automobile engine tribology — approaching the surface · pdf filewear 241 2000...

! .Wear 241 2000 193–203www.elsevier.comrlocaterwear

Automobile engine tribology — approaching the surfaceM. Priest), C.M. Taylor

School of Mechanical Engineering, The UniÕersity of Leeds, Woodhouse Lane, Leeds, LS2 9JT, UK

Abstract

There has been relentless pressure in the second half of the 20th century to develop ever more fuel efficient and compact automobileengines with reduced environmental impact. From the viewpoint of the tribologist this means increasing specific loads, speeds andtemperatures for the major frictional components of the engine, namely, the piston assembly, the valve train and the journal bearings, andlower viscosity engine oils with which to lubricate them. Inevitably, this leads to decreasing oil film thicknesses between the interactingsurfaces of these components and a more crucial role for the topography and surface profile of the two surfaces in determiningtribological performance. This paper reviews the nature of the surfaces encountered in the piston assembly, valve train and journalbearings of the internal combustion engine and how mathematical models of engine tribology are endeavouring to cope with the extremecomplexities the incorporation of surface topography potentially brings. Key areas for future research and the implications for design arehighlighted. q 2000 Elsevier Science S.A. All rights reserved.

Keywords: Automobile; Engine; Tribology; Piston; Piston ring; Cylinder; Engine bearings; Cam; Follower; Valve train; Friction; Lubrication; Wear

1. Introduction

An appreciation of the tribology of the piston assembly,valve train and bearings in an automobile engine mustentail an understanding of the concept of modes of lubrica-tion. This is particularly true if the objective is to facilitateimprovements in aspects of design and performance, sincethis can only be achieved reliably if the underpinningengineering science is satisfactorily identified and compre-hended.It is nearly 100 years since the renowned Richard

Stribeck undertook his experiments on plain journal bear-ing friction, the results of which were subsequently re-ordered by Ludwig Gumbel. Details may be found in

w xDowson 1 . The data has been widely represented on whathas become known as the Stribeck diagram, in the form ofa plot of two non-dimensional groupings: the coefficient of

! .friction m on the ordinate and a variation of the Som-! .merfeld grouping hNrP as abscissa; where h is the

dynamic viscosity, N is rotational speed and P is specificload. With the development of the understanding of regimesof lubrication, this plot has increasingly incorporated the

! .film thickness ratio, or parameter, l on the abscissa asshown in Fig. 1, the modified Stribeck diagram.

) Corresponding author. Fax: q44-113-242-4611.! .E-mail address: [email protected] M. Priest .

The film thickness ratio has proved to be a valuabledesign concept, since it has led to an appreciation of theoccurrence of surface interaction in a range of lubricatedmachine elements, and a recognition that surface topogra-phy can have a highly significant role in the performanceand durability of such components. This is certainly true inregard to the major frictional components of the internalcombustion engine which will be addressed in this paper.

!l is defined as the ratio of the film thickness calculatedthrough the application of classical thin film analysis

.taking the surfaces to be smooth to the composite surfaceroughness. Values of the film thickness ratio appropriate towhat are now called regimes of lubrication have frequently

! .been quoted see Fig. 1 ; however, the position is compli-cated by many factors including the recognition that rough-ness measured in the laboratory may be modified duringoperation andror flattened during a particular load bearingevent.The fact that the data of Stribeck led to a single curve,

as shown in Fig. 1, was a justification of the principles ofdynamic similarity, enabling extrapolation of the predic-tion of performance for a given machine element based onnon-dimensional groupings. Whilst this observation is notstrictly true for more complex lubrication situations wherephysical and chemical actions are important, the shape ofthe curve with its characteristic minimum lent itself toidentification with the regimes of lubrication which have

0043-1648r00r$ - see front matter q 2000 Elsevier Science S.A. All rights reserved.! .PII: S0043-1648 00 00375-6

( )M. Priest, C.M. TaylorrWear 241 2000 193–203194

Fig. 1. The modified Stribeck diagram.

unfolded during the 20th century. These are indicated onFig. 1 and are briefly explained in Table 1.The regimes of lubrication conventionally associated

with the piston rings, camrfollower and engine bearings ofan automobile are shown in Fig. 1. These components relyupon different modes of lubrication for satisfactory perfor-mance and indeed each may enjoy more than one form oflubrication during a cycle. This reflects the challenges thatface the designer in improving operational characteristics,in response to legal and other pressures on emissionscontrol and energy efficiency.Such improvements have to be effected against a back-

ground of engines with higher specific outputs and oftensmaller components with higher surface speeds and tem-peratures. The important issues facing those who developlubricants will be apparent and the trend to lower viscosity

!lubricants e.g. OWr30 SAE grade lubricants, which are.now on the market in some parts of the world is strong.

