- Email: [email protected]

ScienceDirect Materials Today: Proceedings 4 (2017) 8624–8632

www.materialstoday.com/proceedings

ICAAMM-2016

Optimization of Cutting Parameters for Minimization of Cutting Temperature and Surface Roughness in Turning of Al6061 Alloy N. Rajesh*a, M. Yohanb, P. Venkataramaiahc , M. Vani pallavid *a

Research Scholar, Department of ME, J N T University College of Engineering, Ananthapur, A.P, INDIA b Professor, Department of ME, J N T University College of Engineering, Ananthapur, A.P, INDIA c Associate Professor, Department of ME, S V University College of Engineering, Tirupati, A.P, INDIA d , Department of ME, S V University College of Engineering, Tirupati, A.P, INDIA

Abstract In this paper, turning experiments are conducted on Al6061 work material using HSS tool with and with out coolant at different cutting parameter values and Cutting Temperature, Surface Roughness are recorded for each experiment. Regression models for Cutting Temperature and Surface Roughness are developed to analyze the effects of cutting parameters ( Speed, Feed and Depth of Cut) on Machining responses. The developed Regression Equations are solved by using Genetic Algorithm to obtain optimal values for cutting parameters © 2017 Elsevier Ltd. All rights reserved. Selection and Peer-review under responsibility ofthe Committee Members of International Conference on Advancements in Aeromechanical Materials for Manufacturing (ICAAMM-2016). Keywords: Optimization; Turning; Al6061; Taguchi; Cutting Temperature; Surface Roughness; Regression equation, Genetic Algorithrm

Modern machining industries are mainly focused on the achievement of high quality, in terms of work piece dimensional accuracy, surface finish, cutting temperature, high production rate, less wear on the cutting tools, economy of machining in terms of cost saving and increase the performance of the product with reduced environmental impact. Surface roughness and cutting temperature play an important role in many areas and are the factors of great importance in the evaluation of machining accuracy in turning.

* N Rajesh. Tel.: +91-9985289928 E-mail address: [email protected]

2214-7853© 2017 Elsevier Ltd. All rights reserved. Selection and Peer-review under responsibility ofthe Committee Members of International Conference on Advancements in Aeromechanical Materials for Manufacturing (ICAAMM-2016).

N. Rajesh/ Materials Today: Proceedings 4 (2017) 8624–8632

8625

Nomenclature N F Dc T Ra GA

Speed in RPM Feed in m/rev Depth of cut in mm Temperature in 0C Surface Roughness in Genetic Algorithm

μm

Doriana M. [1] proposed an optimization paradigm based on genetic algorithms (GA) for the determination of the cutting parameters in machining operations. Paszkowicz W [3] Optimization algorithms are the branch of intelligent methods used to find optimal machining conditions. Genetic algorithm (GA) is one of the most popular evolutionary optimization algorithms. Artificial neural network and genetic algorithm as evolutionary procedures have been successfully used in the past for typical multi objective optimization problem by researchers [4-7]. Kuldip Singh Sangwan [2] has determined an approach for determining the optimum machining parameters leading to minimum surface roughness by integrating Artificial Neural Network(ANN) and Genetic Algorithm (GA). Rajesh .N [8] has focused on development of an Artificial Neural Network ( ANN) model for predicting surface roughness of a work material in orthogonal turning and also a Taguchi Signal to Noise ratio (S/N) analysis is used to identify the optimum control parameters. Venkataramaih P. [10] have conducted turning experiments on Aluminum Alloy 6061 work material for different values of cutting parameters and experimental responses such as cutting temperature and surface finish are measured and recorded. This Experimental data is used for development of ANN model to predict the cutting temperature and surface finish for different values of cutting parameters. Young Kug Hwang [11] investigated into the MQL (minimum quantity lubrication) and wet turning processes of AISI 1045 work material with the objective of suggesting the experimental model in order to predict the cutting force and surface roughness. This paper focused on development of regression models for the responses in turning of Aluminium and finding the optimum values for the cutting parameters. 2. Experimental work 2.1 Experimental Procedure In this paper, turning experiments are conducted on Al6061 for different speeds, feeds and depth of cuts based on its optimum values are considered from the responses from [8] with HSS cutting tool. Turning is performed on the work material as shown in fig.1 by keeping two input parameters constant and by varying one parameter, under dry and coolant the values of surface roughness and cutting temperature are recorded.

