This article was downloaded by: [Canadian Research Knowledge Network] On: 7 March 2010 Access details: Access

Details: [subscription number 918588849] Publisher Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 3741 Mortimer Street, London W1T 3JH, UK

Vehicle System Dynamics

Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713659010

Investigation into untripped rollover of light vehicles in the modified fishhook and the sine maneuvers. Part I: Vehicle modelling, roll and yaw instability
Nong Zhang a; Guang-Ming Dong b; Hai-Ping Du a a Mechatronics and Intelligent Systems, University of Technology, Sydney, Australia b State Key Laboratory of Vibration, Shock&Noise, Shanghai Jiao Tong University, Shanghai, P.R. China

To cite this Article Zhang, Nong, Dong, Guang-Ming and Du, Hai-Ping(2008) 'Investigation into untripped rollover of light

vehicles in the modified fishhook and the sine maneuvers. Part I: Vehicle modelling, roll and yaw instability', Vehicle System Dynamics, 46: 4, 271 293 To link to this Article: DOI: 10.1080/00423110701344752 URL: http://dx.doi.org/10.1080/00423110701344752

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

Vehicle System Dynamics Vol. 46, No. 4, April 2008, 271293

Investigation into untripped rollover of light vehicles in the modied shhook and the sine maneuvers. Part I: Vehicle modelling, roll and yaw instability
NONG ZHANG*, GUANG-MING DONG and HAI-PING DU
Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

Mechatronics and Intelligent Systems, University of Technology, Sydney, P.O. Box 123, Broadway, NSW 2007, Australia State Key Laboratory of Vibration, Shock&Noise, Shanghai Jiao Tong University, Shanghai, P.R. China
Vehicle rollovers may occur under steering-only maneuvers because of roll or yaw instability. In this paper, the modied shhook and the sine maneuvers are used to investigate a vehicles rollover resistance capability through simulation. A 9-degrees of freedom (DOF) vehicle model is rst developed for the investigation. The vehicle model includes the roll, yaw, pitch, and bounce modes and passive independent suspensions. It is veried with the existing 3-DOF rollyaw model. A rollover critical factor (RCF) quantifying a vehicles rollover resistance capability is then constructed based on the static stability factor (SSF) and taking into account the inuence of other key dynamic factors. Simulation results show that the vehicle with certain parameters will rollover during the shhook maneuver because of roll instability; however, the vehicle with increased suspension stiffness, which does not rollover during the shhook maneuver, may exceed its rollover resistance limit because of yaw instability during the sine maneuver. Typically, rollover in the sine maneuver happens after several cycles. It has been found that the proposed RCF well quanties the rollover resistance capability of a vehicle for the two specied maneuvers. In general, the larger the RCF, the more kinetically stable is a vehicle. A vehicle becomes unstable when its RCF is less than zero. Detailed discussion on the effects of key vehicle system parameters and drive conditions on the RCF in the shhook and the sine maneuver is presented in Part II of this study. Keywords: Untripped vehicle rollover; Light vehicle; Vehicle model; Roll instability; Yaw instability

Nomenclature u0 v zs zui 1, 2, 3, 4 S Sv Constant longitudinal velocity of the sprung mass centre of gravity (CG) Lateral velocity of the sprung mass CG Vertical displacement of the sprung mass CG Vertical displacement of the unsprung masses i = 1, 2, 3, 4 Left-front, Right-front, Left-rear, and Right-rear, respectively The earth-xed inertial reference frame The vehicle-xed non-inertial reference frame Roll angle of the sprung mass around x-axis of Sv Pitch angle of the sprung mass around y-axis of Sv

*Corresponding author. Email: [email protected]

Vehicle System Dynamics ISSN 0042-3114 print/ISSN 1744-5159 online 2008 Taylor & Francis http://www.informaworld.com DOI: 10.1080/00423110701344752

272

N. Zhang et al.

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

p q r ms mui (Ixx )s (Iyy )s (Izz )s (Ixz )s (Izz )ui a b tf tr hs hra Ksi Csi zsui Kt zgi f c r c f r

Yaw angle of the vehicle around z-axis of Sv or S Roll rate of the sprung mass around x axis of Sv Pitch rate of the sprung mass around y axis of Sv Yaw rate of the vehicle around z axis of Sv or S Vehicle sprung mass Vehicle unsprung mass Moment of inertia about the x-axis of the vehicle sprung mass Moment of inertia about the y-axis of the vehicle sprung mass Moment of inertia about the z-axis of the vehicle sprung mass Product of inertia about the xz axes of the vehicle sprung mass Moment of inertia about the z-axis of the vehicle unsprung mass Distance from the sprung mass CG to the front axle Distance from the sprung mass CG to the rear axle 1/2 width of the front axle 1/2 width of the rear axle Height of the sprung mass CG above the roll axis Height of the roll axis above the ground Spring stiffness for each suspension i = 1, 2, 3, 4 Damping coefcient for each suspension i = 1, 2, 3, 4 Change of spring length for each suspension i = 1, 2, 3, 4 Vertical tyre stiffness Road disturbance input for each tyre i = 1, 2, 3, 4 Front tyre steer angle Partial derivative of the roll induced steer at the rear axle Roll induced steer angle at rear Inclination angle coefcient Inclination angle Front tyre slip angle Rear tyre slip angle

1.