Whilst this helps in the fight to reduce friction losses, it

Fig. 2. Fuel energy distribution for a medium size passenger car during anw xurban cycle 3 .

also leads to reduced film thicknesses and potential dura-bility problems. The ability to incorporate more and moreaspects of the physical behaviour of lubricants into analyti-cal modelling is an important and a fast developing fieldw x2 . Uppermost in this regard are the reduction in viscosityat high shear rate, particularly with polymer containingmultigrade lubricants, the rise in viscosity at elevatedpressure and the boundary friction and wear behaviour inthe mixed and boundary lubrication regimes.In the last two decades, there have been many studies

leading to an elucidation of friction in engine components.It is revealing to examine where the energy of the fuel that

w xis burnt actually goes. In Fig. 2, data after Andersson 3 isrecorded representing the distribution of fuel energy for amedium size passenger car during an urban cycle.Only 12% of the available energy in the fuel finds its

way to the driving wheels, with some 15% being dissi-pated as mechanical, mainly frictional, losses. The world-wide economic implications of this are startling and the

Table 1Summary of lubrication regimesRegime of Lubrication Characteristics

Hydrodynamic Full fluid film lubrication in which the surfaces are completely separated. The dynamic viscosity of the lubricant is itsmost important property

Elastohydrodynamic Nominally also full fluid film lubrication with surface separation, but a more concentrated mechanism where elastic deformationof the surfaces and the effect of pressure on viscosity are important

Mixed There is surface asperity interaction to some degree and the characteristics of both Elastohydrodynamic and BoundaryLubrication are influential

! .Boundary The surfaces are in normal contact with behaviour characterised by the chemical and physical actions of thin films ofmolecular proportions

( )M. Priest, C.M. TaylorrWear 241 2000 193–203 195

Table 2Typical tribological and performance parameters for a gasoline engineParameter Engine Piston ringrliner Camrfollower

! . ! .bearing top compression ring nose

Minimum lubricant film thickness -1 mm -0.2 mm 0.1 mmMaximum temperature 1808C 2008C groove, 1208C liner 1508CMaximum pressurerspecific loading 60 MPa 70 MPa 600 MPa

8 y1 7 y1 7 y1Maximum shear rate 10 s 10 s 10 s! .Power loss typical 0.25 kW 0.15 kW 0.04 kW

Minimum dynamic viscosity 0.0025 Pa s 0.0065 Pa s EHLComposite surface roughness 0.35 mm Ra 0.2 mm Ra 0.3 mm Ra

prospect of significant improvement in efficiency by mod-w xest reductions in friction are apparent 4 . Based on the

data in Fig. 2, a 10% reduction in mechanical losses wouldlead to a 1.5% reduction in fuel consumption.A final general point which is worth making relates to

the ever decreasing film thicknesses, which are beingpredicted to occur in engine frictional components. Dow-

w xson 5 has observed that during the 20th century thethickness of lubricating films in machine elements such asbearings, gears, etc., has reduced by several orders ofmagnitude. It has been noted that the automobile engine,as if by nature, seems to operate with oil films in all itsfrictional components of the order of one micrometre inthickness. In fact, films significantly thinner than this areencountered. This serves to emphasise the increasing im-portance of the surface topography of components bearingin mind the Stribeck diagram and the regimes of lubrica-tion already discussed.Before addressing the particular components, which are

the specific interest of this paper, it is helpful to obtain anorder of magnitude feel of important tribological andperformance parameters in each. In the spirit of engineer-ing appreciation, this is made in Table 2. The data relatesto a modern four-cylinder, four-stroke, gasoline enginewith four valves per cylinder and double overheadcamshafts with direct acting followers.

2. Piston assemblies

In terms of overall design concept, pistons and pistonrings have changed very little since the pioneering work of

w x w xRamsbottom 6,7 and Miller 8 in relation to steamengines. Ramsbottom deserves the major credit for hisinnovative design of a single piece, metallic piston ringwith a free diameter ten percent greater than the cylinderbore within which it was to operate. When fitted into asimple circumferential groove in a piston, the ring wasforced against the cylinder bore by its own elasticity toprovide a steam seal. This was a great improvement onprevious piston and ring designs that were composed ofseveral pieces and incorporated leaf or spiral springs to

w xachieve an adequate sealing force. Miller 8 proposed an

ingenious modification to the Ramsbottom ring wherebycylinder pressure was allowed to act on the back face ofthe ring thus providing additional sealing force. This de-sign change produced a self-regulating system as the seal-ing force rises and falls with the cylinder pressure and alsoallowed the use of very light, flexible rings, which wereable to conform to the cylinder.The piston ring is perhaps the most complicated tribo-

logical component in the internal combustion engine. It issubjected to large, rapid variations of load, speed, tempera-ture and lubricant availability. In one single stroke of thepiston, the piston ring may experience boundary, mixed

w xand full fluid film lubrication 9 as illustrated in Fig. 1.Elastohydrodynamic lubrication of piston rings is alsopossible in both gasoline and diesel engines on the highly

w xloaded expansion stroke after firing 10 .The historical development of piston ring analysis em-

phasises the theme of this paper most succinctly.w xIn 1959, Furuhama 11 developed a dynamic hydrody-

namic analysis of piston ring lubrication for a piston ringprofile consisting of a flat central land bounded by twohalf parabolas, which incorporated the effect of the cyclicvariation of both load and sliding speed. This pioneeringeffort correctly identified the importance of squeeze filmaction in maintaining hydrodynamic load capacity but thelikelihood of surface contact was not considered.A key research effort in the experimental field was that

w xof Hamilton and Moore 12 in the 1970s who developedminiature capacitance film thickness transducers mountedflush in the cylinder wall to measure piston ring filmthickness. They complemented their experiments on a

w xmotored engine with a theoretical analysis 13 , whichyielded predicted film thickness values up to eight times

w xgreater than those measured. Brown and Hamilton 14later accounted for this discrepancy by considering theeffect of lubricant starvation on predicted film thickness.Further theoretical analyses subsequently emerged with

increasing degrees of sophistication and fewer limiting! w x.assumptions e.g. Refs. 15,16 .

