Free Academic Seminars And Projects Reports

You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Wire Electric Discharge Machining(WEDM) is one of the most important non-traditional machining process.Taguchi method is supplemented with various supportive techniques such as fuzzy logic, grey-relational analysis,two-phase parameter design, artificial neural network(ANN) and various combination methods to incorporate material removal rate(MRR) and surface roughness(SR)simultaneously .Particle Swarm Optimization(PSO)& Memetic Algorithm(MA) based optimization procedures have been developed to optimize machining parameters viz. machining speed, pulse on time, pulse off time& peak current. Thus improvement of EDM performance is achieved not only by various monitoring and control systems but also by applying parametric optimization with various supportive techniques.
Electrical Discharge Machining (EDM) is a controlled metal removal process that is used to remove metal by means of electric spark erosion. The English scientist Priestley first reported the erosive effect of electrical discharges in 1770.In this process, an electric spark is used as the cutting tool to cut the work piece to produce finished part. The metal removal process is performed by applying an electrical discharge of pulsed, high frequency alternating current or direct current through the electrode to the workpiece.The electrode location is controlled by the machine and is positioned so as not to contact the workpiece.A precise controlled space is maintained, allowing the spark to discharge its current from the electrode to the work piece through an insulated dielectric fluid of oil or water. This removes tiny particles of metal from the work piece.
With the EDM process, both the work piece material and the electrode material must be conductors of electricity. The EDM process can be used in two different ways:
In the EDM process, an electric spark is used to cut the workpiece, which takes the shape opposite to that of the cutting tool or electrode. The electrode and work piece are submerged in dielectric fluid, which is generally light lubricating oil. This dielectric fluid should be a non conductor (or poor conductor) of electricity. A servo mechanism maintains a gap of about 0.01 to 0.02 mm between the electrode and workpiece, preventing them from contacting each other.
The wire-cut EDM is discharge machine that uses CNC movement to produce the desired contour or shape. It does not require a special shaped electrode; instead it uses continuous traveling vertical wire under tension as the electrode. The electrode or cutting wire can be made of brass, copper or any other electrically conductive materials ranging in diameter from0.04 to 0.41 mm.The paths the wire follows is computer controlled along two axes (XY) contour, cutting a narrow slot through the work piece. This controlled movement is continuous and simultaneous in increments of 0.001 mm.Any contour may be cut to high degree of accuracy and is repeatable for any number of successive parts.The dielectric fluid maintains the proper conductivity between the wire and the work piece, and assists in reducing the heat caused by the spark.
During the EDM process the work piece and electrode are submerged in the dielectric oil, which is an electrical insulator that helps to control the arc discharge. The dielectric oil that provides a means of flushing is pumped through the arc gap. This removes suspended particles of work piece material and electrode from the work cavity, insulates against premature discharging and helps to cool the electrode and work piece.
One of the most important factors in a successful EDM operation is removal of the particles (chips) from the working gap. Flushing these particles out of the gap between the work piece and the electrode are very important to prevent them from forming bridges that cause short circuits. These arcs can burn holes in the work piece and in the electrode.EDMs have a built-in power adaptive control system that increases the pulse spacing as soon as this happens and reduces or shuts off the power supply.
Rough machining gives poor surface finish due to micro cracks and pores, also finish machining gives better finish but in that case material removal rate(MRR) or machining speed is very less. Hence various monitoring and control systems were suggested such as continuous gap monitoring system, servo and pulse adaptive control system, knowledge based control system etc.It is very difficult to achieve higher cutting speed and better surface finish simultaneously. Hence it is considered as multi criteria optimization problem. Classical approach suggested by Fisher and Yates is inefficient because it considers one factor only at a time.Taguchi method also can optimize one factor either MRR or surface finish (SF) at a time. Hence it is supplemented with various supportive techniques such as fuzzy logic, grey relational analysis, two-phase parameter design, artificial neural network (ANN) and various combination methods.
