wavenumbers k using an optimization method for the inverse Fourier transform
in 2.5D electrical modeling. Geophys. Prospect. 2000, 48, 789–796. [CrossRef]
33. Pan, K.; Zhang, Z.; Hu, S.; Ren, Z.; Guo, R.; Tang, J. Three-dimensional forward modelling of gravity field vector and its gradient
tensor using the compact difference schemes. Geophys. J. Int. 2021, 224, 1272–1286. [CrossRef]
34. Telford, W.M.; Geldart, L.P.; Sheriff, R.E. Applied Geophysics; Cambridge University Press: Cambridge, UK, 1990.
35. Loke, M.H.; Barker, R.D. Rapid least-squares inversion of apparent resistivity pseudosections by a quasi-Newton method. Geophys.
Prospect. 1996, 44, 131–152. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.