Accepted Manuscript

Title: Comparison of tool life between ceramic and cubicboron nitride (CBN) cutting tools when machining hardenedsteels

Author: Y. Sahin

PII: S0924-0136(08)00620-1DOI: doi:10.1016/j.jmatprotec.2008.08.016Reference: PROTEC 12315

To appear in: Journal of Materials Processing Technology

Received date: 20-7-2007Revised date: 6-8-2008Accepted date: 12-8-2008

Please cite this article as: Sahin, Y., Comparison of tool life between ceramic and cubicboron nitride (CBN) cutting tools when machining hardened steels, Journal of MaterialsProcessing Technology (2007), doi:10.1016/j.jmatprotec.2008.08.016

This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.

Page 1 of 38

Accep

ted

Man

uscr

ipt

1

Comparison of tool life between ceramic and cubic boron

nitride (CBN) cutting tools when machining hardened steels

Y. Sahin, ysahin@gazi.edu.tr, Department of Mechanical Education, Faculty of Technical

Education, Gazi University, 06500-Besevler/Ankara/Turkey

Abstract

This paper describes a comparison of tool life between ceramics and CBN cutting tools when

machining hardened bearing steels using the Taguchi method. An orthogonal design, signal-to-

noise ratio (S/N) and analysis of variance (ANOVA) were employed to determine the effective

cutting parameters on the tool life. First order linear and exponential models were carried out to

find out the correlation between cutting time and independent variables. Second order regression

model was also extended from the first order model when considering the effect of cutting speed

(V), feed rate (f), hardness of cutting tool (TH) and two-way of interactions amongst V, f, TH

variables. The results indicated that the V was found to be a dominant factor on the tool life,

followed by the TH, lastly the f. The CBN cutting tool showed the best performance than that of

ceramic based cutting tool. In addition, optimal testing parameter for cutting times was determined.

The confirmation of experiment was conducted to verify the optimal testing parameter.

Furthermore, the second order regression model and exponential model supported the first order

model regarding the prediction capability. Improvements of the S/N ratio from initial testing

parameters to optimal cutting parameters or prediction capability depended on the S/N ratio and

ANOVA results. Moreover, the ANOVA indicated that the cutting speed was significant but other

parameters were an insignificant effect on the tool life at 90% confidence level. The percentage

contribution of the cutting speed was about 41.63 on the tool life when machining the hardened

steel.

Keywords: Turning; Cutting tool; Tool life; Optimal parameter; Orthogonal design; Hardness

* Manuscript

Page 2 of 38

Accep

ted

Man

uscr

ipt

2

NOMENCLATURE

AISI american iron and steel institute

ANOVA analysis of variance

CBN cubic boron nitride cutting tool

CNC computer controlled lathe machine

CVD chemical vapor deposition process

d depth of cut (mm)

dB decibel

DF degree of freedom

F statistical characteristic

f feed rate (mm/rev)

HRC hardness value by Rockwell

HV hardness value by Vickers

KY1615 mixed alumina ceramic cutting tool

KY4400 coated alumina ceramic cutting tool

MS mean of squares

P contribution (%)

PVD physical vapor deposition process

Ra average surface roughness value (µm)

R2adj% adjusted coefficient of multiple correlation

SS sum of squares

S/N ratio signal-to-noise ratio (dB)

T tool life (min)

TH tool hardness (HV)

V cutting speed (m/min)

WC tungsten carbide cutting tool

ηi average S/N ratio (dB)

ηi,cal calculation test result for tool life and S/N ratio (dB)

ηi,ver verification test result for tool life and S/N ratio (dB)

Page 3 of 38

Accep

ted

Man

uscr

ipt

3

1.0 Introduction

Hardened steels are machined by grinding process in general, but grinding operations are time

consuming and are limited to the range of geometries to be produced (Poulachon et al., 2004). The

hardened steel surfaces have an abrasive effect on the tool material, and the high temperature on

the cutting edge causes diffusion between tool and chip. Therefore, improved technological

processes, optimum tool selection, determination of optimum cutting parameters or tool geometry

should be considered. The developments of new cutting tools have led to the use of higher cutting

speeds compare with conventional machining. High speed cutting reduces machining costs by

increasing production rate. However, high speed cutting leads to the rapid wear of cutting tools,

which is caused by the high temperatures generated at the cutting zone and as a result tool life

decreases (Diniz and Oliveira, 2008). The ability of polycrystalline cubic boron nitride (CBN)

cutting tools to maintain a workable cutting edge at elevated temperature is, to same extent, shared

with several conventional ceramic tools. These tools are characterised by high hot hardness, wear

resistance and good chemical stability and low fracture toughness (Benga and Abrao, 2003). CBN

and ceramic tools are used in the manufacturing industry for hard turning because of its inertness

with ferrous materials and its high hardness. Though CBN particles and binder phases such as TiN

are harder than carbides in steels, it is still possible that the tool will encounter “soft” abrasive

wear. The machining of hardened bearing steel represents groving proportion of applications

involving hard cutting tools such as CBN and ceramics (Lima et al., 2005).

Various studies have been conducted to investigate the performance of CBN and ceramic tools

when machining hard steels or hardened steels. Chou and Song (2004) studied tool nose radius

effects on finish turning hardened AISI 52100 steels. The results showed that large tool nose radii

only gave finer surface finish, but comparable tool wear compared to small nose radius tools.

Grzesik and Wanat (2006) investigated the surface finish generated in hard turning of quenched

alloy steel (60 HRC) using conventional and wiper ceramic inserts. They determined that surfaces

produced by wiper tools contained blunt peaks with distinctly smaller slopes resulting in better

bearing properties. Diniz and Oliveira (2008) investigated the turning of AISI 4340 steel (56 HRC)

interrupted surfaces using three types of CBN cutting tools (low CBN content and high CBN

content) and two cutting edge micro-geometries including chamfered and rounded edge. The

results indicated that the longest tool life was obtained when the low CBN tool was used,

regardless of surface type. Lim et al.(2001) investigated the effects of work materials on the wear

improvement of coated tools by comparing uncoated and TiC coated carbide tools. The

experimental results exhibited that the TiC coating was more effective when machining carbon

1045 grade, decreasing tool wear rates by half an order magnitude. Avila and Abrao (2001) studied

Page 4 of 38

Accep

ted

Man

uscr

ipt

4

the turning hardened AISI 4340 steel using mixed alumina cutting tools. They observed that the

application of a cutting fluid based on an emilsion without mineral oil resulted in longer tool life

compared to dry cutting.

Kumar et al.(2003) studied the machinability of hardened steel (EN24) using alumina based

ceramic cutting tools. It was observed that notch wear and crater wear was higher for mixed

alumina [Ti (C,N)] ceramic tool than zirconia toughened alumina (ZTA) ceramic tool when

machining the steel (40-45 HRC), but the performance of alumina ceramic tool was better than that

of ZTA. Later work by Kumar et al. (2006) found that flank wear increased with increasing cutting

speed in both types of ceramic cutting tools. The flank wear, crater wear in Ti (C,N) mixed alumina

ceramic tool was lower than that of SiCw reinforced alumina cutting tool on machining martensitic

stainless steel-grade 410 (60 HRC) and EN24 steel (45 HRC). Khrais and Lin (2006) studied wear

mechanisms and tool performances of TiAlN PVD coated inserts during machining of AISI 4140

steels at high speeds for both dry and wet machining. Dry cutting was better than wet cutting at

around 200-400 m/min speed.