Wire Electric Discharge Machining (WEDM) process is one of the important non traditional machining processes. It is used to machine hard materials, complex shapes and contours which are difficult by conventional methods. Particle swarm optimization (PSO) and Memetic algorithm (MA) based optimization procedures have been developed to optimize machining parameters viz.machining speed, pulse on time, pulse off time and peak current by using two response equations for material removal rate and surface roughness. The objective function considered for optimization is maximization of material removal rate and minimization of surface roughness. The objective function is solved by taking combined objective function (weight age given 50% to MRR and 50% to SR) i.e. minimization of MRR and SR.The output results of these two algorithms are compared.
Gap monitoring system identifies major gap states and thus differentiates between normal spark and harmful spark. The gap voltage and current signal have been modelled and analyzed mathematically by DDS (data dependent system).Radio and high frequency monitoring detects high frequency signal on the gap voltage. It can also provide pulse control to machine power generator.
Adaptive control for EDM adjusts the machine parameters such as servo settings, pulse off time, flushing rate etc as per the requirements so as to achieve optimal process performance i.e. maximum MRR and minimum tool wear ratio and desired integrity.
EDM fuzzy logic servo control system is capable of monitoring the gap states. Conventional EDM servo control systems, due to the lack of precise information of gap states (such as gap open, normal and harmful discharges etc) are unable to provide any action for avoiding the harmful acing. Servo feed and fuzzy logic strategy together encounters all measured gap parameters and thus makes the system capable to respond to all monitored gap signals in order to avoid arc damage and improve machining rate and work piece quality.
These monitoring and control system were not only complicated bit also costly and hence many times not economically feasible. Hence an experimental approach for parameter design was suggested.
Evaluation of machining performance in EDM is based on performance characteristics such as MRR, SR, electrode wear rate (EWR) and spark gap (SP) often called as uncontrollable factors. Various machining parameters such as peak voltage, pulse on time, pulse off time, peak current spark gap set voltage, wire feed rate, and wire tension over which an operator has sufficient control are referred as controllable parameters.
Fuzzy model was developed with input parameters like tool-work piece combinations, tool area, tool wear, peak current and output parameters such as off time(microseconds),spark gap(mm) and servo sensitivity(milli volt/sec).Information obtained from the experimental model was,MRR is inversely proportional to quality. Increasing current (Ip) increases MRR but increases depth of heat affected zones. For finishing operation, productivity is determined by required surface finish, also for finish machining pulse recurrence frequency can be increased but it increases total and unit energy consumption. Setting off tool wear is sufficient as it determines accuracy and economy of operation and tool consumption. Higher off time decreases machining efficiency while too short off time prevents complete de-ionization of previously formed discharge channel causing abnormal discharges, which adversely affect tool wear, accuracy and surface finish. Hence optimum off time should be maintained. For optimum efficiency spark gap should be constant.
A fuzzy logic unit comprises of a fuzzifier, membership functions, a fizzy rule base, an inference engine and defuzzifier.First the fuzzyfier uses membership functions to fuzzyfy the signal to noise ratios. Next the inference engine performs fuzzy reasoning on fuzzy rules to generate a fuzzy value. Finally the defuzzifier converts the fuzzy value into a multi-response performance index. In the experiment two inputs X1(EWR) and X2(MRR) are given and one output (MRPI) i.e.Y is worked out.
Fig. Structure of the two input one output fuzzy logic unit.
X1=S/N ratio of first quality characteristic.
X2=S/N ratio of second quality characteristic.
Y=multi response performance index (MRPI)
Then MRPI for different levels of parameter is calculated. Larger the MRPI smaller is the variance. Based on ANOVA results it has been found out that work piece polarity; discharge current and open discharge voltage are significant parameters affecting multiple performance characteristics. The levels of these parameters are optimized. Experimental result shows and confirms that EWR is decreased from 29.9% to 20.7% and MRR is increased from 0.00159 to 0.00383 gm/min.
It can also be considered as one of approaches for solving the problem of multiple responses in EDM.A higher value of grey relational grade means that the corresponding process parameter is closer to the optimal value. Thus optimization of the complicated multiple process responses can be converted into optimization of a single grey relational grade.C.L.Lin, J.L.Lin and T.C.Ko carried out experimentation on SKD 11 alloy steel (12 mm diameter) using L-9 orthogonal array to optimize MRR, EWR and SR.The mathematical treatment is given out to calculate grey relational value Xi (k) for EWR, SR and MRR.Grey relational coefficient is then worked out. Averaging all grey relational coefficients, grey relational grade (yi) is obtained.
