Tribological Research and Design for Engineering Systems D. Dowson et al. (Editors) 9 2003 M. Neale. Published by Elsevier B.V. All rights reserved.
Michael Neale OBE. FREng. FIMechE. Neale Consulting Engineers Ltd. Highfield, Pilcot Hill, Dogmersfield, Hampshire. RG27 8SX.
ABSTRACT This paper reviews the contribution that tribology can make to machine design. This is focussed on the three most common types of contact involving machine components. These are area contacts, as in plain bearings, concentrated contacts, as in rolling bearings, and distributed contacts with hard materials, as in abrasive wear. It indicates for each of these how basic understanding can lead to some innovative design solutions. Tribological research itself is also an area where there are considerable opportunities for new methods to be applied, particularly in the rational analysis of extensive field experience data, which can lead to major advances in the understanding of the phenomena involved.
1. TECHNOLOGIES USED IN MACHINE DESIGN Machines consist of an assembly of component parts, and these parts may be connected rigidly together, or may be arranged to allow movement relative to each other. The technologies which underlie this basic structure are therefore, stress analysis, for the strength and rigidity of the parts, zygology for the fixed joints, and tribology for the components which experience relative motion. In addition, since the complete machine is required to perform some particular function, other technologies influence its layout. For example, fluid mechanics is applied to the design of pumps and turbines, heat transfer and combustion technology to engines, and electromagnetism to electric motors and generators. In terms of reliability, tribology is the most ~ritical, because the st/rfaces of machine
components, which operate with relative motion, tend to carry a combination of high loads and sliding speeds. They therefore have the potential, in the event of any malfunction, for the release of large amounts of destructive energy, capable of inducing component failure. Also there is always the possibility of wear of the working surfaces, which can either lead to leakage, as in seals and piston rings, or to the reduced performance of load carrying lubricating films, as in plain bearings. Surface fatigue can also occur on components with concentrated contacts, such as rolling bearings and gears. This paper attempts to review tribology, in terms of its potential impact on machine design, and the guidance that it can provide. The main areas of this impact are considered to be :a) An understanding of the contact conditions at the moving surfaces, the materials from which they can be made, and the action of any separating fluid films.
b) An understanding of methods for collecting information on the tribological performance of machines, such as the wear rates of their components in service, and their need for maintenance.
2. CONTACT C O N D I T I O N S MOVING C O M P O N E N T S
In most machines there are three types of contact with components that are subject to relative movement :a) A contact of substantial area, as in plain bearings and seals. b) A concentrated contact of limited area, as in rolling bearings, gears, and cams and followers. c) A distributed number of contacts with a hard material as in most abrasive wear situations.
2.1. A r e a c o n t a c t s
In any interaction between two surfaces, which operate in direct sliding contact, there is always some risk of seizure between the two. This is particularly likely if the two materials in contact are similar in terms of hardness and melting point, in that any small local surface melting, at a high spot between the contacts, can more easily spread to the whole contact area. While similar materials are sometimes used for manufacturing convenience, as in drive chains between two sprockets, these will only work if the sliding speed between the components is very slow and intermittent, as it often is in that application. The surfaces of the pins and bushes also need to be kept well lubricated. This type of technology does work adequately in applications such as bicycles, where the
chain can be replaced fairly easily at low cost. However if chains are used in continuous service, as in the step drives of public service escalators, the use of steel pins in steel bushes is unsatisfactory. This is because, even if the chains are regularly lubricated, they are still prone to local wear and the resulting debris inhibits the access of new lubricant to the surfaces. This can then develop into major wear and possible seizure on any one bush. The real problem in this application, however, is that each escalator has several hundred chain bushes, any one of which can move into a problem mode at any time. This makes it virtually impossible to carry out effective maintenance at extended regular intervals, and forces the operation into a breakdown maintenance mode, with failures at random intervals. The tribological solution to this problem is to use two different materials in the rubbing contact. The use of reinforced plastic bushes containing solid lubricants eliminates the random failures and results in slow steady and predictable wear, giving a chain life of about 40 years. While it might be expected that this technology would be applied universally, it tends not to happen if the purchase team on a new construction is given the target of achieving the lowest construction cost, as distinct from the lowest total life cost. This is because a chain with plastic bushes tends to cost about 40% more than one without. There are many other applications of plain bearings which operate at relatively low sliding speeds, and often with reciprocating motion, such as in vehicle suspension and steering, and in mechanical handling equipment. In all these applications, unlubricated non-metallic rubbing bearing materials can provide an optimum design solution. Typically these are composed of a relatively soft non-adhesive core material contained in a matrix of stronger material or reinforced by fibres of a strong material.They may also contain added solid lubricants such as graphite or molybdenum disulphide. An outline of their performance is given in
Figure 1. Since they do allow unlubricated operation, and the absence of oil can allow corrosion, it is important that the mating surface is made from a non-corrosive relatively hard material, such as a chromium plated or stainless steel shaft.