(a) (b) (c) Fig.1. a) Experimental setup b) Temperature Gun c) Talysurf – surface measurement tester

8626

N. Rajesh / Materials Today: Proceedings 4 (2017) 8624–8632

2.2. Recording of Experimental data Turning experiments are conducted at dry and coolant conditions and experimental data are recorded. Case (i): Dry condition Case.1 : Experiments are conducted keeping the input parameters Speed and Feed as constant and varying Depth of Cut, the output parameters Temperature (T) and Surface Roughness (Ra) are recorded and shown in Table 1. Table 1. Case - 1, Dry Condition, Al6061, HSS input Parameters Exp Run

output parameters

Speed

Feed

Doc

1

1600

0.13

2

1600

3 4

Temperature (T)

Surface Roughness (Ra)

T-1

T-2

T-3

Avg

T-1

T-2

T-3

Avg

0.05

28

28.5

27.9

28.133

0.65

0.81

0.6

0.6867

0.13

0.1

27.9

28

27.7

27.867

0.64

0.41

0.78

0.61

1600

0.13

0.15

28.1

28.2

27.5

27.933

0.78

0.9

0.8

0.8267

1600

0.13

0.2

27.9

28.4

29.2

28.5

0.94

0.84

0.5

0.76

5

1600

0.13

0.25

28.6

29.2

29.7

29.167

0.75

0.83

0.59

0.7233

6

1600

0.13

0.3

27.6

30.2

31.6

29.8

0.79

0.72

0.82

0.7767

7

1600

0.13

0.35

35.5

36.2

35.3

35.667

0.58

0.67

0.67

0.64

8

1600

0.13

0.4

34.3

31.9

33.4

33.2

0.91

0.89

0.7

0.8333

Case.2 : Experiments are conducted keeping the input parameters Depth of Cut and Feed as constant and varying Speed, the output parameters Temperature (T) and Surface Roughness (Ra) are recorded and shown in Table 2. Table 2. Case - 2, Dry Condition, Al6061, HSS input Parameters Exp Run

output parameters Temperature (T)

Surface Roughness (Ra)

Speed

Feed

Doc

1

228

0.13

2

360

3

450

4

580

5

740

0.13

0.1

28.3

28.8

29.1

28.733

0.67

0.5

0.82

0.6633

6

800

0.13

0.1

28.2

28.6

28.8

28.533

0.84

0.82

0.91

0.8567

7

1150

0.13

0.1

29.9

33.3

34.5

32.567

0.8

0.8

0.8

0.8

8

1600

0.13

0.1

33.5

32.2

32.3

32.667

0.8

0.6

0.84

0.7467

T-1

T-2

T-3

Avg

T-1

T-2

T-3

Avg

0.1

27.9

28.4

28.5

28.267

0.97

0.91

0.73

0.87

0.13

0.1

29.1

28.9

29.5

29.167

0.76

0.94

0.84

0.8467

0.13

0.1

28.2

28.3

28.5

28.333

0.88

0.9

0.9

0.8933

0.13

0.1

29

28.6

28.4

28.667

0.85

0.84

0.91

0.8667

Case.3 : Experiments are conducted keeping the input parameters Depth of Cut and Speed as constant and varying Feed, the output parameters Temperature (T) and Surface Roughness (Ra) are recorded and shown in Table 3.

N. Rajesh/ Materials Today: Proceedings 4 (2017) 8624–8632

8627

Table.3. Case - 3, Dry Condition, Al6061, HSS input Parameters Exp Run

output parameters

Speed

Feed

Doc

1

1600

0.05

2

1600

0.07

3

1600

4

Temperature (T)

Surface Roughness (Ra)

T-1

T-2

T-3

Avg

T-1

T-2

T-3

Avg

0.1

30.09

30.06

33.07

31.073

0.72

0.85

0.99

0.8533

0.1

30.09

30.07

29.09

29.75

1.22

0.46

0.87

0.85

0.09

0.1

31.5

30.2

31.1

30.933

0.61

1

1.02

0.8767

1600

0.1

0.1

29.9

31.5

31.8

31.067

0.92

0.92

0.92

0.92

5

1600

0.13

0.1

30.6

30.1

30.2

30.3

0.91

0.75

0.87

0.8433

6

1600

0.16

0.1

31.6

32.4

33.1

32.367

0.88

1.08

0.99

0.9833

7

1600

0.18

0.1

32.8

31.5

31.7

32

0.83

0.87

0.88

0.86

8

1600

0.2

0.1

31.1

30.9

30.7

30.9

0.78

0.84

0.93

0.85

Case (ii): Coolant Condition (MQL) Case 1 : Experiments are conducted keeping the input parameters Speed and Feed as constant and varying Depth of Cut, the output parameters Temperature (T) and Surface Roughness (Ra) are recorded and shown in Table 4. Table 4. Case -1 , Kerosene, Al6061, HSS input Parameters Exp Run

