Machining Science and Technology

An International Journal

ISSN: 1091-0344 (Print) 1532-2483 (Online) Journal homepage: http://www.tandfonline.com/loi/lmst20

A simplified mathematical model development

for the design of free-form cathode surface in
electrochemical machining

Hasan Demirtas, Oguzhan Yilmaz & Bahattin Kanber

To cite this article: Hasan Demirtas, Oguzhan Yilmaz & Bahattin Kanber (2017) A
simplified mathematical model development for the design of free-form cathode surface
in electrochemical machining, Machining Science and Technology, 21:1, 157-173, DOI:
10.1080/10910344.2016.1275192

To link to this article: http://dx.doi.org/10.1080/10910344.2016.1275192

Published online: 31 Jan 2017.

Submit your article to this journal

Article views: 102

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=lmst20

Download by: [Australian Catholic University] Date: 27 July 2017, At: 02:05
MACHINING SCIENCE AND TECHNOLOGY
, VOL. , NO. , –
http://dx.doi.org/./..

A simplified mathematical model development for the

design of free-form cathode surface in electrochemical
machining
Hasan Demirtasa , Oguzhan Yilmazb , and Bahattin Kanberc
a
Vocational School, Kilis  Aralık University, Kilis, Turkey; b Department of Mechanical Engineering, Gazi
University, Ankara, Turkey; c Department of Mechanical Engineering, Bursa Technical University, Bursa,
Turkey

ABSTRACT KEYWORDS
High-performance machining of free-form surfaces is highly criti- Cathode design; ECM;
cal in automotive, aerospace, and die–mold manufacturing indus- free-form surface; surface
tries. Therefore, electrochemical machining (ECM) process has quality
been used in such cases in that sense. The most important chal-
lenges of using ECM process are the lack of accuracy and difficulty
in designing proper machining tool (cathode) surfaces. In this arti-
cle, a simplified mathematical model is presented to obtain a cath-
ode surface for ECM of free-form surfaces which have high cur-
vatures. In this theoretical approach, the finite-element method
(FEM) is used to solve the 3-D Laplace equation and to deter-
mine the potential distribution between the anode (workpiece)
and cathode (tool) surfaces. A compact and simple program was
developed to obtain a proper cathode surface that only requires
some nodal coordinates on the anode surface and boundary con-
ditions. In this work, a trial cathode surface is constructed for a
given gap distance. For the determined ECM parameters, cathode
shape that satisfies the boundary conditions is obtained for the
45th layer. The results are compared with the literature and ANSYS
Workbench for verification. The developed theoretical approach
benefits simpler and faster FEM solutions, accurate cathode sur-
face, and consequently correct form of machined surface.

Introduction
Electrochemical machining (ECM) is a nontraditional machining process which is
theoretically based on Faraday’s and electrolysis laws. According to these laws, the
electron flow from one electrode to another (connected to a direct current power
supply) is performed with the help of an electrical conductive liquid. In the ECM
process, workpiece and tool (electrode) are charged as positive and negative, respec-
tively. The tool is called a cathode and the workpiece is called an anode. In the
machining mechanism, the material is removed from the anode without a contact
between the anode and cathode. The electrolyte is pumped through the machining

CONTACT Hasan Demirtas hdemirtas@kilis.edu.tr Kilis  Aralik University, Mehmet Sanlı Mah., Dogan Gures
Pasa Bul., No , , Kilis, Turkey.
©  Taylor & Francis Group, LLC
158 H. DEMIRTAS ET AL.

