Uludağ Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi, Cilt 14, Sayı 1, 2009 SURFACE ROUGHNESS EVALUATION WHEN MACHINING CARBON STEEL WITH CERAMIC CUTTING TOOLS Ali Riza MOTORCU* Abstract: The response surface methodology was adopted to investigate the effects of main cutting parameters such as cutting speed, feed rate, and depth of cut on the surface roughness when turning AISI 1050 carbon steel. Machining tests were carried out with uncoated ceramic (KY1615) and coated ceramic cutting tools (KY4400). Optimal machining conditions for the desired surface finish were determined. The adequacy of the second order developed model was analyzed by using analysis of variance. The experimental results indicated that the feed rate was the dominant factor, followed by the depth of cut. The cutting speed showed the minimal effect on the surface roughness. It could be seen that the KY1615 tool produced a better surface roughness than the KY4400 tool. It was shown that average surface roughness’ of Ra values were about 2.515 μm, 2.984 μm for the KY1615, KY4400 cutting tools, respectively. Furthermore, the analysis of variance for the second-order model indicated that squares terms were significant on the roughness, but interaction terms of cutting parameters were insignifi- cant for both cutting tools. Key Words: Turning, ceramic cutting tool, carbon steel, surface roughness, response surface methodology. Karbon Çeliğinin Seramik Kesici Takımlarla İşlenmesinde Yüzey Pürüzlülüğünün Değerlendirilmesi Özet: AISI 1050 karbon çeliğinin tornalanmasında yüzey pürüzlülüğü üzerinde ana kesme parametreleri kesme, ilerleme miktarı ve talaş derinliğinin etkilerini araştırmak amacıyla yanıt yüzey tekniği benimsenmiştir. İşlenebi- lirlik deneyleri kaplamasız seramik (KY1615) ve kaplamalı seramik (KY4400) takımlarla yapılmıştır. İstenilen yüzey pürüzlülüğü için optimum kesme şartları tanımlanmıştır. Deneysel sonuçlarda ilerleme miktarı en etkin faktör iken bunu talaş derinliği izlemiştir. KY1615 takımların KY4400 takımlardan daha iyi yüzey pürüzlülüğü sağladığı görülmüştür. Ra yüzey pürüzlülük değerlerinin ortalaması KY1615 takım için 2.525 μm iken KY4400 takımlar için 2.984 μm’dir. Ayrıca, ikinci dereceden modellerin varyans analizleri terimlerin karelerinin yüzey pürüzlülüğü üzerinde etkili olduğunu göstermiş fakat kesme parametrelerinin etkileşim terimleri her iki kesici takım içinde anlamsız etki yaratmıştır. Anahtar Kelimeler: Tornalama, seramik kesici takım, karbon çeliği, yüzey pürüzlülüğü, yanıt yüzey tekniği 1. INTRODUCTION Ceramic cutting tools are widely used in metal cutting industry for cutting of various hard ma- terials such as, alloy steel, bearing steel, white cast iron and graphite cast iron. The past few decades have witnessed great advancements in the development of these cutting tools. During machining, coated carbide/ceramic tools ensure higher wear resistance, lower heat generation and lower cutting forces, thus enabling them to perform better at higher cutting conditions than their uncoated counter- parts (Sahin and Motorcu, 2005). Surface roughness and dimensional accuracy have been important factors to predict machining performance of any machining operation (Arbizu and Perez, 2003). In material removal processes, however, an improper selection of cutting conditions will cause to obtain surfaces with high roughness and dimensional errors. Therefore, a proper estimation of surface rough- ness has been the focus study of number of researchers in the past three decades. In order to determine the optimal cutting conditions reliable mathematical models have to be formulated to associate the cutting parameters with cutting performance. In literature, Response Surface Method (RSM) has been used by some researchers on the analysis of surface roughness due to its practical, and relatively easy * Uludağ Üniversitesi, Teknik Bilimler Meslek Yüksek Okulu, 16059, Görükle, Bursa. 