The Turbo Trac Traction Drive CVT Jeremy Carter, Brad Pohl SOUTHWEST RESEARCH INSTITUTE

Size: px

Start display at page:

Download "The Turbo Trac Traction Drive CVT Jeremy Carter, Brad Pohl SOUTHWEST RESEARCH INSTITUTE"

Rosemary Lester
5 years ago
Views:

1 4CV-38 he urbo rac raction Drive CV Jeremy Carter, Bra Pohl SOUHWES RESEARCH INSIUE Bill Rutter III, Richar Raney, Bill Saler URBO RAC, INC. Copyright 24 SAE International ABSRAC A unique an attractive variator mechanism has been evelope by urbo rac, Inc. an Southwest Research Institute (SwRI) for initial use in a heavy uty iesel truck application. High efficiency levels have been preicte with analytical moels an confirme with actual test ata. Further, this variator incorporates a very stable an simple control system an has extremely high torque capacity. he prototype of the variator mechanism has also been configure with a moifie Allison 65 series transmission for use as a series application in a Peterbilt truck, the final configuration will be a split power esign. he setup inclues a preliminary control system that allows for highway riving. It is emphasize, however, that Allison i not contribute to this esign or any of the content of this paper. his assembly, along with reasonably low manufacturing costs, has emonstrate that this transmission may be a reasonable alternative for automotive an other mobile or inustrial applications. his paper will present some of the results of an analytical evaluation of the variator as well as actual efficiency an urability test results of an initial prototype. Further, a brief iscussion of the installation of the custom transmission into an actual vehicle is also inclue. INRODUCION A schematic of the urbo rac variator is shown in Figure 1. he variator consists of a central isc, mounte on a coaxial central shaft. he isc engages the central shaft via a ball spline that allows for movement of the isc along the length of the variator while transmitting torque through the central shaft. he variator incorporates an array of cones configure such that the inner surface is parallel to the center axis of the system. he cones operate in constant an irect elastohyroynamic (EHD) contact with a central isk. As the isc is move axially from the one en of the cones to the other, the spee ratio changes steaily. his system has been esigne an patente for use in both a single path an split powere configuration over an array of automotive an even inustrial applications. Southwest Research Institute conucte a preliminary esign stuy of the patente variator concept in a split powere configuration with an 8 range gearbox for use in a 25 hp iesel engine application, as shown in Figure 1. he stuy was conucte to provie etail for builing a prototype of the variator. he analysis inclue stress an fatigue calculations, contact efficiency preictions with elastohyroynamic (EHD) theory, as well as a ynamic moel of the system operating in a iesel truck. o conuct this stuy, a preetermine loa/spee uty cycle was require. SwRI consulte with Detroit Diesel Corp who provie a ata file for the Route 72 Mixe raffic rive cycle. Several of the plots presente in this paper inclue results from that stuy an show preictions of the variator in several of the 8 ranges in both steay state operation an in ynamic operation over the uty cycle profile. Hyraulic Piston Pressure Input Bearing 1 Cone Input Risc Shaft Bearings Rcone, Normal Force Disc Rcrown Cone Roller raction Contact Control Force Bearings φ Central Shaft Bearing 2 Cone Output 1st Stage Planetary Sun Figure 1: Schematic of the urbo rac Variator

Piston area of the cone piston, φ = ilt angle of the cone center axis with reference to the transmission axis his equation is incorporate into a torque balance of the system to etermine the pressure

