Iem

SMa

b

c

d

1

Enmgimbdma

T

0h

Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749

Contents lists available at ScienceDirect

Process Safety and Environmental Protection

journa l homepage: www.e lsev ier .com/ locate /psep

nfluence of workpiece materials on aerosolmission from die sinking electrical dischargeachining process

. Thiyagarajana, S.P. Sivapirakasama,∗, Jose Mathewb,. Surianarayananc, K. Sundareswarand

Department of Mechanical Engineering, National Institute of Technology, Tiruchirappalli, IndiaDepartment of Mechanical Engineering, College of Engineering, Kallooppara, Kerala, IndiaCell for Industrial Safety and Risk Analysis, Central Leather Research Institute, Adyar, Chennai, IndiaDepartment of Electrical and Electronics Engineering, National Institute of Technology, Tiruchirappalli, India

a b s t r a c t

Simultaneous investigation of environmental emissions and machining aspects of electrical discharge machining

process is essential for achieving hygienic and efficient machining. The main objective of the present work is to

experimentally investigate and analyze the aerosol emission rate and the material removal rate from a die sinking

electrical discharge machining process for three commonly used work piece materials viz., tool steel, mild steel and

aluminum using Taguchi methodology of Experimental Design in order to suggest suitable process conditions for

green manufacturing. The aerosol emission profile of all workpiece materials was found to be closely related to the

material removal profile. A significant variation in emission and material removal rate was observed for workpiece

materials which may be accorded to the variation in melting and vaporization temperatures. It was also observed

that majority of aerosol constituents evolved from workpiece materials and that the constituents with low melting

points were having high relative concentration in the aerosol emitted. The study introduced a parameter, the relative

emission rate for comparing the emission for various process parameters and workpiece–tool material combinations.

The favorable machining parameters for each material were then identified by employing signal to noise ratio analysis

of the relative emission rate.

© 2014 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keywords: Electrical discharge machining (EDM); Emission; Aerosol; Taguchi methodology; Relative emission rate

the plasma channel is restricted by the surrounding dielectric

. Introduction

lectrical discharge machining (EDM) is an importanton-traditional manufacturing process, typically used forachining very hard and brittle materials. The schematic dia-

ram of EDM process is shown in Fig. 1. The basic principlen the EDM is the conversion of electrical energy into ther-

al energy through a series of electrical discharges occurringetween the electrodes (work piece and tool) immersed in aielectric fluid. A series of voltage pulses usually of the order ofagnitude 20–100 V are applied between the electrodes with

frequency of 5 kHz to 5 MHz. The peak current and pulse

∗ Corresponding author at: Department of Mechanical Engineering, Natel.: +91 431 2503408; fax: +91 431 2500133; mobile: +91 9944547215.

E-mail addresses: spshivam@nitt.edu, spsivam@yahoo.com (S.P. SivReceived 27 June 2013; Received in revised form 7 January 2014; AccepAvailable online 24 January 2014

957-5820/$ – see front matter © 2014 The Institution of Chemical Engittp://dx.doi.org/10.1016/j.psep.2014.01.001

duration are the main process parameters of EDM governingthe process energy. The dielectric fluid stored in a sump tank isflushed continuously through the inter-electrode gap using apump. The height of dielectric surface above the spark locationis termed as dielectric level. As the pulse in the EDM pro-cess begins, the passage of pre-breakdown current heats thedielectric liquid in the inter electrode gap. When the dielec-tric strength of the liquid in the gap is exceeded, breakdownoccurs and initiates a plasma channel. This plasma channelexpands during the following pulse on time. The growth of

ional Institute of Technology, Tiruchirappall 620015, India.

apirakasam).ted 13 January 2014

fluid, which in turn causes a higher concentration of discharge

neers. Published by Elsevier B.V. All rights reserved.

740 Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749

Fig. 1 – Schematic diagram of EDM process.

energy in a very small volume, which causes the erosion. Thecumulative effect of a succession of sparks spread over theentire work piece surface leads to its erosion to a shape whichis approximately complimentary to that of the tool.

