Multi-body dynamics in full-vehicle handling analysis
S Hegazy, H Rahnejatà and K Hussain Department of Mechanical Engineering, University of Bradford, West Yorkshire, UK Abstract: This paper presents a multidegrees-of-freedom non-linear multibody dynamic model of a vehicle, comprising front and rear suspensions, steering system, road wheels, tyres and vehicle inertia. The model incorporates all sources of compliance, stiffness and damping, all with non-linear characteristics. The vehicle model is created in ADAMS (automatic dynamic analysis of mechanical systems) formulation. The model is used for the purpose of vehicle handling analysis. Simulation runs, in-line with vehicle manoeuvres specified under ISO and British Standards, have been undertaken and reported in the paper. Keywords: multi-body dynamics, suspensions, steering, non-linear characteristics, simulations NOTATION a b c Ck dt Flat Flong Fq Fvert f g [J] k klat K l m MX MY MZ n {q}T s t u distance from vehicle c.g. to the front tyre contact patch distance from vehicle c.g. to the rear tyre contact patch damping constant (tyre) kth constraint function in a joint integration step size lateral tyre force longitudinal tyre force generalized forces in an Euler frame of reference vertical tyre force coefficient of rolling resistance gravitational acceleration Jacobian matrix tyre vertical stiffness tyre lateral stiffness kinetic energy axial length number of constraints overturning moment rolling resistance moment aligning moment number of rigid parts generalized coordinates a scaling factor time longitudinal speed v x, y, z á â ã ä ë ì ø, è, ö ø10 lateral speed displacements in Cartesian coordinates tyre slip angle steer angle camber angle tyre deflection Lagrange multiplier coefficient of friction Euler angles vehicle body roll

Superscript . rate of change with time Subscripts f i, j k r front tyre body i relative to body j kth holonomic constraint function rear tyre



The MS was received on 19 January 1999 and was accepted after revision for publication on 2 June 1999. Ã Corresponding author: Department of Mechanical Engineering, University of Bradford, Bradford, West Yorkshire BD7 1DP UK. ,
K00199 # IMechE 1999

Multi-body dynamics has played an increasingly important role in the analysis of vehicle motions ever since the introduction of linear vehicle dynamic models by Segel [1] for lateral accelerations of up to 0.3 g. However, linear models include significant assumptions: (a) small steering inputs at normal constant vehicle speeds,
Proc Instn Mech Engrs Vol 213 Part K

