CHAPTER 1

BASIC CONCEPTS AND THEORY

1.1 INTRODUCTION :

A bearing is an element of mechanism which allows a force to be transmitted

between two relatively moving parts. Bearings allow relative movement between the
components of machines while providing some location between them and carry the required
load under prevailing conditions, while keeping friction to minimum.
In other words, bearings are highly engineered, precision components with high
reliability and durability enabling machinery to move at extremely high speeds and carry
remarkable loads with ease, efficiency and minimum friction, noise and vibration.
The results of excess friction are excessive wear and heat generation which lead to
loss of efficiency and possible machine failure.
A host of manufacturing industries rely on bearings to enhance machine performance
and prolong the stability of moving parts.
If something twists, turns or moves it probably has a bearing in it.

1.2 HISTORY OF BEARINGS AND BEARING INDUSTRY

TIMELINE:

In ancient times, bearings may have taken the form of roller (cylindrical) bearings in
the wheel hubs of wagons or, more probably, of a conveyor-belt system of logs designed by
the Egyptians to transport large sections of stone for the construction of their pyramids. The
first ball-shaped bearings, made of bronze, were invented by the Romans to assist the
movement of metal turntables on ship decks.

BEARING HISTORY AND TIMELINE

2600 BC – The Ancient Egyptians use a form of roller bearings to help move massive
bricks during construction of the Pyramids.
40 BC - Early example of a wooden ball bearing supporting a rotating table, was
retrieved from the remains of a Roman ship in Lake Nemi, Italy.
1500 AD – Leonardo Da Vinci described a type of ball bearing.
1600 – Galileo describes caged ball bearing to prevent friction.
1794 – First patent for ball race by Phillip Vaughn of Carmarthen.
1883 – FAG begins grinding balls of equal size and roundness forming the creation
of an independent bearing industry.
1898 – First patent issued for Timken Tapered roller bearings.
1907 – Sven Wingquist of SKF invents the modern self-aligning ball bearing
1912 – FAG originates single-row, barrel type, and spherical roller bearings
1913 - Hoover Steel Ball Company is founded by Leander J. Hoover in Ann Arbor.
1916 - NSK inaugurated its business in 1916 and produced the first ball bearings
made in Japan.
1917 – U.S. Bearing Manufacturers create an informal group to aid bearing
manufacturing for World War I which led to the founding of ABMA.
1933 – Article of Incorporation ratified by United States bearing manufacturers to
create AFBMA (Anti Friction Bearing Manufacturers Association).
1960 – Elasto-hydro-dynamic theory explains the mechanism of why bearings and
gears work led to advances in grindings precision and ultrasonic equipment.
1970 – Intel invents the microprocessor and consistent precision control of machine
tools impacting both size and life of bearings.
1980’s – Torrington bearings are used in the Space Shuttle and robot arm to launch
and retrieve satellites in orbit.
1993 – Leading Bearing Manufacturers from throughout the World meet for the first
time in Key Largo, Florida.

1
2006 – ABMA partners with the American Gear Manufacturers Association (AGMA)
for Joint Meeting in Tucson, AZ.
2006 – ABMA, the Japanese Bearing Industrial Association (JBIA) and the
Federation of European Bearing Manufacturers Association (FEBMA) create
the World Bearing Association (WBA) to focus on issues affecting the global
bearing industry.

1.3 RELATIVE MOVEMENT BETWEEN MACHINE

COMPONENTS AND THE CONSTRAINTS APPLIED:
The form of bearing which can be used is determined by the nature of relative
movement required and the type of constraints which have to be applied to it is illustrated as
follows.
For both continuous and oscillating movement, there will be forms of bearing which
allow movement only within a required constraint, and also forms of bearing which allow this
movement among others. The following table give example of continuous movement form of
bearing, and in the case of those allowing additional movement, describe the effect which this
can be have on a machine design.

2
Examples of forms of bearing suitable for continuous movement

3
1.4 NATURE OF SURFACES AND CONTACT:

Fig. 1.1 Surface of metal Fig. 1.2 Ground Surface x 1000

The surface of metal consists of a thin, often transparent oxide film (0.01 – 0.1 μm
thick), containing cracks and pores. Molecules of water, oxygen and grease are weakly
attached to the oxide. Below the oxide may be a layer of mixed-up oxide and metal, often
extremely hard, (perhaps 0.1 μm thick), and below this the metal will be work hardened to a
depth of 1-10 μm. (Fig. 1.1)
All surfaces, even those which feel smooth and give good reflections, are rough. (Fig.
1.2) Chemical measurement show that the real surface area of an abraded metal can be
three or more times the apparent area.

REAL AREA OF CONTACT

0.2 N

12.7 mm

Fig. 1.3 Fig. 1.4 Steel Ball Loaded Against

Aluminised Glass Block,
Viewed Through The
Block X 200

Contacts between flat surfaces at light loads occur at asperity tips only – the scale of
the surface roughness does not matter. (Fig 1.3)
In concentrated contacts, as between a ball and its race, contacts still occurs at
discrete points. (Fig. 1.4)

4
 450 N

38.1 mm

Fig. 1.5 Metal transfer from a copper-coated steel ball loaded against steel plate X 38

At practical loads, contacts are still non uniform. Even in a Brinell indentation contact
occurs at the asperity tips, and the asperities persist, though rather deformed.

