Sec. 4.

2]

2.1)is assu
and the bo
from AB ~
chipveloci
no discont
sameasfOi

4 to the shea
associated
zone mode
point of AI
in the sheal
Parallel-sided shear zone theory equations (

4.2 THE(
4.1 CHIP FORMATION MODEL Fora given
In the previous chapter it was shown that the strain-hardening properties of the work ionemodel
material had a profound effect on the hydrostatic stress distribution in the chip referred to
formation zone. All of the investigations described were made at very slow cutting are therefo]
speeds and it would be expected that at higher, practical speeds account would also stresses alol
have to be taken in determining hydrostatic stresses of variations in flow stress with magnitude;
strain-rate and temperature. In the first attempt to develop a predictive machining tool to be pc
theory for a material of variable flow stress which could take account of such effects the friction<
Oxley and Welsh (1963) introduced the parallel-sided shear zone model of chip shear strain
formation shown in Fig. 4.1. In this the shear plane AB of the shear plane model (Fig. shear strain
by
tl

where Vs is
nuity of the
shear zone.
k= kAB can
AB as draw
equation of (

dp=

U
00:::
For the assun
I!.k = kEF - k,
Fig. 4.1- Parallel-sided shear zone model. equation (4.2
 Theory 51

2.l)is assumed to open up so that the boundary CD between the shear zone and work
.... the boundary EF between the shear zone and chip are parallel to and equidistant
Al3 and like AB are sliplines. The work velocity is assumed to change to the
. chip ve~ocity in the shear zone along smooth geometrically identical streamJines with
. 'discontinuities in velocity. The overall geometry of the shear zone mddel is the
SalJ"... , ..... forthe shear plane model with AB and <I> in Fig. 4.1 geometrically equivalent
plane and shear angle. The force geometry of Fig. 2.2 together with the
aSS;OCllalt:U force relations given in equations (2.2) are equally applicable to :the shear
model. However, the resultant force R will not in general pass through the mid
of AB (Fig. 4.1). Chip curl is again neglected and the overall change ib velocity
the shear zone can be represented by the same velocity diagram as in Fig:. 2.1 with
)ry .'. (2.3) giving the corresponding velocities.

ardening properties of the work
stress distribution in the chip
were made at very slow cutting
tical speeds account would also
of variations in flow stress with
develop a predictive machining
uld take account of such effects
ded shear zone model of chip
B of the shear plane model (Fig.
given rake angle a and undeformed chip thickness tl the geometry of the shear
model (Fig. 4.1) is not completely defined unless the angle <1>, which will still be
.;;.T" ....."'i'I. to as the shear angle, is known. As with the shear plane theories equations
therefore sought for predicting <1>. The method of analysis is to deterP1ine the
_ stresses along AB, in terms of <I> and work material properties, and from these the
-ti magnitude and direction of the resultant force R transmitted by AB; assu~ng the
.;.~:'t66Ito be perfectly sharp <I> is then chosen so that the resultant force is consi~tent with
,:~rthefrictional conditions at the tool-chip interface. From the assumptions Jinade the
l':shear strain will be constant along AB as will the shear strains along CD and EF. The
" "., 'shear strain-rate is assumed to be constant throughout the shear zone and to! be given
,'by

. Vs
Ysz=' (4.1)
&2

where V s is the shear velocity which is equal in magnitude to the velocity disconti
,Imity of the shear plane model (Fig. 2.1) and &2 (Fig. 4.1) is the thickneks of the
~S, - "shear zone. If possible temperature variations along AB are now neglected then
" k = kAB can be assumed constant along AB as can kCD and kEF along CD land EF.
r:""AB as drawn in Fig. 4.1 is a I line and noting that it is straight the appropriate
C
work
r,,:,.
equation of equations (3.1), giving the variation in p along AB, reduces to!

"
"
ii"
dk
dp = ds ds l (4.2)
2

I
For the assumed model dklds 2 can be replaced by the finite difference 11k/11 s~, where
11k = kEF - kCD' which will be constant along AB and substituting this and ~pplying
Ie model. equation (4.2) between A and B gives I
52 Parallel-sided shear zone theory [Ch.4 Sec. 4.2]

(4.3)

where PA and Ps are the hydrostatic (mean compressive) stresses at A and Band
l(l = tl/sin <1 is the length of AB. The variation in P along AB is linear with Ps < PA acan als
when, as is usual, tlk/tJ.s2 > O. The normal force acting on AB is therefore given by the rake

FN PA +Ps lw
(4.4)
2

in the sa
where w is the width of cut. The shear force on AB is given by To r'
material
(4.5~

If the angle made by the resultant force R with AB is 9 (Fig. 4.1) then
where n
'YEFis fo
tan a=FN by the ti
Fs

=
PA+PB

2kAB

and substituting for PB from equation (4.3)

where l
Vsand

In the previous chapter it was described how in the slipline field analysis of
experimental flow fields the hydrostatic stress distribution in the plastic zone could
only be determined reliably if the stress analysis was started from a point on the fre~ which i:
surface near to where plastic deformation first occurred. In view of this Oxley and shear p
Welsh (1963) determined PAin their analysis by assuming that AB would bend to have o(
meet the free surface at the required angle 1tI4 in the region where it could still be
taken to be parallel to the cutting direction. Assuming that the angle that has to be
turned through to meet this condition (1tI4 - <1 occurs in negligible distance and
noting that at the free surface P = k AB , then from the appropriate equation cif
equations (3.1) where.
ko=kc
For
PA = 1 + 2 (~ <1 (4.6)
(4.7) tI
zone tl
kAB 4
kAB,e'
and substituting for PAin the previous equation gives
 r
1 Sec. 4.2] Theory 53
ill

y [Ch.4 ~
,.
~

~
(4.3) I a= 1 + 2 (~ $) - - -
Ilk l
I tan
2kAB fl.s 2
: (4.7)

e) stresses at A and Band

g AB is linear with Ps < PA
I
! ecan also be expressed in terms of the shear angle $, the mean friction angle ;/-.. and

I
the rake angle a by the relation
[l AB is therefore given by

e=$+/-...-a,(4.8)
(4.4)
I:.~'.

-1--. '."
in the same way as foe the shea, plane model. ;
len by
-ti '-:"'r6relate Ak and kAJ?, to the shear flow stress-shear strain curve of the 'work
"material Oxley and Welsh assumed a liriear stress-strain relation such that

(4.5) ~'.
~. Ak=m "YEF ;(4.9)
~
~.
Fig. 4.1) then
i~. -Where m is the slope of the stress-strain curve and YEP is the shear strain along EF.
~ YEP is found by multiplying the strain-rate in the shear zone given by equation: (4.1)
'"
?; by the time taken for a particle of material to flow through the shear zone, that is
.,~>
iii" :-::." ,-'
l<

where VN is the velocity normal to AB as given by equations (2.3). SubstitutiIig for

Vs and VN from equations (2.3) gives

cos a
(4.10)
: slipline field analysis of sin $ cos ($....; a)
III in the plastic zone could
ed from a point on the free which is the same expression as for the shear strain occurring as material crosses the
. In view of this Oxley and shear plane. If now it is noted that for the assumptions made half of this strai~ will
ng that AB would bend to have occurred at AB then
gion where it could still be
lat the angle that has to be
in negligible distance and (4.11)
e appropriate equation of
.. where ko is the shear flow stress at zero plastic strain with, for the machining mbdel,
ko=kcD' ,
For given values of a, J... and t1 it is now possible to determine $ from equ~tions
(4.6) (4.7) to (4.11) if the appropriate values of ko and m and the thickness of the shear
. zone tJ.s 2 are known. $ can then be used together with the corresponding valqes of
kAB' etc., to calculate the cutting forces from equations (2.2). :
 ,.
'.""
I
r.

54 Parallel-sided shear zone theory [Ch.4 Sec, 4.4]

in Fig. 4.2 tht

4.3 EFFECT OF STRAIN-HARDENING ON SHEAR ANGLE from earlier \
In an early application of the parallel-sided shear zone theory, Oxley and Welsh slope. kAB w;
(1963) attempted to explain the wide variation in experimental shear angle values in to be reason::
Fig. 2.5 in terms of the strain-hardening characteristics of the various work materials strain and u
used in obtaining these results. Their calculations showed that for the case m = 0 k = cr/v' 3 anI
(non-hardening material) the relation between $ and A- a, which can be obtained
directly from equations (4.7) and (4.8), closely followed the Ernst and Merchant
relation given in equation (2.5) and represented by a straight line in Fig. 2.5, while
for hardening materials the predicted results fell below this line. When m =1= 0 then it
is necessary to know the value of &2 in order to determine $. To find &2 Oxley and
We!sl! assumed on the basis of the experimental evidence available at the time that
the ratio of the length'to diickness of the shear zone, 1/&2' would be constant and
equal to 10 for the range of cutting conditions and work materials they were
considering. With 1/&2 assumed constant, equations (4.7) to (4.11) show that for a
given A-a an increase in mlkAB is predicted to decrease $. Oxley and Welsh
. therefore estimated mlkAB for the materials represented in Fig. 2.5 to see whether
this factor could be correlated with the observed variations in $. Fortunately
Kobayashi and Thomsen (1959) had provided effective stress-effective natural strain
curves for these materials, which are reproduced in Fig. 4.2, but not surprisingly
.From these r
, ,. . for alpha bn
. steel (anneal
700 - , T4 aluminiux
cE
:;t
this trend. T
-;; 560 does appear
least in prine
others.
-
I f!
.4.4 STRAI
From the ea
relatively hi~
.' appeared th:
_ had a value I
I I I Using estim,
o 0.1 0.2 0.3 O.L. 0.5 0.6 0,7 0,8 O,g equation (4.
effective natural strain, E
turning, the
lOSls or ever
Fig. 4.2 - Effective stress-strain curves: 1, SAE 1112 steel (as received); 2, 2024-T4 aluminium 1O- 1/s norm
alloy; 3, SAE 1112 steel (annealed); 4, 6061-T6 aluminium alloy; 5, alpha brass. be seen, stra
with the flo\
Therefore t(
these were obtained from conventional slow speed (compression) tests. Lacking effects. The
similar data for the high strain-rates of machining (see next section) Oxley and Welsh would be ex
used the stress-strain curves in Fig. 4.2 to measure mlkAB for their comparison with work now to
I
the machining results. In effect they assumed that if materials sorted into a certain I. taken of terr
order on the basis of mlkAB at low strain-rates then they would follow the same order
at the high strain-rates applicable to machining. In calculating mlk AB from the curves
ory [eh.4 .- 'Sec. 4.4] Strain-rate effects 55

in Fig. 4,2 they took m as the average slope above an effective strain of 0.2 reasoning
'l'GLE frornearlier work (Oxley et ai. 1961) that this would be more relevant than t~e initial
,~.~lope:kAB was taken as the value of k at an effective strain of 0.5, this being ~ssumed
,eory, Oxley and Welsh
imental shear angle values in , ',to be reasonable on the basis of strain values calculated from equation (4.1O~. Plane
of the various work materials stniin and uniaxial conditions were related as described in Appendix A2 with
)wed that for the case m = 0 k'/::d/V 3 and y = V3e. The values of mlkAB found in this way are given in Tdble 4.1.
~ - a, which can be obtained
ved the Ernst and Merchant
traight line in Fig. 2.5, while Table 4.1
this line. Whenm:l= 0 then it
line </>. To find ~2 Oxley and Material mlk
'Ice available at the time that 1. SAE 1112 steel (as received) 0.14
l/~2' would be constant and 2. 2024-T4 aluminium alloy 0.10
I work materials they were 3. SAE 1112 steel (annealed) 0.25
~.7) to (4.11) show that for a 4. 6061-T6 aluminium allo.y 0.11
crease </>. Oxley and Welsh ~ '.~' 5. Alpha brass 0.42
.. " Jl~
ed in Fig. 2.5 to see whether '~':~":
.,..", ..
.
'

'ariations in </>. Fortunately ,:,~~.>

,tress-effective natural strain

:;'ig. 4.2, but not surprisingly
~;~~~"<
')trlff >,.
,
'

'f:!,'F'rom these results it would be expected that the range of values of </> would b~ lowest
':~~i~{(bt.alpha brass which has the highest mlkAB value and then in the order S~E 1112

- '::~~'steel (annealed), SAE 1112 steel (as received), 6061-T6 aluminium alloy a~d 2024
:itrL T4 aluminium alloy. The experimental results for </> in Fig. 2.5 can be seen to confirm
:~~}'lhis trend. Therefore in spite of the somewhat speculative nature of the an~lysis it
."iit~::' does appear that once hardening effects are allowed for it is possible to explain at
',,,",' .least in principle why some materials machine with much smaller shear angles than
,others.

