You are on page 1of 13

2002-01-3355

Lubrication, Tribology & Motorsport
Shell Global Solutions (UK), Cheshire Innovation Park, PO Box 1, Chester, CH1 3SH, UK

R.I. Taylor

Copyright © 2002 Shell Research Limited

ABSTRACT
We review some of the key tribological issues of relevance to motorsport applications. Tribology is the science of friction and wear, and in a high performance engine, friction and wear are controlled by good component design (e.g. the engine and the transmission) and also by the use of high performance lubricants with the correct physical (and chemical) properties, matched to the machine they are used in. In other words, design of a specific lubricant for specific hardware can lead to optimised performance. (Tribology is also important in the tire-road contact but is not considered here.) The importance of key physical properties of a lubricant is demonstrated with an emphasis on how the choice of the correct lubricant can help to minimize engine friction (and thus increase available power output) whilst protecting against engine wear. Key lubricant parameters discussed in the paper are the viscosity variation of a lubricant with temperature, shear rate and pressure. The relevance of these parameters in controlling oil film thickness and friction in bearings, the piston assembly, and the valve train is demonstrated with examples showing how a high performance engine differs from more conventional engines.

KEY LUBRICANT PARAMETERS THAT AFFECT PERFORMANCE
The most significant physical property of a lubricant is viscosity. Lubricant viscosity is strongly dependent on temperature, shear rate, and pressure. These effects are often neglected in many simulations of engine components (it is often assumed that the viscosity of the lubricant is constant at the temperature of interest). However, if a realistic assessment of friction and oil film thickness in the contact is required, these effects need to be taken into account. Figure 1 shows the typical way in which lubricant viscosity varies with temperature and shear rate.

INTRODUCTION
The variation of lubricant viscosity with temperature, shear rate and pressure is crucial to how the lubricant performs in an engine. Unfortunately, these effects are typically ignored in models that are used to design engine components. This paper describes the important ways in which lubricant viscosity may vary, and then demonstrates, by using models which take into account these lubricant properties, how the lubricant can have a large impact on both the minimum oil film thickness in key engine components, and on the friction loss of the engine. A conventional engine is compared with a high performance engine, and it is shown that by matching the correct lubricant to the engine, improved performance may be achieved.

Figure 1: Variation of viscosity with shear rate for a SAE10W/50 lubricant Table 1 shows how the viscosity of different SAE viscosity grades differs.

The variation of viscosity with pressure is simply described by the Barus equation: K ( P ) K (0) ˜ exp(DP ) However. which occurs when the limiting shear stress is reached.100 temperature and pressure.3 57. Similarly. T=20qC) -1 D(GPa ). D must be assumed to depend on both temperature and pressure. Jc (s ) is the shear rate at which the viscosity lies precisely halfway between Ko and K . For the above expression to apply to realistic lubricants.8 6.8 10. 150N where K is the viscosity (mPa. 4 cSt Group III.5 10. the Cross equation . typically consisting of around 80% base oil (typically with molecular weights in the range 300-600).5 16. Generally it may be assumed that D decreases with both Simply using the correct value of D (i. We should finally mention that lubricants are not simple fluids. This has quite dramatic consequences for the prediction of friction losses in elastohydrodynamically (EHD) lubricated contacts. T=80qC) -1 D(GPa ). is useful for describing the variation of viscosity with shear rate: 2 Table 2: Values of D (GPa ) at different temperatures and pressures for different base fluids.9 11.s) at infinitely high shear rate. In such circumstances. p=400 MPa.6 11. and N (mPa.4 44. In other words. 5 cSt Group I.9 10.3 Vd (mPa.9 13. the Barus equation will give friction predictions that are too high. then at high enough pressures. the value of D corresponding to the temperature and pressure in the contact) may still lead to errors in the prediction of friction losses in an EHD contact. If the Barus equation is used. according to Infineum log10 J c -1 A  B ˜T where A and B (qC ) are constants for a given lubricant.s).7 PAO 6 Group III. 100N Group I. Ko is the viscosity (mPa. T=80qC) -1 Naphthenic Paraffinic PAO A PAO B Polyglycol 40 26 21 22 24 19 18 14 15 15 -1 35 19 15 16 18 16 14 11 11 11 K § T1 · N ˜ exp¨ ¨T  T ¸ ¸ © 2 ¹ where K is the viscosity (mPa.9 8. there is a maximum friction force (per unit area) of the EHD contact.s) at temperature T (qC).7 13.4 15.900 1.800 Table 1: Typical viscosities of common SAE grades The variation of viscosity with temperature is quite 1 accurately represented by the Vogel equation : D(GPa ).6 15.s) at 20qC 10.7 -1 9. for this simple expression to apply to real lubricants.2 9.9 10.800 3. Base oil D(GPa ). p=400 MPa.9 11.s) at zero shear rate and K is the -1 viscosity (mPa. and estimates based on the limiting shear stress of the lubricant should be used. Alternative expressions for the variation of lubricant viscosity with pressure and temperature include the 6 Roelands’ equation and the Sorab-van Arsdale 7 equations . It is a good approximation to assume that K /Ko is a constant.1 9. T1(qC) and T2(qC) are constants.8 9.8 14. 5 Table 3 reports results from Infineum for base fluids . A simple expression. p=0.SAE grade 20W/50 15W/40 10W/30 5W/30 0W/20 Vk40 (cSt) 144. This is because it is thought that lubricants have a limiting shear stress which cannot 8-10 be exceeded . A good description of the required temperature dependence is: -1 ¡ ¡ ¡ Table 3: Values of D (GPa ) at different temperatures for a range of basestocks.0 11. the viscosity will be high enough that the friction force would exceed this maximum value. according to Larsson D (GPa ) at 60qC -1 Base stock PAO 4 K K    Ko K J 1 Jc D (GPa ) at 80qC -1 D (GPa ) at 100qC -1   10.3 72.3 10. and 20% .8 114.7 13.e. T=20qC) -1 30 91. They are dilute polymer solutions.4 Vk100 (cSt) 17. as originally reported by Larsson . independent of 3 temperature . Table 2 summarises values of D for various different 4 base fluids.200 4.2 9.100 1. p=0.s) at shear rate J (s ). it is found that we have to make Jc dependent on temperature.