Jang et al.(2000) studied tool the wear and maching performance of hardened AISI M2 steel (60

HRC) and determined the flank wear as the dominant wear mode on the ceramic tool insert but

crater wear was very small. Depth of cut was the most important factor to affect cutting force

variation, and the cutting force increased due to the tool wear. Lima et al.(2005) investigated the

machinability of hardened AISI 4340 and AISI D2 cold work tool steel at different levels of that

hardness and using a range of cutting tools. The results indicated that surface roughness of 4340

steel was improved as cutting speed was elevated and deteriorated with feed rate. However, the

surface roughness of AISI D2 steel with mixed alumina inserts allowed a surface finish as good as

that produced by cylindrical grinding. The drastic tool wear occurred for the combination of cutting

speed of 220 m/min with feed rate of 0.15 mm/rev. Xu et al.(2001) showed detail in the effect of

yttrium on mechanical properties and machining performance of Al2O3/TiC(C,N) ceramic tool. The

results exhibited that the adequately addition of yttrium improved the mechanical property of the

ceramic tool material. Barry and Byrne (2001) investigated the mechanism of Al2O3/TiC tool wear

in the finish turning of AISI 4340 hardened steel (52 HRC). The wear resistance of low content of

CBN/TiC composites was found to be superior than high content CBN tools in finish machining of

AISI 4340 steel.

Chou et al.(2002) investigated the performance and wear behaviour of different CBN tools in finish

turning of hardened AISI 52100 steel (62 HRC). The results indicated that low CBN content tools

consistently performed better than high CBN content. The flank wear rates were proportional to

cutting speed and high CBN tools exhibited accelerated thermal wear associated with high cutting

temperatures. Benga and Abrao (2003) studied the machinability of hardened 100Cr6 bearing steel

Page 5 of 38

Accep

ted

Man

uscr

ipt

5

(62-64 HRC) when dry turning using mixed alumina, whisker reinforced alumina and PCBN

inserts. The best tool life results were obtained with the CBN compact, followed by the mixed

alumina at low feed rates and by the whisker reinforced alumina when feed rate was increased.

Arsecularatne et al.(2005) studied the wear mechanisms of cutting tools such as WC, CBN and

polycrystalline diamond (PCD) using the tool life and temperature results. It was concluded that the

most likely dominant tool wear mechanism for the WC was diffusion and that for the CBN was

chemical wear.

Poulachon et al.(2001) presented various modes wear and damage of CBN cutting tools under

different loading conditions on 100Cr6 bearing steel (45-65 HRC), in order to establish a reliable

wear modelling. The wear mechanisms depended not only on the chemical composition of the

CBN, and the nature of the binder phase but also on the hardness value and above all on the

microstructure (percentage of martensite, type, composition of the hard phases, etc.) of machining

work material. Later work by Poulachon et al.(2004) investigated the tool wear mechanism of CBN

cutting tools in finish turning of various hardened steels such as X155CrMoV12 (AISI D2) cold

wok steel, X38CrMoV5 (AISI H11) hot work steel, 35NiCrMo16 hot work steel and 100Cr6

bearing steel (AISI 52100), treated at 54 HRC. They found that tool flank grooves were correlated

with the hard carbide content of the steel workpiece. For steels with only martensitic grains, the

increase in the cutting speed had a greater impact on the tool-wear rate. The last work by

Poulachon et al.(2005) investigated the evaluation of white layers produced during progressive tool

flank wear in dry hard turning with CBN cutting tools at the same workpieces. The surface profile

was influenced by the workpiece microstructure, especially if it was coarse-grained. Luo et

al.(1999) studied the wear of ceramic and CBN tool when turning AISI 4340 steel at various

hardnesses. It was concluded that the flank wear was reduced as work material hardness increased

up to a critical value of 50 HRC, and a further increase in the workpiece hardness accelarated the

tool wear rate. The tool wear was mainly due to abrasion of the tool/binder by hard carbide

particles in the steel workpieces. Kishawy and Elbestawi (2001) investigated the tool wear

characteristics during high speed turning of AISI D2 cold work tool steel. The unfavorable residual

stresses were minimized at high cutting speed and high depth of cut.

Theile et al.(2000) showed that cutting edge geometry had a significant impact on surface integrity

and residual stresses in finish hard turning. Large hone radius tools produced more compressive

stresses, but left white layers on the surface. Ozel and Nadgir (2003) reported that change in edge

geometry, increased cutting speed and depth of cut resulted in increased tool stresses and tool

temperatures at the cutting zone. Hua et al. (2005) indicated that hone edge plus chamfer cutting

edge and aggressive feed rate helped to increase both compressive stresses and penetration depth.

Yen et al. (2004) analysed the effect of various edge geometries of uncoated carbide tools by using

Page 6 of 38

Accep

ted

Man

uscr

ipt

6

FEM cutting simulation. The results exhibited that the tool wear was directly related to cutting

temperature, tool stresses, and chip sliding velocity.

There are very few publications appeared in the literature for predicting tool life using neural

network or other modelling techniques. For example, Choudry and Bartarya (2003) compared the

design of experiments technique and neural networks techniques for predicting tool wear. They

established the relationships between temperature and flank wear. They concluded that neural

networks performed better than design of experiments technique (Schefter et al.,2003; Sick et

al.,2002). Davim (2001) studied the influence of cutting conditions on the surface finish based on

the Taguchi method. The resuts indicated that cutting velocity had a greater effect on the

roughness, followed by the feed rate. Sahin and Motorcu (2008) developed the surface roughness

model using response surface methodology when machining hardened AISI 1050 steel. They

reported that CBN cutting tools produced a better surface roughness than those of KY1615 cutting

tools in all experimental conditions. Later work by Davim and Figueira (2006) investigated the

machinability evaluation in hard turning of cold work steel (D2) with ceramic tools using statistical

techniques. It was concluded that the tool wear was highly influenced by the cutting velocity, and

in a smaller degree, by cutting time. The specific cutting pressure was also strongly influenced by

the feed rate.

Al-Ahmari et al.(2007) investigated that empirical models for predicting of machinability models

(tool life, cutting force and surface roughness) were developed based on the cutting experiments on

austenitic AIS 302 steels. The developed computational neural networks (CNN), response surface

methodology (RSM) and multiple linear regression analysis (RA) are compared and evaluated. It

was found that CNN models were better than RA and RSM models. Also, RSM models were better

than RA models for predicting the tool life and cutting force models. The artificial neural network

(ANN) model of surface roughness analysis by Davim et al.(2007) reveled that cutting speed and

feed rate had significant effects in reducing the surface roughness of free machining steel.