. . . . . .
V, ft/min 100 1000
HAR~)STEEL 100 ~ H A R D
S1EEL- - I0 0 0 0
"'. ~"~-'PTFE FABRIC E I0 z
-,','i" \ ,.... o
'~ r el
w P T,F, E.
REINFORCED ~ THERMOSETS ,' t ~, FILLEI') PTFE
! V, m/s
Figure 1. The application areas of rubbing bearing materials for oscillating movement The major use of bearings with area contact is, however, in fully lubricated applications ~, such as in pumps, turbines, large gearboxes and electrical machines, and in internal combustion engines. In all these cases the tribological understanding of the operation of hydrodynamic films has made it possible to achieve an optimum fluid film plain bearing design. In the case of high speed rotating equipment this has included an understanding of rotor dynamics and its interaction with the bearing design, leading in many cases to the use of tilting pad journal bearings. In the case of engine bearings it is understood how their very high dynamic loads can be carried with adequate oil films, because their short duration is less than the time it takes to squeeze the oil from the bearing clearance. For the same reason, it is recognised that the process of refilling the bearing clearance with oil, at times of load reversal, is also critical and determines where
oil feed drillings and grooves need to be positioned. An understanding of the operating conditions of these bearings also indicates the kind of materials from which they should be made. For the steadily loaded bearings in large machines, the film thickness is small relative to the bearing width and the amount of misalignment that can occur due to shaft or housing deflections. This therefore requires a relatively soft bearing material to allow the shaft to bed into it locally without surface damage. However if it is too soft it could extrude sideways under load and it therefore requires some form of fibre reinforcement. These properties are provided by whitemetal, which consists of a soft tin core reinforced by a matrix of copper-tin needles. It is also common to add antimony to this material to create distributed tin-antimony cuboids, which by providing local harder particles in the surface, can increase the wear resistance of the material. The structure of whitemetal is shown in Figure 2.
Figure 2. The structure of whitemetal For high performance engine crankshaft bearings, with much higher loads, it is necessary to use a different type of material consisting of a strong matrix permeated with a softer material which can smear over the surface in the event of any shaft contact. Materials such as copper-lead, lead-bronze or aluminium-tin are typical materials of this kind. They are harder than whitemetal and
therefore usually require a hardened shaft in order to reduce shaft wear, or any damage in the event of a rubbing contact between the shaft and the bearing. While the development of plain bearings has mainly related to the use of oil lubrication, the fact that the operation of the fluid films is largely understood, makes it possible to lubricate machines with liquids other than oil. This is particularly significant for machines such as pumps and turbines which already operate with some kind of fluid, and if this fluid is used to lubricate the bearings, some significant design improvements can be achieved. A good example of this is in boiler high pressure water circulation pumps where a novel and effective design can be achieved by using the water to lubricate the bearings. This not only eliminates the need to have shaft seals but allows the whole pump, and its driving motor, to be contained in a static pressure vessel, to which the only connections are the input and delivery pipes, and electrical power leads. With no exposed shafts there is no possibility of leakage. There are, therefore some exciting possibilities for producing better and revolutionary machine designs and encouraging some free thinking by designers about component dimensions and layouts. In machine design, the shaft diameter is usually determined by its required torque capacity or by its required lateral stiffness, to avoid excessive deflection or too low a dynamic flexibility. With oil lubrication and a typical operating viscosity of 10 to 20 centipoises, the required bearing length is slightly less than the shaft diameter. This has resulted in established machine layouts, which provide the starting point for most designs. However, when designing for the use of different fluids for lubrication, it is important to remember that the bearing design equations indicate a link between the lubricant viscosity and the required shaft diameter.