gap between the anode and cathode, while direct current is passed through the gap
at a low voltage to dissolve metal from the anode. ECM is more advantageous than
conventional machining, since it can be applicable regardless of material hardness,
tool wear, and cutting forces in some cases. In fact, ECM is very useful for produc-
ing a bright surface finish, machining difficult-to-cut materials, and manufacturing
complex geometry components (Rajurkar et al., 1999). The application of the ECM
process is sometimes limited due to the troubles encountered with peripheral fac-
tors, cathode design, and process monitoring. Among them, the cathode design is
the most important challenging factor that must be overcome. The critical point in
the cathode design is to obtain an optimum cathode surface geometry for a speci-
fied anode surface geometry under the condition of changes in surface normalities,
electrolyte flow rate, type, current density, and material.
Various methods have been applied for the cathode design in ECM process,
such as the boundary-element method (BEM), finite-difference method, and finite-
element method (FEM). FEM is applied more frequently for free-form surfaces than
the others (Mount et al., 2003; Zhitnikov et al., 2004; Sun et al., 2006). Hocheng et al.
(2003) used the iteration integral method for a 2-D cathode design. This model is
based on electric field theory. The errors induced by the model make it inappropri-
ate for short machining times, but the errors decrease with time. Li and Ji (2010)
investigated the model error difference between the 2-D and 3-D Cos θ method
and showed that the 3-D Cos θ method induced fewer errors than the 2-D Cos θ
method. However, this model is based on geometric rules and does not indicate
the ECM parameters (except the cathode feed rate). Kozak et al. (1998) used the
electrochemical shaping theory to make a computer simulation for electrochemical
shaping. The model was developed with considerations of some ECM parameters,
such as electrochemical machinability, applied voltage, electrical conductivity, and
so forth, but the important parameters, such as initial gap distance and cathode feed
rate, were not taken into account. Some studies have utilized BEM to design com-
puter simulations for ECM (Purcar et al., 2004, 2008). These models, however, were
generally suitable only for simple surface forms. Pattavanitch et al. (2010) offered
an ECM process simulation by BEM, and Li and Ji (2009) developed a model to
machine aeroengine blades through the ECM process. The latter model is based on
a BP neural network in which the inputs are the applied voltage, initial machining
gap, feed rate of the cathode, pressure difference between the inlet and outlet of the
electrolyte and electrolyte temperature. ECM cathode designs have also been made
by FEM (Sun et al., 2006; Li and Niu, 2007; Wang and Zhu, 2009). Sun et al. (2006)
used a nonuniform rotational B-spline surface (NURBS) model to explain the work-
piece surface. In this model, the cathode points are found using electrochemical law.
The potential distribution from anode to cathode is described by a Laplace equa-
tion. The Laplace equation is solved using FEM to find the gap which will be used to
obtain the cathode coordinates. But the procedure of the cathode design is too long
and unclear for practical usage.
In this article, a simplified mathematical model is presented for the designing of a
free-form cathode surface in ECM process using FEM. A dedicated FEM algorithm
®
was developed and implemented in Mathematica . The developed mathematical
 MACHINING SCIENCE AND TECHNOLOGY 159

Figure . Mapping of a NURBS surface. Note: NURBS, nonuniform rotational B-spline surface.

model is able to apply the symbolic and numerical computation as a hybrid method
to solve the 3-D Laplace equation with high precision and less computing time. Fur-
thermore, to reduce the CPU time, nodal coordinates were transferred from local
coordinates to a natural coordinate system. To increase the precision of the FEM
results, mesh quality factors, such as skewness ratio and aspect ratio, were consid-
ered. The developed method deals with the control points of free-form surfaces
modeled through NURBS, which are obtained with a 3-D computer-aided design
(CAD) program, and these points are imported into Mathematica . The developed®
method simply uses these control points to trail the cathode shape, with consider-
ation of the electric potential distribution between the anode and cathode surfaces,
which are modeled using a 3-D Laplace equation and an FEM solution.

Mathematical modeling of the cathode surface

In CAD and graphics, NURBS is the most popular tool for defining curves and
surfaces. It can be used for many applications from manufacturing industry to film
industry (Rogers, 2001). Due to capabilities of NURBS modeling for free-form
surfaces, NURBS surfaces have been widely adapted in CAD/CAM software and
graphics applications (Piegl and Tiller, 1997). A NURBS surface includes an intrinsic
rational piecewise polynomial mapping from the 3-D surface to the 2-D parameter
domain as shown in Figure 1. Therefore, NURBS modeling is preferred for describ-
ing the anode and cathode surfaces in this work. Based on this, anode surface is con-
structed and NURBS control points are obtained with using a 3-D surface modeling
software. To construct the trail cathode geometry in Mathematica , these NURBS®
control points are used. To obtain the optimum cathode geometry that satisfies the
Faraday and Ohm’s laws, a model is developed using FEM. To decrease analysis
time, natural coordinate system is used during construction of shape functions.

