A

otsaia©

K

1

areeataowiaiadp

fi

((

0d

Available online at www.sciencedirect.com

Wear 264 (2008) 382–388

Experimental study of abrasive process

Matthieu Barge ∗, Joel Rech, Hedi Hamdi, Jean-Michel BergheauLaboratoire de Tribologie et Dynamiqe des Systemes UMR 5513 CNRS/ECL/ENISE,

58 Rue Jean Parot, 42023 Saint Etienne Cedex 2, France

Accepted 28 August 2006Available online 27 February 2007

bstract

According to most studies dealing with wear, abrasion can be considered on one hand as micro-cutting, leading to material removal (grinding), andn the other hand as micro-ploughing, leading to plastic deformation and lower material removal (abrasive wear). Understanding various flowingransitions around an abrasive particle, under well-established conditions, makes it possible to better control these processes. The aim of the present

tudy is to understand plastic deformation and failure local phenomena induced by an abrasive process. Experimental studies have been carried out onn abrasive process which consists in scratching a soft flat surface (AISI4140 steel) by mean of a turning tool fixed on the periphery of a disc. Regulat-ng angular and feed speeds, successive scratches are expected to study phenomena generated by a single pass of the cutting tool. Scratch topographynd forces are measured in order to highlight the influence of the depth of cut and the cutting speed on these parameters and on specific energy.

2007 Elsevier B.V. All rights reserved.

ftteIlw

wcpadatmi

eywords: Abrasive process; Single grit; Pendular scratch; Specific energy

. Introduction

The grinding process is widely used in the industry to obtainfinal expected shape. In order to be able to fullfill today’s

equirements concerning quality, productivity and economicfficiency, the use of efficient grindings methods is absolutelyssential. Many studies have been undertaken on this processnd most of them focus on the whole process analysis. Thushere still is a lack of knowledge on the mechanical microscopicnalysis of this process. This operation leads to material removalbtained from the abrasive effect of a wheel on the surface of theorkpiece. The microscopic analysis of this process consists

n studying the action of an elementary abrasive particle onn antagonist surface. This approach, known as sclerometry,s widely used to study more general abrasive process (asbrasive wear) and has led to a better understanding of plasticeformations and failure local mechanisms encountered in this

rocess [1].

Nevertheless, considering its working conditions (preform,nishing, etc.), the grinding process can generate cutting speeds

∗ Corresponding author. Tel.: +33 4 77 43 75 38; fax: +33 4 77 43 75 39.E-mail addresses: matthieu.barge@enise.fr (M. Barge), joel.rech@enise.fr

J. Rech), hedi.hamdi@enise.fr (H. Hamdi), jean-michel.bergheau@enise.frJ.-M. Bergheau).

mtgdfta

043-1648/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.wear.2006.08.046

rom 10 to 150 m s−1, and depths of cut from some micrometerso few millimeters. It is thus difficult to use classical experimen-al devices to study the phenomena encountered at those speeds,ven if the use of pendular scratch test can reach high speeds.n fact most studies dealing with this process [2–5] are oftenimited in cutting speed and depth of cut (<37 m s−1, <30 �m)hich are quite far from those used in grinding.Furthermore, the grit, in grinding, or the asperity, in abrasive

ear, exhibits different kind of geometries and can have manyutting edges [5]. Thus the complex material flow around thearticle, which leads to material removal (expected in grinding)nd plastic deformation (to avoid in abrasive wear) (Fig. 1), isifficult to identify. In fact most studies carried out on the localnalysis of grinding use, as abrasive particle, grits coming fromhe grinding wheel itself. This makes it difficult to identify each

echanisms (ploughing, cutting) since the geometry of the grits not controlled.

This paper present a work carried out to qualify the cuttingechanisms in grinding processes. The action of the wheel on

he surface is considered as the successive actions of elementaryrits on this same surface. The aim of the experimental device

eveloped for this study is thus, on the one hand to go one steporward in the cutting speed, and on the other hand to controlhe particle geometry allowing the identification of the cuttingbrasive mode.

