Effect of Tool Material Properties and Cutting ... - Springer Link

10 downloads 26820 Views 4MB Size Report
Feb 6, 2015 - ability of AISI D6 tool steel using three different grades of low-content cubic boron ...... An investigation of worn tools using SEM and. EDAX showed same .... CBN-3 is the best tool due to high thermal conduc- tivity and high ...
Arab J Sci Eng (2015) 40:1151–1164 DOI 10.1007/s13369-015-1578-0

RESEARCH ARTICLE - MECHANICAL ENGINEERING

Effect of Tool Material Properties and Cutting Conditions on Machinability of AISI D6 Steel During Hard Turning Manoj Nayak · Rakesh Sehgal

Received: 16 August 2014 / Accepted: 19 January 2015 / Published online: 6 February 2015 © King Fahd University of Petroleum and Minerals 2015

Abstract Hard turning offers numerous advantages to grinding operation; however, there is a critical need for research to clarify issues related to high cutting forces, high temperatures, and surface roughness to meet the challenges it can offer as an alternate to grinding process. Mathematical models are generated for each response variable (main cutting force, thrust force, cutting temperature, and surface roughness) in terms of actual values of the factors (cutting speed, feed, and tool material) to establish relationships using design expert software for statistical investigation. A 33 full-factorial design with a total of 27 experiments was obtained for parametric analysis and investigation of machinability of AISI D6 tool steel using three different grades of low-content cubic boron nitride (CBN-L) inserts. The parametric analysis study shows that the main cutting force, thrust force, and surface roughness increase with feed. The thrust force and cutting temperature get influenced by tool material properties. It was established that grain size, CBN content, edge geometry, and hardness of the tools affected all the output characteristics. Scanned electron microscopy and energydispersive X-ray of the worn tools showed crater wear, chipping, and fracture of cutting edges, while abrasion and diffusion/dissolutions in CBN tools were the wear mechanisms affirmed in this study. M. Nayak National Institute of Technology, Hamirpur 177005, India M. Nayak (B) Department of Mechanical Engineering, FET, MRIU, Faridabad 121006, India e-mail: [email protected] R. Sehgal Department of Mechanical Engineering, National Institute of Technology, Hamirpur 177005, India e-mail: [email protected]

Keywords Hard turning · Cubic boron nitride · Thrust force · Surface roughness

1 Introduction Cubic boron nitride (CBN) has the highest hardness and thermal conductivity among all materials except diamond, popularly noted for its ability to perform efficient machining and offers low reactivity to ferrous material. In response to the increasing concern for the reduction in grinding sludge, there is paradigm shift from grinding to hard turning. But several issues need to be addressed such as typical failures occurring at the cutting edges during machining of hard materials. Hard materials are considered as difficult-to-machine materials mainly due to high chip–tool interface temperatures and cutting forces generated during the cutting process. Performance of several cutting tools like ceramics and carbides has been investigated for machinability characteristics of hard materials in previous studies [1–3]. Hard turning as a finishing process requires the generation of the machined surface by pure plastic deformation. Therefore, not only the stress, strain, and temperature distribution in the cutting zone are of interest, but the study of the influence of tool material properties on the cutting force and surface roughness is also significant. With the help of developed tool materials like ceramics and CBN, hard turning has become a well-established and profitable alternative to finish grinding operation [4]. To fill this demand, various CBN grades are being developed as they provide a reliable tool life. Hard turning process differs from conventional turning due to relatively low depth of cut, low feeds [5]. Moreover, due to high brittleness of CBN tools along with high strength and high hardness of the work material, these inserts are normally provided with a defined edge geometry. They provide

