Turbulence and Inertia Effects in Bearings
Turbulence and Inertia Effects in Bearings
Turbulence and Inertia Effects in Bearings
in bearings
N. A. Macken* and E. A. Saibel t
Thi s arti cl e surveys t urbul ence and i nerti a effects in bearings. The effects occur f or t wo main
reasons - high-speed operati on and t he use of unconvent i onal lubricants. Experi ment al and
theoreti cal studies are considered and recommendati ons are given f or f ut ur e investigation.
N o t a t i o n
C
f
K
L
N
P
R
Re
Re*
T
U
Uo
U
U~
U t
t t r 2 t t
U , I ~ , I ~ , I ) W
V
t
V
X
Y
y+
8~+
e
Cc
EI
0
/a
p
7" W
7" o
0
Clearance
Friction factor = 2r w/ pU 2
Mixing length constant
Mixing length
Bearing length
Angular velocity
Pressure
Mean radius
Reynolds number, UC/v
Reduced Reynolds number, (UC/v) (C/L)
Torque
Tangential velocity of bearing
Core reference velocity used by Reynolds 2s
Mean velocity in tangential direction
Mainstream velocity (from boundary layer
t heory)
Fluctuating component of velocity in
tangential direction
Average fluctuating components
Axial velocity
Mean velocity normal to tangential velocity
Fluctuating component of velocity in
normal direction
Tangential. direction
Direction normal to tangential
Dimensionless distance y/OV~w/P
Constant in Reichardt' s law of the wall
Eccentricity ratio = eccentricity/clearance
Core eddy viscosity
Eddy viscosity
Angle
Absolute viscosity
Density
Bearing shear stress
Scale shear stress used by Reynolds 2s
Kinematic viscosity
* As s i s t a nt Pr of e s s or o f Mechani cal Engi neer i ng, ~" Pr of es s or o f
Appl i ed Me c ha ni c s , Car negi e- Mel l on Uni ver s i t y, Pi t t s bur gh,
Pe nns yl va ni a 15213, USA
In recent years there has been a great deal of interest in
turbulence in bearings. Previous to this, flows that occurred
in lubricant films were laminar in nature. This was borne out
by experiment and also by the agreement between laminar
theory and experiment.
Non-laminar flow occurs in bearings for two main reasons.
The first is high-speed operation. The second is the use of
unconventional lubricants such as water or liquid metals.
These materials are frequently used as process fluids. The
need to simplify equipment design or the difficulty of over-
coming shaft sealing problems lead t o their use as lubricant
as well. A high velocity combined with a low kinematic
viscosity leads t o high Reynolds numbers and departure
from laminar flow.
For the case of journal type bearings, which are the ones
most attention has been given t o, above the laminar range
there is almost always a flow regime in which a vortex
structure exists. As the Reynolds number increases, de-
parture from laminar flow is first recognized by the forma-
tion of Tayl or vortices, named after G. I. Taylor who first
observed this phenomenon. With a furt her increase,
turbulence appears t o occur but some vortex structure
almost always exists, even with higher Reynolds numbers.
Thus the flow is inherently different from the turbulent
flow in pipes, boundary layers, or jets.
Unfortunately, the experimental work that has taken
place has mostly concerned itself with gross measurements
such as torque and load capacity. Not much has been done
to measure quantities such as fluctuating flow components.
Theoretical descriptions of turbulent flow in bearings
have drawn upon mixing length t heory and laws developed
for fully developed flow in pipes. It is obvious from this
brief discussion that such models cannot adequately des-
cribe the flow field. Also, the convective inertia terms have
almost always been neglected in the equations of motion.
These terms are important in high speed flow and it is
thought that some effects attributed to turbulence may be
caused by inertia. Suprisingly enough, these theoretical
analyses have been reasonably successful in predicting
characteristics of bearings operating in this regime. Perhaps
this is because t hey contain adjustable constants that can
be made to fit existing experimental data.
154 TRIBOLOGY August 1972
E x p e r i m e n t a l i n v e s t i g a t i o n s
The earliest work connect ed with bearings outside the
laminar regime was experimental in nature. Many investi-
gations, particularly the early ones, have been visual in
nature. Gross measurements of such quantities as torque
have been used t o indicate non-laminar behavior. A few
studies have used anemomet ry. Limited in-depth studies
have been made of t urbul ent quantities.
The types of bearings studied have mainly been plain
journal bearings. This type of film is most likely to go
beyond the laminar regime. However, interesting variations
of the plain journal such as step journal, helical-grooved
journal bearings or viscoseals have received some attention.