2 Piston area of the cone piston, φ = ilt angle of the cone center axis with reference to the transmission axis his equation is incorporate into a torque balance of the system to etermine the pressure require for a given input torque as follows: in SR sin( φ) P = µ r A n ( 2 ) Figure 2: Conceptual Drawing of the urbo rac ransmission (U.S. Patent number 6,1,42) Following the esign stuy, a prototype of the variator mechanism was built an installe on a test stan to measure efficiency an urability characteristics of the system. his configuration inclue only the variator by itself an i not incorporate a split power gearbox. hus, the results iffer in format from the preliminary esign stuy. Both the esign calculations an preliminary test results show goo efficiency an reasonably high torque capacity. Further, tests have also been conucte with a prototype variator installe in a Peterbilt truck an configure with a moifie prouction Allison 65 series range box. It is emphasize however, that Allison i not contribute to this research or any of the content of this paper. his paper highlights some of the etails from the concept esign stuy an outlines the test results of the first prototype buil of this novel variator. In-vehicle testing will also be iscusse briefly. MAIN SECION HYDRAULIC REQUIREMENS In orer to transmit torque, the variator requires a sufficient normal force across the rolling contacts to provie a frictional tangential force between the cones an the isc. his normal force is ynamically controlle in an effort to optimize the efficiency of the contacts. he cones are supporte via a sliing fit on cylinrical roller bearings at either en, allowing the cones to move along their center line. Hyraulic pressure is applie to pistons mounte on a thrust bearing on the input sie of the cones as shown in Figure 1. he axial motion of the cones is constraine by contact with the isc an a normal force is applie accoring to the following equation: where in is the input torque to the variator, SR is the variator spee ratio (ω in /ω isc ), µ is the traction coefficient for a given conition, n is the number of cones inclue in the system, an r is the raius of the isc. he cones are esigne with a 15 taper angle that allows for pressure requirements similar to a typical automotive pump elivery as shown in Figure 3. Cone Pressure, kpa Input orque N-m Variator Ratio (ω_in/ω_out) Figure 3: Cone Piston Pressure Requirements Versus Input orque an Variator Ratio (Assuming a friction coefficient =.4 with 8 cones in the system an a isc raius of 127 mm) RAIO CONROL he central isc in this variator is mounte on a housing that constrains the isc to a plane perpenicular to the center axis of the system. he ID of that housing engages the central shaft via a ball spline allowing for movement fore an aft while still transmitting torque to the center shaft. he isc position is controlle with a lever arm attache to the bearing housing that is riven by a linear step motor actuator as shown in Figure 4. P A N = ( 1 ) sin(φ),where N = Normal force across the contact, P = Flui pressure against the cone piston ( #16 in Figure 2), A =

CONAC SPIN he spin velocity (or rill spee) is efine as the ifference in the rotational spee of the riving an riven rollers in a irection perpenicular to the contact plane.

3 CONAC SPIN he spin velocity (or rill spee) is efine as the ifference in the rotational spee of the riving an riven rollers in a irection perpenicular to the contact plane. It is cause by the relative ifference in surface spees of both elements across the contact patch an is major source of power loss in traction rive CV s. he spin velocity is calculate by the following equation Figure 4: Ratio Control Actuator One important factor of this esign is that the ratio control is inherently stable, especially uner increasing loa. he spin in the contact oes cause the isc to want to walk to the large en of the cones, but this bias is always in the same irection an is reacte against by sie slip forces in the contact that ten to resist axial movement. Further, the spin bias is constant over the entire ratio range. So the actuator will never have to eal with suen force reversals. By contrast a toroial variator is an inherently unstable configuration, requiring management of the power roller in at least 2 egrees of freeom. Similarly, other variator esigns may have more intrinsic stability, but the loa-inuce bias can quickly change irection as the system passes certain spee ratios (e.g. belt type CV s an the Kopp Variator at approximately a 1:1 ratio). his is a efinite avantage of this system. SYSEM KINEMAICS he variator incorporates an input gear mesh with a tooth ratio of he overall variator ratio (input spee, ω in, to isc spee, ω ) ranges from approximately 2 to.5, which is sufficiently wie for normal automotive applications. With this variator configure with a final rive ratio of approximately 4:1, vehicle spees from to 6 mph can be obtaine with one pass across of the cones at a constant engine spee of 12 rpm. Figure 5 shows the variation of the variator ratio versus the axial position of the isc, x = is the smallest en of the cone. Variator Ratio (Input Spee/Disc Spee) Variator Ratio Change 5 1 Axial Position, X (mm) Figure 5: Variator Spee Ratio Change Versus Axial Position of the Disc ω = ω sin( φ) ω sin( ) ( 3 ) s1 c θ where ω s1 = spin velocity, ω c = cone rotational velocity, ω = isc rotational velocity, φ = cone taper angle, θ = angle of the center line axis of the isc (equals ) An interesting facet of the urbo-rac variator is that the actual contact spin is constant at all ratios for a given input spee. In most traction variators, the spin characteristic can change substantially over the spee ratio range. For example, toroial variators ten to have a horseshoe shape spin curve versus spee ratio. Other variators may yiel a steeply slope linear spin variation versus spee ratio. SRESS ANALYSIS AND ORQUE CAPACIY he stress analysis of this system was focuse primarily on the rolling element contacts, an two primary stress components were calculate. hese inclue the Hertzian contact stress an the maximum subsurface shear stress. he Hertzian contact stress is important for etermining the tractive capability of the EHD flui in the contact. It represents the stress in the normal irection across each contact. It is calculate as σ Hertz = 3 N 2 π a b ( 4 ) where, N is the normal force in the contact an a an b are the major an minor raii of the contact ellipse he subsurface shear stress is the primary component that etermines the failure point of the contact. Even if the Hertzian contact stress is greater than the yiel strength of the material, it will not fail unless the maximum shear stress is greater than the shear failure limit. his stress is calculate by a lengthy analysis. For the sake of space an clarity, only a portion of the results will be presente hear. More etail may be foun in reference [4]. ypically the shear failure limit is assume to be.5 to.6 of the yiel strength. Figure 6 shows the maximum Hertzian an subsurface shear stress over the simulate riving cycle.

