 Open Access
 Total Downloads : 1306
 Authors : Mahesh Mallampati, Dr.V.Chittaranjan Das
 Paper ID : IJERTV1IS7301
 Volume & Issue : Volume 01, Issue 07 (September 2012)
 Published (First Online): 25092012
 ISSN (Online) : 22780181
 Publisher Name : IJERT
 License: This work is licensed under a Creative Commons Attribution 4.0 International License
Optimization of Cutting Parameters As Speed, Feed & Depth of Cut Based On Surface Roughness In Turning Process Using Genetic Algorithm (Ga) And Particle Swarm Optimization (Pso)
Optimization Of Cutting Parameters As Speed, Feed & Depth Of Cut Based On Surface Roughness In Turning Process Using Genetic Algorithm (Ga) And Particle Swarm Optimization (Pso)
MAHESH MALLAMPATI Asst Professor in Mechanical Department EVM College of Engineering & Technology, Nasararaopet
Dr.V.CHITTARANJAN DAS Professor inMechanical Department RVR&JC College of Engineering, Guntur
Abstract
Abstract: Problem statement: In machining operation, the quality of surface finish is an important requirement for many turned work pieces. Thus, the choice of optimized cutting parameters like speed, feed and depth of cut is very important for controlling the required surface quality. Approach: The focus of present experimental study is to optimize the cutting parameters using one performance measures, surface roughness. Optimal cutting parameters for each performance measure were obtained employing Taguchi techniques. The orthogonal array, signal to noise ratio and analysis of variance were employed to study the performance characteristics in turning operation. Results: The experimental results showed that the work piece surface roughness can be sensed and used effectively as an indicator to control the cutting performance and improves the optimization process. Conclusion: To find the optimum machining parameters in order to get the minimum surface roughness. And then using Particle Swarm Optimization (PSO) algorithm we can obtain the optimization value by using Cprogram and similarly optimized using genetic algorithm by MAT Lab and finally Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) are used to compare the results.
Keywords: Machining Operation (turning); Surface Roughness; Lathes Machines and Mathematical Model

Surface roughness has received serious attention for many years. It has formulated an important design feature in many situations such as parts subject to fatigue loads, precision fits, fastener holes, and aesthetic requirements. In addition to tolerances, surface roughness imposes one of the most critical constraints for the selection of machines and cutting parameters in process planning. A considerable number of studies have investigated the general effects of the speed, feed, and depth of cut on the surface roughness.
To improve the efficiency of these turning processes, it is necessary to have a complete process understanding and model. To this end, a great deal of research has been performed in order to quantify the effect of various hard turning process parameters to surface quality. These factors can be divided into a) setup variables, b) tool variables, and c) work
piece variables. In order to gain a greater understanding of the turning process it is necessary to understand the impact of the each of the variables, but also the interactions between them. It is impossible to find all the variables that impact surface roughness in turning operations. In addition, it is costly and time consuming to discern the effect of the every variable on the output. In order to simplify the problem, one needs to eliminate or select specific variables that correspond to practical applications.
.