Modeling of workpiece (anode) surface

Free-form surfaces have been widely used in aerospace, automotive, and the
die/mold industries. They are generally used to improve the functional require-
ments. The importance of them is increasing with advancing technology and thus
NURBS is the best way to describe these surfaces. Therefore, for describing the
160 H. DEMIRTAS ET AL.

anode surface, NURBS modeling can be expressed as follows:

m n
i=0 j=0 wi j Pi j Ti,k (u) T j,l (v )
S (u, v ) = m n , (1)
i=0 j=0 wi j Ti,k (u) T j,l (v )

where Pi j are the three-dimensional control net vertices, wi j is the weight of the Pi j , u
and v are the biparametric directions, Ti,k (u) and T j,l (v ) are the nonrotational basis
functions in the biparametric u and v directions, and k and l are orders of B-Spline
basis functions.
Using the surface control points, NURBS surface can be plotted in Mathematica , ®
but surface function S(u, v ) cannot be obtained directly. Therefore, all the NURBS
surface function components must be defined clearly. wi j describe the point weight
and for a point that is on the surface, it is equal to 1. Ti,k (u) and T j,l (v ) are the
complex functions and changes with the knots and degrees of the NURBS surface
curves in u and v directions. For obtaining the knots and curve degrees, a com-
mand that is called “BSplineFunction” was used. It represents a B-spline function for
a curve defined by the control points. After defining the knots and curve degrees,
to obtain the nonrotational basis functions Ti,k (u) and T j,l (v ), “PiecewiseExpand”
and “BSplineBasis” commands are used. “PiecewiseExpand” expands nested piece-
wise functions in expression to give a single piecewise function, and “BSplineBasis”
gives the kth nonuniform B-spline basis function of degree d with knots at positions
ui . “PiecewiseExpand” can be used to expand symbolic BSplineBasis functions into
explicit piecewise polynomials.
Due to NURBS surface function S(u, v ), the unit normal theorem can be used to
obtain the function of normal vector for the free-form surface as follows;
Su (u, v ) xSv (u, v )
n (u, v ) = , (2)
|Su (u, v ) xSv (u, v )|
where Su (u, v ) and Sv (u, v ) are the derivatives of the S(u, v ) along u and v direc-
tions. As can be seen from Eq. (2), the normal vector n is a function of u and v and
to obtain its numerical values at all points, u and v should have a value between
one and zero due to the biparametric coordinates of the model as can be seen in
Figure 2, (0 ≤ u ≤ 1, 0 ≤ v ≤ 1). The aim of this is to describe the n function for
every point due to point coordinates in biparametric coordinate system. According
to the electrochemical law, the trail the coordinates of the cathode surface control
points can be obtained due to anode surface control points is as follows:

xc = xa − b. sin θ . sin α (3)

yc = ya − b. sin θ . cos α (4)
zc = za + b (5)

where b describes the distance along z direction between the anode and cath-
ode, xa , ya , za are the coordinates of the anode surface control point, θ is the angle
between the feeding direction of the cathode and normal to the anode, and α is the
 MACHINING SCIENCE AND TECHNOLOGY 161

Figure . Representation of the biparametric coordinates for  ×  points along u and v directions.

Figure . The angles due to normal vector.

angle between normal vector’s project on x − y plane and y direction, as shown in

Figure 3.
As can be seen from Figure 3, θ and α angles associated with the components of
the normal vector can be described as follows:
nx
α = cos−1 (6)
ny

−1 nx + ny
2 2
θ = tan (7)
nz

Finite-element modeling of cathode surface

Flow between anode and cathode was considered as a solid layer for meshing and
numbering as shown in Figure 4. To create such a solid layer, the flow within the
162 H. DEMIRTAS ET AL.

Figure . Solid model of the electrolyte between the anode and cathode.