Speed

Feed

Doc

output parameters Temperature (T)

Surface Roughness (Ra)

T-1

T-2

T-3

Avg

T-1

T-2

T-3

Avg

1

1600

0.13

0.05

28.1

28.6

28.9

28.533

0.8

0.97

0.94

0.9033

2

1600

0.13

0.1

28

28.4

28.3

28.233

0.95

0.76

0.56

0.7567

3

1600

0.13

0.15

29.2

29.1

29

29.1

0.95

0.54

0.8

0.7633

4

1600

0.13

0.2

27.4

27.7

27.9

27.667

0.62

0.64

0.82

0.6933

5

1600

0.13

0.25

29.9

29.2

28.7

29.267

0.79

0.77

0.68

0.7467

6

1600

0.13

0.3

28.4

28.9

29

28.767

0.75

0.89

0.77

0.8033

7

1600

0.13

0.35

28.9

29.2

29.5

29.2

0.87

0.72

0.86

0.8167

8

1600

0.13

0.4

28.6

28.7

28.7

28.667

0.77

0.66

0.8

0.7433

Case 2 : Experiments are conducted keeping the input parameters Feed and Depth of Cut as constant and varying Speed, the output parameters Temperature (T) and Surface Roughness (Ra) are recorded and shown in Table 5.

8628

N. Rajesh / Materials Today: Proceedings 4 (2017) 8624–8632

Table 5. Case -2 , Kerosene, Al6061, HSS input Parameters Exp Run

Speed

Feed

Doc

1

228

0.13

2

360

0.13

3

450

4 5

output parameters Temperature (T)

Surface Roughness (Ra)

T-1

T-2

T-3

Avg

T-1

T-2

T-3

Avg

0.1

26

26.4

26.4

26.267

0.57

0.48

0.28

0.4433

0.1

26

26.3

26.7

26.333

0.67

0.77

0.56

0.6667

0.13

0.1

25.5

25.9

25.5

25.633

0.41

0.4

0.79

0.5333

580

0.13

0.1

26.1

26.2

21

24.433

0.42

0.49

0.35

0.42

740

0.13

0.1

25.6

25.2

26.2

25.667

0.58

0.66

0.85

0.6967

6

800

0.13

0.1

25.7

25.2

24.8

25.233

0.41

0.39

0.48

0.4267

7

1150

0.13

0.1

25.9

26.1

26.2

26.067

0.3

0.49

0.49

0.4267

8

1600

0.13

0.1

25.7

25.9

26.1

25.9

0.63

0.56

0.77

0.6533

Case.3 : Experiments are conducted keeping the input parameters Speed and Depth of Cut as constant and varying Feed, the output parameters Temperature (T) and Surface Roughness (Ra) are recorded and shown in Table 6. Table 6. Case -3 , Kerosene, Al6061, HSS input Parameters

output parameters

Exp Run

Speed

Feed

Doc

T-1

T-2

T-3

Avg

T-1

T-2

T-3

Avg

1

1600

0.05

0.1

27.9

26.9

27.4

27.4

0.24

0.48

0.38

0.3667

2

1600

0.07

0.1

27.4

27.2

27.9

27.5

0.29

0.15

0.22

0.22

3

1600

0.09

0.1

28

29

31.4

29.467

0.35

0..53

0.48

0.415

4

1600

0.1

0.1

30

33.4

29

30.8

0.46

0.38

0.33

0.39

5

1600

0.13

0.1

31

29.8

32.2

31

0.4

0.36

0.33

0.3633

6

1600

0.16

0.1

29

31.2

31

30.4

0.32

0.47

0.54

0.4433

7

1600

0.18

0.1

28.1

27.9

29.3

28.433

0.32

0.45

0.46

0.41

8

1600

0.2

0.1

27.2

27.3

27.1

27.2

0.3

0.31

0.35

0.32

Temperature (T)

