This article was downloaded by: [University of Illinois at Urbana-Champaign]

On: 13 March 2015, At: 11:21

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

International Journal of Production Research

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/tprs20

Correlation of performance and properties of Cermet

cutting tools: a mathematical model
a a
E. H. ROUSHDY & J. I. ELGOMAYEL
a
School of Industrial Engineering, Purdue University , Grissom Hall, West Lafayette, Indiana,
47997, USA.
Published online: 22 Oct 2007.

To cite this article: E. H. ROUSHDY & J. I. ELGOMAYEL (1986) Correlation of performance and properties of Cermet cutting
tools: a mathematical model, International Journal of Production Research, 24:2, 427-438, DOI: 10.1080/00207548608919739

To link to this article: http://dx.doi.org/10.1080/00207548608919739

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Correlation of performance and properties of Cermet cutting tools :
a mathematical model

E. H. R O U S H D Y t a n d J. I. E L G O M A Y E L t

Cermet cutting tool materials are a relatively recent entry in the growing list
of commercial tool inserts but, because of their proven good performance in a
wide range of applications, are increasing their market acceptability. This
paper reports a series of turning tests using three different brands of com-
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

mercially available Cermets to machine two different steel materials. The

experiment was statistically designed to evaluate the effects on the tool life, as
measured by the flank wear and its time rate change, of the following factors:
(a)cutting conditions: speed, feed,time and volume of metal removed,
(b) work material properties represented by its unit horsepower, and
(c) tool material properties: hardness, rupture strength, grain size, specific
weight and thermal conductivity.
A multiple classification analysis of variance wis performed and extended to
regression analysis where the dependency of the flank wear and its time rate of
change on the selected variables was established.
The study concluded t h a t the performance of the Cermet tools is affected,
in order of significance, by: the cutting speed and feed, the grain size and
hardness of the tool material and the machinability of the work material.
Joint functional relationship models of flank wear and its time rate change
with the independent variables were developed.

Nomenclature
F l a n k w e a r (in. x lo3)
F l a n k wear r a t e d W/dt (in. x 103/min)
Surface speed (ftlmin)
Feed r a t e (in./rev)
C u t t i n g time (min)
Volume of metal removed (in.3)
U n i t horsepower ( h p in.-3 min)
Tool material hardness (R,)
Tool material r u p t u r e s t r e n g t h (lb/in. x l o 3 )
Tool material grain size ( p m )
Tool material thermal conductivity ( W c m - K-') '
Tool material specific weight (Iblin.')
Tool g r a d e
Workpiece material

Introduction
Carbide ceramics a r e usually referred t o as Cermets: a combination of t h e
words ceramic a n d metal, a s t h e y consist of a m i x t u r e of 70% aluminium oxide, a

Received October 1984.

t School of Industrial Engineering. Purdue University, Grissom Hall, West Lafayette,
Indiana 47997, USA.
 E. H. Roushdy and J. I. ElGornayel

Physical
property Ceramics Cermets
Rupture strength 80 000-90 000 100 000-120 000
(IbJin.')
Density (g/cm3) 3.90-3.99 4.10-4.4
Grain size luml,
\,
2-4 1-3
Hardness (R,) 9&93 91-94
Composition AIL', AI,O, + Tic
Table 1. Comparison of physical properties of Cermet
and ceramic tools.
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

ceramic, and 30% titanium carbide, a metal. They are very closely related to the
ceramics which have been marketed for some two decades but are tougher and
more tolerant of conditions which produce thermal and mechanical shocks. Their
improved toughness is attributed to the action of the titanium carbide in prevent-
ing alumina grain growth during hot-pressing and inhibiting crack propagation
through the alumina matrix when it is stressed in use. This improvement is
expected to result in a steadily increasing use of Cermets in metal-cutting appliea-
tions.
Compared with conventional ceramic cutting tools, Cermets are harder, have
higher transverse rupture strength and are less likely to fail under either mechani-
cal or thermal shocks. They offer excellent wear resistance and can give a superior
performance in a surprisingly wide range of metal-cutting conditions. The proper-
ties of Cerments are compared to those of oxide ceramics in table 1.
Although ~ e r m e t sare tougher than ceramics, their capability for operating a t
speeds well above those for the conventional carbide still necessitates the use of
rigid machine tools and appropriate edge preparation with special edge chamfer- .
ing. They are usually set up in a negative rake position so that the insert is under
compressive stress. A typical ratio of the transverse rupture strength to the com-
pressive strength for a Cermet tool is 1 : 4, which explains why, when used in a
negative rake setting, Cermets display a great load capacity and can withstand
the increased cutting forces associated with high metal-removal rates. A large,
positive side-cutting-edge angle is also preferable with Cermet tools, to increase
the insert engagement and consequently dissipate more heat, protect the nose
radius and reduce notching and unit load.