The earliest experiments involved long concentric cylinders.
Later refinements included end effects and eccentricity.
Some recent work has also included blockage and tilted-pad
effects.
Non-laminar flow in bearings has several characteristic
features. Some of these are unique to bearings and some are
found in ot her areas of fluid mechanics. These features
include:
1 A development of secondary vortex flow when the
flow becomes non-laminar. Possible exceptions to this
are flows where ~
C/R < 1/2500
Generally, for bearings this value is around 0.001.
Anot her exception is possibly in the largest clearance
of an eccentric bearing 2 .
2 Secondary flows which are oft en laminar for a large
range of parameters despite their compl ex nature.
Since t hey also are originally stable, there is no reason
that t hey would disappear from the flow field
completely by the presence of turbulence. Some modi-
fication, however, is apparently evident 3 .
3 A three-dimensional flow field which, due t o end leak-
age, is composed of an axial flow in addition to the
tangential flow. As pointed out by Burton 4 , the cir-
cumferential type flow is a Couette type flow whereas
the longitudinal flow is a pressure type flow. The two
types cannot be separated since the flow field is really
a combination of both.
4 Geometric factors such as existence of steps, length of
bearings and the way the lubricant enters the bearing
which all have strong influence.
The term turbulent flow is frequent l y applied in bearing
literature to any flow which is non-laminar. From the above
statements, this is perhaps a misnomer.
The following paragraphs will discuss significant exper-
imental developments in more detail. Table 1 also summarizes
work that has been done.
Concentric cylinders
G. I. Tayl or s was the first t o discover the non-laminar nature
of flow between concentric cylinders. Long concentric
cylinders, with the inner cylinder rotating and the outer
cylinder stationary model full journal bearings with two im-
portant exceptions. They do not show end effects and the
Tabl e 1
Author
Exper i ment al studies o f turbulence and inertia effects in bearings
Type of bearings Measurements made Results
Vohr 2 Eccentric cylinders
Taylor s Concentric
cylinders
Taylor 6 Concentric
cylinders
Schultz-Grunow Concentric
and Hein 7 cylinders
Wilcock 9 Journal bearing
Abramovitz 1 Tilting-pad thrust
bearing
Smith and
Fuller H
Kaye and Elgar 12
Yamada~3,14
Yamada ts
Burton and
Carper 16
Journal bearing
Concentric
cylinders
Rotating coaxial
cylinders
Rotating coaxial
cylinders with
grooves
Concentric
cylinders tilted
pads, short journal,
step pads
Torque against speed, critical speed
e, friction factor against Re, visual
observation
Visual observation I
i
Torque against speed
Visual observation
Power loss against speed, bearing
temperature against speed, oil flow
against speed, eccentricity ratio
against speed, friction coefficient
against speed
Friction coefficient against speed
Friction coefficient against Re,
pressure against 0, eccentricity
against load
Flow fluctuations by hot wire
Torque and pressure
Torque and pressure
Velocity profiles, turbulent
intensities, pressure distribution
Determination of critical speed by the
measurements made
Mathematical description of stability and
confirmation by visual observations
Further confirmation of t he stability
criterion
Examination of the flow stability before
and after the critical speed
Confirmation of the stability criterion by
the distinguished happenings at critical
speed
Friction torque increased abnormally when
speed exceeded what appeared to be a
critical value
Comparisons between laminar flow and the
flow with a speed of five times that of
critical speed.
Division of flow patterns with axial flow
Showed effect of axial flow, agreement
with Ref 12
Extend results in Refs 13, 14 to groove
geometry
Discussed intensity results, comput at i on
of wall shear, pressure jumps at dis-
continuities not ed
TRI BOLOGY August 1972 155
effects of eccentricity. Tayl or analysed mat hemat i cal l y the
stability of incompressible viscous flow in a narrow annulus
between rot at i ng concentric cylinders of infinite length. His
analysis predicted the initial instability for the case of
rotating inner cylinder and fixed out er cylinder t o be in the
form of uni form symmet ri cal vortices or eddies. He also
confirmed bot h the value of the critical speed and the
existence of the pairs of votices by means of fl ow visualiza-
tion, using the flow of water bet ween vertical cylinders with
zero axial velocity. Later on, Tayl or 6 was able t o verify his
stability criterion by a number of torque measurement
experiments. Fig 1 is one of his results showing the relation-
ship bet ween log T/pN 2 and log Np/#. Laminar flow would
be indicated from the solid line. Deviations f r om this indi-
cate the onset of vort ex flow.