In hydrodynamic bearings the simplest nondimensional factor for the operating conditions, which relates to the geometry of the lubricant film, ie. the eccentricity ratio is proportional to :ZN P
Where Z = lubricant viscosity N = rotational speed P = bearing load pressure =W DL W=bearing load D=bearing dia. L=bearing length
When designing a bearing with a different lubricant, the rotational speed and the load do not change, ie. N and W are constant. So the lubricant film geometry is proportional to ZDL. For the same film geometry, DL needs to increase by the same proportion that Z is decreased. If the bearing geometry, in terms of D/L remains the same, then:ZD ~ should be constant, ie. D should be proportional to 1/square root of Z Figure 3. shows~ what this means in terms of the possible redesign of two machines, a gas turbine and a fuel pump, which with normal lubricating oils operate with shaft diameters of 100mm and 25mm. An interesting conclusion from this, is that it might be possible to design a gas turbine with air bearings outside the rotor so that the turbine blades were always in compression and therefore able to operate at higher temperatures and higher efficiencies. In fact this diagram does tend to slightly exaggerate the required shaft diameter, because it assumes a constant eccentricity ratio, which corresponds to an increase in minimum film thickness in proportion to the shaft diameter. In practice although the allowable minimum film thickness will need to increase slightly at higher shaft surface speeds, it will not be as much as this, so the optimum shaft diameters are slightly exaggerated.
--I.~2-a~"" WATER I I
I AVTUR I I
! I I
10 cp 20 cp
I [-. <
i "%,,! ; i
100 mm dia. GAS TURBINE
mechanism only works with fluids, such as lubricating oils which have a high pressure coefficient of viscosity. Other fluids such as water do not have these characteristics, so there is a more restricted choice of working fluids that can be used with component contacts of this kind.
Figure 3. The variation of optimum shaft diameter with lubricant viscosity
Tribology has also provided an understanding of some of the wider aspects of the methods of operation of concentrated contacts. Gear teeth, for example, operate with patterns of combined rolling and sliding, as shown in Figure 4. direction . ~ ' '~ ~
The bearing materials commonly used in machines lubricated with fluids other than oil are often similar to the materials used for dry rubbing bearings or in the case of water lubrication may be ceramics which tend to operate well with water. Rubber-like materials can also be used as these tend to deflect in the higher pressure parts of the hydrodynamic film and provide closer clearance sealing zones around the edges.