LITERATURE SURVEY
Parametric Analysis and Optimization of Cutting Parameters for Turning Operations based on Taguchi Method by Dr. S.S.Mahapatra Amar Patnaik Prabina Ku. Patnaik (1) in this paper they have conducted experiment work and done on Genetic Algorithm to optimization the experimental values.
Online optimization of the turning using an inverse process neuro controller, Transactions of ASME, Journal of Manufacturing Science and Engineering by R. Azouzi, M. Guillot, (2) Process modeling and optimization are the two important issues in manufacturing products. The manufacturing processes are characterized by a multiplicity of dynamically interacting process variables
Surface roughness prediction models for fine turning; International Journal of Production Research by A. Mital, M. Mehta (3) a greater attention is given to accuracy and surface roughness of product by the industry these days. Surface finish has been one of the most important considerations in determining the machinability of materials. Surface roughness and dimensional accuracy are the important factors required to predict machining performances of any machining operations.
Present situation and future trends in modeling of machining operations. Progress Report of the CIRP working group on Modeling of machining operations by C.A. Van Luttervelt, T.H.C. Childs, I.S. Jawahir, F. Klocke, P.K.Venuvinod.(4) The predictive modeling of machining operations requires detailed prediction of the boundary conditions for stable machining. The number of surface roughness prediction models available in literature is very limited. Most surface roughness prediction models are empirical and are generally based on experiments in the laboratory. In addition it is very difficult in practice, to keep all factors under control as required to obtain reproducible results. Generally these models have a complex relationship between surface roughness and operational parameters, work materials and chipbreaker types.
Multi machining outputmulti independent variable turning research by response surface methodology, International Journal of Production Research by K.Taraman(5) used Response Surface Methodology (RSM) for predicting surface roughness of different materials. A family of mathematical models for tool life, surface roughness and cutting forces were developed in terms of cutting speed, feed, and depth of cut.
Surface roughness model for turning, Tribology International by R.A. Lindberg and M.Hasegawa (6) conducted 33 factorial designs to conduct experiments for the surface roughness prediction model. They found that the surface rough increased with an increase in cutting speed.
Comparison of a full factorial experiment to fractional and taguchi designs in a lathe dry turning operation by Youssef a. Youssef, yves beauchamp and marc Thomas(7) F~cole de technologies superieure 4750, ave henrijulien, montr6al, Canada This paper presents a comparison of three different Experimental designs aimed at studying the effects of cutting Parameters variations on surface finish. The results revealed that the effects obtained by analyzing both fractional and Taguchi designs (16 trials each) were comparable to the main Effects and towlevel interactions obtained by the full factorial Design (288 trials). Thus, we conclude that screening designs appear to be reliable and more economical since they permit to Reduce by a factor 18 the amount of time and effort required to conduct the experimental design without losing valuable Information.
Investigation Of Cutting Parameter Effects On Surface Roughness In Lathe* Boring Operation By Use Of A Full Factorial Design Yves Beauchamp,ext (8) The main objective of this study is to investigate cutting parameter effects of surface roughness in a lathe dry boring operation. A full factorial design was used to evaluate the effect of six (6) independent variables (cutting speed, feed rate, depth of cut, tool nose radius, tool length and type of boring bar) and their corresponding twolevel interactions. In this experiment, the
dependant variable was the resulting first cut surface roughness (Ra).
Determinations of optimal cutting conditions using design of experiments and optimization Techniques M. S. CHUAT (9)In process planning or NC part programming, optimal cutting conditions are to be determined using reliable mathematical models representing the machining conditions of a particular worktool combination. The development of such mathematical models requires detailed planning and proper analysis of experiments. In this paper, the mathematical models for TiNcoated carbide tools and Rochling T4 medium carbon steel were developed based on the design and analysis of machining experiments. The models developed were then used in the formulation of objective and constraint functions for the optimization of a multipass turning operation with such work tool combinations

PROBLEM DESCRIPTION
To find the optimum machining parameters in order to get the minimum surface roughness. Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) are used to do this and the results are compared.
We have taken 14 samples of turning operation in finishing cut the values of the speed, feed and depth of cut and their respective surface roughness. The value obtained in this by varying three parameter are taken in design of expect V8 software to obtain an equation. In the response surface methodology the linear and second order polynomials were fitted to the experimental data for obtaining regression equations.
And then using Particle Swarm Optimization (PSO) algorithm we can obtain the optimization value by using Cprogram and similarly optimized using genetic algorithm by MAT Lab
In this paper the optimal machining parameters for continuous profile machining are determined with respect to the minimum production time, subject to a set of practical constraints, cutting force, power and dimensional accuracy and surface finish. Due to
complexity of this machining optimization problem, a genetic algorithm (GA) and Particle Swarm Optimization (PSO) are applied to resolve the problem and the results obtained from GA and PSO are compared.
3.1Objective Function:
The full development of machining process planning is based on optimization of the economic criteria subject to technical and managerial constraints. The economic criteria are the objectives of machining operations in terms of quality.
The objectives considered in this paper are surface roughness to be minimized

EXPERIMENTAL PART
The present study has been done through the following plan of experiment.