123

1152

Arab J Sci Eng (2015) 40:1151–1164

strength to the tool nozzle and also prevent chipping and improve wear resistance. The chamfer provides a large negative rake angle resulting in large plastic strains and hence higher cutting temperature, which adversely affects the tool’s performance and the hone exerts high cutting forces. A proper selection of tool geometry along with the cutting conditions for better surface quality of hard materials is suggested [6,7] Tool steels used for hot forging dies, press work, extrusion, and form tools work under extreme condition in a tribological systems. These steels are considered as difficultto-machine steels, and more particularly, it has been reported that the machinability of tool steel is extremely poor. During machining, these steels generate high chip–tool interface temperatures and high cutting forces [8]. The thrust force is highest in hard turning because of low feed rate and low depth of cut, which is much smaller than the nose radius (0.8 mm). Such a large value of thrust force becomes a serious threat to the dimensional accuracies of the component due to greater radial deflection of the workpiece, and it also causes chattering if the dynamic loop stiffness of the machining system is low [9]. Surface integrity such as surface roughness plays an important role in the functionality of the machined component. Surface roughness is worse for high CBN content tool than the low CBN content tools. The reason is plucking out of CBN grains from the sintered tool body by the hard carbide particles of the workpiece, which then produce abrasive effect on the workpiece [10]. Cutting forces data are essential while selecting work material and tool material. It is also required while selecting an appropriate machine tool with adequate power. To avoid excessive distortion of the workpiece and rapid wear of cutting tool and fixtures, the knowledge of cutting forces is must [11]. A study by Huang et al. [12] proposed a cutting force model to consider the tool thermal properties. The model was applied to hard turning of AISI H13 tool steel using CBN tools of varying CBN content, i.e., CBN-L and CBN-H inserts. The model predicted low major cutting forces and low thrust forces, but high chip–tool interface temperatures for CBN-L tools, while high major cutting force and high thrust forces with low chip–tool interface temperatures were predicted for CBN-H tools. This clearly suggests that the cutting forces and the chip–tool temperature are greatly influenced by CBN content, thermal, physical, and mechanical properties of the CBN tools. Tool wear also affects the cutting zone temperature and generates residual stresses on the machined surface [13]. The difference in behavior by the

Table 1 Chemical composition of AISI D6 tool steel

123

different grades of CBN inserts is due to the transformation of boron nitride from hexagonal to cubic form, grain size, shape, porosity, defects, inclusions, and binder material [14]. CBN tools are classified as high CBN content tools (CBNH) constituting more than 70 % CBN with a metallic binder like cobalt, medium CBN tools with 50–70 % CBN content with tin as a binder, and low CBN tools (CBN-L) with less than 50 % CBN content along with a ceramic binder like TiN [15]. RSM-based mathematical models have been proposed for predicting the output characteristic along with parametric analysis on machinability investigation of AISI D2 cold work tool steel [16]. Optimization techniques such as RSM and firefly algorithm [17], genetic algorithm [18], and other statistical investigations [19] for machining study have been reported in the literature. In the present work, an attempt has been made to develop mathematical models based on statistical methodology for each tool grade and to study the effect of cutting parameters and tool material influences on machinability characteristics of AISI D6 steel during hard turning. 2 Experimental Details 2.1 Work Material The work material used is AISI D6 tool steel in hardened condition (54±2HRC). The chemical composition of the materials is given in Table 1. The microstructure observed in the hardened steel is of spheroidal carbides in tempered martensite as shown in Fig. 1, and dispersion of small and large particles of metal carbides in the ferrite matrix is clearly noted. 2.2 Selection of Tool Material CBN compact is classified into two categories: In CBN-L, the compact has the CBN particles bound on contact with a ceramic binder material (TiN, TiC, TiAlN, TiCN), and CBNH compact is formed by binding together the CBN particles using a small amount of metallic binder (Co, Sn). For industrial applications, the former type tools are recommended for machining of hardened steel due to their excellent wear resistance. The later is used for machining cast iron, heat-resistant alloys, and powder metallurgy parts due to their excellent heat resistance and toughness [20]. Most of the researchers have carried out their study using CBN-H tools for machining hardened steel, which is uneconomical in real applications.

