You are on page 1of 25


INTRODUCTION 1.1 Introduction A number of computationally efficient optimization tools are being developed now a day, but the numerical tools for multi-criterion optimization problems have their own significance. The present work aims beyond the mini-max approaches use for such optimization problems. In order to ustify the related work, a machining method, known as !lectric "ischarge #achining $!"#% process has been chosen. The work contained in this minor pro ect report covers the relationships establishment between the various input variables and parameters for the !"# machine. &egression analysis techni'ues have been adopted to obtain the relationships. (nce the problem of !"# process has been modeled, it has been tried to ustify the importance of bicriterion approaches in the manufacturing environment. The results have been obtained for the given set of constraints and the individual ob ective functions and are compared to those obtained by using bicriterion approaches. The generalized genetic algorithms have been used for the nonlinear ob ective functions for a given constraints in the form of crisp value bounded domains. A brief overview of the !lectric discharge machining process is as under. 1.2 Basics of EDM The use of thermoelectric source of energy in developing the non-traditional techni'ues has greatly helped in achieving an economic machining of the extremely low machinability materials and difficult obs. The process of material removal by a controlled erosion through a series of electric sparks, commonly known as !"#, was first started in )*+, in -..&. /hen a discharge takes place between two points of the anode and cathode, the intense heat generated near the zone melts and evaporates the materials in the sparking zone. 0or improving the effectiveness, the workpiece and the tool are submerged in the "ielectric fluid. The basic !"# process has been shown in 0ig. ).). it has been observed that if both the electrodes are made of the same material, the electrode connected to the positive terminal generally erodes at faster rate. 0or this reason, the work piece is generally made the anode. A suitable gap, known as the spark gap, is maintained between the tool and the workpiece surfaces. The sparks are made to discharge at a high fre'uency with a suitable source. .ince the spark occurs at the spot

where the tool and the work piece surfaces are the closest and, since, the spot changes after each spark, the spark ravel all over the surface. This results in a uniform material removal all over the surface and finally the work face confirms to the tool surface. Thus the tool produces the re'uired impression on the workpiece. 0or maintaining the predetermined spark gap, a servo control unit is generally used. The gap is sensed through the average voltage across it and this voltage is compared with a preset value. The difference is used to control the servomotor. The efficiency and accuracy of performance have been found to improve when a forced circulation of the dielectric fluid is provided. In the present work, the tool material is copper for conducting the experiment or investigations. !lectric discharge machining is a chipless metal removal process that uses the principle of metal erosion by an interrupted electric spark discharge between the tool and the workpiece. This process is also called spark erosion process.

The fre'uency ranges from a few hundred to several thousand kilohertz with the application of a suitable voltage across the anode and cathode.Princip ! of op!ration /e know that whenever an arc is caused by an accidental short circuit. pitting erosion occurs on the surface of the shorted material.12mmdepending upon the cutting action re'uired and the current density. It involves the controlled erosion of electrically conducting materials by rapid and repetitive discharge of spark between the electrode tool and workpiece $hence the name spark erosion% the tool is usually made the cathode and the workpiece made the anode. The spark gap ranges from 1. electrons are the emitted from the cathode and cause the ionization of the fluid in the sparkgap. The workpiece and tool separated by a small gap and termed as the sparkgap.112 mm to 1. when more . this spark gap is either flooded or immersed in a dielectric fluid. the spark discharge is produced by the controlled pulsing and direct current. !"# also works on the same principle of erosion by arcing.

Thereby resulting in the metal and carried away by the flowing dielectric fluid. the resistance drops causing an electric spark to ump between the work and the tool gap. #etal removal rate $ii%. !rosion diameter $v%. The spark causes a focused stream of electrons to move with a high velocity and acceleration from the cathode toward the anode .thus creating high compression shock waves . !rosion depth The few variables 5 parameters which are useful in analyzing the !"# process accuracy and efficiency are as below7 $i%. !lectrode wear $iii%. and electrons are emitted from the cathode and the gap is ionized.such shock waves result in local rise in temperature to the order of about )1. . 8ulse interval time $iii%. /orkpiece and the tool are separated by the dielectric fluid in a container.the current density and the power density used is the order of )1. there by causing electrical discharge and machining operation.urface &oughness .11) seconds. 1.111a5cm6 and 211mw5cm6 respectively. The particles of the metal so removed are driven away by the flowing dielectric fluid .the duration of the electric pulse is about 1."# $actor aff!ctin% EDM proc!ss A number of controlling variables play an important role in the entire !lectric "ischarge machining process. The electro-magnetic field cause compressive forces to act on the cathode thus metal removal from the tool is much slower than the workpiece . "ischarge current $iv%. The dielectric breaks down when a proper "3 voltage $21-+21% 4 is applied across the anode and the cathode.electrons are collected in the gap. A few of them are7 $i%. hence the whole cycle of sparking and metal removal take place in a few microseconds. The forces of electric and magnetic fields caused by the spark produced a tensile force and tear off particles of molten and soften metal from the workpiece. 8ulse duration $ii%.111 c and cause melting of the metal.

