Material removal rate and electrode wear study on the EDM of silicon carbide

Journal of Materials Processing Technology 164–165 (2005) 889–896

Material removal rate and electrode wear study on the EDM of silicon carbide C.J. Luis, I. Puertas ∗, G. Villa

Mechanical, Energetics and Materials Engineering Department, Manufacturing Engineering Section, Public University of Navarre, Campus de Arrosadia, Pamplona, Navarre 31006, Spain

Abstract

In this work, a material removal rate (MRR) and electrode wear (EW) study on the die-sinking electrical discharge machining (EDM) of siliconised or reaction-bonded silicon carbide (SiSiC) has been carried out. The selection of the above-mentioned conductive ceramic was made taking into account its wide range of applications in the industrial field: high-temperature gas turbines, bearings, seals and lining of industrial furnaces. This study was made only for the finish stages and has been carried out on the influence of five design factors: intensity supplied by the generator of the EDM machine (I), pulse time (ti), duty cycle (η), open-circuit voltage (U) and dielectric flushing pressure (P), over the two previously mentioned response variables. This has been done by means of the technique of design of experiments (DOE), which allows us to carry out the above-mentioned analysis performing a relatively small number of experiments. In this case, a 25−1 fractional factorial design, whose resolution is V, has been selected considering the number of factors considered in the present study. The resolution of this fractional design allows us to estimate all the main effects, two-factor interactions and pure quadratic effects of the five design factors selected to perform this study. © 2005 Elsevier B.V. All rights reserved.

Keywords: EDM; Manufacturing; Wear; DOE

1. Introduction

high-density parts which have excellent technical properties related to hardness, mechanical resistance, wear, and corro- sion [2].

Electrical discharge machining (EDM) is a non-traditional manufacturing process based on removing material from a part by means of a series of repeated electrical discharges between a tool, called the electrode, and the part being ma- chined in the presence of a dielectric fluid. At present, EDM is a widespread technique used in industry for high-precision machining of all types of conductive materials such as: met- als, metallic alloys, graphite, or even some ceramic materials, of any hardness [1].

In spite of their exceptional mechanical and chemical properties, ceramic materials have only achieved a partial acceptance in the field of industrial applications due to the difficulties in processing and the high costs associated with their manufacture. Over the past few years, the advances in the field of EDM have permitted the application of this tech- nology to the manufacture of conductive ceramic materials. In line with current knowledge, the main inconvenience when applying the EDM technology to the field of ceramic mate- rials is the electrical resistivity of these materials, where the limits are fixed between 100 and 300 (cid:2) cm [2–4].

The term, technical ceramic materials or advanced ce- ramic materials, is a relatively new term, which is applied to a range of various materials generally obtained from inorganic primary materials with a high grade of purity. These primary materials are subjected to typical processes in powder metal- lurgy technology and subsequently, to high-temperature sin- tering processes. With these materials, it is possible to obtain

∗ Corresponding author. Tel.: +34 948 169 305; fax: +34 948 169 099.

In this work, the selected ceramic material has been sil- iconised or reaction-bonded silicon carbide (SiSiC), whose field of applications is in constant growth. In this way, a ma- terial removal rate (MRR) and electrode wear (EW) study has been carried out on the influence of the following de- sign factors: intensity (I), pulse time (ti), duty cycle (η), open-circuit voltage (U) and dielectric flushing pressure (P).

E-mail address: inaki.puerta@unavarra.es (I. Puertas).

0924-0136/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2005.02.045

890 C.J. Luis et al. / Journal of Materials Processing Technology 164–165 (2005) 889–896

This has been performed using the techniques of design of experiments (DOE) and multiple linear regression anal- ysis. In this case, a 25−1 fractional factorial design with resolution V has been employed to carry out the present study.

2. Equipment used and conductive ceramic EDMed

In this section, a brief description of the EDM machine used to perform the experiments and the conductive ceramic material studied in this work will be given.