One major criticism of these analyses is that theyassume the rings operate in either a full fluid film lubrica-tion regime or in an extremely simplified boundary lubri-cation regime. No consideration is given to the transitional


w xFig. 3. Initial top ring profile and topography 21 .

mixed lubrication regime, where surface roughness caninfluence hydrodynamic performance or to the nature ofthe contact occurring between the surfaces in the mixed

w xand boundary regimes. Rohde 17 remedied this situationby developing an innovative piston ring lubrication modelthat incorporated detailed mixed lubrication and surfaceasperity contact models based on the work of Patir and

w x w xCheng 18,19 and Greenwood and Tripp 20 respectively.

2.1. Profile and topography

Piston rings are generally manufactured from cast ironor steel and are often surface treated or coated on theperiphery, and occasionally the flanks, to increase wearresistance. The initial form and topography the piston ringoffers to the cylinder wall is a combination of designfeatures and artifacts of the coating process. Fig. 3 showsthe initial profile and topography of a top compression ring

w xfrom a modern gasoline engine after Priest 21 .The ring is manufactured from spheroidal graphite cast

iron with a flame sprayed molybdenum coating on theperiphery. The barrel-faced, convex form of the profile is adesign feature but the deep valleys in the profile are poresformed in the molybdenum coating during manufacture.Fig. 4 shows the same ring after 120 h running at a

w xconstant speed and load 21 .

w xFig. 4. Top ring profile and topography after 120 h running 21 .

w xFig. 5. Wear of the second compression ring 21 .

It can be seen that the profile has worn significantlywith the initial curvature much reduced. However, thedeep valleys of the porosity persist in the topography,which is an aspect hitherto neglected in analysis of pistonring lubrication.The second ring of this particular engine is a plain cast

iron Napier scraper ring with compression and oil-controlfunctions. This wears much more rapidly in service asshown in Fig. 5, which shows the change in profile andtopography over the first two hours of running.Fig. 5 highlights the complex geometry of the piston

ring as new, with the fine turning marks deliberately left inplace, and the dramatic wear that can occur during run-ning-in with more than 10 mm lost from the peak of theprofile. Fig. 6 gives the brake specific fuel consumption! .bsfc of the engine for the early stages of running andshows a dramatic fall during the first hour. Although, thereare other mechanisms involved, such as changes in the

w xlubricant, it is argued 21 that this dramatic reduction infuel consumption and, hence, friction, is mainly at-tributable to the wear of the second compression ring asshown in Fig. 5.Wear of the other rings in the ring pack is also crucially

important to the performance of the engine as is that of thecylinder wall, although this tends to occur more slowly.Fig. 7 shows the surface topography of the cylinder

wall, manufactured from grey cast iron, as new and after120 h running in the mid-stroke region. No attempt hasbeen made to evaluate the amount of wear in this data; theorigin of the radial coordinate data is simply the mean line.However, there are clearly significant changes evident insurface topography, which have a major effect on thelubricant supply to the ring pack and the nature of surfacecontact between the components. Note that data for the

w xFig. 6. Variation of fuel consumption in the early stages of running 21 .


w xFig. 7. Cylinder wall topography variation 21 .

mid-stroke region has been presented where least wear ofthe liner is encountered and where many traditional analy-ses of piston ring lubrication predict full fluid films andthus no wear. Full details of the experiments undertaken on

w xthis gasoline engine can be found in Ref. 21 .Similar comments could be made about the profile and

topography of the components in a diesel engine and theirchange with running time in an engine. Comparable data

w xfor a diesel engine can be found in Priest et al. 22 .One aspect of the experiments undertaken on the diesel

engine distinct from the gasoline was longer-term runningof the engine. Particularly interesting in this respect wasthe wear of the cylinder wall at top dead centre. Fig. 8

Fig. 8. Diesel engine cylinder wall at top dead centre after long-termw xrunning 21 .

shows the profile of the cylinder at this position after morethan 628 h running marked with the approximate locationsof the ring reversal points, the positions where they cometo rest at the end of the upstroke. Ring 1 is the topcompression ring and ring 4 is the oil-control ring, whichhas two distinct load bearing regions, or lands. Deep wearscars have developed in the cylinder wall surface, manu-factured from induction hardened cast iron in this instance,at the ring reversal positions reflecting the low film thick-nesses and high loads at these points. This is especiallytrue for the top compression ring.

2.2. Wear prediction

The observed wear of the piston rings and cylinderwalls in the above examples has a significant effect on theperformance of the piston assembly. Yet traditionally, nowear modelling has been included in piston ring tribologi-cal analyses. This is because incorporating a considerationof wear in the analysis adds a further layer of complexityto an already sophisticated model. It is further com-pounded by the fact that wear is the least understood of thethree main processes in tribology: friction, lubrication andwear.A piston ring tribology model incorporating prediction

of the change in ring face profile with wear in the enginew xhas recently been reported by Priest et al. 22,23 . It

assumes that the wear of the ring profile may be describedby the Archard wear equation, in the form proposed by

w xLancaster 24VskWxswhere,

! 3.V s worn volume m3 y1 y1! .k s wear factor m m N

! .W s load N! .x s sliding distance ms

.The wear factor, k, is a function of the interacting

materials, their surface topography, the lubricant and theoperating conditions. This can alternatively be expressed

w xFig. 9. Variation of wear factor with film thickness ratio 23 .


w xFig. 10. Predicted and measured ring profiles 22 .