in a typical no-wear !"# . The taper effect reduces gradually to zero after about =2mm penetration ..6micron $generally roughing operations% . :ut for high precision works with use of high fre'uency $211-)111% k.5min and with +11A current it is about +911mm .$iv%. mm can be achieved. Thus the depth of the crater on work piece defines the surface finish and it depends upon the current density. where low park gaps are used. Thus an optimum gap is necessary for higher accuracies tolerances of <-1.12mm per )1 mm depth. usually high fre'uency and low current density give better surface finish.z and low current $)-6A%. but a smaller gaps leads to a lower working voltage and hence a slow metal removal rate.5min .taper effect can be eliminated by the use of vacuum flushing of dielectric fluid.5min. the best surface finish on steel is of the order of 1. which are not desirable. !"# also produces taper. The range of overcut is 2 to )11 microns and depth on the roughening operations . it is defined as the volume of metal removed per unit time per ampere.11. The metal removal rate in roughening operations of steel with a graphite electrode 21 A current is about +11mm. with close control of process variables.112 to 1.the effect of corner radii is e'ual to the spark gap. 8ower consumption by the #achine M!ta r!&o'a rat! #etal removal rate it is direct proportional to the current density used. fre'uency and the electrode material. The taper is of the order of 1. tolerance up to <-1. overcut and corner radii.12 mm can be obtained in normal !"# operations. Its value is lower in finishing operations.+ micron $at )111khz >)A%. The smaller the gap the higher is the accuracy. metal removal rate is as low as 6mm. Accurac( The accuracy of the process mainly depends on the spark gap. The volume of crater is proportional to the energy in the spark. In precision operations. each electrical spark discharge develops a spherical crater in the workpiece. as well as in the electrode. )urfac! finis* In !"# operations.the surface finish is about . Thus the depth of the crater is proportional to the energy in the spark.

This becomes the heat affected zone. Too ! !ctrod! /!ar /hile performing the operations .cost and ease of grinding the electrode most commonly used materials may include brass copper graphite .the hard surface is a source for thermal stresses .al-alloys .&3. Craphite 6. The instant heating and vaporization of metal due to spark. In !"# operations .A? is small in finishing operations which can be removed by producing after !"# the tool material should be a material should which is difficult to machine such as a graphite which goes to vaporization without melting. /ear ratioAmaterial removal from work5material removal from the toolB Typical electrode materials are. . . 3opper. leaves behind a small amount of molten metal on the machined surface which re-solidifies and due to fast cooling action of the dielectric fluid forms a hard surface.plastic deformation and fine cracks at the grain boundaries . ).the depth of .H!at !ff!ct!d +on! etc.A? is about 6 to )1micron )1 micron deep on the work surface and its hardness is about @1. flow hole should be made while making for the circulation of dielectric flow so as to attain large flow rates at low pressure. Tungsten !lectrode material is selected on the basis of wear ratio.the .the tool gets eroded due to sparking action .HA-. metal removal rate .

RE0UIREMENT) O$ A DIE1ECTRIC $1UID The essential re'uirements in the selection of a dielectric fluid are as follows. . Do rust problem and no dangerous gasses are produced with kerosene. so that the tool work piece container etc is not is flushed through the spark gap or supplied to the gap through a hole in the tool or from external ets. To carry away the condensed metal particles and to maintain the gap for continuous and smooth operation. the electrical discharge separates the water into pure hydrogen and pure oxygen. It should have a stable and sufficient dielectric strength to act as an insulation between the tool and the workpiece. To serve as a spark conductor in the spark gap between the tool and the work material 6. In some cases the fluid is also supplied through openings in the work piece wherever feasible. To act as a coolant to 'uench the spark and to cool the tool and the workpiece . 0irst. It should have low viscosity and high wettability. A great deal of heat is generated during the !"# process and the dielectric absorbs the most of it. /ater is best. ). It should be chemically inert.econd. A good compromise then is kerosene. 6. :ut water has a few drawbacks. There are many dielectrics to choose from based on the insulation properties of the fluid.the different flow patterns are discussed in the subse'uent sections. a very explosive pair. ). cooled and re-circulated. Air is not a very good insulator.Di! !ctric f uid As mentioned the dielectric fluid acts as an insulator between the electrode and the mold cavity. it causes rust. The dielectric must be constantly filtered. Additionally the process creates a lot of debris and this need to be filtered out of the system In an !"# process a dielectric fluid in an essential working medium . .. .. the flow of this working medium operations . 0unctions of dielectric fluids The functions of a dielectric fluid in !"# are as follows.