Schultz-Grunow and Hein 7 among others have also made
visual observations of the format i on of Tayl or vortices in
concentric cylinders. The vort ex nature of the flow is still
evident, even at a Reynolds number of 2.13 X 10 4 . Pai s
also studied the vortices bet ween concentric cylinders using
hot wire anemomet r y and found t hat the vortices existed at
several hundred times the critical speed.
Eccentric bearings
Wilcock 9 and Abramovi t z t were among the first t o 9bserve.
non-laminar characteristics in real j ournal beatings. Wilcock
performed some experi ment s with non-concentric journal
bearings under loads operating above and below the
Tayl or transition. He also checked the critical,~peed by the
torque measurement technique which Tayl or 6 empl oyed
earlier.
Smith and Fuller l~ investigated an unloaded full journal
bearing and again verified t hat Tayl or' s critical Reynolds
number was actually very close t o the poi nt of departure
f r om laminar behavior (see Fig 2). They were operating in
the vort ex region with probabl y little superimposed
turbulence.
The effect of eccentricity was discussed most compl et el y
by Vohr 2 . He demonst rat ed t hat the transition speed for
onset of vortices increases significantly with eccentricity
ratio and t hat eccentricity has marked effect on friction
factor. Visual results also showed a breakdown i nt o large,
apparent l y irregular eddies in the wide clearance region.
The vort ex structure is i nt act in the narrow region. The
eddies eventually become finer and a vort ex structure
appeared again and persisted up t o the highest Reynolds
number studied.
O .
. J
4 . 5 N + . h . - Inner cylinder rotati ng
C r i t J c a J v a J ! ~ " . O ~ " o
3. 5 ~ d o-a---<~q~L~-~--
i l i l ~ i = i I
2<) 3"0 4 . 0
Log Np /1~
Fig 1 Vari at i on of t orque wi t h speed (from Ref 6)
- 2 . 0
- 2 . 4 f = 0. 078/ ReO' 43f = 2/Re I
- 2 . 8 " %
%
%.
Cri ti cal Re
!
!
I I I i I I I I I
2.6 3"2 3"8
Log Re
Fi g 2 Frictional characteristics of a journal bearing (from
Ref 10
~-- ~ ~r- Turbulent flow
x~ IOOO I r ~ P lus vr t k Laminar" plus vortices
o~ flow
I
2 0 0 4 0 0 6OO
NR 112 C312/v
Fig 3 Flow regime in a bearing wi t h end leakage, C/R =
0.307 ( f r om Ref 11)
a .
Effects o f bearing length
Kaye and Elgar 12 used photographi c and hot wi re tech-
niques t o investigate the effects of end leakage on con-
centri c cyl i nders. Thei r results are shown on Fi g 3. The
vel oci t y V used i n the Reynol ds number on the ordinate
curves is the axi al vel oci t y. Purel y t ur bul ent f l ow onl y
occurs f or axi al f l ow wi t h reasonably hi gh Reynol ds
number, > 1000, an ext remel y hi gh leakage rate f or bearings.
For no leakage no turbul ence was observed.
Yamada13,14,is di d considerable wor k i n thi s area. He
t ook t orque measurements of concentri c cyl i nders wi t h
axi al f l ow. The effect of rectangular grooves was also
studied Is ,
Since most journal bearings have some side leakage, a
combi nat i on of t urbul ent and vortex flow is indicated
for high rotational speeds.
Measurement o f t ur bul ent intensities
Quant i t at i ve measurements of turbul ence have been
measured by Bur t on and Carper 16 . Experi ments were
carried out using a hot wire anemomet er on a large scale
model of a journal bearing. Velocity profiles showed the flow
to be t urbul ent in nature. Measurement of the fluctuating
component u' indicated agreement with pipe flow near the
wall. Away from the wall there was disagreement, possibly
due to the presence of midstream shear stress, which is
absent in pipe flow.
Traugot t ~7 has performed experiments in induced swirling
156 TRI BOLOGY August 1972
flows with bot h tangential and axial flow fields. Measurements
of fluctuating velocity components and turbulent stresses
were taken.
T h e o r e t i c a l i n v e s t i g a t i o n s
Theoretically speaking, there are several problems associated
with non-laminar flow in bearings.
First there is the problem of mathematically determining
the viscous terms that appear in the Navier-Stokes equations.
