International Journal of Science Engineering and Technology Vol. 1, No. 3, 2008 ISSN: 1985-3785 Available online at: www.ijset.

org 2008 ILRAM Publisher

Application of Taguchi Method in the Optimization of Turning Parameters for Surface Roughness
1
1

G. Akhyar, 1C.H. Che Haron, 1J.A. Ghani

Department of Mechanics and Materials, Universiti Kebangsaan Malaysia

Abstract The quality of design can be improved by improving quality and productivity in company-wide activities. Taguchis parameter design is an important tool for robust design, which offers a simple and systematic approach to optimize a design for performance, quality and cost. Taguchi optimization methodology is applied to optimize cutting parameters in turning Ti-6%Al-4%V extra low interstitial with coated and uncoated cemented carbide tools under dry cutting condition and high cutting speed. The turning parameters evaluated are cutting speed of 55, 75, and 95 m/min, feed rate of 0.15, 0.25 and 0.35 mm/rev, depth of cut of 0.10, 0.15 and 0.20 mm and tool grades of K313, KC9225 and KC5010, each at three levels. The analysis of results show that the optimal combination of parameters are at cutting speed of 75 m/min, feed rate of 0.15 mm/min, depth of cut of 0.10 mm and tool grade of KC9225. The cutting speed and tool grade have a significant effect on surface roughness are 0.000 and have a contribution are 47.146% and 38.881%, respectively. At optimal condition, contribution of each cutting parameter on surface roughness is reached at 20.47 from tool grade, 21.01 from feed rate, 11.54 from depth of cut and 11.17 from cutting speed. KEYWORDS: Taguchi Method, Turning, Ti-6Al-4V ELI and Coated Carbide Tool.

1. Introduction The quality of design can be improved by improving the quality and productivity in companywide activities. Those activities concerned with quality, include in quality of product planning, product design and process design [1, 3]. Robust design is an engineering methodology for obtaining product and process condition, which are minimally sensitive to the various causes of variation, and which produce high-quality products with low development and manufacturing costs [1]. Taguchis parameter design is an important tool for robust design. It offers a simple and systematic approach to optimize design for performance, quality and cost. Signal to noise ratio and orthogonal array are two major tools used in robust design. Signal to noise ratio, which measures quality with emphasis on variation, and orthogonal arrays, which accommodates many design factors simultaneously [1, 2]. Taguchi method offers the quality of product is measured by quality characteristics such as: nominal is the best, smaller is better and larger is better [1, 3]. Optimization using Taguchi method in end milling using conceptual S/N ratio approach and Pareto ANOVA proceed, the Taguchis robust design method is suitable to analyze the metal cutting problem. Ghani et.al [6] found that the conceptual S/N ratio and Pareto ANOVA approaches for data analysis draw similar conclusion in process end milling use at high
Corresponding Author:

cutting speed of 355 m/min, low feed rate of 0.1mm per tooth and low depth of cut of 0.5 mm. Application of Taguchis method for parametric design was carried out to determine an ideal feed rate and desired force combination Although small interactions exist between a horizontal feed rate and desired force, the experimental results showed that surface roughness decreases with a slower feed rate and larger grinding force, respectively [7]. Conceptual S/N ratio approach of Taguchi method provides a simple, systematic and efficient methodology for optimizing of process parameters and this approach can be adopted rather than using engineering judgment. Furthermore, the multiple performance characteristics such as tool life, cutting force, surface roughness and the over all productivity can be improved by useful tool of Taguchi method [8]. This paper describes the turning of Ti-6Al-4V ELI with parameters of turning at three levels and four factors each. The main objective is to find a specific range and combination of turning parameters and interaction to achieve the lowest surface roughness value.

