際際滷

際際滷Share a Scribd company logo

MSC-VPD04 ArvinMeritor_01

Download as PPT, PDF
Ragnar Ledesma, PE, Ph.D.
Ragnar Ledesma, PE, Ph.D.

The document discusses wheel shimmy in heavy duty trucks, which was observed during testing of a prototype vehicle. A simulation model was developed and numerical experiments were conducted to understand the shimmy phenomenon and identify design changes. The analysis found that increased kingpin friction, stiffer steering components and tires, and reduced caster angle could eliminate shimmy by increasing damping. The recommended design modifications were implemented and resolved the issue.

1 of 26
Downloaded 19 times
Wheel Shimmy in Heavy Duty Trucks:Wheel Shimmy in Heavy Duty Trucks: Using Designed Numerical ExperimentsUsing Designed Numerical Experiments to Determine a Robust Solutionto Determine a Robust Solution Ragnar LedesmaRagnar Ledesma Principal EngineerPrincipal Engineer Commercial Vehicle SystemsCommercial Vehicle Systems ArvinMeritor, Inc.ArvinMeritor, Inc.
AbstractAbstract This presentation addresses the wheel shimmy phenomenon commonly encountered in heavy-duty trucks with dependent (solid beam) front axle suspensions. A MSC.ADAMS model of a class-8 tractor-semi-trailer combination has been developed to study the front wheel shimmy observed during road testing of a prototype vehicle. The MSC.ADAMS model was able to reproduce the shimmy event at the observed frequency and vehicle speed. The unstable shimmy mode comprised a coalescence of different vibration modes at the shimmy frequency. These vibration modes include the frame twist mode, frame lateral bending mode, axle roll mode, axle lateral mode, and steering system vibration mode. The solution strategy adopted in this study is to use designed numerical experiments in order to find the optimum combination of possible design changes to the front axle suspension for a given set of design constraints. Possible solutions to the shimmy problem include a combination of the following: increased kingpin friction damping, changing the location of the steering arm ball joint, changing the caster angle, increasing the lateral stiffness of the suspension by various means, and increasing the lateral stiffness of the tires. The DOE/sensitivity study results showed that tire relaxation (lag in the tire response), which results in a negative damping effect, plays a dominant role in wheel shimmy propensity. The sensitivity study identifies kingpin friction damping as the most effective means of controlling wheel shimmy.
Presentation OutlinePresentation Outline Introduction Vehicle Description and Field Test Observations Model Description Designed Numerical Experiments Analysis of Simulation Results Concluding Remarks
IntroductionIntroduction ArvinMeritor is the system supplier of the front axle and suspension for a new series of Class-8 tractors A prototype vehicle was tested on public roads and proving grounds A front wheel shimmy problem was observed at specific vehicle speeds and road conditions Vehicle dynamics simulation and analysis were requested in order to understand the shimmy phenomenon and seek acceptable modifications to the design of the front axle and suspension
Vehicle DescriptionVehicle Description Axle Loads: Front: 10,760 lbm Rear: 8,350 lbm (bobtail) Wheelbase: 230 inches Front axle beam: tubular Front axle suspension: hybrid air spring and mechanical leaf spring Tandem rear axle suspension: trailing arm and air spring Truck frame: 5/16 C-channels Cab: day-cab or sleeper variants
Field Test: ObservationsField Test: Observations Front wheel shimmy occurs at 6-7 Hz, and sustained at vehicle speeds greater than 30 mph Running modes analysis shows that shimmy is a coalescence of frame twist mode, frame lateral bending mode, engine yaw mode, axle roll mode, axle lateral mode, and steering system vibration mode