Note also that a racing piston will differ substantially from a conventional piston in that the piston skirt area will be much smaller (so that friction is reduced).000 or greater.7 1.8 78.5 89. this means that the upper piston rings are often “starved” of oil (this is particularly true at mid-stroke positions). It is thought that under certain conditions.3 litres.3 litres. It is usually found that the effect of oil starvation is to reduce the oil film thickness on the liner.0 litre engine Inline 4 2.0 litres. since Reynolds’ equation does not allow for viscoelastic behaviour. with a V10 engine configuration.4-45. and to increase the friction loss of the ring-pack. using the appropriate piston ring profile. The most common engine configuration would be an in-line 4 cylinder engine. and the liner temperature at the piston ring position (so that the lubricant viscosity may be estimated). when there are often only two piston rings (the oil control ring and the top ring) a compromise must be made. In a high performance engine. the lubricant must also have good anti-foaming properties.0 0. the parameters for the Formula 1 engine were estimated as follows : Since the displacement per cylinder is 0. the more sensitive the piston will be to secondary motion (piston tilt). In general. a Formula 1 engine has a displacement of 3 litres. then if we know that the bore-stroke ratio is in the range 2-2.5 76-132 37-48 at max rpm European 2.4-98. since the larger the bore-stroke ratio. the precise value of the con-rod length is not essential for carrying out preliminary sensitivity studies into the tribology of the key engine components). In Table 4. There is clearly a limit to how far this can go. Table 4 shows a comparison of some of the key engine variables (based on the above figures) for a typical conventional European engine and a Formula 1 engine. then too much oil will be available to the top . These effects cannot be predicted from Reynolds’ equation. these elastic properties can be beneficial in reducing both wear and friction in engine 11-14 bearings . to these physical requirements on the oil. the maximum linear speed is roughly the same. some further analysis is required to determine when the rings are starved.5. In addition. we know that the maximum angle that the con-rod makes with the vertical is around 15q. PISTON ASSEMBLY TRIBOLOGY The calculation of oil film thickness under the piston rings involves solving Reynolds’ equation.14 154. the bore diameter would have to be increased. and we know that is the area swept out by the cylinder multiplied by the stroke. For most conventional European gasoline engines. By assuming that the maximum angle that the con-rod makes with the vertical for the Formula 1 engine lies in the range 10-15q. Formula 1 engine Engine type Displacement (litres) Displacement cylinder (litres) Stroke (mm) Bore-Stroke ratio Con-rod length (mm) Max piston (m/s) speed per V10 3.0-2. and may also cause cavitation damage in bearings. These additives can impart some elastic properties to the lubricant.0 32 at 7500 rpm Bore diameter (mm) Max rpm 17000-18000 7500 Table 4: Key engine parameters for a Formula 1 engine and a conventional European engine Note also that if a higher maximum rpm is desired. then the oil film thickness and 17-22 friction loss of the piston ring may be calculated . Note that for both types of engine. Additives such as Viscosity Modifiers are very large polymers with molecular weights of the order of 10.additives. and to recalculate the oil film thickness and friction accordingly. and since the displacement per cylinder is constant at 0. Needless to say. with a bore-stroke ratio close to 1. the maximum rpm is around 7500 rpm. we obtained a likely range for the conrod length (however.7 2. It is also necessary to know the gas pressures on either side of the piston ring. the piston ring temperature. For the conventional engine.0-2. Therefore. we can calculate the range within which the bore diameter and the stroke must lie.3 91. a bore-stroke ratio in the range 2. In effect. and a maximum rpm of 1700015 18000 (these latter figures were reported by Wright ). If the oil control ring tension is too low. and taking into account the variable speed of the piston ring as the piston moves from bottom dead center to top dead center. typical mid-range displacements are around 2. again making the racing piston more sensitive to secondary motion COMPARISON OF HIGH PERFORMANCE AND CONVENTIONAL ENGINE PARAMETERS There are clearly many differences between a conventional gasoline engine and a high performance engine. In comparison. often with 4 valves per cylinder. because any air that can be entrained in the oil can cause problems in hydraulically actuated valves. whilst at the same time minimizing total engine friction. a lubricant that is suitable for motorsports applications must have a viscosity that separates all the key moving surfaces in the engine.5.5 39. A further complication then arises since there is interaction between the various rings in the piston assembly. then the stroke would need to be reduced. the oil control ring is designed to restrict the oil supply to the upper rings. If all these parameters are known.0 0.