Yang et al. (1998) used the Taguchi method to find the optimal cutting parameters for turning

operations. Analysis of variance was employed to investigate the cutting characteristics of AISI

1045 steels using cemented carbide cutting tools. He found that cutting speed and feed rate were

the significant cutting parameters for affecting the tool life. Feng and Wang (2002) used regression

analysis to develop a complete empirical model of surface roughness, using feed rate, workpiece

hardness, tool point angle, depth of cut, spindle speed, and cutting time. Hypothesis testing

established the adequacy of the model while its performance was deemed satisfactory. Their

analysis concluded that the feed rate was also identified as the most important factor along with the

cutting time. Ozel and Karpat (2005) indicated that an exponential model for both surface

Page 7 of 38

Accep

ted

Man

uscr

ipt

7

roughness and flank wear was developed considering finish hard turning of AISI H13 steels using

CBN tools. In that model, the surface roughness or flank wear was a function of work material

hardness, edge radius of the CBN tool, cutting speed, feed rate and cutting length. The modified

exponential work achieved by the same authors reported that the work material hardness and feed

rate had the greatest effect on the flank wear and surface roughness, respectively. It was followed

by cutting length and feed rate for the flank wear. Kopac et al.(2002) used the Taguchi design to

determine the optimal machining parameters for a desired surface roughness in fine turning of cold

pre-formed steels. They analysed the influence of workpiece material properties, cutting parameters

and TiN (PVD) hard coating on the surface roughness. According to their analysis, cutting speed

was the most significant influence on the surface quality and a higher cutting speed resulted in a

smoother surface. Cheng and Hong (2003) studied a comparison of tool life of tungsten carbide

coated by multi-layer TiCN and TiAlCN for end mills using the Taguchi method. The TiCN coated

tool had the best wear resistance for machining of quenched AISI 1045 carbon steel. In their

investigations, they found that the material of the tool was the main parameter among the four

controllable factors (different coated deposition, feed rate, spindle speed and tool material).

The above review shows that the most of the machining study has been focused on hard and

hardened steels using ceramic or CBN cutting tools. However, relatively few works related to

machining steels and/or hardened steels based on the tool life or surface roughness prediction

models have been reported by the Taguchi design or other methods (Cheng and Hong, 2003;

Davim, 2001; Davim and Figueira, 2006; Kơpac et al., 2002; Sahin, 2006; Sahin et al., 2004; Sahin

and Motorcu, 2008; Yang et al.1998; Ozel and Nadgir, 2002; Ozel and Karpat.2005). The Taguchi

method is a systematic application of design and analysis of experiments for the purpose of

designing and improving product quality. A high quality product can be produced quickly and at

low cost. It is relatively simple method that can be used for optimizing different production stages

with few experimental runs (Sahin, 2001; Montgomery, 2001). The aim of the present study was,

therefore, to develop the tool life model using main cutting parameters such as cutting speed, feed

rate and cutting tool’s hardness, based on the Taguchi method when machining hardened AISI

52100 bearing steels. First order linear and exponential and second order predicting equations for

tool lives were developed within reasonable error. Furthermore, analysis of variance was employed

to investigate the cutting characteristics of steel bars.

Page 8 of 38

Accep

ted

Man

uscr

ipt

8

2.0 Experimental Work

2.1. Materials

The machine used for turning tests was a Johnford TC35 Industrial type of CNC lathe machine.

The lathe equipped with variable spindle speed from 50 rpm to 3500 rpm, and a 10 KW motor

drive was used for the tests. Three types of cutting tools were used for the present work. These are

mixed alumina ceramic tools, coated ceramic cutting tools and CBN cutting tools. One of the tools

was a mixed alumina ceramic with an Al2O3 (70%);TiC (30%) matrix, which is designated by

KY1615. The other insert was coated using a Physical Vapor Deposition (PVD) method. Coating

substance takes place on the mixed ceramic substrate and PVD-TiN coated mixed ceramic with a

matrix of Al2O3 (70%);TiC (30%)+TiN, which is called as KY4400 grade. The insert types were

TNGA 160408-KY1615 and TNGA 160408-KY4400. The geometry angles of insert seating for

ceramics: nominal rake angle -6 o, back rake angle -6 o, clearance angle 6 o, approach angle 75 o,

major tool cutting angle 60 o triangle-shaped inserts and 0.8 mm nose radius. The insert was rigidly

attached to a tool holder of ISO designation of PTBNR 2525-16. These ceramic based tools are

commercially available inserts according to ISO code, supplied by Kennametal Inc. for the turning

tests. The last one was a CBN with an Al2O3+TiC matrix, which is designated by CBN/TiC. The

CBN/TiC tools contained CBN (50%); TiC (40%); WC (6%); AlN, AlB2 (4%). However, the

CBN/TiC insert type was CNGA 120408S-L0 to machine the bearing steel. The tool edge

configuration between the rake and the clearance face consists of the chamfer of 0.1 mm width and

a constant negative rake angle -20 o, side rake angle -6 o, clearance angle 6 o, approach angle 75 o,

edge major tool cutting angle 80 o diamond-shaped inserts with 0.8 mm nose radius. The tool

holder of PCLNR 2525-12A was used. These cutting tools are also commercially available inserts,

supplied by SECO Inc. Hardness’s of these cutting tools are about 2145, 2250 and 3660 HV,

respectively. Details of characteristics of cutting tools types, geometry, designation and some

properties are given in Table 1.

Page 9 of 38

Accep

ted

Man

uscr

ipt

9

The material used throughout this work was an AISI 52100 steel. The material was the AISI 52100

bearing steel containing C (0.99%), Mn (0.39%), Si (0.16%), Cr (1.40%), Ni (1.4%) and balance

Fe. For heat treatment of AISI 52100 steels, the workpieces were austenised at 850 o C for 2 h and

quenched with oil. Several tempering temperatures were selected to prepare specimens of various

hardness values. It was tempered for 1.5 h at 185 oC, resulting in a hardness of 659 HV. Samples to

be cut were in the cylindrical form of steel bars with diameter of 48 mm, length of 240 mm. These

bars were 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.

After each test, the worn cutting tool was measured with the optical tool microscope to determine

the degree of flank wear. The width of the flank wear criteria was taken as 0.3 mm. In general,

experiment was stopped to measure the width of the wear land at each 5, 10, 15 and 30 minutes.

2.2. Experimental design

The Taguchi design was selected to find out the relationships between independent variables and

cutting time. The independent variables were cutting speed, feed rate, depth of cut and tool’s

hardness. The experiments were carried out to analyze the influence of cutting parameters on tool

life for machining hardened AISI 52100 steels. Cutting parameters were selected keeping in mind

that the hard turning operation was generally used as a finishing operation as an alternative to

grinding.

The depth of cut was fixed as 0.2 mm in all test conditions. Three feed rates (f) 0.06, 0.084, 0.117

mm/rev were selected. Three cutting speeds (V) were chosen: 100, 140, 196 m/min. Details of

experimental design, control factors and their levels, and results for tool lives are shown in Table 2.