Elements

C

Si

Mn

Cr

Ni

Mo

V

Cu

Al

Co

W

Weight %

2.130

0.238

0.240

11.10

0.090

0.032

0.130

0.03

0.06

0.025

0.620

Arab J Sci Eng (2015) 40:1151–1164

1153

Fig. 1 Dispersion of small and large particles of carbides in the ferrite matrix

Fig. 2 Machining setup

The three tools used in the present work are CBN-L types and are designated as CBN-1, CBN-2, and CBN-3 based on tool substrate property, thermal, and mechanical properties as given in Table 2. The nose radius of all the tools was 0.8 mm, and an approach angle of 45o was kept throughout the machining.

possible experiment errors. Force measuring dynamometer (make TeLC Germany) was used to measure the main cutting force (Fc ) and the thrust force (Fp ). Cutting temperature of the tool tip was measured with the help of InGaAs radiation sensor (Impact electronic series 300, 24VDC, 4– 20 mA; measuring range 300–800 ◦ C). Software XKM 2000 was used in the dynamometer for cutting force and temperature measurements. A surface roughness tester (SJ-301 Mitotoyo, Japan), X -axis (drive units) with measuring range of 12.5 µm, was used to measure the surface roughness Ra during the experiments. Surface roughness was measured off-line with the profilometer by taking the measurements across the lay. Three measurements (λc = 0.8 mm; N = 5) were taken along the feed length for each sample length machined, and measurements were taken about 120◦ apart; the average Ra value was used in the analysis. Workpiece of AISI D6 tool steel of Ø54.5 and Ø59.5 mm was prepared after machining 1 mm thickness of the top cylindrical sur-

2.3 Experimental Procedure A three-factor three-level full-factorial design was used to determine the effects of tool material properties and cutting conditions on machinability of AISI D6 tool steel. Based on 33 full-factorial designs, a total of 27 experiments were carried out on an all geared DRO Lathe (Model: Bajaj-Pioneer-175 Geared Headed), 8-spindle cutting speeds (8–1,200 rpm), and 24 number of feeds as shown in Fig. 2. Randomization is strictly maintained as per the run order suggested by the DOE, and 14 experiments were repeated to minimize the Table 2 Thermal and mechanical properties of tool materials [9,13]

Tool type

CBN-1

CBN-2

CBN-3

Make

Kyocera

SumiBoron

SumiBoron

Grade

KBN 25M

BNX 10

BNX 20

Binder

TiC

TiCN

TiN

Grain size (µm)

0.5–1

3

3

CBN (%)

45

40–45

55–60

Base material hardness (HV)

2,500

2,800–3,000

3,200–3,400

Transverse rupture strength (Gpa)

1,250

800–900

1,000–1,100

Thermal conductivity (Watt/mK)

44

Unknown

54

Edge geometry

25◦ × 0.12 mm

30◦ × 0.1 mm

30◦ × 0.1 mm

(Chamfered + Honed)

(Chamfered)

(Chamfered)

123

1154

Arab J Sci Eng (2015) 40:1151–1164

Table 3 Factors and factor levels Symbol

Factor

Units

Level-1

Level-2

Level-3

Vc

Cutting speed

m/min

54.03

93.62

132.92

f

Feed

mm/rev

0.08

0.133

0.21

CBN

CBN Grades

Categorical

CBN-1

CBN-2

CBN-3

face, in order to eliminate any surface defects and wobbling, then centered, and faced. Corresponding cutting speeds were equalized at different workpiece diameters to the extent that rotational speeds ratio permit. Short-duration tests were performed (machined length of 20 mm) without coolant. The test conditions in Table 3 span the range of recommended values from tool supplier and are comparable to cutting parameters in previous studies of hard turning. Three CBN grades from two different manufacturers were used, so that a particular tool manufacturer is not endorsed. This cutting condition represents machining at low and medium speeds to ensure that

the CBN tool wear is not rapid and to obtain surface roughness value close to those obtained in grinding. Moreover, constant depth of cut 0.15 mm was used as a usual value of material stock removal in grinding operation. Each test was realized with fresh cutting edge. The obtained experimental data are summarized in Table 4.