and Wright reports that the BMEP of a modern Formula 1 engine is between 13 and 14 bar. but there will be excessive oil starvation for the top ring. which would lead to higher friction (and possibly high wear – leading to failure).81 5. if the oil control ring tension is too high. a stroke of 42. We assume a ring tension of 1000 kPa.51 10. Lubricant is assumed to be SAE15W/40 We can carry out a similar analysis for a Formula 1 engine top ring. a bore diameter of 94. For a conventional engine. in mid-strokes).51 4. assuming fully flooded lubrication OFT (Pm) at -270q OFT (Pm) at -180q OFT (Pm) at -90q OFT (Pm) at 0q OFT (Pm) at 90q R=5 cm 10.5 mm in height. oil consumption may be reduced. with a symmetric. bore distortion will also play a part in controlling oil consumption and friction loss.38 5. are flatter) have a larger oil film thickness in those parts of the engine cycle where the “squeeze” effect dominates (e.46 2.6 mm and a con-rod length of 104 mm – these are the mid values of the parameters estimated in Table 4. at 7500 rpm. for different ring profile radii of curvature. Figure 2 shows a typical combustion chamber pressure curve for a conventional European 2.g. there are usually three piston rings. For such an engine.42 8.84 8. it is assumed that the top piston ring is 1.15 5.55 R=10 cm 8. oil consumption may be too high. .76 5.7 mm. compared to a more curved piston ring. We also assumed that the lubricant was an SAE-15W/40 grade.0 litre gasoline engine operating at 2500 rpm and medium load Note that the brake mean effective pressure (BMEP) corresponding to Figure 2 is approximately 4 bars. around dead center positions) and smaller oil film thicknesses in those parts of the engine cycle where the “wedge” effect dominates (e. assuming fully flooded conditions. parabolic ring profile.g. and although friction may be reduced.17 7.17 9. and is a symmetric parabolic ring. Clearly.23 6. On the other hand.22 3.73 8. Under these conditions. 35 30 Pressure (bars) 25 20 15 10 5 0 -360 -240 -120 0 120 240 360 Crank Angle (degrees) 12 10 8 6 4 2 0 -270 -180 -90 0 90 180 270 360 450 Crank Angle (degrees) R = 5 cm R = 10 cm R = 20 cm Figure 3: Oil film thickness versus crank angle for the top piston ring of a conventional European 2. we assume that the top ring has a height of 1mm. Figure 3 shows the predicted oil film thickness versus crank angle.09 10. OFT (Pm) at 180q OFT (Pm) at 270q OFT (Pm) at 360q Friction Loss (W) 109. For this engine. Figure 3 show that piston rings that have a larger radius of curvature (i. Clearly. Figure 4 shows the predicted oil film thickness versus crank angle for different ring radii of curvature.02 Figure 2: Combustion chamber pressure versus crank angle for a conventional European 2.00 R=20 cm 7.3 137. and oil consumption can be controlled more effectively.08 1. for an SAE-15W/40 lubricant.46 7.e.Oil Film Thickness (microns) ring. for a speed of 7500 rpm.85 1.0 litre gasoline engine. (For these simulations we assumed an engine speed of 18000 rpm. with a bottom dead center liner temperature of 100qC and a top dead center liner temperature of 150qC. the piston ring shape also has an impact on the ring friction.3 Table 5: Effect of ring radius of curvature on oil film thickness and friction. In addition we assumed a peak combustion chamber pressure of around 110 bars – the BMEP of the conventional engine was 4 bars. with a ring tension of 200 kPa. with a peak combustion chamber pressure of just over 30 bars.0 171.0 litre gasoline engine. Therefore we have scaled the combustion chamber pressure accordingly).93 4. Table 5 shows the effect that the ring radius of curvature has on oil film thickness and friction loss.07 5.01 5.

the friction mean effective pressure (FMEP) of the top piston ring. for a conventional engine. is considered to be fairly reliable. and correspondingly higher friction losses. operating at 7500 rpm with an SAE-15W/40 lubricant Finally. For the Formula 1 engine. for different lubricants Lubricant SAE20W/50 SAE15W/40 SAE10W/30 Power loss (W) 402 375 313 Minimum oil film thickness (Pm) 0.0 0. If the piston ring tensions used in the Formula 1 engine are far higher than those used in conventional engines. Also in this complete ring-pack simulation. 12. the effect of oil starvation is to increase the predicted ring frictional loss from approx 110 W to 258 W. so that the total piston assembly friction is equal to around 30 kW. is estimated to be around 8. Table 6 summarises the predicted top ring friction power loss. then the total friction for one cylinder is approximately 3 kW.73 0. then the total ring loss is 1. Table 6 shows that reductions in piston ring friction can be achieved by moving to a lower viscosity lubricant. at 18000 rpm.0 6. If the oil control ring has a similar friction loss. we discuss the effect of lubricant viscometry on piston assembly friction.4 kPa.2 kPa. we would expect an FMEP in the range 25-30 kPa. If we then assume that the ring friction is 60% of the total piston assembly friction. with the remainder coming from piston skirt friction.0 Oil Film Thickness (microns) 3.0 10.0 2. the estimated top ring friction is 880 W (assuming a ring radius of curvature of 10 cm).76 0. If we assume similar numbers for the FMEP of a Formula 1 piston assembly. In practice we know that the interaction of the oil control ring and the top ring leads to lower top ring oil film thicknesses at mid-stroke. using the same data as for Figure 4. the squeeze effect has been neglected. For a conventional engine. and the FMEP of the oil control ring is estimated to be approximately 6. was predicted to be 300 W for a ring radius of curvature of 5 cm. assuming fully flooded conditions The friction loss for the Formula 1 top piston ring. (from the calculations for Figure 4).51 0W/20 Table 6: Effect of lubricant viscosity grade on top ring power loss and minimum oil film thickness (assuming fully flooded conditions) Clearly. and the predicted minimum oil film thickness. In addition.0 8. under fully flooded conditions.0 Oil Film Thickness ( m) Fully flooded Starved 1. and 470 W for the ring with a radius of 20 cm. This equates to an FMEP of approximately 70 kPa. then this estimate will be too low. under starved conditions. and taking into account the effects of starvation in a similar way to that Figure 5: Predicted oil film thickness under the top ring of a conventional European engine. Therefore.0 -270 -180 -90 0 90 180 270 360 450 Crank Angle (degrees) Figure 4: Estimated oil film thickness under the top ring of a Formula 1 engine.4. Although the top piston ring is starved of oil.76 kW. and assuming a top ring radius of curvature of 10 cm. we can see that moving from an SAE-15W/40 lubricant to an SAE- .0 litre gasoline engine at 7500 rpm. In Table 6. For the conventional engine.62 SAE259 0. then a ballpark figure for the total piston assembly friction loss would be 15 kW. There is a trade-off between reduced friction (and greater power available to the wheels) and engine durability. although this is not expected to impact significantly on the predicted friction loss (since the squeeze effect dominates at dead centers where the piston speed is low).0 4. Using the friction estimates for the Formula 1 ring pack.0 for the conventional engine. at 18000 rpm. the total ring power loss is thought to be around 23 60% of the total friction loss of the piston .0 2. Figure 5 shows the effect of oil starvation on predicted top ring oil film thickness for the conventional European 2.0 -360 -240 -120 0 120 240 360 Crank Angle (degrees) ¢ 0. this is at the expense of lower minimum oil film thicknesses. However. the assumption that the oil control ring is fully flooded. 375 W for the ring with a radius of 10 cm.