Introduction

Rollover accidents are dangerous events. According to the National Highway Trafc Safety Administration (NHTSA) of USA [1], although only 8% of light vehicles (passenger cars, pick-ups, vans and sport utility vehicles) in crashes roll over, 21% of seriously injured occupants and 31% of occupant fatalities are involved in rollovers. Rollover accidents can be divided into off-road and on-road rollover, where on-road rollover can be further subdivided into tripped and untripped rollover depending on the mechanism that initiated the on-road rollover. Rollover crash data show that approximately two-thirds of on-road rollovers are untripped [2]. Rollover incidents involve a variety of factors and some of which may be beyond the control of vehicle designers. The on-road untripped rollovers are directly connected with vehicle handling behaviour that is mainly inuenced by vehicle design variables, thus researches on this subject for light vehicles are highly demanding. Experimental evaluations of certain maneuvers on selected vehicle types have been made and published by NHTSA [24], which tried to quantify on-road, untripped rollover propensity of light vehicles and incorporate the test results into the New Car Assessment Program. Baumann and Eckstein [5] performed extensive eld testing and used the recorded wheel inputs in their simulations, and they discovered

Vehicle modelling, roll and yaw instability

273

that the main conditions necessary for untripped rollover were to create high roll rates and maximum side forces at both axles simultaneously. Experimental investigation of vehicle rollover crashes are not only expensive and timeconsuming but also limited to those maneuvers that can be physically reconstructed. Therefore, more attention has been paid to the investigations into rollover crashes based on computer simulations. Many researchers [615] generally studied the effects of vehicle parameters and driving conditions on stability and handling, which is helpful in the preliminary phases of vehicle design. Others [1620] researched on the control of untripped vehicle rollovers by adding electronic equipment, which is beyond the main concern of this paper. In general, vehicle models used by researchers for rollover study can be classied into four categories. Brief descriptions are given for each type of model: (1) yaw-plane model [68], which includes all the axles of the vehicle but no roll degree of freedom (DOF) and is often used in vehicle yaw stability analysis; (2) roll-plane model [711], which consists of a single axle with associated suspensions and body mass free to roll only. This model is often used in heavy vehicle rollover analysis, tripped rollover analysis, or simple static analysis for untripped rollover; (3) roll yaw model [12, 13, 1618], which was initially proposed by Segel [21] , includes both roll and yaw dynamics and is often used by researchers for vehicle rollover analysis and control for its simplicity and connectivity between wheel input and vehicles rollyaw responses; (4) full-car models used in [10, 14, 15, 19, 20], which include all of the bounce, pitch, roll, and yaw dynamics, are often used in commercial vehicle dynamics programs. The mathematical models used in these full-car commercial programs are not publicly and easily accessible. This makes it difcult for researchers to conduct detailed parametric study and in-depth investigation. In this paper, a 9-DOF full-car model is developed. The model includes the roll, yaw, pitch, and bounce modes of vehicle body and passive independent suspensions. This full-car model is constructed under the assumption of constant forward velocity for simplicity and is discussed in detail in section 2. To assess the on-road, untripped rollover propensity of a vehicle, NHTSA uses a set of rollover evaluation maneuvers and examines the severity and frequency of vehicle two-wheel lift (TWL) under these maneuvers [2]. In the simulation of this paper, two types of maneuvers of a light vehicle on a at dry road surface with various parameters are investigated. One is the modied shhook maneuver and the other is the sine maneuver for multiple cycles with constant frequency. NHTSA [2] described their type-1a shhook maneuver in an experiment as follows: To begin this maneuver, the vehicle was driven in a straight line at a speed slightly greater than the desired entrance speed. The driver released the throttle, and when at target speed, initiated the hand wheel commands described in gure 1. Following completion of the counter steer, hand wheel position was maintained for 3 s. The hand wheel was then returned to zero. Clearly the vehicle speed will decrease to some extent during the maneuver because of tyrerolling resistance, scrub during turning, and no supplied drive torque. However, for simplicity, constant speed in the shhook maneuver is assumed in this paper to obtain some aggravated simulation results from a viewpoint of safety for real applications. The second modication on the maneuver is made to the steering amplitude A, which is varied for the purpose of parametric study. The sine maneuver for multiple cycles with constant frequency shown in gure 2 is modied from the single sine steer maneuver and sine sweep steer maneuver used in [18]. In the single sine maneuver, vehicle is run for one cycle at resonant yaw rate frequency, which may neglect something as cycle number increases; in the sine sweep maneuver, results can only be analysed in frequency domain. In the simulations on the sine maneuver presented in this paper, the vehicle is run under the sine maneuver for multiple cycles with constant frequency ranging from 0.11.1 Hz, from which results can be evaluated both in time domain

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

274

N. Zhang et al.

A steering amplitude [degree]

T1

A = steering amplitude T = commanded dwell time

1

Initial steer and conter steer performed at 720 deg/seg 0

-A
Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

0.5

1.5 time [s]

2.5

Figure 1.

Fishhook maneuver description.

Figure 2.

Description for sine maneuver.

for considering multi-cycle effect and in frequency domain for enhanced resolution. Furthermore, it is somewhat like the slalom vehicle test, which aims to evaluate the vehicle handling characteristics. The presentation of this paper is organized as follows: section 2 provides the derivation detail of a 9-DOF vehicle model, which includes the roll, yaw, pitch, and bounce modes and passive independent suspensions. In section 3, the proposed vehicle model is veried with the existing 3-DOF rollyaw model using the same vehicle data and steering inputs. A rollover critical factor is constructed for assessing the rollover resistance capability of a light vehicle. In section 4, vehicle rollovers due to roll instability in the shhook maneuver and yaw stability in the sine maneuver are investigated. Conclusion drawn from Part I of this research is given in section 5.