Their tyre model simulations are based upon a composite slip formulation as a function of lateral and longitudinal slip. Most tyre models currently used in vehicle dynamic simulations involve empirical representations of the measured tyre data. 5] have investigated a three-dimensional vehicle model with 42 rigid-body degrees of freedom. The formula also takes into account the curvature factors. which are functions of vertical load and camber angle. They also proposed a numerical procedure designed to permit efficient vehicle dynamic analysis on a microcomputer of the type in use at that time. full suspension geometry. Pacejka [7] introduced handling diagrams for the analysis of the steady state behaviour of a vehicle. non-linear suspension forces and the non-linear tyre forces and moments. A comprehensive review of the tyre models is given by Pacejka and Sharp [9]. and the longitudinal force (brake force) as a function of longitudinal slip. a stability factor and time of manoeuvre for lateral/ directional control. They have also Proc Instn Mech Engrs Vol 213 Part K presented a driver model to steer the vehicle along a prescribed path during their closed-loop simulation study. lateral and vertical forces. Vehicle motions are described in terms of the fixed global frame of K00199 # IMechE 1999 . Allen et al.20 S HEGAZY. Kortum and Schiehlen È È [15] and Kubler and Schiehlen [16]. These include a non-linear vehicle model and a linear bicycle model that has two degrees of freedom. the tyre cornering forces must be carefully determined. representing an empirical method for fitting tyre data. [11. In order to study the handling behaviour of a vehicle. steering characteristics. The simplest form of tyre modelling involves the computer storage of a large amount of measured tyre data which are used in conjunction with an interpolation method to represent the tyre forces characterized by the measured data. and thus control the directional response of the vehicle during the various manoeuvres. Another function of the suspension system is to maintain the road wheels at correct orientations to the road surface. [10] employed a tyre formula to describe the characteristics of side force and self-aligning torque as functions of slip angle. È 2 DESCRIPTION OF THE VEHICLE MODEL The full vehicle model comprises vehicle mass and inertia.4 s while travelling at a forward tangential speed of 75 kmah. (c) smooth flat roads and (d) lateral tyre forces not altering with small changes in vertical tyre forces. many sources of non-linearity in suspension kinematics. Various literature reviews are provided by Kortum and Sharp [14]. a rack and pinion steering system. roll angle and yaw velocity. The non-linear vehicle model employs a multi-body dynamics formulation and includes a non-linear steering system. performing a double-lane change manoeuvre in order to perform a transient handling simulation. vehicle ride and handling analysis has accounted for one of the largest growth areas in the application of multi-body dynamics. The model has come to be known as the `magic formula'. in fact. a primary function of which is ride comfort. Their analyses include a steady state model for the determination of side force coefficients. H RAHNEJAT AND K HUSSAIN (b) linear tyre behaviour with slip and camber angles. The linear model was considered as a linear time-invariant system with two degrees of freedom and included the evaluation of lateral velocity and yaw rate of the vehicle with constant forward speed. They subjected their model to a severe steering ramp input of 2108 in 0. through multi-body dynamics. There are. Other analyses are concerned with vehicle handling characteristics in response to various intended manoeuvres. Chace [3] and Orlandea et al. [4. Often. camber and normal load) on both paved and off-road surfaces. The simulation results were presented for lateral acceleration. which define the tyre stiffness components. This method is currently used in general purpose dynamics software. The performance characteristics of a vehicle in all such tests are profoundly affected by its suspension. tyre properties and in the vehicle inertial dynamics in roll. adjustable systems and active elements. tyre geometry and peak force variations with slip angle or longitudinal slip. which conflicts with the requirements of good handling analysis which usually calls for a relatively high value of stiffness. [6] reported two vehicle models: a linear and a non-linear dynamic model. with good accuracy. In fact. These conflicting requirements have led to the gradual introduction of independent suspensions. The steady state and dynamic models included a tyre model for comprehensive slip. Allen et al. front and rear double-wishbone suspensions. This tyre model contains 53 coefficients. pitch and yaw motions when it is subjected to longitudinal. A considerable volume of literature deals with the issue of ride comfort of vehicles in single-event perturbations such as negotiating a bump or a ditch. using two different models. 12] investigated an expanded version of vehicle simulation tyre models for a full range of operating conditions (slip. Naude and Steyn [8] have investigated a computer simulation for the handling characteristics of a vehicle. received an impetus with the work of McHenry [2]. Suspensions may also be designed so that the stiffness of the bushes contributes to the overall roll stiffness of the vehicle. The formula is limited to steady state conditions during pure cornering and pure braking. Vehicle handling analysis. A handling diagram was defined as the plot of lateral acceleration versus the difference between the slip angles of rear and front tyres. Pacejka et al. Xia and Willis [13] have studied the effect of tyre cornering stiffness on the vehicle frequency response. for good ride comfort the suspension system should provide a relatively low vertical stiffness. road wheels and tyres. Compliance characteristics in vehicle suspension systems are usually of a non-linear nature.