Study objective
The main objective of the present tests was to study experimentally the
effects of different selected factors (independent variables) on the performance of
Cermet cutting tools, using physical wear as the performance measure (dependent
variable). To evaluate the cutting tools, machining tests were to be performed
including several carefully selected sets of independent variables a t constant
levels spread over their preferred ranges. During the experiments, however, main-
taining each of these variables a t a constant level was not always possible and, as
a result, values of the dependent variable (response) could differ even when mea-
sured under the same test conditions. To minimize this effect:
(1) Experiments were performed within a range of cutting conditions that
provided the highest stability and repeatability of performance.
 Performance of Cermet cutting tools

Tool grade

Hardness (Ro) 94 93t 91t

Rupture strength 96000 85000t 75OOOt
(Iblin.')
Specific weight 4.25 4.33t 5.96t
(Ib/it~.~)
Grain size (pm) 1.5 2 1
Thermal conductivity 1,187 1.20 1.1 16t
a t room temperature
(Wcm-' K - '
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

t Properties measured in the laboratory.

Table 2. Available tool grades and their properties.

(2) Sound statistical and randomization techniques were used to design the
experiment and analyze t h e results.

Factors considered
T h e flank-wear width and its progression with respect t o time were selected
(dependent variables) t o represent the physical wear of t h e tool. Three brands of
Cermet SNG433 inserts were tested. Table 2 shows t h e properties of t h e chosen
tools. Some of these properties were not available from the suppliers and had to
be measured in bhe laboratory.
Although t h e main objective was t o study the effect of tool material properties
on its performance, varying a single property while keeping t h e others fixed was
physically impossible. Each tool grade was, therefore, considered a single vari-
able. When results were obtained, however, regression analyses were applied t o
identify the effect of each property on t h e wear of the tool.
Two grades of steel were chosen for t h e work material: AISI 1080 plain carbon
steel and AISI 4140 alloy steel. Their compositions and properties are shown in
table 3. Surface speed, feed rate and cutting time were t h e variables in the
machining conditions manipulated during the experiment.

Composition Unit horse

Hardness power
Steel % C Alloying elements (BHN) (uh~)
AISI 1080 0.8 Plain carbon steel with 293t 0.892t
hot-rolled no alloying elements
0.3-1.65% manganese
0.1-0.3% silicon
+ 0.04% phosphorus
+ 0.05% s u l ~ h u r

AISI 4140 0.4 Alloy steel 305t 0.912t

hot-rolled -+0.95% chromium
+ 0.30% molybdenum

t Properties measured in the laboratory.

Table 3. Selected work materials
 430 E . H. Roushdy and J. I. ElGomayel

Levels of variables
Work
Independent Tool material,
variables grade AISI steel Cutting conditions
Speed, Feed, Time,
Name A B C 1080 4140 (surface ftlmin) (in./revl (rnin)
Number 7'1 Tz 7'3 MI M2 VI Uz V3 fl f2 f3 I1 tz 13

Values for cutting conditions are selected by preliminary tests.

Quantitative values for properties of tool and work materials are used in regression
analyses.
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

Table 4. Independent variables and their levels.

Statistical design and randomization

Table 4 displays all the factors considered in the experiment. To evaluate the
effect on the wear of changing all these factors, all combinations of the factors a t
their different levels were to be investigated. The factorial design was therefore
a
adapted. The best approach using such design was to completely randomize the
experiment. Table 5 shows the experiment design and its randomization pro-
cedure. For every test (treatment combination), the turning operation was con-
ducted, and the flank wear width measured a t specific time intervals, using a
tool-maker's microscope

Selection of levels for machining variables

From preliminary tests using various combinations of machining conditions,
tool inserts and work materials, the following cutting conditions were chosen for
the experiment: cutting speed, 3 levels: 500, 700 and 900 surface ft/min, feed
rate, 3 levels: 0.005, 0.007 and 0.009 in./rev a n d . d e p t h of cut: constant a t
0.050 in.