These mathematical descriptions are basically of t wo types
as described below. Both types have theoretical bases in
fully developed turbulent flow. One met hod is derived from
mixing length theories, and the other is derived from
logarithmic profiles of fully developed t urbul ent pipe flow.
As pointed out earlier, the flow in bearings is much different
than in pipes or boundary layers, where these theories may
be valid.
In conventional lubrication, the inertia terms in the
equations are negligible to the viscous terms. As discussed
below, at high speeds, bot h inertia and turbulence effects
play a role and it is difficult to distinguish between the two.
A third difficulty with the theories is that many of them
are two dimensional. From the experimental discussion, it
should be clear that a three-dimensional model is called for.
Turbulence models and inertia are discussed in more
detail in the following paragraphs. Table 2 summarizes im-
port ant developments.
Table 2 Theoretical studies of turbulence and inertia effects in bearings
Author Turbulence model Type of Bearing Results
Tao ~s
Ta o 19
Chow and Saibel 2
Constantinescu 2t
Constantinescu 22
Arwas, Sternlicht
and Wernick 2a
Reynolds 2s
Ng ~s
Ng and pan27
Elrod and Ng 2s
Chow and Vohr 29
Vohr and Chow a
Kettleborough a9
Constantinescu as
1/7 velocity power law. Blasius law of
friction
Blasius law of friction
Mixing length:
~2 = K 2 y ( C_ y)
Mixing length:
= Ky O<y <C/ 2
= K( C- - y ) C/ 2<y <C
with K = 0.4
Same as Ref 21 except K = 0.2 to
0,4
Same as Ref 20
Law of wall
I
U = ro i 1/0.41
r o s y
~n + 2.4
o
Core velocity
1
U =: Uo - ro ~ f O' / c)
= y * - - 10.7 tanh
0
Same as Ref 26, except applied to
three dimensional flows, assuming
i sot ropy
Same as Ref 26, except also used
core viscosity
ec = 1/56f* l u =- - u l d y
0
Same as Ref 27
Same as Ref 28
Same as Ref 20
Mixing length 22 or eddy viscosity
model 27 can be used
Short journal
bearing
Slider bearing
Slider bearing
Journal bearing
Journal bearing
Journal bearing
and 180 arc
journal bearing
Slider bearing
Couette and
Poiseuille
flow
Full journal
bearing
Full journal
bearing
Helical-grooved
journal bearing
Spiral-grooved
screw seal
Slider bearing
Slider, step
beating
Closed form expressions for pressure
distribution, load capacity and attitude
angle
Expressions for load capacity, friction
force, and power loss
Expressions for pressure distribution, load
capacity friction force and power loss
Velocity profiles, pressure distribution,
and attitude angle
Load capacity and attitude angle. Good
agreement with Smith and Fuller H
Load capacity, attitude angle, and pressure
distribution. Good agreement with Smith
and Fuller n
Pressure distribution and friction stress
distribution in bearing films
Friction characteristics, velocity profile,
and pressure distribution. Not very
favourable agreement
Linearized equations, pressure distribu-
tion, load capacity, attitude angle, and
load against eccentricity ratio. Results
fall between Constantinescu and Smith
and Fuller 11
Developed working formulas using Ng's 26
t heory and incorporated core viscosity
Used linearized t heory. Load against
eccentricity ratio
Sealing coefficient against Reynolds num-
ber. Favourable agreement with 3t and 32.
Included inertia. Load capacity and
friction characteristics: i f onl y inertia is
considered, the results are in qualitative
agreement with Abramovitz t and
Smith and Fuller It
Included inertia. Used integral equation.
Results for slider and step. Also presented
equations for three-dimensional flow
TRI BOLOGY August 1972 157
Tur bul ence model s based on m i x i n g l engt h
Perhaps the simplest way to deal with the t urbul ent model
is to make use of mixing length theory. Assuming no side
leakage and dropping inertia terms, the equat i on of interest
reduces to:
OZu ~ ,
dig = la + (-- p u'v')
The turbulent stress t erm can be written as:
- p u' v' = p~ au gu
ay ay
where ~ is the mixing length Prandtl used for pipe flow, i.e.
= Ky
with K = 0.4. An alternate approach, presented by Tao is'a9
deals with turbulence in an approxi mat e fashion for the case
of short journal bearings and slider bearings. A short journal
bearing is one in which the axial length is small, and the
pressure t erm in the x direction can be neglected. Ta o
comput ed the average flows by using the Blasius law of
friction and the 1]7 velocity profile.
Chow and Saibel 2e solved equations for a slider bearing.