These are through the use of an equation Fig. The tyre model reported here does not include the overturning moment. 2. and to the lower steering column by a universal joint (see Fig. The bushings provide appropriate longitudinal and torsional compliance which influences the dive (during braking) and squat (during acceleration) characteristics of the vehicle.48 of pinion rotation. 2. Local part frames of reference xi . Z shown in Fig. referred to as the lower and upper control arms respectively. 1. 4a). based upon Lagrange's equation for constrained systems. 2. These are the vertical tyre force. longitudinal traction force and lateral force. Each quarter-suspension comprises two control arms. The steering rack is connected to the tie rods by universal joints. The radial force is dependent on tyre deflection and its rate of change. This component is used to calculate the tyre vertical force. The time rate of change in deflection is obtained by the vector scalar product of the instantaneous tyre radius vector and the wheel centre global velocity. as well as the self-aligning moment.2. 2. Y. The coupler connects between the cylindrical and translational joints. the overturning moment and the rolling resistance moment (see Fig. The steering knuckles. has a component that acts in the direction of the contact normal at the tyre±road contact patch. 2. Typical characteristics for these elements are also shown in Fig. while a rebound stop is located on the upper control arm. yi. z i are attached to all of the moving parts i. Typical characteristic curves for these are shown in Fig. 3). The upper steering column is connected to the steering wheel and to the vehicle body by revolute joints. The tie rods are attached to the steering knuckles by spherical joints which represent the ball joints. The steering rack is connected to the vehicle body by a translational joint. is employed for the derivation of equations of motion for all parts in the model in a body 3±1±3 Euler frame of reference (see Section 3). acting in the radial plane of the tyre. A generic formulation method. The pinion is connected to the steering column by a universal joint.1 Suspensions and steering system models by spherical joints in appropriate locations which define the inclination of the steering axis. 3. both measured along the tyre vertical directional vector [17]. The deflection is obtained by an instantaneous evaluation of the distance between the position of the wheel centre and the road surface plane in the contact patch. The control arms are connected to the vehicle body by elastic rubber bushings with non-linear characteristics. as indicated by the first and second terms in equation (1): Fvert ˆ kä À c 2.2 Tyre model The front and rear suspensions are of double-wishbone configuration. and to the vehicle body by a cylindrical joint. in turn. A bump stop is situated on the lower control arm.2. are connected to the upper and lower control arms There are six components of force and moment generated as a result of tyre interaction with the road. The quantities thus obtained are employed to obtain the stiffness and damping contributions to the tyre vertical force. The shock absorber is attached to the vehicle underbody and the lower control arm.2 dä dt (1) Tyre lateral force There are two alternative methods for determination of the tyre lateral force.1 Tyre vertical force The radial tyre contact force. 1 K00199 # IMechE 1999 Axis system for the vehicle model Proc Instn Mech Engrs Vol 213 Part K . describing their motion relationship which is 166 mm of rack travel resulting from 1148. also shown in Fig.MULTI-BODY DYNAMICS IN FULL-VEHICLE HANDLING ANALYSIS 21 reference X.

2 Sources of compliance in the vehicle model Proc Instn Mech Engrs Vol 213 Part K K00199 # IMechE 1999 .22 S HEGAZY. H RAHNEJAT AND K HUSSAIN Fig.

l ˆ a and á ˆ ár. These forces oppose the motion of the Proc Instn Mech Engrs Vol 213 Part K . with the vertical force. The coefficient of K00199 # IMechE 1999 The rolling resistance force and the traction (or braking) force together constitute the tyre longitudinal reaction force [17]. The equation method.3 Tyre longitudinal force (2) v ‡ lø10 u  Àâ (3) For á ˆ áf .MULTI-BODY DYNAMICS IN FULL-VEHICLE HANDLING ANALYSIS 23 Fig. l ˆ Àb. The rolling resistance force is calculated by multiplying the coefficient of rolling resistance. f. 2. employed in this analysis uses the following simplified equation to generate the lateral force: Flat ˆ ìFvert (1 À eÀ k lat jáj ) where  á ˆ tan À1 friction ì changes from the initial static to the instantaneous dynamic conditions as shown in Fig. 3 Schematic representation of front vehicle suspension and steering system method or by an interpolation process [17].2. The traction force is obtained by multiplying the instantaneous value of the coefficient of friction with the vertical force. 4b.