Checking and preparing the Centre Lathe ready for performing the machining operation.

Cutting S45C bars by power saw and performing initial turning operation in Lathe to get desired dimension (of diameter 59 mm and length 100mm) of the work pieces.

Performing straight turning operation on specimens in various cutting environments involving various combinations of process control parameters like: spindle speed, feed and depth of cut.

Measuring surface roughness and surface profile with the help of a portable stylustype profilometer, Talysurf (Taylor Hobson, Surtronic 3+, UK)
EXPERIMENTAL DETAILS:
Turning is one of the most common of metal cutting operations. In turning, a work piece is rotated about its axis as singlepoint cutting tools are fed into it, shearing away unwanted material and creating the desired part. Turning can occur on both external and internal surfaces to produce an axially symmetrical contoured part.
Parts ranging from pocket watch components to large diameter marine propeller shafts can be turned on a lathe. The capacity of a lathe is expressed in two dimensions. The maximum part diameter, or "swing," and the maximum part length, or "distance between centers."
The generalpurpose engine lathe is the most basic turning machine tool. As with all lathes, the two basic requirements for turning are a means of holding the work while it rotates and a means of holding cutting tools and moving them to the work.
The work may be held on one or by both its ends. Holding the work by one end involves gripping the work in one of several types of chucks or collets. Chucks are mounted on the spindle nose of the lathe, while collets usually seat in the spindle. The spindle is mounted in the lathe's "headstock," which contains the motor and gear train that makes rotation possible. Directly across from the headstock on the lathe is the "tailstock." The tailstock can hold the work by either alive or dead center. Work that is held at both ends is said to be "between centers." Additionally, longer work pieces may have a "steady rest" mounted between the headstock and tailstock to support the work. Typically work pieces are cylindrical, but square and odd shaped stock can also be turned using special chucks or fixtures.
Lathe cutting tools brought to the work may move in one or more directions. Tool movement on the engine lathe is accomplished using a combination of the lathe's "carriage", "cross slide", and "compound rest".
The carriage travels along the machines bed ways, parallel to the work piece axis. This axis is known as the "Z" axis.
Motion perpendicular to the work is called the "X" axis. On an engine lathe this motion is provided by the cross slide mounted on the carriage.
Atop the cross slide is the "compound rest," which can be rotated to any angle and secured. The compound rest also holds the "tool post," where tools are mounted. Tools may also be mounted in the tailstock for endworking operations.
CUTTING TOOL:
Tungsten carbide with the grade of P10
Tungsten carbide also called cemented carbide, hard metal. There are 2 compounds of tungsten and carbon, WC and tungsten semi carbide.
WORKPIECE MATERIAL
S45C OR Equivalent Alloy (1045)
COMPOSITION:
METALS 
COMPOSITION 

MIN 
MAX 

Carbon 
0.42 
0.48 
Silicon 
0.15 
0.35 
Manganese 
0.60 
0.90 
Phosphorus 
max0.03 

Sulphur 
max0.035 

IRON 
REMAINING 
Typical Applications:
Axles, bolts, connecting rods, studs, rams, pins, rolls, spindles, ratchets, crankshafts, torsion bars, sockets, worms, light gears, guide rods etc.
The working ranges of the parameters for subsequent design of experiment, based on Taguchis L27 Orthogonal Array (OA) design have been selected. In the present experimental study, spindle speed, feed rate and depth of cut have been considered as process variables. The process variables with their units (and notations) are listed in Table 4.1
Table 4.1: Process variables and their limits
Variables 
Values of different levels 