The EDM machines used in small and medium sizedindustries are manually operated. An operator will be simulta-neously operating more than one machine. EDM installationsare open systems in which the dielectric tank is exposed towork atmosphere. The high temperature developed in thedischarge channel of the EDM process causes the generationof toxic reaction products and metallic particulates that willbe released into the work environment with the potentialof causing severe occupational and environmental problems(Bommeli, 1983; Tonshoff et al., 1996). Since the discharge tem-perature of the EDM process depends on peak current andpulse duration, the deviations of these parameters can influ-ence the emission. The flushing pressure and dielectric levelthat affect the dynamics in the dielectric may also influencethe emission.

A variety of tools and dielectric materials are beingemployed by the manufacturer based on the work materialand the desired machining performance of EDM process. Gen-erally, one material among, copper/tungsten, alloys of copperor tungsten/graphite is used as the tool electrode and mineraloils/deionised water/commercial hydrocarbon oils are usedas the dielectric fluid. Kerosene, a blend of hydrocarbons, iswidely used in this process due to its high flash point, gooddielectric strength, transparent characteristics, low viscosityand low specific gravity (Bhattacharyya et al., 2007). Differ-ent combinations of workpiece, tool and dielectric fluids alsoresult in the variation in the aerosol emission and its con-stituents (Evertz et al., 2006).

In order to provide suitable control measures to safeguardthe operator, an estimate of the concentration of emissions isrequired. The concentration of aerosol in the work atmospheredepends on the rate at which it is emitted from the dielectricsurface. The emission rate can be controlled by using effec-tive fume extraction systems. Small and medium sized EDMmachines considered in this study are not equipped with fumeextraction systems. Though the sophisticated EDM machinesare using fume extraction systems, there is a lack of a pro-cess specific design due to the insufficient knowledge on theemission with varying process parameters. This draws theattention for a systematic study of emission with varying

process parameters at different combinations of workpiecematerials. Such a study could further help in selecting the

best combination of process parameters that could reduce theemissions.

Good information is available on aerosol exposure onmachine shop workers (NIOSH, 2003a; NIOSH, 2007; Jaakkolaet al., 2009). Nevertheless, literature on experimental investi-gation of emission from the EDM process is scanty. The firstreport on emissions from the EDM process was by Bommeli(1983). Through experimental investigation, he identified thecomponents of emission generated from EDM process whenhydrocarbon based and water based dielectric fluids wereused. The hydrocarbon-based dielectric fluids caused emis-sion of aliphatic, aromatic, polycyclic aromatic hydrocarbons,carbon and metallic particles, whereas water-based dielectricfluids caused emission of toxic constituents like ozone, COand metallic particles. A comprehensive review of safety andenvironmental aspects of the EDM process was presented byTonshoff et al. (1996). Leao and Pashby (2004) presented theenvironmental impact resulting from the use of die sinkingEDM process. It was observed that emission of toxic sub-stances and generation of toxic wastes was the importantenvironmental issues that caused the health problems to theoperators and land and water pollution. The study depictedthat consumption of high amount of energy by the EDMprocess was one of the factors that led to environmental prob-lems.

In a manufacturing industry, though the emission is a crite-rion, productivity and quality aspects of the process are alsovital (Tan et al., 2002). In case of the EDM process, materialremoval rate is one of the significant performance parametersthat need special attention. By varying process parametersat different combination of materials from the perspectiveof reducing exposure may at the same time, considerablyaffect performance characteristics of this process. Therefore,a simultaneous study of the emission rate and the MRR of theprocess become inevitable. Since the analysis of emissions isboth expensive and time consuming, the Taguchi methodol-ogy of Design of Experiment (DoE) was adopted to limit the noof experiments in this work (Peace, 1992).

The present investigation was conducted in the diesinking EDM machine, in a laboratory environmentusing kerosene as a dielectric fluid. This work hadmultiple objectives; the first being to experimentallyinvestigate the aerosol emission as well as the mate-rial removal rate from die sinking electrical dischargemachining process for three different commonly usedwork piece materials viz., tool steel, mild steel and alu-minum. The other objectives were to; analyze and comparethe constituents of aerosol emission generated for each ofthe work piece materials for identifying the intensity ofreactions taking place during process, analyze the influenceof process parameters on the relative emission (i.e. ratio ofemission rate to the material removal rate) using the Taguchimethodology of Experimental Design for suggesting suitableprocess conditions for green manufacturing.