2. Taguchi method, design of experiment, and experimental details 2.1. Taguchi method

G. Akhyar, Department of Mechanics and Material Engineering, University Kebangsaan Malaysia, Kuala Lumpur, Malaysia, E-mail: sati1771@gmail.com

G. Akhyar et al. / Int. J. Sci. Eng. Tech. Vol. 1, No. 3, 2008, 60-66

Taguchi defines as the quality of a product, in terms of the loss imparted by the product to the society from the time the product is shipped to the customer [2]. Some of these losses are due to deviation of the products functional characteristic from its desired target value, and these are called losses due to functional variation. The uncontrollable factors, which cause the functional characteristics of a product to deviate from their target values, are called noise factors, which can be classified as external factors (e.g. temperatures and human errors), manufacturing imperfections (e.g. unit to unit variation in product parameters) and product deterioration. The overall aim of quality engineering is to make products that are robust with respect to all noise factors. Taguchi has empirically found that the two stage optimization procedure involving S/N ratios, indeed gives the parameter level combination, where the standard deviation is minimum while keeping the mean on target [2, 3, 4]. This implies that engineering systems behave in such a way that the manipulated production factors that can be divided into three categories: 1.Control factors, which affect process variability as measured by the S/N ratio. 2. Signal factors, which do not influence the S/N ratio or process mean.

3. Factors, which do not affect the S/N ratio or process mean. In practice, the target mean value may change during the process development applications in which the concept of S/N ratio is useful are the improvement of quality through variability reduction and the improvement of measurement. The S/N ratio characteristics can be divided into three categories when the characteristic is continuous: nominal is the best, smaller the better and larger is better characteristics. 2.2. Experimental details The experiments were carried out with four factors at three levels each, as shown in Table 1. The fractional factorial design used is a standard L27 (313) orthogonal array with 20 degree of freedom [1]. This orthogonal array is chosen due to its capability to check the interactions among factors. The machining trials were carried out on the lathe machine (Colchester T4 with maximum 6000 rpm) in dry condition, as recommended by the tool supplier for the specific work material. The three inserts used were uncoated carbide tool K313 (WC-Co), coated carbide tool KC9225 (TiNAl2O3-TiCN-TiN) CVD and KC5010 (TiAlN) PVD, respectively.

Table 1. Factors and levels used in the experiment Factors A- Cutting speed (m/min) B- Feed rate (mm/rev) C- Depth of cut (mm) D- Tool type Levels 1 75 0.25 0.15 KC9225

0 55 0.15 0.10 K313

2 95 0.35 0.20 KC5010

The maximum flank wear land (VB) was measured at regular interval of one pass machining using Mitutoyo Tool Maker Microscope with 20x magnification. The surface roughness of machined surface was then measured accordingly surface roughness tester model Mpi Mahr Perthometer. The turning process was stop when VB reached 0.2 mm.

depth of cut and tool grade) to achieve low value of the surface roughness. The experimental data for the surface roughness values and the calculated signal-to-noise ratio are shown in Table 2. The S/N ratio values of the surface roughness are calculated, using the smaller the better characteristics [3, 4].

3. Results and Discussions 3.1 Signal to Noise Ratio (S/N) The main objective of the experiment is to optimize the turning parameters (cutting speed, feed rate,

10 log

1 n

y
2

(1)