Surface Roughness (Ra)

3. Development of Regression Models Machining optimization provides optimal or near-optimal solutions in actual metal cutting process. The optimization procedure has two phases. First phase is mathematical modeling of the machining process (cutting performances) where an objective function should be defined. In that phase, all constraints and bounds of the variables, by using equalities and (or) inequalities should be defined too. Second phase is searching for a global minimum of objective function, under all defined limitations. The Regression models are developed using Minitab software for each case and are formulated as in the following

N. Rajesh/ Materials Today: Proceedings 4 (2017) 8624–8632

8629

Objective Functions Case (i): Dry Condition Case 1. f1(Dc) = T = 0.8769 + 1.756 Dc – 9.131 Dc2

(1)

f2(Dc) = Ra = 0.8257 – 2.683 Dc + 3.62 Dc 2

(2)

Combined model, Z = f1(Dc) + f(Dc) = 0.8513 – 0.4635 Dc – 2.755 Dc 2

(3)

Case 2. f1(N) = T = 1.057 – 0.000110 N

(4)

f2(N) = Ra = -0.143 + 0.000903 N

(5)

Combined model, Z = f1(N) + f2(N) = 0.457 – 0.00039N

(6)

Case 3. f1(F) = T = 0.9503 – 4.57 F + 6.47 F2

(7)

f2(F) = Ra = 1.784 -18.48 F + 69.46 F2

(8)

Combined model, Z = f1(F) + f2(N) = 1.367 – 11.52 F +37.96 F2

(9)

Case (ii): MQL Condition Case 1. f1(Dc) = T = 0.6066 - 1.238 Dc +0.64 Dc2

(10)

f2(Dc) = Ra = -0.0797+6.841 Dc - 13.58 Dc 2

(11)

Combined model, Z = f1(Dc) + f(Dc) = 0.263 +2.8015 Dc – 6.47 Dc 2

(12)

Case 2. f1(N) = T = -0.2267+ 0.001643 N

(13)

f2(N) = Ra = 0.3523 + 0.000894 N

(14)

Combined model, Z = f1(N) + f2(N) = 0.06+ 0.0012N

(15)

Case 3. f1(F) = T = 2.965 – 45.73 F + 179.9 F2

(16)

f2(F) = Ra = 1.281 -15.60 F + 55.55 F2

(17)

Combined model, Z = f1(F) + f2(N) = 2.123 – 30.66 F +117.72 F2

(18)

Lower limit & Upper Limit 0.05

4.

Solving the Regression Models for optimum parameter values by using GA

4.1 Genetic Algorithm Genetic Algorithms (GA) are search algorithms based on the mechanics of natural selection and natural genetics. GA then iteratively creates new populations from the old by ranking the strings and interbreeding the fittest to create new, and conceivably better, populations of strings which are (hopefully) closer to the optimum solution to the problem at hand. So in each generation, the GA creates a set of strings from the bits and pieces of the previous strings, occasionally adding random new data to keep the population from stagnating. The end result is a search strategy that is tailored for vast, complex, multimodal search spaces. GA is a form of randomized search, in That the way in which strings are chosen and combined is a stochastic process.

8630

N. Rajesh / Materials Today: Proceedings 4 (2017) 8624–8632

4.2 Basic Genetic Algorithm operations There are three basic operators found in every genetic algorithm: selection, crossover and Mutation. The standard procedure of GA [9] is as shown in the Figure. 2.

Figure.2. procedure of GA

4.3 Searching for optimum values of objective function with GA The developed regression models shown in section -3 are solved by using genetic algorithm tool box. The obtained results are shown in the Table.7. These results are verified experimentally. The solution procedure for the depth of cut case is shown in the following. •

Objective Function Min Z = f1(Dc) + f(Dc) = 0.8513 – 0.4635 Dc – 2.755 Dc 2

•

lower and upper bounds as 0.05

(19)

when the optimization process is terminated, the minimal value for the objective function (19) is found to be Zmin = 0.4 mm. The result is obtained in the 51st iteration. GA parameters of this case is shown in the Figure. 3, are listed below Selected Options: Population type : Double vector Population size: 20 Elite count: 2 Crossover fraction: 0.8 Pareto fraction: [] Migration direction: forward Migration interval: 20 Migration fraction: 0.2 Generations: 100 Time limit: Inf Fitness limit: -Inf Initial population: [15 9] Initial scores: [] Initial penalty: 10 Penalty factor: 100 Plot interval: 1 Creation fcn: @gacreationuniform Fitness scaling fcn: @fitscalingrank