gap had to be distributed in a uniform manner, and this is only possible in laminar
flow.
V.LC .ρ
Re = (8)
μ
Equation 8 describes the Reynolds number that is used to characterize the differ-
ent flow regimes within similar fluids, such as laminar and turbulent flow, where Re
is the Reynolds number, V is the upstream velocity, LC is the characteristic length
of the geometry, ρ is the density, and μ is the dynamic viscosity. For the same elec-
trolyte velocity of the fluid (V), the flow rate defines the flow regimes and therefore
the flow rate of the electrolyte is considered to be constant.
To avoid the effects of over potential and to make Faraday’s law applicable for
obtaining the material removal rate, some assumptions must be made. These are
(i) reaction rates of the electrode are very fast, (ii) to prevent the temperature gradi-
ent and concentration within the electrode gap, the electrolyte solution is completely
mixed, (iii) machining is made by the current at the anode surface without any ECM
(Zhou and Derby, 1995). In the most ECM processes, the cathode is fed at constant
speed toward the anode so that the anode surface maintains its speed as constant.
In this situation, the distance between the anode and cathode can be accepted as a
time-independent value. Therefore, ECM system is acceptable as a quasisteady-state
and free boundary problem (Hardisty et al., 1993; Domanowski and Kozak, 2001;
Sun et al., 2006; Li and Niu, 2007; Wang and Zhu, 2009). According to aforemen-
tioned assumptions, the electric potential distribution inside the gap domain can be
expressed by Laplace’s equation;
∂ 2ϕ ∂ 2ϕ ∂ 2ϕ
+ 2 + 2 =0 (9)
∂x2 ∂y ∂z
Due to Faraday’s and Ohm’s laws, anode and cathode boundaries are shown as
follows:

∂ϕ Vf
= ∗ Cos θ On Anode (10)
∂n kv ∗ kc
ϕA = U } On Cathode (11)
ϕC = 0} On Anode (12)

where ϕ describes the electric potential at each point on the anode, cathode surfaces
and layers of the gap. U is the applied voltage, kc is the electrical conductivity of the
electrolyte, which is directly affected by the electrolyte temperature and electrolyte
 MACHINING SCIENCE AND TECHNOLOGY 163

Figure . Effect of the temperature on the electrical resistivity of the NaCl electrolyte for different
concentrations (ASM Handbook, ).

concentration. The conductivity of an electrolyte changes substantially with tem-

perature. Figure 5 shows the influence of temperature on the electrical resistivity of
the NaCl electrolyte for different concentrations (ASM Metal Handbook, 1989). Any
modification of the electrolyte temperature changes the electrolyte resistivity or con-
ductivity. Based on Eq. (10), electrolyte conductivity was considered to be constant,
and therefore the electrolyte temperature remained constant at room temperature.
Vf is the feed rate of cathode along z direction, kv is the electrochemical machinabil-
ity of the anode material, and θ is the angel between the movement direction of the
cathode and normal to the anode. To satisfy the boundary conditions for the anode
surface, Laplace equation can be minimize based on variation principle and can be
explained as follows:
    2  2  2 
1 ∂ϕ ∂ϕ ∂ϕ ∂ϕ
G (ϕ) = + + dxdydz − ϕds (13)
2 V ∂x ∂y ∂z ∂n

A potential function ϕ(x, y, z), which varies linearly inside each hexahedron ele-
ment is defined as;
8
ϕ x, y, z = Ni x, y, z .ϕi (14)
i=1

where ϕi is electrical potential for each node and Ni is the shape function for each
node and for trilinear hexahedron element in global coordinate system. Substitution
164 H. DEMIRTAS ET AL.

Figure . Hexahedron with  nodes and the mapped cube.

of Eq. (14) into Eq. (13) and then minimization of G(ϕ) give the element stiffness
matrix (Sun et al., 2006);
           
 ∂Nie ∂N ej ∂Nie ∂N ej ∂Nie ∂N ej
K [e] [i] j = + + dV,
Ve ∂x ∂x ∂y ∂y ∂z ∂z
(15)
where K is the element stiffness matrix, e is the element number. Solution of Eq. (15)
requires very long CPU time. Therefore, in this study, numerical solution is made in
natural coordinate system as an isoparametric element to decrease the analysis time.
This method has been used for solving two- and three-dimensional finite-element
problems with great success (Rao, 2011). With using a natural (or intrinsic) coordi-
nate system ξ , η, ζ that is defined by element geometry, the isoparametric element
equations are formulated. In other words, ξ , η, ζ are attached to the solid element
to describe the length of the element along the axial coordinate. For each element
of a specific structure, there is a relationship between the natural coordinate sys-
tem ξ , η, ζ and the global coordinate system x, y, z which must be used in the ele-
ment equation formulations (Gu and Gennert, 1991). The reason of choosing these
particular limits is to simplify the Gaussian quadrate formula for the hexahedron
elements.
Type of FEM is chosen as hexahedron element which has six faces and 8 nodes.
Figure 6 shows the mapped cube of a hexahedron element. In Figure 6, ξ goes from
face 3267 to face 4158, ζ goes from face 4378 to face 1265 and η goes from face 1234
to face 5678. The intersection of the two medians is the center of a face. Node coor-
dinates in natural coordinate system change between −1 and 1 as usual. According
to these data, the shape functions of this element can be explained in Eq. (16).
1
Ni = (1 + ξ ξi ) (1 + η ηi ) (1 + ζ ζi ) , (16)
8
where ξi , ηi , and ζi denote the coordinates of ith node in natural coordinates. The
element stiffness matrix in Eq. (15) can be written as follows:
 1 1 1
K [e] = Be .Be T . det [J] dV (17)
−1 −1 −1