as in the attachment between the upper steering column and the steering wheel.24 S HEGAZY. Fig. the road wheels and tyres. Under dynamic conditions. where the rotational components are the Euler angles in body 3±1±3 successive rotations. for instance between the steering left front knuckle and its upper wishbone. The combined force is referred to as the longitudinal force: Flong ˆ (ì À f )Fvert (4) 2. q.4 ˆ y5.2 Holonomic constraint functions vehicle at the road surface contact patch. C k . H RAHNEJAT AND K HUSSAIN joint. the assembly of parts can be represented mathematically in a manner that conforms to the required dynamic functions of the system. the following three constraints pertaining to an at-point condition exist: C kˆ1À3 ˆ x22. There are 34 parts in the vehicle model. excluding ground (see Table 1). Table 2 lists the different types of joint employed in the assembly of various parts in the vehicle model. equation (6) provides six equations of motion per part in the vehicle system model.24 ˆ z22. It also includes the vehicle body.4 : C kˆ1À3 ˆ x5.1 METHOD OF FORMULATION AND SOLUTION Equations of motion There are 34 parts in the multi-body model. Therefore. These are introduced as holonomic algebraic constraint functions. For a spherical joint. the rear suspension. Tables 1 and 2 provide the list of all parts and all constraints in the full vehicle model.3 Vehicle model The full vehicle model is an assembly of the front suspension and steering system. Each joint or assembly attachment introduces constraint functions in the form of non-linear algebraic equations.4 ˆ 0 (8) K00199 # IMechE 1999 . the motion of each of which can be described in terms of the generalized coordinates. coupler and specified motions is given in Table 2. represented by its mass and inertial components. The Gruebler±Kutzbach expression can be used to È È determine the available degrees of freedom in the vehicle model. the at-point constraint is supplemented by two-axis orthogonality conditions around the rotation ø5. 4 (a) Tyre forces and moments and (b) variation in coefficient of friction with longitudinal speed 3. Typical constraint functions for a number of these joints are given below.24 ˆ y22.24 ˆ 0 (7) For a revolute joint. Thus: nDOF ˆ 6n À ˆ 94 ˆ constraints ˆ 6n À m ˆ 6(34) À 110 (5) 3 3. by Lagrange's equation for constrained systems:   m ˆ dC k d dK dK ˆ0 À Fq ‡ ëk À • dq dt d q dq kˆ1 (6) The generalized coordinates are given by {qgT ˆ fx y z ø è j}.4 ˆ z5. The number of constraints for each Proc Instn Mech Engrs Vol 213 Part K Ideal functions in all mechanisms are assured by appropriate use of constraints in the form of joints or attachments. The reaction forces in the multibody system are given by the summation term in equation (6) along each of the generalized coordinates.

