W. R. D.

Wilson
Professor,
Mechanical and
Hydrodynamic Lubrication in
Nuclear Engineering Department,
Northwestern University,
Evanston, III. 60201
Simple Stretch Forming Processes
Theoretical models for the hydrodynamic lubrication of plane strain and
axisymmetric sheet metal stretch forming processes with cylindrical and spherical
J. J. Wang headed punches, respectively, are developed. The lubricant is treated as an
Graduate Research Assistant, isoviscous Newtonian liquid for both geometries. In addition, the influence of sheet
Mechanical Engineering Department, heating due to plastic deformation with an exponential variation of viscosity with
University of Massachusetts, temperature is analyzed for the plane strain case.
Amherst, Mass. 01003

Introduction
As in most metal forming operations, lubrication and 1(a) shows a plane-strain process where an infinitely wide
friction are of great importance in sheet metal forming sheet is deformed by a cylindrical punch, while Fig. 1(b)
operations. Effective lubrication systems prevent direct metal- shows the equivalent axisymmetric process where a circular
to-metal contact between the sheet workpiece and tooling. sheet is deformed by a spherical punch. In each case the sheet
This reduces tool wear and improves product quality by has an initial half width (radius) of w and is rigidly clamped at
eliminating surface damage as a result of adhesive material its edges and the punch has a radius r and moves normal to the
transfer from sheet to tooling. Lubrication also reduces undeformed sheet with a constant speed V. Initially the sheet
friction at the sheet/tooling interface. In sheet metal forming and punch are completely separated by a lubricant film of
friction has an important influence on the mode of the sheet local thickness h. In this state, which is shown in Fig. 2(a), the
deformation and potential for failure due to tearing. Thus it is film can be divided into two zones: a work zone surrounding
of considerable practical value to understand the mechanisms
of lubrication in sheet metal forming.
As discussed by Wilson [1] four different regimes of
lubrication are possible in sheet metal forming. Of these, the
full film regimes in which the surfaces are completely
separated by a lubricant film with a thickness much larger
than the molecular size, will displace the boundary or mixed
regimes if conditions are favorable. Although many different
aspects of sheet metal forming lubrication have been in-
vestigated [2, 3, 4], there appears to be no published analysis sheet
of full film lubrication in even the simplest stretch forming
operation. Such an analysis, as well as representing the
behavior of the process under full film conditions can be used lubricant film.
to infer under what lubrication regime a particular process
will operate which information is of great utility in improving
lubrication. a Plane-strain
The present paper seeks to develop a mathematical model
of hydrodynamic thick film lubrication in stretch forming
processes. Simple processes such as the one investigated are
not uncommon. In addition similar process geometries arise
in more complex stretch forming processes and in the punch-
sheet contact in deep drawing.

Processes to be Analyzed
The processes to be analyzed are shown in Fig. 1. Figure

Contributed by the Lubrication Division of THE AMERICAN SOCIETY OF

MECHANICAL ENGINEERS and presented at the ASME/ASLE Joint Lubrication
Conference, Hartford, Conn., October 18-20, 1983. Manuscript received by b Axisymmetric
the Lubrication Division, May 3, 1982. Paper No. 83-Lub-5. Fig. 1 Processes to be analyzed