where “e” is the element number and det[J] is the determinant of the Jacobian
matrix, and Jacobian describes the relation between the length of an element in the
 MACHINING SCIENCE AND TECHNOLOGY 165

global coordinate system to the length of an element in the natural coordinate sys-
tem. In general, [ J ] is a function of ξ , η, ζ and depends on the numerical values
of the nodal coordinates (Logan, 2007). According to the polynomial completeness
theory, Jacobian matrix is shown as follows:
⎡ ∂X ∂Y ∂Z ⎤
e e e
∂ξ ∂ξ ∂ξ
⎢ ∂X ⎥
⎢
J = ⎣ ∂ηe ∂Ye ∂Ze ⎥ (18)
∂η ∂η ⎦
∂Xe ∂Ye ∂Ze
∂ζ ∂ζ ∂ζ

where e is the element number, Xe , Ye , Ze are the transformation functions for

mapping the hexahedron element from local coordinate system to natural coordi-
nate system. Be is the matrix that includes derivative of shape functions and for one
degree of freedom, it can be written as;
⎡ ∂N e ∂N8e
⎤
∂x
1
· · · · · · · · · ∂x
⎢ ⎥
⎢ ∂N1e .. ∂N8e ⎥
B = ⎢ ∂y · · ·
e
. · · · ∂y ⎥ (19)
⎣ ⎦
∂N1e ∂N e
∂z
· · · · · · · · · ∂z8
∂N e ∂N e ∂N e
where ∂x1 , ∂y1 , ∂z1 are the derivatives of shape functions, and they can be obtained
using the inverse of the Jacobian matrix for each element as follows:
⎡ e⎤ ⎡ ∂N e ⎤
∂Ni i

⎢ e⎥∂x
⎢ ∂ξ e ⎥
⎢ ∂Ni ⎥ −1 ⎢ ∂Ni ⎥
⎢ ∂y ⎥ = J ⎢ ∂η ⎥ (20)
⎣ ⎦ ⎣ e⎦
∂Nie ∂Ni
∂z ∂ζ

By assembling all element stiffness matrices, global stiffness matrix can be

obtained. The correlation between the potential and global stiffness matrix is shown
as follows;
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
k1 1 k2 1 0 ... 0 ϕ1 C
⎢ k1 2 k2 2 . . . ... . . . ⎥ ⎢ ϕ2 ⎥ ⎢ 0 ⎥
⎥ ⎢ ⎥ ⎢
⎢ ⎥
⎢ 0 ... ⎥ ⎢ ⎥ ⎢ ⎥
⎢ ... ... ... ⎥ . ⎢ . . . ⎥ = ⎢. . .⎥ (21)
⎣ . . . . . . . . . kt−1 t−1 kt−1 t ⎦ ⎣ . . . ⎦ ⎣. . .⎦
0 . . . . . . kt t−1 kn n ϕt 0

where t is the layer number, k is the symmetric stiffness matrix, ϕ1 and ϕn are the
electric potentials on anode and cathode surfaces, respectively. ϕ2 − ϕt−1 are the
electric potentials at layers that are divided in the gap domain. C is the column vector
that is shown as follows:
⎡ ⎤
c1
⎢ c2 ⎥
⎢ ⎥
C=⎢ ⎢ c3 ⎥
⎥ (22)
⎣. . .⎦
cm
166 H. DEMIRTAS ET AL.

where m is the node number at anode surface and cm is the surface area integration
of Eq. (10) and can be written as:
∂ϕ
cm = .S (23)
∂n
where “S” describes the surface area of a hexahedron element. For a free-form sur-
face, it can be calculated as follows:
 
S= |Su (u, v ) x Sv (u, v )| du dv (24)
su sv

where su and sv are the boundaries for an element along u and v directions. Accord-
ing to the conditions that are discussed above, the Eq. (21) can be solved as follows:

ϕ2 = −k−1
12 . (k11 .ϕ1 − C) (25)
ϕi+1 = −k−1
i i+1 . (ki i−1 .ϕi−1 + ki i .ϕi ) i>1 (26)

ki j are the matrix component of the global stiffness matrix for each layer and i and
j are the layer number. For one element and one layer, k11 , k21 , k12 , k22 are shown in
Eq. (27). From Eqs. (25) and (26), the potential distribution inside the gap domain
can be obtained.