1 À74.10 cos ö3.2 À177 À177 92.4 767.12 1.7E‡04 2.7 15.6 14 14 0.4 À767.3E‡04 8.4 ˆ 0 and C kˆ5 ˆ sin è5.5 243.10 sin ö7.5E‡05 3000 3000 8E‡04 8E‡04 27 27 2. with the degrees of freedom z7.10 sin ö7.4E‡05 1.8 6 6 0.6 À39 À39 À141.7 543.4E‡04 3200 1E‡05 1E‡05 300 300 1.6 À166 373 373 À39.10 .10 ˆ 0 (16) (17) (18) (11) (12) (13) C kˆ5 ˆ sin ø3.3 507.10 À sin ø7.4 2700.4 ˆ 0 (9) The z3.1 128.2E‡04 8.6 2469. has the following constraint functions: C kˆ1 ˆ x7.10 ) ˆ0 C kˆ2 ˆ y3.10 ˆ 0 The steering motion for the vehicle manoeuvre is given as a function of time (see Fig.10 cos è7.4E‡04 3200 5E‡04 5E‡04 3000 3000 8E‡04 8E‡04 8200 8200 1.1 555.83E‡08 1.10 À cos ø3.5 À92 À166.9 À574. ø7.10 .9 2526 2526 2513 2513 2520 2520 2700.6 À573 À573 2577 2577 Y Ð 0 0 396 396 333.59E‡06 1.59E‡06 1.10 ˆ 0 C kˆ4 ˆ sin è7.2E‡04 27 27 2.1 2.5E‡05 1.3 À701.4E‡05 6000 6000 1E‡05 1E‡05 1.84E‡05 1.9 À572.10 cos ö7.12 1.3 À556.4E‡05 1.7 0.10 sin è3.4 556.3 13.2 42.4E‡04 80 40 77 1.10 cos è3.7E‡04 200 200 2.10 (sin ø3.10 cos è3.10 cos ö3.1 À543.5E‡04 1.10 translational motion of the steering rack with respect to the vehicle body has the constraint functions C kˆ1 ˆ x3.2 42.3 À516.6 1.3 À701.31E‡05 1.8 574 450.68 1.5 À489.10 sin ö3.31E‡05 8200 8200 1.7 À559.4 sin ö5.1 767.8 0.10 sin ö3.68 1.10 sin è3.5E‡04 3.5 À764 764 À555.6 0.9 À134.10 (cos ø3.12 15.10 ) ˆ0 C kˆ3 ˆ sin è7.2 X Ð 1200 À697 674 295.5 À355.5 À134.10 cos è3.4 2478.1 1.3E‡05 1.59E‡06 and C kˆ4 ˆ sin è5.7 0.10 cos ö7.4 cos ö5.10 (cos ø7.7E‡04 200 200 1.404E‡09 1.5E‡04 7000 7000 8.4 À95.5 À639 À447 À447 À553 À553 À516.10 (sin ø7.4E‡04 2.1 0.3 1.7E‡04 2.4 109.2 42.3E‡04 8.10 sin ö7. for instance between the pinion gear and the vehicle body.5E‡04 3. This presents a single constraint function as Proc Instn Mech Engrs Vol 213 Part K .3 À572.3 489.5 À39.1 À756 756 Z Ð 160 À128.6 2469.3 À538.4 À51 À51 À51 À51 Ixx Ð 4.8 1.84E‡05 1.10 ˆ 0 K00199 # IMechE 1999 C kˆ3 ˆ sin ø3.10 À cos ø7.10 ) ˆ0 C kˆ2 ˆ y7.10 ) ˆ0 (10) (15) (14) A cylindrical joint.10 cos ö3.5E‡04 1.1E‡04 8.10 ‡ cos ø3.5E‡04 1.4E‡05 2.2E‡04 8200 8200 1.4E‡05 6000 6000 9E‡05 9E‡05 9E‡05 9E‡05 Inertia (kgamm2 ) I yy Ð 2.68 1.10 À sin ø3.4E‡05 1.8 À522 522 À579.10 sin ö3.4E‡05 1.59E‡06 1.4 À95. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Part name Ground Vehicle body Steering rack Steering wheel Upper steering column Lower steering column Pinion Lower wishbone left Lower wishbone right Upper wishbone left Upper wishbone right Steering knuckle left Steering knuckle right Tie rod left Tie rod right Upper damper left Upper damper right Lower damper left Lower damper right Lower wishbone left Lower wishbone right Upper wishbone left Upper wishbone right Steering knuckle left Steering knuckle right Tie rod left Tie rod right Upper damper left Upper damper right Lower damper left Lower damper right Front tyre left Front tyre right Rear tyre left Rear tyre right Mass (kg) Ð 1185 4.10 ˆ0 C kˆ4 ˆ Àcos ø3.8 13.8 À141.9 128.4E‡05 6000 6000 1.68 42.MULTI-BODY DYNAMICS IN FULL-VEHICLE HANDLING ANALYSIS 25 Table 1 Mass and inertial properties in the vehicle model Centre of mass location (mm) No.6 92.5 579.4E‡05 6000 6000 9E‡05 9E‡05 9E‡05 9E‡05 Izz Ð 2.2E‡04 8.8 559.4 À549 549 À507.8 1.482E‡09 460 2.2 À74.7 15.1E‡04 2.8 109.10 cos è7.12 15.6 2478.5 138.3E‡05 8200 8200 1.10 cos ö7. 5).9 À574.8 À767.8 538.