2. Materials and methods

Experiments were conducted using the conventional diesinking electric discharge machine manufactured by VictoryElectromech, Pune, India. Since majority of EDM opera-tions include drilling holes of various shapes, a blind

hole of 25 mm diameter was drilled on work pieces of40 mm × 40 mm × 15 mm size. Copper was used as the tool

Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749 741

Table 1 – Composition and properties of workpiece materials.

Components Mild steel Tool steel Aluminum% composition % composition % composition

CompositionIron 98.79 84.77 0.123Chromium 0.137 12.3 0.023Carbon 0.174 2.05 n.dSilicon 0.143 0.215 0.167Manganese 0.603 0.2 0.0015Nickel 0.0819 0.256 0.017Phosphorus 0.023 0.0087 0.015Molybdenum 0.0203 0.058 0.0146Aluminum 0.0094 n.d. 99.27Titanium 0.0043 n.d. 0.0024Sulfur 0.014 0.0336 n.d.Lead 0.0011 n.d. 0.0489Vanadium n.d. 0.0514 n.d.Cobalt n.d. 0.0784 n.d.Copper n.d. n.d. 0.005Zinc n.d. n.d. 0.0081Calcium n.d. n.d. 0.0146Magnesium n.d. n.d. 0.332

PropertiesMelting point (◦C) 1523 1421 660Boiling Point (◦C) 3300 3134 2467Thermal conductivity (W/m-◦K) 51.9 76.2 227Specific heat capacity (J/g-◦C) 0.472 0.461 0.9

n.d. – not detected.

etmcs

2

TtopTTtpsfwmlto

aipt

iwpcse

lectrode. Kerosene, which is the most commonly used dielec-ric fluid, was employed for the present study. Workpiece

aterials used were mild steel, tool steel and aluminum. Theomposition and properties of workpiece materials are pre-ented in Table 1.

.1. Experimental design

he Taguchi Methodology of DoE, which is capable of iden-ifying the influence of input parameters on the emissionf aerosols and material removal rate, employed was in theresent study by conducting least number of experiments.he experiments were designed based on L9 orthogonal array.his basic design makes use of up to four parameters, with

hree levels each. The peak current, pulse duration, flushingressure and dielectric level were the process parameters con-idered in this study. Three levels within the operating rangeor the machining of small and medium sized componentsere selected for each of the factors. A total of nine experi-ental runs were to be conducted, using the combination of

evels for each input factor as indicated in Table 2. The range ofhe input parameters is selected based on the operating rangef small and medium sized EDM machines.

The contribution of each factor on emission of aerosolnd material removal rate were estimated. Since the exper-mental design is orthogonal, the mean effect of each processarameter at different levels can be estimated by averaginghe outputs corresponding to each factor level combination.

In Taguchi method, any repetitive data in an experiments transformed into a consolidated value called the S/N ratio,

hich represents the amount of variation present in the out-ut response. The equation for S/N ratio depends on theriterion of the performance parameter to be analyzed. In this

tudy the S/N ratio was calculated for the parameter, relativemission (ratio of emission rate and material removal rate).

Since a lower value is desirable in the case of relative emis-sion rate, the following equation for ‘lower the better’ type ofS/N ratio was applied.

� = −10 log

[1n

n∑i=1

yi

](1)

where, � is the S/N ratio yi is the measured output value forthe ith repetition and n is the number of repetitions in a trial.

Analysis of variance (ANOVA) was carried out by dividingthe total variability of S/N ratios into contributions by each ofthe process parameter and the error. The following equationswere used to calculate sum of the squared deviations (SS)T andthe sum of squared deviation due to each process parameter(SS)P.

(SS)T =m∑

i=1

�2i − m�m

2 (2)

(SS)P =t∑

j=1

(s�j)2

t− 1

m

[m∑

i=1

�i

]2

(3)

where m represents the number of experiments in an ortho-gonal array, �i the mean S/N ratio for the ith experiment, �m

the total mean of S/N ratio, j the level number of the processparameter p, t the repetition of each level of the parameterp and s�j the sum of the S/N ratio involving the parameterp at level j. The variance of the process parameters (VP) wascalculated by using.

VP = (SS)P(df )P

(4)

742 Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749

Table 2 – L9 orthogonal array.