Designation 
Description 
Low ( 1 ) 
Medium (0) 
High ( +1 ) 
D 
Depth of cut (mm) 
0.6 
1.00 
1.60 
F 
Feed rate (mm/rev) 
.08 
0.2 
0.32 
V 
Cutting speed (m/min ) 
135 
210 
285 
Measuring Surface Roughness:
Roughness measurement has been done using a portable stylustype profilometer,
Talysurf (Taylor Hobson, Surtronic 3+, UK).
Experiments have been carried out using Taguchis L27 Orthogonal Array (OA) experimental design which consists of 27 combinations of spindle speed, longitudinal feed rate and depth of cu. According to the design catalogue prepared by Taguchi, L 27 Orthogonal Array design of experiment has been found suitable in the present work. It considers three process parameters (without interaction) to be varied in three discrete levels. The experimental design has been shown in Table 4 (all factors are in coded form). The coded number for variables used in Table 4.3 and 4.4 are obtained from the following transformation equations:
By obtain Taguchis L27 Orthogonal Array the experiment have be conducted and the value of the particular feed, speed and depth of cut are given below
TABLE 4.4 EXPERIMENTAL RESULTS
Std 
Cutting speed 
Feed rate 
Depth of cut 
Surface Roughness 
m/min 
mm/rev 
mm 
Âµm 

1 
135 
.08 
0.6 
2.086 
2 
135 
.08 
1 
2.338 
3 
135 
.08 
1.6 
2.522 
4 
135 
.2 
0.6 
4.326 
5 
135 
.2 
1 
4.714 
6 
135 
.2 
1.6 
5.044 
7 
135 
.32 
0.6 
6.887 
8 
135 
.32 
1 
7.2362 
9 
135 
.32 
1.6 
7.788 
10 
210 
.08 
0.6 
3.414 
11 
210 
.08 
1 
3.618 
12 
210 
.08 
1.6 
3.773 
13 
210 
.2 
0.6 
5.966 
14 
210 
.2 
1 
6.1983 
15 
210 
.2 
1.6 
6.363 
16 
210 
.32 
0.6 
8.041 
17 
210 
.32 
1 
8.197 
18 
210 
.32 
1.6 
8.303 
19 
285 
.08 
0.6 
4.391 
20 
285 
.08 
1 
4.521 
21 
285 
.08 
1.6 
4.608 
22 
285 
.2 
0.6 
6.868 
23 
285 
.2 
1 
6.994 
24 
285 
.2 
1.6 
7.071 
25 
285 
.32 
0.6 
8.536 
26 
285 
.32 
1 
8.304 
27 
285 
.32 
1.6 
8.653 
EXPERIMENTAL RESULTS AND ANALYSIS
The experimental results are presented in Table given below For the purpose of developing the mathematical model; both the data for the machining responses and factors were logarithmically transformed. Using these sets of data, the parameters for the mathematical models were determined using the multiple regression method and the significance of the models and the parameters were then analyses using analysis of variance. In this work, a commercially available statistical software package DOE was used for the computation of regression and statistical analysis of the constants and parameters. The procedure PROC REG from this package was used to compute values of the mathematical models and to carry out the analysis of variance for the models developed. In the following sections, the significance of each model developed will be discussed.
The experimental value were obtain form the experiment is given the following table
5.1 and 5.2 and by using above softwares the mathematical equation is obtain in term of speed, feed and depth of cut for the surface roughness .
USING DESIGNEXPECT SOFTWARE
SURFACE ROUGHNESS (Ra)
TABLE 5.1 ANOVA for Response Surface Quadratic Model
Analysis of variance table [Partial sum of squares – Type III]
Sum of Squares 
DoF 
Mean 
F 
pvalue 

Source 
Value 
Prob > F 

Model 
110.4894389 
9 
12.2766 
561.2594 
< 0.0001 
Significant 
AA 
15.68472238 
1 
15.68472 
717.0711 
< 0.0001 