A.4 STRAIN-RATE EFFECTS

,,"F-tom the early experimental work of Kececioglu (1958) and Nakayama (959) at
";,relatively high cutting speeds and of Palmer and Oxley (1959) at low cutting speeds it
appeared that the ratio of the length to the mean thickness of the shear zon~ (l/~2)
;,had a value between 6 and 15 for a range of work materials and cutting conditions .
Using estimated values of ~2 found in this way together with typical values bf Vs in
5 0,7 0,8 0.9 " (4.1) shows that for practical cutting speeds, which are typical say of
lin, E .. turning, the average shear strain-rate in the shear zone lies in the range froq. 1()3 to
,:,105/s or even higher. These values are much higher than the strain-rates of ~0-3 to
:ceived); 2, 2024 T4 aluminium ': 1O-:- 1/s normally encountered in conventional tension and compression tests.i As will
1m alloy; 5, alpha brass. , be seen, strain-rate can have a marked effect on a material's stress-strain pr@perties
with the flow stress of many metals increasing rapidly with increase in strain-rate.
"Therefore to be realistic any machining theory should take account of str~in-rate
compression) tests. Lacking :~effects. The temperatures generated in machining and their effect on flo~ stress
ext section) Oxley and Welsh would be expected to be equally important but this was not fully recogniseb in the
AS for their comparison with work now to be considered. Later chapters describe how due account was ev~ntually
,aterials sorted into a certain taken of temperature. I

would follow the same order

lating mlkAB from the curves
56 Parallel-sided shear zone theory [eh.4

Because of the difficulty of making speeded-up conventional tests to measure the

flow stress properties of materials at the high strain-rates appropriate to niIachining a
number of workers have suggested using a machining test itself for this purpose. One
of the first investigations of this kind was carried out by Kececioglu (1958). He made
turning tests on SAE 1015 steel under approximately plane strain, steady-state
conditions in which he measured the thickness of the shear zone i(which he
approximated to a parallel-sided shear zone), the shear angle and the cutting forces
for various values of rake angle, undeformed chip thickness and cutting ~peed. The
shear angle <I> was calculated from equation (2.11) using measured val$es of chip
thickness t2 . Referring to Fig. 2.2 the forces measured were Fe and FTandlthese were
used to calculate e and then Fs from the equations

l
0= <I> + tan- (;:) (4.12)

and

(4.13)

The thickness of the shear zone Lis2 (Fig. 4.1) was measured from photomicrographs

of 'quick-stop' chip sections. From these it was relatively easy to pick out the

boundary between the undeformed work and the shear zone but less easy to pick out

the boundary between the deformed chip and shear zone. These measurements had

therefore to be treated with some caution. Once <1>. Fs and &2 had been determined

then k AB and ysz were calculated from equations (4.5) and (4.1) using the given

. values of a, U, tl and w. The results obtained showed in agreement with previously

reported work at lower strain-rates that flow stress increased with increase in

strain-rate.
Oxley (1963b) pointed out that Kececioglu (1958) had taken no account of
possible differences in shear strain on AB for the different test conditions used and
that this could have masked the effect he was looking for. Oxley overcame this
difficulty by using the equations obtained from the parallel-sided shear zone analysis
to determine additional information from Kececioglu's test results. iFrom the
experimental values of <1>, 0, &2 and kAB' together with the given cutting ~onditions,
he calculated !:l.kfrom equation (4.7) and thenm from equations (4.9) and (4.10). He
then found ko from equation (4.11). Values of ko and m, which define the stress
strain curve for a particular strain-rate, derived in this way are shown plottied against
shear strain-rate values calculated from equation (4.1) in Fig. 4.3. It can be seen that
ko increases with increase in strain-rate which is the expected trend. The v!alues of m
rather surprisingly fall on two curves, both curves showing a decrease lin m with
increase in strain-rate. The reason for the points' falling on two curves is:not clear.
Oxley and Stevenson (1967) pointed out that the top curve represented ipoints for
which the shear strain was approximately 2 and the mean temperature in the shear
zone was approximately 250"C while the corresponding values for the l<>t~er curve
were 4 and 350"C. In the light of this they considered possible explanatibns. They
 reh.4 Sec. 4.4] Strain-rate effects
Cl
iona] tests to measure the
''1riate to machining a
)r this purpose. One
::ecioglu (1958). He made
-
~
~630.-.-r-.-.-.-.-'-'--r--r~-T~~~'-~~~~

1Il~ 560
.VI
.' ~
,lane strain, steady-state
, 'Til 490
~ shear zone (which he o o
3:
gle and the cutting forces
. 5S and cutting speed. The
measured values of chip
-
E

Fe and FT and these were

4 6 8 10 12 14 16 18 20
shear strain-rate, i 110<4 s-'
{a J
(4.12)
70

(4.13)
-
~

E 56
(J)~

...
:>
:::l
U
j from photomicrographs
S 42
!ly easy to pick out the
e but less easy to pick out
These measurements had
-E
I II
I
III

-
III
&2 had been determined ~ 28
md (4.1) using the given III

'ent with previously

'0
;, __.~cd with increase in 0
(J)
0 14 0
0
lad taken no account of iii 0
. test conditions used and
or. Oxley overcame this
sided shear zone analysis
I I I I I I
0 2 4 6 8 10 12 14 16 18 20
, test results. From the
given cutting conditions, shear strain- rate, 1/10<4 s-'
tions (4.9) and (4.10). He
(b l
which define the stress
are shown plotted against ~.,
Fig. 4.3 - Values of ko and m calculated from machining test results: (a) leo results; (b) m results.:
I
ig. 4.3. It can be seen that
~d trend. The values of m .. ~:: ,
ng a decrease in m with c""'quicklyruled out temperature effects as the cause and concluded that the most likely
n two curves is not clear. .:'explanation was to do with the assumption of a linear stress-strain relation whenl in
Ie represented points for . .~eality compression tests on similar steels to that used in the machining tests show:ed
temperature in the shear that experimental points gave a good fit with a relation of the form .
,dues for the lower curve
;sible explanations. They (4.14)
.
i
58 Parallel-sided shear zone theory [Ch.4 . Sec. 4.4J

where 0" and e are the uniaxial stress and natural strain and 0"1 and n (strain-hardening .
index) are constants. Equation (4.14) shows that the slope ofthe stress-strain curve ' 19(y.
decreases as strain increases and if it were applicable to machining conditions this
would explain why for larger strains the use of a linear approximation gives smaller
values of m as noted. In view of this Oxley and Stevenson decided to use the which on simI=
machining results to calculate the constants 0"1 and n in equation (4.14) and in this
way to determine whether the variation of n with strain-rate was more consistent with
expectations than that for m. The method they used is now described.
n=3
Equation (4.14) can be represented as shown in Fig. 4.4 with the points CD, AB

For given vah

can be used to
equation (4.1.
EF against f; (e =

CD ~=---___--L
980

840 I

O"co=v'3k CD
cE
ECD=-E
v'3k CD
-
:::E 700 \
I:>

1/1~
1/1 560
1
{

(l}
19 E
.....
~

III

Fig. 4.4 - Construction for obtaining equations (4.15) and (4.16). ~ 420
0

280

and EF indicating the stress and strain on the corresponding sliplines in Fig. 4.l.
Plane strain and uniaxial conditions have again been related in the usual way as 140
described in Appendix A2. The strain on CD (zero plastic strain) is taken as the
elastic strain just at the point of yielding (note that E is the elastic modulus). From
Fig. 4.4 0

. results certaiI
(4.15)
strain-rate an.
with increase
those for m, ~
Oxleyanc
and
 [Ch.4 Sec. 4.4] Strain-rate effects 59

~ain-hardeoing
ss-strain curve
ning conditions this
nation gives smaller
,_.lI..~ ~nn... simplifying gives
decided to use the
m (4.14) and in this
, -,
nore consistent with
;ribed. ,!, 'n = 3.32 Ig (kEF) (4.16)
the points CD, AB kAB

values of ak (ak = kCD), kAB and kEF

equations (4.15) and (4.16) rEF
l'Ie~luse:Q to find n and the initial flow stress 0'0 (0'0 = V3k cD )' 0' 1 is then found from
....ry.;i.....t.I'nn (4.14). Values of n, 0'1 and 0'0 calculated in this way are shown plotted
,.n .. lin"~" i: (i: = YszlV3) in Fig. 4.5. Although there is considerable scatter in the

0.06
c::
F x~
(JJ
0.05 -c
c::
\
~ 0 \ O'l
~'560
(JJ
0\ 0 0.04 '2
(JJ
.::; \ -c
til \ClO
d
0.03 .c.
U6). O
~ 420 '\.
'\.0 I
- &-.0
280 '~ -__
n
\j 0
Cl CI
-...JJ_ _ _ _ _ _
-..0----- Cl
/iplines in Fig. 4.l. CI CI CI o
140 0.01
n the usual way as
lin) is taken as the
tic modulus). From
o 2 4 6 8 10 12
strain-rate. V 10.5.1 .

Fig. 4.5 - Values of n, O't and 0'0 calculated from machining test results.

certain trends can be clearly seen. 0"0 and 0'1 both increase with increase in
(4.15) , and their difference (0' 1 - 0'0) decreases at higher strain-rates. n decreases
, increase in strain-rate in a similar way to m (Fig. 4.3) but the results, unlike
those for m, appear to follow one line.
, " Oxley and Stevenson compared their results with results obtained by Manjoine
60 Parallel-sided shear zone theory [Ch.4

(1944) from tension tests for strain-rates in the range from 10- 6 to l()3/s. (Mainjoine's
strain-rates are based on engineering strain (change in length divided by!original
length) but for the small strains he considered, the difference from natural stl1ain-rate
would be negligible.) Manjoine described his test material as a commer~ial low
carbon, open-hearth steel. Although this material was not the same as tha~ used in
the machining tests it was similar and Oxley and Stevenson considered it wohhwhile
to compare the two sets of results. The tension tests were made for variOlls initial
temperatures of the test specimens (room temperature, 200, 400 and 6OOcr) and it
was found that the best agreement between machining and tension test resflts was
obtained when the room temperature results were used in making the comparison. In
this connection it is interesting to note that the initial work temperature in the
machining tests was also room
- . -,
temperature and that in both sets of test~I plastic
. - -_. ..
deformation would .have.1eg ~oIlJ~.mpe_rat~~e rise with this more pronoQ-nced at
higher strain-rates. Room temperature values of initial flow stress 0'0 and fldw stress
at a direct strain of approximately 0.19 obtained by Manjoine are given in fig. 4.6.
From these n can be found directly from the relation I

(4.17)

I
where the strain at yield is again taken as the elastic strain just at the point of ~elding.
Once n is known then cr1 can be determined from equation (4.14). Th~ results
obtained from both the machining and tension tests are given in Fig. 4.7. In[spite of th.
!
re!
as

4.:
Fo
in,
fOl
eq
chi
210 k:----~--
_~ 140 re~

(6
70
eql
O~-J--~--~--~ __- L_ _- L__- L__ ~~
wil
10- 6 10- 5 10- 4 10- 3 10- 2 10-1 10 10 2 10 3 val
strain-rate, E/s-1 del
(4.
Fig. 4.6 - Manjoine's (1944) tension test results. sui
y [Ch.4 . Sec; 4.5J Predicting the influence of cutting speed 61

1.(("6 to l()3/s. (Manjoine's

divided by original
'I) 0.18
Jm natural strain-rate
erial as a commercial low 0.16 ;';'
Dt the same as that used in c
n considered it worthwhile 0.14 XC
~,
re made for various initial 6 16 6
-0'
.f:
200, 400 and 600C) and it 6/ 0.12
61 en
111
nd tension test results was /
.51
0.10 c'
making the comparison. In ~I
-0'
work temperature in the a;
-c:,
) both sets of tests plastic
this more pronounced at bi
._1
06 eI,
)w stress 0"0 and flow stress ~'

joine are given in Fig. 4.6.

IJ 1ii!
0"0
IJ
,Ill

01
~ 0.02

4 6
(4.17)
10- 6 10 10- 2 10 2 10" 10
strain-rate, C/S-1

Fig. 4.7 Values of n, 0"1 and 0"0 calculated from machining and tension test results: open:
symbols represent machining values and filled-in symbols represent tension test values. '
just at the point ofyielding.
uation (4.14). The results
iven in Fig. 4.7. In spite of scatter in the machining results it can be seen that the fit between the two set~ of
. , is remarkably good, thus giving some support to the idea of using machining
'. high strain-rate property test.

PREDICTING THE INFLUENCE OF CUTTING SPEED ON SHEAR

ANGLE AND CUTTING FORCES
IlIfl'Wtn,a the introduction of strain-rate into the machining analysis a number: of
vestigl'ltic)fls were made of the influence of cutting speed on shear angle and cutting
and these are now described. In all of these the ratio Ul!.s2 was assumed toibe
.to 10. In the investigations considered first the work material flow st~ess
rat,actensu'(;s were represented by the curves in Fig. 4,3 or by similar curves. :
.. "For

given values of a, A., U and
t1 the method of calculating <I> is as follows!I A
eruionaO.le guess is first made of the value of <I> and this is used to determine &'2
=t1110 sin <1. Ysz is then calculated from equation (4.1), with Vs givenlby
eq1l1atlorls (2.3), and the corresponding values of ko and m are found from Fig. ~.3
a single average curve used to represent the m values. Then, substituting the
OfrEF given by equation (4.10) in equations (4.9) and (4.11), 6.k and kAB are
detennined and used with equation (4.7) to find 6 which in tum is used with equation
to find A.. In general this value of A. will not agree with the given value and ~ is
:suits. suitably adjusted and the process repeated until the difference between the ~o
62 Parallel-sided shear zone theory [eh.4