A higher value of grey relational grade represents a stronger relational degree between the reference sequence and the given sequence. Also the higher value of the grey relational grade indicates the closeness of process parameters closer to the optimum level. Calculations using grey relational analysis are simpler, straight forward than fuzzy based Taguchi method for optimizing the EDM process with multiple process responses.
Two phase parameters designed strategy using Taguchi technique develops a robust high speed and high quality EDM process. A system with dynamic characteristics is no longer suitably designed using the conventional Taguchi approach, which is based on static characteristic. In actual practice the energy transmission of any system does not happen as designed or intended as there may be noise factors disturbing the system. The reality of the system therefore consists of non linear effects between input and output. Hence two phase parameter strategy with double signals for process optimization was proposed.
The result of the two phase dynamic experiment shows that the factor pulse on time, low voltage electric current high voltage sparking current have maximum influence on EDM process robustness. The factor pulse on time and low voltage electric current are controlling factors for EDM machining speed. The final product dimension can be further adjusted to the desired dimension using the second ideal function model. This method is simple, effective and efficient in developing a robust, high speed and high quality machining process.
ANN can also model the multi objective optimization problem. ANN is a logical structure in which multiple processing elements communicate with each other through the interconnections between the processors. A feed forward back propagation learning algorithm that uses a gradient search technique to minimize the mean square deviation between the actual output and the desired output patterns is used to solve multi criteria problem.Dr.Bhattacharya carried out an experimental investigation of two response parameter i.e. cutting speed, surface roughness on Electra supercut-734 with titanium aluminide alloy as a work piece material.
The experimental results are first used to train the neural network. For training the network in cumulative learning; the delta weights are accumulated and the weights are adjusted until a complete set of input and output pairs are presented to the network. ANN model is then tested and varied for its performance by using training data. Initially three levels of six different input parameters (Ton, Toff, SP, Wt, SV, flow rate of dielectric fluid) which are then increased up to five for generating more number of optimum points, so as to give 15625 different combinations and hence the use of ANN model is highly justified as it is not feasible to carry out 15625 experiments. ANN gives important combinations to be worked out and further optimizes the system.

Combination of EDM and ball burnishing machining (BBM) for surface improvement by modifying the micro structure of the machined surface (i.e.minimise surface roughness) eliminates micro cracks and pores. In this arrangement two ZnCr2 balls of 5 mm diameters attached with tool applies force to form a deformation layer and ultimately produces reinforced surfaces. Improved surface roughness ratio (ISRR) is then calculated as
(SRedm SRedm+bbm)*100
ISRR= __ %
Where SRedm+bbm=Surface roughness obtained by combining EDM&BBM
SRedm=Surface roughness obtained by conventional EDM (micrometer Ra)
Thus combination of EDM&BBM is a feasible process for obtaining fine finishing and surface modifications. This method is found to be effective for eliminating the micro cracks and pores caused during machining.
MRR=1.6184-0.0404{[(A-1.375)2 /0.01]-(8/12)}-0.0138(B-20)-0.0465{[(D-3.5)2/0.25]-(8/12)} -- (1)
It is evident from the above response equation for MRR that out of four input operating parameters considered only three parameters namely machining speed, pulse on time and peak current are significant in MRR.
Ra= 1.6592+0.687[A-1.375][1-4.07(D-3.5)-0.0061[C-20]+0.0374[((D-3.5)2-(8/12-- (2)
It is evident that from the above equation out of four operating parameters considered only three parameters viz. machining speed, pulse off time and peak current are significant in surface roughness.
COF= [WF1*Ra/Ra*]-[WF2*MRR/MRR*] -- (3)
Where WF1=weight age factor 1=0.5
WF2=weight age factor 2=0.5
Ra*=2, surface roughness limitation in micro meter.
Ra*=2, MRR limitation in mm3/min.
The COF considered for optimization are maximization of MRR and minimization of surface roughness.