Estimate of contact area

The real area of contact depends on the load and not on the apparent area of
contact. A useful estimate (in mm2) is found by dividing the load (in kg) by the hardness,
(Brielle or Vickers) of the softer member.

Fig. 1.6 Talysurf trace of a bead-blasted surface before and after indentation

ELASTIC AND PLASTIC CONTACT

W

Fig. 1.7

Elastic contact between a ball and a plane, or between two balls, is described by the
Hertz equations:
Contact Radious
1/3
3 WR
a = ---- ---------
4 E

5
 Maximum (or Hertz) pressure
2 1/3
0.6 WE
PH = --------
2
R
2 2
Where 1/E = (1- υ1 )/E1 + (1-υ2 )/E2

E1 & E2 being Young’s Moduli and υ1 & υ2 Poisson’s ratios for two bodies.

1/R = 1/R1 + 1/R2

R1 and R2 the radii of curvature of the two bodies, for a ball in a cup radius of
curvature are taken negative.

Fig 1.8 Load Vs Area

At heavy loads plastic flow takes place, beginning when PH reaches 1.8 x (yield stress
in tension of softer body). The area then increases more rapidly with load, approaching the
value (load/hardness), and on unloading leaves a permanent impression (‘Brinelling’). But
even in this range, the real area of contact is only 1/3 to 1/2 of the area of the impression.

Asperity deformation
Individual asperity contacts probably behave like the ball and plane just described, so
that each contact deforms elastically when it carries a small load and plastically when it
carries high one.
But as the total load increases the surfaces move closer together and the number of
asperity contacts increases. The new small contacts which form balance the growth of the
existing contacts, and the average contact size is unchanged. The number of contacts is
roughly proportional to the load, and the fraction of the load carried by elastic contacts will not
change, even though the original elastic contacts have become plastic. Contact between flat
surfaces is therefore elastic or plastic depending on the surface geometry and material
properties, but does not change from elastic at low loads to plastic at high ones.

Plasticity index
1/2
The plasticity index is (E/H) (σ/ß) where E is Young’s modulus, H is the hardness of
the softer component, is the standard deviation of asperity heights and ß the mean asperity
radius of curvature. Values below 1 indicate mainly elastic contacts; values above 3 indicate
mainly plastic contacts. Very few manufactured surface come below 10 -- ball and roller
bearings being an important exception with an index around 1.
Running-in produces smoother surfaces (σ decreases and ß increases) and contact
then becomes elastic, though this is partly due to better conformity between the surfaces as
well as lower roughness. But non-geometric effects like toughening and surface oxidation are
also involved in running in.

CONTACT SIZES
Individual contacts have a large range of sizes with a few large ones and very many
small ones. But the mean contact width does not very greatly with the pressure, provided this

6
Fig. 1.9 Contact width Vs Frequency Fig. 1.10 Zero crossings and contact widths

is less than one tenth of the hardness, nor even between surfaces finished in different ways.
(Fig. 1.9)
A good approximation is to count the number of times n the surface profile crosses
the mean line per unit length: the mean contact width is about 0.2/n. The contacts in the figure
follow this rule. For many surfaces, especially surface ground ones, n is about 100/mm,
giving a mean contact width of 2 μm- This value can generally be used except at high
pressures when the contacts get bigger and closer together (as in the photographs of contact
areas under heavily loaded balls.)
In practice surface waviness and misalignment can often give high apparent
pressures locally when the nominal pressures are low, and the real contact areas tend to be
ground instead of occurring randomly.

1.5 FRICTION MECHANISMS, EFFECT OF LUBRICANTS

Friction can be defined as the resistance to motion when two moving surfaces are in
contact. Within Limits, friction is proportional to normal pressure between surfaces, and is
independent of the area of contact between surfaces, and of the speed of slipping.
The force required to initially overcome friction is always higher than when two
surfaces are sliding. Hence the terms:-

STATIC FRICTION
Opposing initial movement.

KINETIC FRICTION
Opposing continuous movement
The opposition due to friction depends to a large extent upon the nature of the
surfaces in contact, and these contact surfaces may be improved by lubrication.