values of A. is acceptably small. Oxley and Welsh (1963) using this approach sbowed
a
that their predicted results when plotted on the basis of tP against A. - as in Fig. 2.5
indicated that for a given A. a a decrease in cutting speed from 300 ~o 0.01 mlmin
gave a decrease in tP of approximately 100. Their results also showed a size effect
because of the dependence of Ysz on t 1, a decrease in t1 increasing Ysz. In this
connection for a given A. - a a four-fold increase in t1 was predicted tOldecrease tP by
about 5. Experimental results were presented showing good agreemkmt with these
predictions. Oxley and Welsh (1967) applied the theory to a re-a)Jalysis of the
macbining results of Merchant (1945) which he had used to determine the slope s
(equation (2.6 of his proposed linear shear flow stress against pormal stress
relation. Plotting shear and normal stress values on the shear plane calcnlated from
his machining results Merchant showed that the relation was approximately linear
with s "'" 0.23 and that if this value was substituted in equation (2.6) tlien this gave a
good fit with his experimental shear angle values. In their analysis Oxley and Welsh
started by using Merchant's machining results to calculate m and ko ~s functions of
strain-rate in the way described in the last section. These showed rd.sonably close
agreement with the results in Fig. 4.3 with the m results very close to the average m
curve and with the ko results lying on a line parallel to but about 25% higher than the
ko line in Fig. 4.3. They then used these values to predict tP from the ~arallel-sided
shear zone theory for Merchant's cntting conditions including bis: experimental
values of A.. The line obtained when the predicted results were plotted 'on a tP against
A. - a graph showed excellent agreement with Merchant'S experimentitl shear angle
results. Oxley and Welsh also plotted kAB against (PA +PB)/2, the average normal
stress on AB (Fig. 4.1), using results obtained from their analysis and like Merchant
obtained an approximately linear relation of slope =0.23. They pointed out that
their results were a consequence of the hardening properties of the v{ork material,
which led to P varying along AB, and that they were in no way indicative of a
Bridgman-type relationship in which the shear stress is directly affected by the
normal stress as assumed by Merchant and for which a slope of O~23 would be
unacceptably large. ;
In the applications of the parallel-sided shear zone theory so far cotitsidered A. has
been used as the tool----(;hip interface friction parameter in the same 'fay as for the
shear plane solutions. If J. could be measured by some independent friFtion test or if
its value did not vary greatly with cntting conditions this would not be a great
disadvantage. However, it has only been found possible to measure A. !by machining
tests and the variations in A. with a, U and tl even for a given work!material-tool
material combination are too great to allow the use of an average valud. Thus A. must
be measured from machining tests before 'predictions' can be made, which is clearly
unsatisfactory. This has led Oxley (1966) and others including Ro.Jve and Spick
(1967) to suggest that the shear strength (shear flow stress) in the ~hip material
adjacent to the tool----(;hip interface might be a more effective friction parameter. This
idea is based on experimental observations, such as those described i~ the previous
chapter, which show that plastic deformation occurs in a layer ofthe cb1P adj acent to
the surface which slides over the tool. This can be contrasted with ndrmal metallic
sliding friction conditions where plastic deformation is limited tJ the tips of
contacting asperities and it can usually be assumed that for a given slitling pair and
lubrication condition the coefficient of friction 11, and hence A., will b~ more or less
,
~ory [Ch.4
I) 1,t~ing this approach showed
. --ainst ~ - a as in Fig. 2.5
',. trom 300 to 0,01 mlmin
ults also showed a size effect
. in t1 increasing Ysz. In this
'as predicted to decrease 4> by
kg good agreement with these
teory to a re-analysis of the
Ised to determine the slope s
stress against normal stress
e shear plane calculated from
ion was approximately linear
quation (2.6) then this gave a
teir analysis Oxley and Welsh
Ilate m and ko as functions of
lese showed reasonably close
ts very close to the average m
lut about 25% higher than the
dict 4> from the parallel-sided
s including his experimental
Its were plotted on a 4> against
nt's experimental shear angle
. + Ps)J2, the average normal
!ir analysis and like Merchant
'0.23. They pointed out that
'perties of the work material,
r 1n no way indicative of a
lirectly affected by the
;h .. slope of 0.23 would be
theory so far considered ~ has
~r in the same way as for the
independent friction test or if
lS this would not be a great
Ie to measure ~ by machining
'r a given work material-tool
In average value. Thus ~ must
can be made, which is clearly
s including Rowe and Spick
, stress) in the chip material
ctive friction parameter. This
ose described in the previous
a layer of the chip adjacent to
ltrasted with normal metallic
m is limited to the tips of
it for a given sliding pair and
hence ~, will be more or less
Predicting the inftuence of cotting speed 63
.' .. ; The nature of tool-<hip interface friction is considered in detail in section
.. ' For~the present, attention will be limited to the replacement of ~ by the average
. ,:..,:lIc,:u.stress at the interface as the friction parameter.
.. '. .' shear stress at the tool-<hip interface is given by
(4.18)
2.2) is the frictional force at the interface, h is the tool-<hip co~tact
is the width of cut. Oxley (1966) has obtained an expression for h in the
Taking moments of the normal stress on AB (Fig. 4.1) about B it can
the resultant force R (Fig. 2.2) cuts the tool-<hip interface at a distance
2[1 sin eiPA +WB
cos ~ sin 4> P A + Ps
.' cutting edge, where PA and PB are the hydrostatic stresses at A and )3. If
iecliStribl11tic)O of normal stress along the tool-<hip interface is known then h
For example if, as assumed by Oxley, the distribution is triangular!with
lQ.:l!o.UUUUJ value at B then x = hl3 and therefore
[1sin e 2PA +PB
(4.19)
cos ~ sin 4> PA + PB
r'etlla(:ed by't as the friction parameter the method of calculation is the same as
the point where ~ is found from equation (4.8). Then, Fis found trom
. (2.2) and h from equation (4.19), withpA andpB given by equations (4.6)
; and these are substituted in equation (4.18) to give 'to The procdss is
'with 4> suitably adjusted until the calculated and given values of 't are in
agreement. Working backwards from machining results Oxley (~966)
above equations to determine 't and showed that this was reasonably
equal to about half the value of kAB for the conditions he considered.
and Oxley (1967) using 't as the friction parameter investigated, the
ofcutting speed on the cutting forces Fc and FT (Fig. 2.2) over a very large
Qt:s:bec~ds (30 to 3000 mlmin). t They represented the work material How stress
:teristi,csusing equation (4.14) with the required values of 0'1 and n found from
While recognising that 't would most likely vary with cutting speed, etc.,
",,"'11,.,...,,1 for the purpose of their analysis that it could be taken to be con~tant
. to approximately half the value of k AB' Their results showed that over the
speed range considered 4> was predicted to increase with increase in speed
the tool-<hip contact length h and force normal to the cutting velocity F{' iv'ere
, '
be noted that, even at a cutting speed of 6000 mlmin, Fe is increased by less than 3% fiy the
ce.resultinlil from the rate ofchange of momentum in chip fonnation. However, as speed is increased
this force increases rapidly. .
64 Parallel-sided shear zone theory [eh.4

predicted to decrease. The force in the cutting velocity direction Fe was predicted to
decrease up to a speed of approximately 300 mlmin and then to increase rapidly with
further increase in speed. They explained this result by noting that up to 300 'mlmin
the reduction in Fe was caused by the increase in <I> which reduced the area! of the
plane AB and hence the force F swhich is given by the product of k AB and the area of
AB (see equation 4.5). However, at higher cutting speeds although <I> still ind:eased
with increase in speed the reduction in area of AB was more than compensatedlby the
increase in kAB resulting from the increase in strain-rate, and Fs , an4 as a
consequence Fe. no longer decreased but actually increased. Fenton and:Oxley
considered that their results cast great doubt on the suggestion made by a nu~ber of
workers that superhigh cutting speeds could lead to a dramatic fall in Fe. They
5
recognised,. however,
. .
that their analysis had limitations
. . . .
particularly those resulting
I
from the lack of consideration given to temperature.
Fenton and Oxley (1968-1969) in a later investigation used the parallel-sided
i
Ex
shear zone theory to calculate 1: from the machining results of Merchant (1945~. They sp~
found that 1: varied markedly (a variation of approximately 60% from the Imyest to
the highest value) with a, U and tl and that this variation could only be explained if fOI
account was taken of the strain-rates and temperatures in the plastic zone lat the
tooi-chip interface. They therefore suggested that 1: should be expressed in terms of
strain-rate and temperature and showed how this might be done using the velocity
modified temperature concept of MacGregor and Fisher (1946) which combines
strain-rate and temperature in a single parameter. This approach is considered in
detail in later chapters. Assuming a linear relation between 1: and velocity-m~dified
temperature to represent the tool-chip interface frictional conditions, Fent<?n and 5~1
Oxley (1969-1970) made predictions of <1>, Fe, FT and A. (which it will be noted is now In a1
a derived result) for a very wide range of cutting conditions. Their analysis was' based char
on the experimental machining results for SAE 1112 steel obtained by Eggleston to ttl
et at. (1959) and Kobayashi and Thomsen (1960). The approach used ..;vas to n"ot"
determine the work material flow stress parameters cr 1, nand 1: from a small number assu
of machining test results using the machining theory in reverse and then to; apply beca
these in making predictions for a much wider range of cutting conditions. Working in the
this way they predicted results which were in good agreement with experirPental zOOt
results over very wide ranges of a (5 to 35), U (0.03 to 300 mlmin) and tl (Q.05 to free:
0.254 mm). In particular, their predicted results showed in agreement with ~xperi the 1
mental results that <I> increased with increase in a, tl and U and that Fe aind FT mat,
decreased with increase in a and with increase in U. A. was predicted to decrease with ZOOt
increase in tl and with decrease in a and the experimental results largely confirmed 'qui,
this. " " as tl
alIo'
io.6.
resu

mak
toge
The
und
witt
theory [Ch.4 "

:ity direction Fc was predicted to

7 '-'..lten to increase rapidly with
oting that up to 300 mlmin
q, which reduced the area of the
Ie product of k AB and the area of
speeds although q, stilI increased
as more than compensated by the
strain-rate, .and Fs , and as a
y increased. Fenton and Oxley
suggestion made by a number of
to a dramatic fall in Fc. They
tions particularly those resulting I

e.
)tigation used the parallel-sided
i11vestigation of the influence of
"""'_JO"JIla...._ ......_ ..

results of Merchant (1945). They
<imately 60% from the lowest to
iation could only be explained if
atures in the plastic zone at the
c should be expressed in terms of
night be done using the velocity
I Fisher (1946) which combines
, This approach is considered in
between 1: and velocity-modified
rictional conditions, Fenton and
.d A (which it will be noted is now
ditions. Their analysis was based
1. P steel obtained by Eggleston
~e approach used was to
t.. " and 1: from a small number
}ry in reverse and then to apply
Jf cutting conditions. Working in
,d agreement with experimental
,03 to 300 mlmin) and tl (0.05 to
lowed in agreement with experi-
1, tl and U and that Fc and FT
Awas predicted to decrease with
mental results largely confirmed
.'\'fo',vp,n"r'on and Oxley (1969-1970) largely overcame the above diffic~lties by
nde:tormed chip thickness on the size of the chip formation zone and the str1n-rates
and scale on the strain-rate in the chip
oolvmlfl the parallel-sided shear zone model (Fig. 4.1) described in the previous
, it was assumed that the thickness of the shear zone lls2 would be prop,ortional
length I of this zone. In other words that lls2 would depend on scale bnly and
"nught be expected on cutting speed. The main investigations on which this
is based (Kececioglu 1958, Nakayama 1959) are open to criticism
of the methods used in measuring the size of the shear zone. For example, in
comprehensive experiments, which are those made by Kececioglu, this
was measured by using a mechanical, spring-activated, 'quick-stop' device to
the chip sections which were then mounted, polished and etched. In this way
bOlmdary between the shear zone (deformed material) and work (und~formed
," could be picked out reasonably well but the boundary between the shear
and chip (deformed material) was far more difficult to identify. Also, the
, device used allOWed cutting to continue (or a distance of the sathe order
undelform chip thickness during the retraction of the tool which cohld have
changes to occur duJ;ing stopping. Kececioglu's results show some ~ecrease
with increase in cutting speed but in view of the method of measuring lls2 this
must be treated with caution. '
orthogonal machining experiments using an explosive 'quick-stop' device
with a printed grid to measure the deformation in the chip fomiatibn zone.
used the results obtained to determine the inftueqce of cutting spbed and
this zone. A detailed description of this investigation is now given. !
 ,
66 The influence of speed and scale [Ch.5

5.2 EXPERIMENT
The explosive 'quick-stop' device used in the experiments was basically the same'as
that developed by Hastings (1967) and referred to in Chapter 3. In this a block which
holds the tool is located in vertical guides and rests on a shear pin. On detonation the
explosion exerts pressure through a piston on the top of the tool-block:. The stress on
the shear pin therefore builds up and at a certain stress the shear pin falls and the tool
is accelerated away from the work. Hastings had originally used hardened steel shear
pins but Stevenson and Oxley found that cast iron shear pins resul.ted in a more
consistent type of failure in the shear pin and more consistency in thd time between
firing the explosive charge and failure of the pin. Hastings measured the average
acceleration of the tool block in the initial 13 mm of its movement using a high speed
cine camera, 'obtaining a: value of '15 K lOS rnls'Z. If this accelerati~n is uniform
throughout the movement then at a cutting speed of 240 mlmin the icutting action
would only continue for a distance of about 0.005 mm during the r~lative decele
ration of the tool and work. However, measurement of the tool-bloc~ acceleration
over the first 0.05 mm by Stevenson and Oxley using a wire strain gauge glued
between the body of the 'quiCk-stop' device and the tool-block gave an acceleration
of 2.5 x lOS m/s2 and hence a deceleration distance of 0.03 mm at 240 mlmin. That
this was approximately correct was shown by slight changes in the undeformed chip
thickness over this distance in specimens obtained at this speed. No figure for
acceleration over the initial travel distance is available for Kececioglu'~ device. Cine
picture measurements, however, show the acceleration of the explosiVie device to be
about ten times greater. Also, as Hastings has shown, the explosive deVice causes far
less distortion of the chip. A flow field which is typical of those obtained by Stevenson
and Oxley is given in Fig~ 3.3. It can be seen that the chip thickness reniains uniform,
the rake angle is maintained and the undeformed chip thickness is uniform up to
approximately 0.025 mm before the end of the cut.
The work material used in the experiments was a resulphurized lov;, carbon steel,
CS 1114 (Australian standard) of chemical composition 0.13% C, 1.4'Yo Mn, 0.25% .
S, 0.019% P. This was chosen because it gave a continuous chip, with very little or no
built-up edge over a wide range of cutting conditions. Discs of this material were
parted off from bright bar of 0.11 m diameter and turned down to 0.1'0 m diameter
and fine ground to a width of approximately 3.2 mm. In a test three s~ch discs were
clamped tightly together on a mandrel, in order to obtain approximately plane strain
conditions on the centre disc. The experimental machining arrangem~nt is given in .
Fig. 5.1; With this the diameter of the discs was reduced by a straight-edged cutting
tool which was fed radially inwards. The cutting edge was set normal to the cutting
and feed directions. The cutting tools used were brazed carbide recessing tools with
rake angles of 10, 20 and 30 and a clearance angle of6. The cutting fa~e was lapped . ~
on a diamond lapping machine before each test. The experiments co,'vered a wide ~i
range of cutting speeds (5 to 250 mlmin) and undeformed chip thicknesses (0.127 to ..\
0.274 mm) with the latter equal to the radial feed measured in mm per ,evolution of:
the workpiece. The lathe used in the tests was a Churchill-Denhams SR-lOV I
with a ",...
35 h:p. variable speed motor and spindle speeds infinitely variable from 15 to 1500 :
rev/min. . .
The centre disc was polished on one side and thinly plated with !copper. The
copper was then etched to produce a grid using the photographic proc~ss described .'
d scale [Ch.5 Analysis of experimental flow fields 67

i.. s was basically the same'as

>

1 Chapter 3. In this a block which grid area

10 a shear pin. On detonation the
p of the tool-block. The stress on
!ss the shear pin fails and the tool
ginally used hardened steel shear
n shear pins resulted in a more
consistency in the time between
Hastings measured the average
its movement using a high speed
If this acceleration is uniform
of 240 mlmin the cutting action
mm during the relative decele centre disc:
mandrel
nt of the tool-block acceleration
asing a wire strain gauge glued
! tool-block gave an acceleration
of 0.03 mm at 240 n1Imin. That
t specimen discs
lathe chuck