139 Motorcu, A. R.: Surface Roughness Evaluation When Machining Carbon Steel With Ceramic Cutting Tools for use (Feng and Wang, 2002; Hasegawa et al., 1976; Lambert, 1983; Mital and, Mehta, 1998;; Sata et al.,1985; Sundaram and Lambert, 1981; Petropoulos, 1974; Sahin and Motorcu., 2004a; Sahin and Motorcu, 2004b; Taraman, 1974). However, there were numbers of surface roughness prediction mod- els produced by ceramic cutting tools on steels and other hard materials in literature (Darwish, 2001; Davim, 2001; Davim 2007; Escalona and Cassier, 1998; Kơpac et al., 2002; Lee and Tarng, 2001; Lin et al., 2001; Sahin and Motorcu 2008; Suresh et al., 2002; Ozel and Karpat 2005; Yang et al., 1998). The most of the surface roughness prediction models given above are the Taguchi method or empirical while some of them are generally based on experiments in the laboratory (Choudhury and El-Baradie, 1997; Beauchamp, 1996; Chou and Song 2004; Grzesik and Wan 2006; Lima et al., 2005; Noordin et al., 2004; Motorcu, 2006). The aim of the present study was, thus, to develop a mathematical model for the surface roughness prediction using the response surface methodology based on the main cutting parameters such as cutting speed, feed rate, depth of cut. The machining tests were carried out on AISI 1050 car- bon steel with uncoated ceramic and coated ceramic cutting tools. Second-order predicting equation for the surface roughness was developed within ±5 % standard error. Furthermore, analysis of variance (ANOVA) was employed to investigate the cutting characteristics of steel bars. 2. EXPERIMENTAL WORK 2.1. Materials The machine used for turning tests was a Johnford TC35 Industrial type of CNC lathe ma- chine. The lathe equipped with variable spindle speed from 50 rpm to 3500 rpm, and a 10 KW motor drive was used for the tests. Cutting tools tested were mixed ceramic and coated ceramic tools. One of tools was a mixed ceramic with an Al2O3 (70%) +TiC (30%) matrix, which was designated by KY1615. The other inserts were coated using a Physical Vapour Deposition (PVD) method. Coating substance took place on the mixed ceramic substrate and TiN coated mixed ceramic with a matrix of Al2O3 (70%) +TiC (30%) +TiN, which was called as KY4400 grade. The insert type was TNGA 160408-KY1615 and TNGA 160408-KY4400. All tools were commercially available inserts accord- ing to ISO code. The cutting tools were supplied by Kennametal Inc. for the machining tests. The ma- terial used throughout this work was an AISI 1050 steel. The chemical analysis of AISI 1050 steel used in this study is presented in Table I. The steel bar stock was 40 mm diameter, 250 mm in length and these bars are machined under dry condition. The work material bars were trued, centered and cleaned by removing a 1 mm depth of cut from the outside surface, prior to the actual machining tests. The surface roughness of the carbon steel was measured by aid of a stylus instrument. The equipment used for measuring the surface roughness was a surface roughness tester, MAHR Perthometer-M1 type of portable. The surface roughness measured in the paper is the arithmetic mean deviation of the surface roughness of profile Ra. In collecting the surface roughness data of the shaft with the surface profilometer, three measurements are taken along the shaft axis for each sample. Table I. Chemical analysis of AISI 1050 steel Elements Wt.% Elements Wt.% Elements Wt.% C 0,47 Si 0,176 Mn 0,658 P 0,0144 S 0,0053 Cr 0,0540 Mo 0,0250 Ni 0,133 Al 0,0201 Co 0,0193 Cu 0,169 Nb <0,002 Ti <0,001 V <0,001 W <0,005 Pb <0,002 Sn <0,0048 Mg - Sb <0,002 Fe 98,24 140 Uludağ Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi, Cilt 14, Sayı 1, 2009 2.2. Experimental design To develop a second-order model, a design consisting of 18 experiments was conducted. De- tails of the model and design are given elsewhere (Sahin and Motorcu, 2004; Motorcu, 2006). 