61

G. Akhyar et al. / Int. J. Sci. Eng. Tech. Vol. 1, No. 3, 2008, 60-66

Table 2. Experimental results for surface roughness and its calculated S/N ratios.
Exp. run. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Factor A 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 B 0 0 0 1 1 1 2 2 2 0 0 0 1 1 1 2 2 2 0 0 0 1 1 1 2 2 2 C 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 D 0 1 2 1 2 0 2 0 1 0 1 2 1 2 0 2 0 1 0 1 2 1 2 0 2 0 1 Designation A0B0C0D0 A0B0C1D1 A0B0C2D2 A0B1C0D1 A0B1C1D2 A0B1C2D0 A0B2C0D2 A0B2C1D0 A0B2C2D1 A1B0C0D0 A1B0C1D1 A1B0C2D2 A1B1C0D1 A1B1C1D2 A1B1C2D0 A1B2C0D2 A1B2C1D0 A1B2C2D1 A2B0C0D0 A2B0C1D1 A2B0C2D2 A2B1C0D1 A2B1C1D2 A2B1C2D0 A2B2C0D2 A2B2C1D0 A2B2C2D1 Surface roughness, Ra (m) 1.71 1.17 2.50 1.09 4.94 3.48 6.01 6.49 2.82 0.53 1.56 1.44 4.67 3.02 0.97 3.94 2.56 6.19 2.02 1.73 1.10 3.56 2.37 4.31 1.29 4.26 5.20 S/N ratio for surface roughness -4.66 -1.364 -7.959 -0.749 -13.875 -10.832 -15.578 -16.245 -9.067 5.514 -3.863 -3.168 -13.387 -9.601 0.264 -11.91 -8.165 -15.834 -6.108 -4.761 -0.828 -11.029 -7.495 -12.69 -2.212 -12.589 -14.321

Table 2 shows the actual data of surface roughness along with its computed S/N ratio value. Whereas the S/N ratio for each levels of the surface roughness as shown in Table 3. In the standard L27 (313) orthogonal array, factor A, B, C and D are arranged in columns 1,2, 5 and 9, respectively. Whereas interaction factors between the cutting speed and feed rate (AxB), the cutting speed and depth of cut (AxC) and the feed rate and depth of cut (BxC) are arranged in columns 3, 6 and 8, respectively. 3.2 Analysis of variance for S/N ratio Taguchi recommends to analyze data using the S/N ratio that will offer two advantages; it provides a guidance for selection the optimum level based on least variation around on the average value, which closest to target, and also it offers objective comparison of two sets of experimental data with respect to deviation of the average from the target [3]. The experimental results are analyzed to investigate the main effects and differences between the main effects of level 0, 1 and 2 on the variables.

Average S/N ratio for each level of experiment is calculated based on the value of Table 1, and is shown in Table 2. The different values of the S/N ratio between maximum and minimum (main effect) are also shown in Table 2. The feed rate and the tool grade are two factors with the highest different in values of 9.553 and 8.445, respectively. Based on Taguchi prediction that the bigger different in value of S/N ratio shows a more effect on surface roughness or more significant. Therefore, it can be concluded that, increase changes the feed rate reduces the surface roughness significantly. Furthermore, the tool geometry changes, mainly tool nose radius, increase or decrease the surface roughness significantly. The result of S/N ratio analysis for the surface roughness values, which was calculated using Taguchi Method, is shown in Table 3. Then, analysis of variance is shown in Table 4, which consists of DF (degree of freedom), S (sum of square), V (variance), F (variance ratio) and P (significant factor) [3, 4]. In most engineering cases, the significant value selected was 5% ( = 0.05).

62

G. Akhyar et al. / Int. J. Sci. Eng. Tech. Vol. 1, No. 3, 2008, 60-66

Table 3. Average for S/N ratio and main effect of surface roughness
No. columns 1 2 5 9 Variable Factors Cutting speed Feed rate Depth of cut Tool grade Designation 0 A B C D -8.918 -3.022 -7.063 -9.403 Level average 1 2 -7.221 -8.213 -8.757 -12.574 -8.961 -8.329 -3.252 -11.697 Max. - Min. 1.697 9.553 1.897 8.445 Rank 4 1 3 2

Table 4. ANOVA analysis of S/N ratio for surface roughness

No 1 2 3 5 6 8 10 Varible Factors Cutting speed Feed rate Cutting speed x feed rate Depth of cut Cutting speed x depth of cut Feed rate x depth of cut Tool grade Error Total Designation A B AxB C AxC BxC D DF 2 2 4 2 4 4 2 6 26 Sum of Variance Square (S) (V) F Percent (P) 0.327 0.000 0.170 0.253 0.862 0.614 0.000 Contribution (%) 1.482 47.146 5.082 1.903 0.675 1.552 38.881 3.279 100