This table showed that the experimental plan had three levels. A standard Taguchi experimental

plan with notation L9 (34) was chosen. The rows in the L9 orthogonal array used in the experiment

corresponded to each trial and the columns contained the factors to be studied. The first column

consisted of cutting speed, the second contained the feed rate and the consecutive column consisted

Page 10 of 38

Accep

ted

Man

uscr

ipt

10

of the cutting tool’s hardness. The experiments were conducted twice for each row of the

orthogonal array to circumvent the possible errors in the experimental study. In the Taguchi

method, the experimental results are transformed into a signal-to-noise (S/N) ratio. This method

recommends the use of S/N ratio to measure the quality characteristics deviating from the desired

values. To obtain optimal testing parameters, the-higher-the-better quality characteristic for

machining the steels was taken due to measurement of the tool life. The S/N ratio for each level of

testing parameters was computed based on the S/N analysis. This design was sufficient to

investigate the three main effects. With S/N ratio analysis, the optimal combination of the testing

parameters could be determined.

Apart from this method, another model with transformation technique was developed and

compared. This model is an exponential model with logarithmic transformed variables such as V, f

and TH. The functional relationship between tool life and independent variables under

investigation could be postulated as;

pmn THfVCTmean ...= (1)

where V is cutting speed, f is feed rate and TH is hardness of tool. A logarithmic transformation

can be applied to convert the non-linear form of equation into the linear form of Eq.(1);

THpfmVnCTmean ln.ln.ln.lnln +++= (2)

A logarithmic transformation can be applied to convert the non-linear form of equation into the

linear form of Eq.(1). Details of the logarithmic transformation equations and estimations can be

found in previous work (Sahin, 2001; Sahin, 2006).

Page 11 of 38

Accep

ted

Man

uscr

ipt

11

3. 0 Results and Discussion

3.1. Analysis of control factor

Analysis of the influence of each control factor (V, f, TH) on the tool life was performed with a so-

called signal-to-noise (S/N) response table, using a Minitab 15 computer package. The

experimental design, results for tool lives and S/N ratios are shown in Table 2. The control factors

and their un-coded tool lives were included in this table. Table 3 shows the S/N response table of

tool lives for machining the hardened steels. It indicated the S/N ratio at each level of control factor

and how it was changed when settings of each control factor were changed from level 1 to level 2.

The influence of interactions between control factors was neglected here. The control factor with

the strongest influence was determined by differences value. The higher the difference, the more

influential was the control factor. The control factors were sorted in relation to the difference

values. It could be seen in the Table 3 that the strongest influence was exerted by cutting speed,

followed by hardness of cutting tool, lastly feed rate, respectively. Since the first level of the

cutting speed was about 41.4 dB while the third level of the cutting speed was about 27.18 dB the

difference being the most highest of 13.58 dB. It is followed by the hardness of cutting tool. The

difference between the first level of the tool hardness and third level of the tool hardness was found

to be about 12.06 dB, which is significant level again. The feed rate showed the least effect on the

tool life since the difference between the first level and third level were about 10.51 dB. A similar

observation in the mean table was also found for the tool life of machining these types of hardened

steels.

3.2. Main effect on the tool life

Fig. 1a, b shows the main effect plots for tool lives of the cutting tools for S/N ratios and mean

values, respectively. The greater is the S/N ratio, the smaller is the variance of the tool life around

the desired value. Optimal testing conditions of these control factors could be very easily

Page 12 of 38

Accep

ted

Man

uscr

ipt

12

determined from the response graph. The best tool life value was at the higher S/N value in the

response graph. For main control factors, Fig.1a indicates the optimum condition for the tested

samples (V1 f1 TH3). Thus, it could be concluded that the best tool life of cutting tools can be

achieved and their optimal setting of control factors for tested samples are shown in Table 4. From

the results of control factors, higher tool life was obtained under cutting conditions of V=100

m/min, f=0.06 mm/rev when machining AISI 52100 workpiece by CBN/TiC cutting tool. The

experimental work was carried out on the same bearing steel using the determined optimal control

factors. The tool life was found to be about 165.510 min. Then this value was transferred to the S/N

ratio (dB), average value of S/N ratio was calculated and it was about 44.376 dB. These values

were lower than that of theoretical values of tool lives in Table 5. Moreover, the mean tool life of

cutting tool is shown in Fig.1b. It was evident that the cutting speed had the greatest effect on the

optimal testing conditions. It might be that higher cutting speed led to higher flank wear width

because of easy for removal of particles from the place. It is followed by the cutting tool’s

hardness. It was observed that the tool life obviously increased highly as cutting tool changed from

level 1 to level 3 due to the having considerable hardness’s between these two cutting tools. The

feed rate was also effective on the tool life of the cutting tool (see Table 1). Effects, however, were

lower compared to those of cutting speeds. The effects of cutting parameters and their interaction

effects on the tool life are shown in Fig.2 as a three dimensional surface contour graph. As shown

in Fig.2a, longer tool life was observed when reducing the cutting speed. Increasing the feed rate

was not so effective on the tool life when machining at lower cutting speed and feed rate. However,

the tool life decreased with increasing the feed rate when machining medium cutting speed. It was

obvious that performance of all tools were good at lower cutting speeds. The tool life for the

KY1615 and KY4400 insert decreased considerably with increasing the cutting speed, but the

CBN/TiC cutting tool showed the best performance because of retaining their hardness at elevated

temperature and fracture toughness (Fig.2b). In addition, performances of all cutting tools were

good when machining at lower feed rate (Fig.2c). The cutting time reduced significantly when

tested with the KY1615 cutting tool. Moreover, the tool life decreased with increasing the feed rate,

especially for the KY1615 and KY4400 cutting tools because of not having high enough chemical

Page 13 of 38

Accep

ted

Man

uscr

ipt

13

stability of these tools. In the current study of machining hardened bearing steel, an orthogonal

design, S/N ratio and ANOVA were employed to determine the effective cutting parameters such

as V, f and tool’s hardness on the tool life. It was concluded that the cutting speed was found to be

the most important parameters on the tool life among control parameters. Similar results were

obtained with previous works carried out by Lima et al.(2005), Jang et al.(2000); Sahin (2006),

Benga and Abrao (2003), Chou et al.(2002), Yang and Tarng (1998), Davim and Figueira (2006).

For example, flank wear of the mixed alumina ceramic tools increased with increasing cutting

speed and depth of cut. It was concluded that surface roughness or tool flank wear models were

developed as a function of cutting speed, feed rate, and cutting time (Lima et al.2005). Higher

speed accelerated tool wear and increased the cutting temperature of the ceramic tool when

machining the hardened AISI M2 steel (Jang et al.2000). It was concluded that the tool wear and

surface roughness were highly influenced by the cutting velocity (Davim 2001; Davim and

Figueira 2006). In a similar study, it was reported that the cutting speed was the most dominant

factor affecting the tool life when turning of hardened 100Cr6 bearing steel (Benga and

Abrao,2003). The flank wear rates were proportional to cutting speed and high CBN tools exhibited

accelerated thermal wear when finish turning of hardened AISI 52100 steels (Chou et al.2002,

Barry and Byrne, 2001). The proposed model allowed to prediction of the tool life as a function of

the cutting parameters and workpiece hardness (Poulachon et al.2001). However, Chou et

al.(2002), Ozel and Nadgir (2002); Ozel and Karpat (2005), Poulachon et al.(2001), Poulachon et

al.(2004), Luo et al.(1999), Hua et al.(2005); Theile et al.(2000), Yen et al.(2004), pointed out that,

the most likely dominant parameters were workpiece hardness, steel microstructure, CBN content,

and edge geometry. On the other hand, Cheng and Hong (2003) indicated that material of tool was

the main parameter among other factors in milling quenched carbon steel.