M. Barge et al. / Wear 26

2

skmitt

wtptca

actiTd

smctitstp

3

Fig. 1. Action of an abrasive particle on an antagonist surface.

. Experimental device

The abrasive process is a pendular scratch test. It consists incratching a material with a cutting insert (which geometry isnown) fixed on the periphery of an aluminum disc (Fig. 2). Thisakes it possible to be close to the grinding process kinematic

n which one the wheel would have only one active grit. Finallyhe use of a machine tool allows to reach high cutting speeds (upo 50 m s−1) for depths of cut of 80 �m.

In the scratching configuration, the disc is moved in rotationith a (ω) speed and in translation with a (Va) speed. Regulating

hem successive scratches can be obtained, generated by a single

ass of the cutting insert and sufficiently spaced (of a ll distance)o have no interaction with eachothers. The feed speed (Va) ishosen in order to make the cutting insert working in swallowings shown on Fig. 2.

ombk

Fig. 2. Configuration and parameters

4 (2008) 382–388 383

The radius of the tool (defined as the combination of the discnd the cutting insert) is defined as the distance between theenter of rotation of the disc and the cutting insert tip and is equalo 250 mm. The cutting insert is oriented in a 45◦ configurationn order to reproduce the negative rake angle of grinding grits.he scratch is characterised by its length (lr) and its maximumepth (amax).

The sample is made of an AISI4142 steel. The scratchedurface is grinded and polished to obtain a good flatness and aean roughness of 0.114 �m. Furthermore, its dimensions are

hosen to obtain a significant number of scratches at each pass ofhe tool. It is fixed on a dynamometer to measure forces appliedn the three directions of space. In the following the normal andangential forces refer to the referential linked to the scratchedurface. In addition to force measurement, the topography ofhe scratches are measured to obtain further informations on therocess as it will be shown later.

. Parametrisation of the process

Two parameters have been chosen to study their influence

n the cutting process: the maximum depth of cut amax and theinimum cutting speed Vcmin (defined in Eq. (2)). Indeed, it will

e shown in this section that the process is completely definedinematically by these two parameters.

of the pendular scratch tester.

384 M. Barge et al. / Wear 26

Table 1Summarizes the cutting conditions and parameters for the chosen values of(amax) and (Vcmin)

amax (�m) Vcmin (m/s) ω (rd s−1) Va (m s−1) lr (mm) Tc (ms)

50 13 52 0.1 9.92 0.76426 105 0.2 0.38239 157 0.3 0.25552 209 0.4 0.191

80 5 20 0.05 12.52 2.50310 40 0.1 1.25515 61 0.15 0.83620 81 0.2 0.62730 121 0.3 0.41840 161 0.4 0.31450 202 0.5 0.251

C

rsIotfcfi

V

amdbctc(

V

fisdpi

l

eced

m

T

4

4

4

oaa

4.1.2. Longitudinal profileThe longitudinal profile first allows to check for a good agree-

ment between the imposed and measured maximal depth of cut.

Fig. 3. The 3D topography result for a scratch.

utting conditions and parameter values.

The feed speed is calculated to satisfy the no-scratch-ecovering condition. It insures that successive scratches areufficiently spaced not to have any interaction with eachothers.ts expression is given in Eq. (1). This condition only dependsn the maximum depth of cut and the chosen distance betweenwo successive scratches (2 mm in all cases presented in theollowing). It must be noticed too that this condition gives aonstant ratio between (Va) and (ω) when (amax) and (ll) arexed:

a = ll + 2√

2Ramax − a2max

2π + 2A cos[1 − (amax/R)]ω (1)

The cutting speed is defined as the vectorial sum of feednd rotation speeds. The swallowing condition gives rise to ainimum speed obtained when the tool tip is at the maximal

epth of cut (Fig. 2). This minimum cutting speed is thus giveny Eq. (2). As it has been noticed, when the maximum depth ofut and the distance between two successive scratches are fixed,he ratio between (Va) and (ω) is constant and thus the minimumutting speed can be changed for a given maximum depth of cutTable 1):

cmin = Rω − Va (2)