! .as a variation of wear factor with film thickness ratio lrelative to the wear factor in the boundary lubricationregime, k , as shown in Fig. 9.0The wear factor in the boundary lubrication regime, k ,0

is determined from bench test rig experiments using actualcomponents and lubricant at operating conditions of load,speed and temperature indicative of boundary lubrication.This empirical input to the model clearly exposes our lackof fundamental understanding of the wear processes takingplace in such tribological interfaces. This approach, how-ever, has been applied successfully in automotive valve

w xtrain wear modelling 25,26 .With this relationship and the cyclic variation of mini-

mum film thickness predicted by the lubrication analysis,the wear factor can be determined at each crank angle inthe engine cycle. Thus, it is possible to predict, interac-tively, the changes in wear and lubrication of the pistonring that take place with running time in the engine.

w xAn example, taken from Priest et al. 22 , of the applica-tion of this model is given in Fig. 10. The measured newand worn ring profiles are shown overlaid using twodifferent methods. Firstly by geometry, visually matchingthe unworn outlying regions of the profile, and secondlyby mass, converting the weight loss of the ring to anevenly distributed volume loss around the ring circumfer-ence. The predicted ring profile, after 120 h simulatedrunning, is also shown and this correlates well with themeasured data.

3. Valve train

The improvement in engine breathing was one of theways in which increased power for the internal combustionengine was gradually realised 100 years ago. It continuesto be an important area for study with the poppet valvetrain, the first effective means of introducing air and fuelinto the combustion chamber and for exhausting the burntgases, still dominant in applications. Attempts to replacethe poppet valve train by other devices, notably rotary and

sleeve valves have consistently failed though efforts stillsporadically continue.Although the poppet valve train has established itself as

the favoured method of introducing the combustible chargeand exhausting the used gases, primarily because of tribo-logical problems with the alternatives, it too has experi-enced severe difficulties. The introduction of the overheadcamshaft exacerbated design difficulties since the lubrica-tion, or tribological performance, of such an arrangementproved to be inherently poor. In the last 20 years, mostautomobile manufacturers have experienced operatingproblems with cam and follower lubrication and the engi-neering science background to this has been widely stud-

w xied. Details may be found in Taylor 27,28 .Analytical developments to enable the prediction of the

cyclic variation of important parameters such as minimumfilm thickness, maximum Hertzian stress and power loss,based upon elastohydrodynamic lubrication theory, havetaken place to enable designers to effect the most advanta-geous mechanical schemes to promote cam and follower

! w x.durability e.g. Refs. 29,30 . However, although it hasbeen convincingly demonstrated that lubrication of a ‘hy-drodynamic’ nature does have a role to play, the moderncam and follower has traditionally been associated with theboundary lubrication regime where the role of chemicalactions in thin surface films is vital. This is linked to theadditive package of the lubricant and in particular toextreme pressure additives, of which forms of zinc di-

! .alkyldithophosphate ZDDP are the most common. Thisserves to emphasise that, at least for parts of the cam andfollower cycle, surface interaction takes place. Clearly,surface topography may have a very significant role and itis this aspect of valve train studies which will be high-lighted in the remainder of this section.The most common materials for cams and followers are

irons and steels with a variety of metallurgies according topreparation. There is a range of possible surface treatmentsto assist running-in and prevent early failure. Ceramicfollowers are, however, becoming more common. Thefailure modes are pitting, polishing and scuffing all ofwhich are influenced by materials, lubrication, design andoperating conditions. The durability and type of failure canvary considerably depending upon the combination ofmaterials chosen to work together, their surface treatmentand the lubricant and its additive package.Only very small dimensional tolerances are permitted

on finish ground cams. The development of the lift profileto promote satisfactory engine breathing over a range ofoperating conditions is a complex process in its own right,and one that does not seem to have been linked directly tothe tribological operating environment. The surface finishof automotive metallic cams and followers would be typi-cally 0.2 mm Ra, with the implicit view that smoother

!finishes are better to promote elastohydrodynamic lubrica-.tion . However, the picture in reality is not that simple

with evidence that both roughening of smooth surfaces can


Table 3Cam surface roughness before and after 100-h tests in a laboratory

w xsimulator 4! .Cam roughness Average Ra across cam at specific points

! . ! .mm nominal Nose Flanks Base circle

before after before after before aftera0.1 0.20 0.50 0.14 0.15 0.14 0.14

0.2 0.28 0.17 0.27 0.28 0.28 0.270.4 0.31 0.22 0.42 0.41 0.48 0.48

a0.8 0.52 0.55 0.78 0.77 1.27 1.271.6 0.96 0.34 1.47 1.45 2.41 2.40

aDamage on the nose caused by run-out on cam.

occur and that lubricant retention in surface troughs can beimportant. A few aspects of surface topography considera-tions in relation to cam and follower performance andanalysis will be briefly touched upon.