Fig, 5.1 - Experimental machining arrangement.

changes in the undeformed chip
~d at this speed. No figure for
Jle for Kececioglu's device. Cine
tion of the explosive device to be
[1, the explosive device causes far
II of those obtained by Stevenson
chip thickness remains uniform,
...... thickness is uniform up to
1 resulphurized low carbon steel,
ition 0.13% C, 1.4% Mn, 0.25%
inuous chip, with very little or no
ons. Discs of this material were
turned down to 0.10 m diameter
,',LCl<.UUi!:>'" (1967). The radial grid master from which the grids were printed on the
cirineIls was prepared in the following way. A segment of the grid was drawn up in
times larger than the required size. This segment was then reduced tdone-fifth
a lithographic camera, five identical segments being made. These were then
together to form about 70 of arc. This composite was then reducefi to one
size on a microfilm camera, making a total reduction of 50. The depth of grid
. ' obtained in this way was approximately 9.5 mm. The resulting iinage, on
, was printed on a glass plate. To this plate was glued a 10catiQg spigot,
with the grid to allow location on the specimen for contact printi*g. It was
.",.",t1P',.,.r1 most convenient not to make a complete grid circle on the spedmen but
-cplllcentrate attention on the 70 segment to get good quality. This req~ired that
. . . . . I~ +~~, device would have to be 'fired' at a rotational position whjch, after

n. In a test three such discs were
btain approximately plane strain
lchining arrangement is given in
luced by a straight-edged cutting
Ige was set normal to the cutting
lzedcarbide recessing tools with
)f 6. The cutting face was lapped
:be experiments covered a wide
)rmed chip thicknesses (0.127 to
,easured in mm per revolution of
urchill-Denhams SR-IOV with a
finitely variable from 15 to 1500
delay involved, would cause the cutting action to be stopped aboutihalf-way
the grid. A photo-cell was used to sense the rotational positicin of the
:t:!j;~;DJ.~m for this purpose. To control the total reduction in disc diameter before
iiff.:'v_"tnT,n;na' a mitroswitch actUated by the in~feed was provided in series with the
Do<){ocel:l. The 9.5 mm depth of grid segment allowed about four te$ts to be
pellform:ed on each specimen, with about 10 revolutions of cutting for eacn test.
ANALYSIS OF EXPERIMENTAL FLOW FIELDS
,. .o.J'LU'JU~;Jl the grid boundaries of the experimental flow fields obtained, suqh as that
in Fig. 3.3, are actually arcs and radial lines it can be seen that when analysing

thinly plated with copper. The

photographic process described
68 The influence of speed and scale [Ch.

the small area of the deformation zone it is reasonable to assume the arcs to b
straight lines and the sides of the "squares" to be equal. Likewise it is reasonable tl
treat the process as if it were straight line cutting with the cuttibg velocity constan
across the undeformed chip thickness and equal to the disc circutnference multipliet
by the number of revolutions per minute of the workpiece. Fig. 5.2 gives a tracing 0

I
0.254 mm

boundary of start of
observable plas tic flow

Fig. 5.2- Tracing of deformed grid given in Fig. 3.3: cr = 20"; U = 105 mlIrtin; t, = 0.264 mm.

the deformed grid in Fig. 3.3 with the grid lines which were originally parallel to the
work velocity represented by smooth curves which average o~t the irregularities
actually occurring in these lines in the chip formation zone and chfP. Fig. 5.2 is typical
of the traced flow fields obtained from the experiments and the analysis of the
deformation will be described by reference to it.
In principle the strain-rates in the chip formation zone cati be determined by
measuring the velocities of flow and hence the velocity gradient~ from experimental
flow fields such as that in Fig. 5.2. However, as with the slipline field analyses
described in Chapter 3 it was found that the velocities could no~ be measured with
sufficient accuracy for this purpose. In view of this an appr~ximate method of
calculating strain-rates based on the parallel-sided shear zd,ne model of chip
formatidn(Fig. 4.1) was developed. With this method the direction of maximum
shear strain-rate in the chip formation zone is assumed to be c6nstant as with the
parallel-sided shear zone model and is determined in the sa~e way as with this '
model. Referring to Fig. 5.2, a straight line AB has been drawn wpich passes through
(approximately) the intersections of the streamlines in the work! and chip (assumed
parallel to the tool cutting face) and this is taken as the directionl of maximum shear
strain-rate within the chip formation zone. The velocity diagramiin Fig. 2.1 can now
be used to represent the velocity changes in the chip formation *one (Fig. 5.2) with .'
lie [Ch.5
Ie to assume the arcs to be
T ~wise it is reasonable to
. -, Jtting velocity constant
,. ,rcumference multiplied
ce. Fig. 5.2 gives a tracing of
O.25~~~
boundary of start of
observable plastic flow
1 = 105 mlmin; II = 0.264 mm.
)riginally parallel to the
out the irregularities
ne aliLl chip. Fig. 5.2 is typical
~nts and the analysis of the
zone can be determined by
gradients from experimental
:h the slipline field analyses
could not be measured with
an approximate method of
shear zone model of chip
d the direction of maximum
:d to be constant as with the
I the same way as with this
I drawn which passes through
the work and chip (assumed
direction of maximum shear
r diagram in Fig. 2.1 can now
rmation zone (Fig. 5.2) with
Analysis of experimental How fields 69
the total change in shear velocity, i.e. in the direction AB, in the zone. Wit~ the
. plane (Fig. 2.1) and parallel-sided shear zone (Fig. 4.1) models the velocity
(Fig. 2.1) is automatically satisfied but this is not generally true fot the
... ""nt~(l chip and work velocities and angle <p (Le. the angle between AB ana the
) in Fig. 5.2. Because of errors in construction and measurement~ and
from the idealised model used, e.g. some chip curl, the velocity dia~ram
satisfy both measured velocities and <p simultaneously but this can easily be
by changing anyone of these parameters by a small amount (less !than
For most of the cutting conditions considered the velocity diagram giving the
with the velocities in the chip formation zone measured from the defOlTmed
found to be that defined by <p and the chip velocity V together with the given
a. The velocity diagram defined in this way (Fig. 5.3) was therefore bsed
analysis.
==u
Fig. 5.3 - Velocity diagram used in strain-rate analysis.
the centre streamline in the chip formation zone (Fig. 5.2), lines
to AB representing directions of maximum shear strain-rate have peen
at approximately equal time intervals along the streamline (i.e. throug~ the
of the grid) and hence approximately equal distances As apart, where l6s is
",""UH:;U normal to AB. The change in shear velocity AVs (i.e. in the direction AB)
an element of thickness As can be represented as shown by the broken lines in
. diagram given in Fig. 5.3 where ~ is the angle turned through in reaching
element, measured from the work velocity, and A~ is the further angle tutned
.' in crossing the element. The mean value of the maximum shear strain-dte in
" ..;:,un'ilL is given by
AVs
'Ymax = As (5.1)
,
the geometry of Fig. 5.3, AVs can be expressed in terms of the chip velocity V,
angie <p and the angles ~ and A~. Doing this and substituting in equation (5.1)
70 .The influence of speed and scale [eb.5

Vcos (<p - 0:) sin (~~)

(5.2)
1ma" = &-sin(<p+~)sin(<p+~+~~)

The strain occurring across the element ~1 can be found by multiplying .y max by the
time a particle takes to flow across the element, which gives

~ = sin (~13)
(5.3)
1 sin(<p + ~+ ~13)sin(<p +~)

Equations (5.2) and (5.3) are now sufficient to calculate from an experimental
flow field the strain-ratesand strains in the chip formation zone.

5.4 RESULTS AND DISCUSSION

In making calculations the procedure was as follows. A transparency of an experi
mental flow field was projected onto a drawing board to give an approximately 300
times magnification. The actual magnification in each case was found by measuring
an interval on the undeformed grid with a toolmaker's microscope. The deformed
grid boundaries were drawn to give the streamlines and points along them represent
ing approximately equal time intervals. The velocity of the chip Vwas found by the
change in lengths of the grids along the streamlines and the angl~ <p found from the
construction in Fig. 5.2. For equal time intervals (i.e. equal distances &- measured
normal to AB) along a streamline the angles 13 and ~13 were measured. These were
then used with the corresponding values of V, <p, 0: and &- to de~ermine Ymax from
equation (5.2) and ~1 from equation (5.3). With the larger values of undeformed
chip thickness, Il and ~13 were measured using three streamlines near the centre of
the chip formation zone, these all giving approximately the same results, but for
smaller values of undeformed chip thickness this had to be redttced to one or two
streamlines near the centre ofthe zone. .
The distributions of maximum shear strain-rate in the chip forination zone found
in this way for four different cutting speeds are given in Fig. 5.4. :These results have
been plotted taking AB (Fig. 5.2) as the origin of the horizontal scale with the units of
this scale equal to increments of &-. In the tests <p did not vary &reatly and tls was
therefore very nearly constant (&- = 0.023 mm) in all tests. The! strain-rate can be
seen to increase with increase in cutting speed (Fig. 5.4) as would ~e expected and to
have a maximum value which occurs in the region defined by the plane AB (Fig. 5.2).
The maximum values of 1max are plotted against shear velocity l(s (found from the
velocity diagram constructed from the measured values of V and <p and the given
value of 0:) in Fig. 5.5. Shear velocity is used as it might be expe~ted that this is the
velocity to which Ym,!'X is most likely to be related. The results sho~ (Fig. 5.5}that the .
maximum value of "( max increases with increase in Vs but that th~ undeformed chip
thickness also influences strain-rate, a decrease in 11 increasing tM strain-rate. This is
also shown in Fig. 5.6 where the experimental distributions of Ymm.: are given for two
values of 11 , In Fig. 5.7 the maximum values otYmax have been pl~tted against VslIl
and it can be seen that the results are less scattered than whe~ Vs is used as the ;
:ale [Ch.5 Results and discussion 71

"; 20
(5.2) til

'"0
..
-.
18
16
nd by multiplying Ymax by the .)GO
"e
gives
-Ql 11;
c 12
~
I
.~ 10
(5.3)
-~

til
8
.6
Iculate from an experimental

.tion zone. 'I;

2
0
-12 -10 -8 -6 -4 -2 2 4 6 8
A transparency of an experi AS
to give an approximately 300
case was found by measuring distance travelled through zone, t.;.s fO.023mm
's microscope. The deformed
j points along them represent
,- Pistribution of Yrna. in chip formation zone shOwing influence of cutting speed: ex =
20"; 1, = 0.264 mm .
f the chip V was found by the
Id the angle <I> found from the
equal distances !ls measured
3 were measured. These were
nd !ls to determine Ymax from
lar~er values of undeformed
rtlines near the centre of
;:; same results, but for -U! 60 b. 0
I to oe reduced to one or two '"s;! 50
-. )( 0
the chip formation zone found .)GO
e 40
in Fig. 5.4. These results have
E 30
b.
0
)rizontal scale with the units of
.S
::1
20
... ~ D
d not vary greatly and !ls was x
c
1 tests. The strain-rate can be E 10
J) as would be expected and to
led by the plane AB (Fig. 5.2). o 30 60 90 120 150 180 210 240
Ir velocity Vs (found from the
shear velocitYJ Vs f(m min-I)
ues of V and <I> and the given
~t be expected that this is the - Variation of maximum Ymax values with shear velocity: 6., 0, 0 represent results fbr
results show (Fig. 5.5) that the' e
.10"aOO t1 = 0.127, 0.175, 0.274 mmrespectively;.,., represent results for ex = 20"aitd
but that the undeformed chip -- the same values of t 1
creasing the strain-rate.This is
I tions of Ymax are given for two
,ave been plotted against Vsltl
t than when Vs is used as the
72 The influence of speed and scale (Ch.5

-; 36
III I I I TI
PI

I 1tl ::0.127
:2 mm
--.>< 32
.
e I I
.;>co 28
~~ I I
24
e I
I I
c:
. 20 I I
iii 16
I
....
0
..c.
Q}
12
III
8
E
:::I
.~ 4
x
0
E 0
-8 -6 -4 -2 2 4 6 8
AB
distance travelled through zone, t:.s 10.023mfn

Fig. 5.6 - Distribution of Ymax in chip formation zone showing influence of und*formed chip
thickness: (l = 10; U = 120 mimin.

70
I
... VI 60
:2 50
--.><

..! 40
..
E 30
:::I
E
x 20
0
E 10

o 2 4 6 8 10 12 14 16 18 20 22 24 i
~/103s-1 (
t1 a

Fig. 5.7 - Variation ofmaximumymax values with shear velocity divided by und~formed chip
thickness: symbols same as in FIg. 5.5.
d scale rCh.5 Results and discussion 73

" ',...' horizontal scale (Fig. 5.5). From the results in Fig. 5.7 the line of best fit passing
~"'through the origin is
mm

(5.4)

'the units of VSlt1 are s -1.