Cenerally. some parameters or variables are generally specified such as erosion diameter. contamination by metal particles and products of decomposition. 0or example. 2.112 to 1. )! !ction of di! !ctric f uid The main consideration in the selection of a dielectric fluid is the operating conditionsE the choice of dielectric fluids depends on the size of workpiece. applied working voltage. the same machine is to be run for various investigations with in the controlled parameter variation. The commonly used dielectrics are kerosene. "ielectric fluid and the current density used. white spirit is used to machine tungsten carbide and metals for with intricate shapes and re'uiring high surface finish.+. It should not emit toxic vapours and should not have unpleasant odours. . It should have a high flash point. once the operating conditions and workpiece-tasking is defined. their machining rate and wear ratios. surface finish re'uired and material removal rate. It should not alter its basic properties under operating conditions of temperature variations. It should be economical for use. erosion depth.ome of the dielectric fluids a. It should be able to deionise immediately after the spark discharge. The choice of a particular dielectric fluid depends on the tolerance re'uired. paraffin transformer oil or mixture of various oils. low viscosity fluids are used for very high surface finishesE the used dielectric fluid is recirculated and after proper filtering action remove the metal particles. :efore solving the optimization problem for this process. . The parameter variation ranges are specified in upcoming chapter on problem formulation and solutions. dielectric pressure etc.12mm. 0iltering medium like wound cotton yarn cartridge diatomaceous earth filter are used to remove wastes from the medium for normal precision works filters better than 6 microns sizes are used. work material.this gap depends on the type of tool. complexity of the shape. )par2 %ap The spark gap between the tool and the work is in the range of 1. size and shape of the work material removal rate and the type of electrode.

automobile. extruding. As well as dimensional factors of size and shape. since the material of the electrode has to be specially matched.Ad'anta%!s !lectrical discharge machining has advantages over other machining techni'ues due to its ability to create complex and intricate parts with a high degree of accuracy. stamping.ome of the common applications for !"# include producing plastic molds. . an important consideration before undergoing this procedure is the material of the workpiece. The workpiece is burr-free after completion and saved from heat damage because very little is actually generated during the procedure that would harm the material. the workpiece is not deformed from impact because there is no direct contact between the electrode and the material. die casting dies from hardened steel and forging dies. . defense. /hile using this process. aerospace and medical. Another advantage of !"# is its ability to machine parts on an extremely small scale. (ther purposes include the manufacturing of engine parts like compressor blades of titanium alloys and nickel based super alloys. Industries that benefit from the use of the electrical discharge machining process include food and beverage. This process is able to machine hard materials that other machining processes would have difficulties with. electronics.

. #any operating variables are considered as parameters with fixed working values over a given erosion depth and erosion rate on the work piece.CHAPTER 2 1ITERATURE RE3IE4 !fforts are being made now a day to solve complex engineering problems through mathematical algorithms with high computations accuracy. I have tried to use the concept of genetic algorithms for solving the complex formulated problems of the !lectro "ischarge #achining process. Fain 4. G6H has formulated the generalized !lectro "ischarge #achining method with the limited constraints related to 8ulse interval time and pulse duration. A number of research papers exist in the literature related to these newly developed optimization techni'ues.H states that operating working voltage and the pulse interval plays an important role in obtaining the re'uired surface finish. 0luttering of the edges in the !"# region has been investigated for the variation in the controlling parameters. The spark gap control has also been explained to obtain the desired level of surface roughness for a given set of operating variables. Iahng. 4 G)H has tried to model and formulate the !lectric "ischarge #achining process with the optimization of ob ective functions related to moving tra ectories of machine tool electrode. The formulated problem has been analysed by using a simple optimization algorithm by keeping other ob ective functions unaffected and the results are concluded to give the suitable operating variable value selection on the basis of output obtained. The flow movement of the dielectric fluid controls the homogeneous surface characteristics in the entire !"# controlled region. G. A moving frame reference has also been used to locate the tool electrode at any instant along its traversed tra ectories. I. Cradient based methods have been used to optimize the single ob ective function variable. The stress has been given on understanding the problem from the core and some of the research papers used for formulation and understanding the problem are outlined as below7 Foopelli. 3. .