13.083 6.542 1.36 416.179 208.089 43.13 44.858 11.214 2.32 16.800 8.400 1.74 5.958 1.490 0.31 13.697 3.424 0.71 243.223 171.611 35.57 28.949 4.825

Table 4 shows that the significant value of the feed rate and tool grade (P) is 0.000. It means that the feed rate and tool grade influences significantly on the surface roughness value at significant value of 0.05. In addition to P value for the cutting speed and depth of cut are insignificant. The feed rate and the tool grade have a contribution for the surface roughnesses are 47.146 % and 38.881%, respectively. From this result, it can be concluded that the feed rate is more significant factor and give most contribution on the surface roughness. Bhattacharyya found that the surface roughness was primarily dependent on the feed rate and the nose radius of tool [9]. The nose radius related to tool grade and tool geometry. Since three types of tool were applied in this experiment have different tool nose radius, so effect of tool nose geometry changes on surface roughness was significant. The interaction between the cutting speed and feed rate (AxB), the cutting speed and depth of cut (AxC) and the feed rate and depth of cut (BxC) are also insignificant. These significant values of interaction are

0.170 from AxB, 0.862 from AxC and 0.614% from BxC. While, a contribution for each interaction is small. The most significant factor, which affects the surface roughness measured in turning Ti-6Al-4V, is the feed rate therefore the quality of surface roughness can be controlled by a suitable feed rate value. Previous researchers suggest similar results. They claimed that the surface roughness well strongly depends on the feed rate followed by the cutting speed. Jaharah et al. [6] recommended to obtain better surface finish for specific test range in end milling was use of high cutting speed (355 m/min), low feed rate (0.1 mm/tooth) and low depth of cut (0.5 mm). Table 5 shows level of contribution each parameter and interaction on surface roughness values for estimating optimum condition. The biggest contribution is from feed rate of 0.15 mm/rev (21.01%), and then followed by tool grade of KC9225 (20.47%). The contribution from depth of cut and cutting speed were 11.54% and 11.17%, respectively.

63

G. Akhyar et al. / Int. J. Sci. Eng. Tech. Vol. 1, No. 3, 2008, 60-66

The optimum condition in turning of aerospace material Ti-6Al-4V ELI which produces a low surface roughness is at cutting speed of level 1 (0.75m/min), feed rate of level 0 (0.15 mm/rev), depth of cut of level 0 (0.10 mm) and tool grade of level 1 (KC9225). Meanwhile, optimum condition for interaction

factors is the cutting speed and feed rate of level 1, the cutting speed and depth of cut of level 0, and the feed rate and depth of cut of level 0. The total contribution for optimum condition is 383,898 and its expected result is 376.028.

Table 5. Estimate of optimum condition for the smaller the better characteristics.
Factors description Cutting speed (A) Feed rate (B) Cutting speed x federate (AxB) Depth of cut (C) Cutting speed x depth of cut (AxC) Feed rate x depth of cut (BxC) Tool grade (D) Contribution for all factors (total) Current grand average of performance Expected result at optimum condition Level of description 0.75 0.15 0.1 KC9225 Levels Contribution Contribution (%) 1 0 1 0 0 0 1 42.881 80.672 51.780 44.303 43.300 42.360 78.602 383.898 -7.87 376.028 11.17 21.01 13.49 11.54 11.28 11.03 20.47 100.00

Main Effects Plot (data means) for SN ratios

Cutting speed -4 -6 -8 Feedrate

Mean of SN ratios

-10 -12 0 -4 -6 -8 -10 -12 0 1 2 0 1 2 1 Depth of cut 2 0 1 T ype of tool 2

Signal-to-noise: Smaller is better

Figure 1. Main effects for factors machining verse S/N ratio of surface roughness.