They neglected the laminar friction t erm and also
modified the mixing length model. Constantinescu 21 in an
early work empl oyed a similar mixing length model where:
= Ky O < y < C / 2
= K ( C - y ) C/2 ~ y <~ C
Results were comput ed with constant and variable
viscosity for a journal bearing wi t hout leakage. A comparison
with Smith and FuUer' s experimental results showed poor
agreement, as shown in Fig 4. The aut hor suggested t hat per-
haps a smaller value of K, i.e. K = 0.2 to 0.4 might be more
appropriate because of the small film thickness. In a later
paper, Constantinescu n considers the possibility of side
leakage. This leads t o an additional equation and an
additional turbulence term. The value of the mixing length
is taken to be the same as in his earlier work, 21 . In this
paper, he also recomput es his results for K = 0.3 and shows
bet t er agreement with Smi t h and Fuller. This is surprising
in light of the previous discussion t hat indicated t hat fully
t urbul ent flow was probabl y never reached by the exper-
iments.
Arwas, Sterlicht, and Wernick ~-3 used the same mixing
length approach as Constantinescu 2a in application to a full
journal bearing and a 180 arc journal bearing. Side leakage
was neglected and K = 0.4. Agreement appears as good as
t hat of Constantinescu 2~ .
Tur bul ence model s based on e d d y vi scosi t y
In dealing wi t h fully developed pipe flow, it is well
established 24 t hat the simple mixing length theories given
above do not adequately describe the phenomenon of tur-
bulence. Logarithmic ' laws of the wall' suggested by
Boussinesq, Reichardt and others are used. These laws were
arrived at by fitting fully developed pipe flow data. The
flow field is found to be composed of several regions, each
with a different logarithmic law.
Reynolds zs was one of the first to use this approach in
dealing with bearings. His turbulence model included a wall-
law region and a core flow. His met hod, applied t o a slider
bearing, showed pressures produced in the t urbul ent film
were ten times greater than those in a laminar flow.
04
E
12
u
Q.
- - - T . r , . q = 0
-----Laminar, q - O o / / o~, ~
81 o Experimental, Smith / / ',]
and Fuller o / / i
Re =3310 ~ "" / tl
/ / 0 1
.,."10 / | ,
4 / 1 0
~ J 0
" ' ~ ' ~ ' ~ " ~ 1 I I I I I I
4 0 8 0 120 160
e[o]
F i g 4 Pressure di st r i but i on f o r a journal bearing wi t h o u t
leakage ( f r om Ref 20)
8O0
6 0 0
4OO
2OO
!
I
Cl R 2.93 x I O - 3 i
I
x Smi t h end Fuller I
I
Present theory I
. . . . Constantinescu's analysti~t 7
7 4 5 0 r e v / m l ~ / / /
/ / / / x
. " I aooo .~
, ,
0. 2 0. 4 0 . 6
D
0"8
E c c e n t r i c i t y r a t i o ,
F i g 5 Load against eccent ri ci t y rat i o ( f r om Ref 26)
300O
2 0 0 0
IOOO
z
A widely used t heory for bearings at present is t hat
developed originally by Ng 26 . The viscous t erm in the
Reynolds equation is rewritten in terms of eddy viscosity.
For eddy viscosity, Reichardt' s formula gives
e~ _ K[ . V. _ 6 + ~ t a n h ( ~ + ) ]
Ng found t hat with a value of K 0.4 and 5~ = 10.7 the
above formula reasonably fits all regions of pipe flow data.
This idea has been furt her expanded by Ng and Pan 27 .
They comput ed results for a three-dimensional journal
bearing. To simplify comput at i on, a linearized form of the
equations was used. Results showed much bet t er agreement
with Smith and Fuller' s experi ment t han did Constantinescu' s
analysis as shown in Fig 5.
Elrod and Ng 28 have modified the above t heory by using
a const ant eddy viscosity in the core region, as suggested by
158 T RI BOL OGY August 1972
Clauser. Their comput at i ons also di d not use t he linearized
equat i ons. Results are compar ed t o t he linearized case.
Chow and Vohr 29'3 have applied t he linearized and non-
linearized analyses descri bed above t o hel i cal -grooved and
spiral-grooved bearings. Reasonabl e agreemerit was s hown wi t h
results o f Pape and Vr akki ng 31, as well as Ket ol a and
McGrew 32 . The bearings st udi ed have st ep confi gurat i ons,
and t he results are surprising in t erms of what is ment i oned
bel ow about inertia effects.