2. Vehicle model description The vehicle model shown in gure 3 consists of a rigid sprung mass supported by four independent suspensions, which includes wheel assemblies as four unsprung masses. Due to the constraints between the rigid bodies, the unsprung masses translate laterally together with the sprung mass, which means that the unsprung masses have no independent lateral DOF compared with the sprung mass. In yaw rotation, the unsprung masses rotate together with the sprung mass relative to the earth-xed inertial reference; however, the roll and pitch rotations are restricted only to the sprung mass. Under the assumption of a constant forward speed and the above constraints, the number of DOF of the vehicle model is set to nine: sprung mass lateral and vertical centre-of-mass

Vehicle modelling, roll and yaw instability

275

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

Figure 3.

Nine-DOF vehicle model.

motions, roll and pitch rotations of the sprung mass about the axis at its CG, yaw rotation for the total vehicle, and four centre-of-mass vertical motions of the unsprung masses. Denote with S the earth-xed inertial reference frame and with Sv a vehicle xed noninertial reference frame rotating with the angular velocity v = [0 0 r]T and translating with the velocity vo = [u0 v 0]T . Equations of motion of the 9-DOF vehicle model are derived as follows, under the assumption of the xed roll axis concept. 2.1 Kinematics

2.1.1 Translation of CG of sprung mass. The position of CG of sprung mass in the reference coordinate Sv is: 0 hs cos sin 0 hs sin (1) los = Ry ( )Rx () 0 + 0 = zs hs cos cos + zs hs where 1 0 0 Rx () = 0 cos sin 0 sin cos cos 0 sin 1 0 Ry ( ) = 0 sin Then: d dt

(2)

(3)

cos

los =

d dt

los + v los
v

(4)

where denotes the cross-product operator. The velocity of vehicle sprung mass CG point with respect to S is expressed as the sum of the translational velocity vo of the vehicle reference

276

N. Zhang et al.

system Sv and the time derivative of the position vector los : vs = vo + d dt los = vo + d dt los + v ros = v o + los + v los
v

(5)

The acceleration of vehicle sprung mass CG point with respect to S is computed as: as = d dt vs = d dt
vs + v vs = vs + v vs v

(6)

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

2.1.2 Translations of CGs of unsprung masses. The positions of CGs of unsprung masses in the reference coordinate Sv are: lou1 = , hu1 + zu1 b tr lou3 = , hu3 + zu3 a tf lou2 = , hu2 + zu2 b tr lou4 = hu4 + zu4 a tf

(7)

Here, we denote with 1 the left-front unsprung mass, 2 the right-front, 3 the left-rear, and 4 the right-rear unsprung mass. Then the velocities and accelerations of vehicle unsprung mass CG points with respect to S , vui and aui , i = 1, 2, 3, 4, can be obtained using equations (5) and (6). 2.1.3 Rotations of sprung mass and unsprung masses. The angular momentum of the sprung mass can be dened as: H s = ms ls los vs + Is s = (HCG )s + (H )s (8)

where (HCG )s is the angular momentum of the sprung mass due to the motion of the CG, (H )s is the angular momentum of the mass due to its rotation about the CG, and s = [p q r]T . The angular momentum of unsprung mass can be dened as:
4

Hu =
i=1

[mui (roui vui ) + Iui u ]

(9)

As unsprung masses only have yaw rotation, then Hu = 0 0 Huz

T

(10)

The angular momentum of total vehicle can be written as: H = Hs + Hu (11)

Vehicle modelling, roll and yaw instability

277

2.2 2.2.1

Kinetics Lateral equation for the entire vehicle mass.

4

Fty = ms asy +
i=1

mui auyi

(12) (13)

Fty = (Fy1 + Fy2 ) cos(f ) + (Fy3 + Fy4 ) cos(r ) where Fyi , i = 1, 2, 3, 4, is the tyre lateral force. 2.2.2 Vertical equation for the sprung mass.
4

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

i=1

Fsi = ms asz

(14)

where Fsi is the suspension force, we denote with 1 the left-front suspension, 2 the right-front, 3 the left-rear, and 4 the right-rear suspension. Fsi can be computed as: Fsi = Ksi zsui + Csi zsui i = 1, 2, 3, 4 (15)

where Ksi is the spring stiffness, Csi the damping coefcient, and zsui the change of spring length for each suspension. Zsui can be computed as: left-front : zsu1 = (zs tf a ) zu1 right-front : zsu2 = (zs tf a ) zu2 left-rear : zsu3 = (zs tr + b ) zu3 right-rear : zsu4 = (zs + tf + b) zu4 (16) (17) (18) (19)

In an effort to maintain generalized applicability to different types of suspension, suspension forces can be modied to have their equivalent forces acting at the track width (i.e., at each wheel) as above. 2.2.3 Vertical equations for the unsprung masses. Fsi Kt (zui zgi ) = mu zui i = 1, 2, 3, 4 (20)

where Kt is the vertical tyre stiffness and zgi the road disturbance input for each tyre. 2.2.4 Roll equation for the sprung mass. Mx = Hx rHy = (Fs1 Fs2 )tf + (Fs3 Fs4 )tr + ms ghs 2.2.5 Pitch equation for the sprung mass. My = Hy + rHx = (Fs1 Fs2 )a (Fs3 + Fs4 )b + ms ghs (22) (21)