and there may be two or three positions around the bearing where the oil film thickness is low. it is also important to know what flow rate of oil is required to lubricate the bearings. to be lower than the values shown in Figure 7. the predicted oil film thicknesses for the higher speed condition were greater than those for the lower speed condition because the oil temperature was assumed to be the same in both simulations. The “squeeze” effect is also included. which allows for lubricant shear 30 thinning . Firstly. so the minimum oil film thickness is likely. the MercedesBenz M111 2. 120 240 360 -360 -240 In addition to oil film thickness. at 2500 rpm and 7500 rpm Figure 7 shows the predicted oil film thickness for both load curves using a modified version of the Short Bearing Approximation. In the absence of data for Formula 1 con-rod and main bearings.0W/20 lubricant will lead to a 30% decrease in top ring minimum oil film thickness. the load needs to be calculated from the adjacent con-rod bearing loads. for bearing loads that are dominated by gas pressure.0 litre gasoline engine. and in particular. the minimum oil film thickness in the bearing will occur at a position corresponding to the peak combustion chamber pressure. Figure 6 shows the calculated con-rod load of the M111 engine at two speeds. the minimum oil film thickness position will occur elsewhere. Two important conclusions can be drawn from the results. for the Formula 1 engine it would be necessary to know the angle of the “vee” in the V10 configuration. For inertially dominated loads. BEARING TRIBOLOGY Con-rod bearing loads may be estimated by knowing the combustion chamber pressure and the engine speed. whereas at the higher speed. If the flow rate is not sufficient. we try to illustrate how lubricant viscosity can influence oil film thickness and friction by analyzing bearings from a conventional engine. The simulations were carried out for an SAE-15W/40 lubricant. Figure 8 shows the flow rates predicted by the modified Short Bearing Approximation. in practice. In the calculations carried out above. 18 16 Oil Film Thickness ( m) 2500 rpm 7500 rpm 14 12 10 8 6 4 2 0 -360 -240 -120 0 £ 120 240 360 Crank Angle (degrees) 5000 0 -120 0 -5000 -10000 -15000 -20000 Crank Angle (degrees) Figure 7: Predicted con-rod bearing oil film thickness based on load data from Figure 6. (The standard Short Bearing Approximation often assumes . For the main bearing. inertial effects begin to dominate.0 litre gasoline engine. conventional engines. For low speed. and will also give a decrease in top ring friction of approximately 30%. However. At the lower speed the con-rod load is dominated by the gas pressure in the combustion chamber. 2500 rpm and 7500 rpm. lubricant starvation can occur which may lead to catastrophic damage to the bearings. Figure 6: Con-rod bearing loads for conventional European 2. 25000 20000 15000 10000 Load (N) 2500 rpm 7500 rpm that it is only the “wedge” effect that is important in the lubrication of engine bearings). inertial effects clearly start to become more important. In practice. the con-rod bearing load is often dominated by the combustion gas pressure. The benefits of moving to lower viscosity lubricants to obtain better fuel efficiency via lower engine friction have been well 24-29 documented . it is likely that the oil temperature at the higher speed condition is likely to be higher. as engine speed increases.

L is the bearing length.75 3 (17500/7500) *0.4 0. and also bearing deformation. This means that a ball-park estimate of the con-rod bearing friction is 1.25 Q v Z RLc where K is the lubricant viscosity.8 rather than 0. c is the radial clearance.1 Average Flow Rate (litres/min) 0.3 kW.01 Average Friction Power Loss (W) 73. (ii) the shear rate varies around the bearing.5 63.5W 1.75 FB v c 0. friction loss. an SAE-15W/40. assuming that the temperature of the oil is the same. required to supply the bearing : hmin v KZ RL3 W Figure 8: Flow rate (litres/min) required to avoid lubricant starvation of con-rod bearing. the maximum pressure in the bearing. Z is the angular speed of the bearing.6 Flow Rate (litres/min) 0. based on load data of Figure 6 Table 7 summarises the results for minimum oil film thickness. Clearly models that assume the viscosity of the lubricant is constant around the bearing will have limited usefulness. Therefore for the same lubricant.568 Table 8: Sensitivity of con-rod bearing results to lubricant viscosity grade at 7500 rpm It is recognized that the model used here is too simplistic for use as a detailed design tool of bearings.158 0. the friction power loss of the bearing. we can estimate the power loss of a typical Formula 1 bearing by comparing with the results on a conventional bearing. the loads in a Formula 1 engine are far greater – therefore we use 0.171 K 0.55 348. R is the bearing radius. and the lubricant flow rate.152 0. A useful model would need to take into account differences in temperature around the bearing.15 Average Friction Power Loss (W) 671.4 46.0 litre gasoline engine at 2500 rpm.63 37.25 K 0. FB(W). and W is the load on the bearing.7 425.75 L0. the Formula 1 con-rod bearing power loss would be expected to be a factor of 1.8 higher.25 R 2. and Table 8 summarises the results at 7500 rpm.1 0 -360 -240 -120 0 120 240 360 Crank Angle (degrees) 2500 rpm 7500 rpm change around a “real” bearing for a number of reasons.89 6.25Z 0. Since there are 10 con-rod bearings.5 litres.496 0. Although the short bearing approximation has it’s limitations. Q. hmin (Pm). The viscosity will vary because (i) the temperature of the oil changes around the bearing (cool at the inlet oil hole. the conventional bearing had a power loss of approximately 580 W at a speed of 7500 rpm.0.548 SAE-20W/50 SAE-15W/40 SAE-10W/30 SAE-0W/20 4. then for a given viscosity. If we simply assume that all the bearings dimensions of a Formula 1 engine are 0. for an SAE-15W/40 lubricant.8 times those of a conventional engine (although the displacement per cylinder for the Formula 1 engine is 0.514 0. for different lubricants (an SAE-20W/50.5 0.2 0.25W 0.5 577. it is useful for identifying trends. a Formula 1 engine may be expected to have a power loss of 13 kW from these SAE-20W/50 SAE-15W/40 SAE-10W/30 SAE-0W/20 2.13 3.6). an SAE-10W/30 and an SAE-0W/20) for the conventional European 2.7 0. and average flow rate.3 0. For the con-rod bearing considered in Table 8.179 Table 7: Sensitivity of con-rod bearing results to lubricant viscosity grade at 2500 rpm Lubricant Minimum OFT (Pm) 6. The lubricant viscosity may be expected to .4 0.75 R1.27 5.7 Average Flow Rate (litres/min) 0. hotter as it shears during flow around the bearing). In particular.75Z 1. The derivation of the above expressions is detailed in Appendix 1.4 0. and (iii) the pressure varies around the bearing.25 L1. Note that the expressions above are only valid if the minimum oil film thickness is small compared to the bearing radial clearance.5 PMAX v c 0. Lubricant Minimum OFT (Pm) 4. PMAX (MPa).72 3.3 litres. the bearings will be proportionately smaller. whereas that for the conventional engine we are considering is 0. the model can be used to predict the variation of key parameters such as the minimum oil film thickness. Although a Formula 1 engine will have a much greater engine speed than a conventional engine.