Inertia effects
I n t he earl y experi ment al wor k wi t h non-l ami nar fl ow i n
bear i ngs t here appeared i n t he literature muc h discussion as
t o why a depar t ur e f r om l ami nar fl ow occur ed 9 . I t was
suggested t hat t he presence o f t he Tayl or vort i ces i ndi cat ed
an increase in inertia forces in t he fluids. As i ndi cat ed, t he
analytical investigations have by and large i gnored this fact
and have consi dered onl y t urbul ence.
Ar gument s have been made t hat t he inertia forces are still
insignificant compar ed t o t he viscous forces is . Kulinski and
j our nal bearings, st ep j our nal bearings, helical-grooved
beatings and spiral-grooved bearings. The experi ment s
shoul d also st udy t hree-di mensi onal effect s of side leakage,
and di fferent met hods f or i nt r oduci ng t he l ubri cant i nt o t he
bearings.
On t he t heoret i cal side, it seems t hat inertia effects must
be mor e careful l y dealt wi t h and i ncl uded in t he analyses.
There seems little need t o cont i nue t o solve probl ems o f
mor e and more compl exi t y while still neglecting these
terms. Also, t he t urbul ence model s do not adequat el y des-
cribe t he phenomenon t hat is occuri ng in t he non-l ami nar
range o f most bearings. A new model whi ch takes i nt o
account t he vor t ex nat ur e of the fl ow and t hree-di mensi onal
effect s is needed. The measur ement s suggested above can
be used t o devel op these new theories.
Acknowledgements
The survey above was part i al l y suppor t ed by NSF grant
GK-27831. This suppor t is gr at ef ul l yacknowl edged.
Ost rach 39 have per f or med a careful order o f magni t ude
analysis. The y have shown t hat i f t he r educed Reynol ds number Ref erences
Re* = UC/ o( C/ L) ~ 0(1)
t hen t he i nert i a t erms shoul d be ret ai ned in t he equat i ons.
Onl y when Re *<< 1 can t he inertia t erms be omi t t ed f r om
t he equat i ons. Zuk et a134 also discussed inertia effect s and
came t o t he same concl usi ons. I n addi t i on, t hey c omput e d
t he modi f i ed Reynol ds number f or onset of t ur bul ence f or
various experi ment s. The values were of or der one whi ch
suggests t hat inertia effect s may be pl ayi ng a significant role.
It woul d appear t her ef or e, t hat any t heoret i cal st udy o f
t urbul ence woul d have t o i ncl ude inertia as well.
I n addi t i on, Zuk et al a4 , as well as Bur t on (see discussion
in Ref 35) have poi nt ed out t hat near steps or ot her dis-
cont i nui t i es significant inertia effect s are felt. Bur t on 3s also
i ndi cat es t hat i nert i a may be i mpor t ant in shor t bearings as
well.
Pi nkus and St ernl i cht 36 have discussed t he t wo classical
approaches t o t he pr obl em o f how t o deal wi t h inertia. The
averaging me t hod was first f or mul at ed by Osterle and
Saibe137 . The i t erat i on met hod was suggested by Kahl ert 3s ,
The dr awback t o these above approaches is t hat t hey are
bot h appr oxi mat e i n t echni que. Tur bul ence is also not
i ncl uded. Ket t l ebor ough 39 consi dered bot h inertia and
t urbul ence i n his discussion o f a slider bearing. He used
Chow and Saibel' s 2 defi ni t i on o f mi xi ng l engt h. His results
showed t hat t he i ncl usi on of t urbul ence had a compar at i vel y
small effect and caused a slight decrease in l oad capaci t y, in
cont r adi ct i on t o ot her t heoret i cal analyses. The increase in
l oad-carryi ng capaci t y was shown t o be due t o inertia and
not t o t ur bul ence effect s. Wi t h inertia, not t ur bul ence,
i ncl uded i n t he analysis, qualitative agr eement was shown
wi t h the results of Abr amovi t z 1 .
Const ant i nescu in a recent paper 3s has discussed a new
appr oach t o t urbul ence and inertia effects. The integral
equat i ons are used rat her t han t he differential equat i ons of
mot i on. Appr oxi mat e const ant s are used in t he vel oci t y
profiles. The skin fri ct i on vari at i on can be arrived at f r om
mi xi ng l engt h or El rod and Ng' s 2s t heor y. The met hod is
applied t o a slider and st ep bearing. Equat i ons are present ed
f or t hree-di mensi onal flows as well.