2.1. Die-sinking EDM machine

approximately 10% of free silicon, which fills the pores. The resulting microstructure has a low level of porosity and a very fine grain, giving a density of 3.1 g/cm3. The mechan- ical resistance of the material (2000–3500 MPa to compres- sion and 310 MPa to tension, up to 1350 ◦C), combined with a hardness (25–35 GPa, Vickers hardness) which is higher than that of tungsten carbide (WC), explains its use as an element in high-temperature gas turbines as well as bearings and seals. Furthermore, it has a high thermal conductivity (150–200 W m−1 K−1, 20 ◦C) and a low thermal expansion coefficient (4.3–4.6 × 10−6 K−1, 20–1000 ◦C), which pro- vides it with a good resistance to thermal shock. Reaction- bonded silicon carbide performs better under chemical corro- sion than other ceramic materials, such as tungsten carbide or alumina (Al2O3). Therefore, it is frequently used as industrial furnaces lining.

The samples of silicon carbide used in the experi- ments were ground sheets of the following dimensions: 50 mm × 50 mm × 5 mm. Moreover, as was mentioned ear- lier, the electrodes used in this case were made of electrolytic copper and subjected to a negative polarity as, according to the bibliography in the EDM field and these authors’ expe- rience [2,5], it is the most stable and recommended way to carry out the EDM process of silicon carbide.

The equipment used to perform the experiments was a die-sinking EDM machine of type ONA DATIC D-2030-S (Fig. 1). Also, a jet flushing system in order to assure the adequate flushing of the EDM process debris from the gap zone was employed. The dielectric fluid used for the EDM machine was a mineral oil (Oel-Held Dielektrikum IME 82) with a flash point of 82 ◦C. The electrodes used were made of electrolytic copper (with a cross-section of 12 mm × 8 mm) and the polarity was negative.

2.2. Reaction-bonded or siliconised silicon carbide

3. Design of the experiments

In the present section, the design factors and response vari- ables selected for this work, as well as the methodology em- ployed for the experimentation, will be described.

The ceramic material used was REFEL F® silicon car- bide. This reaction-bonded silicon carbide (SiSiC) is manu- factured by infiltrating silicon into a porous block made of silicon carbide powder and carbon, which is then submitted to a firing process at a specific temperature. This produces

3.1. Design factors selected

There are a large number of factors to consider within the EDM process, but in this work the level of the genera- tor intensity (I), pulse time (ti), duty cycle (η), open-circuit voltage (U) and dielectric flushing pressure (P) have only been taken into account as design factors. The reason why these five factors have been selected as design factors is that they are the most widespread and used amongst EDM resear- chers.

The intensity (I) depends on the different power levels that can be supplied by the EDM machine generator. It represents the maximum value of the discharge current intensity. The intensity values used in the EDM machine programming are power levels of the generator, these corresponding with val- ues of the peak intensity (Ip), which is applied between the electrode and the part to be EDMed.

Fig. 1. EDM machine used to carry out the experiments.

Pulse time or on-time (ti) is the duration of time (in (cid:1)s) the current is allowed to flow per cycle. On the other hand, duty cycle (η) is the percentage of pulse time relative to the total cycle time. This third factor is calculated by dividing pulse time by the total cycle time (i.e., pulse time plus pause time), where pause time or off-time (to) is the duration of time (in (cid:1)s) between two consecutive sparks.

C.J. Luis et al. / Journal of Materials Processing Technology 164–165 (2005) 889–896 891

Table 1 Levels selected for the five design factors of the 25−1 design Factors Levels

−1 0 +1

Open-circuit voltage (U) is the value of the electric tension applied between the part to be machined and the electrode just before the discharge is produced. Finally, the dielectric flush- ing pressure (P) is the pressure of the dielectric jet removing the EDM splinter or debris from the gap zone. This pressure value is measured by a pressure gauge in the EDM machine.

