Chinese

Journal of
Chinese Journal of Aeronautics 20(2007) 570-576
Aeronautics
www.elsevier.com/locate/cja

Convergence Analysis of the Numerical Solution for Cathode

Design of Aero-engine Blades in Electrochemical Machining
Li Zhiyong*, Niu Zongwei
School of Mechanical Engineering, Shandong University of Technology, Zibo 255091, China

Received 25 September 2007; accepted 19 October 2007

Abstract

As a main difficult problem encountered in electrochemical machining (ECM), the cathode design is tackled, at present, with
various numerical analysis methods such as finite difference, finite element and boundary element methods. Among them, the finite
element method presents more flexibility to deal with the irregularly shaped workpieces. However, it is very difficult to ensure the con-
vergence of finite element numerical approach. This paper proposes an accurate model and a finite element numerical approach of cath-
ode design based on the potential distribution in inter-electrode gap. In order to ensure the convergence of finite element numerical ap-
proach and increase the accuracy in cathode design, the cathode shape should be iterated to eliminate the design errors in computational
process. Several experiments are conducted to verify the machining accuracy of the designed cathode. The experimental results have
proven perfect convergence and good computing accuracy of the proposed finite element numerical approach by the high surface quality
and dimensional accuracy of the machined blades.

Keywords: electrochemical machining; aero-engine blade; cathode design; convergence analysis

1 Introduction* charge machining, die casting, forging and numeri-

cal control machining might result in poor surface
With the rapid progresses of aero-engine per-
quality, residue stresses and surface micro-cracks[1-4].
formance, the machining quality and dimensional
In comparison with them, electrochemical machin-
accuracy of aero-engine blades become ever-in-
ing (ECM), by using anodic dissolution to shape
creasing important. In order to obtain excellent
metal, has been developed to machine high-strength,
aerodynamic performance, the shape of aero-engine
heat-resistant alloys, which are otherwise extremely
blades turns more and more complex. In addition,
difficult to be worked by conventional methods, to
because of their high strength, high temperature
which, naturally, ECM has turned to be the ideal
stability, and high corrosion resistance, hard passive
alternative in aeronautic and astronautic industries,
alloys, such as nickel-based superalloys, titanium
particularly in production of aero-engine blades.
alloys and molybdenum alloys have found wider
Fig.1 shows a selection of aero-engine blades pro-
application than ever as aero-engine blade materials,
duced by ECM[5-7].
for which traditional machining methods of manu-
However, further application of ECM is limited
facturing aero-engine blades such as electro-dis-
by the difficulties in tool (cathode) design and proc-
ess monitoring.
*Corresponding author. Tel.: +86-533-2768608. In ECM, cathode design is to fix the shape of a
E-mail address: [email protected]
Foundation item: Foundation of the Chinese People’s Liberation Army cathode to produce a required shape of workpieces.
General Armament Department (41318.3.2.4)
 Li Zhiyong et al. / Chinese Journal of Aeronautics 20(2007) 570-576 · 571 ·

cathode design, the cathode shape should be iterated

in computing process to eliminate the errors. Finally,
experiments for verification have been conducted
with satisfied results.

2 Principle and Model of ECM

Fig.3 shows the schematic diagram of ECM
system. The ECM process connects the workpiece
(anode) to the tool (cathode) via an electrolytic cell,
Fig.1 ECM aero-engine blades.
through which an electrolyte is pumped. A low DC
Rapid advancements in computer and numerical voltage of 10-20 V is applied across the gap be-
technologies have rendered it possible to develop a tween the tool and the workpiece. A high ampere
more complicated approach of cathode design. Cur- current (30-200 A/cm2) is made to pass the gap
rent efforts are primarily made to address this issue through electrolyte flowing at high speed of 6-30
based on the Laplace equation for potential distribu- m/s. The anode material dissolved electrochemically
tion in inter-electrode gap by using numerical solu- is flushed away by flowing electrolyte. By feeding
tions. Various numerical solutions such as boundary the tool continuously, the shape of the workpiece is
element, finite difference and finite element have becoming a mirror of the tool electrode. In an ECM
been applied to the cathode design in ECM. Among process, the inter-electrode gap distribution depends
them, finite element method presents more flexibil- on electric filed distribution in function of many
ity to deal with the irregularly shaped workpiec- process parameters, which vary in space and time
es[8-12]. However, it is not easy to ensure the con- during electrochemical dissolution process. Fig.4
vergence of the finite element method. By taking an shows the integrated mathematic model of electric
aero-engine compressor blade as the object (model field distribution within ECM gap domain :[13-15].
blade) of the research shown in Fig.2, this paper Based on the ECM fundamental theories, the
proposes an accurate model and finite element nu- electric potential distribution within : can be de-
merical approach for the cathode design of blades in scribed by Laplace equation
ECM based on the potential distribution in in-
w 2M w 2M w 2M
ter-electrode gap. All boundary conditions which   0 (1)
wx 2 wy 2 wz 2
should be satisfied in the numerical approach are
analyzed. In order to ensure the convergence of this According to Faraday’s law and Ohm’s law,
numerical approach and increase the accuracy of anode and cathode boundaries should satisfy par-