based on rigid bearing models. The effects of elastic deformation on friction were not reported. These are probably of larger dimension than the con-rod bearings. It is also in the valve train that friction modifier additives are thought to have the largest effect.6 times higher than the peak pressure in the conventional conrod bearing. Friction losses may then also be calculated. due to temperature variations. This approach has been used successfully by 39 40 41 researchers from Shell . once the pressure in the bearing exceeds a certain value (in the case of the bearing analysed by Fenner et al this pressure was 200 MPa). K is the lubricant viscosity (at the appropriate temperature) and Z is the camshaft angular speed. A lower viscosity lubricant will give a higher boundary friction torque. for both hydrodynamic and elastohydrodynamic lubrication conditions. under normal operating conditions. The reason is that the pressures become so high that elastohydrodynamic lubrication occurs. and these are effective at reducing the boundary friction torque. In the term for hydrodynamic friction. there is a central portion of the bearing where the oil film thickness is essentially constant. and due to the pressure variation around the bearing. For high pressures (those greater than about 50 MPa) it is likely that a rigid bearing model will start to become inaccurate. and from knowing the cam and follower profiles. since it is generally accepted that the valve train is in the mixed or boundary regime for most of its operating conditions. and it is necessary to solve both the Reynolds’ equation and the equation that describes the elastic deformation of the bearing surfaces. The boundary friction torque may be calculated by estimating the oil film thickness using elastohydrodynamic theory. Therefore the main conclusion was that the peak pressure in the con-rod bearing was much lower than would have been predicted from a rigid bearing model (the rigid bearing model predicted a peak pressure of 1200 MPa. In a valve train lubrication model.0 litre gasoline engine (with a direct acting bucket tappet valve train system). Figure 9 shows the valve lift curve for a typical European 2. The simple equations above are very approximate. For the detailed design of bearings. robust 31-34 methods available . not many take full account of the way in which the lubricant viscosity may vary around the bearing. (It should also be noted that the peak pressure in the Formula 1 conrod bearing is expected to be approximately 1. due to the changing shear rate. there are many fast. and should be used for identifying trends only. the valve train friction torque increases as the viscosity of the oil decreases. Measurements of valve train friction torque frequently show that. C is a constant. This is certainly true as far as conventional engines are concerned. whereas the model which took elastic deformation into account predicted a peak pressure of only 200 MPa). . Ford and Nissan . whereas the hydrodynamic friction torque will be lower. where an approximate expression is : *TOTAL *B  CKZ where *B is the friction torque due to boundary friction. and the second term represents hydrodynamic friction in the valve train system (both from the cam-follower contact and from the camshaft bearings). ignoring any change in the bearing load). Total friction torque in valve train is *TOTAL.bearings. Concluding this section. This figure may have to be reduced if oil temperatures are substantially higher than for the conventional engine. and assigning an effective friction coefficient. and then comparing this value to the combined surface roughness of cam and follower. it effectively became constant. it is suggested that although there are many sophisticated models of journal bearings in widespread use. and friction. So we could assume that each main bearing may contribute 2 kW. This is generally a time consuming calculation that requires a numerical treatment either by 35-38 the finite element or finite difference techniques . the oil film thickness may be calculated from elastohydrodynamic 1 lubrication theory . The authors find that under high loads. Therefore a ball-park figure of the engine bearing power loss is 25 kW. and where the pressure distribution is effectively that of a Hertzian contact. There are 6 main bearings for the V10 engine. The paper on the elastohydrodynamic lubrication of 36 journal bearings by Fenner et al is particularly illuminating. The expressions are not recommended for use in the design of bearings. are expected to have a significant effect on predictions of oil film thickness. and is found to give good agreement with experimental measurements. Friction modifier additives can be added to the lubricant. the kinematics of the cam-follower system need to described. VALVE TRAIN TRIBOLOGY Lubrication in the valve train differs from that in the bearings and the piston assembly. The variation of lubricant viscosity around the bearing. In essence.