As can be seen from Eq. (27), stiffness matrix components are written as a 3-D
array due to advantages like easy to understand, searching and sorting the elements,
and easy to type the algorithm. First dimension of “K” describes the element num-
ber, second dimension is the column, and third dimension is the row of stiffness
matrix.

Case study: A free-form cathode surface design

In this section, the developed mathematical model is implemented for a simple free-
form surface to create a corresponding cathode surface. Anode surface, shown in
Figure 7, was modeled using the control points of the free-form surface. To obtain
FEM results, the points on the surface were obtained with using a 3-D surface mod-
eling software. Mesh quality has a significant role to get good accuracy and stability
for the numerical computation. In this study, the aspect ratio and skewness of the
 MACHINING SCIENCE AND TECHNOLOGY 167

Figure . Sample free-form surface.

elements are chosen to check the mesh quality. Skewness can be defined as follows:
 
γmax − γe γe − γmin
Max , (28)
180 − γe 180 − γe
where γmax is the largest angle in the cell, γmin is the smallest angle in cell, and γe
is the angle for an equiangular cell (90° for hexahedron element). Table 1 gives the
cell quality and the corresponding range of skewness values. The range of skewness
value of this work is 0.0108 < γ < 0.012.
The other main factor that affects the mesh quality is the aspect ratio. An ele-
ment aspect ratio is the ratio of its maximum to its minimum width. To obtain good
numerical results, aspect ratio must be near to 1. The aspect ratio range of this study
is taken between 1.00013 ≤ AR ≤ 1.032. Due to these rules, anode surface and the
mesh points are shown in Figure 8.
The anode surface mesh points that have good mesh quality are imported from
®
a 3-D modeling software to Mathematica . With the help of the mesh points, sur-
face knot vectors along u and v directions, B-Spline basis functions, NURB surface
model S(u, v ) are obtained, and normal vectors of all points are calculated through
S(u, v ). After all these calculations, layer points are determined for a given gap dis-
tance that is shown in Table 2 and Eqs. (3)–(5).

Table . Skewness values for mesh quality.

Value of skewness Excellent Good Acceptable Poor Sliver Degenerate

Mesh quality –. .–. .–. .–. .–. .–.

Table . ECM process parameters.

Electrical Electrochemical
conductivity (k) machinability Scalar potential Feed rate (Vf ) Equilibrium gap
mm3
Electrolyte type (mS/cm) (kv ) ( 100A.min ) on anode (U) (V) (mm/min) (mm)

NaNO .    .

ECM, electrochemical machining.

168 H. DEMIRTAS ET AL.

Figure . Free-form surface with mesh points.

For obtaining the cathode geometry with the boundary conditions in Eqs. (9)–
(11), the mathematical model is obtained. The FEM parameters and ECM param-
eters are shown in Tables 2 and 3. According to the boundary conditions of this
model, the potential of the cathode surface must be zero. Due to this, the potential
distribution must be obtained for all layers. In this study, fifty layers are used.
After obtaining the layer points that are discussed above, the transformational
calculations (from global coordinate system to natural coordinate system) and
assembling of global stiffness matrix are made by the developed program. The pro-
gram can be used for all surface types. For this example, the cathode surface that
satisfies the boundary conditions is the 45th layer, and the potential distribution is
shown in Table 4. For manufacturing the cathode, calculated surface points can be
®
exported as a .dxf file format by Mathematica and solid model of the cathode can
be obtained as shown in Figure 9.
To compare the results of this study with the literature data, the model which is
developed by Li and Diu (2016) is improved by adding anode and cathode layers
in ANSYS Workbench as shown in Figure 10. During ECM process, the electrolyte
flows between anode and cathode. Therefore, electrolyte geometry does not per-
fectly fit to the geometry of anode and cathode. To investigate the effect of geometric
idealization, second finite-element model is developed by extending the electrolyte
as shown in Figure 11.
The analysis type is selected as Electric and more than 400 thousands hexahedral
solid elements are used in both models. The distribution of voltage potential along

Table . FEM parameters for the freeform surface.