Cutting time and duration of tests

Three readings were .taken of flank wear over its repeatable linear range
because this range did not occur for the two extremes of the cutting conditions
during the same cutting time (figure 1). The volume of metal removed' (VMR)

Figure 1. Linear ranges for the extremes of the cutting conditions.

 Performance of Cermet cutting tools 431

Tool Workpiece material

Cutting
conditions AISI 1080 steel AISI 4140 steel

Set- Speed, V, Feed, Cutting

up (surface F, time,
number ftlmin) (in./rev) (min) A B C A B C
1 fl 11 35 29 41 3 20 1
1, 1 - 10 - 19 - 28 - 37 - 46 -
t3 al b4 c3 a3 bI CI
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

y z set-up number
x - y experiment number
z z insert number
Table 5. Randomization of the experiment

using t h e selected cutting conditions was then plotted against the cutting time
(figure 2). Fixing the VMR and varying the duration of t h e experiment allowed
wear measurements t o be taken a t the linear ranges of the curve. Wear values
were recorded after cutting times corresponding t o 40, 50 and 60 in.' of metal
removed.

Experimental results
The tests were conducted and values of flank wear ( W ) on each tool a t each
treatment combination measured (table 6). Since t h e results were recorded within
 432 E. H. Roushdy and J. I. ElGornayel

the linear range of the wear curve, the specific fixed time intervals chosen, 10, 20
and 30 minutes, were accurately estimated by extending the wear line (figure 3).
These d a t a are shown in table 7. As the wear rate (dW/dt) is a significant measure
of tool performance, its values were calculated (table 8) from the transformed
data.

Statistical analysis of the results

Multiple classification analysis of variance (ANOVA) and regression analyses .
were conducted on the data. The multiple classification analysis of variance was
applied on both collected and transformed data. The fixed-effect model was
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

adopted using the BMD2V statistical computer program with the 1% level for t h e
F-test. All the ANOVA results are summarized in ~ a b l e . 9 .Regression ' analyses
were conducted to establish the dependency of the measured variables on the
selected factors. The SPSS 15 statistical computer program was employed t o
conduct a stepwise multiple regression analysis. A separate model was developed
for each of the transformed d a t a sets examined by the ANOVA tests.
The joint functional relationships were found to be:
w =7 10-6~52.8vl.ZBt0.61 G - 0 . 7 1 7 ~ i r 0 . 5 2 ~ 2 . 6 4

where the value of the R Z coefficient with all the terms included was 0.792, and
W' = 1.4 1 0 - 2 ~ 1 . 4"29.3~-0.506~0.55

where the value of the R Z coefficient with all the terms included was 0.680 (see
nomenclature for definitions of terms). These regression models are commonly
used t o predict the behaviour of a tool, either its flank wear or its flank wear rate,
from the value* of the factors stated in the model. This method should, however,

Fig. 2. Variation of volume of metal removed with cutting time for different metal
removal rates.
 Performance of Cermet cutting tools 433

Flank wear (in. x lo3)

Cutting conditions AISI 1080 steel AISI 4140 steel

Duration
Set- V
up (surface F VMR t
number ftlrnin) (in./rev) in."rnin) A B C A B C
1 0.005 40 26.6 2.5 1.5 4.5 10.0 12.0 17.0
50 33.4 3.0 2.0 6.0 12.0 13.5 20.0
60 40.0 3.2 2.5 7.0 13.5 15.5 22.0
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

Table 6. Flank wear corresponding to a fixed volume of metal removed.

be considered as limited to predicting tool behaviour over the narrow ranges of

values of t h e factors in the models.

Discussions and conclusions

The selected factors did significantly affect t h e performance of t h e tool

Cutling conditions
Any increase in the cutting speed, feed or experiment duration increased the
flank wear. Yet, only t h e speed and feed affected t h e rate of change of the wear
 434 E . H . Roushdy and J . I . ElGomayel
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

with time. These results are supported by all past experience and well established
metal-cutting theories.