Exp. no. Current (A) Pulse duration (�s) Dielectric level (mm) Flushing pressure (kg/cm2)

1 2 2 40 0.32 2 261 60 0.53 2 520 80 0.74 4.5 2 60 0.75 4.5 261 80 0.36 4.5 520 40 0.57 7 2 80 0.58 7 261 40 0.7

9 7 520

where (df)P was the degree of freedom of the process param-eter = t − 1. Since it was a saturated design where all columnswere assigned with factors, the variations due to error wereestimated by pooling the estimates of the factors having leastvariance. The corrected sum of squares (S)P was calculated as:

(S)P = (SS)P − (df )pVe (5)

The percentage contribution � was calculated as:

� = (Sp)(SST)

(6)

Fig. 2 – Effect of p

60 0.3

2.2. Sampling of aerosol

In order to estimate the emission rate of the EDM process,the machine was completely enclosed by a transparent hood.Bottom portion was connected to fresh air inlet to compensatethe account of air sucked using the pump. Air was allowedto pass through a glass fiber filter paper to collect airborneaerosols. Weight of the filter paper was taken before and aftersampling using a sensitive balance (accuracy ±0.01 mg). Therate of emission of aerosols (AE) into the work atmosphere wascalculated using the following equation.

AE = wb − wa

ts(7)

eak current.

Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749 743

of pu

wam

2

I(alrs“pag5lT(l

C

Fig. 3 – Effect

here wa and wb are the weight of filter paper before andfter the sampling (in mg) and ts is the sampling duration (ininutes).

.3. Analysis of metallic constituents

nductively Coupled Plasma-Atomic Emission SpectrometryICP-AES) was used to analyze the metallic constituents oferosol samples because of its special characteristics such asow detection limits, multi element detection, wide dynamicange and good precision. The analysis was done as per thetandard procedure developed by NIOSH for the analysis ofElements by ICP” (method no.7301) (NIOSH, 2003b). The filteraper samples with aerosol were completely digested usingquaregia (1HNO3:3HCl) and diluted to 25 ml using 5% aquare-ia solution. The analysis was done by ICP-AES model OPTIMA300DV of make Perkin Elmer. The aerosol samples were ana-yzed for the constituents of workpiece and tool materials.he concentrations of metallic constituents of the sample, Cs

mg/ml), were obtained and the rates of emission of the metal-ic particles were calculated using the following equation.

M = Cs × Vs

ts(8)

lse duration.

where, CM is the rate of emission of metallic particles (mg/min)and Vs is the volume of sample(ml).

The percentage of each metallic constituent was calculatedusing the following equation.

PM = 100 × CM

AE(9)

where, PM is the percentage of metallic particle in emission.Three samples of aerosol from each workpiece materials wereanalyzed for its metallic constituents and its average percent-age was calculated.

2.4. Analysis of material removal rate

The material removal rate (MRR) was calculated by taking theweights of the work piece before and after the experiment.

MRR = (Wwb − Wwa)tm

(10)

In which, WWb and WWa are weights of the workpiece in

mg before and after machining and tm is the machining timein minutes.

744 Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749

le3

–Ex

per

imen

talr

esu

lts.

.no.

Alu

min

um

Mil

dSt

eel

Tool

Stee

l

MR

R(m

g/m

in)

Emis

sion

(mg/

min

)R

elat

ive

emis

sion

rate

MR

R(m

g/m

in)

Emis

sion

(mg/

min

)R

elat

ive

emis

sion

rate

MR

R(m

g/m

in)

Emis

sion

(mg/

min

)R

elat

ive

emis

sion

rate

102.

681.

170.

011

65.4

50.

490.

007

82.6

70.

790.

0127

.67

0.93

0.03

420

.69

0.29

0.01

439

.07

0.63

0.01

620

.33

0.67

0.03

319

.35

0.24

0.01

236

.60.

520.

014

304.

672.

200.

007

93.6

51.

220.

013

127.

521.

650.

013

311

3.93

0.01

314

2.05

2.36

0.01

617

5.7

2.87

0.01

622

5.37

3.23

0.01

410

3.72

2.11

0.02

152.

22.

460.

016

260.

672.

270.

009

132.

452.

080.

016

147.

72.

420.

016

365.

675.

870.

016

245.

194.

450.

018

252.

4.75

0.01

929

0.58

5.13

0.01

821

9.36

4.28

0.01

922

6.76

4.69

0.02

2.5. Relative emission rate

The quantity of aerosol emission from the process as com-pared with that of the work piece erosion is referred to asthe relative emission rate (RER). This parameter indicates theamount of emission per unit material removal and was calcu-lated using the equation.