3.3. Confirmation tests

Once the optimal level of design parameters was selected, the final step was to predict and verify

the improvement of the quality characteristic using the optimal level of the design parameters.

Page 14 of 38

Accep

ted

Man

uscr

ipt

14

Initial cutting parameters for machining the hardened steels were V2 f2 TH2, but it corresponded to

35.022 dB. Table 5 indicates the testing conditions used to obtain the tool life of cutting tools for

the confirmation tests. The established S/N ratio (ˆηi) using the optimal level of the design

parameters could be calculated. The mean S/N ratio was found to be about 52.905 dB at the

optimal level. The increase of the S/N ratio from the initial cutting parameter to the optimal cutting

parameter was about 17.88 dB. It meant that the tool life was increased by about 1.51 times.

Therefore, based on the S/N ratio analysis, the optimal testing parameters for the tool lives of

different cutting tools were the cutting speed at 1 level, feed rate at level 1, and type of material at 3

level. This table showed the results of the confirmation experiment for the tool life of the cutting

tool. In the table, ηi,ver indicates the verification test results, and ηi,cal shows the theoretical

calculated values in terms of optimal cutting parameters and significant factors. Table 5 also

showed a comparison of predicted tool life with the actual tool life using the optimal testing

parameter, but it is not a good agreement between the predicted and actual tool life. The optimal

testing parameter for cutting time seen in this table showed that the difference |ηi,ver -ηi,cal| was

about 8.53 dB. However, the optimal cutting time could be calculated based on ANOVA results

using significant factors. In this case, the S/N ratio was about 41.4 dB and its corresponding value

was about 117.489 min (see Table 5). The difference in the S/N ratio was about 2.97 dB. It is the

indication of a good agreement between the predicted and actual tool life. Consequently, it clearly

showed that the required cutting performance characteristics in dry turning testing processes had

great improvements through this proposed model when machining the hardened bearing steels.

3.4. Correlations

The correlations between the main factors V, f, TH and the cutting time of the tools were obtained

by multiple regressions. The obtained correlations were as follows;

The first order model equation for tool life prediction is given by;

THfVTmean *0324.0*955*790.0190 +−−= (3)

Page 15 of 38

Accep

ted

Man

uscr

ipt

15

where Tmean is the mean tool life. Eq.(3) indicated that the feed rate had the most significant effect

on the tool life of the sample when machining the hardened steel. The model had an adjusted R2

value of 87.7% and standard error was about 0.17.

The prediction model of exponential multiple regressions are given by;

THfVTmean ln*07.2ln*80.1ln*32.2269.5ln +−−−= (4)

An adjusted R2 value of 86.7% indicated that 86.7% of the variability in the tool life lnTmean is

explained by the model with factors lnV, lnf, lnTH and standard error was about 0.38.

A second order model was postulated to increase the sensitivity of the predicting equations by

taking into account of interaction effects.

The second order model equation for the tool life prediction is given by;

THfTHV

fVTHfVTmean

**09.1**00060.0

**2.6*144.0*4313*63.2656

+++−−−=

(5)

Eq.(5) indicated that again the feed rate had the most significant effect on the tool life. This

equation showed that the cutting speed was effective parameter, followed by interactions of f*V,

f*TH variables. The value of regression fitness coefficient was about 0.752. This showed that the

second order model could be explained the variation to the extent of 75.2%. The standard error for

the tool life was about 0.24.

The predicted tool lives produced by different cutting tools are calculated from the regression

equations 3, 4, 5 and tabulated in Table 6. The results are combined with some errors of the

prediction model for the first order model linear regression, exponential model in addition to the

second order models. There were some differences between theoretical and experimental values for

the first order, exponential model and second order model. For example, average error was about

24.1% for the exponential model while the average error was about 31% for the second order

model, respectively. For the most of cases, however, absolute average error was less than 20% for

Page 16 of 38

Accep

ted

Man

uscr

ipt

16

the first order and second order model, except for trial 6, 8 and 9. It was observed that the multiple

regression models were supported by the exponential model and second order model. The

regression model could be found to be better prediction capability, except for trial 6, 8, 9 when

considering each experiment separately (see Table 6). It was considered that less than 20% error

was reasonable, considering that there was an inherent randomness in metal cutting process

(Risbood, 2003; Sahin, 2006). Therefore, the model constructed in the present work could be used

to predict the tool lives of the cutting tools.

3.5. Uncertainty analysis of results

The most common estimated parameter is the population means, µ.

δµ ±= T or δµδ ±≤≤− TT (6)

when δ is an uncertainty and T is the sample mean. The interval δ−T to δ+T is called the

confidence interval on the mean. However, it depended on a confidence level. The confidence level

was the probability that the population mean would fall within the specified interval:

Confidence level= αδµδ −=±≤≤− 1)( TTP (7)

α is then the probability that the mean will fall outside the confidence interval. Since the

population distribution is normal, but n<30 and σ is not known, the t-distribution must be used

with n-1 degrees of freedom (df). This concept can be restated as

[ ] ααα −=≤≤− 12/2/ tttP (8)

Substituting for t, we obtain,

=

≤−≤− 2/2//

ααµ

tnS

TtP αµ αα −=

+≤≤− 1** 2/2/n

StT

n

StTP (9)

Page 17 of 38

Accep

ted

Man

uscr

ipt

17

which can also be stated as

n

StT n *)1.(2/ −∝±=µ (10)

Confidence interval

05.02/

90.01

==−

αα

(11)

To find the 90% confidence interval for the unknown mean of the tool life of the entire populations

of cutting tools, we first found the value of )1.(2/ −∝± nt for n-1=9 df. This was obtained from

student’s t distribution table by moving down column headed 0.05 to 8 df (Sahin, 2001). The value

we got was 1.86.

Sample mean, T

02.799

1.151.10........1215.117114 =+++++== ∑n

TT i min

Variance, 2σ

n

TTin

i

2

12

)−=∑

=σ , 175.24072 =σ

Population standard deviation, 25.46=σ .

Sample standard deviation (S ) can be calculated from the following equation.

1

)( 2

12

−

−=∑

=

n

TTi

S

n

i and 06.49=S .

Estimated standard error, n

Sx =σ , 35.16=xσ .

We could get an internal estimation of a population parameter (µ) using Eq.(10).