70 / Vol. 106, JANUARY 1984 Copyright © 1984 by ASME Transactions of the ASME

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 02/19/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Assumptions Used
A number of assumptions have been introduced to render
the analyses tractable.
(1) The punch is rigid.
(2) The deformation resistance of the sheet can be
characterised by a constant engineering stress a.
(3) The initial thickness d of the sheet is small compared
work with the punch radius r.
zone (4c) In the infinitely wide case the deformation is
a Early stages homogeneous plane-strain.
(4b) In the axisymmetric case the stress state in the un-
supported area is balanced biaxial tension.
(5) The lubricant is newtonian with viscosity \i.
(6a) In the ispviscous case the viscosity fx, is constant.
(6b) In the thermoviscous case the viscosity /i at tem-
perature AT (measured relative to the initial condition)
is given by
H = H0e~aAT (1)
where /n0 is the viscosity at the initial temperature and
a is the temperature coefficient of viscosity.
(7) The lubricant is incompressible.
(8) The surfaces are smooth.
(9) The pressures generated in the inlet zone do not in-
fluence the deformation of the sheet.
(10) The pressure gradient in the work zone does not in-
fluence the lubricant flow.
Later stages
(11) The film thickness h in the work zone is small com-
Fig. 2 Zones of lubrication
pared with the punch radius r.
(12) The friction in the work zone does not influence strip
deformation.
the center of the punch where the sheet conforms to the (13) The punch temperature is constant.
punch; and an inlet zone surrounding the work zone where the (14) The sheet temperature rises due to adiabatic plastic
sheet is in unsupported deformation. As deformation heating.
proceeds the film thickness at the boundary between the inlet (15) The dominant mode of heat transfer in the lubricant
and work zones decreases to zero and with further defor- film is conduction across its thickness.
mation a portion of the work zone becomes unlubricated (16) Viscous heating within the lubricant film is negligible.
(h = 0) while the lubricated work zone persists near the center For the most common forming operations the deformation
of the punch as shown in Fig. 2(b). of the sheet is so large compared with the deformation of the

Nomenclature

D = aAT, nondimensional surface At = difference between sheet and

temperature difference punch temperatures w = initial half width of sheet
F = Tr/iilnVwio2/3d2/\ non- U = sheet surface velocity relative X = distance along punch from
dimensional friction stress to punch center line
F, = l2(eD-2eD-D2eD + l)/ = sheet surface velocity in inlet X = distance from boundary
D*(eD-\) function of D ux zone between inlet and work zones
D D
F2 = (De -2e +D + 2)/
2D(eD-\) function of D
u = coordinate
mean surface velocity in x
system
y = depth of punch penetration
Z = depression of sheet
TT h ( awd\ 2/3 = sheet velocity in inlet zone in x
ux coordinate system z = depression of sheet at bound-
H = 0.448 — ary between inlet and work
r KfiVr/
nondimensional local film u2 = coordinate
punch velocity in inlet zone in x
system
zones
a = temperature coefficient of
thickness = effective mean surface velocity viscosity
u*
V = punch velocity
Hi = 0.448^
/i, /
(-
awd\m
c = sheet specific heat
e = and
angle between local normal
punch axis
r \ fj. Vr) ' d = initial sheet thickness 0> = angle at boundary between
nondimensional inlet film h = local lubricant film thickness inlet and work zones
thickness h0 = film thickness corresponding 02 = angle at outer edge of sheet
K = aa/pc, thermal plastic to zero pressure gradient V- = lubricant viscosity
parameter hi = film thickness at boundary Mo = lubricant viscosity at zero
L = l/w, nondimensional work- between inlet and work zones pressure and punch tem-
piece half width I = current half width of sheet perature
^ = r/w, nondimensional punch P = pressure in lubricant film P = sheet density
radius r = punch radius a = sheet flow stress
T = tV/w, nondimensional time t = time T = friction stress