5 0. and under motorway conditions (7500 rpm . Figure 12 details the engine friction breakdown at idling. the valve lift curve will differ. under urban conditions (2500 rpm. 0. total friction for the engine can be calculated.0 litre gasoline engine. and the type of lubricant used. . with a temperature of 100qC. Each cam has an approximate power loss of 37.00 -180 -90 0 Cam Angle (degrees) 90 180 Figure 9: Valve lift for typical European 2.0 litre gasoline engine Figure 11 shows the predicted friction torque for this valve train. the spring loads are likely to be far higher. the valve train friction power loss will be approximately 599.00 2. Table 9 shows how the results differ when different lubricants are used in the engine (for urban conditions). For a high performance engine.high load).10 2. at an engine speed of 2500 rpm. of which 30 W is due to boundary friction).0 litre gasoline engine For the conventional European 2.50 0 -180 -90 0 Cam Angle (degrees) 90 180 0. 2.8 W (of which 481. given the engine operating conditions. the total number of con-rod and main bearings. These calculations were performed assuming a standard SAE15W/40 oil. medium load).5 W. at 2500 rpm.50 TOTAL ENGINE FRICTION By counting up the total number of pistons in an engine.0 litre gasoline engine Figure 10 shows the predicted oil film thickness for this engine. For the typical European engine.00 -180 -90 0 Cam Angle (degrees) 90 180 Figure 10: Predicted valve train oil film thickness in conventional European 2. and the components making up the valve train will be lighter. assuming an SAE-15W/40 lubricant.00 ¤ 1. and the total number of cams.50 Figure 11: Predicted boundary friction torque for one cam of the conventional European 2.50 1. Oil Film Thickness ( m) 2.8 W is due to boundary friction.00 0.50 Valve Lift (mm) 1.00 7.5 5 Friction Torque (Nm) 1.

you would use a lower viscosity lubricant (again with a friction modifier). the valve train friction is likely to make a smaller contribution than at lower speeds.5 Piston assembly 743.5 467. Figure 13 summarises this very rough analysis for a Formula 1 engine at 18000 rpm.2 Total 1942. We estimated piston assembly losses of around 30 kW. However. The reason why the SAE-0W/20 oil has a sharp decrease in valve train friction compared to the other oils is that it is assumed that it contains a friction modifier (if it didn’t contain a friction modifier the valve train friction power loss would have been 641. Power Loss = 392. with consequent improvements in the available power to the wheels.6 554. The author would also like to thank Shell Global Solutions (UK) for permission to publish this work. whereas if the engine has less boundary friction. that include accurate information about how the lubricant viscosity varies with temperature. It is important to use models for the piston assembly.0 litre engine under urban operating conditions (* Note that the SAE0W/20 oil is assumed to be friction modified) In Table 9.4 Bearings 610. This entails choosing the lubricant viscosity which gives the lowest friction over the range of operating conditions appropriate for the engine. and the associated friction losses.1 SAE410. and choosing an optimum friction modifier for reducing friction in boundary lubricated contacts. and bearing friction losses of 25 kW. we would expect valve train friction losses to be less than 10 kW (and that this would mainly be hydrodynamic friction losses in the valve CONCLUSION In this paper I have attempted to demonstrate that the physical properties of a lubricant are extremely important in determining both the minimum oil film thickness in key lubricated contacts. However.9 1647. for European 2.train). Unfortunately. Figure 12: Engine friction breakdown for various operating conditions for European 2. with a friction modifier. To optimize power output from a high performance engine. We have compared a conventional gasoline engine with a high performance Formula 1 engine. At 18000 rpm.2 1282.2 380. If the engine has a lot of boundary friction. Glyn Roper and Mark Wakem for useful discussions.5 1844. it is possible to decrease engine friction by 10 or 20% by using the 41-47 correct lubricant. There are a number of papers which the reader can refer to for more information on total engine friction calculations. you would use a higher viscosity lubricant. it can be seen that reducing the lubricant viscosity is effective at reducing friction in the bearings and piston assembly.0 litre engine Oil grade SAE20W/50 SAE15W/40 SAE10W/30 Valve train 588. we know that at the maximum engine speed.0 492. we do not have sufficient details of a Formula 1 valve trains system to enable us to calculate the total engine friction of a Formula 1 engine. and shown that matching the lubricant to the engine can lead to significant reductions in engine friction. 18000 rpm. By looking at engine friction results for the conventional engine. reducing the lubricant viscosity causes an increase in the valve train friction. and the valve train.4 0W/20* Table 9: Sensitivity of engine and component friction to lubricant viscosity grade.9 677. ACKNOWLEDGMENTS The author would like to thank Simon Dunning.1 567.1 W Idling Conditions Power Loss = 1844. . the engine bearings.8 W Motorway Conditions Valve Train Bearings Piston Figure 13: Approximate estimate of engine friction in a Formula 1 engine at 18000 rpm (valve train friction estimate is uncertain due to lack of information) The strategy for minimizing engine friction in a Formula 1 engine is fairly simple.2 W). and pressure. shear rate.9 W Urban Conditions Total Friction = 65 kW Valve Train Bearings Piston Valve Train Bearings Piston Valve Train Bearings Piston Assembly Power Loss = 7461. it is necessary to choose a lubricant which gives the lowest possible friction.5 599.8 625.