N. Rajesh/ Materials Today: Proceedings 4 (2017) 8624–8632

Selection fcn: @selectionstochunif Crossover fcn: @crossoverscattered Mutation fcn: [1x1 function_handle] [1] [1] Distance measure fcn: [] Hybrid fcn: [] Display: diagnose Plot fcns: @gaplotbestf @gaplotbestindiv Output fcns: [] @gatooloutput Vectorized: off Use parallel: never Diagnostic information: Fitness function = @ dry1 Number of variables = 1 Nonlinear constraint function = 0 0 Inequality constraints 0 Equality constraints 0 Total numbers of linear constraints

Fig. 3 GA toolbox of Matlab with selected options

Table. 7 optimum parameter values obtained from GA

Case

Case

Dry -1

Dry-2

Dry-3

Dc 0.4

N 1589.6

F 0.152

MCF-1

MCF-2

MCF-3

Dc 0.05

N 228

F 0.13

8631

8632

5.

N. Rajesh / Materials Today: Proceedings 4 (2017) 8624–8632

Conclusions

Turning experiments are successfully conducted and optimum cutting parameters values are determined for minimizing cutting temperature and surface roughness using Genetic Algorithm at dry and MQL condition. This work is more useful to model the cutting responses in view of predicting the optimum cutting parameter values. References [1] Doriana M., et.al., Genetic algorithm-based optimization of cutting parameters in turning processes, Forty Sixth CIRP Conference on Manufacturing Systems 2013, Procedia CIRP 7 , 2013, 323 – 328. [2] Kuldip Singh Sangwan, et.al., Optimization of Machining Parameters to Minimize Surface Roughness using Integrated ANN-GA Approach, The 22nd CIRP conference on Life Cycle Engineering, Procedia CIRP 29, 2015, 305 – 310. [3] Paszkowicz W. Genetic algorithms, a nature-inspired tool: a survey of applications in materials science and related fields: Part II, Materials and Manufacturing Processes, (2013), 28(7), 708–725. [4] Pettersson F., Biswas A., Sen P.K., Saxena H., Chakraborti N. Analyzing leaching data for low-grade manganese ore using neural nets and multi objective genetic algorithms, Materials and Manufacturing Processes, 2009, 24(3), 320–330. [5] Nakhjavani O. B., Ghoreishi M..Multi Criteria Optimization of Laser Percussion Drilling Process Using Artificial Neural Network Model Combined with Genetic Algorithm, Materials and Manufacturing Processes, 2006, 21(1), 11-18. [6] Pal Sukhomay, Pal Surjya K., Samantaray Arun K. .Determination of Optimal Pulse Metal Inert Gas Welding Parameters with a Neuro-GA Technique, Materials and Manufacturing Processes, 2010,25(7),606-615. [7] Somashekhar K. P., Ramachandran N., Mathew Jose. Optimization of Material Removal Rate in Micro-EDM Using Artificial Neural Network and Genetic Algorithms, Materials and Manufacturing Processes, 2010, 25(6), 467 -475. [8] Rajesh. N, et.al., Prediction of Surface Roughness for Optimized Control Parameters in Turning using ANN, National Conference on Condition Monitoring (NCCM), October 4-5, 2013, Bangalore, NCCM-2013-23 [9] Radovanovic, M., Marinkovic, V., Graphical-Analytical Method for Determining the Optimal Cutting Parameters, Proceeding of the International Conference Mechanical Engineering in XXI Century, University of Nis, Faculty of Mechanical Engineering, Nis, Serbia, 2010, p. 171-174. [10] Venkataramaih. P, et.al., Prediction and study on cutting temperature and surface finish in turning of aluminum alloy 6061 using ANN, National Conference on Condition Monitoring (NCCM), September 25-26, 2015, Vishakhapatnam, NCCM-2015-4 [11] Young Kug Hwang, et.al., Surface roughness and cutting force prediction in MQL and wet turning process of AISI 1045 using design of experiments, Journal of Mechanical Science and Technology 24 (8), 2010, 1669~1677.