Thus,

413.3002.7935.16*86.102.79 ±=±=µ

Page 18 of 38

Accep

ted

Man

uscr

ipt

18

438.109413.3002.79 =+=uppµ min

602.48413.3002.79 =−=lowµ min

Therefore µ is between 109.43 and 48.602 min with a 90% degree of confidence. Another method

was to use Thompson τ test. The data to be rejected could be found from the lowest and upper

points of data ),( 21 δδ ;

TT lowerupper ±= ,2,1δ (12)

42791211 =−=−= TTupperδ

9.689.68791.102 =−=−=−= TTlowerδ

Thompsonτ test can also be used to determine the rejectable points

( 06.49,86.1,9 ==== Sn στ ) (Sahin, 2001).

SST *τ= (13)

25.9106.49*86.1 ==TS

Since 4225.911 >=> δTS , the points taken as the data are not rejectable. Otherwise,

XS, values would have been re-calculated again by assuming, n=8.

3.6. Analysis of variance

The ANOVA was used to investigate which design parameters significantly affect the quality

characteristics of the tool life for the turning process and to check the adequacy of the models

under development. Examination of the calculated value of variance ratio (F), which is the variance

of the factor divided by the error variance for all control factors. The results of ANOVA of tool

lives in machining hardened steels are shown in Table 7. In addition to degree of freedom (DF),

mean of squares (MS), sum of squares (SS), F-ratio and P-values associated with each factor level

Page 19 of 38

Accep

ted

Man

uscr

ipt

19

were presented. This analysis was performed for a confidence level of 90%. The F value for each

design parameters was calculated. The calculated value of the F showed a high influence of the

cutting speed (V) on the tool life since F-calculation was equal to 9.178 while F-table was about

9.0, but the feed rate (f), and hardness of the tool (TH) had an insignificant effect on the tool life

since F-test was equal to 5.563, 7.209, respectively. The last column of the above table indicated

the percentage of each factor contribution (P) on the total variation, thus exhibiting the degree of

influence on the result. It was important to observe the P-values in the table. From the analysis of

Table 7, the only factor A (P≈41.63%) showed a significant effect, cutting tool’s hardness

(P≈32.68%), and feed rate (P≈25.22%) were not significant on it.

4.0 Conclusions

The following conclusions could be drawn from results of tool lives of different cutting tools when

machining hardened bearing steels.

The L9 (34) orthogonal arrays were adopted to investigate the effects of cutting speed, feed rate and

hardness of cutting tools on the tool life. The results showed that the cutting speed exerted the

greatest effect on the tool wear, followed by the hardness of cutting tool, lastly the feed rate. The

estimated S/N ratio using the optimal testing parameter for the tool life was calculated. The

regression model was also supported by the exponential model and second order model as well.

Furthermore, CBN/TiC cutting tools showed the best performance than those of other tools. The

improvements of the S/N ratio from the initial testing parameters to the optimal cutting parameters

were ranged from 18% to 51% depending on the ANOVA results or S/N ratios. Moreover, the

ANOVA indicated that the cutting speed was significant but other parameters were insignificant

effect on the tool life at 90% confidence level. The percentage contribution of cutting speed was

about 41.63 on the tool life when machining the hardened steels.

Page 20 of 38

Accep

ted

Man

uscr

ipt

20

Acknowledgements

This research project was financed by the National Turkish University of Gazi in Turkey. The

author also wishes to acknowledge the technical assistance by Dr. Motorcu for carrying out this

work.

Page 21 of 38

Accep

ted

Man

uscr

ipt

21

References

Abouelatta, O.B., Mádl, J., 2001. Surface roughness prediction based on cutting parameters and

vibrations in turning operations. J.Mater.Process.Technol.118, 269-277.

Al-Ahmari, A.M.A., 2007. Predictive machinability models for a selected hard material in turning

operations. J.Mater.Proces.Technol.190, 305-311.

Arsecularatne, J.A., Zhang, L.C., Montross, C., 2006. Wear and tool life of tungsten carbide,

PCBN and PCD cutting tools. Int.J.Mach.Tools&Manuf. 46, 482-491.

Avila, R.F., Abrao, A., 2001. The effect of cutting fluids on the machining of hardened AISI 4340

steel. J.Mater.Process.Technol.119, 21-26.

Barry, J., Byrne, G., 2001. Cutting tool wear in the machining of hardened steels, Part I: cubic

boron nitride cutting tool wear. Wear 247, 139-151.

Benga, G.C., Abrao, A.M., 2003. Turning of hardened 100Cr6 bearing steel with ceramic and

PCBN tools. J.Mater.Process.Technol. 143-144, 237-241

Che Haron, C.H., Ginting, A., Goh, J.H., 2001. Wear of coated and uncoated carbides in turning

tool steel. J.Mater.Proces.Technol.116, 49-54.

Cheng, T.-C., Hong, H., 2003. Comparison of the tool life of tungsten carbide coated by multi-

layer TiCN and TiAlCN for end mills using the Taguchi method.

J.Mater.Process.Technol.123, 1-4.

Chou, Y.K., Evans, C.J., Barash, M.M., 2002. Experimental investigation on CBN turning of

hardened AISI 52100 steel. J.Mater.Process.Technol. 124, 274-283.

Chou, Y.K., Song, H., 2004. Tool nose radius effects on finish hard turning.

J.Mater.Process.Technol.148, 259-268.

Choudry, S.K., Bartarya, G., 2003. Role of temperature and surface finish in predicting tool wear

using neural network and design of experiments. Int.J.Mach.Tools&Manufac.43,747-753.

Davim, J.P., Figueira, L., 2006. Machinability evaluation in hard turning of cold work tool steel

(D2) with ceramic tools using statistical techniques. J.Mater.&Des.28, 1186-1191.

Page 22 of 38

Accep

ted

Man

uscr

ipt

22

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.

Davim, J.P., Gaitonde, V.N., Karnik, S.R., 2007. Investigations into the effect of cutting conditions

on surface roughness in turning of free machining steel by ANN models. J.

Mater.Proces.Technol. (in press).

Dawson, T.G., Kurfess, T.R., 2002. Machining hardened steel with ceramic-coated and uncoated

CBN cutting tools. Soc.Manuf.Eng. 156, 1-7.

Diniz, A.E., Oliveria, A.J., 2008. Hard turning of interrupted surfaces using CBN tools.

J.Mater.Proces.Technol. 195, 275-281.

Feng, X., Wang, X., 2002. Development of empirical models for surface roughness prediction in

finish turning. Int.J.Adv. Manuf. Technol. 20, 348-356.

Grzesik,W., Wanat,T., 2006. Surface finish generated in hard turning of quenched alloy steel parts

using conventional and wiper ceramic inserts. J.Mach.Tools&Manufac.46, 1988-1995.

Hua, J., Shivpuri, H., Cheng, X., Bedehar, V., Matsumoto, Y., Hashimoto, Y., Watkins, T.R., 2005.

The effect of feed rate, workpiece hardness and cutting edge on the residual stress.

Mater.Sci.&Eng. A 394, 238-248.

Huang,Y., Dawson, T.G., 2005. Tool crater wear depth of modelling in CBN hard turning. Wear

258, 1455-1461.

Jang, D.Y., Hsiao, Y.T., 2000. Use of ceramic tools in hard turning of hardened AISI M2 steel.