Journal of Tribology JANUARY 1984, Vol. 106/71

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 02/19/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 tooling that assumption (1) is justified. Assumptions (2), (3), h, /iiU\-
(4a), (4fc), and (12) simplify the plasticity problem mainly by — =2.231-T ) 0)
eliminating inhomogeneous deformation due to strain r \ ad /
localization or friction effects. Assumption (4b) has been From assumptions (1), (2), (3), and (11) the local mean
shown to be fairly accurate by Chakrabarty [5]. Assumptions velocity C/is given by
(5) and (6b) are quite realistic for a wide range of liquid
U=r'6l -(r6xl)/2l (8)
lubricants. Non-newtonian effects and pressure dependence
of viscosity are unlikely to be important at the relatively low where the angle dx defines the position of the boundary
pressures found in sheet forming operations. Compressibility between the inlet and work zones as shown in Fig. 1(a) and the
is neglected in assumption (7) for similar reasons. Assumption current half width / of the sheet are given by
(8) eliminates the influence of surface roughness on lubricant
transport investigated by Wilson and Delmolino [6] and 0, = tan" 1 ((y-r)/w) + tan~l(r/-Jy2l2ry+w2) (9)
Wilson and Farrell [7]. Assumptions (9) thru (11) are all quite
accurate and have been used in previous models of forming and
processes. Assumptions (13) and (14) neglect heat transfer
from the sheet to the punch or the surroundings. These l=rdx + \[Jr^2ry + w2 (10)
assumptions maximize the influence of temperature on
lubricant transport and are probably not too far from reality. respectively, and y is the depth of penetration of the punch.
Assumption (15) is justified because of the relatively low Since the punch speed is constant
speeds and thin lubricant films found in most stretch forming
y=Vt (11)
processes. Assumption (16) is valid provided the speeds and
lubricant viscosities are not too high. where / is the time measured from the start of deformation,
equations (9) and (10) can be differentiated to yield
V ( r(Vt-r) \
(12)
Isoviscous Plane-Strain Analysis w2 r)2 V
+(Vt-r)
The amount of lubricant carried into the work zone is and
controlled by the inlet zone. From assumptions (1) and (9) the V(Vt-r)
shape of the lubricant film in the inlet zone is decided by the l=rd, +- (13)
tooling and workpiece geometry while the pressure is decided
by hydrodynamics. respectively.
The conditions in the inlet zone are unsteady. However if a It is convenient to continue by using nondimensional
coordinate system which moves along with the boundary variables defined by
between the inlet and work zones is used, then the conditions h, / awd \ in

change only slowly with time and the problem can be treated #i=0.448-!-(—— ) (14)
r V ixVr /
as quasi-steady. Thus assumptions (5), (6a), (7), and (8) yield
the Reynolds equation L = l/w (15)
l R = r/w (16)
h dp
U{h h (2)
w^= -^ and
where h is the local film thickness, (i the lubricant viscosity, p T=tV/w (17)
the local pressure, x the distance along the film from the Thus eliminating 0 between equations (7) and (8) and non-
boundary, U the mean surface velocity in the x coordinate dimensionalizing yields
system, and h0 the film thickness where the pressure gradient
is zero. H,
From assumption (10) h 0 is given by \dTdT 1L
2L dT
dT I
K=hx (3) and equations (9), (10), (11), (12), and (13) reduce to
where hx is the film thickness at the boundary between the
= tan- 1 (7 , --K) + tan-'(.R/V7^-277? + l) (19)
inlet and work zones and from assumptions (1), (9), and (11)
the local film thickness h is given by the parabolic ap-
proximation
L=Rd,l^T2-2TR +l (20)
h = hx+x2/2r (4)
where r is the punch radius. dd R(T-r)
The boundary conditions on the inlet zone are that far out = \ ( (21)
in the inlet zone
dT 1+(T-R) 2
\ •4Tl-2TR TT)
and
/i = o o , p = 0 (5)
dL „ ddx (T-R)
and at the boundary between the inlet and work zones = R—L + (22)
dT dT V7*-277? + l
h = hi,p = ad/r (6) Substitution for 6X, L, dd/dT, and dL/dT from equations
(19), (20), (21), and (22), respectively, in equation (18) allows
where a is the sheet flow stress and d is the initial sheet the nondimensional entrained film thickness Hx to be
thickness. calculated as a function of the nondimensional time T for a
The inlet problem which has now been defined has already given value of the nondimensional punch radius R.
been solved by Blok and Van Rossum [8] for a foil bearing.
In the work zone because of assumptions (5), (6a), (7), (8),
Integration of the Reynolds' equation (2) and substitution of
and (10) the Reynolds' equation reduces to
the boundary conditions (5) and (6) yields an equation for the
entrained film thickness hx which may be written in the rt dh dU „ bh
U +h — + 2 =0 (23)
present notation as dx ax dt