278

N. Zhang et al.

2.2.6 Yaw equation for the entire vehicle mass. Mz = Hz = a(Fy1 + Fy2 ) cos(f ) b(Fy3 + Fy4 ) cos(r ) + tf (Fy2 Fy1 ) sin(f ) + tr (Fy4 Fy3 ) sin(r ) + Mz1 + Mz2 + Mz3 + Mz4 where Mzi , i = 1, 2, 3, 4 is the tyre self-aligning moment. 2.3 Tyre force calculation (23)

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

Magic-Formula (MF) tyre models are considered the state-of-the-art for modeling tyreroad interaction forces and moments in vehicle dynamics. Since 1987, Pacejka and others have published several versions of this type of tyre model. The latest version of MF published by Pacejka [22] is used in this paper. The input for the MF tyre model consists of the wheel load (Fz ), the longitudinal and lateral slip (, ), and inclination angle ( ). The output are the forces (Fx , Fy ) and self-aligning moment Mz in the contact point between the tyre and the road. To calculate these forces and moments, the MF equations use a set of MF parameters, which are derived from tyre testing data. For pure slip conditions, the MF equations can be described as: Y (x) = D sin[C arctan{Bx E(Bx arctan(Bx))}] (24)

where Y (x) is either Fx with x the longitudinal slip or Fy with x the lateral slip . The coefcient B, C, D, and E depend on the wheel load Fz and the inclination angle . The self-aligning moment Mz is due to the offset of lateral force Fy , called pneumatic trail t, from the contact point. A cosine version the MF equation is used to calculate the pneumatic trail t: Y (x) = D cos[C arctan{Bx E(Bx arctan(Bx))}] (25) in which Y (x) is the pneumatic trail t as a function of lateral slip angle . In combined slip conditions, the lateral force Fy will decrease due to the longitudinal slip , and the longitudinal force Fx will decrease due to the lateral slip . The forces and moments in the combined slip conditions are based on the pure slip characteristics multiplied by some weighing functions. The front and rear tyre slip angle can be computed as follows from gure 4: front tyre slip angle : f = v + ar f u0 v br r rear tyre slip angle : r = u0 (26) (27)

where f is the front tyre steer angle and r = , C is the roll-induced steer angle at the rear tyre, C is determined by geometric properties of the rear suspension. The tyre longitudinal slip is set to zero because of the constant vehicle longitudinal velocity assumption, and the inclination angle is computed as the multiplication of roll angle by a coefcient. A set of scaling factors is available in the MF tyre model to examine the inuence of changing tyre properties without the need to change one of the real MF coefcients. F z0 is the scaling factor for tyre nominal load. As can be seen in gure 5, tyre lateral stiffness and maximum tyre lateral force will increase for a given tyre normal load if Fz0 is increased.

Vehicle modelling, roll and yaw instability

279

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

Figure 4. Top view of the model.

4000 Tyre Lateral Force (N)

Fz = 4000N =0 increase of

2000

Fz0

0
increase of

2000

Fz0

4000 20 10 0 10 Tyre Slip Angle (Degree) 20

Figure 5.

Inuence of scaling factor Fz0 on tyre lateral force.

3.

Model comparison and description of rollover critical factor

A light passenger vehicle, of which parameters are listed in Appendix A, is used to conduct the simulation of 9-DOF vehicle model. The 3-DOF rollyaw model presented by Segel [21] is used for cross-checking the accuracy of the 9-DOF model. For both the 3- and 9-DOF models, the MF tyre model is used. Transformed vehicle parameters from 9- to 3-DOF model are also listed in Appendix A. The two maneuvers described in section 1 are used for the comparison simulations between the existing 3- and 9-DOF models proposed in this paper. Common parameters used for the two models are: front and rear tyre 195/65 R15 car tyre; constant vehicle longitudinal velocity u0 = 75 km/h; steer ratio from hand wheel to road tyre is 20; maximum wheel angle for two maneuvers is 270 ; steering frequency for sine steer is 0.3 Hz.

280

N. Zhang et al.

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

Figure 6.

Comparison of vehicle lateral acceleration between the two vehicle models under shhook maneuver.

The vehicle lateral acceleration, roll angle, roll rate, and yaw rate for the shhook maneuver are obtained from the two vehicle models, respectively, and are shown in gures 69 for comparison. Differences of vehicle lateral accelerations and yaw rates between the two models are small. However, roll angle and roll rate of the 9-DOF model are obviously larger than that of the 3-DOF model. By analysing the suspension force equations (15)(19) and roll equation (21), it can be deduced that the differences between vertical displacements of the unsprung masses zui , i = 1, 2, 3, 4, cause the roll angle and roll rate differences between the two models. If zui , i = 1, 2, 3, 4, are set to zero in the 9-DOF model, the differences become very small, as shown in gures 10 and 11. The vehicle lateral acceleration, roll angle, roll rate, and yaw rate for the sine maneuver are obtained from the two vehicle models, respectively, and are shown in gures 1215 for comparison. Differences of vehicle lateral acceleration and yaw rate between the two models are small. However, roll angle and roll rate of the 9-DOF model are obviously larger than that of the 3-DOF model, as shown in gures 13 and 14. If zui , i = 1, 2, 3, 4, are set to zero
6 4 Roll Angle (Degree) 2 0 2 4 6 0 2 4 6 Time(s); [maxmum error: 10.4218%] 8 9d of vehicle model 3d of vehicle model

Figure 7.