18 ex- periments constitute 23 factorial designs with an added center point repeated four times, the added center point being used to estimate pure error. An augment length of 2 was chosen depending on the capacity of the center lathe. The augments point consists of three levels for each of the independent variables denoted by -2, -1, 0, +1, +2. Table II shows levels of independent variables. Table II. Levels of independent variables Levels Lowest Low Center High Highest Coding number -2 -1 0 +1 +2 Cutting speed, V (m/min) 306 408 510 612 714 Feed rate, f (mm/rev) 0.145 0.20 0.255 0.310 0.365 Dept of cut, d (mm) 0.43 0.57 0.71 0.85 0.99 3. RESULTS AND DISCUSSION 3.1. Analysis of experiments The KY1615 and KY4400 cutting tools were used for these experimental results. Analysis of the influence of each independent variable on the surface roughness was performed with a Minitab computer package. Factorial design of experiments, cutting conditions and experimental results are given in Table III. Each coefficient was calculated and then formed the final linear regression Eq.(1) and (2), respectively. The positive value of the Ra from Eq.(1) shows an increase in roughness value while its negative value indicates a decrease in roughness value. The second-order model was postulated in obtaining the relationship between the surface roughness and the machining independent variables. Table III. Experimental results produced by the KY1615, KY4400 cutting tools and their theoretical values with errors for the model Experimental cutting Results Trial conditions KY1615 cutting tool KY 4400 cutting tool Number V f d Experimental Theoretical Average Experimental Theoretical Average Ra Ra-theo error,% Ra Ra-theo error,% 1 408 0.200 0.57 1.449 1.459 0.69 1.893 1.902 0.84 2 612 0.200 0.57 1.478 1.498 1.35 1.93 1.945 0.48 3 408 0.310 0.57 3.409 3.561 4.46 4.053 4.095 1.07 4 612 0.310 0.57 3.416 3.570 4.51 3.835 3.944 2.83 5 408 0.200 0.85 1.490 1.473 1.14 1.952 1.916 1.78 6 612 0.200 0.85 1.502 1.488 0.93 1.94 1.970 1.55 7 408 0.310 0.85 3.527 3.645 3.35 4.16 4.229 1.55 8 612 0.310 0.85 3.503 3.630 3.63 4.036 4.088 1.43 9 306 0.255 0.71 2.297 2.234 2.74 2.909 2.896 0.26 10 714 0.255 0.71 2.333 2.258 3.21 2.867 2.795 2.30 11 510 0.145 0.71 1.089 1.158 6.34 1.445 1.472 1.84 12 510 0.365 0.71 5.609 5.402 3.69 5.883 5.783 1.70 13 510 0.255 0.43 2.384 2.285 4.15 2.78 2.729 1.86 14 510 0.255 0.99 2.396 2.358 1.59 2.909 2.887 0.75 15 510 0.255 0.71 2.354 2.314 1.70 2.776 2.764 0.55 16 510 0.255 0.71 2.339 2.314 1.07 2.76 2.764 0.02 17 510 0.255 0.71 2.343 2.314 1.24 2.794 2.764 1.20 18 510 0.255 0.71 2.356 2.314 1.78 2.786 2.764 0.91 Absolute average error, % 2.64 Absolute average error, % 1.18 141 Motorcu, A. R.: Surface Roughness Evaluation When Machining Carbon Steel With Ceramic Cutting Tools The model equation for the surface roughness of the Ra prediction value for the KY1615 tool is derived from the data shown in Table III and given by; R 2a  2.185  0.002349 *V  22.407 * f  0.3729* d  0.0000001.V  79.8898* f 2 (1)  0.0978* d 2  0.0013*V * f  0.0004 *V .d  2.2727 * f * d Eq.(1) indicates that the feed rate has the most significant effect on the surface roughness of sample when using ceramic tools. The correlation coefficient R2 is very close to unity (99.2%), which shows the high correlation that existing between the experimental and predicted values. This shows that the second order model can be explained the variation to the extent of 99.2%. The model equation for the surface roughness of the Ra prediction for the KY4400 tool is giv- en by; Ra  2.9953 0.0007 *V 15.5138* f 1.7599* d  0.00000001*V 2  71.6198* f 2 (2)  0.6008* d 2  0.0082*V * f  0.0004*V * d  3.8799* f * d The model has an adjusted R2 value of 99.5 %, standard error for the surface roughness is about 0.157. Interactions of V*d and V*f in the range given in Eq. (2) had no significant effect on the surface roughness while the interaction of f*d had a slight effect on it. 3.2. Effect of too type The effect