4 3 25 F t = the tangential force across the contact, ω c an ω = rotational spee of the cone an isc respectively. Stress, MPa Duty Cycle ime, sec Subsurface Shear Stress Hertz Stress Figure 6: Maximum Hertzian an Subsurface Shear Stress Over the Simulate Duty Cycle he torque capacity of the urbo-rac variator is etermine primarily by the subsurface shear stress in contacting elements. he maximum shear stress inuce by the normal forces in the contact must be lower than the shear failure limit of the material to prevent any permanent amage. his can be accomplishe in two ways. First, the overall isc raius may be increase to limit the tangential forces an thus the normal forces impose. Secon, the crown raius (shown as R crown in Figure 1) may be increase to increase the area of the contact patch, reucing the stress in the contact (See Figure 7). Conversely, this coul have a negative affect on efficiency. Some of the factors influence by these two raii inclue: angential force transferre across the contacts. Normal force require at the contacts Stress an strain in the contacts Power loss in the contact Packaging space of the entire variator Rotational spee of the central shaft Spee ratios throughout the system Disc rotational inertia All of these factors influence the torque capacity of the transmission, which is irectly relate to the torque an spee of the central shaft accoring to the following equations: cs = r F ( 4 ) t Variator orque, N-m Crown Ra = 295 mm Crown Ra = 147 mm Crown Ra = 74 mm Max Sub-Surface Shear = 11.3 MPa Number of Cones = 8 raction Coefficient =.4 Disc Ra = mm Variator Ratio Figure 7: orque Capacity (Output orque) of the Variator for Crown Raii of 295 mm, 147 mm an 74 mm he torque capacity of the current variator esign is limite to about 41 N-m of output torque. his is clearly low if all of the power is transmitte through the variator, but will be acceptable in a split powere arrangement an coul be increase by increasing the isc raius, increasing the crown raius or increasing the number of cones in the system. Design improvements an urability stuies are uner way to increase the system torque capacity. EFFICIENCY PREDICION AND MEASUREMENS he torque losses in this variator come from the following sources: Elastohyroynamic (EHD) losses in the contacts (i.e. longituinal slip, sie slip an spin) Bearing losses losses Hysteretic loss of the contacting materials Viscous Churning he EHD losses are ue to the loa an geometry characteristics of the contact as well as the properties of the EHD flui incorporate. he normal force across the contact, the slip in both the traction an sie slip irections an the geometry of the contact all influence the level of EHD power losses. hese losses were preicte with the moel use in ref [4] for operating conitions in the 8 spee split powere configuration. Results are shown for efficiency in ranges 1 through 6 versus the variator ratio in Figure 8. rc ω = ω c ( 5 ) r where r c an r = the raii of the cone an isc respectively, cs = the torque in the central shaft

