A TOPOGRAPHIC DESCRIPTION OF THE BEARING PROPERTIES

OF ELECTRO-DISCHARGE MACHINED SURFACES

G. Petropoulos1, N. M. Vaxevanidis2, A. Koutsomichalis2 and A. Iakovou1

1. Department of Mechanical and Industrial Engineering, University of Thessaly, Volos, Greece
2. Hellenic Air Force Academy, Dekelia, Attica, Greece

ABSTRACT

The present study concerns with the investigation of a set of “non–common”

surface topography parameters of EDMed surfaces that describe their bearing
capacity in tribological applications; a topic somehow overlooked in technical
literature. The parameters considered are the Abbott (bearing) curve parameter at
10 % of the raw unfiltered and the roughness profile height Ptp10% and Rtp10%,
respectively, the Rk family of parameters (DIN EN ISO 13565-2), the skewness Rsk
and the kurtosis Rku of the profile height distribution and the fractal dimension D.

KEYWORDS: Electro-discharge machining, Surface topography, Rk set of

parameters, Abbot curve

1. INTRODUCTION
Among the various non-conventional processes electro-discharge machining (EDM) is the most
widely and successfully applied for high precision machining of various electrically conductive
materials regardless of their mechanical properties. The material is removed by means of
repetitive spark discharges that cause local melting and/or evaporation of the workpiece
material and the resulted surface is characterized by overlapping craters and features indicative
of the intense thermal impact involved /1/.
Besides the study of surface integrity changes (including surface topography) induced by
EDMachining /2/, of major interest is the correlation of the surface topography parameters with
the machining conditions towards the advanced control and optimisation of the EDM. 3-D
profilometry is a promising tool for this purpose /3/, together with the application of design of
experiments methodology /4/ and neural network models /5/.
In previous works /2, 6, 7/ emphasis was directed towards the multi-parameter analysis of the
EDMed surface texture interrelationship to process parameters and statistical regression
models were developed to correlate the machining conditions with the imparted surface finish
characteristics.
The present study concerns with the investigation of a set of “non-common” surface topography
parameters of EDMed surfaces that describe their bearing capacity in tribological applications;
to the authors’ knowledge this topic is somehow overlooked in technical literature. Surface
analysis followed deals with three aspects, namely: (a) correlation of bearing curve parameters
with machining conditions and other surface texture parameters, (b) representation of the
bearing curves through polynomial fitting models and (c) description of bearing curves by the set
of Rk parameters. The influence of the process variables on a number of amplitude surface
roughness parameters (Ra, Rsk, Rku) is also examined.

2. EXPERIMENTAL
EDMachining was performed on a HOSTEK SH-38GP (ZNC-P type) electro-discharge machine-
tool with working voltage (Ve) of 30V and open circuit voltage of 100V. Experiments were

Proceedings of the 2nd International Conference on Manufacturing Engineering (ICMEN)

5-7 October 2005, Kallithea - Chalkidiki, Greece
Edited by Prof. K.-D. Bouzakis, Director of the Laboratory for Machine Tools and Manufacturing Engineering,
Aristoteles University Thessaloniki, 541 24 Thessaloniki, Greece

159
conducted in a typical oil dielectric (BP250) with electrolytic copper being used as the tool
electrode (anode).
The pulse current, ie and the pulse-on time, tp considered to be the main operational parameters
varied over a range from roughing to finishing, namely: ie: 5, 10, 20, 30 A and tp: 100, 300, 500
µsec, thus resulting in 12 discrete pulse energies; the pulse energy was calculated by the
formula: We = Ve Ie tp (Ve = 30 V).
Specimens of plain carbon steel Ck60 and an AISI D2 tool steel (Sverker 21) in the form of
square plates of dimensions 70mm x 70mm x 10 mm were used as workpieces (cathode).
The surface texture parameters under study are the Abbott (bearing) curve parameter at 10 per
cent of the raw unfiltered and the roughness profile height Ptp10% and Rtp10%, respectively and the
Rk family of parameters defined in DIN EN ISO 13565-2 (1998) /8/, namely: Rk: core roughness
depth, MR1 and MR2: material portion, Rpk: reduce depth height and Rvk: reduced valley height;
see /5/. A number of amplitude surface roughness parameters, i.e. the arithmetic average Ra,
the skewness of the amplitude distribution Rsk and the kurtosis of the amplitude distribution, Rku
as well as the fractal dimension D were also monitored.
The multi-parameter surface texture analysis was performed using a Rank Taylor-Hobson
Surtronic 3+ profilometer equipped with the Talyprof software. The cut-off length was selected
at 0.8 mm whilst 40 measurements were conducted on every specimen at random directions as
it is known that EDMachining generates geometrically isotropic textures /9/.