Comparison of roll angle between the two vehicle models under shhook maneuver.

Vehicle modelling, roll and yaw instability

281

20 10 0 10 20 30 0 2 4

9d of vehicle model 3d of vehicle model

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

Roll Rate (Degree/s)

Time(s); [maxmum error: 13.07%]

Figure 8. Comparison of roll rate between the two vehicle models under shhook maneuver.

in the 9-DOF model, the roll angle and roll rate differences become very small, as shown in gures 16 and 17. To evaluate a vehicles rollover resistance capability, static stability factor (SSF) is a basic indicator, which has been incorporated into NHTSAs rollover rating program [23]. A vehicles SSF is calculated using the formula SSF = T/2H, where T is the track width of the vehicle and H the height of the CG of the vehicle. The lower the SSF number, the more likely the vehicle is to roll over. In SSF, vehicle lateral acceleration level represents the only, or the rst order, characteristics for vehicle rollover stability. Hac [8] suggested many modications to SSF. These include the effects of lateral movement of vehicle CG during body roll, suspension jacking forces, changes in track width due to suspension kinematics, tyre lateral compliance, tyre gyroscopic forces, and dynamic overshoot in the roll angle. In [24], Dynamic Stability Index (DSI) is proposed and discussed. The DSI is based on the SSF model with the addition of the roll moment of
40 30 Yaw Rate (Degree/s) 20 10 0 10 20 30 0 9d of vehicle model 3d of vehicle model 2 4 6 Time(s); [maxmum error: 5.6684%] 8

Figure 9.

Comparison of yaw rate between the two vehicle models under shhook maneuver.

282
6 4 Roll Angle (Degree) 2 0 2 4 6
Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

N. Zhang et al.

9d of vehicle model 3d of vehicle model

Time(s); [maxmum error: 1.028%]

Figure 10.

Comparison of roll angle between the two vehicle models under shhook maneuver.
20 9d of vehicle model 3d of vehicle model 10 Roll Rate (Degree/s)

10

20

30 0 2 4 6 Time(s); [maxmum error: 1.2185%] 8

Figure 11.

Comparison of roll rate between the two vehicle models under shhook maneuver.
Lateral acceleration in global coordinates (m/s2)

10

9d of vehicle model 3d of vehicle model

10

2 4 6 8 Time(s); [maxmum error: 0.14776%]

10

Figure 12.

Comparison of vehicle lateral acceleration between the two vehicle models under sine maneuver.

Vehicle modelling, roll and yaw instability

283

10 8 6 Roll Angle (Degree) 4 2 0 2 4

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

9d of vehicle model 3d of vehicle model

6 0 2 4 6 8 Time(s); [maxmum error: 10.2145%] 10

Figure 13.

Comparison of roll angle between the two vehicle models under sine maneuver.

the sprung mass (Ixx )s , where denotes the roll acceleration. In addition to DSI, vehicle half-track width is modied to account for the lateral CG offset of a vehicle. The rollover critical factor (RCF) that quanties the rollover resistance capability of a light vehicle is dened as follows: Factor = g (tf + tr ) hs || |ay |[(hs + hra ) zs ] (Ixx )s 2 ms (28)

where ay is vehicle lateral acceleration and g the gravity acceleration. RCF accounts for the decrease of vehicle half-track width due to roll angle, the increase of vehicle CG height due to the CG vertical displacement, and the inuence of roll moment of the sprung mass. As described by equation (28), smaller RCF indicates low rollover resistance capability of a vehicle and rollover will occur when RCF becomes negative.
40 30 Roll Rate (Degree/s) 20 10 0 10 20 30

9d of vehicle model 3d of vehicle model

2 4 6 8 Time(s); [maxmum error: 5.2144%]

10

Figure 14.

Comparison of roll rate between the two vehicle models under sine maneuver.

284
50 40 Yaw Rate (Degree/s) 30 20 10 0 10 20 30
Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

N. Zhang et al.

9d of vehicle model 3d of vehicle model

10

Time(s); [maxmum error: 2.745%]

Figure 15.

Comparison of yaw rate between the two vehicle models under sine maneuver.

8 6 Roll Angle (Degree) 4 2 0 2 4 6 0

9d of vehicle model 3d of vehicle model

2 4 6 8 Time(s); [maxmum error: 0.048058%]

10

Figure 16.

Comparison of roll angle between the two vehicle models under sine maneuver.

40 30 Roll Rate (Degree/s) 20 10 0 10 20 30 0 2 4 6

9d of vehicle model 3d of vehicle model

10

Time(s); [maxmum error: 3.2592%]

Figure 17.

Comparison of roll rate between the two vehicle models under sine maneuver.