4.. This paper is mainly meant for skilled workmanship towards achieving the desired surface characteristics in minimum time and with safety. it has been claimed that consistency and repeatability of the machine towards maintaining the minimum deviation in the operating conditions helps a lot in the #achining accuracy in the process. . 8. 8andit. #etal removal rate and dielectric material effect on the !"# process.teel alloys. IG @ H has developed the governing e'uations for the analysis of !lectro "ischarge machining process under a controlled environment. The results obtained by the formulation used with the help of 'uadratic elements have shown a good convergence with those obtained by the commercial packages.3. .. Jim.. The relationships obtained have been used to plot graphs for the variation in the operating controlling parameters and their effects on the conse'uents such as metal removal rate. Ju G2H has specified the basic thumb rules for the analysis of surface features of !lectro discharge machining process.J. 8andit.owever. "ischarge 3urrent. This paper is recommended for peer mainly. These graphs can be directly used for the selection of given constrained condition of operating variables for the desired ob ectives. The saving of the production cost is ustified for the !"# process carried out. . A suitable hard alloy material is selected as electrode tool material. .. surface roughness and the power consumption by the machine etc. .. 3ommon measures and precautions which are helpful in carrying out the !"# process for efficient operation are also been suggested.. #adhu. # G9H ahs stated the critical factors affecting the performance of the !lectro "ischarge machining process when the workpiece material is 3emented carbide. And the dielectric fluid is given turbulent flow in and around the !"# region. The operating variables like 8ulse duration. Fain. The energy parameters and the #etal removal rate relationships have been developed for the given set of operating voltages and the dielectric pressure. A computer program has also been developed in the form of subroutines for the calculation of electrode wear rate. (perating voltage. 8ulse Interval time and heat dissipation rate differ in operating ranges considerably as compared to electro discharge machining of . # G*H has considered the theoretical aspects of the !lectro "ischarge machining 8rocess.

Cuha.pedding. H has stated a practical approach for #achining the :eryllium 3opper alloys as workpiece by !lectro "ischarge #achining process. ..&a urkar. ?. has developed a network based flow diagram for the !lectric "ischarge machining processes. the tra ectory of the electrode movement plays a vital role in improving the surface characteristics of the !3# process.pedding. T. .A G))H have used the concept of conformal transformation of the operating characteristic variables.G). The surface characteristics of the wire3ut !"# process have been analysed for the sensitivity of the operating variables. and ?hu. "G)1H have elaborated a number of alternatives for the improvement in the electrochemical machining process by changing the sensitivity parameters and have stated that in comparison to other operating variables. . and /ang. T. The interpretation of output variables variations has been carried out for the wire 3ut !"# process and the suggested ranges of the input variables are given for a desired set of output variables in terms of metal removal rate and power consumption etc. . I. #ethods have been suggested for obtaining the desired level of surface characteristics by using this !"# method. The flow diagram had been used to optimize the !"# process by selecting the independent paths for a given set of parameter values. The dependencies have also been evaluated at a particular instant by assuming some of the operating conditions fixed.8 . A.. surface roughness has been mapped onto parametric surface.. The variables are parameterized and the parametric representation of the metal removal rate. 8.K G)6H has considered the theoretical aspects of the modelling of the wire3ut !"# process. The intermediate processes are shown in the diagram along with their precedence events and successors. . &a urkar. A brief note is also given for specifying the operating characteristics and safety precautions for the !"# process to be carried out. A user friendly approach has been adopted for the definition of process parameters and these parameters are compared to other various ranges of operating variables. .A. I.afety precautions and the indicative measures are suggested for the fruitful implementation of the process.myers. The dependencies of the decision variable on each other are represented and a computation algorithm has been proposed to evaluate the mapped point for specified surface characteristics onto parametric plane.

everal (b ective 0unctions.G)+H states that spark gap and the controlling parameters for a sensitive !"# process layout can be controlled in number of ways. ?hang. . circuitry. the surface roughness and the power consumption by the machine.. encoders etc. The effect of the selection of dielectric fluid has also been analysed for a given set of electrode tool and material combination. The results are tabulated and graphs have been recommended for use for the machining of steel materials.the results have recommended for the machining parameters. Indurkhya. F G)=H. has developed 0uzzy 8rogramming and Jinear 8rogramming with . The author also states that transducers./ang /. The performance Index evaluated by using this method helps in analysing the adverse and positive gradient effects of the variation in the #etal removal rate. : G)2H the report submitted by the author and his team members have calculated the effect of motion and turbulence level in the dielectric material during various stages of the electo discharge machining process. A feedback system with real time stability analysis and process monitoring through digital modern sensors and transducers can give an efficient responding mechanism for the !"# process control. ?immermann . ?hang 3 G)@H and his team members have developed tolerances for the different machining parameters. The performance index evaluation for the !"# process for a given specified crisp sets of the operating variables helps in understanding the efficiency of the process to be carried out. C has developed an artificial neural network for the entire !lectro "ischarge #achining process. The relationships between the operating intermediate processes along with decision variables have been framed.#. can be selected to give influence of simultaneous variation of operating variables and their response data storage facility.