Modeling and Simulation ObjectivesModeling and Simulation Objectives Reproduce shimmy phenomenon observed in the field Develop an understanding of the shimmy phenomenon Determine the sensitivity of shimmy propensity to various front axle and suspension design parameters Recommend modifications to the design of the front axle and suspension
MSC.ADAMS Model DescriptionMSC.ADAMS Model Description Truck frame: modal neutral file from MSC.NASTRAN Front suspension: mechanical leaf spring (beam elements), air spring (nonlinear force-deflection curve), and shock absorbers (nonlinear force- velocity curves) Rear suspension: trailing arm (beam elements) and air springs Steering system: pitman arm, drag link, steering arm, tie rod arms, cross tube Cab, engine, transmission: rigid bodies Tires: Pacejka magic formula tire model
Truck Frame Vibration ModesTruck Frame Vibration Modes First Lateral Bending Mode
Truck Frame Vibration ModesTruck Frame Vibration Modes Second Lateral Bending Mode
Truck Frame Vibration ModesTruck Frame Vibration Modes Combined Lateral Bending and Frame Twist Mode
Model Verification: Natural FrequenciesModel Verification: Natural Frequencies vertical acceleration response lateral acceleration response
Model Verification: Shimmy SimulationModel Verification: Shimmy Simulation knuckle rotation: time history knuckle rotation: frequency spectrum
Model Verification: Shimmy SimulationModel Verification: Shimmy Simulation frame lateral acceleration: time history frame lateral acceleration: frequency spectrum
Designed Numerical ExperimentsDesigned Numerical Experiments Factors: Tire relaxation length (tire characteristic) Kingpin friction (steering system damping) Steering arm length (steering geometry) Kingpin caster angle (axle geometry) Vehicle speed (operational parameter) Front axle suspension lateral stiffness (suspension characteristic) Responses: Knuckle rotation about kingpin axis (standard deviation value) Truck frame lateral acceleration (standard deviation value)
ANOVA Table: Knuckle RotationANOVA Table: Knuckle Rotation Sum of Mean F Source Squares DF Square Value Prob > F Model 49.83 14 3.56 436.45 < 0.0001 B 16.57 1 16.57 2031.95 < 0.0001 C 7.43 1 7.43 910.62 < 0.0001 D 0.64 1 0.64 78.11 < 0.0001 AE 5.26 1 5.26 644.54 < 0.0001 BC 7.39 1 7.39 905.63 < 0.0001 BD 0.63 1 0.63 76.85 < 0.0001 BE 0.13 1 0.13 16.11 0.0002 CD 0.26 1 0.26 31.91 < 0.0001 ABE 5.16 1 5.16 632.47 < 0.0001 ACE 2.96 1 2.96 362.97 < 0.0001 ADE 0.10 1 0.10 12.31 0.0010 BCD 0.26 1 0.26 31.68 < 0.0001 ABCE 2.96 1 2.96 363.30 < 0.0001 ABDE 0.097 1 0.097 11.90 0.0012 Residual 0.40 49 8.155E-003 Cor Total 50.23 63
ANOVA Table: Frame Lateral AccelerationANOVA Table: Frame Lateral Acceleration Sum of Mean F Source Squares DF Square Value Prob > F Model 0.72 18 0.040 313.99 < 0.0001 A 4.470E-003 1 4.470E-003 35.17 < 0.0001 B 0.22 1 0.22 1766.13 < 0.0001 C 0.11 1 0.11 829.25 < 0.0001 D 7.787E-003 1 7.787E-003 61.29 < 0.0001 E 3.517E-003 1 3.517E-003 27.68 < 0.0001 F 1.444E-003 1 1.444E-003 11.37 0.0015 AB 4.221E-003 1 4.221E-003 33.22 < 0.0001 AE 0.076 1 0.076 600.86 < 0.0001 BC 0.11 1 0.11 827.49 < 0.0001 BD 7.614E-003 1 7.614E-003 59.92 < 0.0001 BF 1.942E-003 1 1.942E-003 15.28 0.0003 CD 3.045E-003 1 3.045E-003 23.96 < 0.0001 ABE 0.077 1 0.077 603.25 < 0.0001 ACE 0.045 1 0.045 354.47 < 0.0001 ADE 1.630E-003 1 1.630E-003 12.83 0.0008 BCD 2.989E-003 1 2.989E-003 23.53 < 0.0001 ABCE 0.045 1 0.045 353.18 < 0.0001 ABDE 1.651E-003 1 1.651E-003 12.99 0.0008 Residual 5.718E-003 45 1.271E-004 Cor Total 0.72 63
Response Surface: Knuckle RotationResponse Surface: Knuckle Rotation Effect of vehicle speed and tire relaxation length on wheel shimmy 0.939374 1.50165 2.06393 2.62621 3.18849 KingpinRotation 500.00 625.00 750.00 875.00 1000.00 30.00 37.50 45.00 52.50 60.00 A: tire relaxation leng th E: vehicle speed
Response Surface: Knuckle RotationResponse Surface: Knuckle Rotation Effect of kingpin friction and steering arm length on wheel shimmy 0.0883032 0.840693 1.59308 2.34547 3.09786 KingpinRotation 0.00 0.03 0.05 0.08 0.10 280.20 286.55 292.90 299.25 305.60 B: king pin frictionC : steering arm leng th