3 Results and Discussions Analysis of the results of these trials was done using standard statistical methodology [21] using Design Expert Software (version 7) as per the full-factorial DOE model as shown in Table 5. All responses were analyzed using 2FI polynomial design model with ANOVA standard deviation. Graphical analyses are explained with the help of 3D response surface graphs obtained during the regression analysis. The statistical methodology consists of first assessing statistical normality of the data and, wherever necessary, applying appro-

Table 4 Experimental data along with standard order and run as per DOE matrix Stand. order

Run order

CBN grades

Cutting speed (m/min)

Feed (mm/rev)

Main cutting force (Fc )

Thrust force (Fp )

Cutting temperature (T)

Surface roughness (Ra )

1

1

CBN-1

54.03

0.080

67

184

482

0.47

2

10

CBN-1

54.03

0.133

72

187

484

0.71

3

26

CBN-1

54.03

0.210

88

148

479

1.75

4

2

CBN-1

93.62

0.080

68

201

551

0.79

5

13

CBN-1

93.62

0.133

86

223

549

0.81

6

14

CBN-1

93.62

0.210

93

183

534

1.83

7

4

CBN-1

132.92

0.080

45

172

511

0.56

8

20

CBN-1

132.92

0.133

67

192

515

0.77

9

24

CBN-1

132.92

0.210

83

128

490

1.81

10

19

CBN-2

54.03

0.080

64

158

458

0.58

11

23

CBN-2

54.03

0.133

75

156

483

0.83

12

12

CBN-2

54.03

0.210

92

177

481

1.65

13

25

CBN-2

93.62

0.080

62

178

537

0.53

14

18

CBN-2

93.62

0.133

79

178

554

0.93

15

17

CBN-2

93.62

0.210

99

198

536

1.72

16

3

CBN-2

132.92

0.080

57

160

515

0.48

17

9

CBN-2

132.92

0.133

66

158

509

0.89

18

16

CBN-2

132.92

0.210

88

165

504

1.66

19

8

CBN-3

54.03

0.080

62

165

429

0.54

20

11

CBN-3

54.03

0.133

90

215

478

0.92

21

5

CBN-3

54.03

0.210

97

196

463

2.14

22

15

CBN-3

93.62

0.080

71

195

530

0.57

23

27

CBN-3

93.62

0.133

82

244

535

0.92

24

22

CBN-3

93.62

0.210

98

205

498

2.39

25

7

CBN-3

132.92

0.080

69

188

515

0.54

26

6

CBN-3

132.92

0.133

76

190

512

0.84

27

21

CBN-3

132.92

0.210

85

195

524

2.53

123

Arab J Sci Eng (2015) 40:1151–1164

1155

Table 5 Full-factorial DOE model Study type

Factorial

Experiment

27

Initial design

Full factorial

Blocks

No blocks

Design model

2FI-analyzed as a polynomial using ANOVA standard deviation

Response

Name

Units

Obsr.

Min.

Max.

Model

Fc

Major cutting force

N

27

45

99

RQuadratic

Fp

Thrust force

N

27

128

244

RQuadratic

T

Temperatures

◦C

27

429

554

RQuadratic

Ra

Surface roughness

µm

27

0.47

2.53

RQuadratic

Table 6 ANOVA table (partial sum of squares) for reduced quadratic model (response Fc ) Source Model A-Cutting speed

Sum of squares

DF

Mean squares

4,496.99

5

899.40

280.06

1

280.06

3,659.95

1

3,659.95

C-CBN grades

229.41

2

114.70

A2

327.57

1

13.63

Residuals

504.87

21

24.04

Cor Total

5,001.85

26

B-Feed

SD Mean C.V.% PRESS

4.90

F value

Prob > F

37.41

< 0.0001

11.65

0.0026

6.9

152.24

0.0001

90.26

0.0196

2.83

4.77

R-squared

0.8991

Adj R-squared

0.8750

6.36

Pred R-squared

0.8342

829.13

Adeq Precision

20.105

3.1 Major Cutting Force Models Statistical analysis of Fc was made with response surface reduced (backward transformation) quadratic regression model

Remarks Significant

0.0014

77.07

priate mathematical transformations to ensure normality of that data. These normalized data were used to produce a series of “best fit” regression equations. The ANOVA table of each output characteristic contains the usual sum of squares, degree of freedom, mean squares, and the test statistics “F.” Values of “Prob > F” less than 0.050 indicate that the model terms are significant. In addition to the basic analysis of variance, the diagnostic checking is done by the software, and the program displays some useful information like “R-squared (R 2 ), adjusted R-squared (Adj R 2 ), prediction R-squared (Pred. R 2 ), adequate precision (Adeq. R 2 ), standard deviation (SD), mean, degree of freedom (DF) percentage contribution (% Contr.), coefficient of variation (C.V.%), and prediction error sum of squares (PRESS) values. The mathematical regression equations can be used to predict the main cutting force, thrust force, cutting temperature, and surface roughness during hard turning of AISI D6 steel using CBN inserts.