Fig.2 Solid model of an aero-engine blade. Fig.3 Schematic diagram of ECM system.
· 572 · Li Zhiyong et al. / Chinese Journal of Aeronautics 20(2007) 570-576

enclosed ECM gap domain :. Failing to satisfy

those boundary conditions mentioned above though,
the initial cathode shape provides an enclosed gap
domain as an analysis model in FE.
(2) FE model. : is divided into small hexahe-
dron elements, as shown in Fig.5. There are five
layers in : with mn (n lines×m columns) being the
number of nodal points at each layer. A potential
function M(x, y, z) that varies linearly inside each
hexahedron element is defined as
8
M ( x, y , z ) ¦ Ni ( x, y, z )Mi (5)
Fig.4 Electric field distribution within ECM gap. i 1

where Ni(x, y, z) (i=1,2,···,8) is a linear functions of

ticular machining conditions expressed by
x, y and z, which can be expressed by
wM vf
cos T on anode boundary (2)
wn K vN N i ( x, y , z ) ai  bi x  ci y  di z 
Ma U on anode boundary (3) ei xy  fi yz  gi xz  hi xyz (6)
Mc 0 on cathode boundary (4)
where M is the electric potential within :, Ma the
anode electric potential, Mc the cathode potential, U
the applied voltage, vf the feed rate of cathode, K
the current efficiency, Kv the volumetric electro-
chemical equivalent of anode metal, N the electro-
lyte conductivity (constant), T the angle between
the normal to the anode and the feed direction of
cathode. The aim of cathode design is to find a
cathode boundary which can satisfy Laplace equa-
tion for the electric potential distribution within Fig.5 Meshing and numbering elements.
ECM gap domain : and also all boundary condi-
According to FE computing method, the coef-
tions listed in Eqs.(1)-(4). This can be achieved by
ficient matrix of each element, Ke, can be calculated
specific numerical solution, which will be discussed
as
below.
§ wȃ e wȃ ej wȃ e wȃ ej wȃ e wȃ ej ·
Ȁ ije i i i
³³³ V e ¨¨ wx ˜ wx  wy ˜ wy  wz ˜ wz ¸¸ dV
3 Finite Element Numerical Solution for
© ¹
Cathode Design (7)
As mentioned above, the aero-engine blade Assembling all elements coefficient matrix to-
shown in Fig.1 should be regarded as a workpiece, gether can be given
i.e., an anode in ECM. The general process of nu-
ª k11 k12 0 0 0 º ªM1 º ªb1 º
merical solution based on finite element (FE) « » « » «0 »
method are summarized in following three stages: « k21 k22 k23 0 0 » «M 2 » « »
«0 k32 k33 k34 0 » ˜ «M 3 » «0 » (8)
(1) To form an initial ECM gap domain. An « » « » « »
«0 0 k43 k44 k45 » «M 4 » «0 »
approximate initial cathode boundary (shape) can be «0
¬ 0 0 k54 k55 »¼ «¬M 5 »¼ «0 »
¬ ¼
achieved with the “cos T method”. Combining this
initial cathode shape and the blade shape forms the where kij(i, j=1,2,···,5) is mn×mn matrix with non-
 Li Zhiyong et al. / Chinese Journal of Aeronautics 20(2007) 570-576 · 573 ·