Conclusions
Experi ment al i nf or mat i on is sorel y needed t o bet t er des-
cribe t he nat ure of t ur bul ent fl ow in bearings. Such exper-
i ment s shoul d measure mean and f l uct uat i ng component s
1 DiPrima, R. C. ' A note on the stability of flow in loaded
journal bearings', TransAm Soc Lubric Engrs, Vol 6, No 3
(1963) pp 249-253
2 Vohr, J. H. ' An experimental study of Taylor vortices and
turbulence in flow between eccentric rotating cylinders',
Trans Am Soc mech Engrs, J Lubric Technol, Ser F, Vol 90,
No 1 (Jan 1968) pp 285-296
3 Coles, D. 'Transition in circular Couette flow', J Fluid Mech,
V.ol 21, Pt 3 (1965) pp 385-425
4 Burton, R. A. ~rurbulent film bearings under small dis-
placements', Trans Am Soc L ubrqc Engrs, Vol 7, No 4 (1964)
pp 322-332
5 Taylor, G. I. 'Stability of a viscous liquid contained between
two rotating cylinder', Trans R Soc, London, Ser A, Voi 223
(1923) p 289
6 Taylor, G. I. 'Fluid friction between rotating cylinder',
Proc R Soc, London, Set A, Vol 157 (1936) pp 546-564
7 Schultz-Grunow and Hein, H. 'Beitrag zur Couettestromung'
Z/w" Flugwiss, Vol 4 (1956) p 28
8 Pal, S. I. 'Turbulent flow between rotating cylinders',
NACA TN 892 (1943)
9 Wilcock, D. F. 'Turbulence in high speed journal bearings',
TransAm Soc mech Engrs, Vol 72 (1950) pp 825-833
10 Abramovitz, S. 'Turbulence in a tilting pad thrust bearing',
TransAm Soc mech Engrs, Vol 78 (1956) pp 7-11
11 Smith, M. I. and Fuller, D. D. 'Journal bearing operation at
super laminar speeds', TransAm Soc mech Engrs, Vol 73,
(1956) p469
12 Kaye, J. and Elgar, E. C. 'Modes of adiabatic and diabatic
fluid flow in an annulus with an inner rotating cylinder',
TransAm Soc mech Engrs, Vol 80 (1958) pp 753-765
13 Yamada, Y. 'Resi~ance of flow through an annulus with an
inner rotating cylinder', Bull Jap Soc mech Engrs, Vol 5,
No 18 (1962) pp 302-310
14 Yamada, Y. 'Torque resistance of flow between rotating co-
axial cylinders having axial flow', Bull .lap Soc mech Engrs,
Vol 5, No 20 (1962) pp 6344541
15 Yamada, Y. ' On pressure loss of flow between rotating co-
axial cylinders with rectangular grooves', Bull Jap Soc mech
Engrs, Vol 5, No 20 (1962) pp 6424551
16 Burton, R. A. and Carper, H. J. ' An experimental study of
annular flows with applications in turbulent film lubrication',
Trans Am Soc mech Engrs, J Lubric Technol, Vol 89, Set F,
No 3 (Jul 1967) pp 381-391
17 Traugott, S. C. ' Influence on solid-body rotation of screen-
produced turbulence' , NACA Tech Not e, No 4135 (1958)
TRIBOLOGY August 1972 159
18 Tao, L. N. ' A t heory of lubrication in short j ournal bearings
wi t h turbulent fl ow' , Trans A m Soc mech Engrs, Vol 78
(1958) pp 1734-1740
19 Tao, L. N. ' A t heory of lubrication with turbulent flow and
its application to slider bearings' , Tr ans Am Soc mech Engrs,
JappMech, Vol 27, Set E, No 1 (Mar 1960) pp 1-15
20 Chow, V. T. and Saibel, E. ' The effect of turbulence on slider
bearing lubrication' , Trans A m Soc mech Engrs, J appl Mech,
Vol 26, Set E, No 1 (Mar 1959) pp 122-126
21 Constantinescu, V. N. ' On turbulent lubrication' , Proc l nst
mech Engrs, Vol 173, No 38 (1959) pp 881-900
22 Constantinescu, V. N. ' Analysis of bearings operating in
turbulent regime' , Trans A m Soc mech Engrs, J bas Engng,
Vol 84, Ser D, No 1 (Mar 1962) pp 139-151
23 Arwas, E. B., Sternlicht, G. and Wernick, R. J. ' Analysis of
plain cylindrical j ournal bearings in turbulent regime' , Trans