113. 1979 21. pp 91-106. Colloid Sci. June 2001 6. pp 305314. Vol 213.F. Bair. Delft.C. 73. Taylor. Engrs. K. Baker.. Saito & R. Proc. P. T.C.P. “The Effects of Lubricant Rheology on the Performance of Journal Bearings”. Engrs. pp 1-17. Instn.W. Jr. Papay. Bates. Ruddy. Yokohama. Conf. J. article in Infineum Insight.E. Series F. J. ASME Journal of Lubrication Technology. Cheng. Roelands. Larsson. Proc. “Development and Field Test Performance of Fuel Efficient SAE 5W-20 Oils”.V. Eriksson. “Recent Advances in Engine Bearing Design Analysis”. Coy.J.A. “Rheology of Non-Newtonian Fluids: A New Flow Equation for Pseudo Plastic Systems”. Brown. 26. P. “Lubricant Properties for Input to Hydrodynamic & Elastohydrodynamic Lubrication Analyses”. Taylor. J. Dursunkaya & M. “Facing the Challenges of the Future in Crankcase Lubricants”. “Piston Assembly Friction & Wear: The Influence of Lubricant Viscometry”. “Fuel Economy Factors in Lubricants”. Mech. 7. Part J. Bell. pp 185-189.I. R. “The Elastohydrodynamic Solution of Journal Bearings Under Dynamic Loading”. Vol 211. Mech. Issue No. Booker. T. pp 382-389. Ting & J. Johnson.. Y-R. M. Engrs. pp 825-834. Wright. pp 115-126. Part J. ASME Journal of Lubrication Technology. pp 1-15.S. Part J. Energy Cons. ASME. J. Int.I. pp 427-433. Rohde.A. K.I. 8 Asia Fuels & Lubricants Meeting. “The Inclusion of Lubricant Shear Thinning in the Short Bearing Approximation”. 10. Leeds-Lyon Symposium on Tribology. 1964 12.J. Jap. Coy. pp 147-152. Jeng. T.M. Milton. R. “The Influence of Lubricant Rheology on Friction in the Piston Ring Pack”. “Fuel Economy Engine Oils: Present & Future”. 171. K. Yaguchi & K. 93. pp 417-437. 1996 27. van Os. 107.. Mech. Lyons. Inst. Yamada. 1965 3. 1999 31. Mech. “The Viscoelastic Properties of Multigrade Oils and Their Effect on Journal Bearing Characteristics”. Taylor.. 1966 7. M. Goenka. pp 17-27.. Instn.P.N.M. Tribology Int. Cross. N. Netherlands. Int. SAE 821226 26. Trans. 20. Keribar.J. B. Part J. Paranjpe. B. no 11. 1997 34. Williamson. Non-Newtonian Fluid Mechanics. van Arsdale.L.F.M. ASLE Preprint No. R.F. 17. III: The Role of Elasticity in Bearing Performance”. Fleming. 2001 16. B. “Piston Ring Lubrication & Cylinder Bore Wear Analysis”.P. Wright.L. SAE 952341 29. 2002 9. Hoglund. J. “Dynamically Loaded Journal Bearings: Numerical Application of the Mobility Method”. Engrs. pp 23-52. Oh & P. P. “An Integrated Model of Ring-Pack Performance”. K. Jan 2002 30. R. 1984 15. A. pp 35-46. Walters. London) 2.H. 1993) 17. Thompson & J. “Rapid Performance Evaluation of Journal Bearings”. Vol 214. C. C. published by the Society of Automotive Engineers. R. Kitahara.O.P. Moore. R. Tribology International. Taylor & R. 1971 (Errata : p 315. Jackson & B.L. 34. P.P. “The Principles of Lubrication”. “The Development of Fuel Economy th Lubricants”.. 1991 23. Mech. 84-LC-1C-1. “A Study of Piston Ring Lubrication”. Instn. Mech.G. Proc. E. “Correlational Aspects of the Viscosity-Temperature-Pressure Relationship of Lubricating Oils”. 1984 28. ASLE Trans. K. Theoretical Analysis of a Single Ring and a Complete Ring-Pack”. Rao. Mech. “The Behaviour of Lubricants in Elastohydrodynamic Contacts”. Part J. Mayer. S. “The Influence of Multigrade Oil Rheology on Friction in Dynamically Loaded Journal Bearings”. 96. Coy & A.I. 1999 35.J. 1985 . M.M. J. SAE 920489 33. M. Sjoberg & E. Z. Proc. “A Review of Engine Bearing Analysis Methods at General Motors”.L.I. Vol 216. D. Goenka & R.K. Williamson.K.A.. Cameron. Sorab & W. Proc.A. “Piston Ring Lubrication – Part II.G. Taylor (Editor). Engrs. Engrs. T. Larsson. 30. Saunders. “A Correlation for the Pressure & Temperature Dependence of Viscosity”. B. Xu. Economou. Part J. 1995 24. Proc. Technology : Frontiers in Research & Design. “Improved Fuel Efficiency by Lubricant Design: A Review”. Through Fluid Film Lubr.C. L. “Formula 1 Technology”. J. “European Activity Concerning Engine Oil Viscosity Classification – Part IV – The Effects of Shear Rate and Temperature on the Viscosity of Multigrade Oils”. Instn. ASME Journal of Tribology. “Engine Tribology”. Strachan & A. R.. 1974 20. 1997 13. SAE 830027 4. Trib.N. Damrath & A. Wilcock & H. P. H. PhD Thesis. Instn. A.. 1971) 32. 1966 (published by Longmans Green. Proc..E. pp 604-610. Singapore. 1957 19.L. Proc. 2000 25. Dowson. Engrs. “The Shear Rheology of Thin Compressed Liquid Films”. Vol 214. D. ASME Publication. Tribology Transactions. Taylor. “Engine Friction & Bearing Wear. Athre & S. “Reducing Friction Losses in Automobile Engines”. 1997 10. 1991 8. edited by S. A. M. SAE 1999-01-3670 14. Okrent. S. Inoue.S. Instn. D. Hutton. R. 30.REFERENCES 1. 213. Hirani. Engrs. pp 389395. Hoshi. H. SAE 920492 22. Technische Hogeschool. “Non-Newtonian Effects in Elastohydrodynamic Lubrication”. 1992 11.M. Williamson.S. pp 168-176. Part I – Theory. Mech. pp 239-251. pp 1423-1428. SAE 941981 18.C. p 1189. “Friction & Lubrication Analysis of a Piston Ring-Pack”. E.. J.F. published by Elsevier (Tribology Series.V. Soc.Biswas. 2000 5. Eilon & O.