3.1. Initial surface roughness

Few studies seem to have been directed to the influenceof initial surface roughness on the subsequent performance

w xof automotive cams and followers. Roylance et al. 31undertook a study to examine wear behaviour during therunning-in period as influenced by surface topography andhardness. Two test apparatus were adopted, a motoredcylinder head and a laboratory cam and tappet machine.With the motored head it was noted that the cam androcker enjoyed a smoothing of their surfaces during their‘flushing’ runs but that both roughened up during the 40-hrun-in period. It was suggested that the smoothing processmight not have encouraged lubricant retention in the con-tact, hence, promoting subsequent damage. For the labora-tory apparatus, less wear was reported for smoother sur-faces and for cases where the hardness of the cam was lessthan that of the follower.

w xTaylor 4 has reported studies using a laboratory appa-ratus to test a direct acting mechanism. The cams weresteel; induction hardened to 2.5 mm depth, stress relieved,ground and phosphated. Cams with different surfaceroughness initially were run against nominally similarfollowers for 100 h at a fixed camshaft speed and tempera-ture. The data is recorded in Table 3.It can be seen that on the cam flanks, there was little

change in roughness reflecting healthy oil film thickness atthese sites. The cam and follower were mechanically sepa-rated over the base circle and, hence, there was no changein this region. Over the nose region, there was evidence onthe rougher cams of an improvement in surface finishreflecting removal of asperity tips and surface flattening. Itwill be noted that the smoother cam roughened up, andthere was a tendency for the final roughness at the nose tomove towards the same value irrespective of initial rough-ness. It is worthy of note that in another series of tests thetemperature of the bulk apparatus was varied up to 1208C,but this did not appear to be an influential factor.

3.2. Surface roughnessrwaÕiness-analysis

w xAs has been noted, Patir and Cheng 18,19 developedin the late 1970s a stochastic analysis approach to studythe effects of surface roughness height and distributionbetween components upon the lubrication of machine ele-ments. By incorporation of an asperity interaction model

w xdeveloped by Greenwood and Tripp 20 , it proved possi-ble to investigate mixed lubrication situations. For exam-

w xple, Ruddy et al. 32 have applied the approach to studyoil-control piston ring effects on oil consumption. Theapproach has also been used to assess cam and follower

! w x.performance e.g. Refs. 33,34 . For the direct acting camw xand follower situation shown in Fig. 11, Dowson et al. 33

studied the predicted effects of surface roughness uponnominal film thickness, power loss and load carried by theasperities.Whilst some interesting predictions emerge from such

analysis, for example, the small proportion of the loadcarried by the asperities because of the non-conformalnature of the camrfollower contact, it must be observedthat currently the value of such analyses is limited inrelation to the prospect of influencing design. One reasonfor this is the use of data extrapolated beyond the filmthickness ratio limit for which it was established.Potentially of more value in the longer term, is the

!application of deterministic studies of waviness rough-

w xFig. 11. Roughness considerations in a cam and follower 33 .


w xFig. 12. Effects of waviness orientation in EHL contacts 35 .

.ness in elastohydrodynamic lubrication. At the presenttime with the exponential advance in computing power,and the development of increasingly more powerful nu-merical analysis techniques, there is rapid development ofnon-smooth surface, elastohydrodynamic lubrication analy-sis. This has enormous significance in the design anddevelopment of concentrated contacts to operate under

w xincreasingly severe conditions. Ehret et al. 35 have stud-ied the effects of various surface textures, described byeither an orientated waviness or a uniform distribution ofasperities, for rolling and sliding conditions. Fig. 12 showsthe predicted variation of the minimum film thickness tomaximum Hertzian deformation ratio with waviness orien-tation angle, zero degrees being transverse waviness, for

w xpure rolling and pure sliding, after Ehret et al. 35 . It isclear that whilst the orientation of waviness is only moder-ately influential upon film thickness with pure rolling, thisis not the case for pure sliding. Surface texture can lead toimportant differences and in pure sliding it is shown thatthe best lubrication condition is produced for transversewaviness. The definitive linking of surface texture to im-proved machine element performance in, say, a cam andfollower, would be a powerful technique.

3.3. Wear prediction

The prediction of wear in machine elements is a notori-ously difficult task. In Section 2 of this paper, a techniquewas described with regard to the piston ring and cylinderliner contact. The same technique was first applied to the

!analysis of pivoted follower valve train systems e.g. Refs.w x.25,26 . It is encouraging to report that such modellinghas proved of value in identifying critical conditions forwear, including the positions of maximum wear. Thereasons for the observed severity of wear can be identifiedand this has proved of value in developing laboratorytesting apparatus with appropriate kinematic conditionsw x36 . Of course, the influence of the dimensional wearcoefficient in predicting wear magnitudes is crucial as withthe piston ringrcylinder liner contact.It is also worthy of note that empirical techniques for

predicting cam follower wear as influenced by material

w xproperties have been investigated. Tanimoto et al. 37used statistical methods of multi-regression analysis, basedupon theoretical and experimental studies, to establish thecorrelation between material properties and wear resis-tance.

3.4. WaÕiness

w xO’Connor and Spedding 38 undertook an experimentalstudy of the effect of surface waviness on a cam profile.Such waviness was influenced by the form imposed ontothe grinding wheel by the roller dresser. Wear and oilretention characteristics of the automobile camshaft sur-face were investigated. It was suggested that the wavy camsurface could offer good oil retention characteristics butthat with time the release of hard wear particles due to thewave peaks could be problematical. It was proposed that acam surface having the same roughness, but zero or re-duced waviness, would show improved performance.Waviness on automotive cams remains an issue and it isinteresting to note that analysts are now in a position to

!contribute to the debate on the overall influence e.g. Ref.w x.37 .

4. Engine bearings

The Mobility technique for the analysis of dynamicallyw xloaded engine bearings 39,40 was established some 30

years ago and remains the most common approach. Themethod assumes that the hydrodynamically generated loadcapacity due to pressures in the lubricant film balances theapplied load, thus precluding the possibility of instabilityeffects as inertia loading is neglected. The technique hasproved amenable to simple computerised analysis yieldingthe cyclic minimum film thickness, an important designparameter, amongst other predictions.It is, however, important to recall that there is a wide

range of assumptions implicit in the use of the Mobilitymethod as it is normally applied. These include,

! .i Short bearing analysis — highly inaccurate athigh eccentricity ratios.