C o n ce n t r a te d contacts
The other common type of tribological contact is where the load is transmitted across a small concentrated contact area, as in rolling bearings and gears. The surfaces therefore need to be hard to carry the loads without permanent deformation, and the material needs to have the minimum number of internal defects to reduce the risk of fatigue failure. The lubrication mechanism is elastohydrodynamic which means that there is some elastic flattening, and local concavity, of the surfaces, which traps the lubricant in this zone of very high pressure. Under these conditions typical lubricating oils increase their viscosity to the point of becoming semisolid, and produce an effective separating film between the surfaces. It is important, however, for designers to remember that this
"~/~.'~/~Opposing surfoce slides Dedendum / ~~.. in these directions _K_ .__~'~ ~ - x , , _ _ t ---~-'--'--" ."t--,~, 9----Line of action ~
Figure 4. The contact conditions between gear teeth The sliding tends to drag the surfaces to form inclined cracks, and if the rolling component of the movement is in the same direction, it will tend to push out any oil that gets into these cracks. If, however, the rolling is in the other direction, the oil will be pushed into the crack and will tend to extend it and generate surface failure by lifting out portions of the surface. It can be seen from Figure 4. that this situation arises in the dedendum of the teeth, and explains why pitting of the teeth usually occurs in these areas. Another rolling contact application where tribological knowledge can help in understanding the detailed nature of the contact, is in guide wheels and rollers used to support large rolling structures. For example, large astronomical telescopes, are often
mounted on rotary tables with wheels about 1 metre diameter. If these wheels do not track properly, large sideways forces can be generated, which can damage the structure, or can cause shocks, which disturb the telescope settings, when the wheels jump sideways to correct their position. This problem arises because cylindrical wheels made from a relatively rigid material, such as metal, will always roll along a line exactly at right angles to their axis of rotation, as shown in Figure 5. It is very difficult, if not impossible, to set them to the required standard of tracking accuracy.
Figure 5. The track patterns of rolling wheels Once this problem is understood, design solutions can be devised, such as the use of a central location bearing, and wheels, at a larger track diameter, which have a spherical outer profile, and bearings which allow the wheels to tilt and thus steer themselves back to the correct track, under the influence of any sideways forces that begin to be generated. 2.3. D i s t r i b u t e d material.
A major cause of wear of the parts of machines arises from their contact with abrasive materials. These materials are typically very hard and may also be sharp. The general tribological solution to this component wear problem is to make the components as hard as possible. However,
this can result in them being brittle and therefore liable to fracture. It is therefore generally better to try to obtain a hard surface on a tough core This can be achieved by conventional surface hardening processes or by the use of materials which work harden when subject to surface impacts and deformation. These solutions are particularly effective when the abrasive material is primarily moving along the component surface as in ploughing tools, or the inside of pipes carrying abrasive slurries. If however the movement is more at right angles to the surface due to the direction of the local flow, it is generally better to use an elastic, rubberlike, material which can cause the abrasive to bounce off the surface with the minimum damage. Another common type of problem with abrasive wear occurs when abrasive material gets into the working surfaces of plain bearings. The type of bearings which commonly suffer from this problem are pivot pins and bushes operating in an abrasive environment, as on construction equipment. Both the pin and its mating surface are generally made from iron or steel to obtain adequate strength and the abrasive tends to embed in both surfaces and result in their mutual abrasion and damage. This type of wear is particularly sensitive to the relative hardness of the two surfaces, as shown in Figure 6. It can be seen from this that it is essential to avoid the use of materials of a similar hardness for the pin and its mating surface. The wear resistance is greatly increased if one surface, usually the pin is made as hard as possible, and the other left relatively soft. The abrasive dirt then embeds in the softer material, protecting it from wear, and attempts to wear the very hard mating surface. This results in much lower total wear rates.
To develop new methods of research which concentrate on the collection of data points from the field, as the only real source of a lot of this information
b) To analyse machine performance from a tribological point of view in order to determine patterns of machine deterioration. 3.1 N e w r e s e a r c h m e t h o d s There is a tendency for tribological research to become concentrated on detailed work in laboratories and to regard the collection of data from the field as low grade and empirical because the data does not derive from a controlled experiment. This probably arises from peer review pressure from scientifically oriented reviewers of research projects. It is, however, worth noting t h a t in the social sciences, the analysis of data from the field is quite normal:
Figure 6. The wear rate of two surfaces with abrasive particles trapped between them
3. THE TRIBOLOGICAL P E R F O R M A N C E OF M A C H I N E S When machines are in operation most of the time related features, such as maintenance intervals and machine life, are determined by the tribological components. The main deterioration processes tend to be wear and fatigue, which usually take hundreds or thousands of hours to produce critical levels of deterioration. Because of this situation, there are two areas where tribologists can make original and useful contributions:-
In research on wear, in a laboratory, it is difficult to carry out tests which relate to real situations, because of the need to accelerate the tests in order to match practical time scales. It can be argued t h a t such work is really a waste of time, other t h a n to obtain a few reference points, when there is a large amount of real data out in the field waiting to be collected. It is of course true t h a t any one data point, from the field, or even a small group of them can be unreliable, particularly because the test conditions are not precisely defined. But if enough data is collected, the overall trends can be seen clearly, and a range of useful numerical values can be determined.