72 / Vol. 106, JANUARY 1984 Transactions of the ASME

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 02/19/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 where U is the sheet velocity relative to the punch and x is the DeD-2eD+D +2
distance along the surfaces from the punch center line. As (35)
2D(eD - 1 )
described by Wilson [9] this equation can be decomposed by
Lagrange's method to yield two simultaneous ordinary and
differential equations D = aAT (36)
dx U where t7, and U2 are the surface velocities in the x coordinate
r- (24) system, a is the temperature coefficient of viscosity, and ATis
dt
the surface temperature difference.
and From assumptions (13) and (14) the temperature difference
dh h dU AT is given by
(25)
It ~2 ~dx AT=a(l-w)/pcw (37)
which imply that the lubricant film is transported at half the where a is the sheet flow stress, p its density, and c its specific
surface speed and thins out as the surface stretches. heat. Substituting for T2 - Tx from equation (37) in equation
Equations (24) and (25) can be nondimensionalized yielding (36) and nondimensionalizing
dd 6 dL
D= K(L-\) (38)
dT 2L dT (26)
where
and
K=uo/pc (39)
dH H dL
(27) Proceeding in the same manner as in the previous isoviscous
dT 1L dT
2L analysis yields a modified versus of equation (18)
respectively, where 8 defining the local position as shown in dd .„ .8,dL^ \ 2/i
Fig. 1(«) is given by H,= + (F2 - 1) (40)
( ( •
dT L dT ) ' " • ) •

6=x/r (28)
Since 8lt L, d8x/dT, and dL/dT are still given by equations
and the local nondimensional film thickness His defined by (19), (20), (21), and (22) and since Fx and F2 can be calculated
from equations (34), (35), and (38), equation (40) can be used
h ( owd\
(29) to calculate the nondimensional entrained film thickness Hx
tf=0.448-( — )
as a function of the nondimensional time T for particular
r \ fxVr/
values of the nondimensional punch radius R and thermal
The initial value of 8 in equation (26) is 0j, given by plastic parameter K.
equation (19) as a function of nondimensional time T and
If Wilson and Aggarwals' method [11] is applied in the
punch radius R. Similarly the initial value of H in equation
work zone, equations equivalent to equations (26) and (27)
(27) is Hx, given by equations (18) thru (22) as a function of T
can be derived for the thermoviscous case. These are
an&R. Thus equations (26) and (27) can be integrated to yield
values of H over the 8, T space for a given value of R. This dd /I X 6 dL
was done using a digital computer with a fourth order Runge bl (41)
dT \2 ) L dT
Kutta scheme. Further details are given by Wang [10].
and
From assumptions (6a) and (10) the local friction stress Tin
the lubricated part of the work zone is given by dH (\ \ H dL
F2 (42)
T=nU/h (30) dT ' \2 J L dT
or in nondimensional form respectively. Since F2 is given by equations (35) and (38) and
the initial values of 8\ and / / , are given by equations (19) and
Rd dL (40), respectively, a similar numerical procedure to that
F=2.23 (31)
LH dT described for the isoviscous case was used to calculate values
where the nondimensional friction stress F is of//over the 8, Tspace for given values of R and A".
From the work of Mahdavian and Wilson [12] the friction
F=Tr/ixU31/3
Vl T/l/3„2/3 J2/3 (32) stress T is the current thermoviscous problem is given by
Since L, H, and dL/dT are known, F can be calculated as a T=n0UD/h(eD-l) (43)
function of 8 and T for a given value of R.
or in nondimensional variables from equations (30), (31), and
(43)
Thermoviscous Plane Strain Analysis DR8 dL
(44)
With a lubricant whose viscosity varies with temperature (a (ld-\)LH dT
thermoviscous lubricant), the variation in viscosity across the Since all the variables in equations (44) are known, the
lubricant film thickness generally precludes the use of the
nondimensional friction stress F was readily calculated as a
conventional Reynolds' equation. However Wilson and
function of 8 and T for given values of R and K.
Aggarwal [11] have shown that under the conditions
established by assumptions (6b), (7), (8), (15), and (16) it is
possible to use a modified Reynolds equation in which the
mean surface velocity (U in the present case) is replaced with Isoviscous Axisymmetric Analysis
an effective mean surface velocity U* which is given by The geometry of the axisymmetric process is shown in Fig.
1(b). The problem is more difficult than the plane strain
U*=(U+F2(Ul-U2))/Fl (33) problem because the sheet in the inlet zone is no longer flat.
where An exact calculation of the shape of the unsupported sheet is
very complex but Chakrabarty [5] has used assumptions (1),
\2(e2D-2eD-D2eD + l) (2), (3), and (4b) to derive some approximate relationships
F,= (34)
Dl(eD-\) which form the basis of the present axisymmetric model.