RER = AE

MRR(11)

3. Results and discussion

The experiments are conducted as per the orthogonal arrayand the results of MRR, aerosol emission and relative emissionrate are presented in Table 3. It can be observed from this tablethat the output responses MRR and emission vary significantlywith the process parameters. Both these output responsesshow a similar trend. Also these parameters are dependent onworkpiece material. The relative emission rate also is found tobe varying with the process parameters. Detailed discussionon the influence of workpiece materials on aerosol emissionis presented in the following sections.

3.1. Effect of peak current

The effect of peak current on the emission of aerosol, MRR,and relative emission rate for different workpiece materialsare presented in Fig. 2. Emission of Aerosol and MRR increasedwith increase in peak current due to the increase in meltingand vaporization of electrodes (workpiece and tool) at highvalues of peak current. These values were the highest for theworkpiece material with the lowest melting and vaporizationtemperatures (aluminum) and the lowest for the workpiecematerial with the highest melting and vaporization tempera-tures (mild steel).

For aluminum, the relative emission rate was very high atlow value of peak current due to which, a significant portion ofthe eroded material was escaped into the atmosphere as emis-sions. Due to the low melting point and high heat conductivityof aluminum, more materials would be melted and removedfrom the workpiece, compared to other materials. At low val-ues of peak current, the size of molten metallic particles wouldbe low. So a significant portion of the removed materials wouldbe carried away by the emission, thereby increasing the rela-tive emission rate. At high values of peak current (>2 A) sincethe size of molten particles was increasing, emission is foundto depend mainly on vaporization of workpiece materials. Thevaporized fraction of the molten material for aluminum waslow because of high latent heat of vaporization due to whichthe relative emission rate significantly reduced at mediumvalues of temperature (at a current of 4.5 A). As the peak cur-rent increased above 4.5 A, the vaporization fraction graduallyincreased due to increase in heat energy. This may be thereason for the increase in relative emission rate at higher val-ues of peak current.

For other materials (mild steel and tool steel) the relativeemission rate increased slightly with increase in the peak

current. This trend may be due to the increase in vaporizedfraction of the workpiece with the increase in peak current. Ta

b

Exp

1 2 3 4 5 6 7 8 9

Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749 745

of di

3

Tsw(ttsadurt1a

iticmvsr

Fig. 4 – Effect

.2. Effect of pulse duration

he effect of pulse duration on the output parameters is pre-ented in Fig. 3. The aerosol emission and MRR increasedith increase in pulse duration up to a medium value

261 �s) and then slightly decreased. As the pulse-on dura-ion was increased, the amount of heat energy transferredo the workpiece and tool surfaces were increased and con-equently the material removal and the aerosol emissionlso increased. The reduction after a medium value of pulseuration may be due to the fact that, beyond a partic-lar value of spark radius, the temperature of the outeregion would be considerably reduced, causing a reduc-ion in the workpiece and tool vaporization (Eubank et. al.,993) and consequent material removal and emission oferosol.

The relative emission rate was found to be increasing withncrease in pulse duration. As the pulse duration increased,he energy transferred to the workpiece and tool surfacencreases. So, more molten material would be vaporizedausing an increase in relative emission. Eventhough theaterial removal and emission was reduced after medium

alue of pulse duration, the relative emission increased

lightly due to the increase in vaporization fraction of theemoved material.

electric level.

3.3. Effect of dielectric level

Graphs presented in Fig. 4 depict the effect of dielectric levelon aerosol emission rate, MRR and relative emission raterespectively. For all materials the aerosol emission showed adecreasing tendency with an increase in the dielectric levelabove the spark location. As the dielectric level increased,more vapors would be condensed and/or precipitated in thedielectric itself which caused a decrease in the emission oftotal aerosol. The material removal rate was found to bedecreasing with an increase in the dielectric level. This wasdue to the reason that the flushing is more effective at lowdielectric level. However, this variation was not significant. Forall materials the relative emission was found to be higher ata medium value of flushing dielectric level (60 mm) and lowerat a higher value of dielectric level (80 mm). This is due to thesignificant reduction of material removal rate from 40 mm to60 mm of dielectric level. The decrease in relative emissionrate at 80 mm was due to the effect of condensation of vaporsduring transport.