5 Efficiency 1% 98% 96% 94% 92% 88% 86% Variator Ratio, ω_in/ω_ Position Figure 8: Efficiency Preiction for Each Range in the 8 Spee Split Powere Application, with 678 N-m of Input orque In aition to the EHD losses, a small amount of energy is expene in the elastic eformation of the material on either sie of the contact. his is calle the hysteretic loss an is calculate as 3 b N ( 1 ) Hysteretic = ϕ 16 R where ϕ = the hysteresis loss factor of the material (equals about.1 for harene steel), b = contact ellipse minor axis length, R x = equivalent raius in the traction irection, N = normal force across the contact Further, the bearings are another substantial power loss in a system such as this, especially the cone bearings, which have to react all of the normal force at the contact. hese were calculate with equations provie by SKF [6]. hey inclue both viscous losses as well as loa epenent losses an are efine as viscous 7 = 16 1 f ( 2 ) a loa = f P1 o b m 1 ( 3 ) Losses ue to flui churning an inertial losses have not been inclue. Figure 9 shows a breakown of all of preicte power losses, both maximum an average, over the aforementione Detroit Diesel City Driving cycle with the engine running at a constant 18 rpm. Aitionally, Figure 1 shows the results of the overall variator efficiency analysis over the city riving cycle in the 8 range, split powere configuration. It is emphasize that this is the efficiency of the variator only, uner the spee an torque conitions seen in conceptual split powere configuration. Losses in the range box are not inclue. However, in the split power configuration the overall system efficiency will not change ramatically because these efficiency numbers are applie only to the power that passes through the variator. herefore the losses in the variator branch are ae in parallel with the mechanical branch. here will x m also be some ae losses for gear reuction ownstream when bringing the power back together. % of Input Power 16% 14% 12% 1% 8% 6% 4% 2% % EHD Hysteretic Piston Bearing Input Cone Brg Average: Output Cone Brg Max: Disc Central Brg Cone Input Figure 9: Power Loss Breakown as a % of Input Power in the 8 Spee, Split Power Application Variator Efficiency, % Duty Cycle ime, sec Figure 1: Overall Variator Efficiency Over the Detroit Diesel 72 Mixe raffic Driving Cycle In aition to power loss preictions, the variator itself has been run on a test stan at Southwest Research Institute an teste for efficiency (See Figure 11). he variator was run at 6 input spees, ranging from 75 rpm to 2 rpm, 5 input torques, from 6 N-m to 34 N- m, an 7 spee ratios, from.6 up to 2.. he efficiency tests were conucte on the variator only, with all of the power passing through the isc an out the central shaft. Figure 12 through Figure 14 show the efficiency results at three ifferent spee ratios versus spee an input torque. he results shown in Figure 12 through Figure 14 show variator efficiency ranging from about 65% at the absolute lowest point up to about 92% maximum, which is very similar to the range shown in Figure 1 preicting variator efficiency in the split powere application. It shoul be note that the torque applie to the variator in the split powere setup in most of the simulate conitions ranges from about 5 N-m to 135 N-m, which represents the highest efficiency levels shown in the actual test ata. Aitionally, it shoul be note that these efficiency curves are relatively flat, much like a conventional automatic. In contrast to other variators,

all of the power is riving through the variator. Figure 15 shows a schematic of the current transmission esign as installe in the vehicle.

6 these trens o not show steep efficiency changes over the spee ratio range. IN VEHICLE CONFIGURAION AND ESING In evelopment an preparation for the efficiency measurements, the variator was configure for use in a 2 range, non-split powere gearbox in a Peterbilt truck where all of the power is riving through the variator. Figure 15 shows a schematic of the current transmission esign as installe in the vehicle. Similarly, Figure 16 shows a picture of the actual transmission incluing the variator an the moifie Allison gearbox. Figure 11: urbo rac Variator Assembly Mounte on the SwRI est stan for Efficiency esting Efficiency 1% 95% 85% 8% 75% 7% 65% 6% Input Spee, rpm Input orque, N-m Figure 12: Measure Efficiency of the Variator Alone on the SwRI est Stan Operating at a Spee Ratio (ω in /ω ) of.6 Figure 15: Schematic of the Current In-Vehicle Design 1% 95% 85% 8% 75% 7% 65% 6% Input Spee, rpm Efficiency Input orque, N-m Figure 13: Measure Efficiency of the Variator Alone on the SwRI est Stan Operating at a Spee Ratio (ω in /ω ) of 1. Efficiency 1% 95% 85% 8% 75% 7% 65% 6% Input Spee, rpm Input orque, N-m Figure 14: Measure Efficiency of the Variator Alone on the SwRI est Stan Operating at a Spee Ratio (ω in /ω ) of 2. Figure 16: Current In-Vehicle Emboiment of the Variator Configure with the Allison 65 Series Range Box his custom transmission has been installe in the truck an function teste for normal city riving. he transmission was initially manually controlle by the river with a simple linear electric actuator for ratio control. A more sophisticate, automatic control system is currently being evelope for emonstration purposes. he split-powere emboiment inclues a geare neutral position of the isc, where all of the clutches can be engage, but the output spee is fixe at zero. As the isc is move off of the geare neutral position, the shear in the flui provies a smooth an steay launch. However, at positions very close to geare neutral, the system experiences a very near infinite torque ratio an the rolling contacts will experience slip. Further, evelopment an testing is being conucte to confirm the urability of the system with this launch metho.