of cutting tool’s type on the surface roughness in machining the AISI 1050 steel was investigated, as shown in Fig.1. The surface roughness was produced by the KY1615, KY4400 cutting tools under different conditions. A comparison was made between these two types of cutting tools. It can be seen in Fig.1 that the KY1615 tool produced a better surface roughness value than that of the KY4400 tool for all experimental conditions. KY1615 tool is uncoated and it has sharper edge. 7,000 KY1615-Experimental 6,000 KY1615-Theoretical KY4400-Experimental KY4400-Theoretical 5,000 4,000 3,000 2,000 1,000 0,000 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Trial number Figure 1: Effect of cutting conditions on the average surface roughness of Ra produced by two ceramic cutting tools On the other hand, coating layer was removed from the substrate material when the KY4400 coated cutting tool was used. It is confirmed that the KY1615 tool was more appropriate than that of the KY4400 tool when machining the steel. For example, the average surface roughness’ of Ra values were about 2.515 μm, 2.984 μm for the KY1615, KY4400 cutting tools, respectively. This study 142 Surface roughness, Ra (m) Uludağ Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi, Cilt 14, Sayı 1, 2009 showed that it was appropriate when machining the carbon steel with ceramic based cutting tools un- der high speed conditions. But, these values are a little bit high. The experimental results and their theoretical values are shown in Fig.1 and Table III. As shown in this figure, some differences between the calculated and actual values of the surface rough- ness’ can be seen, but both values for the KY1615, KY4400 cutting tools are within a reasonable limit. For example, absolute average error was about 2.64% for the KY1615 cutting tool, but the average error was about 1.18% for the KY4400 cutting tool. Thus, the results indicate that the model con- structed using the multiple regression analysis can be used to provide accurate prediction of the sur- face roughness in machining the steels within ±5%. 3.3. Analysis of Variance (ANOVA) The ANOVA was used to investigate which design parameters significantly affected the cut- ting characteristics of the steels. Examination of the calculated values of variance ratio (F) which is the variance of the factor is divided by the error variance for all independent variables. The result of the analysis of variance for the second-order model is shown in Table IV. This analysis is carried out for a level of significance of 5% i.e., for a level of 95%. From the analysis of Table IV, it was apparent that the F calculated value was greater than the F table value (F 0.05, 9, 8=3.39) except for interaction ef- fects. Thus, the second order developed model was quite adequate. Furthermore, this table showed that quadratic terms were significant for this model, but the interaction terms of cutting parameters were not significant on the surface roughness of the Ra value produced by KY1615 cutting tool. Table IV shows the result of the analysis of variance for the second order model of the Ra equation in machin- ing the steel with the KY4400 cutting tool. The same table indicated that the quadratic effects were significant on the surface roughness for both equations, but the interaction terms were not significant on the surface roughness of the tested steels. Previous work carried out by Sahin and Motorcu (2005) showed that first order effect of feed rate and cutting speed was significant while depth of cut was insignificant. The ANOVA for the second order model exhibited that interaction terms and square terms were statistically significant while cutting speed and depth of cut was insignificant for turning AISI 1040 steel with PVD-coated ceramic tool (Sahin and Motorcu, 2004a, 2004b). However, interaction of feed rate and tool’s nose radius produced statistically significant impact on the surface roughness