Response Surface: Frame Lateral AccelerationResponse Surface: Frame Lateral Acceleration Effect of vehicle speed and tire relaxation length on wheel shimmy 0.0950708 0.169451 0.243832 0.318212 0.392593 FrameLateralAccel 500.00 625.00 750.00 875.00 1000.00 30.00 37.50 45.00 52.50 60.00 A: tire relaxation leng th E: vehicle speed
Response Surface: Frame Lateral AccelerationResponse Surface: Frame Lateral Acceleration Effect of kingpin friction and steering arm length on wheel shimmy 0.0230786 0.110925 0.198771 0.286617 0.374463 FrameLateralAccel 0.00 0.03 0.05 0.08 0.10 280.20 286.55 292.90 299.25 305.60 B: king pin friction C : steering arm leng th
DOE ResultsDOE Results All factors were found to be significant to wheel shimmy propensity Kingpin friction was the most significant factor increasing kingpin friction helps reduce wheel shimmy propensity Interaction between vehicle speed and tire relaxation length (negative damping effect) was a dominant contributor to wheel shimmy 2, 3, and 4-factor interactions were present it is difficult to make general statements regarding the effect of any one specific factor on shimmy propensity
Design RecommendationsDesign Recommendations These recommendations are specific to the prototype vehicle: Increase kingpin friction Increase length of steering arm Use tires with higher lateral stiffness (bias ply tires) Reduce caster angle Increase suspension lateral stiffness
Design ModificationDesign Modification Initial Design of Front Axle and Suspension
Design ModificationDesign Modification Final Design of Front Axle and Suspension
Summary and ConclusionSummary and Conclusion A MSC.ADAMS model was used in simulating the front wheel shimmy phenomenon Using designed numerical experiments, the model was utilized in determining a robust solution to the shimmy problem Design modifications were made based on the findings of the DOE/sensitivity study Vehicle tests confirmed that the shimmy issue has been eliminated

More Related Content

MSC-VPD04 ArvinMeritor_01

  • 1. Wheel Shimmy in Heavy Duty Trucks:Wheel Shimmy in Heavy Duty Trucks: Using Designed Numerical ExperimentsUsing Designed Numerical Experiments to Determine a Robust Solutionto Determine a Robust Solution Ragnar LedesmaRagnar Ledesma Principal EngineerPrincipal Engineer Commercial Vehicle SystemsCommercial Vehicle Systems ArvinMeritor, Inc.ArvinMeritor, Inc.
  • 2. AbstractAbstract This presentation addresses the wheel shimmy phenomenon commonly encountered in heavy-duty trucks with dependent (solid beam) front axle suspensions. A MSC.ADAMS model of a class-8 tractor-semi-trailer combination has been developed to study the front wheel shimmy observed during road testing of a prototype vehicle. The MSC.ADAMS model was able to reproduce the shimmy event at the observed frequency and vehicle speed. The unstable shimmy mode comprised a coalescence of different vibration modes at the shimmy frequency. These vibration modes include the frame twist mode, frame lateral bending mode, axle roll mode, axle lateral mode, and steering system vibration mode. The solution strategy adopted in this study is to use designed numerical experiments in order to find the optimum combination of possible design changes to the front axle suspension for a given set of design constraints. Possible solutions to the shimmy problem include a combination of the following: increased kingpin friction damping, changing the location of the steering arm ball joint, changing the caster angle, increasing the lateral stiffness of the suspension by various means, and increasing the lateral stiffness of the tires. The DOE/sensitivity study results showed that tire relaxation (lag in the tire response), which results in a negative damping effect, plays a dominant role in wheel shimmy propensity. The sensitivity study identifies kingpin friction damping as the most effective means of controlling wheel shimmy.