The procedure is detailed as follows7 !ach ob ective function can be represented as an function in terms of minimization or of maximization type as under.tep )7 0ormulate the bi-criterion model .tep 67 . Archimedean and non-Archimedean goal programming methods suffer from the difficulties in determining the weights for deviation variables within a lower priorityE and the ranking of the goals in a preemptive preference order.olve the problem with one ob ective at a time . Archimedean goal programming. we obtain sumLcost) MA valueLcalculated and sumLtol) .CHAPTER " METHODO156 ADOPTED .everal multiob ective optimization methods are available in the literature including the weighted multiob ective method. The weighted multiob ective method involves sub ectivity and bias in specifying the weights in order to aggregate several non commensurable and conflicting ob ectives into a single e'uivalent function. such as the intangible cost mentioned above. and the original 3hebyshev goal programming methods. It should also be noted that the #3C8 is in fact a fuzzy programming approach and hence can be used to effectively deal with ob ectives that are imprecise or fuzzy in nature. The #3C8 can be implemented using the Jingo software.olving #in sumLcost with all the constraints. . the final solution could be dictated by a single goal. /ith the original 3hebyshev approach. non-Archimedean goal programming. The modified 3hebyshev goal programming #3C8 approach helps to avoid the above difficulties and will be used in this study.

olving this problem.7 0ormulate and solve the modified 3hebyshev goal programming 8roblem The notion is that the best deviation from the two worst values of two ob ectives is obtained by maximizing L. . #inimizing sumLcost leads to a lower sumLtol and maximizing sumLtol causes an increase in sumLcost. sum-costM A )9. the two ob ectives conflict with each other. . and sum-tolM A 1. .MAvalueLsumLcalculated . .96@. (ther outputs are omitted due to space limit.*2).olving #ax sumLtol with all the constraints. /e obtain sumLcost6 MAvalueLsumLcalculatedM and sumLtol6 MA value calculatedM the ob ective values of the above solutions are summarized below7 As we can see. Deither situation is desirable particularly when the differences are as substantial as shown in the above table.tep . we obtain A 1. This calls for a compromised solution by solving the modified 3hebyshev goal programming problem in the next step.+*@.

This model describes the functional dependence between input variables and out variables. &ecently most of the research work is devoted to the methods for finding a statistically. 0or the !"# process formulation same techni'ues have been applied and the relationships have been developed as specified ob ective functions in the problem. The data thus obtained are analyzed by means of regression methods and then the mathematical model of the system is then obtained. These methods are based on the experiments and their aim is to determine an investigation program which s compromise between a re'uired number of investigations in real life condition and their informativeness. #etal removal rate ii%. /hereas the main output 'uantities are i%.CHAPTER 7 $ORMU1ATION O$ THE EDM PROCE)) PROB1EM AND IT) )O1UTION In many engineering activities. The forms of approximation functions can be considered different. "imensional shape accuracy of the work piece A theoretically recommended approach to the problem of finding a mathematical description of the !"# process would be to carry out investigations in the whole region . !rosion depth etc. 8ower consumption iv%.experimental model of such systems. !lectrode wear iii%. . 8ulse "uration b%. In !lectric "ischarge machining process. 8ulse Interval c%. it is very difficult to obtain an explicitly or implicitly formalized description of the system which could them be optimized. Amplitude of "ischarge current d%. main input 'uantities are7 a%. !rosion surface e%.urface roughness v%.