As soon as the two chosen parameters (amax and Vcmin) arexed the scratch length can be obtained from Eq. (3). The con-tant ratio between (Va) and (ω) makes it possible to notice noependency of the scratch length to the cutting speed. It is thusossible for a given (amax) to keep a constant scratch length withncreasing minimum cutting speed (Table 1):

r = 2√

2Ramax − a2max − 2

(Va

ω

)A cos

(1 − amax

R

)(3)

The contact time between the tool and the sample is a param-

ter allowing to define the cutting condition for which the forcesan be measured (considering the dynamometer response timequal to 150 �s). Its expression, given in Eq. (4), exhibits itsependency to both rotation speed (and thus Vcmin) and maxi-

Fe

4 (2008) 382–388

um depth of cut:

c = 2

ωA cos

(1 − amax

R

)(4)

. Results

.1. Topography of scratches

.1.1. Topography resultsThe 3D scratch topography (Fig. 3) is obtained by mean

f an interferometer. The computation of longitudinal (in thebscise direction) and lateral profile (in a given lateral section)re computed with Toposurf software (Fig. 4).

ig. 4. Theoretical and measured longitudinal profile of the scratches for differ-nt parameters of the process.

ar 26

Ago⎧⎪⎪⎪⎨⎪⎪⎪⎩

attwiwFefcfhip

4

ttrha(s

Q

oseviz

waeicdbafss

4

cttsettcaq

4

4

tTptp

i

F

M. Barge et al. / We

comparison of the measured and theoretical (Eq. (5)) profilesives interesting informations concerning the phenomenologyf the process:

l(t) = R sin(ωt − A cos

(1 − amax

R

))

− Vat + √2Ramax − a2

max

a(t) = R(

1 − cos(ωt − A cos

(1 − amax

R

)))− amax

(5)

First it can be noticed that measured and theoretical profilesre very similar. This highlights that the elastic recovery behindhe tool is neglectable. This mechanism being characteristic ofhe ploughing process, it makes it possible to conclude that,ithin the high speed range used, the predominant action mode

s cutting all along the scratch in contrast with low speeds inhich ploughing appear at the beginning as shown by Lortz [6].urthermore it can be noticed that there is no frontal ridge at thend of the scratch (or neglectable one for the scratch obtainedor amax = 80 �m and Vcmin = 5 m s−1). This makes it possible toonclude that the whole material volume is removed in the chipormed during the process. Nevertheless, these first conclusionsave to be checked by a lateral profile analysis in order to maket sure that the material do not flow on the sides as observed inloughing.

.1.3. Lateral profileThe study of lateral profiles makes it possible to confirm

he last conclusions made. Indeed, this profile is often used inhe case of linear scratch test to compute the material removalate during the process [1] neglecting elastic recovery [7] andardening [8]. In the case of pendular scratching, Brinksmeiernd Giwerzew [4] have defined a relative chip volume parameterEq. (6)) as a function of the abrasive particle position in thecratch length (l) as shown in Fig. 5:

R = AR(l) − (A1(l) + A2(l))

AR(l)(6)

This parameter is defined as the ratio of the differencef removed and pile-up material areas (AR − (A1 + A2) by thecratch area (AR). In ideal cutting the volume of the scratch is

qual to the chip volume, there is no material pile-up and QRalue is unity. In ideal ploughing the whole material remainsn the side pile-ups, there is no chip and the QR value isero. Thus this parameter gives an interesting information on

wbag

Fig. 5. Theoretical (left) and measured (right) later

4 (2008) 382–388 385

hich mode is effective during the process if it is computedlong the scratch. This value is thus calculated from the lat-ral profiles along each scratch. It is shown (as representedn Fig. 5) that the areas of side pile-ups are neglectable inomparison of the scratch area. This exhibits that the pre-ominant mode all along the process is cutting as it haseen noticed by the longitudinal profile study. These resultsre in agreement with those of Brinksmeier and Giwerzewor lower speeds which have shown that increasing cuttingpeed leads to a cutting process instead of ploughing at lowpeeds.