Tribol.Transac. 43, 641-646.

Jindal, P.C., Santhanam, A.T., Schleinkofer, U., Shuster, A.F., 1999. Performance of PVD TiN,

TiCN, and TiAlN coated cemented carbide tools in turning. Int.J.Recfrac.Met.&Hard

Mater.17, 163-170.

Kishawy, H.A., Elbestawri, M.A., 2001. Tool wear and surface integrity during high speed turning

of hardened steel with cubic boron nitride tool. J.Eng.Manuf. 215, 755-767.

Page 23 of 38

Accep

ted

Man

uscr

ipt

23

Kơpac, J., Bahor, M., Sokovic, M., 2002. Optimal machining parameters for achieving the desired

surface roughness in fine turning cold pre-formed steel work processing.

Int.J.Mach.Tools&Manuf.42, 707-716.

Khrais, S.K., Lin, Y.J., 2006. Wear performance and tool performance of TiAl PVD coated inserts

during machining of AISI 4140 steel. Wear 262, 64-69.

Kumar, A.S., Durai, A.R., Sornakumar, T., 2003. Machinability of hardened steel using alumina

based ceramic cutting tools. Int. J.Refrac.Met.& Hard Mater.21, 109-117.

Kumar, A.S., Durai, A.R., Sornakumar, T., 2006. Wear behavior of alumina based ceramic cutting

tools on machining stainless steel and EN24 steel. Tribol.Inter. 39, 191-197.

Lim, C.Y.H., Lau, S., Lim, C., 2001. Work materials and effectiveness of coated tool.

Surf.&Coat.Technol.146-147, 298-304.

Lima, J.G., Avila, R.F., Abrao, A.M., Faustin, M., Davim, J.P., 2005. Hard turning:AISI 4340 high

strength alloy steel and AISI D2 cold work tool steel. J.Mater.Process.Technol.169, 388-395.

Luo, S.Y., Liao, Y.S., Tsai, Y.Y.,1999. Wear behaviour of CBN cutting tools when turning AISI

4340 hardened alloy steel. J.Mater.Proces.Technol. 88, 114-121.

Montogomery, D.C., 2001. Design and Analysis of Experiments, fourth ed.,Wiley, New York.

M’Saoubi, R., Chandrasekaran, H., 2004. Investigation of the effects of tool micro-geometry and

coating on tool temperature during orthogonal turning of quenched and tempered steel.

J.Mach.Tools&Manufac.44, 213-224.

Ozel, T., Karpak, Y., 2005. Predictive modeling of surface roughness and tool wear in hard turning

using regression and neural networks. Int.J.Mach.Tools&Manuf.45, 467-479.

Ozel, T., Nadgir, A., 2002. Prediction of flank wear by using back propagation neural network

modelling when cutting hardened H-13 steel with chamfered and honed CBN tools.

Int.J.Mach.Tools&Manufac. 42, 287-297.

Page 24 of 38

Accep

ted

Man

uscr

ipt

24

Pawade, R.S., Joshi, S.S., Brahmankar, B.K., 2008. Effect of machining parameters and cutting

edge geometry on surface integrity of high speed turned Inconel 718.

J.Mach.Tools&Manufac.48, 15-28.

Poulachon, G., Moisan, A., Jawahir, I.S., 2001. Tool wear mechanisms in hard turnining with

PCBN tools. Wear 250, 576-586.

Poulachon, G., Bandyopadhyay, B.P., Jawahir, I.S.,Pheulpin, S., Seguin, E., 2004. Wear behavior

of CBN tools while turning various hardened steels. Wear 256, 302-312.

Poulachon, G., Albert, A., Schluraff, M., Jawahir, I.S., 2005. An experimental investigation of

work material microstructure effects on white layer formation in PCBN hard turning.

Int.J.Mach.Tools&Manufac. 45, 211-218.

Risbood, K.A., Dixit, U.S., Sahasrabudhe, A.D., 2003. Prediction of surface roughness and

dimensional deviation by measuring cutting forces and vibrations in turning process. J.

Mater.Proc.Technol.132, 203-214.

Sahin, Y., 2001. Metal Cutting Prenciples Vol. 2. Nobel Yayın&Dağıtım, Ankara, Turkey

(Turkish).

Sahin, Y., 2006. Machinability of AISI 1050, 4140 and E52100 steel’s by different cutting tools

and development of statistical models. Research Report of Gazi University, TEF 07/2003-38

Number Coded Project, Ankara, Turkey.

Sahin, Y., Motorcu A.R., 2004. Prediction of surface roughness in the machining the carbon steels

by cutting tools. In: S.Ghosh., J.K.Lee., J.C.Castro (Eds.), AIP Conf.Proc.712, Proceedings

of 8th Int.Confer.on Numerical Methods in Industrial Forming Applications, Ohio State

University, OH, USA, pp.1414-1422.

Sahin, Y., Motorcu, A.R., 2008. Surface roughness model in machining hardened steel with cubic

boron nitride cutting tools. Int.J.Refrac.Met.&Hard Mater. 26, 84-90.

Sahin, Y., Motorcu, A.R., 2004. A model for surface roughness in turning AISI 4140 steel using

coated carbide cutting tools. In: M.Akkok., A.Erdem., S.Engin Kılıc., E.I.Konuksever.,

Page 25 of 38

Accep

ted

Man

uscr

ipt

25

E.Tonuk (Eds.), Proceedings of 11th Int. Conf.on Mach.Des.&Prod. UMTIK2004, Middle

East Technical University (METU), Antalya/Turkey, pp.269-284.

Sahin, Y., Sur, G., 2005. The effects of various multilayer ceramic coatings on the wear of carbide

cutting tools when machining metal matrix composites. Surf.&Coat.Technol.199, 112-117.

Sick, B., 2002. On-line and indirect tool wear monitoring in turning with artificial neural

netwworks: a review of more than a decade of research. Mechan. Syst.&Sig.Proces.16, 487-

546.

Scheffer, C., Kratz, H., Heyns, P., Klocke, F., 2003. Development of a tool wear monitoring system

for hard turning. Int.J.Mach.Tools&Manufac. 43, 973-985.

Thiele,. J.D., Melkote, S.N., R.A.Reascoe., T.R.Watkins., 2000. Effects of cutting edge geometry

and workpiece hardness on surface residual stresses in finish hard turning of AISI52100

steel. ASME J.Manufac. Sci.& Eng.122, 467-479.

Yang, W.H., Tarng, Y.S.,1998. Design optimization of cutting parameters for turning operations

based on the Taguchi method. J.Mater.Proces.Technol. 84, 123-129.

Yen, Y.C., Jain, A., Altan, T., 2004. A finite-element analysis of orthogonal machining using

different tool edge geometries. J.Mater.Proces.Technol.146, 72-81.

Page 26 of 38

Accep

ted

Man

uscr

ipt

26

FIGURES

Fig. 1. Main effect plots for tool lives in machining steels: a) S/N ratio (dB), b) Mean (min)

Fig.2. Surface plots of tool lives (Tmean) vs machining varibles (V, f, TH). a) Cutting speed versus

feed rate, b) Cutting speed versus hardness of tool, c) Feed rate versus hardness of tool.