% Cont.

by selecting the backward elimination procedure to automatically reduce the insignificant terms. The resulting ANOVA table for the reduced quadratic model for Fc is shown in Table 6. The quadratic model with P value > 0.05 suggests that the model is significant. Various R 2 statistics of Fc are given in Table 6, R 2 = 0.8991 indicates that 89.91 % of the total variations are explained in the model. Adjusted R 2 = 0.8750 indicates that 87.50 % of the total variability is explained by the model after considering the significant factors. The predicted R 2 = 0.8342 is in reasonable agreement with the adjusted R 2 of 0.8750, and the model would be expected to explain 83.42 % of the variability in new data. Lower value of C.V. = 6.36 % indicates improved precision and reliability of the conducted experiments. The P value indicates the probability that the factor or interaction is significant or insignificant. The result indicates that A, B, C, A2 are significant model terms. Examination of the P value shows the feed to be dominant factor with 90.26 % contribution to Fc followed by cutting speed and CBN grades. The mathematical models in the form of regression equations to predict the main cutting force (Fc ) with cutting speed (Vc ), feed rate (f) for three different tools are given below:

123

1156

Arab J Sci Eng (2015) 40:1151–1164

Fig. 5 3D surface plots of Fc for CBN-3

Fig. 3 3D surface plots of Fc for CBN-1

Fig. 4 3D surface plots of Fc for CBN-2

FcCBN−1 = 16.354 + 0.788Vc + 218.14 f −4.749 × 10−3 Vc2

(1)

FcCBN−2 = 17.81 + 0.788Vc + 218.14 f −4.749 × 10−3 Vc2

(2)

FcCBN−3 = 23.132 + 0.788Vc + 218.14 f −4.749 × 10−3 Vc2

(3)

Equations (1), (2), and (3) are valid for finish hard turning in continuous cutting of AISI D6 tool steel using CBN-1, CBN-2, CBN-3, respectively, under the conditions 54.03 ≤ Vc ≤ 132.92 m/min; 0.08 ≤ f ≤ 0.21 mm/rev; and constant depth of cut = 0.15 mm. From the ANOVA table of Fc , it is observed that the feed has the most significant effect on it, which is in agreement with the study presented by Thiele [22]; Quin and Hosson [23]. The regression equations of Fc suggest that Fc CBN−3 > Fc CBN−2 > Fc CBN−1 , clearly indicating the influences of tool material properties. In Figs. 3, 4, 5, the change in major cutting force can be studied de-

123

pending on feed rate and cutting speed for different CBN tools. It can be seen that the major cutting force increases in parallel with the increase in feed rate compared to a curvilinear profile (quadratic effect) with the increase in cutting speed (Vc2 ). The curvilinear profile in the figures is in accordance with the quadratic model fitted. The quadratic effect of the model term Vc2 can be explained for the single bend of the curve, i.e., the curvilinear profile, and is concave due to negative signs of the coefficients (Eqs. 1–3). This can be explained in a simple way; as the cutting speed increases, the friction between the workpiece and the tool cutting edge increases, and hence, the temperature at the cutting zone increases. This induces a thermal softening effect, and workpiece gets softened, which then requires less cutting force. The other reason is at a high cutting speed, situation improves the formation and removal of chips at the cutting region. Also, as the feed rate increases, the undeformed chip thickness increases, and since the cutting force is directly proportional to undeformed chip thickness, more forces is required for chip formation; therefore, the feed rate becomes a major contributory to the Fc component of the cutting forces in accordance with a previous study [22]. Studies also suggest that there is no appreciable change in magnitude of the main cutting force and feed force (axial force) during hard turning, when the cutting speed was increased from 140 to 240 m/min [23,24]. Therefore, to minimize Fc , it is necessary to select low feed and high cutting speed. The other observation that can be made from these graphs is that higher Fc is encountered by CBN-3, possibly because of softening effect. Since the hardness of CBN-1 < CBN-2 < CBN-3, the hardness differences between the workpiece and the tool are less in case of CBN-3, thereby requiring more of this force than the other two tools CBN-1 and CBN-2. Therefore, tool with low CBN content along with low hardness should be selected for minimizing Fc .