BB 
18.69230265 
1 
18.6923 
854.571 
< 0.0001 

CC 
0.724005556 
1 
0.724006 
33.09994 
< 0.0001 

AB 
0.889549653 
1 
0.88955 
40.66826 
< 0.0001 

AC 
0.182698201 
1 
0.182698 
8.352561 
0.0102 

BC 
4.18053E05 
1 
4.18E05 
0.001911 
0.9656 

A^2 
0.437292007 
1 
0.437292 
19.99203 
0.0003 

B^2 
73.37755104 
1 
73.37755 
3354.661 
< 0.0001 

C^2 
0.001906597 
1 
0.001907 
0.087165 
0.7714 

Residual 
0.37184638 
17 
0.021873 

Cor Total 
110.8612852 
26 
TABLE 5.2 Analysis of variance (ANOVA) for Surface Roughness
Actor 
Coefficient 
DoF 
Standard 
95% CI 
95% CI 
Estimate 
Error 
Low 
High 
VIF 

Intercept 
3.680614352 
1 
0.077022 
3.518111 
3.843118 

AA 
0.936539474 
1 
0.034974 
0.862751 
1.010328 
1.006579 
BB 
1.022395833 
1 
0.034974 
1.09618 
0.94861 
1.006579 
CC 
0.200555556 
1 
0.034859 
0.127008 
0.74103 
1.013333 
AB 
0.272266667 
1 
0.042694 
0.18219 
0.362343 
1 
AC 
0.122574561 
1 
0.042412 
0.21206 
0.03309 
1.006579 
BC 
0.001854167 
1 
0.042412 
0.09134 
0.087628 
1.006579 
A^2 
0.269966667 
1 
0.060378 
0.39735 
0.14258 
1 
B^2 
3.497083333 
1 
0.060378 
3.369696 
3.624471 
1 
C^2 
0.01869213 
1 
0.063312 
0.15227 
0.114885 
1.013333 
Final Equation in Terms of Coded Factors:
Ra =+ 6.184 + 0.9711* A + 2.2593* B + 0.1811* C 0.1991* A * B
– 0.0989 * A * C – 0.0018* B * C – 0.2677 * A2 – 0.21277 * B2 0.06667* C2
Final Equation in Terms of Actual Factors:
Ra = 5.257 + 0.0402 * A + 29.4195 * B + 1.50896 * C – 0.0221* A
* B – 0.0016637 * A * C – 0.0311 * B * C 4.8 e 5 * A2 – 14.776 * B2 – 0.266666* C2

OPTIMIZATION OF CUTTING PARAMETERS

PARTICLE SWARM OPTIMIZATION (PSO)
In recent years, the collective behavior of large numbers of moving cooperative agents, frequently called particles, has proven to be useful in the fields of optimization and control. The collection of these particles is called a swarm and its application is referred to as swarm intelligence. The power of swarm intelligence is that simple behavior on the local level can result in useful, often more complex, and behavior on the global level. Even if the individual agents are too simple for the label intelligent, the swarm often does manifest intelligent behavior. The global behavior of the swarm is difficult to predict based on the local dynamics of the particles.

GENETIC ALGORITHM
Machining optimization problem requires a robust search method (Robustness means numerical stability and ability to find a solution for a wide range of algorithmic parameters), which runs well in complex situations. The genetic algorithm (GA) approach has been selected firstly because their behavior is robust and so far efficient and secondly more and more attention is being is drawn to GA in a variety of disciplines. Evolution program started in the sixties when a group of biologists used digital computers to simulate genetic systems

PROCEDURE FOR GENETIC ALGORITHM


The sample values given below are taken and these values are used to train the network.

Then the network is exported to the Mat lab workspace and this network is selected and tested by using the same values as input and the layers and numbers of neurons per layer are varied in order to get the best network.

Then this network works as Regression equation and it is used in the genetic algorithm program in order to get the optimum result.

Here we are using minimization problem and the objective is minimized and the optimum parameters to get the minimized objective is found
Procedure of Single Objective Optimization

At first random strings are generated such that the string is divided into the inputs such as speed, feed and depth of cut.

Then the strings are sent into a fitness function, we know that genetic algorithm is based on survival of the fittest.

Then the minimum fitness function is generated, generally GA is used for maximization problems, so the minimum fitness function is converted in maximum fitness function

Then find the cumulative fitness function

By using the cumulative fitness function find the normalized fitness function.