3. RESULTS AND DISCUSSION

3.1. Variation of amplitude parameters

Arithmetic average roughness (Ra) is by far the most commonly used parameter in surface
finish measurement and for general quality control. Despite its inherent limitations, is easy to be
measured and offers a good overall description of height characteristics of a surface profile. The
variation of average roughness, Ra with pulse energy is presented in figure 1. For EDMed
surfaces the variation of Ra with process operational parameters follows well-known patterns
/10/; it increases when the pulse energy increases with a gradually lower rate. For medium and
high pulse energies (We ≥ 150 mJ), EDMed surfaces of AISI D2 tool steel (Sverker 21) are
systematically rougher than the corresponding ones of Ck 60. This observation can be
correlated with differences in crater dimensions for the certain pulse energies, which in turn are
physically linked with corresponding alterations of the heat source, i.e. the plasma channel and
the resulting melting isothermal as well as with the thermal properties (conductivity, diffusivity) of
the material being machined /11/.
Skewness parameter (Rsk) is typically used to measure the symmetry of the profile about the
mean line and is sensitive to deep valleys or high peaks /12/. The variation of skewness of

Sverker 21 Ck 60

20

15
Ra [µm]

10

0
0 100 200 300 400 500
Pulse Energy We [m J]

Figure 1: Variation of average roughness Ra, with pulse energy, We.

160 2nd ICMEN - 2005

EDMed surfaces with pulse energy is illustrated in figure 2; in general, it appears uncorrelated to
the pulse energy. Judging from the measured skewness values, the EDMed profiles are
revealed to be “empty” of material as indicated from the positive values, excluding some low
pulse energies; at higher energies skewness is almost stabilized; see also /7/.

Sverker 21 Ck 60

0,5
Rsk [µm]

0
0 100 200 300 400 500
-0,5

-1
Pulse Energy We [m J]

Figure 2: Skewness of the heights distribution Rsk, versus pulse energy, We.

Kurtosis (Rsk) typically describes the sharpness of the probability density of the profile /7, 12/.
Measured values of this parameter are in the range of around 3, see figure 3 indicating normally
distributed high peaks and low valleys. Note also that for both Rsk and Rku parameters higher
values were measured for AISI D2 than the Ck 60 specimens for the same medium and high
pulse energies.

Sverker 21 Ck 60

3
Rku [µm]

0
0 100 200 300 400 500
Pulse Energy We [mJ]

Figure 3: Kurtosis of the heights distribution Rku, versus pulse energy, We.

3.2. Correlation of bearing curve parameters with machining conditions

The Abbott (material ratio or bearing) curve is nowadays established as a mean for providing a
working representation of the portions of the surface at different depths combining texture
aspects related to contact area and contact mechanics, wear, lubricant retention and others
/13/. Other views of the surface bearing capacity can be provided by the skewness of the profile
height distribution and the fractal dimension. These topographic parameters apart from
connecting the EDM process to functional behaviour of the machined surfaces can also
describe useful surface features and generally to give useful insights into the nature of
roughness regarding the EDM process control.
The bearing curve parameters corresponding to the raw and the roughness profile were
selected at 10% depth in order to be functionally significant. The variation of Ptp10% and Rtp10%
parameters with pulse energy is presented in figures 4 and 5 respectively. Both parameteres
increase when the pulse energy increases, over a relatively wide range. This behaviour implies

EDM 161
that the bearing capacity of the surface becomes pronounced at the same level, when
intensifying the machining condition.
Worth mentioning that for medium and high pulse energies, both Ptp10% and Rtp10% values
measured for AISI D2 specimens are systematically higher than the ones measured for Ck 60,
see Figures 4 and 5; a trend similar to one observed in Ra measurements.
A similar trend is followed if the Abbot parameters are considered with respect to Ra and Rsk.
Note, that Rsk (skewness) can be considered as a rough measure of the surface bearing
capacity and as it is shown generally correlates with the Rtp (Abbot curve) parameter.