! .ii Circumferential symmetry — grooverhole ef-fects neglected.

! .iii Rigid surfaces — manifestly not so in manycircumstances.

! .iv Perfect alignment — little known about misalign-ment issues.

! .v Newtonian lubricant behaviour — very high shearrates encountered.

! .vi Non vibrational instability — shaft inertia omit-ted

! .vii Isothermal lubricant film — heat transfer effectscan be important.


These restrictive assumptions clearly imply that anypredictions can only be benchmark and point to the impor-tance of field experience being used alongside predictionsin order to make sound design prognostications for im-

w xprovements. Taylor 4 has detailed a range of researchstudies in which some of the assumptions detailed abovehave been relaxed.

w xThe ‘thinning film’ alluded to by Dowson 5 is evidentin relation to engine bearings. In the early 1970s, thebenchmark prediction for a satisfactory minimum lubricantfilm thickness was about 2.5 mm. Twenty-five years later,predictions of minimum film thickness a factor of 5 ormore less than this are being made for engine bearings inpassenger cars. A minimum film thickness in the range0.5–1.0 mm implies that asperity interaction may occurbetween the journal and bearing for at least part of theengine cycle. This is one of three surface topographyaspects that will be addressed in relation to automotivebearings.

4.1. Asperity interaction

The likelihood that an engine bearing would have toperform part of its operational envelope with some contactbetween the journal and bearing for a part of each cycle

w xwas addressed by Conway-Jones and Gojon 41 . Thissituation is being influenced by the introduction of smallerengines and, hence, smaller components such as bearings,whilst the specific power output is increasing. Thus, veryhigh maximum specific loads are being encountered inengine bearings.The effect of such asperity interaction would be to

increase friction and, hence, give a larger average torqueover a cycle. Evidence of this was presented by the authorsand is shown in Fig. 13. With an increase in bearingtemperature, the predicted torque and minimum film thick-ness fall according to hydrodynamic theory. However, at apredicted oil film thickness of 1.2 mm an increase inmeasured torque was encountered, this being attributed to

w xasperity interaction. Conway-Jones and Gojon 41 dis-

w xFig. 13. Surface asperity influence in engine bearings 41 .

w xFig. 14. Lobe size and frequency effects in engine bearings 42 .

cussed the effects of surface topography upon the addi-! .tional friction and heat generation noting that for their

experimental data the run-in roughnesses of the compo-nents were 0.35 and 0.26 mm Ra. They developed anempirical method of calculating the asperity loading, hence,enabling heat generation to be determined on a mPVbasis, V being the surface velocity.

4.2. Journal waÕiness

w xMehenny et al. 42 have considered the effect ofcircumferential waviness on engine bearing performance.There is a clear evidence that the machining process for,say, big-end bearings may have a significant effect onbearing failure due to the creation of a lobed crankpin.Experimental evidence of such lobing with any frequencybetween 1 and 21 on a journal has been obtained and lobesmay be of the order of 5 mm in amplitude on a 50-mmnominal diameter shaft. Clearly, this is large comparedwith predicted minimum film thickness. Mehenny et al.w x42 developed an analytical approach to assess the effectsof lobe size and frequency and presented data to indicatethe influence upon minimum film thickness and maximum

! .pressure Fig. 14 .Comparison with some early data of DeHart and Smiley

w x43 , who had measured bearing weight loss due to wearwith imperfect journal geometries, was encouraging.

4.3. Bearings with microgrooÕes

w xKumada et al. 44 reported the continuing developmentof bearings with circumferential microgrooves, with theintention of enabling engine bearings to operate underincreasingly severe conditions. Their evaluation revealedthat the initial wear process and deformation at the mi-


crogroove peaks led to good conformability, lower temper-atures because of higher oil flow and resistance to seizurebecause of the oil retention properties. The improvedconformability was assessed through friction measure-ments, which with stepwise decreases of bearing filmthickness, stabilised in a much shorter time than with plainbearings. Temperatures 10%–20% lower were obtainedwith the microgrooved bearings used as main bearingswhilst the claim to improve seizure characteristics wasjudged by the time taken to failure when the supply oflubricating oil was cut off. Further references to associated

w xresearch in this field can be found in Ref. 44 and it isinteresting to note that the use of grooving in cylinder

w xliners is also adopted 45 , though the orientation of thegrooves is normal to the piston primary motion.Since the attention paid to engine bearing design in the

1960s and 1970s, it is interesting to observe that researchin the tribology of the piston assembly and valve train hasbeen more prevalent. The increasingly severe operatingconditions, however, are leading to an increasing propen-sity to bearing failures and attention to all aspects ofdesign is appropriate. The role of surface topographyconsiderations alongside improved consideration of elasticand thermal effects is needed.

5. Future challenges

There remain extensive challenges for those interestedin the tribological design of automotive piston assemblies,valve trains and engine bearings. The following are impor-tant aspects meriting detailed research and development inpursuit of improved performance and durability.

! .i Improved surface profile, surface roughness andmixed lubrication considerations.

! .ii Development of a linkage between lubricationmechanics and chemical mechanics, with a betterunderstanding of the role of additives in reactionfilms.

! .iii Consolidation of the developments in the under-standing of lubricant rheology to make more ef-fective design prognostications.

! .iv Wear modelling linking to failure, materials, lu-brication and thermal considerations.