in terms of the key problem variables.1.1) where W is the load (N). Jeng. PO Box 1. Chester. Cheshire Innovation Park.. STLE Journal of Lubrication Engineering. pp 235-246. 7. Masuda. K. H.5Z 0. Part J. Vol 202. Mech. maximum pressure.(A1. Fenner. Leeds.J. R.5¨ 0. J. World Tribology Congress. Proc. Rebora & F. Bell. The eccentricity ratio is defined as: H § hmin · ¨1  ¸ c ¹ © …(A1.s).5 L1.T.. 2001 39.D. 1997 43. 1990 42. Z is the angular speed (rad/s). T.3) and so: hmin § K 0. 1996 38. the lubricant flow rate and the maximum pressure. 19 LeedsLyon Symposium on Tribology. McIvor. P. “Development of a Friction Prediction Model for High Performance Engines”. R. R. 1992 37. Staron & P.1 is solved numerically. In this limit. M. Bishop. Paranjpe & A.C. Z.S. 1988 APPENDIX 1 In this Appendix we derive the relationships quoted in the paper for oil film thickness. Cusenza.I. In the standard Short Bearing Approximation (see for example reference A1). Vol 73. Taylor.C. L.A. R.36. the numerical solution has the disadvantage that it does not explicitly reveal how the minimum oil film thickness in the bearing varies with the key variables (W. A. c). SAE 920488 45. Kai. SAE 920487 44. K.S.5 R 0. Conway-Jones. 47. “An Analysis of Valve Train Friction in Terms of Lubrication Theory”. SAE 890836 47. Paranjpe & Y-R. and (1-H) = hmin/c. SAE Transactions. Instn. Engrs. Applying these approximations to equation A1. H. Xu. D.Taylor@Shell.N. “FLARE: An Integrated Software Package for Friction and Lubrication Analysis of Automotive Engines – Part I: Overview & Applications”. pp 567-573.5 ¨ ¸ W © ¹ . H | 1. and H is the eccentricity ratio. I. “Elastohydrodynamic Analysis of Connecting Rod Bearing for High Performance Engines”. Leeds-Lyon Symposium on Tribology. friction power loss.K Goenka. from the Short Bearing Approximation.5 · ¸ | 0. “Development & Evaluation of a Friction Model for Spark-Ignition Engines”.(A1. the friction power loss. T. equation A1. and lubricant flow rate.G. Proc.A. we often find the following relationship between load W and eccentricity ratio H: W KZH RL3 S § 16 · 2 ¨ 2  1¸H  1 2 2 2 c (1  H ) 4 © S ¹ …(A1.L.M.1 and derive analytical expressions for the minimum oil film thickness. This expressions are valid when hmin<<c. “Engine Friction – A Change in Emphasis”. SAE 830165 41.I. K is the lubricant viscosity (Pa. Engrs.N. CH1 3SH. Vol 211. Frequently. K. J. T. “Engine Friction : The Influence of Lubricant Rheology”. pp 215-226. c is the radial clearance (m). “Elastohydrodynamic Lubrication in Plain Bearings”. R is the bearing radius (m). Colgan & J. Stefani. L is the bearing length (m). Goto & S. Nitschke & J. “FLARE: An Integrated Software Package for Friction and Lubrication Analysis of Automotive Engines – Part II: Experimental Validation”. Although this is efficient and fast. Hamai. R. Patton. “A Predictive Model for Wear in Automotive Valve Train Systems”. Xu. Vienna. Monaghan. J. UK E-mail : Robert. “Effect of Design Variables on Friction and Economy”.B. Heywood.4) .. “The Effect of Compliance on Peak Oil Film th Pressure in Connecting Rod Bearings”... we obtain the simplified equation below: CONTACT Ian Taylor is in the Automotive Lubricants Group at Shell Global Solutions and can be contacted at : Shell Global Solutions (UK).com W| KZ RL3 4hmin 2 . Willermet.2) where hmin is the minimum oil film thickness in the bearing. Mech. Instn. 1965 46. In this Appendix we start with equation A1. SAE 892145 40.

K = 10 mPa. The relations derived in this Appendix give : hmin = 2.25 1.7) PMAX | 125 c 0. Tm.25 Pm. we consider the maximum pressure in the bearing.. we find that FB (W) is given by : FB | 2S K 0. FB(W). P..25 1. H | 1.25 R 2.We also know that the friction power loss.5 … (A1. The maximum pressure will occur when the derivative of the above expression (with respect to T) is equal to zero. Stachowiak and A..2 MPa. in the bearing: P · 3ZKH sin T § L2 ¨  y2 ¸ 2 3 ¨ ¸ c (1  H cos T ) © 4 ¹ . consider the case where : L = 21 mm. by G.. since we are assuming that hmin << c.5) 1 hmin §h · cosT m | 1   O¨ min ¸ 5 c © c ¹ 2 .8) As an example of the use of these relationships. in the standard short bearing approximation by: calculations show that the angle at which this occurs.. “Engineering Tribology”. PMAX = 229. .75 L0.10) and By using the expression for hmin derived in A1. is given by: FB 2S KZ 2 LR 3 c(1  H 2 ) 0. is given.11) Therefore.5 . PMAX = 227.s.W. c = 30 Pm..7 W.(A1.25 0. Detailed . so: Q | Z RLc This completes the derivation of the relationships quoted in the section on bearing modeling.25 c 0.5 W.7 MPa.. the lubricant flow rate required to avoid lubricant starvation is given by: Q Z RLcH . Elsever Tribology Series No.5W 1.(A1.. Finally. FB = 125. R = 25 mm. the maximum pressure in this model will occur on the bearing center line.75W 0.12) In the approximation we are using here.(A1.75 36 5 Z K R L .4.. A numerical solution of the Short Bearing Approximation gives : hmin = 2.W. FB = 121.25 0..(A1.(A1.22 Pm. where y = 0..(A1. 24 Clearly.75Z 1..9) REFERENCE A1.6) sin 2 T m | 2 § hmin ¨ 5© c · §h · ¸  O¨ min ¸ ¹ © c ¹ 2 .. Batchelor. W = 20000 N. and assume the engine speed is 1700 rpm. The standard Short Bearing Approximation gives the following expression for the pressure.(A1.. the expression for the maximum pressure in the bearing becomes: We also know that in the Short Bearing Approximation.