when machining AISI 4140 alloy steel with CVD-coated cutting tools (Sahin and Motorcu, 2004b). In their work, a better surface finish was obtained with these CVD-tools than those of the ceramic cutting tools used for the present work. This might be due to selecting lower cutting conditions and more appropriate for machining the steel with coated carbide tool in previous study. Although the higher cutting condition was used for machining the carbon steel by ceramic based cutting tool, higher surface roughness was obtained for current work due to related to chip curvature of the tested material. Higher cutting speed resulted in softening the steel. Therefore, plastic deformation ability of the steel increased and the chip curvature decreased due to high speed for machining the steel by ceramic tool. For the carbide tool, however, the steel did not very softened as much as ceramic tool since the cutting speed was lower. However, Ko- pac et al., (2002) found that the surface roughness decreased with an increase in cutting speed. 4. CONCLUSIONS The RSM was adopted to investigate the effects of machining factors such as cutting speed, feed rate, and depth of cut on the surface roughness when turning the AISI 1050 carbon steel. The feed rate was the dominant factor, followed by the depth of cut. The cutting speed showed the minimum effect on the surface roughness. Furthermore, it could be seen that the KY1615 cutting tool produced a better surface roughness than that of the KY4400 tool. The average surface roughnesses of Ra values were about 2.515 μm, 2.984 μm for the KY1615, KY4400 cutting tools, respectively. It was found that it was appropriate when machining the mild steel with ceramic based cutting tools under high speed conditions. Moreover, the ANOVA showed that the quadratic effects were significant on the rough- ness while the interactions terms of cutting parameters were statistically insignificant for both cutting tools. 143 Motorcu, A. R.: Surface Roughness Evaluation When Machining Carbon Steel With Ceramic Cutting Tools Table IV. Analysis of variance for machining the AISI 1050 steels by the KY1615 and KY4400 cutting tools Source Degree of free- Sum of squares Mean squares dom (DF) (SS) (MS) F statistic Contribution, % (P) K1615 cutting tool Regression 9 19.578 2.175 115.19 0.000 Linear 3 18.015 6.005 317.98 0.000 Squares 3 1.559 0.519 27.53 0.000 Interaction 3 0.003 0.001 0.06 0.981 Residual 8 0.151 0.018 Lack of fit 5 0.150 0.030 430.96 0.000 Pure error 3 0.0002 0.00007 Total 17 19.729 K4400 cutting tool Regression 9 19.777 2.1974 400.83 0.000 Linear 3 18.6221 6.20738 1000 0.000 Squares 3 1.1306 0.37687 68.74 0.000 Interaction 3 0.0242 0.00808 1.47 0.293 Residual 8 0.0439 0.00548 Lack of fit 5 0.0432 0.00864 40.26 0.006 Pure error 3 0.0006 0.00021 Total 17 19.820 ACKNOWLEDGEMENT This research has been carried out with financial support from University of Gazi in Turkey through, 07/2003-38 number coded which is gratefully acknowledged. REFERENCES 1. Abouelatta, O.B. and Mádl, J.,(2001) Surface roughness prediction based on cutting parameters and vibra- tions in turning operations, J. Mater. Proces. Technol., 118, 269-277. 2. Arbizu, I.P. and Perez, C.J.L., (2003) Surface roughness prediction by factorial design of experiments in turning processes, J. Mater. Process .Technol., 143-144, 390-396. 3. Beauchamp, Y., Thomas, M., Youssef, A. and Masounave, J.,(1996) Investigation of cutting parameter ef- fects on surface roughness in lathe boring operation by use of a full factorial design, Comp.& Ind. Engineer., 31(3-4), 645-651. 4. Choudhury, I.A. and El-Baradie, M.A., (1997) Surface roughness prediction in the turning of high-strength steel by factorial design of experiments, J.Mater.Proc.Technol., 67(1-3), 55-61. 5. Davim, J. P., (2001) A note on the determination of optimal cutting conditions for surface finish obtained in turning using design of experiments, J. Mater. Process. Technol., 116, 305-308. 