The main effects for each level of parameter on surface roughness are shown in Figure 1. The best choice for machining titanium alloy is based on S/N ratio as followed; at cutting speed of -8.918 (A1), feed rate of -3.022 (B0), depth of cut of -7.063 (C0) and tool grade of -3.252 (D1). The best combination is A1B0C0D1 that means at low high cutting speed, low feed rate, low depth of cut and CVD coated carbide tool.

It can be seen from Figure 1 that B0 is the maximum value with 3.022 of S/N ratio, and decrease dramatically to B1 (-8.757), and then follow to B2 (12.574). For the graph of feed rate, the slope between the horizontal and feed rate line is bigger. It means that the feed rate changes effected significantly on surface roughness, and the same trend can also be

64

G. Akhyar et al. / Int. J. Sci. Eng. Tech. Vol. 1, No. 3, 2008, 60-66

observed on the graph of tool grade factor for each level. Figure 2 shows the interaction between the cutting speed and feed rate (AxB), the cutting speed and depth of cut (AxC) and the feed rate and depth of cut (BxC). The S/N ratio value at (AxB)1 is a best interaction because of it gives the biggest delta value, and then followed by interaction (AxC)0. The cutting speed at level 1

(A1) and the feed rate at level 0 (B0) have a maximum value. It can be also seen from Table 5 that the optimum predicted result for each interaction gives contribution is 13.49% from AxB, 11.28% from AxC and 11.03% from BxC. The contribution for all factors is 383.898, while the expected result at optimum condition is 376.028.

Interaction Plot (data means) for SN ratios

0 0 1 2

-5 C utting spee d -10 0 -5 F eedr ate -10 0

C utting speed 0 1 2

Feedrate 0 1 2

-5 Depth of cut -10

Depth of cut 0 1 2

Signal-to-noise: Smaller is better

Figure 2. Interaction S/N ratio for surface roughness with smaller is better

4. Conclusions From the findings of the following can be concluded: 1. Taguchis robust design method is suitable to optimize the surface roughness in turning Ti-6Al-4V ELI. 2. The significant factors for the surface roughness in turning Ti-6Al-4V ELI were the feed rate and the tool grade, with contribution of 47.146% and 38.881%, respectively. 3. The optimal condition for surface roughness in turning Ti-6Al-4V ELI was resulted at cutting speed of 75 m/min, feed rate of 0.15 mm/rev, depth of cut of 0.10 mm and CVD coated carbide with KC9225. 4. The optimal interaction parameter was between the cutting speed and feed rate at level 1 (75 m/min).

References [1] Park, S.H. Robust Design and Analysis for Quality Engineering; Chapman & Hall, London, 1996. [2] Phadke, M.S. Quality Engineering Using Design of Experiment, Quality Control, Robust Design and The Taguchi Method; Wadsworth & Books, California, 1988. [3] [Ranjit, R. Design of experiment Using The Taguchi Approach; John Wiley & Sons Inc., New York, 2001. [4] Ranjit, R. A Primer on The Taguchi Method, Society of Manufacturing Engineers; Dearborn, Michigan, 1999. [5] Bendell, T. Taguchi Method, Proceedings of the European Conference of Taguchi Method, Elsevier, Amsterdam, July 13-24, 1988. [6] Ghani, J.A., Choudhury, I.A., Hasan, H.H. Application of Taguchi Method in Optimization of End Milling Parameters, Journal of Materials

65

G. Akhyar et al. / Int. J. Sci. Eng. Tech. Vol. 1, No. 3, 2008, 60-66

Processing Technology (2004) 145: 84 92. [7] Liu, C.H., Andrian, C., Chen, C.A., Wang, Y.T. Grinding Force Control in Automatic Surface Finish System, Journal of Materials Processing Technology (2005) 170: 367373. [8] Mohan, N.S., Ramachndra, A., Kulkarni, S.M. Influence of Process

Parameters on Cutting Force and Torque During Drilling of Glass-Fiber Polyester reinforce Composite, Journal of Composite Structures (2005) 71: 407413. [9] Bhattacharyya. Metal cutting theory and practice, New Central Book Agency, Calcutta 1998.

66