Tool material properties

Only grain size and hardness significantly affected the wear and its rate of
change with time. The regression equations indicated that increasing the grain
size and decreasing the hardness reduces the wear. This contradicts all the data
already established through metal-cutting practice and research. This inconsis-
tency could be due to:
(i) the influence o f variables not investigated having overshadowed the effect
of the variables considered; or
(ii) the quantitative values of the properties measured a t room temperature
not reflecting the actual values a t the high- pressures and temperatures
-

existing a t the cutting edge.

Workpiece material
Of all the selected factors, the workpiece material properties, as measured by its
unit horsepower (U), had the most significant effect on the performance of the
tool as measured by either its flank wear or its flank wear rate. The regression
models indicate that any minute increase in the material unit horsepower will
directly accelerate both the flank wear and its wear rate. The difference in U
between the two materials tested is relatively small t o change the wear so signifi-
cantly. A rational explanation for this uncommon behaviour would be that the
unit horsepower, as a measure of the machinability of'the work material is not the
 Performance of Cermet cutting tools 435.

Flank wear (in. x lo3)

Cutting conditions AISI 1080 steel AISI 4140 steel

Set- V
up ' (surface F t
number ftlmin) lin.lrev) (minl A B C A B C
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

Table 7. Flank wear corresponding to selected intervals of cutting time.

appropriate characteristic for judging the impact of t h e work material properties

on the life of the tools.

Interaction between tool and work material

Interaction of the tool and the work material has been proved to affect the
tool behaviour significantly, indicating t h a t there is always a combination of
Cermet and work material which, when properly selected, leads t o optimum tool
performance. The analysis of variance clearly indicate t h a t the regression equa-
tions used had a highly significant effect. Since the tool was represented as a
single factor in the ANOVA b u t its properties were broken into individual-values
in the regression analysis, the interaction effect between the tool and workpiece
 436 E . H . Rovshdy and J . I . ElGomayel

Rate of flank wear ((in. x 103)/min)

Cutting conditions AISI 1080 steel AISI 4140 steel

Set- V Time
UP (surface F range
number ftlmin) (in./rev) (min) A B C A B C
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

9 0.009 5-15 0.51 0.51 0.49 0.51 0.37 0.96

Table 8. Rate of flank wear (in the linear range).

Flank wear

Volume- Time- Rate of

Source of variability dependent dependent wear
Factors 1 Tool, T 00 00 00
2 Speed, V 00 00 00
3 Feed. F 00 00 00

- -

6 Material, M 00 00 00

Two-way 7 TV o o -
interaction 8 Tf - - -
9 TR - x x
10 Tt x - x
11 TM 00 00 00
12 Vf - - -
13 VR - x x
14 Vt x o x
15 VM o o -
16 fR - x x
17 fl x - X
18 fM 00 o -
19 RM o x x
20 tM x 00 x

oo very significant
- not significant
o significant .
x not considered in analysis
Table 9. Summary of ANOVA results.
 Performance of Cermet cutting tools

was probably a result of t h e unique combination of the Cermet material proper-

ties a n d n o t of each p r o p e r t y considered separately.
Since t h e tool behaviour h a s been proved to depend heavily o n i t s unique
interaction w i t h t h e workpiece material, any evaluation of tool performance
should consider t h e chemical a n d physical properties of the workpiece material as
well a s those of t h e tool.

Lea madriaux d'outillage trenchant Cermet sont une arrivbe relativement

kcente dans la liste croissante des pieces rappordes d'outils commerciaux.
E t a n t donne leur performance Bprouvbe de bonne qualid, dans toute une
gamme d'applications, les Cermets sont en train de devenir plus acceptables
dans le commerce. Cet article decrit une s(?ried'essais de chariotage qui ont 6 6
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

conduits e t qui utilisaient trois marques diffbrentes de Cermets disponsibles

dans le commerce pour usiner deux madriaux en aciers diff6rents.
L'experience a e t i conpue de fapon statistique pour &valuer les effets des fac-
teurs suivants sur la longevid de I'outil mesuhe par l'usure du flanc e t le
changement de taux dans le temps de celle-ci: (a) les conditions de coupe:
vitesse, avance, d u r k e t volume de metal reti&; (b) ProperiBt& du madriau
travaille represendes par sa puissance unitaire; (c) proprieds dea matC.riaux de
I'outil: dured, resistance a la rupture, taille du grain, densid e t conductiviti
thermique. Une analyse de classification multiple de la variance a A d effectube
et etendue en analyse par la mBthode d e kgression la od la dbpendance de
I'usure du flanc e t son taux de changement dans le temps sur lea variables
choisies a e t i etabli. L'etude a conclu que la performance des outils Cermet est
affecthe, par ordre d'importance, par: la vitesse de coupe et I'avence, la taille
du grain e t la d u r e d du mathriau de I'outil, e t la facilid d'usinage du madriau
travaill6. Des modeles de relation fonctionnelle jointe, avec des variables ind8-
pendantes, ont &tidivelopp6s pour I'usure du flanc e t le taux de changement
dans le temps de celle-ci.