Vehicle modelling, roll and yaw instability

285

4. Vehicle rollover due to roll instability or yaw instability In some cases, vehicle rollovers take place due to roll instability. This instability is demonstrated in the shhook maneuver simulation, of which the vehicle velocity is set to 75 km/h and the maximum steering angle is specied as 270 . Three sets of vehicle parameters are used in the simulation: in addition to the normal vehicle parameters in Appendix A, suspension stiffness is reduced to 60% of its original value for the second simulation case and increased to 140% of its original value for the third case. To enhance the comparison effect, the sprung mass height hs is increased to 0.4 m with hra reduced to 0.2 m for a constant SSF value in the decreased stiffness case. The RCF for the three cases under the shhook maneuver are computed and shown in gure 18. For the reduced suspension stiffness case, the RCF value is obviously smaller than that of the normal and increased stiffness cases. Figures 1922 show that the major effect

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

7 6 Rollover Critical Factor 5 4 3 2 1 0 0 1 2 3

Ks Decrease Normal Ks Increase

Time (s)
Figure 18. Comparison of RCF for different suspension stiffness under shhook maneuver.

10 Vehicle Lateral Acceleration (m/s2) Ks Decrease Normal Ks Increase 5

-5

-10

3 Time (s)

Figure 19.

Comparison of vehicle lateral acceleration for different suspension stiffness under shhook maneuver.

286
0.015 CG Vertical Displacement (m) 0.01 0.005 0 0.005 0.01 0.015

N. Zhang et al.

Ks Decrease Normal Ks Increase

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

3 Time (s)

Figure 20.

Comparison of CG vertical displacement for different suspension stiffness under shhook maneuver.

15 10 Roll Angle (Degree) 5 0 5 10 15 0 1 2 3 4 Time (s) 5 Ks Decrease Normal Ks Increase 6 7

Figure 21.

Comparison of roll angle for different suspension stiffness under shhook maneuver.

200 150 Roll Acceleration (Degree/s2) 100 50 0 50 100 150 200 0 1 2 3 4 Time (s) 5 6 7 Ks Decrease Normal Ks Increase

Figure 22.

Comparison of roll acceleration for different suspension stiffness under shhook maneuver.

Vehicle modelling, roll and yaw instability

287

30 20 Yaw Rate (Degree/s) 10 0 10 20 30

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

Ks Decrease Normal Ks Increase

3 4 Time (s)

Figure 23.

Comparison of yaw rate for different suspension stiffness under shhook maneuver.
8 6 RCF 4 2 0 5 10 15 Time (s) 20 25

Min per cycle

0.5

0 0.2

6 Cycles

10

Figure 24.

Comparison of RCF between the increased and the normal suspension stiffness under sine maneuver.
Acceleration (m/s2) 10

10

10

15 Time (s)

20

Max per Cycle

9.8 9.7 9.6 2 4 6 Cycles 8 10

Figure 25. Comparison of vehicle lateral acceleration between the increased and the normal suspension stiffness under sine maneuver.

288

N. Zhang et al.

2 Zs (mm) 1 0 1 5 1 Min per Cycle 1.2 1.4 1.6 2 4 6 Cycles 8 10 10 15 Time (s) 20

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

Figure 26. Comparison of CG vertical displacement between the increased and the normal suspension stiffness under sine maneuver.

Roll Angle (Deg)

5 0 -5 5 10 15 Time (s) 20 25

Max per Cycle

7 6 5 2 4 6 Cycles 8 10

Figure 27.

Comparison of roll angle between the increased and the normal suspension stiffness under sine maneuver.

Roll Acceleration

200 0 200 5 10 15 Time (s) 20

Max per Cycle

200 160

2
Figure 28. maneuver.

6 Cycles

10

Comparison of roll acceleration between the increased and the normal suspension stiffness under sine

Vehicle modelling, roll and yaw instability

289

60 40 Yaw Rate (Degree/s)

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

20 0 20 40 60 5 10 15 Time (s) 20

Figure 29.

Comparison of yaw rate between the increased and the normal suspension stiffness under sine maneuver.

on RCF comes from roll angle variation shown in gure 21; inuences of other factors are positive but insignicant. Figure 23 shows that the differences of vehicle yaw rates for these three cases are small. The results of the simulation case with increased suspension stiffness in the shhook maneuver show higher rollover resistance capability than that of the normal stiffness case. However, the same vehicle with increased suspension stiffness rolls over due to yaw instability in the sine maneuver simulation, in which the vehicle runs under the sinusoidal steering input with maximum angle 270 at a frequency of 0.46 Hz for 11 cycles, maintaining a constant vehicle velocity 75 km/h. In gures 2431, the results of the vehicle, with which the suspension stiffness is increased to 140% of its normal value, are represented in solid lines and that of the normal vehicle

Front Tyre Slip Angle ( __ ); Rear Tyre Slip Angle (---) Angle (Degree) 10 0 10 0 Angle (Degree) 5 10 15 20 25 Ks

10 0 10 0 5 10 15 Time (s) 20 25 1.4 Ks

Figure 30.

Comparison of tyre slip angle.

290

N. Zhang et al.

0.5 0.45 Time Lag per Cycle(s)

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

0.4 0.35 0.3 0.25 0.2 2 4 6 Time (s) 8 10

Figure 31.