A m Soc mech Engrs, J bas Engng, Vo186, Ser D, No 2
(Jun 1964) pp 387-395
24 Schlichting, H. Boundary layer theory, McGraw Hill, 4th
Edition (1960)
25 Reynolds, A. J. ' Analysis of turbulent bearing films' , J mech
EngngSci, Vol 5, No 3 (1963) pp 258-272
26 Ng, C. W. ' Fl ui d dynamic foundat i on of turbulent lubrication
t heory' , Trans A m Soc L ubric Engrs, Vol 7, No 4 (1964)
pp 311-321
27 Ng, C. W., and Pan, C. H. T. ' A linearized turbulent lubrication
t heory' , Trans A m Soc mech Engrs, J bas Engng, Vol 87,
Ser D, No 3 (1965) pp 675-688
28 Ekod, H. G. Jr and Ng, C. W. ' A t heory of turbulent fluid
films and its application to bearings' , TransAm Soc mech
Engrs, JLubr i c Technol, Vol 89, Ser F, No 3 (Jul (1967)
pp 346-362
29 Chow, C. Y. and Vohr, J. H. ' Hefical grooved j ournal bearing
operated in turbulent regime' , Trans Am Soc mech Engrs, J
Lubric Technol, Vol 92, Set F, No 2 (Apr 1970) pp 346-358
30 Vohr, J. H. and Chow, C. Y. ' Theoretical analysis of spiral-
grooved screw seal for turbulent operation' , Trans A m Soe
mech Engrs, JLubr i c Technol, Vol 91, Ser F, No 4 (Oct 1969)
pp 675-686
31 Pape, J. G. and Vrakking, W. J. 'Viscoseal-pressure generation
and friction loss under turbulent conditions' , Trans A m Soc
Lubric Engrs, Vol 11, No 4 (1968) pp 310-320
32 Ketoia, H. N. and McGrew, J. M. ' Turbulent operation of the
viscoseal', Trans Am Soc Lubric Engrs, Vol 10, No 3 (1967)
pp 256-272
33 Kulinski, E. S. and Ostrach, S. ' A critical evaluation of the
status and trends in high speed fluid film lubrication' ,
NASA Cr 1058 (1968)
34 Zuk, J. et al. ' Convective inertia and gas ingestion effects on
flow regimes of the viscoseal t heory and experiment' , Trans
A m Soc Lubric Engrs, Vol 10, No 3 (1967) pp 273-293
35 Constantinescu, V. N. ' On the influence on inertia forces in
turbulent and laminar self-acting films' , Trans A m Soc mech
Engrs, JLubr i c Technol, Vol 92, Ser F, No 3 (Jul 1970)
pp 473-481
36 Pinkus, O. P. and Sternficht, B. Theory o f hydrodynamic
lubrication, McGraw Hill ( 1961)
37 Ostede, F. and Saibel, E. ' On the effect of lubricant inertia
in hydrodynami c lubrication' , ZAMP, Vol 6 (1955) p 334
38 Kahlert, W. ' Der einflurs der tragheitskrafte bei der hydro-
dynamischen schmiermiteltheorie' , Ing-Arch, Vol 16 (1947)
p 321
39 Kettleborough, C. F. ' Turbul ent and inertia flow in slider
bearings' , Tr ans Am Soc Lubric Engrs, Vol 8, No 3 (1965)
pp 286-295
cr ieal
r e v i e ws i n
t " rlb ogy
1970
The f i r st issue of a new annual publ i cat i on
cri t i cal reviews in t r i b o l o g y
is now available.
The Leeds Industri al Uni t of Tr i bol ogy have scanned all the worl d' s l i terature in t r i bol ogy
f or 1970 and classified i t i nt o si x categories: devel opments - cont act , f r i ct i on and wear
- l ubri cat i on and l ubri cants - materials - mechanisms - i ndustri al appl i cati ons. Thi s
l i terature has been cr i t i cal l y reviewed by leading aut hori t i es in the f i el d, under the
General Edi t orshi p of Professor D. Dowson. In addi t i on t o the reviews, the publ i cat i on
contai ns a compl ete bi bl i ography, and a subject i ndex. 100 pages A4 size.
cri t i cal reviews in t r i b o l o g y 1 9 7 0 c o s t s 6 a copy, or 4 f or subscribers t o TRI BOLOGY.
Ai r mail postage 1. 50 extra. Order your copy now from"
Sales Manager, IPC Science and Technol ogy Press Lt d,
32 Hi gh Street, Gui l df or d, Surrey, England
160 TRIBOLOGY August 1972