Bei den Cermetschneidewerkzeugmaterialien handelt es sich um eine rela-

tive kurzliche Eintragung in die wachsende Liste kommerzieller Werkzeugein-
satze. Wegen ihrer gezeigten guten Leistung bei einer weiten Anzahl von
Anwendungen erholt sich die Marktakzeptierbarkeit von Cermets. Diese
Abhandlung berichtet uber eine Reihe von Drehtests, die mit drei verschiede-
nen Marken kommerziell erhaltlicher Cermets durchgefuhrt wurden, um zwei
verschiedene Stahlmaterialien zu bearbeiten. Das Experiment war statistisch
ausgelegt, urn die Auswirkungen der folgenden Faktoren auf die Lebensspanne
des Werkzeugs zu bewerten, wie sie durch die Flankenabnutzung und ihrer
Zeitratenanderung gemessen werden: (a) Schneidebedingungen: Geschwindig-
keit, Beschickung, Zeit und entferntes Metallvolumen; (b) Arbeitsmaterialei-
genschaften reprasentiert durch die Einheitspferdestiirke; (c)
Werkzeugmaterialeigenschaften: Harte, Bruchstarke, Korngrok, spezifisches
Gewicht und spezifische Warmeleitfahigkeit. Eine mehrfache Klassifizierung-
sanalyse der Veranderlichkeit wurde durchgefuhrt und zur Fkgressionsanalyse
ausgedehnt, wo die Abhangigkeit der Flankenabnutzung m d ihrer Zeitrate-
nanderung von den gewahlten Variablen festgestellt wurde.
Die Untersuchung ergab, da5 die Leistung der Cermetwerkzeuge betroffen
ist und zwar der Bedeutung nach durch: Schneidegeschwindigkeit und
Beschickung, KorngroDe und Harte des Werkzeugmaterials und
Bearbeitungsfahigkeit des Arbeitsmaterials. E s wurden gemeinsame funk-
tionelle Beziehungsmodelle von Flankenabnutzung und ihrer Zeitratenan-
derung rnit den unabhangigen Variahlen entwickelt.

References
DRAPERM,W., and SMITH,H.. 1966, Application Regression Analysis (John Wiley and
Sons, Inc.).
 438 Performance of Cermet cutting, toola

FENTON,R. G., and JOSEPH,N. D., 1979, The effect of the statistical nature of tool life on
the economics of machining, International Journal of Machine Tool Design and
Research, 19,43.
FUJIWARA, Y . , and Dor, H., February 1971, A systematic investigation of mechanical
properties of cemented carbides, Proceedings from the First I n l e r d i o n a l Cemented
Carbides Conference, MR-71-907, pp. 1-15.
HICKS,C. R., 1964, Fundamental Concepts on the Design of Experiments (Holt, Rinehart
and Winston).
NUKHEILJEEM, P. N., and BASU,S. K:, 1973, Statistical evaluation of metal cutting par-
ameters'in hot machining, I n t e r n a t i d Journal of Produetion Research, 11, 21.
OKUSHIMA. K., and FUJUE. Y., November 1960, Physical properties of ceramic tool
materials, Japanese Society of Mechanical Engineers, 3.
RANE,G. E., and SCHEITHAUER, W., J r , 1976, A correlation of cemented carbide cutting
Downloaded by [University of Illinois at Urbana-Champaign] at 11:21 13 March 2015

tool material properties with performance parameters, NAMRC, 4,469.

ROUSHDY,E. H., May 1980, CERMET cutting tools; properties and prediction of per-
formance, Ph.D. thesis, Purdue University.
SHULL,E . H., BIOGERS,J . V., and BINDT,J . M., 1975, Mechanical properties and per-
formance evaluation of model family of ceramic cutting tools, NAMRC, 3.