Figure

Page 27 of 38

Accep

ted

Man

uscr

ipt

27

196140100

42

39

36

33

30

0,11760,08400,0600

366022502145

42

39

36

33

30

V

Mea

no

fSN

rati

os

(dB

)

f

TH

Main Effects Plot for SN ratios

a)

196140100

120

100

80

60

40

0,11760,08400,0600

366022502145

120

100

80

60

40

V

Mea

no

fto

ol

life

(min

)

f

TH

Main Effects Plot for Means

b)

Fig. 1. Main effects plot for tool lives in machining steels: a) S/N ratio (dB), b) Mean (min)

Page 28 of 38

Accep

ted

Man

uscr

ipt

28

0

0

0,12

,100

40

80

0,08

120

10150 0,06

200

Tmean

f

V

a)

3

3500

0000

40

80

2500100

120

1502000

200

Tmean

TH

V

b)

Page 29 of 38

Accep

ted

Man

uscr

ipt

29

3

3500

0000

40

80

25000,06

120

0,080,10 2000

0,12

Tmean

TH

f

c)

Fig.2. Surface plots of tool lives (Tmean) vs. machining varibles (V, f, TH). a) Cutting speed versus

feed rate, b) Cutting speed versus hardness of tool, c) Feed rate versus hardness of tool

Page 30 of 38

Accep

ted

Man

uscr

ipt

30

TABLES

Table 1. Cutting tool’s type, geometry , designation and some properties

Table 2. Experimental design and results for tool lives and their S/N ratios

Table 3. S/N response table of tool lives in machining hardened steels

Table 4. Optimumum levels of control factors

Table 5. Confirmation test results and comparison with calculated values

Table 6. Experimental results produced by different cutting tools and their theoretical values with

absolute average errors for first and second order models

Table 7. Results of analysis of variance for tool lives in machining hardened steels

Table

Page 31 of 38

Accep

ted

Man

uscr

ipt

31

Table 1. Cutting tool’s type, geometry, designation and some properties

Type of cutting

tool

Tool

designation

Chemical composition of

material

Hardness

(HV)

Thermal

conductivity

(W.m-1K-1)

Mixed ceramic

tool (KY1615)

TNGA

160408

Al2O3 (70 %)

+TiC (30%)

2145 28

Coated ceramic

tool (KY4400)

TNGA

160408

Al203 (70%)+ TiC (30

%)+TiN

2250 32

Cubic boron

nitride

(CBN/TiC)

CNGA12040

8-L0

CBN (50 %)+TiC

(40%)+WC(6%)+AlN,AlB

2 (4%)

3660 44

Page 32 of 38

Accep

ted

Man

uscr

ipt

32

Table 2. Experimental design and results for tool lives and their S/N ratios

Control factors and their uncoded values

Experimental and theoretical values

Trial

number

V(m/min) f (mm/rev) TH

(HV)

Measured tool life ,T

(min)S/N ratio

(dB)

1 100 0.06 2145 114 41.138

2 100 0.084 2250 117.5 41.401

3 100 0.1176 3660 121 41.656

4 140 0.06 2250 108.5 40.709

5 140 0.084 3660 110 40.828

6 140 0.1176 2145 17.5 24.861

7 196 0.06 3660 97.5 39.780

8 196 0.084 2145 10.1 20.086

9 196 0.1176 2250 15.1 23.580

Mean S/N ratio (dB) 34.893

Page 33 of 38

Accep

ted

Man

uscr

ipt

33

Table 3. S/N response table of tool lives in machining hardened steels

Average S/N ratio (dB)Symbol Control factors

Level 1 Level 2 Level 3Max.-Min.

V Cutting speed (m/min) 41.4 35.47 27.82 13.58

f Feed rate (mm/rev.) 40.54 34.11 30.03 10.51

TH

Cutting tool’s

hardness (HV) 28.70 35.23 40.75 12.06

Page 34 of 38

Accep

ted

Man

uscr

ipt

34

Table 4. Optimumum levels of control factors

Main control factors Symbol Optimum level Optimum value

Cutting speed V 1 100

Feed rate F 1 0.06

Hardness of cutting tool TH 3 CBN/TiC

Page 35 of 38

Accep

ted

Man

uscr

ipt

35

Table 5. Confirmation test results and comparison with calculated values

Optimal control

factors

Verification test results for tool

life and S/N ratio, ηi,ver

Calculation test results

for tool life and S/N

ratio, ηi,cal

V f TH Tver (min) S/N ratio (dB) S/N ratio

(dB)

Tcal (min)

Difference

(dB)

S/N ratio (dB) based on optimal parameters (1,1,3)

1 1 3 165.510 44.376 52.905 398.107 8.529

S/N ratio (dB) based on ANOVA results (1)

1 - - - - 41.40 117.489 2.976

Page 36 of 38

Accep

ted

Man

uscr

ipt

25

Table 6. Experimental results produced by different cutting tools and their theoretical values with some errors for first and second order models

Experimental Regression

equation

Exponential equation Second order equation n

terms of main in terms of

interaction effects

Trial

num.

Average

measured

tool

life,T

(min)

Theoretic

al tool

life (min)

Absol.

errors,

%

lnTmean Theoretic

al tool

life

Absol.

errors,

%

Theoretical

tool life

(min)

Absol.

errors,

%

Theoretical

tool life

(min)

Absol.

errors, %

1 114.0 123.33 -8.1 4.736 5.00 -5.57 148.41 -30.00 131.72 -15.54

2 117.5 103.82 11.69 4.766 4.496 5.69 89.65 23.37 100.00 14.95

3 121.0 117.36 3.00 4.796 4.899 -2.17 134.10 -10.82 120.56 0.36

Page 37 of 38

Accep

ted

Man

uscr

ipt

26

4 108.5 95.13 2.18 4.687 4.319 -7.85 75.11 30.77 93.36 13.95

5 110.0 117.84 -7.12 4.70 4.723 -0.47 112.50 -2.27 113.88 -3.52

6 17.5 36.74 -109.9 2.862 3.010 -5.17 20.28 -15.94 29.13 -66.40

7 97.5 96.51 1.01 4.580 4.546 0.74 94.25 3.33 96.99 0.52

8 10.1 24.58 -143.33 2.313 2.833 -22.52 16.99 -68.31 20.16 -99.60

9 15.1 -4.10 127.15 2.715 2.327 14.29 10.25 32.11 -5.38 64.3

Page 38 of 38

Accep

ted

Man

uscr

ipt

28

Table 7. Results of analysis of variance for tool lives in machining hardened steels

Symbol Degree of

freedom

(DF)

Sum of

squares (SS)

Mean of

squares (MS)

F-calculation F-table Contribution, P

(%)

V 2 278.102 139.051 9.178 9.00 41.63

f 2 168.450 84.225 5.563 9.00 25.22

TH 2 218.312 109.156 7.209 9.00 32.68

Error 2 30.28 15.14 - - 4.53

Total 8 667.892 83.486 100