Comparison of time lag for each cycle.

are in dashed lines. Figure 24 shows that the minimum RCF values per cycle for the vehicle with normal value remain constant and positive after the second cycle; however, the RCF for the vehicle with increased suspension stiffness decreases while the number of steering cycle increases and it becomes negative after the seventh cycle steering. Factors inuencing the RCF of a vehicle are studied, and the obtained results are given in gures 2527. For the vehicle with increased stiffness, its maximum lateral acceleration, highest vehicle CG vertical position, maximum roll angle, and roll acceleration per cycles all increase until certain cycles and then remain constant, which result in the reduced RCF. For the normal vehicle in dashed lines, these values generally remain unchanged after the second cycle. Figure 29 shows that the yaw rates for the vehicle with increased suspension stiffness have an obvious ascending trend with respect to time, and this indicates the vehicles yaw instability. The lower part of gure 30 shows that the rear tyre slip angle is larger than the front tyre slip angle after 15 s for the vehicle with increased stiffness, which indicates the over steering characteristics. At the same time, time lag per cycle between steer input and vehicle lateral acceleration becomes larger and then remains constant after a certain number of cycles, which is shown in gure 31. The above presented simulation results may partly explain the rollover accident reported in [25], which described that a sports utility vehicle suddenly rolled over (untripped) after having crossed several cones in a slalom test.

5.

Conclusions

To evaluate a vehicles rollover propensity through simulation, a 9-DOF vehicle model has been developed. The model includes the roll, yaw, pitch, and bounce modes of a vehicle body and passive independent suspensions. Model verication against the existing 3-DOF rollyaw vehicle model has been carried out, and the validity of the proposed 9-DOF vehicle model has been conrmed. The RCF has been constructed for evaluating a vehicles rollover propensity through simulation based on the SSF and taking into account the inuence of other key dynamic factors.

Vehicle modelling, roll and yaw instability

291

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

The RCF depends explicitly on those key dynamic state variables of a vehicle dynamic system such as the vertical displacement, the roll angle, the lateral and the roll angular accelerations of sprung mass. In comparison to the 3-DOF vehicle model, the proposed 9-DOF model provides the more accurate prediction of the outboard shift of the sprung mass CG, increase in CG height, and roll acceleration in extreme shhook and sine maneuvers, which affects the vehicle RCF signicantly. The obtained simulation results show that the shhook maneuver mainly tests the vehicles roll instability, whereas the sine maneuver evaluates the vehicles yaw instability. A vehicle with increased suspension stiffness can improve its rollover resistance capability in the shhook maneuver. However, rollover may occur to the same vehicle due to its increased yaw instability in the sine maneuver. It has been found that the proposed RCF well quanties the rollover resistance capability of a vehicle for the two specied maneuvers. The RCF takes into account both roll and yaw instabilities. In general, the larger the RCF, the more kinetically stable is a vehicle. A vehicle becomes unstable when its RCF is less than zero. Detailed discussion on the effects of key vehicle system parameters and drive conditions on the RCF is presented in Part II of this study. Acknowledgements Financial support for this research was provided jointly by the Australian Research Council (DP0560077) and the University of Technology, Sydney, Australia. References
[1] National Highway Trafc Safety Administration, 2003, Initiative to address the mitigation of vehicle rollover. Docket No. NHTSA-2003-14622-1. [2] Forkenbrock, G.J., Garrott, W.R., Heitz, M. and OHarra, B.C., 2002, A comprehensive experimental examination of test maneuvers that may induce on-road, untripped, light vehicle rollover - phase IV of NHTSAs light vehicle rollover research program. DOT HS 809 513, National Highway Trafc Safety Administration. [3] Forkenbrock, G.J., OHarra, B.C. and Elsasser, D., 2003, An experimental evaluation of 26 light vehicles using test maneuvers that may induce on-road, untripped rollover and a discussion of NHTSAs rened test procedures - phases VI and VII of NHTSAs light vehicle rollover research program. DOT HS 809 704, National Highway Trafc Safety Administration. [4] Forkenbrock, G.J., OHarra, B.C. and Elsasser, D., 2004, A demonstration of the dynamic tests developed for NHTSAs NCAP rollover rating system - phase VIII of NHTSAs light vehicle rollover research program. DOT HS 809 705, National Highway Trafc Safety Administration. [5] Baumann, F.W. and Eckstein, L., 2004, Effects causing untripped rollover of light passenger vehicles in evasive maneuvers. SAE 2004-01-1057. [6] Nguyen, V., 2005, Vehicle handling, stability, and bifurcation analysis for nonlinear vehicle models. Masters thesis, Department of Mechanical Engineering, University of Maryland, College Park. [7] Allen, R.W., Myers, T.T., Rosenthal, T.J. and Klyde, D.H., 2000, The effect of tire characteristics on vehicle handling and stability. SAE 2000-01-0698. [8] Allen, R.W., Klyde, D.H., Rosenthal, T.J. and Smith, D.M., 2003, Estimation of passenger vehicle inertial properties and their effect on stability and handling. SAE 2003-01-0966. [9] Hac, A., 2002, Rollover stability index including effects of suspension design. SAE 2002-01-0965. [10] Hac, A., 2005, Inuence of chassis characteristics on sustained roll, heave and yaw oscillations in dynamic rollover testing. SAE 2005-01-0398. [11] Eger, R. and Kiencke, U., 2003, Modeling of rollover sequences. Control Engineering Practice, 11, pp. 209216. [12] Takano, S., Nagai, M., Nagai, M., Taniguchi, T., and Hatano, T., 2003, Study on a vehicle dynamics model for improving roll stability. JSAE Review, 24, pp. 149156. [13] Whitehead, R., Travis, W., Bevly, D.M. and Flowers, G., 2004, A study of the effect of various vehicle properties on rollover propensity. SAE 2004-01-2094. [14] Garrott, W.R. and Heydinger, G.J., 1992, An investigation, via simulation, of vehicle characteristics that contribute to steering maneuver induced rollover. SAE 92085. [15] Frimberger, M., Wolf, F., Scholpp, G. and Schmidt, J., 2000, Inuences of parameters at vehicle rollover. SAE 2000-01-2669. [16] Chen, B.C. and Peng, H., 2001, Differential-braking-based rollover prevention for sport utility vehicles with human-in-the-loop evaluations. Vehicle System Dynamics, 36(4), pp. 359389.