3.4. Effect of flushing pressure

Fig. 5 represents the effect of flushing pressure on the emis-sion rate of aerosol, MRR and relative emission rate. It can be

746 Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749

f flus

Fig. 5 – Effect o

observed from this figure that the emission of aerosol and MRRdecreased with increasing flushing pressure up to 0.5 kg/cm2.Beyond 0.5 kg/cm2 both these parameters got increased. Theincreasing flushing pressure at low range (0.3–0.5 kg/cm2)caused increase in turbulence in the dielectric medium, whichin turn influenced the reduction in heat transfer rate in theworkpiece material. This reduction caused, a reduction inmaterial removal as well as emission. At higher values offlushing pressure (0.5–0.7 kg/cm2) more molten materials werecarried away by the dielectric level resulting in an increase inthe material removal and emission rate of aerosol.

The relative emission rate was found to be higher at amedium value of flushing pressure due to the significantreduction in material removal. This parameter was found to belower at a lower level of flushing pressure for low melting pointmaterial (aluminum) and at high level of flushing pressure fora high melting point material (mild steel). The reduction foraluminum was due to the high rate of material removal at lowflushing pressures.

3.5. Effect of workpiece materials

From the above discussion it is clear that the influence ofprocess parameters was similar for all workpiece materials.So the interaction between process parameters and work-

piece materials was not significant. The emission of aerosolwas the highest when aluminum was used as the workpiece

hing pressure.

and was the lowest when mild steel was used as the work-piece. Aluminum was the material with the lowest meltingand boiling temperatures and mild steel was the material withthe highest melting and boiling temperatures. These resultsindicate that there is a strong correlation between aerosolemissions and melting and boiling temperatures of workpiecematerial. As these workpiece parameters decreased, moreworkpiece constituents were melted and evaporated resultingin an increasing the aerosol emission.

The average values of relative emission rate for the work-piece materials under consideration are presented in Fig. 6. Itcan be observed that the relative emission was higher for a lowmelting point material (aluminum) and lower for a high melt-ing point material (mild steel). For low melting point materialsa significant portion of the heat supplied by the spark was usedto increase the vaporization of the electrode and dielectricmaterials that aids in the increase of relative emission.

3.6. Constituents of aerosol samples

The constituents of aerosol samples were analyzedand the average concentration was calculated. Themajor metallic constituents of aerosol generated whilemachining different workpiece materials were presentedin Figs. 7–9. From these figures it is clear that the major

portion of the constituents belongs to the workpiece material(aluminum, iron, chromium etc.). This was due to the fact

Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749 747

Fig. 6 – Effect of workpiece material of relative emissionr

twacotfia

siit

3

SoaTpu

tp

dbiwro

Fa

Fig. 8 – Composition of aerosol generated while machining

The surface roughness is improved in final stage with low

ate.

hat, while machining more heat was concentrated in theorkpiece side as explained by Eubank et al. (1993). The ratioluminum:copper (Fig. 7) was higher compared to that of iron:opper ratios (Figs. 8 and 9). This was because of lower valuesf melting and boiling temperatures of aluminum comparedo iron. 30–40% of the aerosol samples include unidenti-ed components. It may include carbon and hydrocarbonsttached to the aerosol.

In the case of alloys (tool steel) the composition in the emis-ion was found to be dependent on the melting temperature ofndividual components. Since the melting point of chromiums high compared to that of iron, the fraction of chromium inhe emission was lower compared to tool steel.

.7. Statistical analysis on relative emission rate

tatistical analysis of the relative emission rate was carriedut in order to find out the optimum parameter settings andnalyze the contribution of each factor to output responses.he signal to ratio (S/N) was helpful in optimizing the processarameters. The significance of each factor was then analyzedsing analysis of variance (ANOVA).

The S/N ratio calculated for the relative emission rate andhe mean S/N ratio at three levels of process parameters areresented in Table 4.

Regardless of the type of performance characteristicsesired (higher the better, lower the better or nominal theetter), a large S/N ratio implied better performance character-

stics. Therefore, the optimal level of the process parametersas that with the highest S/N ratio. Since lower value of

elative emission rate is desirable, lower output response is

ptimal. It can be observed from tables that, for the relative

ig. 7 – Composition of aerosol generated while machiningluminum.

mild steel.

emission rate, a lower value of peak current is desirable fortool steel and mild steel whereas a medium value is desirablefor aluminum. Lower values of pulse duration, dielectric leveland flushing pressure were found to be favorable.

