492

Int. J. Mech. Eng. & Rob. Res. 2014 K K Dhande et al., 2014

A STUDY OF THE EFFECT OF STEERINGPARAMETERS ON ROLLOVER PROPENSITY

K K Dhande1*, N I Jamadar1 and Vinod Sakhala1

*Corresponding Author: K K Dhande,kkdhande@gmail.com

Rollover propensity is an important safety issue which should be considered early in the designof a vehicle. Although there is a trend toward higher-tech solutions to mitigate rollover risk, wepropose that a vehicle designer should also be fully aware of the impact many of the vehicle'sdesign parameters have on rollover propensity. Such awareness is essential to makingappropriate engineering trade-offs throughout the vehicle development process. This Studyinvestigates the effect of various vehicle parameters on rollover propensity by performingsimulations. The accuracy of the simulation was verified by comparing with experimental datafrom On-field testing of rollover of sport utility vehicles. The vehicle model used in the simulationstudy considers the non-linear, transient dynamics of both yaw and roll motion. The vehiclemodel was subjected to the specific steering input defined by the National Highway Traffic SafetyAdministration (NHTSA), the Slalom Test. Investigations were aimed at identification of correlationsbetween the vehicle steering parametersand rollover propensity using validated vehicle simulation.During the study it was observed that one of the steering performance parameter aids in reducingrollover propensity. All such results are presented in this paper.

Keywords: VDC, ESP, RSC, ABS, NHTSA

INTRODUCTIONOver the past decade, the occurrence ofrollover incidents in vehicle accidents hasreceived much attention due to the increasein Sport Utility Vehicle (SUV) miles driven onthe roadways and litigation claiming thatengineers are to blame for designing unsafevehicles. The NHTSA reported in 2010 that 3%

ISSN 2278 – 0149 www.ijmerr.comVol. 3, No. 3, July 2014

© 2014 IJMERR. All Rights Reserved

Int. J. Mech. Eng. & Rob. Res. 2014

1 Mech. Engg. Dept. Pad. Dr. D.Y. Patil Institute of Engg &Tech, Pimpri, Pune, India.

of all light passenger vehicle crashes on Indianroads involve rollover, yet rollover accidents areresponsible for 1/3 of all passenger vehicleoccupant fatalities (Ponticel, 2003). Thestatistics reveal that SUVs rollover fatalitiesare more prevalent than in cars; 59% of totalfatalities in SUV accidents occurred in rollovercrashes, while rollovers accounted for only 23%

Research Paper

493

Int. J. Mech. Eng. & Rob. Res. 2014 K K Dhande et al., 2014

of total fatalities in car accidents. For thepurposes of a dynamic rollover resistancerating test, NHTSA selected the Fishhook andSlalom steering manoeuvre as a primarycandidate.

The rollover propensity of a vehicle isdetermined from the highest speed for whichit can complete the selected manoeuvrewithout undergoing two-wheel lift. Since thevehicle testing is conducted on-road, theresults are more repeatable and give morecontrol over the test environment than do off-road tripped tests (Forkenbrock et al., 2003).Even though the evaluation procedure is onlymeant to test vehicles for on-road, untrippedrollover propensity (which accounts for a smallpercentage of rollover crashes), it is believedthat the results are still a valuable measure ofoverall rollover stability for relative comparisonof various vehicles (Viano and Chantal, 2003).

It is well known that steering performancedepend on steering ratio and Ackermanpercentage. These are the major vehicleparameters that contribute to rolloverpropensity. In this study, the effects of steeringperformance is evaluated to determine theirinfluence on rollover propensity, while holdingother vehicle properties such as suspensionconfiguration and weight distribution constant.A detailed description of the basic simulationmodel, including validation testing, and adiscussion of the results for several parametricvariation studies are provided in the followingsections.