Journal of Tribology JANUARY 1984, Vol. 106/73

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 02/19/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Chakrabarty found that the depression z of the un- which describe the transport and thinning of the lubricant in
supported part of the sheet surface below its edge could be the work zone must be modified to
described by the equation d8 U
z = wsin02 ln(tan(0/2)/tan(0, /2)) (45) (59)
~dt~ Tr
where 82 is the angle of slope at the sheet edge, 0, is the angle and
of slope at the boundary between the inlet and work zones,
and 0 is the angle formed by the surface normal and the axis dh h ( dU
— ({/cot0+—— (60)
of symmetry. The angles 8{ and 02 are related by ~dT 2r V 88
wsmd2=rsin2vd1 (46) respectively. By using equation (51) these can be rewritten in
nondimensional form as
At the boundary between the inlet and work zones
dd sin0 / cosfl2 \
6 = 6l,z=zl (47) (61)
~dT2 2sin02 VV COS0i
cos0j /
where from geometry
and
Zi=y-r(l-sin8l) (48)
dd Hcosd /( C0S
coscP22 A
and y is the punch penetration. Substituting this condition in (62)
dT2 sin02 V COS0i /
equation (45) yields the following relationship
The process for solving equations (61) and (62) is similar to
y = r((\ - c o s f l , )
that using for the equivalent plane strain equations (26) and
+ sub 2 0, ln(tan(0, /2)/tan(0 2 /2))) (50) (27) except that the solution is conducted in the 0 2 , 8 space
In the work zone the velocity U of the sheet relative to the rather than the T, 8 space. The initial value of 6 in equation
punch is given by Chakrabarty's analysis as (61) is 8X obtained from equation (46) while the initial value of
H in equation (62) is / / , as given by equation (58). Thus a
rsin0 / cos0<
cos02 \ dd2 fourth order Runge Kutta program was used to obtain
U= (51) solutions for H in the 02> 8 space for different values of R.
sin02 \^ cos0
cos0, / dt
These were immediately mapped into the T, 8 space by using
Since from assumption (11) the inlet zone must be relatively equation (50).
short and the surface velocity is continuous this expression
can also be used to calculate the sheet velocity £/, relative to Results and Discussion
the punch in the inlet zone by substituting d\ for 0
Perhaps the most useful results of the present analyses are
/•sin0! (/ cosd
COS02 \ dd
dd2 the solutions for lubricant film thickness since these can be
U, (52)
sin02 ^\ cos0i
cos0, / ~di
dt used to identify the regime (1) in which a particular process
As in the previous isoviscous analysis the important velocity operates. The initial film thickness at the center in a plane
in the inlet zone is the mean surface velocity U relative to the strain process is given by
boundary between the inlet and work zones given by M '
[ht],-. = 2.24 (63)
U=r8l-Ul/2 (53) olnwlnd2'1'
Differentiating equation (46) yields It is interesting to note that the film thickness is proportional
to the (ixV/a)1/3 in the present process while as described by
ddi _ cot02 d62 Wilson [1] the film thickness in steady "wedge" processes
(54)
~dt~ 2cot0, dt such as rolling is proportional to fxV/a while in "squeeze"
Substituting for 0 from equation (54) and for U] from processes such as upsetting the film thickness the film
equation (52) in equation (53) yields thickness is proportional to (^K/a) 1/3 . Process geometry
expressed through r, w, and d also has an important influence
r sm0, dd2 on the lubricant film thickness in stretch forming.
£/= (55)
2 sin02 dt While equation (63) gives the initial central film thickness in
a plane strain process, the film thickness entrained by the inlet
Chakrabartys' analysis also shows that in the inlet zone the
zone varies as the process proceeds and the sheet wraps round
meridional curvature of the sheet is equal to the punch radius
the punch. Figure 3 shows the variation of the non-
r. Thus equation (4) must be replaced with
h = h{+x2/r (56)
and when this change together with the new relationship for U
is used, equation (18) is modified to
2/3
//?sin0,
'Rsmd, dddd
2^2\-
H, =0.198 (57)
V sin0, ~dl
dt /
Differentiating equation (50) with respect to 62 and using
equation (54) to eliminate dd] /d82 yields
dT _ sin0|COt02
dd2 ~ 2cot0,