minimizing tool electrode wear rate. The space of variables can be restricted to the region which is physically sensible and sets of e'uations are obtained. so these two 'uantities may be considered as constraints or they may be omitted.1+=lnTo<6.6lnTi<1. The output 'uantities usually chosen as ob ective functions are7 maximizing metal removal rate.61=lng N)6.)22lng N-).=++I)).1@@lnTi-1..=6@-1. The surface roughness and the dimension and shape accuracy of the workpiece can be taken as the third and fourth ob ective functions in case of the accurate machining.lngg1.1.<1.. The (b ective 0unctions of the problem formulated after regression analysis of the investigations are as below KA e)).=@lnN<1.1+2lnN<1.+-1.)*=lnN.6.112 The pulse duration in .62.lngTo)..of the space of input variables. the decision variables are those input 'uantities whose values are set on the machine i.166lnTo<1.)16 N1. In the problem formulation a particular case of !"# process has been analyzed.)=)lngg-. The choice of the ob ective and constraint functions depends on the user re'uirement. These sets of e'uations give a mathematical description of the !"# process which is the basis for building the optimization model..1=))nTo-1.@. 8ulse duration..21*I2.1@=-1.ub ected to 3onstraints Ti P 211 x )1-@ .))*lnTo<1..9+21.@1*6.222<1. 8ulse interval time and discharge current.129lng .))*lnN<1. If the erosion surface does not change in the machining process and depth of erosion is also known in advance.)1=1.imilarly.1612lnN<1.)6+lngg-1.1..@@.12)nTo-1. these two input 'uantities are treated as parameters.1)@lngTo1.=++<1.11.2lnTo<1.6)*-1.1.)+lnN<1.)=+lngTi.1+9lnN<1.12..12)lngTo-1.++lnN<1.22=-1.16@lngTi-).I). In the present model. In this process rough machining has been considered.econds does not exceed DA e-1.)=+lnN<1. It is assumed to have constant dielectric pressure and average working voltage while taking the experimental readings which are to be used further for regression analysis.+*lnTi-1. .lngTi1.=+6 O A e-9). It is granted that cylindrical copper electrode is to be used as a tool and tool steel as the machining material. the power consumption by the machine may be either a new ob ective function or the constraint or it may be discarded.e.+)-1.

9 h8a These minimum and #aximum Jimits The pulse Interval time in . . 8A @1. in each case there is no guarantee to get optimized results of the parameters when other ob ective functions are not taken into account. 0or a given set of constraint each ob ective function can be evaluated by using any nonlinear programming techni'ue or by using Cenetic algorithms.econds does not exceed These minimum and #aximum Jimits The "ischarge 3urrent in Amperes does not exceed These minimum and #aximum Jimits The !rosion "iameter in mm does not exceed These minimum and #aximum Jimits The !rosion "epth in mm does not exceed These minimum and #aximum Jimits The applied working 4oltage is kept almost constant The "ielectric 8ressure is kept almost constant and e'ual to this value for a case when cylindrical 3opper electrode is the tool material and Tool steel is the machined material This problem is a case of #ulti. The integrated problem is formulated as a bicriterion model to handle both tangible and intangible costs. .owever.Ti Q 6111 x )1-@ To Q 621 x )1-@ To P )62 x )1-@ I Q )69 I P @+ N Q =1 NP 21 g Q )1 gP2 4A . /hile solving such cases.2 R .3riteria (ptimization type with different #ini-#ax constraints and ob ective 0unctions 3onditions. The model is solved using a modified 3hebyshev goal programming method to achieve a preferred compromise between the two conflicting and noncommensurable criteria. the approximate solution is achieved but is considerably away from the exact solutions. A number of some other approaches exist to solve a #ulti-ob ective function problem but all of them are either derivative function based or assumes suitable penalty functions for the account of other ob ective function.

.+)-1.++lnN<1.1)@lng To1.6lnTi<1.)22lngN-).16@lngTi-).<1.1612lnN<1.12. O.9+2-1.222<1.)=+lngTi.62.1+=lnTo<6.. such combination of #ini-#ax type ob ective functions along with constraints is considered and the optimal values of the parameters and ob ective functions are evaluated at particular instances.=@lnN<1.=+6 #inimize O A e-9).61=lngN)6.@@.6lnTi<1.1.=+6 #inimize D A e-1.112 . D.12)lng To-1.)22lngN-).1.=++ I)).16@lngTi-1..6.)=)lngg-..))*lnN<1..lng To).1+2lnN<1.1@=-1. power consumption in /atts 7 To be maximized 7 To be #inimized 7 To be #inimized /hen multi.1@@lnTi-1.1612lnN<1.12)nTo-1.)*=lnN-1.166lnTo<1.1+=lnTo<6.=++<1. the nature of the ob ective functions are conflicting type and the combination of ob ective 0unction for #etal removal rate with other two ob ective functions in terms of electrode wear and power consumption makes ob ective function mini-max type. According to the :i-3riteria #odel formulation to the above problem.)=+lnN<1. )t!p 17 0ormulation as per bicriterion model Cas! I7 considering ob ective functions $i% and $i% #aximize K A e)). the ob ective functions are of following types7 $i%.)=+lnN<1.lngg1.)1=-1.)6+lngg-1.21*I2.ub ected to 3onstraints Ti P 211 x )1-@ To Q 621 x )1-@ I Q )69 N Q =1 g Q )1 Ti Q 6111 x )1-@ To P )62 x )1-@ I P @+ NP 21 gP2 Cas! II# considering ob ective functions $i% and $iii% #aximize K A e)).=6@-1.129lngN1.+*lnTi-1. it is clear that for a given set of entry constraints variables.criteria model is developed.=++ I))..=@lnN<1.1=))nTo-1. #aterial removal rate $ii%.1.)6+lngg-1. !lectrode /ear in percentage $iii%. K.))*lnTo<1.1+9lnN<1.)16 .1@=-1..)+lnN<1.@.@1*-6.In the above problem formulation.6)*-1..11.)1=-1.+-1.lngTi1.166lnTo<1.22=-1.222<1.=++<1.ub ected to 3onstraints .12)lng To-1.I).2lnTo<1.