PSO was developed by Dr.Eberhart and Dr.Kennedy in 1995.PSO is initialized with a group of random particles and then searches for optima by updating generations. In every iteration each particle is updated by following two best values. The first one is the best solution (fitness) it has achieved so far. This value is called pbest.Another best value that is tracked by the particle swarm optimizer is the best value, obtained so far by any particle in the population. This best value is a global best called gbest.When a particle takes a part of the populations its topological neighbours,the best value is a local best and is called lbest.After finding the two best values, the particle updates its velocity and position with following equation (a)&(b).
V[ ]=C1*rand( )*(pbest[ ]-present[ ])+C2*rand( )*(gbest[ ]*(gbest[ ]-present[ ]) -- (a)
Present [ ] =persent [ ] +V [ ] -- (b)
V [ ] is the particle velocity.
Persent [ ] is the current particle (solution).
Rand ( ) is a random number between 0&1.
C1, C2 are velocity factors.
Velocity values for C1=2.25, C2=3.25
The four WEDM parameters such as machining speed, pulse on time, pulse off time and peak current are considered as particle. The four particles are initialized using the following formulae.
Machining speed (A) =A min-{(A max-A min)*ran (0-1)}
Pulse on time (B) =B min + {(B max B min)*ran (0-1)}
Pulse off time© =C min+ {(C max-C min)*ran (0-1)}
Peak current (D) =D min+ {(D max-D min)*ran (0-1)}
After initializing each particle in the first iteration, MRR, SR is calculated. Then pbest &gbest are chosen in the first iteration. In the second iteration each particle of the first iteration are updated using velocity formula (a) & (b).Then pbest & gbest are chosen in the second iteration. This process is continued in all iteration.
The combination of local search operators with a global search technique has provided very good results in certain optimization problems. The resulting algorithm from such an approach is termed as memetic algorithm. Particle swarm optimizer (global search) &simulated annealing (local search) are combined. The memetic approach takes the concept of evolution. It combined with an element of local search. PSO employs the basic operational steps of population initialization, updating the particles position by acceleration. An additional component of the algorithm is the notion that each individual can be readily improved upon.
It is a point by point method. The algorithm begins with an initial point and a high temperature T.A second point is created at random in the vicinity of the initial point and the difference in the function values (E) at these two points is calculated. If the second point has small function value, otherwise point is accepted with the probability exp {-E/T}.This completes one iteration of the simulated annealing procedure. In the next generation another point is created at random in the neighborhood of the current point & the metropolis algorithm is used to accept or reject the point. The algorithm is terminated when a sufficiently small temperature is obtained or a small enough change in function values is found.
The four WEDM parameters are initiated the initial population as explained in PSO.MRR, Ra&COF are calculated using the formulae (1), (2) & (3) respectively and algorithm explained in simulated annealing. Then pbest&gbest values were chosen using PSO algorithm. In the second iteration each particle of first iteration is updated using velocity formulae (a) & (b).Then MRR, Ra, COF are calculated using SA.Then pbest, gbest values are chosen. Thus the entire iteration is continued with the same procedure.
PSO technique produces maximum combined objective function value at initial iteration and minimum COF value in the subsequent iterations. MA produces varying COF value in the initial stages &middle stages. At the end of the iterations MA produces better results i.e. minimum COF values.
Taguchi method which is supplemented with various supportive techniques minimizes the complexities involved in setting the process parameters so as to satisfy multi objective optimization for maximum MRR&minimum SR and power consumption simultaneously. PSO and MA based procedures used to optimize WEDM parameters viz. machining speed, pulse on time, pulse off time and peak current by taking COF.From the test analysis, it is evident that PSO technique yields better results than MA.This optimization process is easy to use and very simple to implement and efficient in handling COF.
CMTI, A NICMAP Publication, Volume 3, July 2004. (pp 11-21).
Books: 1. R.K.Jain, Production Technology, Khanna Publishers.
2. HMT Production Technology, Tata McGraw-Hill.
3. Steve Krar& Arthur Gill, Exploring Advanced Manufacturing
Technology, Industrial Press Inc.