  • 3. Presentation OutlinePresentation Outline Introduction Vehicle Description and Field Test Observations Model Description Designed Numerical Experiments Analysis of Simulation Results Concluding Remarks
  • 4. IntroductionIntroduction ArvinMeritor is the system supplier of the front axle and suspension for a new series of Class-8 tractors A prototype vehicle was tested on public roads and proving grounds A front wheel shimmy problem was observed at specific vehicle speeds and road conditions Vehicle dynamics simulation and analysis were requested in order to understand the shimmy phenomenon and seek acceptable modifications to the design of the front axle and suspension
  • 5. Vehicle DescriptionVehicle Description Axle Loads: Front: 10,760 lbm Rear: 8,350 lbm (bobtail) Wheelbase: 230 inches Front axle beam: tubular Front axle suspension: hybrid air spring and mechanical leaf spring Tandem rear axle suspension: trailing arm and air spring Truck frame: 5/16 C-channels Cab: day-cab or sleeper variants
  • 6. Field Test: ObservationsField Test: Observations Front wheel shimmy occurs at 6-7 Hz, and sustained at vehicle speeds greater than 30 mph Running modes analysis shows that shimmy is a coalescence of frame twist mode, frame lateral bending mode, engine yaw mode, axle roll mode, axle lateral mode, and steering system vibration mode
  • 7. Modeling and Simulation ObjectivesModeling and Simulation Objectives Reproduce shimmy phenomenon observed in the field Develop an understanding of the shimmy phenomenon Determine the sensitivity of shimmy propensity to various front axle and suspension design parameters Recommend modifications to the design of the front axle and suspension
  • 8. MSC.ADAMS Model DescriptionMSC.ADAMS Model Description Truck frame: modal neutral file from MSC.NASTRAN Front suspension: mechanical leaf spring (beam elements), air spring (nonlinear force-deflection curve), and shock absorbers (nonlinear force- velocity curves) Rear suspension: trailing arm (beam elements) and air springs Steering system: pitman arm, drag link, steering arm, tie rod arms, cross tube Cab, engine, transmission: rigid bodies Tires: Pacejka magic formula tire model
  • 9. Truck Frame Vibration ModesTruck Frame Vibration Modes First Lateral Bending Mode
  • 10. Truck Frame Vibration ModesTruck Frame Vibration Modes Second Lateral Bending Mode
  • 11. Truck Frame Vibration ModesTruck Frame Vibration Modes Combined Lateral Bending and Frame Twist Mode
  • 12. Model Verification: Natural FrequenciesModel Verification: Natural Frequencies vertical acceleration response lateral acceleration response
  • 13. Model Verification: Shimmy SimulationModel Verification: Shimmy Simulation knuckle rotation: time history knuckle rotation: frequency spectrum
  • 14. Model Verification: Shimmy SimulationModel Verification: Shimmy Simulation frame lateral acceleration: time history frame lateral acceleration: frequency spectrum
  • 15. Designed Numerical ExperimentsDesigned Numerical Experiments Factors: Tire relaxation length (tire characteristic) Kingpin friction (steering system damping) Steering arm length (steering geometry) Kingpin caster angle (axle geometry) Vehicle speed (operational parameter) Front axle suspension lateral stiffness (suspension characteristic) Responses: Knuckle rotation about kingpin axis (standard deviation value) Truck frame lateral acceleration (standard deviation value)
  • 16. ANOVA Table: Knuckle RotationANOVA Table: Knuckle Rotation Sum of Mean F Source Squares DF Square Value Prob > F Model 49.83 14 3.56 436.45 < 0.0001 B 16.57 1 16.57 2031.95 < 0.0001 C 7.43 1 7.43 910.62 < 0.0001 D 0.64 1 0.64 78.11 < 0.0001 AE 5.26 1 5.26 644.54 < 0.0001 BC 7.39 1 7.39 905.63 < 0.0001 BD 0.63 1 0.63 76.85 < 0.0001 BE 0.13 1 0.13 16.11 0.0002 CD 0.26 1 0.26 31.91 < 0.0001 ABE 5.16 1 5.16 632.47 < 0.0001 ACE 2.96 1 2.96 362.97 < 0.0001 ADE 0.10 1 0.10 12.31 0.0010 BCD 0.26 1 0.26 31.68 < 0.0001 ABCE 2.96 1 2.96 363.30 < 0.0001 ABDE 0.097 1 0.097 11.90 0.0012 Residual 0.40 49 8.155E-003 Cor Total 50.23 63