tep 67 . The problem can be formulated by using this bicriterion approach for each case as below7 3ase I 7 #aximize S .2+.9*=@2+2 #aximize K.91+. The values obtained for each case are tabulated as below7 3ase I 4alue L#in O 4alueLmax K 3ase II 4alueLmin D 4alueLmax K #inimize O.)96.ub ected to $valueLmaxK-valueLmaxKM% SQ ---------------------------------.Ti P 211 x )1-@ To Q 621 x )1-@ I Q )69 N Q =1 g Q )1 Ti Q 6111 x )1-@ To P )62 x )1-@ I P @+ NP 21 gP2 .2@= #aximize K.@*1992 2=9.2@+*66 "ifference +6+). This calls for a compromise solution by solving modified 3hebyshev Coal programming problem in the next step. in case II.olve the problem with one ob ective at a time The problem is solved by using generalized genetic algorithm for the given set of constraints and only one ob ective function at a time.*9@@=9* *2).+*9 +)9. mm5min 1. .=6.A $valueLmaxK-+)9.2+.6.2@=+.2@= "ifference 1. #inimize O leads to a lower value of K while maximizing K leads to an considerable increase in the value of O in case I. mm 1.*=9@+2 #inimize D.tep III 7 0ormulation and solution by modified 3hebyshev Coal programming problem.2@+*66 . *2).6*) 2.6.9=@2+2 . mm5min =112..@+2*)62 As we see the two ob ectives conflict to each other in each case. #inimize D leads to a lower value of K while maximizing K leads to an considerable increase in the value of D. Deither solution can be considered desirable particularly when the differences are substantial as shown in the table above.*=9@+2%5 2. /att 6=@.

6*) $valueLminO6M-valueLminO)M% Ti P 211 x )1-@ To Q 621 x )1-@ I Q )69 N Q =1 g Q )1 . . for case II7 #inimize T .A $1.9*=@2+2%5 2=9. /e obtain the following results .=6.@+2*)62 $valueLmaxK6M-valueLmaxK)M% $valueLminD6M-valueLminD% T Q ---------------------------------.@*1992 $valueLminD6M-valueLminD)M% Ti P 211 x )1-@ To Q 621 x )1-@ I Q )69 N Q =1 g Q )1 Ti Q 6111 x )1-@ To P )62 x )1-@ I P @+ NP 21 gP2 Ti Q 6111 x )1-@ To P )62 x )1-@ I P @+ NP 21 gP2 The notion is that the best deviation from the two worst values of two ob ectives is obtained by maximizing S and minimizing D.91+.A $valueLmaxK-.*9@@=9*.valueLminO%5 1.ub ected to $valueLmaxK-valueLmaxKM% TQ ---------------------------------..olving this problem.$valueLmaxK6M-valueLmaxK)M% $valueLminO6M-valueLminO% SQ ---------------------------------.imilarly.valueLminD%5 +6+).A $=112.2@=+.

+*= )6.2 3ase II 7 minimizing T 8revious calculated value 0inal calculated values "ifference 4alueLmin D 6=@.62 1.=6. At any stage.+* +21.)22 0rom the above result computation it is clear that the results are close enough to take a decision about the behavioural response of the ob ective functions and the constraints.9*=@2+2 . the dominance factor of the input variables and output .=.6)+.)*1+.@2+=9 2. CHAPTER 8 RE)U1T) AND CONC1U)ION The present method adopted to solve the optimization problem of !"# process is simple enough and is flexible in selection of ob ective functions and the constraints for such machining processes.*9@*+2 4alueLmax K .3ase I 7 maximizing S 8revious calculated value 0inal calculated values "ifference 4alue L#in O 1. The developed bicriterion method for the !lectric "ischarge machining process optimization is also 'uite useful for observing the sensitivity of the ob ective functions.. This techni'ue seems to be 'uite useful for solving such constraint based machining parameter optimization problems.+*9 1.@=@).*=9@+2 +6+.*.9@.9=@2+2 .1.9+.)96. Any set of the ob ective functions with respect to any set of constraint can be analyzed for their significant behavioural response.119196= 4alueLmax K +)9.