Number of layer Element type Number of nodes Number of elements

 Hexahedron , ,

FEM, finite-element method.

 MACHINING SCIENCE AND TECHNOLOGY 169

Table . Potential distribution inside the gap domain.

Point number
Layer number          

st layer          
nd layer − . − . − . − . − . − . − . − . − . − .
rd layer − . − . − . − . − . − . − . − . − . − .
th layer − . − . − . − . − . − . − . − . − . − .
th layer − . − . − . − . − . − . − . − .  − .
th layer − . − . − . − . − . − . − . − . − . − .
th layer − . − . − . − . − . − . − . − . − . − .
th layer − . − . − . − . − . − . − . − . − . − .
th layer − . − . − . − . − . − . − . − . − . − .
th layer − . − . − . − . − . − . − . − . − . − .
th layer − . − . − . − . − . − . − . − . − . − .

the electrolyte thickness in normal direction is compared in Figure 12. It is shown

that the linearity in potential distribution is changed when the finite-element model
is used with extended electrolyte. So, an extended model can give more realistic
cathode design.

Figure . Solid model of cathode surface.

Figure . Finite element model of anode, cathode, and electrolyte developed in ANSYS (Model ).
170 H. DEMIRTAS ET AL.

Figure . Finite element model of anode, cathode, and extended-electrolyte developed in ANSYS
(Model ).

Figure . Potential distribution along the electrolyte thickness.

Conclusion
A mathematical model was developed to design a cathode surface for ECM of free-
form surfaces. The main aim of this work is to obtain an accurate cathode surface
that satisfies the Laplace equation to account for some boundary conditions that
arise when using FEM in ECM. Thus, a computer program was developed for easy
and practical usage in solving the encountered equations. A case study was con-
ducted for a free-form surface, and the cathode surface coordinates were obtained.
Two different ANSYS Workbench models were used considering with the anode
and cathode surfaces. This work was also verified by comparing the results with the
literature (considering the linearity situation) and ANSYS Workbench. It was also
shown that a cathode surface can easily be obtained using zero or near zero potential
points that were obtained from the results of the developed program. The developed
theoretical model proposes robust FEM solutions, best-fit cathode surface, and it
will lead to machine correct form of free-form surfaces. Experimental and work-
shop applications are being conducted and the verification tests are to be future
works.
 MACHINING SCIENCE AND TECHNOLOGY 171

Funding
This work was supported by the Gaziantep University Scientific Research Project (BAP) Depart-
ment under the grant number MF.14.19.

Nomenclature
C Column vector for anode surface control points
e Element number
J Jacobian matrix
K Stiffness matrix
k, l Orders of B-Spline basis functions
ki i Global stiffness matrix component
kc Electrical conductivity of the electrolyte
kv Electrochemical machinability of the anode material
LC Characteristic length of the geometry
m Node number
N Shape function
n Normal vector of control points
Pi j Three-dimensional control net vertices
Re Reynolds number
S Surface area of a hexahedron element
S(u, v ) NURBS surface function
Su (u, v ) Derivative of S(u, v ) along u direction
Sv (u, v ) Derivative of S(u, v ) along v direction
T Nonrotational basis function
t Layer number
U Applied voltage
u, v The biparametric directions
V Upstream velocity of the fluid
Vf Feed rate of cathode
wi j The weight of the Pi j
Xe , Ye , Ze Transformation functions for mapping the hexahedron element
xa Coordinate of control points on anode surface along x coordinates
xc Coordinate of control points on trial cathode surface along x coordi-
nates
ya Coordinate of control points on anode surface along y coordinates
yc Coordinate of control points on trial cathode surface along y coordi-
nates
za Coordinate of control points on anode surface along z coordinates
zc Coordinate of control points on trial cathode surface along z
coordinates
α Angle between the normal vector’s project on x − y plane and y
direction
172 H. DEMIRTAS ET AL.