SIMULATION MODELDynamic simulation has proven to be anefficient and accurate method for analyzingvehicles and evaluating their dynamic

behaviour (Allen et al., 1990; and Garrott andHeydinger, 1992). As a part of the presenteffort, a simplified model that captures theessential vehicle dynamics associated withun-tripped rollover was developed andvalidated.

MODEL DEVELOPMENTThe primary dynamics that are of concern arethose associated with the yaw and roll motions.And yaw and roll can be easily evaluated onthe bases of dynamic testing of the vehicle.Simplified vehicle model was developed in IPGCar Maker (multi-body dynamic software)along with its all sub systems. Simulink wasused for developing the vehicle control system.

MODEL VALIDATIONIn order to validate the vehicle model describedabove, simulation results were compared withphysical testing and experimental data for theSlalom manoeuvre. The Slalom manoeuvreuses a steering input such that driver has tonegotiate the specified course at a set entrancevelocity. The velocity profile of this manoeuvreis characterized by the vehicle reaching adesired steady state speed, known as theentrance speed, and coasting through the restof the manoeuvre once the initial pylon isbegun. The profile consists of an initial zerosteer angle up to the target speed followed bygoing to a steer angle at a rate such thatvehicle can negotiate the slalom coursewithout any wheel lifts or without hitting any ofthe pylon.

In the Slalom manoeuver steeringperformance have more control on vehiclebehaviour. If there is any unexpected behaviourof the vehicle during testing one can tune the

494

Int. J. Mech. Eng. & Rob. Res. 2014 K K Dhande et al., 2014

various vehicle parameters to achieve goodresults in virtual world followed by real world.

The simulation model contains a ‘virtualgarage’ that consists of ‘virtual vehicles,’ eachwith a configuration to match those ofexperimental vehicle used in the physicaldynamic testing. Parameters in the virtualgarage, such as wheelbase, track width, rollstiffness, roll damping, roll centre location,

brake characteristics are held constant. Thevirtual vehicles varied from one another byhaving different steering performance and totalweight.

Below Figures graphically compare theactual testing data with corresponding resultsfrom the simulation model. The inputs to thesimulation are the steer angle, gear step,master cylinder brake pressure and velocity

Figure 1: Slalom Test Steer Angle Graph

Figure 2: Slalom Test Longitudinal Speed Graph

495

Int. J. Mech. Eng. & Rob. Res. 2014 K K Dhande et al., 2014

data from the actual testing. The statescompared are lateral acceleration, yaw rate,and roll angle.

SIMULATION STUDIESOnce the simulation model was validated,vehicle parameters were varied in order toassess their effect on rollover propensity. Thedynamic test used to determine the effects ofthese properties is the Slalom manoeuvre. The

Slalom is a highly repeatable manoeuvre ascited by NHTSA’s research. For eachparametric variation, a new Slalom constanthad to be determined. This was accomplishedby using the simulation to find the steer angleat maximum speed through which the vehiclenegotiates the Slalom course. Rollover velocityis calculated by determining the vehicle speedat which a steering manoeuvre causes “twowheel lift or vehicle hitting the pylon.” “Two

Figure 3: Slalom Test Lateral Acceleration Graph

Figure 4: Slalom Test Yaw Angle Graph

496

Int. J. Mech. Eng. & Rob. Res. 2014 K K Dhande et al., 2014

wheel lift” is defined to occur when the normalforces on both inside tires drops to zero. Bycomparing the rollover velocities with thevarious vehicle properties, a correlation isidentified between rollover propensity andthose properties.

From the above graphs it was clear that thevehicle is skidding at after last slalom haspassed, and this was not acceptableaccording to the standard protocol. To avoid

this hazards behaviour of the vehicle it ismandatory to tuned the vehicle parametersfrom Vehicle Dynamic Controller (VDC). Fromthe past experience I have modified the VDCparameter called LwGr. LwGr is the thresholdfor the steering angle front axle. LwGr wastuned from 0.1 to 0.02 in order to achieve asexpected behaviour.