COS0 2
+ cos02 ln(tan(0 [ /2)/tan(0 2 /2)) + (58)
COS0,
04 0.6 08 1.0 1.2
Using equation (57) in conjunction with equations (46) and non-dimensional time T
(58) allows the calculation of Hx for any value of 02 and/?. Fig. 3 Film thickness entrained by the inlet zone in a plane strain
Because of the radial symmetry equations (24) and (25) process

74/Vol. 106, JANUARY 1984 Transactions of the ASME

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 02/19/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 1Di -~_^T=0.3 R=0.1
T=0.6
£08
u T=1.2
gQ6

~ ^^=1.8
gQ4
o
C \
f 0.2
•o
c \
o n 1 1 1 1 I I _ I I I \
O 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
angle 6
Fig. 4 Variation of film thickness distribution with time in a plane
strain process

1.0
Fig. 7 Influence of plastic heating on the film thickness distribution in
a plane strain process

Fig. 5 Influence of nondimensional punch radius on film thickness

distribution in a plane strain process

0.6 0.8 1.0 1.2

non-dimensional time T
Fig. 8 Comparison of entrained film thickness in equivalent plane
strain and axisymmetric processes

failure. Thin film or microasperity lubrication associated with

surface roughness or boundary lubrication associated with a
chemically active lubricant may still provide adequate
lubrication. These mechanisms are beyond the scope of the
present paper.
As the process proceeds the film thickness entrained by the
inlet zone is modified by stretching and wiping action in the
0.6 0.8 1.0 1.2
non-dimensional time T work zone. Figure 4 shows how the local nondimensional film
Fig. 6 Influence of plastic heating on entrained film thickness in a thickness H in the work zone varies with the position coor-
plane strain process dinate angle 8 for various nondimensional times T and a
particular nondimensional punch radius R = 0.1. For this
value of R the film thickness decreases with increasing 6 and
dimensional inlet film thickness Hx as a function of non- T. However this behavior is dependent on the value of R as
dimensional time T for different values of the non- shown in Fig. 5 which again plots H against 6 but this time for
dimensional punch radius R. Hi can be interpreted as the various values of R and a fixed value of T. For larger values
ratio of the film thickness entrained by the inlet zone divided of R, H tends to remain relatively constant over a large range
by the initial central film thickness. The entrained film of 6 values in the center of the work zone and then decreases
thickness at first generally increases with time though this rapidly near the edge of the conjunction.
effect is very small with small non-dimensional punch radii. Figure 6 shows the influence of plastic heating of the sheet
As the process proceeds stretching of the sheet off the punch on lubricant entrainment in the plane strain process. The
head becomes more important than the wrapping action, the results are plotted as nondimensional inlet film thickness Hx
velocity U decreases and so does the entrained film thickness. on a base of nondimensional time T for various values of the
Eventually U falls to zero and the inlet zone can no longer thermal plastic parameter K and a fixed value of the non-
entrain a lubricant film. This breakdown in thick film dimensional punch radius R = 0.5. A' is a measure of the
lubrication will have similar consequences to those described reduction in lubricant viscosity produced by plastic heating.
by Wilson [1] in upsetting and extrusion. However in the At small values of T, before the sheet heats up the entrained
present case the breakdown is due to a failure in the film film thickness is close to the isothermal solution (A" = 0). As
generation mechanism while in the previously cited cases it is the process proceeds the sheet heats up and the lowered
due to film transport. In practice the breakdown in thick film viscosity of the lubricant near the sheet surface tends to
lubrication does not necessarily imply a complete lubrication reduce the negative influence of the sheet stretching over the

Journal of Tribology JANUARY 1984, Vol. 106/75

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 02/19/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 R=0.5 U.ID
R = 0.5
T = 0.8
T=1.8
S = 1.0
plane
strain/

i=0.6
E
0.10
o Q4

axisymmetric L-K~-°

% Q2 Q4 0.05

Fig. 9 Comparison of film thickness distribution in equivalent plane

strain and axisymmetric processes
K=5

non- d mensional f r i c t i o n stress F K = 10 \ .