By using the normalized fitness function find the minimum

Then give string numbers to each string.

Then generate a mat pool based on the string numbers.

Then apply the genetic operators on the mat pool such as crossover mutation etc so that a new population is generated. This population is used as an input for the next generation.

Then find the optimum values for finding the minimum fitness function


CONCLUSIONS
GENETIC ALGORITHM
A neural network is created based on experimental values is used instead of mathematical models and optimized machining parameters are found using Genetic algorithm.
By training the network using single objective function and using genetic algorithm optimum speed, feed and depth of cut are evaluated and the corresponding values of objective functions for corresponding speed feed and depth of cut are found and the results are as follows.

By optimization of surface Roughness
Optimum speed = 145.405 m/min Optimum feed = 0.0876 mm/rev Optimum depth = 0.6057 mm Minimized surface roughness =2.494 Âµm
PARTICLE SWARM OPTIMIZATION
By obtained equation form the design of expect software and written particle swarm optimization in clanguage the program obtain the corresponding value of objective function for corresponding speed feed and depth of cut are found and the results are as follows.

By optimization of surface Roughness

Optimum speed = 145m/min Optimum feed = 0 .08 mm/rev Optimum depth = 0.6 mm
Minimized surface roughness =2.601 Âµm
Based on the conducted experiments and accomplished analysis, the following conclusions can be made

The speed and feed rate are the most significant factors in surface roughness model.

All types of lathe machines have been used to produce continuous finished profiles. A continuous finished profile has many types of operations such as turning. To model the machining process, several important operational constraints have been considered. These constraints were taken to account in order to make the model more realistic. A model of the process has been formulated with nontraditional algorithms; GA and PSO have been employed to find the optimal machining parameters for the continuous profile. GA produces better results.
REFERENCES

D. C. MONTGOMERY, Design and Analysis of Experiments, 5TH Edition. Wiley, New York (1984).

P.N.Rao, Text book of Manufacturing Technology, vol.2.

DesignExpert software for design of experiments (DOE) version 8 (v8).

Jiju Antony, Design of Experiments for Engineers and Scientists.

D.R. COX Honorary The Theory of the Design of Experiments by Fellow Nuffield College Oxford, UK.

Dr. S.S.Mahapatra Amar Patnaik Prabina Ku. Patnaik, Parametric Analysis and Optimization of Cutting Parameters for Turning Operations based on Taguchi Method. Proceedings of the International Conference on Global Manufacturing and Innovation – July 2729, 2006

R. Azouzi, M. Guillot. Online optimization of the turning using an inverse process neurocontroller, Transactions of ASME, Journal of Manufacturing Science and Engineering.

A. Mital, M. Mehta Surface roughness prediction models for fine turning; International Journal of Production Research.

V. S. R. K. Prasad et al., Optimal selection of process parameters for turning operations in a CAPP system, International Journal of Production Research, 35, pp.14951522, 1997.

G. C. Onwubolu et al., Multipass turning operations optimization based on genetic algorithms, Proceedings of Institution of Mechanical Engineers, 215, pp. 117124, 2001.

R. Gupta et al., Determination of ptimal subdivision of depth of cut in multipass turning with constraints, International Journal of Production Research, 33(9), pp.25552565, 1995.

C.A. Van Luttervelt, T.H.C. Childs, I.S. Jawahir, F. Klocke, P.K.Venuvinod Present situation and future trends in modeling of machining operations. Progress Report of the CIRP working group on Modeling of machining operations.

K.Taraman Multi machining outputmulti independent variable turning research by response surface methodology, International Journal of Production Research.

Yves Beauchamp, Marc Thomas, Youssef A. Youssef & Jacques Masounave Investigation Of Cutting Parameter Effects On Surface Roughness In Lathe* Boring Operation By Use Of A Full Factorial Design in Computers ind. Engn S, VoL 31, No. 3/4, pp. 645 651,1996 Copyright O 1995 China Mach~ Press Published by Elsevier Science Ltd. Printed in great Britain.