Sverker 21 Ck 60

30
parameter Ptp 10%
25
Bearing curve

20
15
10
5
0
0 100 200 300 400 500
Pulse Energy We [m J]

Figure 4: Variation of Ptp10% parameter of unfiltered profile with pulse energy, We.

Sverker 21 Ck 60
parameter Rtp 10%

25
Bearing curve

20
15
10
5
0
0 100 200 300 400 500
Pulse Energy We [m J]

Figure 5: Variation of Rtp10% parameter of roughness profile with pulse energy, We.
In addition to Abbot parameters fractal-based methods for describing surface texture have
attracted great interest since they can provide information that conventional surface roughness
parameters cannot. The fractal dimension D was calculated via the power spectrum method and
it appears almost insensitive to the pulse energy variation, see figure 6; such a phenomenon is
consistent with results reported in /14/.

Sverker 21 Ck 60
Fractal Dimension D

1,5

0,5

0
0 100 200 300 400 500
Pulse Energy We [m J]

Figure 6: Variation of fractal dimension D with pulse energy, We.

162 2nd ICMEN - 2005

3.3. Representation of bearing curves through polynomial fitting models

Preliminary analysis and previous researches indicated that 3rd order polynomial fitting model is
the best representation of bearing curves /15/. Due to limitation of space only the raw profile
bearing curves fitting, for Ck 60 steel specimens, according to equation, y = ko + k1x + k2x2 +
k3x3 is presented. Calculated fitting coefficients are shown in figure 7. An intense dropping trend
of the cubic coefficient appears at high pulse energies. All these changes in the shape of the
bearing curves are closely related to contact mechanics and wear behaviour /16/.

Figure 7: Polynomial fitting coefficient against pulse energy, We (ko: constant, k1: linear, k2:
square, k3: cubic, material: Ck 60).

3.4. Description of bearing curves by the set of Rk parameters

The variation of these parameters with pulse energy for all EDMed specimens is plotted in
figures 8-12. Judging from these plots it is obvious that ISO 13565-2 standard suits EDM, as the
bearing curves are of a general s-shape and the parameters vary in a monotonous way. Note
however, that this standard was originally formulated for stratified textures and therefore present
results need further clarification, e.g. through friction and wear test on a pin-on-disc tribometer.

Sverker 21 Ck 60

40
30
Rk [µm]

20
10
0
0 100 200 300 400 500
Pulse Energy We [m J]

Figure 8: Variation of Rk with pulse energy, We.

EDM 163
 Sverker 21 Ck 60

10
8

Rpk [µm]
6
4
2
0
0 100 200 300 400 500
Puls e Ene rgy We [m J]

Figure 9: Variation of Rpk with pulse energy, We.

Sverker 21 Ck 60

30
Rvk [µm]

20

10

0
0 100 200 300 400 500
Pulse Energy We [m J]

Figure 10: Variation of Rvk with pulse energy, We.

Sverker 21 Ck 60

30
25
20
MR1 [%]

15
10
5
0
0 100 200 300 400 500
Puls e Ene rgy We [m J]

Figure 11: Variation of MR1 parameter with pulse energy, We.

Sverker 21 Ck 60

120
100
80
MR2 [%]

60
40
20
0
0 100 200 300 400 500
Puls e Ene rgy We [m J]

Figure 12: Variation of MR2 parameter with pulse energy, We.

164 2nd ICMEN - 2005

From the plots presented in Figures 8-12 it is evident that for low pulse energies (We ≤ 90 mJ)
measured values for the two steels are very similar alike; for medium and high pulse energies
measured values for AISI D2 tool steel (Sverker 21) are systematically higher than the
corresponding ones for Ck 60. This trend was identified for all five parameters of Rk family.
Moreover, for high pulse energies (We ≥ 270 mJ) Rvk, MR1 and MR2 are almost constant and
characteristic for each material, see Figures 10, 11 and 12 respectively.