SURFACE INTERACTION AND THE CAUSE OF FRICTION

Most surface are rough on an atomic scale and when placed in contact touch only at
the tips of their asperities. The real area of contact will generally be much smaller than the
apparent. At these regions of real contact, if the surfaces are clean, the atoms on one surface
will attract those on the other and produce strong adhesion. With metals this may be referred
to as cold welding and is particularly marked in mechanisms operating in high vacuum, e.g.
outer space. When sliding occurs these adhesions have to be overcome, that is, the junctions
have to be sheared. The force to shear the junctions is the primary cause of the friction
between clean surfaces. If in addition one surface is harder than the other the roughness on it
will plough out grooves in the softer and this constitutes a second cause of friction. In general
we may write:
Force of Friction=Force to shear junctions + Force to plough the asperities on one
surface through the other,
Or F = F adh + F ploughing … (1)

7
 For clean surfaces the adhesion easily dominates. For lubricated surfaces the
ploughing term may be an appreciable fraction of F.
With metals the junctions have shear strength comparable to the bulk shear strength
of the softer of the two metals in sliding contact. With similar metals, the junctions, due to
heavy work-hardening during shear, are stronger than both and large fragments may be torn
out of both surfaces. For this reason it is generally bad practice to slide similar metals
together. If they must be similar it is desirable to choose hard materials which do not work
harden further during sliding and which have limited ductility: or to use materials of non
homogeneous structure such as cast iron.

THE LAWS OF DRY FRICTION

The adhesion component of friction can be estimated as follows. If the real area of
contact is A and the specific shear strength of the junction is s, we have
F = As … (2)

For a given sliding pair s is approximately constant. Since A is nearly proportional to

the load for a very wide range of surface conditions (see Section F1), and does not depend
markedly on the overall size of the bodies, the frictional force too is proportional to the load
and independent of size of the bodies. These are the two basic laws of dry friction.
With metals where the contact regions flow plastically
W
A= … (3)
þ
Where p is the yield pressure or hardness of the softer metal.
Then the coefficient of friction µ is given by
F As s
µ= = = … (4)
W W þ
For ductile solids þ is of order 5 s so that this gives a value µ = 0.2. Most metals in air have a
value nearer µ = 1. There are two reasons for this. First, work-hardening during shear may
increase s in the surface layers without appreciably affecting þ. Secondly, under the influence
of the combined normal and tangential stresses junction growth may occur producing an area
of contact much larger than that given by eqn(3). This can be very marked with clean ductile
metals and can lead to enormous coefficient of friction or even to gross seizure.
With non-metals the laws of deformation may give a different dependence of A on W
than that given by eqn (3). In addition “s” itself may depend on the contact pressure. With
polymers and rubbers both these factors usually lead to a coefficient of friction which
decreases with increasing load.

EFFECT OF LUBRICANTS AND SURFACE FILMS

The most effective way of reducing friction and wear is to prevent contact between
the surfaces. This is achieved, ideally, if the surfaces can run under hydrodynamic or elasto-
hydrodynamic conditions with the equilibrium film thickness greater than the height of the
surface asperities. In this case the chemical nature of the lubricant is immaterial, only its
viscous properties are relevant. However, if contact occurs through the liquid film it is
desirable to add to the lubricant, additives of specific chemical properties. For example, polar
molecules can adsorb on the surfaces to give a thin protective film which will prevent the
surfaces themselves from coming into contact. If the rubbing surfaces are metals, fatty acids
which can react to form metal soaps are particularly effective. Again, if the lubricant contains
reactive chlorine of sulphur compounds these will react to give protective metal chlorides and
sulphides.
The action of these surface films is, in principle; provide a surface layer which is very
much easier to shear than the underlying solids. For example, if hard metal surfaces are
covered with a very thin film of a softer metal, the area of contact A will be determined by the
hardness of the underlying metal whereas shearing will occur in the weaker metal layer. Thus
a thin film of indium on tool steel can reduce the friction by a factor of 10. Such films are
gradually worn away. There are two ways of replenishing them. One is by incorporating the
soft metal as a separate phase in the harder matrix. The soft metal is extruded by the rubbing

8
process itself. This is essentially the way in which copper-lead alloys function. The other is to
fall back on a chemically-formed surface film which can replenish as it is worn away. This
involves the use of ‘oiliness’ or extreme pressure additives dissolved in the lubricating oil.
Here it is important to avoid too reactive an additive since it may lead to excessive corrosion.
Another approach to the reduction of friction and wear is to coat the rubbing surfaces
with lamellar materials such as graphite or molybdenum disulphide. These substances are
strong in compression and weak in shear. However, because the cleavage planes are low-
energy surfaces one of the major difficulties is to achieve strong adhesion to the underlying
solids.
Yet another way of reducing friction, and especially wear, is to coat the solids with
layers of very hard materials such as those that are produced by plasma or flame spraying.
There is evidence that because of their extreme hardness the surfaces are scarcely deformed
during sliding so that even adsorbed films of oxygen or water vapour can provide some
lubricating action. In addition, because of their limited ductility they will tend to be wear-
resistant. On the other hand, if ductility is completely absent the films may crack and fragment
during sliding and the consequent behavior will be catastrophic.
Broadly speaking, the coefficient of friction for metals in air, when dry, is of order µ =
1 for the softer metals and µ » 0.4 for harder metals such as steel. For surfaces operating with
a good boundary lubricant such as a fatty acid, µ is of order 0.05 to 0.1, and the amount of
wear may be 1000 to 10000 times less than for unlubricated surfaces. However, this is only
achieved when the adsorbed film is in a condensed form. As the temperature is increased
these films are melted or desorbed and for most surface-active materials the films cease to be
effective as friction- and wear-reducing agents when the surface temperature reaches 150 to
0
200 C. At higher temperatures E.P. additives containing sulphur or chlorine or phosphorus
must be used. The friction is of order µ = 0.1 to 0.2. Under very severe running conditions
oxygen can often provide a useful protective film and in many cases this is provided by air
dissolved in the lubricant itself.