However, the series esign oes not incorporate a geare neutral conition, but rather uses a clutch to launch. Emphasis on the Hyromechanical ype, SAE Paper #22-1-589. 6. SKF Bearing Catalogue, 1999 7.

7 However, the series esign oes not incorporate a geare neutral conition, but rather uses a clutch to launch. Emphasis on the Hyromechanical ype, SAE Paper # SKF Bearing Catalogue, Zaretsky, E. V., Selection of Rolling-Element Bearing Steels for Long-Life Applications, Effects of Steel Manufacturing Processes on the Quality of Bearing Steels, ASM SP 987, J. J. C. Hoo, E., American Society for esting an Materials, Philaelphia, 1988, p Machia, H., sutomu, A., Fatigue Life Analysis of raction Drive CV, SAE Paper # CONAC Figure 17: Peterbilt ruck Use for In-Vehicle esting of the urbo rac Variator CONCLUSION he urbo rac variator offers a reasonable alternative for any application requiring continuous spee variation. It is a relatively simple esign an provies for easy stable ratio control. Further, the efficiency levels are very esirable for any application an the spee ratio range is thoroughly sufficient. he torque capacity is reasonable, especially when incorporate in a split powere configuration. Further evelopment is necessary an being conucte for practicality an urability in high power vehicle applications. So far the functionality has been emonstrate in a 25 hp iesel truck an continue evelopment will provie more steay control an smooth stable operation. Aitionally, urability tests will be conucte in the lab to confirm the longevity of the system an improve efficiency an control capability. REFERENCES 1. Greenwoo, J. an Williamson, J., "Contact of Nominally Flat Surfaces, " Proc. R. Soc. Lonon A295, 3-319, Harris,. A., "Rolling Bearing Analysis," 4 th e. John Wiley an Sons, Inc., Loewenthal, S. H., Zaretsky E. V., "Design of raction Drives," NASA Reference Publication 1154, Miller, D. an Carter, J., he esign an Analysis of an Alternative raction Drive CV, SAE Paper Pohl, B, CV Split Power ransmissions, A Configuration versus Performance Stuy with an Author: Jeremy Carter Research Engineer Drivetrain Engineering Vehicle Systems Research Department Southwest Research Institute phone: Co-authors: Bra Pohl Research Engineer Drivetrain Engineering Vehicle Systems Research Department Southwest Research Institute phone: Bill Rutter III Presient, urbo-rac, Inc PO Box 3186 Milan, X 7972 Phone: Fax: GrinNog@aol.com Richar Raney urbo-rac, LD 28 Fannin Ave Milan, X 7975 Phone: Fax: raneyrc@yahoo.com Bill Saler Prototype Design Engineer, urbo-rac, LD 8225 E. Montebello Ave Scottsale, AZ 8525 Phone: Fax: wsaler@cova.net

INFLUENCE OF SHOCK ABSORBER MODEL FIDELITY ON THE PREDICTION OF VEHICLE HALF ROUND PERFORMANCE Chris Masini Oshkosh Corporation Oshkosh, WI

INFLUENCE OF SHOCK ABSORBER MODEL FIDELITY ON THE PREDICTION OF VEHICLE HALF ROUND PERFORMANCE Chris Masini Oshkosh Corporation Oshkosh, WI 2010 NDI GROUND VEHICLE SYSTEMS ENGINEERING ND TECHNOLOGY SYMPOSIUM MODELING & SIMULTION, TESTING ND VLIDTION (MSTV) MINI-SYMPOSIUM UGUST 17-19 DERBORN, MICHIGN INLUENCE O SHOC BSORBER MODEL IDELITY ON