For any of the vehicle tuning process it hasto be checked that whether the results from

Figure 5: Slalom Test Steer Angle Graph After Simulation

Figure 6: Slalom Test Longitudinal Speed Graph After Simulation

497

Int. J. Mech. Eng. & Rob. Res. 2014 K K Dhande et al., 2014

Figure 7: Slalom Test Lateral Acceleration Graph After Simulation

Figure 8: Slalom Test Yaw Angle Graph After Simulation

actual testing are matching with the resultsafter simulation, once this correlation has beenverified then go for the parameter tuning. FromFigures 5 and 6, it was clear that the correlationresults are matching.

CONCLUSIONThis study has shown that a simple vehiclemodel developed to study the transientdynamics of SUV can capture the dynamics

seen in events leading to rollover. This wasdemonstrated with a correlation betweensimulation results and field test data; thereforegiving validation to using the simulation tostudy rollover propensity.

To improve vehicle performance and itsstability the control system plays major role.So by tuning the vehicle control system in thiscase VDC, vehicle stability was improved. Andthe unexpected behaviour what we were facing

498

Int. J. Mech. Eng. & Rob. Res. 2014 K K Dhande et al., 2014

at the last slalom during testing has beenavoided at some acceptable range. Futureresearch will investigate the effect of othervehicle properties, such as suspension setup,tire models, etc., on rollover propensity usingsimulation and scaled experiments.

FUTURE WORKAs rollover is an unwanted phenomenon, oncethe roll angle is calculated, some correctiontechniques may be developed, like steeringcorrection required or braking torque requiredbringing the vehicle back from roll condition.Also a roll sensor which will work constantlywith a controller taking steering input canpredict the approach of serious rolloversituation or may be indicate the driver aboutcornering speed and/or steering rate for safecornering. Suspension is key element; rigidsuspension would cause discomfort and softsuspension increases roll moment. Hence useof active or semi active suspensions canprove landmark in avoiding rollovers. All thischanges can be implemented in the vehicleconjunction with vehicle control system likeElectronic Steering Performance (ESP), VDC,Anti-lock Brake System (ABS), Roll StabilityControl (RSC), etc.

REFERENCES1. Allen R W, Szostak H T, Rosenthal T J and

Klyde D H (1990), “Field Testing andComputer Simulation Analysis of GroundVehicle Dynamic Stability”, SAE PaperNo. 900127.

2. Christos J P and Guenther D A (1992),“The Measurement of Static RolloverMetrics”, SAE Paper No. 920582.

3. Cooperrider N K, Hammound S A andThomas T M (1990), “Testing and Analysisof Vehicle Rollover Behavior”, SAE PaperNo. 900366.

4. D’Entremont K L, Zhengyu L and NaleczA G (1993), “An Investigation into DynamicMeasures of Vehicle Rollover Propensity”,SAE Paper No. 930831.

5. Forkenbrock G J, Garrott W R, Heitz Mand O’Hara B C (2003), “An ExperimentalExamination of J-Turn and FishhookManeuvers that May Induce On-Road,Untripped, Light Vehicle Rollover”, SAEPaper No. 2003-01-1008.

6. Garrott W R and Heydinger G J (1992),“An Investigation, Via Simulation, ofVehicle Characteristics that Contribute toSteering Maneuver Induced Rollover”,SAE Paper No. 920585.

7. Gillespie T D (1992), “Fundamentals ofVehicle Dynamics”, Society of AutomotiveEngineers, Warrendale, PA, ISBN: 1-56091-199-9.

8. Ponticel P (2003), “Dynamic TestingRollover on the Way”, AutomotiveEngineering International, November,pp. 26-28.

9. Viano D C and Chantal P (2003), “CaseStudy of Vehicle Maneuvers Leading toRollovers: Need for a Vehicle TestSimulating Off-Road Excursions,Recovery and Handling”, SAE Paper No.2003-01-0169.