FRICTION AND LUBRICATION OF POLYMERS

Because polymers are viscoelastic materials their deformation characteristics and
hence their frictional properties are load, speed and temperature dependent. However, a
representative value for most thermoplastics at room temperature is µ » 0.4 and there is often
lumpy transfer. In this class of materials PTFE (Teflon) and high-density Polythene are
exceptional. Once some orientation has been produced at the sliding interface the friction
may be extremely small (µ » 0.06) and the transfer consists of extremely thin drawn-out
highly-oriented polymer fibers. If the molecule is modified to incorporate bulky side groups
(e.g. Teflon-HFP copolymer, or low-density Polythene) the low friction and light wear
properties are lost. The special behavior of the unmodified materials is associated with a
smooth molecular profile.
Cross-linked polymers generally have a lower friction than thermoplastics, but they
tend to be brittle and to fragment during sliding.
Because polymers are viscoelastic materials, deformation losses (i.e. the ploughing
term) may be appreciable. At engineering speeds this may lead to excessive heating beneath
the surface causing subsurface failure and spelling. By contrast, frictional heating with metals
may produce softening of the surface layers or structural changes or even the formation of
surface compounds such as carbides oxides or nitrides.
Polymers may be lubricated with boundary films as with metals, though the reduction
in friction is not in general as marked. With some polymers such as Polythene it is possible to
incorporate surface active materials in the polymer itself. (e.g. stearamide or oleamide); these
can diffuse to the surface and provide a lubricating film which is replenished by further
diffusion as the film is worn away.
Rubber is an extreme case. The friction is usually high and increases with softness of
the rubber. This is mainly because a soft rubber gives a larger area of true contact. Values of
µ greater then 1 are often observed, e.g. with types on road surfaces. At low speeds grease
greatly reduces the friction: at higher speeds water provides effective hydrodynamic
lubrication since the contact pressures are so low.

9
1.6 VISCOSITY AND RHEOLOGY
Rheology is the study of the flow of materials under stress. In tribology, the flow
properties of liquid lubricants are of primary interest.
Viscosity is a measure of the internal friction of a fluid. It is the most important
physical property of a fluid in the context of lubrication. The viscosity of a lubricant varies with
temperature and pressure and, in some cases, with the rate at which it is sheared.
Dynamic viscosity is the lubricant property provides a relationship between the shear
stress and the rate of shear which may be expressed as:
Shear stress = Coefficient of Dynamic Viscosity x Rate of Shear

RHEOLOGY OF NON-NEWTONIAN LUBRICANTS

If the viscosity of a fluid is independent of its rate of shear, the fluid is said to be
Newtonian. Minerals lubricating oils and synthetic oils low molecular weight are Newtonian
under almost all practical working conditions.
Polymeric liquids of high molecular weight (e.g. silicones, molten plastics, etc.) and
liquids containing such polymers may exhibit non Newtonian behavior at relatively low rates of
shear. This behavior is shown diagrammatically.
Grease are non Newtonian in the above sense but, in addition, they exhibit a yield
stress the magnitude of which depends on their constitutions. The stress/strain rate
characteristic for typical grease is also indicated in fig – 1.11.

Fig. 1.11 Shear stress / Viscosity / shear rate characteristics of non-Newtonion liquids.

1.7 MODES OF LUBRICATION

Before considering the methods by which fluid films can be formed between bearing
surfaces it is necessary to distinguish between the three major modes of lubrication;
boundary, mixed and fluid film. The friction and wear characteristics of bearing often indicate
the mode of lubrication and a useful further guide is given by the ratio.
Thickness of lubricating film or protective layer
R = ---------------------------------------------------------
∑surface roughnesses (cla)

This ratio not only indicates the mode of lubrication, but it has a direct bearing upon
the effective life of lubricated machine components.

Boundary lubrication
The friction and wear characteristics of the lubricated contact are determined by the
properties of the surface layers, often of molecular proportions, and the underlying solids. The
viscosity of the bulk lubricant has little effect upon the performance of boundary lubricated
contacts and the frictional behavior broadly follows the well-known laws for unlubricated
surfaces. This mode of lubrication is encountered in door hinges and many machine tool slide
ways.