zero determinants assembled from Ni(x, y, z), Mi (i= “sharp”, a momentous change of curvature happens
1,2,···,5) the column vector potential. Particu- at blade edges, which may results in excessive de-
larly,M1 and M5 are column vectors of potential at formation of hexahedron elements. All these will
the blade shape and cathode shape respectively. M2 reduce the computing accuracy of potential distribu-
toM4 are center gap potential and b1 is the column tion in ECM gap domain, which in turn affects the
vector generated by path integration in cathode boundary. Therefore, in order to improve
vf computing accuracy and accelerate convergence of
³³ K v ˜ N cos T ˜ M ds (9)
the FE method, might be taken the following steps:
(3) Boundary condition treatment and cathode ķ to work out and implement an FE cathode design
shape selection. One boundary condition on anode program to obtain some equi-potential shape. The
(blade) shape is applied in Eq.(9) by calculating b1, initial cathode boundary generated by “cos T
another boundary condition, Ma = U, is also applied method” requires to be replaced by a selected cath-
by letting all elements in vector M1 equal U in ode shape (an equi-potential shape); ĸ to repeat the
Eq.(8). Thereafter, Eq.(8) can be solved successfully, FE cathode design program after the initial cathode
which means the electrical potential distribution boundary has become an equi-potential shape se-
within : can be determined. From the known po- lected in above step to adjust the cathode boundary;
tential distribution, may be found several equi-po- Ĺ repeat step ĸ till the adjusted amplitude of
tential shapes, out of which an appropriate one can cathode boundary is no more than the designated
be selected as the cathode shape according to actual value, for example, 0.5%. This means a stable po-
machining parameters. tential distribution in ECM gap domain and a cath-
ode boundary.
4 Convergence Analysis of FE Algorithm As mentioned above, every step of iteration
In principle, the FE numerical solution for can increase the design accuracy of cathode shape
cathode design of aero-engine blades mentioned and accelerate convergence of FE algorithm. The
above does not need iteration. After computation, rapid convergence of FE algorithm is vital to obtain
the potential distribution in ECM gap will satisfy a high quality cathode shape. Slow convergence of
the Laplace equation and all boundary conditions. FE algorithm in cathode design makes lead-time and
Though all these equi-potential shapes found from machine downtime longer and cost higher. The FE
the known potential distribution are the cathode computing program generally takes about one hour
shapes required, their potential distribution in ECM for iterating 3-5 times to obtain an appropriate
gap is not stable and, moreover, these equi-potential cathode shape. Fig.6(a), (b), (c) and (d) show four
shapes are not precise. This means the need for im- sections from sixty shape cathode boundaries in an
provement of convergence of this FE numerical so-
lution. The main causes that affect convergence of
FE numerical solution can be summarized as fol-
lows:
(1) The problem of boundary insulation. Ac-
cording to FE numerical solution, it is considered
that the ECM gap boundaries have been insulated,
as shown in Fig.4. However, in actual ECM process,
all gap boundaries are “open”, meaning absence of
insulation.
(2) Curvature change. Because the pressure
side and suction side of the model blade are very
· 574 · Li Zhiyong et al. / Chinese Journal of Aeronautics 20(2007) 570-576

stable. Fig.7(a) and (b) are the solid models of

cathode shape designed for machining pressure side
and suction side of the model blade (as shown in
Fig.2) respectively. The two cathode models were
completed on UG software platform.

Fig.7 Disigned cathode shapes for the model blade.

5 Experimental Work
In order to verify the design accuracy of cath-
ode shape, machining experiments on the pressure
side of model blade were conducted on an indus-
try-use ECM machine, Model JAPAX-300B, as
shown in Fig.8. Prior to machining experiments, a
Fig.6 Comparison of cathode sections before and after
iteration with FE algorithm. specific cathode tool was made. Then the cathode
model, shown in Fig.7(a), was incorporated into a
attempt to make a comparison of them before and
CAM system to produce a digital profile, which was
after iteration with FE algorithm.
set into the controller of a CNN machine. Thus the
From Fig.6, it is clear that FE iteration algo-
required cathode tool can be machined out of
rithm for cathode design of blade is able to effec-
stainless steel 1Cr18Ni9T in accordance with the
tively improve design accuracy of the cathode shape.
digitized data, as shown in Fig.9. Fig.10 shows the
After iteration, the sections of cathode shape be-
machined blade samples. According to the require-
come more smooth and stable, which in turn make
ments, among the given nine sections of model
the potential distribution in ECM gap domain more
 Li Zhiyong et al. / Chinese Journal of Aeronautics 20(2007) 570-576 · 575 ·

blade, each one’s dimension should be within 0.12

mm of any other one’s. The final measure results of
each section of two blade samples are plotted on
Fig.11(a) and (b), which evidence an entire agree-
ment between all of the measured data and the di-
mension accuracy requirements. This in turn proves
the presented numerical solution in cathode design
is valuable to produce a required tool shape, and
therefore, improve ECM process performance.