Arab J Sci Eng (2015) 40:1151–1164

1157

Table 7 ANOVA table (partial sum of squares) for reduced quadratic model (response: Fp ) Source

Sum of squares

DF

Mean squares

F value

Prob > F

% Contr.

12,429.55

8

1,553.69

8.77

< 0.0001

A-Cutting speed

80.22

1

80.22

0.45

0.5096

0.875

B-Feed

31.24

1

31.24

0.18

0.6796

0.35

C-CBN grades

4,035.19

2

2,017.59

11.38

0.0006

22.12

BC

2,558.88

2

1,279.44

7.22

0.0050

13.99

A2

4,195.85

1

4,195.85

23.67

0.0001

46.017

B2

1,528.17

1

1,528.17

8.62

0.0038

16.75

Residuals

3,190.30

18

Cor total

15,619.85

26

Model

SD Mean C.V.% PRESS

13.31 182.93

Remarks Significant

177.24

R-squared

0.7958

Adj R-squared

0.7050

7.28

Pred R-squared

0.5546

6,956.79

Adeq Precision

10.586

3.2 Thrust Force Models Statistical analysis of Fp was made with response to surface reduced (backward transformation) quadratic regression model. Proceeding with the backward elimination procedure to remove the insignificant terms, the resulting ANOVA table for the reduced quadratic model for thrust force is shown in Table 7. The quadratic model with P value > 0.05 suggests that the model is significant. The predicted R 2 value is 0.5546 and is in reasonable agreement with the adjusted R 2 value of 0.7050. Both predicted R 2 and adjusted R 2 are within the range of 0.20 of each other as obtained in this case. The adequate precision is 10.586 as desired and indicates an adequate model discrimination. This justifies the accuracy of the predicted model. The result indicates in this case that C, BC, A2 , and B2 are all significant model terms. The quadratic effect of cutting speed is the most significant term with 46.017 % contribution. The effect of CBN grades alone is a significant factor with 22.12 % and at the same time, it is interesting to note that the interaction of feed with CBN grades is also significant. It implies that feed as a secondary factor affects the thrust force. Cutting speed and feed as a lone factor has little influence on the thrust force. It is reported that feed alone has no influence on thrust force. The equations for predicting thrust force using three different tools CBN-1, CBN-2, CBN-3 are as follows: Fp CBN−1 = 19.535 + 3.124Vc + 877.732 f − 0.017Vc2 −3,932.755 f 2 Fp CBN−2 = −46.51 + 3.124Vc + 1,275.190 f −3,932.755 f 2

(4) − 0.017Vc2 (5)

Fp CBN−3 = −13.960 + 3.124 Vc + 1,253.190 f − 0.017Vc2 −3,932.755 f 2

(6)

Fig. 6 3D surface plots of FP for CBN-1

In conventional turning of soft annealed steel, the major cutting force is highest among the three cutting forces (the other two being thrust force and feed force). This is due to the fact that turning is done with relatively high depth of cut and larger nose radius. However, the thrust force is highest in hard turning because of small nose radius and feed rate. The cutting action in hard turning occurs on a very small area of the tool nose often at the chamfer resulting in high thrust force. Response surface graphs show the evolution of the thrust force with increase in cutting speed and feed rates for all the three CBN tools as shown in Figs. 6, 7, 8. It can be seen from the graphs that the thrust force follows a curvilinear profile with the increase in cutting speed as well as the feed. Again, the curvilinear profile along the feed and cutting speed is in accordance with the quadratic model fitted. At low speed and low feed, the thrust force in case of CBN-2 and CBN-3 is low but increases as the feed and cutting speed

123

1158

Arab J Sci Eng (2015) 40:1151–1164

increase and then further decrease with further increase in feed and cutting speed. The thrust force on CBN-1 is high at low feed because of the presence of the chamfer and hone but decreases as the feed rate increases. This finding is consistent with the previous research for inserts with chamfer angles and hones [25,26]. The large negative effective angle formed due to the chamfer results in high thrust force initially and further decreases due to thermal softening. The initial thrust force is highest for CBN-1 (KBN 25M) tool because of geometry effect, i.e., availability of the chamfered and honed edge, whereas CBN-2 and CBN-3 had less initial thrust force due to non-availability of the hone at the cutting edge.