4. CONCLUSIONS

Close correlation exists between the tp bearing curve parameters of the raw and the roughness
profile and the pulse energy regarded as the main machining variable. It was found that the Ptp
and Rtp parameters increase monotonously with increase in pulse energy, whilst most of the Rk
parameters exhibit a similar trend. Rsk and Rku parameters appear uncorrelated to pulse energy
and moreover, the fractal dimension D appears insensitive to the pulse energy variation. Note
that Rsk can be considered as a rough estimate of the surface bearing capacity and as it is
shown, generally is correlated with the tp (Abbot) parameters.
The bearing curves are represented by a 3rd degree polynomial model which fits them
satisfactorily. An intense falling trend of the cubic coefficient appears at high pulse energies. All
these changes in the shape of the bearing curves are closely related to contact mechanics and
wear behaviour of machined surfaces.
As far as the description of bearing curves by the set of Rk parameters defined in DIN EN ISO
13565-2 (1998) is concerned it is indicated that these parameters can provide an adequate
description of crucial portions of the resulted surface profile in relation to the machining
condition; this is expectable owing to the near s-shaped bearing curves of the EDMed surfaces.

5. REFERENCES

1. Snoeys R., Staelens F., Dekeyser W., Current trends in non-conventional material removal
processes, Annals of the CIPR, 35(2) (1986) 467-480.
2. Vaxevanidis N.M., Psylaki P., Petropoulos G.P., Hassiotis N., Surface integrity and
microstructural phenomena of Ck 60 steel due to Electro-Discharge Machining, Proc.
EUROMAT 99 - Steels and materials for power plants, 1999, pp: 240-247.
3. Ramasawmy H., Blunt L., Effect of EDM process parameters on 3D surface topography,
Journal of Materials Processing Technology, 148 (2004) 155-164.
4. Petropoulos G., Vaxevanidis N., Iakovou A., David K., Multi-parameter model of surface
texture for EDMachining of tool steel using design of experiments, Proceedings of 1rst
Manufacturing Enginnering Society International Conference “SIC-CISIF” 19-21 September
2005 Calatayud, Spain, 2005, paper S1-046.
5. Tsai K.M., Wang P.-W., Prediction of surface finish in electro discharge machining based
upon neural network models, International Journal of Machine Tools and Manufacture 41
(2001) 1385-1403.
6. Mamalis A.G., Vosniakos G.C., Vaxevanidis N.M., Xiong J.Z., Residual stress distribution
and structural phenomena of high-strength steel surfaces due to EDM and ball-drop forming,
Annals of the CIRP, 37(1) (1988) 531-535.
7. Petropoulos G., Vaxevanidis N., Pandazaras C., Statistical multi-parameter analysis of
EDMachined surface textures, Journal of Materials Processing Technology, 155-156 (2004)
1247-1251.
8. DIN EN ISO 13565-2, Geometrical Product Specifications (GPS) - Surface texture: Profile
method - Surfaces having stratified functional properties - Part 2: Height characterization
using the linear material ratio curve. (1998)
9. Thomas T. R., Rosen B. G., Amini N., Fractal characterization of the anisotropy of rough
surfaces, Wear, 232 (1999) 41-50.

EDM 165
10. Rebelo J.C., Dias A., Kremer D., Lebrun J.L., Influence of EDM pulse energy on the surface
integrity of martensitic steel, Journal of Materials Processing Technology, 84 (1998) 90-96.
11. Mamalis A.G., Vaxevanidis N.M., Karafillis A.P., Surface Integrity and Formability of Steel
Sheets, VDI Verlag, Düsseldorf. (1990)
12. Gadelmawla E.S., Koura M.M., Maksoud T.M.A., Elewa I.M., Soliman H.H., Roughness
parameters, Journal of Materials Processing Technology, 123 (2002) 133-145.
13. Petropoulos G. P., Τοrrance Α., Pandazaras C., Abbott curves characteristics of turned
surfaces, International Journal of Machine Tools and Manufacture, 43 (2003) 237-243.
14. Hasegawa M., Liu J., Okuda K., Nunobiki M., Calculation of the fractal dimensions of
machined surface profiles, Wear, 192 (1996) 40-45.
15. Prostrednik D., Osanna P.H., The Abbot curve-well known in metrology but not on technical
drawings, International Journal of Machine Tools and Manufacture, 38 (1998) 741-745.
16. Torrance A. A., A simple datum for measurement of the Abbott curve of a profile and its first
derivative, Tribology International, 30 (1997) 230-244.

166 2nd ICMEN - 2005