.1.4. Conclusion on the topographyThe scratch topography analysis makes it possible to con-

lude that the pendular scratch test carried out in this work leadso a cutting mode despite the highly negative rake angle of theool. In fact the use of a cutting insert as abrasive particle ishown to be responsible of this result since it exhibits sharp lat-ral edges leading to cutting. Finally these first results show thathe whole material is removed from the sample and exited inhe chip. Other studies could be helpful to characterise the pro-ess, as those of Wang and Subhash [9] using a microscope tonalyse the scratches and chips generated by the process with auick-stop test apparatus.

.2. Forces

.2.1. Forces during the processThe force measurement device makes it possible to obtain

heir variations in the normal and tangential directions.heir computation using a least square scheme exhibits aarabolic variation as shown in Fig. 6. This parabolic varia-ion makes it possible to conclude to a perfect symmetry in therocess.

The forces (normal or tangential) are numerically expressedn Eq. (7) as function of maximal forces and contact time:

(t) = −4Fmax

T 2c

t2 + 4 Fmax

Tct (7)

In order to compare the evolution of forces during the process

ithout any dependence on contact time (which is dependant onoth amax and Vcmin), the theoretical expression can be explaineds a function of the instantaneous depth of the scratch (a). Eq. (8)ives this relation and its approximation considering the depth

al profiles at different abscise in the scratch.

386 M. Barge et al. / Wear 264 (2008) 382–388

tial fo

o

F

ttcsTw

Fig. 6. Normal and tangen

f scratch sufficiently small in comparison to the tool radius:

(a) = Fmax

⎡⎢⎢⎢⎣1−

⎛⎜⎜⎝

A cos

(1 − a+amax

R

)

A cos(

1−amax

R

)⎞⎟⎟⎠

2⎤⎥⎥⎥⎦

(a+amax)/R = 0

a /R = 0

max

≈ −Fmaxa

amax+ ε

(a + amax

R

)2

(8)

dft

Fig. 7. Normal and tangential force measur

rce measurement vs. time.

This equation exhibits a linear dependence of the forces tohe instantaneous depth of cut. Each slope has a value equalo the ratio of the maximum force and the maximum depth ofut. Nevertheless it must be noticed that Fig. 7 exhibits that thelope seems to decrease with increasing maximal depth of cut.his show that the maximal force increases more than linearlyith this parameter. In fact, as soon as, for a given instantaneous

epth of cut, the force can be determined only by the maximumorce, it is interesting to study its variation as a function of thewo process parameters (amax and Vcmin).

ement vs. instantaneous depth of cut.

M. Barge et al. / Wear 264 (2008) 382–388 387

Fd

4

tbdltmb[tiodfip

4

ria

E

tif

g

F

odliifalnbiFscatle

5

dttsttIt(

dmNliirial removal in order to be able to qualify their impact on the

ig. 8. Normal and tangential force vs. minimum cutting speed and maximumepth of cut.

.2.2. Forces versus process parametersThe dependence of normal and tangential maximum forces to

he process parameters can be observed on Fig. 8. First, it muste noticed an increase in the forces with increasing maximumepth of cut. This result is in agreement with those related in theiterature [4] for lower speeds and depths of cut. Neverthelesshe variations of the forces with the minimum cutting speed are

ore peculiar. Indeed for the lowest speeds a few decrease cane observed what is in agreement with the Matsuo and coworkers2,3] results. Then, after a speed threshold which value is abouthe highest speeds of the last authors’ studies, a great increasen forces can be observed. This threshold seems to be dependentn maximum depth of cut. Indeed the critical speed seems toecrease when maximum depth of cut increase. It is thus dif-cult to uncouple the dependency of maximum forces to eacharameter of the process.