3.3 Cutting Temperature Models

Fig. 7 3D surface plots of FP for CBN-2

Statistical analysis of “T” was made with response surface reduced (backward transformation) quadratic regression model. Table 8 shows the ANOVA table with model F value of 16.84, which implies that the model is significant. In this case, A and A2 are significant model terms. R 2 = 0.8821 indicates that 88.21 % of the total variations are explained in the model. Adjusted R 2 = 0.8297 indicates that 82.97 % of the total variability is explained by the model after considering the significant factors. The predicted R 2 = 0.7282 is in reasonable agreement with the adjusted R 2 of 0.8297, and the model would be expected to explain 72.82 % of the variability in new data. Lower value of C.V. = 2.53 % indicates the improved precision and reliability of the conducted experiment. After removing the insignificant terms, the equations for predicting the cutting temperature (T) with cutting speed (Vc ), feed rate (f) as the process parameters are obtained as follows:

Fig. 8 3D surface plots of FP for CBN-3

Table 8 ANOVA table (partial sum of squares) for reduced quadratic model (response: T) Source

Sum of squares

Model A-Cutting speed B-Feed

DF

Mean squares

F value

Prob > F

% Contr.

22,060.56

8

2,757.57

16.84

TCBN−2 > TCBN−3 . The variation in temperatures is due to variation in thermal conductivity of the tools. The tool material CBN grades have very little influence on the cutting temperature with 1.86 % contribution as per the ANOVA. It can be seen from the response surface graphs, i.e., Figs. 9, 10, 11, that increase in temperature occurs due to increase in cutting speed for all the three tools, and increase in feed has little influence on the cutting temperature. This may be due to increase in friction at higher cutting speeds, which then induces a temperature rise in the cutting zone. At higher cutting speed situations, the formation and removal of chip improves in the cutting region, thereby reducing the cutting temperature at the cutting zone. The cutting temperature generated also depends on the thermal conductivity of the tools. Since there is a slight difference in thermal conductivity of the three tools owing to their CBN content, the slight variation observed in the response surface graph Fig. 11 is obvious. Higher the CBN content, higher is the thermal conductivity, and lower the CBN content higher is the initial cutting temperature of the cutting zone [24]. The amount of heat transferred to the workpiece by all the three tools depends on the thermal conductivity of tool material. The thermal

Fig. 11 3D surface plots of “T” for CBN-3

conductivity of CBN-1 is low as compared to CBN-2 and CBN-3, and therefore, the amount of heat transferred to the workpiece by CBN-1 is high and the material gets softened due to high heat content of the workpiece. The main requirement of a cutting tool for machining hard materials is thermal conductivity and hardness [27]. Therefore, CBN-3 is the best tool in terms of thermal conductivity and hardness, and the sharpness of the cutting edge is retained for a longer time. From the response surface analysis Figs. 9, 10, 11, the lowest cutting temperatures are observed during machining using CBN-3, which agrees well with the model equations. ANOVA indicates that the CBN grades are not significant nor do have any secondary effect on temperature within the cutting speed range of 54.03 ≤ V ≤ 132.92 m/min. But at higher cutting speeds, the effect will be prominent. 3.4 Surface Roughness Models Statistical analysis of “Ra ” was made with response to surface reduced (backward transformation) quadratic regres-

123

1160

Arab J Sci Eng (2015) 40:1151–1164

Table 9 ANOVA table (partial sum of squares) for reduced quadratic model (response Ra ) Source

Sum of squares

Model

10.41

DF

Prob > F

Mean squares

F value

% Contr.

120.62