3 30 0.4 4 50 0.5 5 70 0.6

3.2. Response variables selected

I ti ((cid:1)s) η U (V) P (kPa) −120 20 −160 40 −200 60

The response variables selected for this study refer to the speed of the EDM process, i.e., material removal rate (MRR), and the efficiency of the copper electrode used, i.e., volumet- ric electrode wear (EW). These response variables are defined in Eqs. (1) and (2), respectively: MRR = volume of material removed from part

(1)

time of machining

(2)

EW = volume of material removed from electrode

volume of material removed from part

were made. In case the first-order model turned out not to be adequate for modelling the behaviour of the response variable to be studied, this was widened by adding 10 star points, thus giving a central composite design with the star points located in the centres of the faces. Thus, the case of the second- order model consisted of a total of 30 experiments, i.e., the previous 20 of the first-order model plus the new 10 of the star points.

Although there are some other ways of measuring MRR and EW [6], in this work we have calculated them by the weight difference of the sample and the electrode just before and after being subjected to the EDM process. Moreover, density values of 3.10 and 8.96 g/cm3 were chosen for silicon carbide and electrolytic copper, respectively, in order to assess MRR and EW.

3.3. Fractional factorial design employed

The low and high levels selected for intensity, pulse time, duty cycle, open-circuit voltage and dielectric flushing pres- sure were: 3 and 5, 30 and 70 (cid:1)s, 0.4 and 0.6, −120 and −200 V and finally, 20 and 60 kPa, respectively. The lev- els of the intensity factor (3 and 5) are equivalent to 2 and 6 A, respectively. A summary of the levels selected for the factors to be studied is shown in Table 1. The previous se- lection of variation levels was made taking into account that the authors wanted this study to be focused on the area of the finish machining stages, owing to the influence that a good surface quality has on such important properties, in the case of ceramic materials, such as fatigue and wear resistance [8,9].

The design of experiments (DOE) technique is a pow- erful work tool which allows us to model and analyse the influence of determined process variables over other speci- fied variables, which are usually known as response variables. These response variables are unknown functions of the for- mer design variables, which are also known as design factors [7–9].

Table 2 shows the design matrix for the second-order mod- els as well as the values obtained in the experiments for the response variables studied in this work, i.e., MRR and EW. As can be observed in this table, rows 1–16 correspond to the fractional factorial design, rows 17–26 correspond to the star points and finally, the central points are placed in the four last rows of the design matrix. On the other hand, the design matrix for the first-order model of MRR and EW would be obtained simply by eliminating the 10 rows corresponding to the star points.

The graphs, models and tables that are to be presented in the following two sections were produced using the DOE module of the software STATGRAPHICS® plus v. 5.0.

Within the design of experiments, there are various types that can be considered. One of the most widely known ones is the factorial design; this consists in experimenting with all the possible combinations of variables and levels. Neverthe- less, if the number of design factors is reasonably high, as in this case, and it is possible to assume that certain high- order interactions are negligible, the information on the main effects and the low-order interactions may be obtained by running only a fraction of the complete factorial experiment. These fractional factorial designs are among the most widely used types of designs for product and process design and for process improvement [7].

4. Results and analysis of MRR

The design finally chosen was a 25−1 fractional factorial one with four central points. This design has a resolution of V, which means that no main effect or two-factor interaction in this model is aliased with any other main effect or two-factor interaction. The addition of four central points allows us to carry out lack-of-fit tests for the first-order models proposed, where a total of 20 experiments for these first-order designs

A first-order model was proposed for the response vari- able MRR, where this was rejected as a result of the values obtained for the curvature test that can be seen in Table 3. As in this case the P-value, which is equal to 0.0002, is lower than 0.05, the null hypothesis that there are no pure quadratic effects in the model is therefore rejected, accepting