10
 The thickness of the protective layers formed by physical and chemical reaction
between the solids and the surrounding bulk lubricant, additives or atmosphere is usually
small compared with the roughness of the solid surfaces. The length of fatty acid molecules,
which are frequently used as boundary lubricants, and the thickness of protective oxide films
-7
is often as small as 2nm (10 in). Hence, for boundary lubrications.( Fig. 1.12)
R≤1

Fig. 1.12 Boundary lubrication Fig. 1.13 Mixed lubrication

Mixed lubrication
Although it is usual to classify the mode of lubrication in many machine elements as
either ‘boundary’ or ‘fluid film’, it is not generally known that a very large proportion of them
operate with a mixture of both mechanisms at the same instant. There may be region of close
approach where surface interactions and boundary lubrication contributes to the overall
friction and wear characteristics in addition to a substantial fluid film lubrication action from
most of the contact. In addition it is recognized that local hydrodynamic effects between
surface irregularities can contribute to the total load carrying action; a mechanism known as
asperity lubrication.
In mixed lubrication it is necessary to consider both the physical properties of the bulk
lubricant and the chemical interactions between the bulk lubricant or additives and the
adjacent solids.
This mode of lubrication is encountered in many gears ball and roller bearing, seals
and even some conventional plain bearings. It is now recognized that it is difficult to eliminate
‘fluid film’ action from boundary lubrication experiments and ‘boundary’ effects occur in ‘fluid
film’ investigations more often than is generally acknowledged. This indicates the growing
importance of recognition of the regime of ‘mixed’ lubrication.
If mixed lubrication is to be avoided there must be no possibility of asperity interaction
and hence the total fluid and boundary film thickness must exceed the sum of the surface cla
values by a factor which varies from about two to five depending upon the method of surface
manufacture. There is a possibility of mixed lubrication whenever, (fig. 1.13)
R≤5

Fluid film lubrication

The best way to minimize wear and surface damage in rolling or sliding contacts in
machines is to separate the solids by a film of lubricant. The lubricant can be a liquid or a gas
and the load supporting film can be created by the motion of the solids (self-acting or
hydrodynamic) or by a source of pressure outside the bearing (externally pressurized or
hydrostatic).
The main feature of this mode of lubrication is that the bearing solids are separated
by a fluid film which is considerably thicker than the dimension of the surface irregularities or
protective surface films formed by boundary lubricants or chemical reaction. The films are
normally many thousands of times thicker than the size of single molecules and this allows
them to behave like the bulk lubricant applied to the bearing. They can nearly always be
analyzed according to the laws of slow viscous flow and the frictional resistance arises solely
from the viscous shearing of the fluid. Viscosity of the bulk lubricant is the most important
physical property of fluid film lubricants, but density is also important in gas bearings and
some very highly stressed liquid lubricated contacts. It can usually be assumed that liquid
lubricants adequately wet solid bearing surfaces and the main role of surface tension in fluid

11
film lubrication is found in the mixed phase flow regions where cavitations has occurred in
bearings, its effect on foaming and on viscous lifting in devices like ring oiled bearings.
Since a requirement for successful fluid film lubrication is the absence of asperity
interaction the film thickness is usually at least two to five times greater than the sum of the
surface cla values. It is more difficult to give an upper limit to the film thickness in relation to
roughness, since there is no physical boundary to the definition presented earlier. However,
thick films do not normally carry much load and in engineering situations it is rare to find a
fluid lubricating film thicker than one hundred times the sum of the surface rough nesses.
Fluid film lubrication can reliably be expected when R ≥ 5 and hence, in engineering
situations,
5 ≤ R ≤ 100

A special form of fluid film lubrication in which the development of effective films is
encouraged by local elastic deformation of the bearing solids is known as
‘elastohydrodymanic lubrication’
It is now recognized that this is the principal mode of lubrication in many gears, ball
and roller bearings, cams and some soft rubber seals. In counter formal contacts where the
local pressures are high an additional important feature is the effect of high pressure upon the
viscosity of many liquid lubricants. In all cases the elastic deformation creates a near parallel
lubricating film in the central region of the conjunction with a restriction giving the minimum
film thickness near the outlet or sides of the effective load bearing region.
Many nominal elastohydrodynamic contacts operate with film thickness which is small
by conventional fluid-film lubrication standards and there is often an associated ‘mixed’
lubrication action. If local elastic distortion of the bearing surfaces occurs in the load carrying
region the range of effective fluid film lubrications is often extended well into the region
normally associated with ‘mixed’ or even ‘boundary’ lubrication. A typical range of film
thickness ratios for elastohydrodynamic conditions (Fig. 1.15) is given by,
1 ≤ R ≤ 10