.3. Specific energy

Specific energy is defined as the ratio of the energy needed toemove an elementary volume of material. Thus its expressions given by the ratio of the energy given to produce a scratch (E)nd the volume of this scratch (Vr) (Eq. (9)):

spe = E

Vr(9)

The volume of the scratch can be computed from the 3Dopography and the energy from the force measurement. Thus

t is possible to study the variations of the specific energy as aunction of amax and Vcmin parameters as shown in Fig. 9.

The specific energy is often considered as independent ofeometrical parameters [10]. In our case the few differences

p(se

ig. 9. Specific energy vs. minimum cutting speed and maximum depth of cut.

bserved between the results obtained for the two maximumepths of cut can be explained by the influence of strain (highlyinked to the maximum depth of cut) on the material behav-or. The interest of specific energy computation is to study thenfluence of the minimum cutting speed. The curve obtainedor an 80 �m maximum depth of cut exhibits the same vari-tion trends as the forces. First a decrease is observed forowest cutting speed and then a quite linear increase can beoticed. These variations being highly linked to the materialehavior, it is difficult to explain them clearly from an exper-mental study. Nevertheless, some assumption can be made.irst an increasing cutting speed is associated to an increase intrain rate (which is recognized to harden the material) whatould explain the increase in the specific energy. Secondly,n other influent factor is temperature. This one is charac-erised by its influence on softening the material. Thus forowest speed it can be responsible of the decrease in specificnergy.

. Conclusion

The first conclusion on this study concerns the experimentalevice itself. Indeed it has been shown that the use of a cut-ing insert as abrasive particle leads to an ideal cutting mode inhe topography study. This characteristic has been confirmed byymmetric results obtained in forces study. This symmetry inhe process can lead to consider it as successive linear scratchests at different depth of cut with quite constant speed (Vcmin).t is thus possible to take advantage of linear scratch tests (isola-ion of cutting and ploughing modes) and pendular scratch testshigher cutting speeds).

Concerning the first results obtained with this experimentalevice, the forces and energy variation trends with the maxi-um depth of cut are the same as those obtained in the literature.evertheless, the influence of the cutting speed is more pecu-

iar to qualify. Internal variable, as temperature or strain rate,nterfere with the variation of forces and thus with energy. Its important to understand the local phenomenology of mate-

rocess. The competition between strain rate and temperaturehardening and softening) is certainly the key point to under-tand to be able to explain the variation observed in forces andnergy.

3 ar 26

R

88 M. Barge et al. / We

eferences

[1] K. Kato, Micro-mechanisms of wear–wear modes, Wear 153 (1992)277–295.

[2] T. Matsuo, S. Tayoura, E. Oshima, Y. Ohbuchi, Effect of grain shape oncutting force in superabrasive single grit tests, Ann. CIRP 38 (1) (1989)323–326.

[3] Y. Ohbuchi, T. Matsuo, Force and chip formation in single-grit orthogonalcutting with shaped CBN and diamond grains, Ann. CIRP 40 (1) (1991)

327–330.

[4] E. Brinksmeier, A. Giwerzew, Chip formation mechanisms in grinding atlow speeds, Ann. CIRP 52 (1) (2003) 253–258.

[5] H. Hamdi, M. Dursapt, H. Zahouani, Characterization of abrasive grain’sbehavior and wear mechanisms, Wear 254 (2003) 1294–1298.

[

4 (2008) 382–388

[6] W. Lortz, A model of the cutting mechanisms in grinding, Wear 53 (1979)115–128.

[7] V. Jardret, H. Zahouani, J.-L. Loubet, Understanding and quantificationof elastic and plastic deformation during a scratch test, Wear 218 (1998)8–14.

[8] M. Barge, G. Kermouche, P. Gilles, J-.M. Bergheau, Experimental andnumerical study of the ploughing part of abrasive wear, Wear 255 (2003)30–37.

[9] H. Wang, G. Subhash, An approximate model upper bound approach for

single grit rotating scratch with conical tool on pure metal, Wear 252 (2002)911–933.

10] H. Wang, G. Subhash, Mechanics of mixed-mode ductile material removalwith a conical tool and the size dependence of the specific energy, J. Mech.Phys. Solids 50 (2002) 1269–1296.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具