892 C.J. Luis et al. / Journal of Materials Processing Technology 164–165 (2005) 889–896

Table 2 Design matrix for the second-order models of MRR and EW η P (kPa) MRR (mm3/min) I ti ((cid:1)s)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 3 5 3 5 3 5 3 5 3 5 3 5 3 5 3 5 3 5 4 4 4 4 4 4 4 4 4 4 4 4 30 30 70 70 30 30 70 70 30 30 70 70 30 30 70 70 50 50 30 70 50 50 50 50 50 50 50 50 50 50 0.4 0.4 0.4 0.4 0.6 0.6 0.6 0.6 0.4 0.4 0.4 0.4 0.6 0.6 0.6 0.6 0.5 0.5 0.5 0.5 0.4 0.6 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 U (V) −120 −120 −120 −120 −120 −120 −120 −120 −200 −200 −200 −200 −200 −200 −200 −200 −160 −160 −160 −160 −160 −160 −120 −200 −160 −160 −160 −160 −160 −160 60 20 20 60 20 60 60 20 20 60 60 20 60 20 20 60 40 40 40 40 40 40 40 40 20 60 40 40 40 40 EW (%) 10.29 7.34 19.82 10.01 11.87 9.83 13.55 12.16 12.66 7.48 16.06 7.91 12.84 11.63 16.33 9.29 13.29 7.47 6.61 6.74 6.92 10.50 10.92 8.00 7.24 6.65 6.61 7.06 7.25 7.51 0.033 0.104 0.020 0.116 0.049 0.091 0.030 0.136 0.051 0.312 0.030 0.267 0.092 0.344 0.055 0.354 0.032 0.283 0.244 0.177 0.147 0.168 0.087 0.258 0.195 0.201 0.189 0.199 0.199 0.201

The equation of the second-order proposed model is that

shown in Eq. (3):

MRR = −0.985466 + 0.151912 · I − 0.00642768 · ti

that there is statistical evidence of curvature in the first-order model, for a confidence level of 95%. Thus, that the pro- posed first-order model is suitable for a significance level α of 0.05 is rejected and so, the second-order model is selected.

Table 4 shows the ANOVA table for the case of the second-order model proposed, where now, the total num- ber of degrees of freedom is equal to 29. As can be ob- served in this table, there are three effects with a P-value less than 0.05, which means that they are significant for a confidence level of 95%. These significant effects, ar- ranged in order of importance, are: the main effects of in- tensity and voltage and finally, the interaction effect between them.

+3.08858 · η − 0.000251344 · U−0.00216784 · P −0.0344143 · I2 + 0.00035 · I · ti+0.02125 · I · η +0.00114688 · I · U + 0.0000375 · I · P +0.0000464643 · t2 i + 0.0020625 · ti · η −0.00000921875 · ti · U + 0.00001125 · ti · P −3.44143 · η2 + 0.002375 · η · U −0.0020625 · η · P − 0.0000121339 · U2 +0.00000859375 · U · P + 0.0000152143 · P 2

On the other hand, a value of 0.9794 was obtained for the R2-statistic, which signifies that the model explains 97.94% of the variability of MRR, whereas the adjusted R2-statistic (R2

(3)

adj) is 0.9335.

Table 3 Analysis of variance table for the curvature test of the first-order model of MRR Source of variation Sum of squares Degrees of freedom Mean square F-ratio P-value 486.06 0.0002 0.014258 0.000029 Curvature Pure error Total error 0.014258 0.000088 0.014346 1 3 4

C.J. Luis et al. / Journal of Materials Processing Technology 164–165 (2005) 889–896 893

Table 4 Analysis of variance table for the second-order model of MRR Source of variation Sum of squares Degrees of freedom Mean square P-value

F-ratio 219.53 1.53 4.81 101.29 0.12 4.41 1.19 0.11 51.02 0.01 1.29 0.41 1.32 0.49 4.41 2.19 0.41 1.40 1.15 0.14 0.0000 0.2468 0.0560 0.0000 0.7354 0.0652 0.3041 0.7483 0.0001 0.9096 0.2862 0.5367 0.2805 0.5012 0.0652 0.1732 0.5367 0.2666 0.3123 0.7190 0.144901 0.001013 0.003173 0.066856 0.000080 0.002909 0.000784 0.000072 0.033672 0.000009 0.000848 0.000272 0.000870 0.000324 0.002909 0.001444 0.000272 0.000926 0.000756 0.000091 0.000660 A:I B:ti C:η D:U E:P AA AB AC AD AE BB BC BD BE CC CD CE DD DE EE Total error Total 0.144901 0.001013 0.003173 0.066856 0.000080 0.002909 0.000784 0.000072 0.033672 0.000009 0.000848 0.000272 0.000870 0.000324 0.002909 0.001444 0.000272 0.000926 0.000756 0.000091 0.005940 0.287711 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 9 29

where the variable values have been specified according to their original units. The ANOVA table shown in Table 4 can be used in order to simplify the models, especially for the case of polynomial ones. In this way, the simplified model which presents the highest value for the adjusted R2-statistic is that shown in Eq. (4):