Fig. 1.14 Fluid Film lubrication Fig. 1.15 Elastohydrodynamic lubrication

1.8 MECHANISMS OF WEAR

Wear can be defined as the progressive loss of substance resulting from mechanical
interaction between two contacting surfaces. In general these surfaces will be in relative
motion, either sliding or rolling, and under load. Wear occurs because of the local mechanical
failure of highly stressed interfacial zones and the failure mode will often be influenced by
environmental factor. Surface deterioration can lead to the production of wear particles by a
series of events characterized by adhesion and particle transfer mechanisms or by a process
of direct particle production akin to machining or, in certain cases, a surface fatigue form of
failure. These three mechanisms are referred to as adhesive, abrasive and fatigue wear and
are the three most important.
In all three cases stress transfer is principally via a solid-solid interface, but fluids can
also impose or transfer high stresses when their impact velocity is high. Fluid erosion and
cavitations are typical examples of fluid wear mechanisms. Chemical wear has been omitted
from the list because an environmental factor, such as chemical reaction, influences almost
every aspect of tribology and it is difficult to place this subject in a special isolated category.
Chemical reaction does not itself constitute a wear mechanism; it must always be
accompanied by some mechanical action to remove the chemical products that have been
formed. However, chemical effects rarely act in such a simple manner; usually they interact
with and influence a wear process sometimes beneficially and something adversely.

12
Fig. 1.16 A simplified picture of adhesive wear Fig.1.17 Some aspects of adhesive wear

FLUID AND CAVITATION EROSION

Both these wear mechanisms arise from essentially the same cause, namely the
impact of fluids at high velocities. In the case of fluid erosion, the damage is caused by small
drops of liquid, whilst in the case of cavitations; the impact arises from the collapse of vapor
or gas bubbles formed in contact with a rapidly moving or vibrating surface.
Fluid erosion frequently occurs in steam turbines and fast flying aircraft though the
impact of water droplets. The duration of impact is generally extremely small so that very
sharp intense compression pulses are transferred to the surface material. This can generate
ring cracks in the case of such brittle materials as Perspex, or form plastic depression in
surface. As the liquid flows away from the deformation zone, it can cause strong shear
deformation in the peripheral areas. Repeated deformation of this nature gives rise to a
fatigue form of damage and pitting or roughening of the surfaces soon becomes apparent.
With cavitations erosion, damage is caused by fluid cavities becoming unstable and
collapsing in regions of high pressure. The cavities may be vaporous, or gaseous if the liquid
contains a lot of gas. The damage caused by the latter will be less than the former. The
physical instability of the bubbles is determined by the difference in pressure across the
bubble interface so that factors such as surface tension and fluid vapour pressure become
important. The surface energy of the bubble is a measure of the damage which is likely to
occur, but other factors such as viscosity play a role. Surface tension depressants have been
used successfully in the case of cavitations attack on Diesel engine cylinder liners. Liquid
density and bulk modulus, as well as corrosion, may be significant in cavitations, but since
many of these factors are interrelated it is difficult to assess their individual significance.
Attempts to correlate damage with material properties have lead to the examination of the
ultimate resilience characteristic of a material. This is essentially the energy that can be
dissipated by a material before any appreciable deformation or cracking occurs and is
measured by ½ (tensile strength) 2/elastic modulus. Good correlation has been shown with
many materials. The physical damage to metals is of a pitting nature and obviously has a
fatigue origin.( Fig. 1.18)

13
 Fig. 1.18 Wear by fluids containing abrasive particles

1.9 HEAT DISSIPATION FROM BEARING ASSEMBLIES

Heat is dissipation from a bearing assembly by:
(a) Heat transfer from the bearing housing, Hh
(b) Heat transfer along and from the shaft, Hs
(c) Heat transfer to a lubricant/coolant flowing through the assembly, H1

HEAT TRANSFER FROM THE BEARING HOUSING Hh

Hh = h Asf (θb – θa)
Where h is the total heat transfer coefficient (see Fig.1.21), As is the housing surface
area and can be estimated from the surface area of an annular disc of similar overall size
(see examples in Fig. 1.23), f is a factor which depends and can housing internal and external
thermal resistances and can be estimated from the dimensions of the equivalent annular disc
(see Fig. 1.24), θa is the temperature of the surroundings and θb is the bearing temperature.

Notes: 1. θb is not normally known; values of θb must be assumed initially to obtain a curve
relating Hh to θb ; this curve, together with other relating Hs, H1 and the heat generation
in the bearing to θb, can then be used to determine the actual value of θb (see
paragraph entitled Bearing Operating Temperature).
2. For housing which are an integral part of a machine or structure some experience
and judgment may be required to decide appropriate effective boundaries of the
housing and hence the size of the equivalent disc. Joints between the housing and
the rest of the structure can only be taken as boundaries of the housing if they lie
across the direction of heat flow and have low conductivity. Factors which reduce joint
conductivity are low interface pressure (e.g. bolted joints), rough surface finish and
low temperature difference across the joint.

14
HEAT TRANSFER ALONG AND FROM THE SHAFT, Hs
The shaft may either supply heat to the bearing assembly or remove heat from it,
depending on the temperature and location on the shaft of any other heat sources or sinks
(such as the rotor of an electric motor or the impeller of an air circulating fan). In the absence
of heat sources, the shaft will remove heat from the bearing. In estimating the heat transfer to
or from the assembly the parts of the shaft extending on each side of the bearing must be
considered separately and their contribution added.
Most practical situation is covered by one or other of the two cases following.