Figs 5.4 and 5.6 itcan be seen that plastic deformation starts well in advance
5.2) and continues well beyond it. The chip formation zone is therefore
thickness with the thickness appearing to change only sli~tly with
(Fig. 5.4) but significantly with undeformed chip thickness (fig. 5.6) .
1.274 analysis of all of their experimental re;sults Stevenson and Oxley
.......(11.J... "

concluded that the average thickness of the chip formatibn zone

\.... over its centre portion changed little with speed but a great dbal with
'9-~ chip thickness, thus supporting the assumption of the parallel-sided
theory that &2 (Fig. 4.1) would depend on scale but not on speed. The
468
! do in fact show that the ratio of the length of the chip formation zope to its
, is approximately constant for the range of cutting conditions coilsidered
one) 6.s /0.023mm
average value of this ratio equal to 2.6 which can be compared to the values
owing influence of undeformed chip 8 and 15 obtained by Kececioglu (1958) and the constant val~e of 10
m1min. in the parallel-sided shear zone analysis. Although supporting the ~ssump
" U&2 is constant the results (Figs. 5.4 and 5.6) clearly show that the
is not constant in the chip formation zone as it is assumed to be in the
~kSlOe:a shear zone theory. ,
strain distributions obtained by Stevenson and Oxley:confirm
:UUIPtllon made in the parallel-sided shear zone analysis that half the total strain
in the chip formation zone has occurred by the time material reaJhes AB
o The results also show that the values of total strain in the chip fol:mation
," from the experimental Bow fields are close to the values det~rmined
~auati(m (4.10) using the corresponding experimental values of <I> and the given
t:.
results to those in Figs. 5.4 and 5.6 have been obtained by Goriani and
(1967) and Crookall and Richardson (1969). These, however, show that
increases towards the cutting edge which is consistent Jrith the
in curvature of streamlines near the cutting edge shown in ~jg. 5.2.
J"VJl>,(U"", Oxley (1969-:-1970) accepted thatthe strain-rate would be ~e~ter near

and referred to the work of Enahoro and Oxley (1966) whlchlshowed

statgnlihcJD point of the Bow existed at the cutting edge (see Chapter 3) which
16 18 20 22 24
"u~""'''','.. extremely high strain-rates in this region. In spite of t~s they
that their results, such as those in Figs. 5.4 and 5.6, would be represent
. the Bow over most of the chip formation zone.
velocity divided by undeformed chip
Fig. 5.5.
 Thl
cut

6 kAl
tha
equ
is do
equ
Work material properties: the influence or: the
strain-rate and temperature dkh
J
effe
straJ

,
6.1 FLOW STRESS DATA DETERMINED FROM MACHINING TEST;
RESULTS lade
The strain-rate distributions in the chip formation zone given in Figs. 5.4 and 5.6 cast the s
doubt on the method of calculating crl and n ,used by Oxley and Stevenso~ (1967) resul
which is based on the parallel-sided shear zone model and is described in section 4.4. tem~
In this the shear flow stress k AB and the change of shear flow stress across tpe shear the tl
zone Me are assumed to correspond to the average shear strain-rate in the sh~ar zone errOl
given by equation (4.1). This is clearly in conflict with the results in Figs. 5.4 and 5.6
which show that the strain-rate in the region of AB (Fig. 5.2) is much greate~ than at
the entry and exit boundaries of the chip formation zone. To allow fOF this in
calculating cr 1 and n Stevenson and Oxley (1971}-1971) made an analysis similar to the
parallel sided shear zone analysis but in which the strain-rate on AB (Fig. p.2) was
determined from an equation of the form of equation (5.4) instead of from equation whict
(4.1). This analysis is now described. :
AB as established in Fig. 5.2 is assumed to be a direction of maxim4m shear
strain-rate and hence maximum shear stress with the shear strain-rate constant along
ABandgiven' by' ,

where
byeql
(6.1)
t Anir
aneq
tbatt
where, for the results in Chapter 5, C' = 2.59. It is implicit in using this equa:tion that varial
the maximum values ofy max in Figs. 5.4 and 5.6 are assumed to occur at AB ~hich for expel
histOI
the given results appears reasonable. The shear strain is also assumed to be!constant seriOl
along AB and, from the results described in Chapter 5, to be given by mainl

~ influence of
MACmNING TEST
e given in Figs. 5.4 and 5.6 cast
lOxley and Stevenson (1967)
and is described in section 4.4.
:ar flow stress across the shear
?- -"rain-rate in the shear zone
.ults in Figs. 5.4 and 5.6
Ie, . I is much greater than at
'n zone. To allow for this in
made an analysis similar to the
ain-rate on AB (Fig. 5.2) was
(5.4) instead of from equation
direction of maximum shear
hear strain-rate constant along
(6.1)
.Iicit in using this equation that
umed to occur at AB which for
is also assumed to be constant
5, to be given by
Flow stress data determined from mac;hining test results 75
1 cos <X
(6.2)
2 sin 4> cos (4) - <X)
is again assumed to be perfectly sharp with AB transmitting the ~esultant
force. Possible temperature variations along AB are again neglected land k =
" assumed constant along AB. Following these assumptions a similar analysis to
,;",,,....,,,n out in Chapter 4 for the parallel-sided shear zone model, whic~ yielded
(4.7), can be made. In the new analysis the hydrostatic stress at A (Fig. 5.2)
'<;;lilUll."'.... in the same way as for the parallel-sided shear zone model, ile. from
. However, the rate ofchange of k normal to AB can now be related to
,., strain-rate at AB and it is possible to replace Ak/As2 in equation!(4.7) by
determined in the following way. .,
general k will vary with strain, strain-rate and temperature and if history
" are neglected so that k = f( Y,1, 1)t where y, 1 and T are the maximu\m shear
, maximum shear strain-rate and temperature then
(6.3)
~etern[lining dk/ds 2 at AB from equation (6.3) for substitution in equation (4.7)
term can be neglected as strain-rate can, on the basis of expeiimental
:OU.i1111-1,Q.U;;
(Figs. 5.4 and 5.6), be assumed to pass through a maximum at 413. The
ioerature term is also neglected because of the difficulties involved in cal~lating
:telmpercltulre gradient across AB but as will be seen this should not lead to large
Therefore
dk dkdy dt
(6.4)
ds2 = dy dt ds2
t is time. If now the flow stress properties of the work material are repxiesented
'Jb<;;qu,<1u',JU (4.14), that is 0' = 0'1" then at AB
An implicit assumption in the treatment of material flow stress properties throughout this Jok is that
an equation of state exists for the materials considered. That is, for a particular material it i~assumed
that the flow streSs will be unique for given values of strain, strain-rate and temperature. ThUs possible
;:variatioru resulting from the paths by which the strain, strain-rate and temperature are reached, which
;,ex:pelimc~nts show can be significant, are neglected. It might reasonably be expected that negl~ct ofsuch
effects, which at this stage would appear prohibitively difficult to allow for, will not ~ntroduce
,,,,~,~,,:-.... errors particularly in the case of the steady-state machining process to which attclntion has
been limited.
 -'1>'<'

~:'

76 The inOuence of strain-rate and temperature [eh.6

Se(

or dkJdr = nkA.FJrAB (6.5) eXf

wo:
where r AB is given by equation (6.2). The second term on the right-hand side of obs
equation (6.4) is simply the shear strain-rate given by equation (6.1) and t~e third vah
term is ,the reciprocal of the velocity normal to AB which from equations (2.3) can be
taken as '

dtl ds2 = II U sin <\> : (6.6)

,
Substituting from equations (6.1), (6.2), (6.5) and (6.6) in equation (6.4) and noting
from equations (2.3) rbar -

Vs = U cos wcos (<\> - (X)

the following relation is obtained:

(6.7)

and substituting this for WLls2 in equation (4.7) and notingthatl = tlsin <\> (4.1) gives

C'n
(6.8)
sin <\>

Experimentally determined values of 9, <\> and C' can now be used to calculate n from
equation (6.8). The values thus obtained will correspond to the actual strai~-rate at ,keepi
AB as will values of kAB determined from the resolved forces along AB. ~ further eqmll
advantage of using equation (6.8) is that in calculating 0"1 and n it is n6 longer oonvt
necessary to assume a value of strain at the boundary representing the start qf plastic '~i:hl
flow as it was in the parallel-sided shear zone analysis - see Fig. 4.4. If itf is again . lflitsi
assumed that for each test condition (each strain-rate) the derived values o~ n, kAB pu~h
and r AB define a single stress-strain curve of the form given in equation (4.~4) then caicb.I
0"1 in this equation is given by , whi~h
along
wasa.
O"AB
0" -- (6.9) exper:
I - e~
out v(
',were 1
Stevenson and Oxley (197~1971) have calculated 0"1 and n in the ab6ve way :;::'; ,In
using the machining test results considered in the previous chapter wliich also eq~ati
included cutting forces measured during the tests. They had intended a,t first to thatin
calculate 0"1 and n using values of <\> and strain-rate found from the 'quick-stop' flow (6.8) II
fields together with the forces measured at the instant of 'quick-stop'. How~ver, the given t
variations in these measured cutting forces, strain-rates, etc., caused *artly by
Itlperature [Ch.6
(6.5)
. on the right-hand side of
y equation (6.1) and the third
ch from equations (2.3) can be
(6.6)
5) in equation (6.4) and noting
(6.7)
oting that l = t/sin <\> (4.7) gives
(6.8)
1 "used to calculate n from
<- the actual strain-rate at
ed forces along AB. A further
lting cr 1 and n it is no longer
representing the start of plastic
is - see Fig. 4.4. If it is again
::) the derived values of n, kAB
1 given in equation (4.14) then
(6.9)
:d cr1 and n in the above way
previous chapter which also
fhey had intended at first to
md from the 'quick-stop' flow
of 'quick-stop'. However, the
rates, etc., caused partly by
Flow stress data determined from machining test results 77
error, but which also reflect real variability of the propertie~ of the
material within the very small test specimen (chip formation zone), tedded to
the trends. It was therefore decided to use average force and shear angle
(Figs. 6.1 and 6.2) found over relatively long cuts at a number of ispeeds
4
!II
<lJ 2
~
-
o
en
c
-::;
u
o 75
cutting speed U/(m min-')
Fig. 6.1 - Experimental cutting forces: ex = 20"; t\ =- 0.264 mm; w '" 9.5 mm.
~ and tl constant and to determine the corresponding strain-rateS from
(6.1). The forces were measured using a dynamometer which \vas a
,~"", ... t'''u.'> arrangement of semiconductor strain gauges bonded to a tube integral
the explosive 'quick-stop' tool holder. The dynamometer was calibrated: while
operating position on the lathe with dead weights acting through lever arms and
rods which bore directly on the tool edge. The shear angle valuesl were
from equation (2.11) using experimentally measured chip thickness values
were determined for each cutting condition from about 20 measurelnents
a representative length of chip. Further smoothing of the experimental rbults
by working with values obtained from the curves drawn throu~ the
points in Figs. 6.1 and 6.2. In this way Stevenson and Oxley ave~aged '
variations in work material properties and effectively obtained results which
representative of larger specimens. '
,In their calculations Stevenson and Oxley determined 9, Fs and kAB!from
,,,",,''v,,,, (4.12), (4.13) and (4.5) using experimental values of <\>. Fc and FT' (Note
in determining kAB from equation (4.5) l was taken equal to t 1/sin<\>.) Equ~tions
and (6.9) were then used to find n and cr 1 with the corresponding straintrates
by equation (6.1) with C' taken as 2.59. Values of cr1 and n obtained in thi~ way
78 The inftuence of strain-rate and temperature {Ch.6 Sec. 6.:

01 30 E
(IJ E
-..
~ .r.
9 .r.~

~ 25 0,
(IJ c: 8
0, ~ a
c:
a
Ci(IJ
20 h to -
E
u
CL
:::e:
-..
\:) 7(
c:
.r.
!II
0
u
0.5
51

o 4.
cutting speed, U/(m min-1)
.. 28
Fig. 6.2 - Experimental shear angles and tool-ehip contaCt lengths: IX = 100; t1 = 0.264 mm.

are shown plotted against strain-rate in Fig. 6.3, where &=&AB =.y dY3. Also givetit Ii
. in Fig. 6.3 are values of 0'1 and n, for the same material as used by Stevenson and
Oxley, determined from slow speed compression tests and calculated as above fronil
machining results obtained at very low cutting speeds. The values of cr 1 and n found
from the machining and compression tests can be seen to fit together very well and
show similar trends with strain-rate to the results in Fig. 4.7. The new results should
be far more reliable and do, indeed, show far less scatter than the earlier ones (Fig~
4.7) owing to the improvements made in both the experiments and analysis and alsa
I
because the same material was used for the complete range of tests. As before it waS
possible to represent 0'1 and n by single curves over the entire range of strain-rat~
although, as will be seen, the temperature in the tests (along AB in the machining
tests)varied from room temperature to approximately 200C. However, if willb~
noted that there is a dip and subsequent rise in the 0- 1 curve between strain-rates of
102ls and lQ4/s and in explaining this it is found that consideration must be given t6
temperature as well as strain-rate. Methods for calculating temperatures in machin~
ing are now considered. . !

6.2 TEMPERATURE CALCULATION METHODS ,

Heat is generated in machining by the work done in" the plastic zone in which the chi~
is fonned and by friction at the tool-chip interface. In this connection it should b~
noted that for the shear plane model the total work rate FeU is equal to FsVs + Pv;
and that for the assumptions made this also applies to the parallel-sided shear zon$
model and to the chip fonnationmodel given in Fig. 5.2 with FsVs the work rate i~
the chip formation zone and FV the work rate at the tool-chipinterface. If the tool i$
I
Iture rCh.6 Temperature calculation methods 79

E 0.3
E
--
..c
..c~
1

c
"En
c
x~
Q)
~ "0

to --
u
~ 0.2
.5
01
c C
0
u c
Q)

0.5
-
"0
C!
..c
I

--
0.1 C!

VI

0: 10"'; tl = O.264mm.