MRR = −0.99879 + 0.152933 · I − 0.00529342 · ti

+3.05541 · η − 0.000528938 · U − 0.001375 · P −0.0330263 · I2 + 0.00035 · I · ti +0.00114687 · I · U + 0.0000499342 · t2 i

Fig. 2. Main effects plot for MRR.

−0.00000921875 · ti · U − 3.30263 · η2 +0.002375 · η · U − 0.0000112664 · U2 +0.00000859375 · U · P

(4)

giving the latter a better flushing of the EDM debris. With respect to duty cycle, although its influence is not significant, MRR tends to increase until a maximum peak (near its central value of 0.5) after which MRR decreases. If pause time (to)

where now, the adequacy statistics are 0.9755 (R2) and 0.9555 (R2 adj). Nevertheless, the first model (with no simplifications) is to be considered in this study on MRR in order to take into account the contribution of all the possible effects.

Fig. 3. Interaction plot for MRR.

Figs. 2 and 3 show the main effects plot and interaction plot for material removal rate, respectively. As can be ob- served in Fig. 2, the most influential factors are intensity and open-circuit voltage, where the three remaining factors for a 95% confidence level have no influence on MRR. More- over, material removal rate greatly increases when these two factors are increased, where these tendencies are the ones that one could expect a priori, as more energetic pulses usu- ally lead to a higher material removal and on the other hand, an increase in the ionisation or breakdown voltage results in an increase in both the discharge energy and the work gap,

894 C.J. Luis et al. / Journal of Materials Processing Technology 164–165 (2005) 889–896

mial fitted for the model, it is shown in Eq. (5):

i − 0.339688 · ti · η

Fig. 4. Estimated response surface of MRR vs. I and U.

EW = 108.049 − 26.5842 · I + 0.801646 · ti − 141.888 · η −0.371658 · U + 0.0366417 · P + 2.83293 · I2 −0.0469063 · I · ti + 9.00625 · I · η −0.00842188 · I · U + 0.0172187 · I · P −0.00218018 · t2 −0.000877344 · ti · U − 0.000664063 · ti · P +116.293 · η2 + 0.0942187 · η · U −0.0809375 · η · P + 0.00119558 · U2 +0.000363281 · U · P − 0.00150518 · P 2

(5)

is decreased, MRR will increase but only until a certain value of this factor, as it is always necessary a minimum amount of time between sparks to carry out the flushing of the gap zone.

where the values of the variables are specified in their original units. Taking into account Table 6, the simplified model with the highest adjusted R2-statistic is shown in Eq. (6):

Furthermore, as can be seen in Fig. 3, the most important interaction effect is produced between intensity and voltage, where the rest of interactions are not significant for a confi- dence level of 95%. Fig. 4 presents the estimated response surface of MRR versus intensity and voltage, whereas the other factors are held constant at their central values.

EW = 103.89 − 25.4856 · I + 0.809412 · ti − 131.394 · η −0.339209 · U − 0.157444 · P + 2.69561 · I2 −0.0469063 · I · ti + 9.00625 · I · η −0.00842188 · I · U + 0.0172187 · I · P −0.00252346 · t2

As Fig. 4 clearly shows, there is a strong interaction be- tween intensity and voltage. Thus, the tendency of MRR to increase when intensity is increased depends on the value of open-circuit voltage, in such a way that the latter smoothes as we move in the plot towards lower values of the voltage factor.

i − 0.339688 · ti · η −0.000877344 · ti · U + 102.561 · η2 +0.0942187 · η · U + 0.00110976 · U2 +0.000363281 · U · P

(6)

5. Results and analysis of EW

where here, R2- and adjusted R2-statistics are 0.9558 and 0.9014, respectively. As in the case of MRR, the former model is the one that will be employed in the present study in order to take into account all the considered effects.