! .v Satisfactory provision of lubricant, especially inthe cam and follower interface.

! .vi Improved materials, surface coatings and surfaceengineering.

6. Conclusions

It is about a quarter of a century since the energy crisisof the 1970s. This led to an enhanced awareness of the

need to use scarce natural resources more efficiently andprecipitated an intense study of the efficiency of theinternal combustion engine: the piston assembly, valvetrain and engine bearings. Such studies have remainedvibrant and have been further driven by the increasingrecognition of the fragility of our environment and theneed to accommodate growth in the automobile sector in asustainable manner. This paper has reviewed the currentposition regarding the tribological design and friction asso-ciated with the tribological components of the engine witha specific focus upon surface topography and surfaceinteraction considerations. Much remains to be achieved inthis important field and significant areas for future atten-tion have been identified.

References

w x1 D. Dowson, History of Tribology, Professional Engineering Publish-ing, 1998, 768 pp.

w x2 R.C. Coy, Practical applications of lubrication models in engines,! .New Dir. Tribol. MEP 1997 197–209.

w x3 B.S. Andersson, Company perspectives in vehicle tribology —Volvo, 17th Leeds–Lyon Symposium on Tribology — VehicleTribology, Tribol. Ser. vol. 18 Elsevier, Amsterdam, 1991, pp.503–506.

w x4 C.M. Taylor, Automobile engine tribology — design considerations! . ! .for efficiency and durability, Wear 221 1 1998 1–8.

w x5 D. Dowson, Developments in lubrication — the thinning film, J.! .Phys. D., Appl. Phys. 25 1992 A334–A339.

w x6 J. Ramsbottom, On an improved piston for steam engines, Proc.! .Instn. Mech. Eng. 1854 70–74.

w x7 J. Ramsbottom, On the construction of packing rings for pistons,! .Proc. Inst. Mech. Eng. 1855 206–208.

w x8 G.M. Miller, On a packing for pistons of steam engines and pumps,! .Proc. Inst. Mech. Eng. 1862 .

w x9 D. Dowson, Piston assemblies; background and lubrication analysis,! .in: Taylor Ed. , Engine Tribology, Tribol. Ser. vol. 26 Elsevier,

Amsterdam, 1993, pp. 213–240, Chap. 9.w x10 J.E. Rycroft, R.I. Taylor, L.E. Scales, Elastohydrodynamic effects in

piston ring lubrication in modern gasoline and diesel engines, in:! .Dowson Ed. , Elastohydrodynamics — Fundamentals and Applica-

tions in Lubrication and Traction, Proc. 23rd Leeds–Lyon Sympo-sium on Tribology, September 1996, Elsevier, Amsterdam, 1997, pp.49–54.

w x !11 S. Furuhama et al., A dynamic theory of piston ring lubrication 1st. ! . ! .report, calculation , Bull. JSME 2 7 1959 423–428.

w x12 G.M. Hamilton, S.L. Moore, The lubrication of piston rings, firstpaper, measurement of the oil-film thickness between the pistonrings and liner of a small diesel engine, Proc. Inst. Mech. Eng., 188! . ! .20r74 1974 253–261, D51–D60.

w x13 G.M. Hamilton, S.L. Moore, The lubrication of piston rings, secondpaper, comparison between measured and calculated thicknesses ofthe oil-film lubricating piston rings, Proc. Inst. Mech. Eng., 188! . ! .20r74 1974 262–268, D51–D60.

w x14 S.R. Brown, G.M. Hamilton, The partially lubricated piston ring, J.! . ! .Mech. Eng. Sci., Inst. Mech. Eng. 19 2 1977 81–89.

w x15 Y. Wakuri, M. Soejima, T. Taniguchi, On the oil film behaviour of! .piston rings correction of effective pressure region of oil film ,! . ! .Bull. JSME 21 152 1978 295–302.

w x16 D. Dowson, P.N. Economou, B.L. Ruddy, P.J. Strachan, A.J.S.Baker, Piston ring lubrication — Part II, theoretical analysis of asingle ring and a complete ring pack, in: Rohde, Wilcock, Cheng


! .Eds. , Energy Conservation Through Fluid Film Lubrication Tech-nology: Frontiers in Research and Design, ASME, 1979, pp. 23–52.

w x17 S.M. Rohde, A mixed friction model for dynamically loaded con-tacts with application to piston ring lubrication, Surface RoughnessEffects in Hydrodynamic and Mixed Lubrication, Rohde, Cheng! .Eds. , ASME, 1980, pp. 19–50.

w x18 N. Patir, H.S. Cheng, An average flow model for determining effectsof three-dimensional roughness in partial hydrodynamic lubrication,

! . ! .J. Lubr. Technol., Trans. ASME 100 1 1978 12–17.w x19 N. Patir, H.S. Cheng, Application of average flow model to lubrica-

tion between rough sliding surfaces, J. Lubr. Technol., Trans. ASME! . ! .101 2 1979 220–230.

w x20 J.A. Greenwood, J.H. Tripp, The contact of two normally flat rough! .surfaces, Proc. Inst. Mech. Eng. 185 1971 625–633.

w x21 M. Priest, The Wear and Lubrication of Piston Rings, Ph.D. Thesis,Univ. Leeds, 1996, Chap. 8, 246 pp.

w x22 M. Priest, D. Dowson, C.M. Taylor, Predictive wear modelling of! .lubricated piston rings in a diesel engine, Wear 231 1999 89–101.

w x23 M. Priest, D. Dowson and C.M. Taylor, Prediction of the lubricationand wear of piston rings — theoretical model, World Tribology