Case 1 Heat source or sink of known temperature on the shaft

This case includes any situation where the shaft temperature is known or can be
assumed at some known distance from the bearing.

Fig. 1.19 Heat source or sink of known temperature on the shaft

Heat flow from the bearing is given by:

kA
Hs = ------ {C1 (θb-θa) + D1 (θa-θs )}
L
Where k = thermal conductivity of shaft material
A = shaft cross-sectional area
θa = temperature of surrounding air
θs = temperature of heat source or sink
θb= bearing temperature (see Note 1 opposite)
C1 and D1 are factors (Fig.1.22) depending on mL
1/2
Where m = (4h/kD) for solid shafts, diameter D,
and h = total heat transfer coefficient for shaft (Fig. 1.21)
Note that H2 may be negative, denoting heat flow into the bearing, if θs is sufficiently high.

Case 2 Shaft end free or insulated

This case includes situation in which either the shaft is thermally insulated at some
section, or the heat flow along the shaft can be assumed negligible at a specified section, or
the shaft end is free.

Fig. 1.20 Heat transfer when shaft end free or insulated.

15
Heat flow from the bearing is given by :

kA
Hs = -------- {C2 (θb-θa)}
L
Where the terms have the same meaning as in Case 1 except for:
L = L' + D/4 if the shaft end is free,
Or L = L' if the shaft end is insulated,
And C2 is another dimensionless factor depending on mL

(b)
Fig 1.21 Total heat transfer coefficient h = h c + h r : hc is obtained from (a) and hr from(b)

16
 Fig 1.22 Correlations between C1 C2 D1 Vs mL

HEAT TRANSFER TO A LUBRICANT / COOLANT. Hl

Hl = Qρc (θO – θ1)

Where Q = volume flow rate

ρ = density of lubricant / coolant
c = specific heat of lubricant / coolant
θ1= inlet or supply temperature
θO = outlet temperature

Provided the flow rate is not excessive, θ O will normally be nearly equal to the bearing
temperature, θb

Fig 1.23 Equivalent discs for three simple housing types

17
 Fig. 1.24 f Factors for housings

EXPERIMENTAL METHOD FOR DETERMINING THE HEAT DISSIPATION

FROM BEARING ASSEMBLIES
An experimental method can be used where calculation methods are unreliable, and
normally consists of substituting an electrical heating source for the frictional heat generation
at the bearing surfaces. For good accuracy the installation and operating conditions used in
the test must approximate those intended for service as closely as possible and direct heat
transfer from the heater to the surroundings must be avoided. The ready bearing
temperatures (measured near the bearing surface, corresponding to a series of recorded
electrical power inputs, should be noted, and can be used to plot a curve of heat dissipation
against bearing temperature rise above ambient, which is the required heat dissipation
characteristic for the assembly.

BEARING OPERATING TEMPERATURE

The normal purpose of determining the heat dissipation characteristics of bearing
assemblies to estimate the bearing operating temperature for which the procedure is as
follows:
1. Decide the range within which the bearing temperature, θb is expected to lie.
2. For several values of θb covering this range, determine the total heat dissipation
from the bearing assembly Hd, either by the experimental method (see above) or
by adding together Hh , Hs and Hl as calculated by the methods given earlier. Plot
Hd against θb.
3. For the type and design of bearing under consideration use a appropriate
sections this Handbook to estimate the friction coefficient, and hence frictional
heat generation, Hf in the bearing. Estimate any other heat inputs to the
assembly, for example, due to friction at bearing seals, and add these to Hf to
obtain to total heat generation, Hg . If Hg varies with θb calculate Hg at several
values of θb against on the same axes us used for Hd above.
4. The bearing operating temperature is that corresponding to the intersection of the
curves for Hd and Hg.

18
Methods for reducing the operating temperature of bearings:
These include:
1. Positioning the housing for unrestricted access of the surroundings air to as much
as possible of its surface.
2. Provision of substantial heat flow paths, unimpeded by joints, between the
bearing surface and housing exterior.
3. Use of a material for the housing having a high thermal conductivity.
4. Fins on the housing surface.
5. Forced air cooling – often conveniently achieved by means of a fan on the shaft.
6. Internal cooling – for example by water coil or passages.
7. Fitting thermal insulation in any joints between the housing and hot structure or
components.
8. Maintain cleanliness and recommended viscosity of oil.

1.10 OIL WHIRL AND WHIP INSTABILITIES - WITHIN

JOURNAL BEARINGS
Oil whirl is probably the most common cause of sub synchronous instability in
hydrodynamic journal bearings. Typically, the oil film itself flows around the journal to lubricate
and cool the bearing. This develops an average speed slightly less than 50 percent of the
journal surface speed (Figure 1.25).
Normally, the shaft rides on the crest of an oil pressure gradient, rising slightly up the
side of the bearing somewhat off vertical at a given, stable attitude angle and eccentricity. The
amount of rise depends on the rotor speed, rotor weight and oil pressure. With the shaft
operating eccentrically relative to the bearing center, it draws the oil into a wedge to produce
this pressurized load-carrying film.