2.5

o
1.0

o° ^1 ui b =•
I 1
! T </1 -1
n
0.2 0.4 0.6 0.8 1.0 1.2
s5
0.2

angle 8
Fig. 11 Influence of plastic heating on friction distribution in a plane
strain process
0.4
0.6

thickness at T=0. This is obviously not a realistic solution.

angl

01
The analysis fails because at small values of T a number of the
assumptions used including (2), (4b), (9), (10), and (11)
7a become suspect. For this reason the theoretical predictions are
a>P 6 o 6 shown with a dashed line at small values of T.
r r Figure 9 compares the film thicknesses in the work zones of
b
1 /
equivalent plane strain and axisymmetric processes. While the
1.2

\/6 variation of film thickness with position is similar in each

X ID case, the axisymmetric film is still about an order of
/
1.4

magnitude thinner than the plane strain film. The axisym-

Fig. 10 Influence of nondimensional punch radius on friction metric film extends to larger values of the position angle 6
distribution in a plane strain process than the plane strain film because the sheet tends to wrap
more round the punch in the former case.
Figure 10 shows the influence of punch radius on friction in
punch head on the entrainment process. Thus despite the the isothermal plane strain process. Results are plotted as
overall reduction in lubricant viscosity the net influence of nondimensional friction stress f o n a base of the angular
plastic heating of the sheet is to increase the entrained position coordinate 6 for a variety of different non-
lubricant film thickness. dimensional punch radii R at a fixed nondimensional time T
In addition to the increased entrainment plastic heating also = 1.8. At small values of 6 near the punch center, .Fincreases
tends to reduce the outward transport, stretching and thinning almost linearly with 6. This is because the film thickness in
of the lubricant film due to the motion of the sheet surface is this area is essentially constant (Fig. 5) while the sheet surface
the work zone. Figure 7 plots the nondimensional film velocity is proportional to 6. At larger values of 6 where the
thickness H in the work zone of a plane strain process on a film thickness is decreasing, F increases rapidly. Similar
base of the position angle 6 for various values of the plastic results are obtained for the isoviscous axisymmetric case.
thermal parameter K and a fixed nondimensional punch Plastic heating of the sheet dramatically reduces friction
radius R = 0.5. Increasing K tends to increase the film stresses as shown in Fig. 11. In the presence of plastic heating
thickness H. viscous friction stresses are so low as to be generally
Figure 8 compares the film thickness entrained by the inlet negligible. Thus the chief component of friction will be due to
zone in an isoviscous plane strain process with that in an asperity interactions in the mixed or boundary regimes (1)
equivalent (same p, V, a, r, w and d) axisymmetric process. particularly in regions where the inlet zone fails to generate a
The film thickness entrained in the axisymmetric process sufficiently thick film to ensure adequate surface separation.
decreases monotonically with increasing nondimensional time
T and is generally about an order of magnitude smaller than
that in the plane strain process. This is due to less favorable Conclusions
geometry in the inlet zone and higher interface pressures in the A mathematical model for the hydrodynamic lubrication of
axisymmetric case compared with the plane strain case. The plane strain and axisymmetric stretch forming process has
entrained film thickness in the axisymmetric case also reaches been developed. The model indicates that a hydrodynamic
zero (indicative of lubrication breakdown) at a much earlier lubricant film can be entrained by a wedge shaped inlet zone
stage than in the plane strain case. It is interesting to note that formed at the outer edge of the conjunction between punch
the axisymmetric analysis predicts an infinite entrained film and sheet as the sheet wraps round the punch. The film is