Fig.8 Machining experiment of ECM.

Fig.11 Measured data of dimensional accuracy of given
nine sections.

6 Summary and Conclusions

A numerical solution for cathode design of
aero-engine blades in ECM is proposed. The essen-
tial of this approach is to find an appropriate cath-
ode (tool) shape, which can satisfy the Laplace
equation for electrical potential distribution in ECM
gap domain and all other electrode boundary condi-
tions to achieve the desired blade shapes and sizes.
Moreover, the convergence of this FE numerical
solution has been discussed in detail. In order to
verify the machining accuracy of the designed cath-
ode, the experiments have been conducted. The ex-
perimental results demonstrate that the machined
blades have high surface quality and dimensional
accuracy. This proves the proposed FE numerical
Fig.9 Cathode tool for the pressure side of model blade. approach has good convergence and computing ac-
curacy.

References
[1] Yang C T, Song S L, Yan B H. Improving machining performance

of wire electrochemical discharge machining by adding SiC abra-

sive to electrolyte. International Journal of Machine Tools and

Manufacture 2006; 46(15): 2044-2050.

Fig.10 Experimental samples of machined blades. [2] Pajak P T, Desilva A K M, Harrison D K. Precision and efficiency
· 576 · Li Zhiyong et al. / Chinese Journal of Aeronautics 20(2007) 570-576

of laser assisted jet electrochemical machining. Precision Engi- Materials Processing Technology 2004; 149(1-3): 398-403.

neering 2006; 30(3): 288-298. [11] Purcar M, Bortels L, den Bossche B V. 3D electrochemical ma-

[3] Wüthrich R, Fascio V. Machining of non-conducting materials chining computer simulations. Journal of Materials Processing

using electrochemical discharge phenomenon—an overview. In- Technology 2004; 149(1-3): 472-478.

ternational Journal of Machine Tools and Manufacture 2005; [12] Mount A R, Clifton D, Howarth P. An integrated strategy for

45(9): 1095-1108. materials characterisation and process simulation in electrochemi-

[4] Peng W Y, Liao Y S. Study of electrochemical discharge machin- cal machining. Journal of Materials Processing Technology 2003;

ing technology for slicing non-conductive brittle materials. Jour- 138(1-3): 449-454.

nal of Materials Processing Technology 2004; 149(1-3): 363-369. [13] Zhu D, Rajurkar K P. Modeling and verification of interelectrode

[5] Sun J J, Taylor E J, Srinivasan R. MREF-ECM process for hard gap in electrochemical machining with passivating electrolyte.

passive, materials surface finishing. Journal of Materials Process- Transactions of the ASME 1997; 119(4B):770-775.

ing Technology 2001;108(3): 356-368. [14] Li Z Y, Zhu D. Numerical solution for cathode design of

[6] Lee S J, Lee Y M, Du M F. The polishing mechanism of electro- aero-engine blades in electrochemical machining. IEEE Interna-

chemical mechanical polishing technology. Journal of Materials tional Conference on Mechatronics and Automation. 2006; 1418-

Processing Technology 2003; 140(1): 280-286. 1423.

[7] Westley J A, Atkinson J, Duffield A. Generic aspects of tool de- [15] Zhou Y M, Derby J J. The cathode design problem in electro-

sign for electrochemical machining. Journal of Materials Process- chemical machining. Chemical Engineering Science 1995; 50(17):

ing Technology 2004; 149(1-3): 384-392. 2679-2689.

[8] Sun C H, Zhu D, Li Z Y. Application of FEM to tool design for

Biography:
electrochemical machining freeform surface. Finite Elements in
Li Zhiyong Born in 1976, he received
Analysis and Design 2006; 43(2): 168-172.
B.S. from Qingdao University in 1999, and
[9] da Silva-Neto J C, da Silva E M. Intervening variables in electro-
Ph.D. degree from Nanjing University of
chemical machining. Journal of Materials Processing Technology
Aeronautics and Astronautics in 2004. His
2006; 179(1-3): 92-96.
main research interest includes non-tradi-
[10] Zhitnikov V P, Fedorova G I, Zinatullina O V. Simulation of
tional machining technology.
non-stationary processes of electrochemical machining. Journal of
E-mail: [email protected]