292

N. Zhang et al.

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010

[17] Lee, A.Y., 2002, Coordinated control of steering and anti-roll bars to alter vehicle rollover tendencies. Journal of Dynamic Systems, Measurement, and Control, 124(1), pp. 127132. [18] Johansson, B. and Gafvert, M., 2004, Untripped SUV rollover detection and prevention. 43rd IEEE Conference on Decision and Control, Atlantis, Paradise Island, Bahamas. [19] Hac, A, 2002, Inuence of active chassis systems on vehicle propensity to maneuver-induced rollovers. SAE 2002-01-0967. [20] Ungoren, A.Y. and Peng, H., 2004, Evaluation of vehicle dynamic control for rollover prevention. International Journal of Automotive Technology, 5(2), pp. 115122. [21] Segel, L., 1956, Theoretical prediction and experimental substantiation of the response of the automobile to steering control. The Institution of Mechanical Engineers, Proceedings of the Automobile Division, No. 7, pp. 310330. [22] Pacejka, H.B., 2002, Tyre and vehicle dynamics. SAE, ISBN 0 7680 1126 4. [23] National Highway Trafc Safety Administration, 2000, DOT announces proposal to add rollover ratings to auto safety consumer information program. NHTSA Now 6(7). [24] Dorohoff, M.D., 2003, A study of vehicle response asymmetries during severe driving maneuvers. Masters thesis, Department of Mechanical Engineering, The Ohio State University. [25] Wilson, K.A. A jeep rollover proves hard to understand. www.autoweek.com, 11/19/2001.

Appendix

Table A1. Name and value Ms = 1350 kg Mu1 = 42 kg Mu2 = 42 kg Mu3 = 42 kg Mu4 = 42 kg Ixx s = 380 kg m2 Iyy m = 1000 kg m2 Izz s = 2240 kg m2 Ixz s = 0 kg m2 Izz u = 1 kg m2 a = 1.253 m b = 1.508 m tf = 0.75 m tr = 0.75 m hs = 0.3 m k basic = 20000 N/m Ks1 = k basic Ks2 = k basic Ks3 = k basic Ks4 = k basic c basic = 1500 N.s/m Cs1 = c basic Cs2 = c basic Cs3 = c basic Cs4 = c basic Kt = 200000 N/m Pdr Pfai = 0.07 rad Pdf Pfai = 0.8

Basic vehicle parameters for 9-DOF yawrollpitchbounce model. Expression Vehicle sprung mass Vehicle unsprung mass of the left-front suspension Vehicle unsprung mass of the right-front suspension Vehicle unsprung mass of the left-rear suspension Vehicle unsprung mass of the right-rear suspension Moment of inertia about the x-axis of the vehicle sprung mass (about CG) Moment of inertia about the y-axis of the vehicle sprung mass (about CG) Moment of inertia about the z-axis of the vehicle sprung mass (about CG) Product of inertia about the xz axes of the vehicle sprung mass (about CG) Moment of inertia about the z-axis of the 1/4 vehicle unsprung mass (about CG) Distance from the sprung mass CG to the front axle Distance from the sprung mass CG to the rear axle 1/2 width of the front axle 1/2 width of the rear axle Height of the sprung mass CG above the roll axis

Spring stiffness of the left-front suspension Spring stiffness of the right-front suspension Spring stiffness of the left-rear suspension Spring stiffness of the right-rear suspension Damping coefcient of the left-front suspension Damping coefcient of the right-front suspension Damping coefcient of the left-rear suspension Damping coefcient of the right-rear suspension Vertical tire stiffness Roll steer angle at the rear axle Relation between roll angle and camber angle

Vehicle modelling, roll and yaw instability Table A2. Name and value Ms = 1350 kg Mu = 168 kg I xx s = 380 kg m2 I xz s = 0 kg m2 I zz s = 2240 kg m2 I zz u = 416.83 kg m2 THITA R = 0 rad a = 1.26711 m b = 1.49389 m c = 0.01411 m e = 0.1134 m Tw = 1.5 m hs = 0.3 m K R = 45000 Nm/rad c R = 3375 Pdr Pfai = 0.07 rad Pdf Pfai = 0.8 Rolling sprung mass Non-rolling unsprung mass Basic vehicle parameters for 3-DOF yawroll model. Expression

293

Moment of inertia about the x-axis of the rolling sprung mass (about CG) Product of inertia about the xz axes of the rolling sprung mass (about CG) Moment of inertia about the z-axis of the rolling sprung mass (about CG) Moment of inertia about the z-axis of the non-rolling unsprung mass (about CG) Inclination angle of the roll axis pointing down Distance from the vehicle CG to the front axle Distance from the vehicle CG to the rear axle Distance from the CG of Ms to the vehicle CG Distance from the CG of Mu to the vehicle CG Average vehicle track width Distance from the CG of Ms to the roll axis Roll stiffness Roll damping coefcient Nm.s/rad Roll steer angle at the rear axle Relation between roll angle and camber angle

Downloaded By: [Canadian Research Knowledge Network] At: 16:30 7 March 2010