~B = YAJV3. Also given

) used by Stevenson and
:alculated as above from
values of CT1 and n found
it together very well and .
'. The new results should
'~ earlier ones (Fig.
j analysis and also
of tests. As before it was
ltire range of strain-rate
-ng AB in the machining
)0e. However, it will be
! between strain-rates of ','
eration must be given to
temperatures in machin
tic zone in which the chip
; connection it should be
:Vis equal to FsVs + FV
parallel-sided shear zone
ith F sV s the work rate in
lip interface. If the tool is
strain -rate, E/s-I
6.3 Values of nand 0', calculated from machining and compression test results.
-------J sharp, e.g. if wear has occurred on the clearance face, then addition~l
be generated by friction between the wear scar and the newly machined
have been a number of attempts to calculate temperatures in machining
shear plane model. In these, and in fact in all of the temperature calculatidn
to be considered, it is assumed that all ofthe work-done in chip formation is
into heat with only a negligible amount of energy being retained within tHe
metal. Experimental work supporting this assumption has been reportdd
and Quinney (1934, 1937) and by Bever et ai. (1953). Hahn (195t)
the shear plane temperature by assuming the shear plane to be a uniform.
moving obliquely through an infinite workpiece. Leone (1954) and
Shaw (1954) also assumed the shear plane to be a uniform band sourqe
that it moved over a semi-infinite workpiece with the proportions ~f
heat entering the work and chip determined by applying Blok's (1938)
principle. Trigger and Chao (1951) and Lowen and Shaw (1954) both used
partition principle to calculate the average tool-chip interface temperatur:e
that the heat developed at the tool-chip interface was uniformly distri
. . Weiner (1955) obtained a solution for the shear plane temperature distrib~
by assuming that the chip velocity was perpendicular to the shear plane and that
i
I
80 The influence or strain-rate and temperature
Sec
the heat conduction in the directions of motion of the work and chip couM be
neglected. Using relaxation techniques Rapier (1954) calculated temperature distri
He
butions in the work, chip and tool which he treated as three separate systems. also
assumed a constant shear plane temperature and a plane uniform heat source ~t the
tool-chip interface with all the interface heat flowing into the chip and none into the
tool. Dutt and Brewer (1964) improved the analysis by treating the work, chi~ and
tool as one system, but after making some approximations found that they wer~ able
to dispense with the tool region altogether. In this way they were able to detetmine .~
the proportions of shear plane heat entering the work and the chip, an~ the
proportions of tool-chip interface heat entering the chip and tool. Chao and TJ)'igger
(1955) improved their earlier analytical solution for the interface tempetiature
distribution by allowing the fraction of interface heat flowing into the tool td vary "
along the interface although still assuming a uniform heat source. In their ca)cula
tions they made use of an iterative analytical procedure involving a grid of rel:ll and
fictitious point heat sources. ;
The chief disadvantages of the above methods of temperature calculation result
mainly from the simplifications made in the shear plane model. This assumes a
velocity discontinuity across the shear plane while in real materials as clearly s~own
in Chapters 3 and 5 the transition from the work to chip velocity occurs gradually
over a finite plastic zone. Also, the velocity of the chip material adjacent to the
tool-chip interface is less than the chip velocity, resulting in the charact~ristic (196
deformation usually observed in this region. It has been assumed in all the icases zon~

considered that the generated heat is confined uniformly within the shear plane and tern]
at the tool-chip interface rather than being spread over finite plastic zones, and as tern)
shown by Boothroyd (1963) this will certainly result in the temperatures peing Sha\
overestimated. Methods of calculating temperatures which are based on Bpoth solu'
royd's work and which have been widely applied by the au thor and his co-workers are bed.
now described. ' basi~

The temperature rise in the plastic zone in which the chip is formed is found by
considering the plastic work done in this zone and is given by

(!6.1O)
The
aver:
where p is the density of the work material, S its specific heat and Pis the propqrtion
of heat conducted into the work. There have been a number of attempts to pre9ict P
theoretically including that by Weiner (1955) based on the shear plane model of chip
formation. Weiner's results are represented by the line in Fig. 6.4 which sh~ws P
plotted against RT tan <I> where RT is a non-dimensional thermal number given by wher
, the f.
(6.11) aver;
flow

with K the thermal conductivity of the work material. Experimental valuesofip are
also given in Fig. 6.4. These were obtained by Nakayama (1956) using a thermo
couple technique to measure the heat carried away by the work and by Boot~royd
I perature Temperature calculation methods 81

he work and chip could be

,,' 'ted temperature distri
:'- yarate systems. He also
Ie uniform heat SOurce at the
to the chip and none into the
, treating the work, chip and
IDS found that they were able
they were able to determine
ork and the chip, and the
) and tooL Chao and Trigger
the interface temperature
flowing into the tool to vary
leat source. In their calcula 30
~ involving a grid of real and Rrtan <ll

mperature calculation result
the~)retlcal
ane model. This assumes a
:d materials as clearly shown
lip velocity occurs gradually
lip material adjacent to the
mlting in the characteristic
en assumed in all the cases
y within the shear plane and
r finite plastic zones, and as
in the temperatures being
Nhich are based on Booth
ufhr,r and his co-workers are
t ..... lIp is formed is found by
'en by
(6.10)
further limits that ~ should not exceed 1 or be less than 0are also imposed~ The
heat and ~ is the proportion
lber of attempts to predict ~
Ie shear plane model of chip
~ in Fig. 6.4 which shows ~
thermal number given by
fact that not all of the plastic work of chip formation has occurred at AB.; The
(6.11)
:xperimental values of ~ are
Ima (1956) using a thermo
:he work and by Boothroyd
6.4 - Theoretical and experimental results for ~: line represents Weiner's (1955~
results; 0 represents Nakayma's (1956) experimental results; T, ., represent
(1963) experimental results, + represents results calculated by Tay et aL (1974~
using finite element method.
. who measured temperature distributions in the work, chip, deformktion
and tool using an infrared radiation method and calculated ~ froni the
distribution in the work. (Details of the different methods of measUring
in machining have been given by Boothroyd (1965), Trent (1977) and
1984).) From the experimental results in Fig. 6.4, it can be seen that Weiner's
underestimates ~. In view ofthis ~ has usually been estimated in the work to
""''''V,",''' from the following empirical equations which have been derived oh the
the experimental results in Fig. 6.4:
~ 0.5-0.35 19 (RTtan <\ for 0.04 ~ RT tan <\> ~ 10.0
(0.12)
~ = 0.3--0.15Ig(RT tan<\ for RT tan <\> > 10.0
temperature along AB (Fig. 5.2) is taken to be given by .
(0.13)
Tw is the initial work temperature and 11 (0 < 1] :::;; 1) is a factor which alloJs for
temperature at the tool-chip interface which will help determine the av~rage
stress at the interface is assumed to be given by
(6.14)
 .t~.
2i
.4...J4!Qi!.

.82 The influence or strain-rate and temperature [Ch.6

where dTM is the maximum temperature rise in the chip which occurs at the interface
and 'I' (0 <: 'I' ~ 1) is a factor which accounts for the possible variation oftemperature
along the interface. Boothroyd (1963) has calculated 6.TM using numerical methods
for both triangular (maximum thickness at the cutting edge) and rectangulctr plastic
zones (heat sources) at the interface with no sliding at the interface and hence with all
of the frictional work dissipated in these zones. The results of these calcula~ions are
represented by the lines in Fig. 6.5. The experimental results given in Fig. 6.5 for the

t
t
c
s
n
s
11
n
Sl
v.
Sl
t~
sl
0.1L--L~ __ ~-L~L-~-L __L--L~ c(
o 0.5 2.0 2.5 H
re
ju
k,
Fig. 6.5 - Calculated and experimental tool-cbip interface temperature results: fulliline A
represents numerical results assuming rectangular plastic zone and broken line repreSents
numerical results assuming triangular plastic zone; symbols represent experimental res~lts. fit
d~
m.
"
ratio of ATM to 6.Tc , the average temperature rise in the chip, were obtained by m]
Boothroyd from his experimentally measured temperature distributions. Ii can be sp
syen that the calculations based on the rectangular plastic zone give the bes~ fit with re:
experiment. This model has therefore been favoured in the work to be descIibed. If fit
ot
the thickness of the rectangular plastic zone is taken as 2 , where 0 is the ratio of this tel
thickness to the chip thickness, then, as shown by Stevenson (1970), Boo~hroyd's cal
calculated results for this case (full line in Fig. 6.5) can be representedl by the by
I
equatIon ~.

(15
Ig (!~:) = 0.06-0.1950 (Rt2rl2 +0.51g (R~t2) (6.15)
mil
ten
net
i thi:
where 6.Tcis the average temperature rise in the chip which is assumed to be given by ass
perature reh.6
occurs at the interface
t.,. ..riation of temperature
r M using numerical methods
,,.,,.,'.....:.;Tn""
:dge) and rectangular plastic
~ interface and hence with all
ults of these calculations are
:sults given in Fig. 6.5 for the
2.5
tr;....pcrature results: full line
e and broken line represents
)resent experimental results.
the chip, were obtained by
ture distributions. It can be
:ic zone give the best fit with
the work to be described. If
t 2 , where b is the ratio of this
venson (1970), Boothroyd's
can be represented by the
R~t2) (6.15)
ich is assumed to be given by
Temperature calculation methods 83
(6.16)
tool-chip contact length.
of the restrictive assumptions associated with the temperature calculation
reported so far were overcome by Tay et al. (1974) who applied tae finite
"nj~mf'trlnt1 to determine temperature distributions using the experimer1tal flow
which the distributions of velocity and strain-rate could be found, and
forces obtained by Stevenson and Oxley (1969-1970) as a basis] In this
"'mnf'l~tll"f' distributions were obtained numerically by solving th~ steady
~enslonal energy equation using the finite element method with the distribu
sources in the chip formation and tool-chip interface plastic zones
from the strain-rate and flQw. stress distributions in4hese zones. 1lhe flow
used were those obtained by Stevenson and Oxley (1970-1971, 19'i3) from
test results. In considering the interface, account was taken of ~oth the
at the interface and the deformation in the plastic zone adjacen~ to the
with the associated velocities and strain-rates measured from the !experi
fields. In the calculations flow stress was taken to be a function of strain,
and temperature. The chip, work and tool were treated as one system,
taken of theactual shape and size of the tool, and material prqperties
specific heat and thermal conductivity were considered as func~ions of
. Results were obtained for three experimental flow fields ana these
that the temperature along AB, with AB found as in Fig. 5.2, was fairly
for most of its length but increased rapidly near to the cutting edge.
, as pointed out by Tay et aI. (1974), the strain-rate also increases in this
this would to some extent compensate for the temperature rise ahd help
assumption, made in the machining analysis in tlle previous secti~n, that
is"constant along AB. The results also allow the temperature gradient normal to
be determined and hence an indication to be given of the error introduced by
. the temperature term in equation (6.3) when obtaining an expression for
the conditions considered the temperature term is about 25% the
'I'.U.....,,'''' of the strain term. Neglect of this term is, however, less important than it
at first appear to be. This follows from the fact that in any test such as a high
, compression test, the results of which might be used to compare machining
. with, there will also be a temperature-time gradient similar to that in
~rullllOlg.This would therefore to some extent compensate for neglect; of the
nperaltUlre term in the machining analysis. TaY'et al. (1974) also used their r~sults to
the proportion of heatconducted into the work. These values are indicated
crosses in Fig. 6.4 and can be seen to fit in well with the experimentalrdults for
i
I
et al. (1976) noted the laborious nature of the analysis carried out by T~y et al.
and the substantial computation time involved and concluded that iIi deter
temperatures for a wide range of cUtting conditions or in deterptining
;.........tnr...' for use with a predictive machining theory a simpler meth~d was
which did not need an experimental flow field to start with. They a~hieved
using a model in which the streamlines in the chip formation zon~ were
to be hyperbolic with the velocities along streamlines given by the diagram
84 The influence of strain-rate and temperature [Ch.6

in Fig. 5.3. The hyperbola used was of a form which gave strain-rate distributions
similar to those found experimentally in Figs. 5.4 and 5.6 with the strain-rate passing 120e
through a maximum at AB. The constant in the streamline equation was selected so
that the corresponding strain-rates at AB were consistent with those given by
looe
equation (6.1). In this way Tay et al. (1976) were able to determine the strain-rate
distribution in the chip formation zone and neglecting variations in flow stress within u
0
the zone assumed that the heat generation at any point would be proportional to the -... SO(
strain-rate at that point. In modelling the tool-chip interface flow account was taken c:
,..;.
of both sliding and plastic deformation. The sliding velocity was given by an equation m
based on measurements made from experimental flow fields. The interface plastic .....<1: 601
zone was taken to extend over the full contact length and to be triangular in shape
with maximum thickness at the cutting edge. The shear stress was assumed to be
constant along the interface. For the chip formatien modeldeveloped in this way it 40
was shown that the velocities and strain-rates could be determined for given values of
U and a: if q, (Fig. 5.2) was known. Temperatures could then be calculated given the 20
cuttingforces from which the values of average flow stress in the chip formation zone
and at the interface, needed in the calculations, could be determined. The method of
determining temperature distributions using the finite element method was essen
tiallythe same as that used byTay eta/. (1974). Account was again taken ofthe shape
and thermal properties of the tool and of the influence of temperature on the work
material thermal properties. In the calculations the thickness of the chip formation
zone and the maximum thickness of the tookhip interface plastic zone had to be
known and these were taken as 112.6 and t.J20 respectively on the basis of measure
ments made from experimental flow fields.
Tay et al. (1976) used their method to calculate temperature distributions from
experimental cutting force and shear angle data obtained when machining S1016
steel (0.19% C, 0.16% Mn, 0.016% P, 0.027% S) using a Sandvik S6 grade carbide
tool. In the calculations it was necessary to know the tool--chipcontact length. This
was found from the equation 40

t1 sin e 20
h = ~~~~----~ (6.17)
sin q, cos (9 + a: q,)

using experimental values of <I> and e. This equation is derived by drawing a line
through point A in Fig. 4.1 parallel to R and assuming that no force is transmitted
above the point where the line cuts the tool face, i.e. -that the distance from the 1"'-"""";;:'4';;'';'.'"
cutting edge to this point is equal to h. Average values of TAB and Tint obtained by
numerical integration of temperatures along AB and the interface found byTay et al.
are given in Fig. 6.6. Also given in Fig. 6.6 are the values ofT AB and Tint calculated, .,." "

..~~~~.;.
using the same experimental machining results and work material thermal proper , ,,:

~..:~;---;
ties, from equations (6.13) and (6.14) taking the temperature factors 11 and 'JI as ::;
unity. In these calculations 0 in equation (6.15) was taken as 0.05 which meant that' " '']~:1listributior
the thickness of the rectangular tookhip interface plastic zone was assumed equal to ; ,c "~,sured expe'
the maximum thickness of the triangular plastic zone used by Tay et al. The tool--chip i co' ."this the cha
contact length was again calculated from equation (6.17). Stevenson et al. (1983) : with tempe
using a similar approach to that of Tay et al. (1976) calculated temperature;
ure [eh.6 Temperature calculation methoc:ls 85

train-rate distributions 1200.-~--~----.---~----~--~----~~

rain-rate passing (0) _---0
1-~ was selected so ./"- - - Tint

///~+-
t with those given by 1000
termine the strain-rate
)Os in flow stress within
I be proportional to the 800
/ :+
flow account was taken 0/