In the same way as MRR, a first-order model was proposed for EW, where this is rejected as a result of the curvature test shown in Table 5. As in this case the P-value, which is equal to 0.0002, is lower than 0.05, it is accepted that there is statistical evidence of curvature in the first-order model, for a confidence level of 95% and then, the second-order model is selected.

Table 6 presents the ANOVA table of the second-order model for EW. As can be observed in this table, there are nine effects with a P-value less than 0.05, which means that they are significantly different from zero at the 95% confidence level.

Fig. 5 shows the main effects plot for electrode wear. As can be observed in this figure, the most influential fac- tor on EW is intensity in such a way that EW decreases when this factor is increased, at least until a point after which it tends to increase within the work interval consid- ered in this study. In practice, this is the behaviour that one could expect a priori considering the experience on metallic materials, as an increase in intensity (above all, for low values of this factor) is usually associated with a decrease in the electrode wear, although, if the intensity density through the section of the electrode is excessive, then an increase in EW is produced. Regarding the influ- ence of pulse time on EW, this is not the anticipated one, as EW usually decreases when this last factor is increased

On the other hand, the values obtained for the R2-statistic and the R2-statistic adjusted for the degrees of freedom are 0.9652 and 0.8878, respectively. With respect to the polyno-

Table 5 Analysis of variance table for the curvature test of the first-order model of EW Source of variation Degrees of freedom Mean square F-ratio P-value 492.76 0.0002 70.9703 0.1440 Curvature Pure error Total error Sum of squares 70.9703 0.4321 71.4024 1 3 4

C.J. Luis et al. / Journal of Materials Processing Technology 164–165 (2005) 889–896 895

Table 6 Analysis of variance table for the second-order model of EW Source of variation Degrees of freedom P-value

F-ratio 80.99 19.37 3.85 0.55 5.12 15.12 10.80 9.96 1.39 1.46 1.43 5.67 6.05 0.87 2.55 1.74 0.32 6.90 1.04 0.68 0.0000 0.0017 0.0812 0.4775 0.0499 0.0037 0.0094 0.0116 0.2681 0.2583 0.2618 0.0412 0.0362 0.3763 0.1449 0.2193 0.5845 0.0275 0.3351 0.4299 Mean square 105.56000 25.25240 5.02445 0.71601 6.67342 19.71170 14.08130 12.97800 1.81576 1.89751 1.86792 7.38481 7.88206 1.12891 3.32169 2.27256 0.41926 8.98774 1.35141 0.89033 1.30337 A:I B:ti C:η D:U E:P AA AB AC AD AE BB BC BD BE CC CD CE DD DE EE Total error Total Sum of squares 105.56000 25.25240 5.02445 0.71601 6.67342 19.71170 14.08130 12.97800 1.81576 1.89751 1.86792 7.38481 7.88206 1.12891 3.32169 2.27256 0.41926 8.98774 1.35141 0.89033 11.73030 336.75200 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 9 29

or, what it is the same, when the sparks frequency is dimin- ished.

Fig. 6. Interaction plot for EW.

With respect to the third significant factor for EW, i.e., flushing pressure, an increase in the latter tends to improve the wear on the electrode, at least for the work interval stud- ied. An adequate flushing of the gap zone is really important for the stability of the EDM process and therefore, an in- crease in the dielectric pressure leads to a better behaviour of the electrode wear. Of course, excessively high values for the flushing pressure are expected to create turbulence in the dielectric and, as a consequence of that, a worsening in the electrode wear. On the other hand, as can be observed in Fig. 5, the lowest values for EW are obtained when duty cy- cle and voltage take values near their central values (0.5 and −160 V, respectively).

The interaction plot for the electrode wear can be observed in Fig. 6. As has been stated earlier, the strongest interactions are those between intensity and pulse time (AB) and intensity

and duty cycle (AC). As can be appreciated in Fig. 6, the interaction effects between pulse time and voltage (BD) and pulse time and duty cycle (BC) are also significant (the curves for their corresponding low and high levels are not parallel in the interaction plot) but to a lesser degree.