The ANOVA results for output responses are presented inTables 5–7. From these tables the contribution of each machin-ing parameter on output responses could be observed.

From the S/N ratio tables and ANOVA results, it can beobserved that the peak current was the most influentialparameter in case of mild steel and tool steel, whereas pulseduration was the most significant parameter in case of alu-minum. The second most influential parameters were; pulseduration in case of tool steel and mild steel and peak currentin case of aluminum. The significance of other factors (flush-ing pressure and dielectric level) was comparatively less in allthe cases.

Optimum values of current and pulse duration discussedabove could lead to a substantial reduction in productivity.Consequently, sheer optimization of process parameters isnot a feasible solution for occupational exposure to emissionsfrom EDM process. Control methods such as fume extractionsystem, local exhaust ventilation system and administrativecontrols should be in force to reduce the risk of exposure.Major quality aspect of the process is surface roughness.

Fig. 9 – Composition of aerosol generated while machiningtool steel.

748 Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749

Table 4 – S/N ratios.

Workpiece material Factor Level I Level II Level III Max–min

Mild steel I 45.17172 41.11139 40.33319 4.838532tp 44.62927 41.24702 40.74001 3.889268h 42.66146 41.81516 42.13968 0.846296fp 43.09571 41.0551 42.46548 2.04061

Tool steel I 43.43684 41.95596 39.87175 3.565088tp 43.69765 40.70819 40.85871 2.989459h 42.48875 41.17698 41.59881 1.311772fp 42.14298 41.21328 41.90828 0.929701

Aluminum I 37.61453 45.15001 43.04405 7.53548tp 47.17663 39.63946 38.9925 8.18413h 42.84777 41.18695 41.77387 1.660823fp 42.94623 41.25836 41.60401 1.687865

I – peak current, tp – pulse duration, h – dielectric level, fp – flushing pressure.

Table 5 – Results of ANOVA for mild steel.

Source Df SS V S % contribution

I 2 40.50326 20.25163 39.40948 52.56657tp 2 26.82312 13.41156 25.72933 34.31923ha 2 1.093782 0.546891 5.835792fp 2 6.550446 3.275223 5.456664 7.278405

Total 8 74.9706 37.4853 100

a Factor pooled into error.

Table 6 – Results of ANOVA for tool steel.

Source Df SS V S % contribution

I 2 296.4057 148.2029 281.2075 62.04tp 2 124.2692 62.13458 109.0709 24.06ha 2 15.19823 7.599113 13.41fp 2 17.41206 8.70603 2.213829 0.49

Total 8 453.2852 56.66065 100

a Factor pooled into error.

Table 7 – Results of ANOVA for aluminum.

Source Df SS V S % contribution

I 2 90.6982 45.3491 86.44213 38.60202tp 2 124.2075 62.10373 119.9514 53.56607ha 2 4.256073 2.128037 7.602451fp 2 4.769918 2.384959 0.513844 0.229465

Total 8 223.9316 111.9658 206.9074 100

aFactor pooled into error.

current values. The rate of emission during this step is lowsince current is very low. So the optimization of processparameters will not affect the quality aspects of the process.

4. Conclusions

An experimental investigation on aerosol emission, and MRRof EDM process for three different work piece materials viz.,tool steel, mild steel and aluminum was conducted in thisstudy. The metallic constituents of aerosol emissions wereanalyzed and compared with that of workpiece materials

used. Also the influence of process parameters on the rela-tive emission rate (i.e. ratio of emission rate to the MRR) was

analyzed. The following conclusions were derived from thisstudy.

• The aerosol emission profile of all workpiece materials wasclosely related to the material removal profile. The influenceof process parameters on aerosol emissions was similar forall the workpiece materials considered.

• There is a significant variation in emission and MRR for theworkpiece materials due to the variation in melting andvaporization temperatures. These parameters were higherfor aluminum (low melting point material) and lower formild steel (high melting point material).