/
as given by an equation
600 /
>. The interface plastiG ~
be triangular in shape
ess was assumed to be
400 __ - - 0 - - - - _ _ __
,eveloped in this way it --- --- -0
____ ~B
I

ined for given values of ~-------+---- !

be calculated given the 200
he chip formation zone
rmined. The method of
~nt method was essen
200 300 400 500 600 700 800
gain taken of the shape cutting speed J Uf(m min -1)
nperature on the work
1000.-~---r----~---.----r----.----~~
~ of the chip fonnation
(b )
plastic zone had to be
1 the basis of measure ~---a:

-
-------
_--- Tint
ture distributions from
vhen machining S1016
v' :1_ S6 grade carbide
.tact length. This ~
3Cl 400
_o~
-+
-- --
0-

..
+

OJ
E ...qr=
---0----------0
.!!! 200 + +- TAB
(6.17)
OL-i-__ ~ ____ ~ __ ~L_ _ __ L_ _ _ _L __ _ ~ __ ~

ved by drawing a line 200 300 400 500 600 700 800
[l0 force is transmitted cutting speed} U/(mmin-1) ';.. "

the distance from the

Fig. 6.6 - Comparison of temperatures calculated from equations (6.13) and (6.14) with those

.B and Tint obtained by

,Calculated by Tay et at. (1916); broken lines represent results given by equations (6.13) and

face found by Tay et al. (6.14) with 11 'I' = 1 and full lines represent results obtained by Tay etat.: (a) a:: = 5", tl =

AB and Tint calculated, O.264mm; (b) a:: = 10", tl O.127mm.

lterial thermal proper

Ire factors T( and '11 as
0.05 which meant that ::i'li~:trilhl1tir",.,
and compared their results with tool temperature distributions mea
e was assumed equal to experimentally using the technique developed by Wright and Trent (1973). in . .!
ray et al. The tool-chip the change in temper of the high speed steel tool used is measured and correlatJd
;tevenson et at. (1983) temperature. Good agreement was shown between calculated and experimenthl
dculated temperature
86 The influence of strain-rate and temperature [Ch.6

results. If on this basis the results ofTay etai. in Fig. 6.6 are accepted as re.~listic then
the use ofthe factors 11 and", in equations (6.13) and (6.14) to give bettere~timates of
TAB and Tint is clearly justified. From the results in Fig. 6.6 it can be seen tat 11 and '"
would need to have values falling in the range 0.7 to 0.95 in order to make ~he values
of TAB and Tint found from equations (6.13) and (6.14) equal to the values found by
Tay et ai.

6.3 THERMAL PROPERTIES

In the applications of the machining theory described in the following chapters
attention is mainly limited to plain carbon steel work materials and in 9alculating
temperatur~s the appropriate ten1perature~dependent thermal properties pave been
determined in the f9U9wmg.w.ay-, Th~ iufl~c:mce of carbon content on specific heat S is
found to be small and from data given by Woolman and Mottram (1964) th~ equation

S/(J/(kgK = 420+0.504Tf'C (6.18)

can be used for all of the steels considered. However, there is a marked influence of
carbon content on thermal conductivity Kand allowance must be made fQr this and
the influence of other elements on K. Hastings (1975) has shown how this can be
achieved using experimental results for thermal conductivity given by Woolman and
Mottram (1964). In this approach two equations are used which are derived from the
experimental results. In the first equation the thermal conductivity iat OOC is
expressed in terms of chemical composition. This is then used with the second
equation which relates the variation in thermal conductivity to tempedture. The
relevant equations found from the data of Woolman and Mottram are

Ko = 1/(5.8 + 1.6[C] + 4.1[Si] + 1.4[Mn] + 5[P] + [Ni] + 0.6[Cr] +

+ 0.6[Mo]) I (6.19)

and

"" K = 418.68[0.065 + (Ko - 0.065)(1.0033-11.095 x 10- 4 1)] (6.20)

where Ko is the thermal conductivity at OC, K is the thermal conductivity at rc with

the units of thermal conductivity, W/(mK), and [C], [Si], etc., are the perdmtagesof
the various elements presenqn the work material considered. Equations ~6.19) and
(6.20) yield equations of the form .

KI(W/(mK = 54.17-0.0298Trc (6.21)

I
i
which is for a steel of chemical composition 0.20% C, 0.15% Si, @.015%S,
0.72% Mn, 0.015% Ai, and
perature [Ch.6 Interpretation of flow stress data 87

are accepted as realistic then K/(W/(m K = 52.61 - 0.0281 Tf'C (6.22)

I " give better estimates of
:an be seen that 11 and",
; in order to make the values
equal to the values found by
a steel of chemical composition 0.38% C, 0.1 % Si,0.77% Mn,0.015% P.
temperature calculations the effect of temperature on work mkterial
which will be ne~gibly small, has been ignored and p has been taken as
f{)r all of the steels considered.

d in the following chapters ATION OF FLOW STRESS DATA IN TERMS OF STRAIN

materials and in calculating AND TEMPERATURE
hennal properties have been TAB calculated by Stevenson and Oxley (1970-1971) from equations '(6. 10)
n content on specific heat S is . ,using the same experimental machining results (Figs. 6.1 and 6.2) as Jsed in
.1ottram (1964) the equation 01 and n in Fig. 6.3, are given in Fig. 6.7. In these calculations \11 was

(6.18)

tere is a marked influence of

;e must be made for this and
has shown how this can be
tivity given by Woolman and
d which are derived from the 3Q
mal conductivity at O"C is :1,
then used with the second
....
250
.ctivity to temperature. The
1_

,
I
-CIl
ld Mottram are 20! ow<f.

,~
I'J < Li-li] + 0.6[Cr] + 15 . E,
(6.19) c:
10 '0
i:
11\

5
(6.20)

150 225 300

mal conductivity at roc with
,etc., are the percentages of cutting speed J U/(m min- 1 )
dered. Equations (6.19) and
Fig. 6.7 - Calculated values of temperature and strain-rate.

(6.21) unity and Twas 22C. An iterative calculation procedure was necessarY as S
were both assumed to vary with temperature as described in the prejvious
1% C, 0.15% Si, 0.015% S, '-. and to have values corresponding to a temperature T = TAS' Also gi~en in
'6,7 are values of EAB where AB = 1Ar/V3 with 1AS found from equation: (6. 1)
 88 The influence of strain-rate and temperature [Ch.6
Sec.6.{
taking C' = 2.59. From the results in Fig. 6.7 it can be seen that TAll rises rapidly with
in the IT
speed at slow cutting speeds. This is because ~, the proportion of heat conducted intd "
purpose
the work, is also changing rapidly. At higher speeds, the temperature rise tends to
In Fi/
become adiabatic with little heat conducted into the work and T A.S become~
approximately constant. The strain-rate EAB increases approximately linearly with!
cutting speed over the entire range. Following Stevenson and Oxley it is now possible'
to explain the dip and subsequent rise in the 0" 1 curve in Fig. 6.3 in terms of thesd
temperature and strain-rate results by referring to suitable flow stress data.
Some experimental results of Ohmori and Yoshinaga (1966,1%8) showing th~
variation of lower yield stress with temperature and strain-rate for a mild steel of:
similar carbon content (0.14%) and static lower yield stress (182 MPa) to those used:
in Stevenson and Oxley's tests are given in Fig. 6.8. These results were obtained from!

/"
/'
/'
560 /"
21C /'
.7
/'
0 /'
a.. " 90 /'
~ /'
-..
0
c 420
..
200o~--

---
III
III
Q)
Z 350
III

-
:9 300s;..._
.~ 280
>,
to..
Q)
s:
.
210

140
10 10 2 10 3 100 300
strain-rate, temperature, i/oC

Fig. 6.8 Experimental lower yield stress results.

high-speed tension tests and cover temperatures up to 300C (these are starting
temperatures and no account has been taken of the small temperature rise in the
tests). The tension tests went to strain-rates of up to 3.7 x 102/s, which is somewhat
lower than the minimum value in the conventional speed machining tests (Fig. 6.3).
From the results in Fig. 6.8 it can be seen that, for the ranges of strain-rate and
temperature considered, the yield stress increases linearly with the logarithm of the
strain-rate and decreases approximately linearly with temperature. The broken lines
in Fig. 6.8 are extrapolations of the tension test results to strain-rates of the order met
Iperature [Ch.6
en that TAS rises rapidly with
,,' 'n of heat conducted into
mperature rise tends to
lework and T AS becomes
approximately linearly with
n and Oxley it is now possible
in Fig. 6.3 in terms of these
.ble flow stress data.
ga (1966, 1968) showing the
train-rate for a mild steel of
ress (182 MPa) to those used
ie results were obtained from
o 100 200
temperature} T1C
s results.
to 300C (these are starting
mall temperature rise in the
7 x 102 /s, which is somewhat
:d machining tests (Fig. 6.3).
he ranges of strain-rate and
.rly with the logarithm of the
mperature. The broken lines
I strain-rates ofthe order met
Interpretation of How stress data 89
machining tests. The right-hand side of Fig. 6.8 was constructed for the
of interpolation to any temperature between OC and 300C.
Fig. 6.9 a curve of lower yield stress has been constructed from the data:in Fig.
420
&:
~
-..
~
0'0
IJI~
IJI
...
CII
Ui
~
53 280
strain- rate} ifs- 1
Fig. 6.9 Comparison of machining and tension test results.
the values of T AS andeAsgiven in Fig. 6.7. This curve is of the same fprmas
cr1 in Fig. 6.3 and in particular shows the same dip and subsequent rise in
However, the cr1 curve is higher than the lower yield stress curve, whicIl is not
as the strain corresponding to cr1 (e = 1) is far greater than the sttain at
. To make a more direct comparison it was assumed that the value~ of cr1
6.3) found from the machining test results could be used to represent the
curve for the corresponding values of temperature and strain-rate iover a
range to allow equation (4.14), i.e. cr = cr1sn , to be used to find the stress at
near to that at initial yield. (It should be noted that in deterritining O't from
and n using equation (6.9) it has only been necessary to assume that
(4.14) held over a small strain range as eAB was of the order ofl,the\strain
teSDolllOlngto 0'1') A curve of stress found in this way for a strain of 0.01 is gi~en in
and can be seen to be of the same general shape as the 0'1 curve butlmuch
actual stress values to the lower yield stress curve. The differences [in the
I!rcun,es in Fig. 6.9 could reflect real differences in materials but might ~o be
!,lamera by limitations in the methods used in calculating stresses. In this connec
linear extrapolation used in Fig. 6.8 (broken lines) should. be treate9 with
caution as results for wider ranges of strain-rate, e.g. those in Fig~ 6.10
by Campbell and Ferguson (1970) from dynamic shear tests on EN 38 mild
show that at higher strain-rates the rate of increase of flow stress with the
of the strain-rate increases markedly. (It should be noted that the reshlts in
.7 also show this effect.) This would mean that the values of stress ob~ained
the extrapolated results would be underestimates. Also, the methbd of
I
90 The inftuence of strain-rate and temperature [Ch;6
Sec. 6.4

400 occurs..
constan
c
-
:::E
....
VI~
VI
300
again a:
tempen
rapidly
effect.
~
VI
Furt
. stress w
i3
(J)
200 .' machini
.c: .interfac
VI

"0
materia:
Qj iilterfac1
>.
Jollowin
~

(J) . '"withno
~
.S!
0
10- 4

shear strain -rate,i / 5-'

Fig. 6.10 Experimental lower yield shear stress results: chain dotted line represents stress
values corresponding to values of TAB and tAB in Fig. 6.7.

calculating TAB in which 11 was taken as unity is open to question as this assumes tl'l.at
all the plastic work of chip formation has occurred at AB. In this way TAB would hlive
been overestimated and hence the stresses'found from Fig. 6.8 would be underesti
mated. Ferguson in written discussion to the paper by Stevenson and Oxley
(1970-1971).used the results in Fig. 6.10 to construct the curve (shown chain dotted)
corresponding to the values of TAB and &AB in Fig. 6.7. It caube seen that this cuwe
(Fig. 6.10) again shows a dip and subsequent rise in stress similar to that observed in
the <11 curve. ,
In the light of the above results the following qualitative explanation for the
inftuence of cutting speed on the <11 curve in Fig. 6.3 can now be given. At low cutting
speeds the increase in stress caused by the increase in strain-rate resulting from ian
increase in cutting speed predominates over the decrease in stress caused by tlbe
corresponding,rise.intemperature (thermal softening). This is because the logaritfim
of the strain-rate, with which stress increases linearly (Fig. 6.8), changes rapidly with
increase in cutting speed at low cutting speeds. For the same increase in cutting spded
at higher cutting speeds, the increase in the logarithm of the strain-rate is less a6d
consequently the rise in stress resulting from the increase in strain-rate is reduced. iAt
a certain stage the decrease in stress caused by thermal softening can become of tbe
same O1::der, or even greater, than the increase in stress caused by the increasel in
strain-rate. For the results in Fig. 6.3 thermal softening clearly predominates o~er
strain-rate hardening at a strain-rate&AB "'" 2.0 x 103/s which corresponds to acutttng
speed of about 23 mlmin and it is here that the lowest point of the dip in the <11 cutve
Iture [Ch.6
:49U"LU'.n
:ted line represents stress
:jg.6.7.
stion as this assumes that
this way TAB would have
(- 0 would be underesti
~nson and Oxley
~~ ~ l ;>llOWn chain dotted)
m be seen that this curve
milar to that observed in
tive explanation for the
I be given. At low cutting
m-rate resulting from an
. in stress caused by the
;is because the logarithm
,.8), changes rapidly with
increase in cutting speed
he strain-rate is less and
,train-rate is reduced. At
ening can become of the
aused by the increase in
early predominates over
I corresponds to a cutting
of the dip in the 0"1 curve
Interpretation of How stress data
As cutting speed is increased further , TAB (Fig. 6.7) becomes approximat~ly
while SAB continues to rise. The strain-rate effect therefore takes over OIice
. . the 0"1 curve starts to rise. It is also at about the point where the
.. becomes approximately constant that n (Fig. 6.3) starts to decre~se
. . and it therefore appears that this is more a strain-rate than a temperattire
results showing the influence of strain-rate and temperature on fl~w
obtained by Stevenson and Oxley (1970-1971) from their experimen'tal
results (Figs. 6.1 and 6.2) by considering the flow at the tool-c~ip
To do this they assumed that a plastic state of stress existed in the chip
adjacent to the tool-chip interface over the full contact length with ~he
a direction of maximum shear stress and maximum shear strain-rate and,
Boothroyd (1963), that the plastic zone at the interface was rectangular
sliding at the interface. From these assumptions it follows that the shear flbw
kint and maximum shear strain-rate at the interface are given by
(6.23)
(6.24)
I
the values of interface temperature Tint can be calculated from equatiQns
to (6.16). Stevenson and Oxley calculated k int , .y int and Tint (Fig. 6.7) frlm
. equations using the smoothed experimental results in Figs. 6.1 and 6.2
the experimental values of tool-chip contact length h (Fig. 6.2) which w~re
from the length of the wear scar on the tool cutting face. In the calculations
ratio of the thickness of the interface plastic zone to the chip thickness tz , Was
to be 0.125. The values of F, 12 and V which were needed in determining kint,
found from the equations
F = Fe sin (X + FTcos (X (6f5)
i
I
(6.i6)
,
"/ I
..equations (2.3) using experimental values of <1>, Fe and FT together with the
cutting conditions. In calculating Tint an iterative procedure was required a~ S
K were temperature dependent with in this case the temperature T determining
I
I
 _ .. .i
,fT
92 The influence of strain-rate and temperature (Ch.6
,
Sand K taken as the mean chip temperature given by Tc = T w + ~ Tst. + ~ Tc. The
temperature factor'll in equation (6.14) was taken as unity. i
Values of flow stress O'int (O'in! = y3k ind found by Stevenson and Oxl~y in this way
together with values of 0'1 taken from Fig. 6.3 are given in Fig. 6.11. Tolcompare the
i