. 2.. . 4ol. This techni'ue helps in getting the reliable multiob ective decisions under constrained penalties for the constrained optimization of such processes.variables contained in the constraints and ob ective functions can be computed.=. 4ol. the absolute values of the ob ective function differ significantly for their absolute values under maxmax or min-min condition. pp. )**+. 6. it has been found that the results obtained by the bicriterion approach show their convergence towards the exact solutions obtained by optimization of ob ective functions under min-max condition..Foopelli. )-)6.-.owever.. 4. U#ulti-(b ective (ptimization of !lectro discharge #achining 8rocess. "uring the solution of the problem. )**1. Fain.U #icrotechnic ournal issue. pp. U#ulti-(b ective (ptimization of 8arameter 3ombinations in !lectrical "ischarge #achining with (rbital #otion of Tool !lectrode. I..U Fournal of 8rocessing of Advanced #aterials. RE$ERENCE) ). 4. +. .

W Int.3. ". Cuha. I. Fournal of #aterials 8rocessing Technology.*-. . ))2X)6=. /ang. 7. )**2..pedding.A. pp. 4..#.. 3. and ?hu.-26. )*9). )**= 12. Do ). .J.!#-)).urface 0eatures of !lectro-dischargeW. Ju..pedding.urface 3haracterization of the /ire !"# 8rocessU Fournal of 8recision !ngineering.. :.cheduling with 0uzzy 8rocessing Time and 0uzzy "ue date through Cenetic Algorithms. 1=. #asatoshi. U8arametric (ptimization and . <.. . Do )-.8andit. )-=.#adhu. #. /. 4olume )61.4ol.). Iubota..U Fournal of processing of Advanced #aterials. )**1 9. )*==. V"evelopment of 3onstrained (ptimization Analyses and . 8..urface 3haracteristic :ehavior "ue to &ough and 0ine 3utting by !"#.".afely and !fficiently.6. U.hop . #ach. )***. pp.. == -96..A. U.W !uropian Fournal of (peration research.. U!lectrodischarge #achining of :eryllium 3opper Alloys .X+1=.. )**). UImprovement of !lectrochemical #achining by -sing (rbital !lectrode #ovement. 11.urface &oughness.15). Tools #anuf.pp. 6. pp. 4ol. I..*..+2. 6)=-66+. A. 4ol 61. 4ol.tudy on #odelling of /ire !"# 8rocessU International Fournal of #aterials 8rocessing Technology 4ol @*. 4ol. UA #athematical #odel for !lectro-"ischarge #achined . pp..U Trans. F.Iahng. )*=9. . ...U Annuals of the 3I&8.. American . Fain. Iee. Fanuary )**=.ystems -sing #odern 3ontrol 3oncepts. ... . and 8roc. .myers. pp..8andit. Do. #.+95).trategies for #ulti8ass &ough Turning (perations.*-)+6.2). . ..ymposium on !lectro -#achining. 17.I..#!.U Annuals of the 3I&8. 4ol. pp.. of the 9th DA#&3. T.8 . .ociety of 8recision !ngineering. . 6@5). )**@ 8. and &yo. pp. UAnalysis of !lectro-"ischarge #achining of 3emented 3arbides. ?.T 6112 Transactions of the A. T.U Annuals of the 3I&8. )@)-)=.U International Fournal of !lectro machining... 2++ 5 4ol. and /ang. A-C-.&a urkar. 6+. . 8. . V0uzzy 8rogramming for #ultiob ective Fob .. U0inite !lement Analysis of !"# 8rocess. )6=. UAdvances in !"# #onitoring and 3ontrol .. pp.U 8roceedings of the International .. 2-)2.K.. )**= 1".)9-69.. pp. )))-))@. Jim.U Towards a :etter -nderstanding of the .

?hang.I..yst.18. ).ets . ?hang. U!ffect of "ielectric 0luid 3haracteristics on !"# 8erformance. 1:.U a report from C! &esearch and "evelopment. 8. VIntegrated Tolerance (ptimization with .. Chose A. 3.?immermann. .W 0uzzy . .imulated. 6112 . pp.everal (b ective 0unctions..cienceW. F. mallik A. V0uzzy 8rogramming and Jinear 8rogramming with .3... and /ang. The !nglish -niversity 8ress. .. "ecember )**=. !/8 limited. +2X22...)*=9 R!f!r!nc! >oo2s 1.. V Don Jinear (ptimizationW. Jondon 2. 2.. :. 19."ixon J. V #anufacturing .