• Major constituents of aerosol belong to the workpiece mate-rials. The melting point of the constituents has a strong

Process Safety and Environmental Protection 9 2 ( 2 0 1 4 ) 739–749 749

•

•

•

(sopfpdscTrsw

A

TfPa

influence on the emission. Constituents with low melt-ing points were having high relative concentration in theaerosol emitted.The study introduced the parameter relative emission ratefor comparing the emission on various process parame-ters and workpiece tool material combinations. The relativeemission varies with process parameters and workpiecematerials. It was found that the relative emission rate washigher for materials with low melting temperature (alu-minum).It was observed from the S/N ratio analysis for relative emis-sion rate that a lower value of peak current is desirablefor tool steel and mild steel whereas a medium value isdesirable for aluminum. Lower values of pulse duration,dielectric level and flushing pressure were favorable.ANOVA results on the relative emission rate show that thepeak current and pulse duration were the most influentialprocess parameters.

Optimum values of peak current (<4.5 A) and pulse duration<261 �s) which could reduce the emission may lead to a sub-tantial reduction in material removal rate which is a measuref productivity in EDM process. The optimum values of thesearameters for increased production were based on manu-acturing condition and machine specifications. Decrease inroduction rate increase the cost of operation and exposureuration. So, alternate control methods like fume extractionystem, local exhaust ventilation system and administrativeontrols should be employed to reduce the risk of emission.he relationship between process parameters and emissionate identified in this study could help to develop a processpecific fume extraction system that operate according to theorkpiece material and process parameters.

cknowledgements

he authors are grateful to the Director, NIT Tiruchirappallior providing the facilities to carry out the experimental work.

art of this work is supported by the Ministry of Environmentnd Forests, Government of India (F.No.19/102/2008-RE).

References

Bhattacharyya, B., Gangopadhyay, S., Sarkar, B.R., 2007. Modellingand analysis of EDMed job surface integrity. J. Mater. Process.Technol. 189, 169–177.

Bommeli, B., 1983. Study of the harmful emanations resultingfrom the machining by electro-erosion. In: Proceedings of theSeventh International Symposium on Electromachining (ISEMVII), pp. 469–478.

Eubank, P.T., Patel, M.R., Barrufet, M.A., Bozkurt, B., 1993.Theoretical models of the electrical discharge machiningprocess. III. The variable mass, cylindrical plasma model. J.Appl. Phys. 73 (11), 7900–7909.

Evertz, S., Dott, W., Eisentraeger, A., 2006. Electrical dischargemachining: occupational hygienic characterisation usingemission based monitoring. Int. J. Hyg. Environ. Health 209(55), 423–434.

Jaakkola, M.S., Suuronen, K., Luukkonen, R., Jarvela, M., Tuomi, T.,Alanko, K., Makela, E.A., Jolanki, R., 2009. Respiratorysymptoms and conditions related to occupational exposuresin machine shops. Scand. J. Work Environ. Health 35 (1), 64–73.

Leao, F.N., Pashby, I.R., 2004. A review on the use ofenvironmentally-friendly dielectric fluids in electricaldischarge machining. J. Mater Process Technol. 149, 341–346.

NIOSH, 2003a. Aerosol Program Assessment Committee Report,Available at:www.cdc.gov/niosh/topics/aerosols/pdfs/ACGreport.pdf(accessed 18.08.09).

NIOSH, March 2003b. NIOSH Manual for Analytical Methods.Elements by ICP, Method 7301. Issue 1.

NIOSH, 2007. Health Hazard Evaluation Report: Report onRespiratory and Dermal Conditions Among Machine ShopWorkers, Pittsburg, Kansas. U.S. Department of Health andHuman Services, Public Health Service, Centers for DiseaseControl and Prevention, National Institute for OccupationalSafety and Health, NIOSH HETA No. 2007-0263-3069,Cincinnati, OH.

Peace, G.S., 1992. Taguchi Methods – A Hands-On Approach.Addision-Wesley, MA, USA.

Tan, X.C., Liu, F., Cao, H.J., Zhang, H., 2002. A decision-makingframework model of cutting fluid selection for greenmanufacturing and a case study. J. Mater. Process. Technol.129, 467–470.

Tonshoff, H.K., Egger, R., Klocke, F., 1996. Environmental and

safety aspects of electrophysical and electrochemicalprocesses. Ann. CIRP – Manuf. Technol. 45 (2), 553–568.