6. Davim, J.P. and Figueira, L., (2006) Machinability evaluation in hard turning of cold tool steel (D2) with ceramic tools using statistical techniques, Materials & Design, 28(4), 1186-1191. 7. Darwish, S.M., (2001) The impact of the tool materials and the cutting parameters on surface roughness of supermet 718 nickel alloy, J. Mater. Process. Technol., 97, 10- 15. 8. Diniz, A.E. and Micaroni, R., (2002) Influence of refrigeration/lubrication condition on SAE 52100 hardened steel turning at several cutting speeds, Int. J. Mach. Tools& Manufact., 42, 899-904. 9. Escalona, P.M. and Cassier, Z., (1998) Influence of the critical cutting speed on the surface finish on turned steel, Wear, 218, 103-109. 10. Feng, C.X.J. and Wang, X., (2002) Development of empirical models for surface roughness prediction in finish turning, Int. J. Advan. Manuf. Technol., 20, 348-356. 144 Uludağ Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi, Cilt 14, Sayı 1, 2009 11. Kơpac, J., Bahor, M. and Sokovic, M.,(2002) Optimal machining parameters for achieving the desired sur- face roughness in fine turning cold pre-formed steel work processing, Int. J. Mach. Tools & Manuf., 42, 707- 716. 12. Lima, J.G., Avila, R.F., Abrao, A.M., Faustin, M. and Davim, J.P., (2005) Hard turning: AISI 4340 high strength low alloyed steel and AISI D2 cold work steel, J. Mater. Proc. Technol., 169, 388-395. 13. Lin, W.S., Lee, B.Y. and Wu, C.L., (2001) Modeling the surface roughness and cutting force for turning, J. Mater. Process. Technol., 108, 286-293. 14. Mital, A. and Mehta, M.,(1998) Surface finish prediction models for fine turning, Int. J. Prod. Res., 26-12, 1861-1876. 15. Montgomery, C. D., (1996) Design and Analysis of Experiments, John Willy and Sons, USA. 16. Motorcu, A.R., (2006), Machinability of AISI 1050, AISI 4140 and AISI E 52100 steels’ by different cutting tools and development of models, Ph.D. Thesis, Gazi University Institute of Science and Technology, Anka- ra. 17. Noordin, M.Y., Venkatesh, V.C., Sharif Elting, S. and Abdullah, A.,(2004) Application of response surface methodology in describing the performance of coated carbide tools when turning AISI 1045 steel, J. Mater. Process. Technol., 145(1), 46 -58. 18. Sahin, Y. and Motorcu, A. R., (2004) Prediction of surface roughness in machining of carbon steel by CVD coated cutting tools, Proc. 8th Intern. Conf. on Numerical Methods in Industrial Forming Processes, Uni- form 2004, Ohio State University, Mechanical Engineering Department, USA, 141-147. 19. Sahin, Y. and Motorcu, A.R, (2004) A model for surface roughness in turning AISI 4140 steel using coated carbide cutting tools, 11th Int. Conf. on Mach. Des. & Prod., Antalya/Turkey, 11-15. 20. Sahin, Y. and Motorcu, A. R., (2005) Surface roughness model for machining of mild steel by coated cutting tools, J. Mater. & Des., 26, 321-326. 21. Sata, T., Li, M., Takata, S., Hiraaka, H., Li, C. Q., Xing, X.Z and Xiao, X.G.,(1985) Analysis of surface roughness generation in turning operation and its applications, CIRP Annals. 341, 473-476. 22. Suresh, P.V.S, Venkateswara, R.P and Desmukh, S.G.,(2002) A genetic algorithmic approach for optimiza- tion of the surface roughness prediction model, Int. J. Mach. Tools & Manuf., 42, 675-680. 23. Yan, D., Popplewell, N., Balkrisnan, S. and Kaye, J.E.,(1996) On-line prediction of surface roughness in finish turning, Eng. Des. Auto., 2(2), 115-126. 24. Yang, W.H., Tarng, Y.S., (1998) Design optimization of cutting parameters for turning operations based on the Taguchi method, J. Mater. Process. Technol., 84, 123-129. 25. Yen, Y.C., Jain, A. and Altan, T., (2004) A finite-element analysis of orthogonal machining using different tool edge geometries, J. Mater. Process. Technol.,146, 72-81. Makale 15.01.2009 tarihinde alınmış, 06.07.2009 tarihinde düzeltilmiş, 07.07.2009 tarihinde kabul edilmiştir. İletişim Yazarı: A. R. Motorcu (armotorcu@uludag.edu.tr). 145