  • 17. ANOVA Table: Frame Lateral AccelerationANOVA Table: Frame Lateral Acceleration Sum of Mean F Source Squares DF Square Value Prob > F Model 0.72 18 0.040 313.99 < 0.0001 A 4.470E-003 1 4.470E-003 35.17 < 0.0001 B 0.22 1 0.22 1766.13 < 0.0001 C 0.11 1 0.11 829.25 < 0.0001 D 7.787E-003 1 7.787E-003 61.29 < 0.0001 E 3.517E-003 1 3.517E-003 27.68 < 0.0001 F 1.444E-003 1 1.444E-003 11.37 0.0015 AB 4.221E-003 1 4.221E-003 33.22 < 0.0001 AE 0.076 1 0.076 600.86 < 0.0001 BC 0.11 1 0.11 827.49 < 0.0001 BD 7.614E-003 1 7.614E-003 59.92 < 0.0001 BF 1.942E-003 1 1.942E-003 15.28 0.0003 CD 3.045E-003 1 3.045E-003 23.96 < 0.0001 ABE 0.077 1 0.077 603.25 < 0.0001 ACE 0.045 1 0.045 354.47 < 0.0001 ADE 1.630E-003 1 1.630E-003 12.83 0.0008 BCD 2.989E-003 1 2.989E-003 23.53 < 0.0001 ABCE 0.045 1 0.045 353.18 < 0.0001 ABDE 1.651E-003 1 1.651E-003 12.99 0.0008 Residual 5.718E-003 45 1.271E-004 Cor Total 0.72 63
  • 18. Response Surface: Knuckle RotationResponse Surface: Knuckle Rotation Effect of vehicle speed and tire relaxation length on wheel shimmy 0.939374 1.50165 2.06393 2.62621 3.18849 KingpinRotation 500.00 625.00 750.00 875.00 1000.00 30.00 37.50 45.00 52.50 60.00 A: tire relaxation leng th E: vehicle speed
  • 19. Response Surface: Knuckle RotationResponse Surface: Knuckle Rotation Effect of kingpin friction and steering arm length on wheel shimmy 0.0883032 0.840693 1.59308 2.34547 3.09786 KingpinRotation 0.00 0.03 0.05 0.08 0.10 280.20 286.55 292.90 299.25 305.60 B: king pin frictionC : steering arm leng th
  • 20. Response Surface: Frame Lateral AccelerationResponse Surface: Frame Lateral Acceleration Effect of vehicle speed and tire relaxation length on wheel shimmy 0.0950708 0.169451 0.243832 0.318212 0.392593 FrameLateralAccel 500.00 625.00 750.00 875.00 1000.00 30.00 37.50 45.00 52.50 60.00 A: tire relaxation leng th E: vehicle speed
  • 21. Response Surface: Frame Lateral AccelerationResponse Surface: Frame Lateral Acceleration Effect of kingpin friction and steering arm length on wheel shimmy 0.0230786 0.110925 0.198771 0.286617 0.374463 FrameLateralAccel 0.00 0.03 0.05 0.08 0.10 280.20 286.55 292.90 299.25 305.60 B: king pin friction C : steering arm leng th
  • 22. DOE ResultsDOE Results All factors were found to be significant to wheel shimmy propensity Kingpin friction was the most significant factor increasing kingpin friction helps reduce wheel shimmy propensity Interaction between vehicle speed and tire relaxation length (negative damping effect) was a dominant contributor to wheel shimmy 2, 3, and 4-factor interactions were present it is difficult to make general statements regarding the effect of any one specific factor on shimmy propensity
  • 23. Design RecommendationsDesign Recommendations These recommendations are specific to the prototype vehicle: Increase kingpin friction Increase length of steering arm Use tires with higher lateral stiffness (bias ply tires) Reduce caster angle Increase suspension lateral stiffness
  • 24. Design ModificationDesign Modification Initial Design of Front Axle and Suspension
  • 25. Design ModificationDesign Modification Final Design of Front Axle and Suspension
  • 26. Summary and ConclusionSummary and Conclusion A MSC.ADAMS model was used in simulating the front wheel shimmy phenomenon Using designed numerical experiments, the model was utilized in determining a robust solution to the shimmy problem Design modifications were made based on the findings of the DOE/sensitivity study Vehicle tests confirmed that the shimmy issue has been eliminated