Figure 1.25 Oil Film within a Journal

If the shaft receives a disturbing force such as a sudden surge or external shock, it
can momentarily increase the eccentricity from its equilibrium position. When this occurs,
additional oil is immediately pumped into the space vacated by the shaft. This results in an
increased pressure of the load-carrying film, creating additional force between the oil film and
shaft. In this case, the oil film can actually drive the shaft ahead of it in a forward circular
motion and into a whirling path around the bearing within the bearing clearance. If there is
sufficient damping within the system, the shaft can be returned to its normal position and
stability. Otherwise, the shaft will continue in its whirling motion, which may become violent
depending on several parameters.

19
OIL WHIRL INSTABILITY

Oil whirl demonstrates the following characteristics:

1. Oil whirl can be induced by several conditions including:
· light dynamic and preload forces
· excessive bearing wear or clearance
· a change in oil properties (primarily shear viscosity)
· an increase or decrease in oil pressure or oil temperature; improper bearing design
(sometimes an over design for the actual shaft loading)
· fluid leakage in the shroud of blades and shaft labyrinth seals (so-called “Alford force”
or “aerodynamic force”)
· change in internal damping (hysteretic, or material damping, or dry (coulomb) friction)
· Gyroscopic effects, especially on overhung rotors with excessive overhang.
· Any of these conditions can induce oil whirl after a disturbing force induces an initial
rotor deflection.
2. Sometimes machines exhibit oil whirl intermittently due to external vibratory forces
transmitting into the unit or from sources within the machinery itself. In these cases, these
vibratory forces have the same frequency as the oil whirl frequency of that bearing and can
contribute just the right magnitude of disturbing force at just the right tuned frequency to set
the shaft into the whirl motion. This vibration transmits from other machinery through attached
structures such as piping and braces, or even through the floor and foundation. If this occurs,
it may be necessary to either isolate this machine from surrounding machinery, or to isolate
the offending machine itself.

3. Oil whirl is easily recognized by its unusual vibration frequency which 40 percent to
48 percent is generally of shaft RPM. (Reference 2 states that pure oil whirls occur at 43
percent of shaft speed, but that the instability may occur at the first critical speed.)

4. Figure 2 shows the development of oil whirl just after the shaft is brought up to speed.
Note that the shaft went into whirl at a machine speed of approximately 1,800 RPM and
remained in whirl until about 4,000 RPM. At this point, note that the shaft 1X RPM speed
entered resonance, which actually generated sufficient force to overcome oil whirl. However,
once the machine passed through resonance, whirl once again occurred just above 5,200
RPM.

Figure 1.26 Development of Oil Whirl Just After Startup; Followed by Oil Whip
from 9,200 to 12,000 RPM
5. Oil whirl is considered severe when vibration amplitudes reach 40 to 50 percent of the
normal bearing clearance. At this point, corrective action must be taken.

20
6. Temporary corrective measures include changing the temperature of the oil (thus, the
oil viscosity), purposely introducing a slight unbalance or misalignment to increase the
loading, temporarily shifting the alignment by heating or cooling support legs, scraping the
sides of or grooving the bearing surface to disrupt the lubricant wedge, or changing the oil
pressure.

7. Permanent corrective steps to resolve the oil whirl problem include installing a new
bearing shell with proper clearances, preloading the bearing by an internal oil pressure, or
completely changing the bearing type to oil film bearings that are less susceptible to oil whirl
(including axial-groove bearings, lobed bearings or tilting pad bearings). The tilting pad
bearing is a good choice because each segment or pad develops a pressurized oil wedge
tending to center the shaft in the bearing, thereby increasing the system damping and overall
stability. Whip frequency never changed even though the machine continued up in speed to
12,000 RPM. When a shaft goes into oil whip, its dominant dynamic factors become mass
and stiffness in particular; and its amplitude is limited only by the bearing clearance. Left
uncorrected, oil whip may cause destructive vibration resulting catastrophic failure – often in a
relatively short period of time.

DRY WHIP
Dry whip occurs in journal bearing machines subjected to either a lack of lubrication
or the use of the wrong lubricant. When this occurs, excessive friction is generated between
the stationary bearing and rotating journal. This friction can excite vibration in the bearing and
other components. This kind of vibration is called dry whip. Dry whip can also be caused by
journal bearings having excessive clearance as well as those having insufficient clearance.

The dry whip condition is similar to rubbing a moistened finger over a dry pane of
glass. It will generate a frequency specifically dependent upon the shaft and construction
materials, geometries and lubricant properties. Normally, this frequency will produce a high
squealing noise similar to that generated by dry rolling element bearings. The frequency
content itself will not be an integer multiple of the machine speed. When dry whip is
suspected, it must be taken care of quickly in order to prevent a potential catastrophic failure.
And when dry whip is suspected, both the lubricant itself and the lubrication system should be
closely inspected and the bearings should be checked to ensure they have proper
clearances.

21