γ Skewness ratio of a hexahedron element

θ Angle between the z direction and normal vector
μ Dynamic viscosity of the fluid
ξ , η, ζ Global coordinate system directions
ρ Density of the fluid
ϕ Electric potential

References
ASM International Handbook Committee. (1989) ASM Handbook Volume 16: Machining. ASM
International, Metals Park, OH.
Domanowski, P.; Kozak, J. (2001) Inverse problems of shaping by electrochemical generating
machining. Journal of Materials Processing Technology, 109(3): 347–353.
Gu, G.G.; Gennert, M.A. (1991) Boundary element methods for solving poisson equations
incomputer vision problems. IEEE Computer Society Conference, Maui, HI.
Hardisty, H.; Mileham, A.R.; Shirvarni, H. (1993) A finite element simulation of the electrochem-
ical machining process. CIRP Annals, 42(1): 201–204.
Hocheng, H.; Sun, Y.H.; Lin, S.C.; Kao, P.S. (2003) A material removal analysis of electrochemi-
cal machining using flat-end cathode. Journal of Materials Processing Technoloygy, 140(1–3):
264–268.
Kozak, J.; Budzynski, A.F.; Domanowski, P. (1998) Computer simulation electrochemical shap-
ing (ECM-CNC) using a universal tool electrode. Journal of Materials Processing Technology,
76(1–3): 161–164.
Li, Z.; Ji, H. (2009) Machining accuracy prediction of Aero-engine blade in electrochemical
machining based on BP Neural Network. Proceedings of the 2009 International Workshop on
Information Security and Application, China, Qingdao.
Li, Z.; Ji, H. (2010) The effects of normal gap distribution on cathode design of Aero-engine blades
in electrochemical machining. Advanced Materials Research, 97–101, 3583–3586.
Li, Z.; Diu, Z. (2006) Numerical solution for cathode design of Aero-engine blades in electro-
chemical machining. International Conference on Mechatronics and Automation, China.
Li, Z.; Niu, Z. (2007) Convergence analysis of the numerical solution for cathode design of Aero-
engine blades in electrochemical machining. Chinese Journal of Aeronautics, 20, 570–576.
Logan, D.L. (2007) A First Course in the Finite Element Method, 4th edn.; Nelson Press, Canada.
Mount, A.R.; Clifton, D.; Howarth, P. (2003) An integrated strategy for materials characteriza-
tion and process simulation in electrochemical machining. Journal of Materials Processing
Technology, 138(1–3): 449–454.
Pattavanitch, J.; Hinduja, S.; Atkinson, J. (2010) Modelling of the electrochemical machining pro-
cess by the boundary element method. CIRP Annals - Manufacturing Technology 59(1): 243–
246.
Piegl, L.A.; Tiller, W. (1997) The NURBS Book, 2nd edn.; Springer Press, Germany.
Purcar, M.; Bortels, L.; Bossche, B.V.; Deconinck, J. (2004) 3D electrochemical machining com-
puter simulations. Journal of Materials Processing Technology, 149, 472–478.
Purcar, M.; Dorochenko, A.; Bortels, L.; Deconinck, J.; Bossche, B.V. (2008) Advanced CAD inte-
grated approach for 3D electrochemical machining simulations. Journal of Materials Process-
ing Technology, 203, 58–71.
Rajurkar, K.P.; Zhu, D.; McGeough, J.A.; Kozak, J.; De Silva, A. (1999) New developments in
electro-chemical machining. Annals of the ClRP, 48(2): 567–579.
Rao, S.S. (2011) The Finite Element Method in Engineering. Elsevier/Butterworth Heinemann
Publicatios, Holland.
 MACHINING SCIENCE AND TECHNOLOGY 173

Rogers, D.F. (2001) An Introduction to NURBS with Historical Perspective, 1st edn.; Morgan Aca-
demic Press, Kaufmann Publishers, USA.
Sun, C.; Zhu, D.; Li, Z.; Wang, L. (2006) Application of FEM to tool design for electrochemical
machining freeform surface. Finite Elements in Analysis and Design, 43, 168–172.
Wang, M.H.; Zhu, D. (2009) Simulation of fabrication for gas turbine blade turbulated cool-
ing hole in ECM based on FEM. Journal of Materials Processing Technology, 209(4): 1747–
1751.
Zhitnikov, V.P.; Fedorova, G.I.; Zinatullina, O.V. (2004) Simulation of non-stationary processes of
electrochemical machining. Journal of Materials Processing Technology, 149(1–3): 398–403.
Zhou, Y.; Derby, J.J. (1995) The cathode design problem in electrochemical machining. Chemical
Engineering Science, 50(17): 2679–2689.