into the wear process in these engines the results also give a useful reference to the performance to be expected from the range of engines. This study involved the collection of over 600 data points from engine manufacturers and users all over the world. They provided the data willingly in exchange for a promise that they would receive a copy of the final result, in which only they would be able to recognise their own performance from the large mass of data points. It may be worth pointing out t h a t any research student doing this type of work can find the process very interesting and instructive and can include a reasonable amount of foreign travel. The student will also be extremely employable as soon as the work is over.
Figure 7. The wear rate of the cylinders of internal combustion engines
An example of some work of this kind is shown in Figure 7. which gives a clear indication of the patterns of cylinder liner wear in internal combustion engines, in normal service. It is particularly interesting that the lowest rates of wear occur on the larger engines, which indicates that it is primarily determined by the number of times that the piston rings visit the top dead centre position, and this is less on larger engines because they run more slowly. It is also clear why modern very high speed motorcycle engines tend to have a limited life between overhauls. This type of work can be supported by the examination of specimens of the liners and rings collected in the field to note any significant material or detail design effects, and link in to the metallurgical factors involved. Apart from providing an insight
There are plenty of opportunities for work of this kind and in 2002 perhaps the most obvious would be a similar study of wheel/rail interactions on the worlds railways. The main variables would probably be the radius of the curves, the axle loads, the wheel profiles, the wheel and rail materials, the degree of steering damping on the bogies, and the frequency of traffic. There are hundreds of data points out there waiting to be collected, and the patterns could then be matched to the results of more fundamental work on the tribological conditions in contacts of this kind, and could give them full validation.
3.2 Machine d e t e r i o r a t i o n analysis There is another major contribution waiting to be made by tribologists, and this is to understand the trend patterns of progressive deterioration of machines. This is a tribological study because most of the deterioration mechanisms are tribological. At first sight this might sound rather boring, but it has tremendous significance. This is because, while some machines deteriorate immediately if their maintenance attention is reduced, there are many others with longer response times, which may be up to four years. A high quality study and analytical
report on this subject could become a major international reference document. The reason for this is that current trends in company m a n a g e m e n t place the main emphasis of performance review on the results obtained over one year, to match the standard period for financial reports. If a company owns and operates a large amount of engineering equipment (such as a railway) an examination of its accounts by the management will show that a large amount of the company's expenditure is being spent on maintenance. It may be concluded that this expenditure is excessive and the almost standard way of dealing with this is to reduce the maintenance budget by about 15% and see whether this has any effect. If the equipment is robust, which in many industries it is, there will predictably be no effect. A further reduction will therefore be tried the following year with probably only minor effects. This is the point where the board members should move on using their excellent record of increased profitability. Then after a further one or two years, disaster will strike, possibly with fatalities, and with a very large recovery cost, greatly exceeding the total of any savings that had been made. In any official enquiry, the board will claim that they acted sensibly in the interest of the shareholders, and no reference data was available to prove otherwise. An effective way of preventing this situation in the future is to produce the reference data, and the threat of corporate liability will then ensure that sensible decisions are made, with engineering advice.
4. C O N C L U S I O N This paper has attempted to review the contribution that tribology can make to machine design and operation. It is an extremely important contribution and can involve very interesting and challenging work for those involved. To achieve it, however, tribologists must lift their horizons above work such as the study of two high spots in contact, however important that may be. They should apply their skills to the large amount of internationally important and intellectually challenging work, on practical problems, that is out there waiting to be done. It is to be hoped that this paper may have given some leads on how to do this, and there are a number of excellent papers being presented at this conference, which are moving in this direction.