!Congress, Abstracts of Papers, September, 1997, p. 82 full paper in.preparation for J. Eng. Tribol., Proc. Inst. Mech. Eng., Part J, 1999 .

w x24 J.K. Lancaster, Dry bearings: a survey of materials and factors! .affecting their performance, Tribology 1973 219–251, December.

w x25 J.C. Bell, P.J. Davies, W.B. Fu, Prediction of automotive valve trainwear patterns with simple mathematical models, Proc. 12th Leeds–Lyon Symp. on Tribology: Mechanisms and Surface-Distress, 1985,

! .Paper XI iii , pp. 323–333.w x26 J.C. Bell, T. Colgan, Pivoted follower valve train wear: criteria and

! .modelling, Lubr. Eng., J. STLE 1991 114–121, February.w x27 C.M. Taylor, Valve train lubrication analysis, Proc. 17th Leeds–Lyon

Symposium on Tribology — Vehicle Tribology, Tribol. Ser. vol. 18Elsevier, Amsterdam, 1991, pp. 119–131.

w x28 C.M. Taylor, Fluid film lubrication in automobile valve trains, Proc.! .Inst. Mech. Eng., J. Eng. Tribol., Part J 208 1994 221–234.

w x29 A.D. Ball, D. Dowson, C.M. Taylor, Cam and follower design, Proc.15th Leeds–Lyon Symposium on Tribology — Tribological Designof Machine Elements, Tribol. Ser. vol. 14 Elsevier, Amsterdam,1989, pp. 111–130.

w x30 D. Dowson, P. Harrison, C.M. Taylor, The lubrication of automotivecams and followers, Proc. 12th Leeds–Lyon Symposium on Tribol-ogy — Mechanisms and Surface Distress, Butterworth, London,1986, pp. 305–322.

w x31 B.J. Roylance, G. Wang, C.H. Bovington, A. Hubbard, Running-inwear behaviour of valve train systems, Proc. 17th Leeds–LyonSymposium on Tribology — Vehicle Tribology, Tribol. Ser. vol. 18Elsevier, Amsterdam, 1991, pp. 143–147.

w x32 B.L. Ruddy, D. Dowson, P.N. Economou, A theoretical analysis ofthe twin-land type of oil control ring, J. Mech. Eng. Sci., Inst. Mech.

! . ! .Eng. 23 2 1981 51–62.w x33 D. Dowson, C.M. Taylor, G. Zhu, Mixed lubrication of a cam and

flat faced follower, Proc. 13th Leeds–Lyon Symposium on Tribol-ogy — Fluid Film Lubrication, Tribol. Ser. vol. 11 Elsevier, Ams-terdam, 1987, pp. 599–609.

w x34 L.S. Yang, A. Ibo, H. Negishi, A valve train friction and lubricationanalysis model and its application in a camrtappet wear study, SAETechnical Paper 962030, 1996, 9 pp.

w x35 P. Ehret, D. Dowson, C.M. Taylor, Time-dependent solutions withwaviness and asperities in EHL point contacts, Proc. 23rd Leeds–Lyon Symposium on Tribology — Elastohydrodynamics ’96, Tri-bol. Ser. vol. 32 Elsevier, Amsterdam, 1997, pp. 313–324.

w x36 J.C. Bell, Reproducing the kinematic conditions for automotivevalve train wear in a laboratory test machine, Proc. Inst. Mech. Eng.,

! .J. Eng. Tribol., Part J 210 1996 135–144.w x37 I. Tanimoto, M. Kano, M. Sasaki, Establishment of a method for

predicting cam follower wear in the material development process,SAE Technical Paper 902087, 1990, 8 pp.

w x38 R.F. O’Connor, T.A. Spedding, Investigation into the effects ofwaviness present on the cam surface profile of an engine camshaft,

! . ! .Precis. Eng. 11 2 1989 83–88, Butterworth.w x39 J.F. Booker, Dynamically loaded journal bearings: mobility method

! .of solution, Trans. ASME, J. Basic Eng., D 187 1965 537–546,Sept.

w x40 J.F. Booker, Dynamically loaded journal bearings: maximum film! .pressure, Trans. ASME, J. Lubr. Technol. 1969 534–543, July.

w x41 J.M. Conway-Jones, R. Gojon, Heat flow in crankshaft bearings,Proc. 17th Leeds–Lyon Symposium on Tribology — Vehicle Trbol-ogy, Tribol. Ser. vol. 18 Elsevier, Amsterdam, 1991, pp. 33–42.

w x42 D.S. Mehenny, C.M. Taylor, G.J. Jones, H. Xu, The Influence ofcircumferential waviness of the journal on the lubrication of dynami-cally loaded journal bearings, SAE Technial Paper 970216, 1997, 9pp.

w x43 A.O. DeHart, J.O. Smiley, Imperfect journal geometry — its effecton sleeve bearing performance, SAE Paper 813C, 1964.

w x44 Y. Kumada, K. Hashizume, Y. Kimura, Performance of plain bear-! . ! .ings with circumferential microgrooves, STLE Trans. 39 1 1996

81–86.w x45 S. Saburi, C.M. Taylor, D. Dowson, An analytical study of the oil

film thickness between the piston rings and cylinder liner of a longstroke, low speed, marine diesel, Jpn. Inst. Tribol., ITC, Yokohama,Japan, 1995, pp. 1435–1440.

automobile engine tribology — approaching the surface · pdf filewear 241 2000...

Documents