76/ Vol. 106, JANUARY 1984 Transactions of the ASME

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 02/19/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 thickest near the punch center and decreases near the edge of Acknowledgments
the conjunction and as the process proceeds breakdown of the
hydrodynamic film may occur near the outer edge of the The work described in this paper was conducted while the
conjunction in the later stages of the process. The initial authors were at the University of Massachusetts. The authors
wish to thank the University Computing Center for the use of
central film thickness in an isoviscous plane strain process is
its facilities.
given by
References
lr/2/3,
[A,],_o=2.24 (63) 1 Wilson, W. R. D., "Friction and Lubrication in Bulk Metal-Forming
a1' 3^2/3 Processes," J. App. Metalworking, Vol. l , N o . 1, 1981, pp. 7-19.
2 Newnham, J. A., "Sheet Metal Forming," in Metal Deformation
Processes: Friction and Lubrication, ed. Schey, J. A., Marcel Dekker, New
York, 1970, pp. 703-770..
where n is the lubricant viscosity, V the punch velocity, r the 3 Fogg, B., and Owais, M. K. A., "Calculation of Pressure Distribution
punch radius, a the sheet flow stress, w the initial sheet half and Oil Flow and Film Thickness in Externally Pressurized Lubrication in Sheet
width, and d the initial sheet thickness. The film thickness in M<*&[," Annals of the C.I.R.P., Vol. 27,No. 1, 1978, pp. 183-188.
an axisymmetric process is substantially less than that in an 4 Nine, H., "Draw Bead Lubrication," in Mechanics of Sheet Metal
Forming, ed. Koistenen, D. C , and Wang, N. M., Plenum Press, New York,
equivalent plane strain process due to the higher interface 1978.
pressures and less favorable inlet geometry in the axisym- 5 Chakrabarty, J., " A Theory of Stretch Forming over Hemispherical
metric case. Plastic heating of the sheet tends to increase the Punch Heads," Int. J. Mech. Sci., Vol. 12, 1970, pp. 315-325.
lubricant film thickness and to decrease friction. 6 Wilson, W. R. D., and Delmolino, W. P., "The Influence of Surface
Roughness on the Lubrication Breakdown in Upsetting Between Overhaning
Hydrodynamic friction levels are so low as to be generally Dies," Wear, Vol. 29, 1974, pp. 1-10.
negligible. Thus the main frictional effects in practical 7 Wilson, W. R. D., and Farrel, W. J., "Lubricant Transport and
processes will be due to asperity interaction in the mixed or Breakdown in a High Speed Forging Operation," Proc. N.A.M.R.C. IV, 1976,
boundary regimes. pp.130-137.
8 Blok, H., and Van Rossum, J. J., "The Foil Bearing - A New Departure
Future theoretical work should be directed towards in Hydrodynamic Lubrication," Lub. Engr., Vol. 9, No. 6, 1953, pp. 316-320.
achieving a better understanding of the practically more 9 Wilson, W. R. D., "An Isoviscous Model for the Hydrodynamic
important axisymmetric process, including allowing for the Lubrication of Plane Strain Forging Processes wth Flat Dies," ASME JOURNAL
influence of friction on the mode of deformation and OF LUB. TECHNOLOGY, Vol. 96, 1974, pp. 539-546.
10 Wang, J. J., "Hydrodynamic Lubrication of Stretch Forming
developing a model for thermal effects. The influence of Processes," M.S. Project Report, University of Massachusetts, 1981.
surface roughness on the lubrication process should also be 11 Wilson, W. R. D., and Aggarwal, B. B., "Thermal Effects in
explored. On the experimental side, measurements of Hydrodynamically Lubricated Strip Rolling," Proc. 5th Leeds-Lyon Sym-
lubricant film thickness and frictional conditions would be posium on Tribology, I. Mech. E., London, 1978.
12 Mahdavian, S. M., and Wilson, W. R. D., "Lubricant Flow in a Plasto-
most useful in testing the validity of the present theoretical Hydrodynamic Work Zone," ASME JOURNAL OF LUB. TECHNOLOGY, Vol. 96,
models and guiding their improvement. 1976, pp. 16-21.

Journal of Tribology JANUARY 1984, Vol. 106/77

Downloaded From: http://tribology.asmedigitalcollection.asme.org/ on 02/19/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use