980

910

o 840
CL
~I:> 770 \
"
::f 700 "
'\
....
Q)

1ii 630
~
a 560
1,90

I, 20

350L---~------~------~------~-------7
300 1,00 500 600 700
velocity modified temperature, Tmod/K i
Fig. 6.11- Values of cr 1 and crin, calculated from machining test results.~

two sets of results, which from Fig. 6.7 can be seen to be for differentitemperature
ranges, and to combine the effects of strain-rate and temperatur~ in a single
parameter the results have been plotted against velocity-modified temprature. This
parameter was first introduced by MacGregor and Fisher (1946) and as fnentioned in
section 4.5 was used by Fenton and Oxley (1968-1969) in their machinibg analysis in
considering the variations in shear stress at the tool-chip interface wi~h strain-rate
and temperature. As used in machining the velocity-modified temperafure Tmod can
be expressed as i

Tmod = 1{1- V Ig(IsQ)l (6.27)

where T(K) and S (s -1) are the testing temperature and strain-rate an4 v and Eo are
constants for a given material and range of testing conditions. It is assumed that for a
given strain the flow stress for a particular material will be a unique fun~ion of Tmod
defined in this way. In calculating Tmod for the machining results So w~ taken as lis
erature
::: = Tw + ATsz + ATe The
, and Oxley in this way
n FIg. 6.11. To compare the
hining test results.
'e tor different temperature
d temperature in a single
modified temperature. This
. (1946) and as mentioned in
1 their machining analysis in
ip interface with strain-rate
lified temperature Tmod can
(6.27)
strain-rate and v and Eo are
ions. It is assumed that for a
Ie a unique function of T mod
Ig results Eo was taken as lis
Interpretation of flow stress data 93
v was selected so that two points of equal 0'1 in Fig. 6.3 had the same Tmod raIue
. .gave v equal to .0.09. For ~he conditions on AB, T in equation (6.27) was taken
+Z73"C with g equal to gAB while for the tool-chip interface the values! used
i+ 273C and Eint (Eint = .y inIV3). \
two sets of results (Fig. 6.11) can be seen to fit together quite well, following
over the entire range of conditions. This is at first somewhat surprising
strains at the interface will be significantly greater than one, the ~train
to 0'1' The average shear strain at the interface is in fact !siven
by the ratio hI'Ot2 which noting that h = t2 gives a shear strain of about
a')i'Uu..",u by Stevenson and Oxley, '0 is equal to 0.125. In reality the ~train
adjacent to the interface can be many times greater than this as shown
of ~hip sections and printed grid experimental flow fieldsl The
of infl\lence of strain 00. the (J'int values in Fig. 6.11 is, how~ver,
results obtained in materials testing in general which show !that,
of the order of one, strain has little influence on flow stress. For Imost
is true for all testing conditions; it is particularly true for the, high
and strain-rates encountered at the interface. :
. in Fig. 6.11 show a decrease in stress with increase in veldlcity
. temperature, as would be expected, up to a modified temperature of ~bout
above this value, and up to a modified temperature of just over 600 K, the
t'TP'''''''''' with increase in modified temperature. This effect is well known from
of both high and low speed tension and compression tests on mild stetls
example, Manjoine (1944), Ohmori and Yoshinaga (1966) and Tanak~ and
(1967) - and is attributed to dynamic strain ageing or as it is sometimes
blue-brittleness. Manjoine's results show that as strain-rate is increased the
at which dynamic strain ageing occurs is also increased. Notini this
and Oxley (1970-1971) showed that the interface strain-rates and terbper
at which dynamic strain ageing was observed to occur in machining rere
with those observed by other workers for dynamic s~rain ageing conditions
speed tension and compression tests. Stevenson and Oxley (1973) ilater
nn,oh'.,f'",rl how dynamic strain ageing conditions could also be achieved itt the
f"n'...... <.h,-,'n zone by pre-heating the work to a sufficiently high temperature .
be concluded from the work described so far in this chapter that a car~fully
machining test can provide an effective method for measuring a matenal's
properties. In fact for the high strains (1 and over), strain-rates (lbs to
and temperatures (200 to lOOOC) encountered in machining it is diffi~t to
of a more suitable testing method. The philosophy of obtaining a work
flow stress properties from relatively few machining test results andithen
these to make predictions of cutting forces etc. over a much wider range of
lQ1tlOnS as proposed, for example, by Fenton and Oxley (1969-1970), therefore
a sound one. However, the approach of using machining results to prbdict
'-.!.J1UUJ'/:; results clearly invites accusations of tail chasing. It would be far tiftter
the viewpoint of attempting to verify a predictive machining theory if this Gould
using flow stress properties obtained from an independent test. The\next
considers some of the few such data available which are suitable fOJ this
i
94 The influence of strain-rate and temperature (Ch.6

6.S FLOW STRESS DATA DETERMINED FROM HIGH SPEED

COMPRESSION TEST RESULTS
Oyane et al. (1967) and Oyane (1973) have obtained flow stress data froniI high speed
compression tests for a range of plain carbon steels which are suitable for use in
making machining predictions. This is achieved by obtaining values df 0"1 and n,
which through equation (4.14) are taken as defining the stress-strai~ curve of a
material for a given strain-rate and temperature, from the compression test results
and then expressing these as functions of velocity-modified temperature!. Machining
predictions were first made in this way by Hastings et aI. (1974) and Oxley and
Hastings (1976). The method of determining 0"1 and n as functions: of velocity
modified temperature is now described. i
In the compression tests the strain-rate (= 450/s) was lower than that usually
encountered in machining (l()1to lfJ'/s) but the testing temperatures co~ered a wide
range (0 to 1100C), which makes it possible to extrapolate the comiJ,ression test
results into the machining range by using the velocity-modified temperature para
meter given in equation (6.27). This parameter has been used many timbs before to
represent the opposing effects of strain-rate and temperature but ~ormally in
situations where temperature is the main variable and there are rela:tively small
variations in strain-rate with the strain-rates not very high. In machining,: strain-rates
are very hjgh and both temperature and strain-rate vary over large ianges. The
compression test results cannot be used to test the validity of equation (6J27) for such
conditions because the strain-rate in these tests was' constant. However, the results of
Campbell and Ferguson (1970) given in Fig. 6.10 which show the effects 6f very wide
ranges of strain-rate and temperature on the. lower yield point of a 0.12% plain
carbon steel are suitable for this purpose. LoWer yield stress pointstak~n from the
curves in Fig. 6.10 at strain-rates of 1, 10, 102 , 103 , 104 and 3.2 x 104/s ate replotted
eo
against Tmod' with as before v = 0.09 and = lIs, in Fig_~.12and can b9 seen to fit a
single curve exceptionally well, thus giving support to the use of equation (6.27) for
conditions where strain-rate and temperature vary over large ranges and are of the
same order of magnitude as in machining. I

Oyane et al. (1967) and Oyane (1973) have given results for plain c~bon steels
with carbon contents in the range 0.16 to 0.55% and their res~lts.for a 0.16% carbon
steel, plotted on a log-log basis, are given in Fig. 6.13. In the strain range from 0.2 to
0.4 straight lines of best fit have been drawn to approximate the cunles for each
temperature. That these lines are a good approximation (over a rest~cted strain
range) supports to some extent the use of the linear logarithmic stress against strain
relation of equation (4.14) in machining analyses. (Normally the use of this equation
has been justified on the basis of slow speed (low strain-rate) compression test
results.) A good fit is not obtained over the much wider strain range of the
compression tests, chiefly because of the near adiabatic conditions in thd tests which
cause variations in flow stress with the resulting temperature rise in additlon to those
caused by strain-hardening. In deriVing the constants 0"1 and n for each test it was
therefore decided to use the straight lines drawn for the 0.2..:0.4 strain irtterval. The
value ofn was found directly from the slope and 0'1 by extrapolating the line to a
stra~n e = 1. In calculating the correspond.ing Tmo~ values the constants viand eo were
agam taken as 0.09 and lis. The stram-rate E was constant at 45t1J/s and the
temperature T (K) was taken as the test starting temperature plus J calculated
I
I,
,
emperature Flow stress data 95
1000
I HIGH SPEED

,[ress data from high speed, B C

800
s which are suitable for use in x
~

obtaining values of 0'1 and n, , d QJ

Q.. \:)

-
C'
ng the stress-strain curve of a :::E 03
)m the compression test results 600 A 01
+
~ /..-+ n ...-+-/+\
b
.dified temperature. Machining
+ 02 ,
c
VI
+ QJ
~s et al. (1974) and Oxley and VI o1 \:):
QJ
400 0
md n as functions of velocity ....11'....1 \:
x lower yield point (ao=f3 1
'+
'0 .c:
0 I'
s) was lower than that usually .3: result~ of Compbell & .S:
Ig temperatures covered a wide '
trapolate the compression test
- 0
-200 . Ferguson (1970)
'

~
e'
1ii

tty-modified temperature para

)een used many times before to
temperature but normally in o 200 400 600 800 1000
and there are relatively small velocity modified temperature! Tmod I K
high. In machining, strain-rates
e vary over large ranges. The Fig. 6.12 - Flow stress results plotted against velocity-modified temperature.
idity of equation (6.27) for such
)nstant. However, the results of ,
ch show the effects of very wide 1000.--------r----,---,--r-r-r-rTi
r yield point of a 0.12% plain
~ld stress points taken from the
04 and 3.2 x lO4/s are replotted
F' ".12 and can be seen to fit a
Ise of equation (6.27) for d
)Vet-large ranges and are of the a..
:E
b
n results for plain carbon steels
their results for a 0.16% carbon VI
VI
L In the strain range from 0.2 to
........
QJ

'proximate the curves for each 11)

nation (over a restricted strain 200

3:
logarithmic stress against strain ~
ormally the use ofthis equation
v strain-rate) compression test
lch wider strain range of the lOOL-------~--~---L--~~~~
,tic conditions in the tests which a' 02 03 04 05 0607 '0
lerature rise in addition to those strain, e:
.ts 0'1 and n for each test it was
the 0.2-0.4 strain interval. The , Fig. 6.13 High speed compression test results for a 0.16% plain carbon steel: 0,0"(::;.0
1 by extrapolating the line to a
lOO"C; x, 200"C; A, 400"C; 0, 500"C; 6., 600"C; 'V, 100"C; 0, 800'C; +, 900"C; 'Y, lOOO"C;
+ , llOO"C. I
tIues the constants v and 80 were
vas constant at 450/s and the
temperature plus a calculated
96 The influence of strainrate and temperature , [Ch.6

temperature rise appropriate to the mid-point of the 0.2-0.4 strain interval. The
temperature rise was calculated by assuming that all of the plastic work was
converted to heat and that the conditions were adiabatic. The rise!1T (K) ~as then
given by

!1T = -
1 fo. 3
O'dE (6.28)
pS
In determining !1T the density p was taken as 7862 kglm3 and S was calculated from
equation (6.18) with T (0C) in this equation taken as the test starting temp~rature.
The results for 0'1 and nfoLthe 0.16% carbon steel are plotted against Tm~d in Fig.
6.12 and can be seen tQc-Sbowca.-clear..d:y.namic. strain ageing (blue-brittle ~ region, F
where flow stress increases with increase in temperature. This effect is not eVident in
the results of Campbell and Ferguson (1970), which are also given in Fig. 6.12, f
because of the very low strain (E < 0.01) associated with these results. The crves for e
0'1 and n have been represented mathematically using sections of polynoniials and
these have been used with rescaling functions to represent the 0'1 and n curv~s for the
other carbon steels. The rescaling functions, which give an increase in ct 1 and a ;;
decrease in n with increase in carbon content, were chosen to give a good fit :With the ~ 7.
experimental values of 0'1 and n for the 0.35, 0.45 and 0.55% carbon steels: derived
from the compression test results in the same way as described above. In this way Tt
continuous changes in 0'1 and n over the ranges ofvelocity-modified temperature and an
carbon content considered were represented by a relatively simple set of fUnctions. an
The method of obtaining these functions is described in Appendix A3. . an