Fig. 7 shows the estimated response surface of the elec- trode wear as a function of the design factors of intensity and

Fig. 5. Main effects plot for EW. Fig. 7. Estimated response surface of EW vs. I and ti.

896 C.J. Luis et al. / Journal of Materials Processing Technology 164–165 (2005) 889–896

techniques have been employed to model the previously men- tioned response variables by means of equations in the form of polynomials.

Fig. 8. Estimated response surface of EW vs. I and P.

The design finally chosen to accomplish the present study was a fractional factorial one of type 25−1 and whose res- olution is V. The design factors selected in this case were: intensity (I), pulse time (ti), duty cycle (η), open-circuit volt- age (U) and flushing pressure (P), where all of them, except for the last one, are parameters widely used by the machin- ists to control the EDM machine generator. First-order mod- els were proposed for MRR and EW but both of them were rejected from the results obtained in their corresponding lack- of-fit tests. Thus, second-order models were finally selected to carry out this study for both response variables MRR and EW.

In the case of MRR, the only influential design factors, for a confidence level of 95%, were: intensity and voltage. The variation tendencies of the latter were those expected a priori and according to these, in order to obtain a high value of MRR within the work interval of this study, both design factors (I and U) should be fixed as high as possible.

Fig. 9. Estimated response surface of EW vs. ti and P.

pulse time. As is depicted in this figure, EW diminishes when intensity is increased but to a lesser extent for low values of pulse time, due to the existence of a statistically significant interaction between the factors plotted in the graph.

With regard to EW and arranged in descending order of importance, intensity, pulse time and flushing pressure turned out to be the influential factors for a confidence level of 95%. The variation tendency of EW obtained in the case of inten- sity was the one that was expected in advance, whereas the opposite behaviour was obtained in the case of pulse time. Moreover, in the case of flushing pressure, it was verified that an increase in the latter (within the considered work in- terval, i.e., 20–60 kPa) resulted in a decrease in the wear on the electrode.

Figs. 8 and 9 depict the estimated response surfaces for EW versus intensity and flushing pressure and versus pulse time and flushing pressure, respectively. As one can see in these figures, the wear on the electrode tends to decrease when the dielectric flushing pressure is increased, for any value of the two above-mentioned design factors.

References

[1] I. Puertas, C.J. Luis, J. Mater. Process. Technol. 143–144 (2003)

6. Conclusions

521–526.

[2] I. Puertas, C.J. Luis, Rev. Met. Madrid 38 (5) (2002) 358–372. [3] J.H. Zhang, T.C. Lee, W.S. Lau, J. Mater. Process. Technol. 63 (1997) 908–912. [4] S.H. Lee, X.P. Li, J. Mater. Process. Technol. 115 (2001) 344– 358.

The enormous interest in the study of die-sinking EDM when machining conductive ceramics is a consequence of the problems encountered when utilising other manufacturing processes such as: laser machining, abrasive jet machining, ultrasonic machining or grinding, among others.

[5] I. Puertas, C.J. Luis, Proceedings of the International Conference on Advances in Materials and Processing Technologies (AMPT’03), vol. II, Dublin, Ireland, 2003, pp. 877–880. [6] E.B. Guitrau, The EDM Handbook, Hanser Gardner Publications, Cincinnati, 1997. [7] D.C. Montgomery, Design and Analysis of Experiments, 5th ed., Wi- ley, New York, 2001.

[8] C.J. Luis, Int. J. Prod. Res. 40 (10) (2002) 2245–2268. [9] C.J. Luis, Int. J. Prod. Res. 40 (12) (2001) 2865–2881.

In this work, a study on the influence of the most relevant EDM factors over material removal rate (MRR) and elec- trode wear (EW) has been carried out. The study has been made for a conductive ceramic such as siliconised silicon carbide (SiSiC) and only for the finish stages. In order to achieve this, DOE and multiple linear regression statistical