6 À573 À573 2577 2577 Y 396 396 396 271 234.9 À367 367 À531.5 À574. the scalar constraint functions (Section 3.4 À705. similar to constraint functions (8) to (19).4 À191.5 0 À731.6 À140.1).8 À29.7 2569.8 À705.9 À128.3 À550.6 2471.8 À550.9 À51 À51 À51 À51 No.9 730.8 À192 À192 100 100 À148.9 À743. b 0X2 2 X 0X22ð. of constraints 5 5 4 4 4 5 3 3 3 3 3 3 4 4 4 4 3 3 3 3 3 3 4 4 5 5 5 5 1 1 1 Fig.9 À49. 3.2 556. t < 0X1 0X1 < t < 0X3 t > 0X3 (19) There are.4 742.8 À756 À756 Z 574 450.8 À128. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Constraint type Revolute joint/26 Revolute joint/33 Universal joint Universal joint Cylindrical joint/54 Translational joint/61 Spherical joint Spherical joint Spherical joint Spherical joint Spherical joint Spherical joint Universal joint Universal joint Cylindrical joint Cylindrical joint Spherical joint Spherical joint Spherical joint Spherical joint Spherical joint Spherical joint Cylindrical joint Cylindrical joint Revolute joint Revolute joint Revolute joint Revolute joint Coupler Motion Motion Part I Steering wheel Upper steering column Upper steering column Lower steering column Pinion Steering rack Lower wishbone left Lower wishbone right Upper wishbone left Upper wishbone right Steering knuckle left Steering knuckle right Steering rack Steering rack Upper damper left Upper damper right Lower wishbone left Lower wishbone right Upper wishbone left Lower wishbone right Steering knuckle left Steering knuckle right Upper damper left Upper damper right Steering knuckle left Steering knuckle right Steering knuckle left Steering knuckle right Joint 54 Part J Upper steering column Vehicle body Lower steering column Pinion Vehicle body Vehicle body Steering knuckle left Steering knuckle right Steering knuckle left Steering knuckle right Tie rod left Tie rod right Tie rod left Tie rod right Lower damper left Lower damper right Steering knuckle left Steering knuckle right Steering knuckle left Steering knuckle right Tie rod left Tie rod right Lower damper left Lower damper right Front tyre left Front tyre right Rear tyre left Rear tyre right Joint 61 Joint 26 Joint 33 X 674 295.26 S HEGAZY.7 À49.5 À114 À597 À639 À697.7 À148.4 À556.8 780. H RAHNEJAT AND K HUSSAIN Table 2 Assembly constraints in the vehicle model Joint location (mm) No. 110 constraint functions in the vehicle model.2 2471. 5 Steering wheel angle input with time V b 0.9 À670.2).4 376.9 731.5 317 À40 À92 À128.9 À742.6 376.3 À580.8 À191.9 531.9 À140.5 À697. & !' ` ð(t À 0X1) ð ⠈ 0X11ð 1 ‡ sin À . the applied forces K00199 # IMechE 1999 Proc Instn Mech Engrs Vol 213 Part K .7 2584.5 À574.2 2704.2 À780.4 743.5 À580.9 705.1 À705.9 2584.4 À730.4 670.8 À29.1 À697.5 2569.3 Formation of the Jacobian matrix The set of differential equations of motion (Section 3. in fact.9 2704.

and step-by-step integration using a `stiff' algorithm for widely split eigenvalue problems [18±20]. Figure 6 shows an animated output for the transient manoeuvre during a simulation time of 5 s. The Jacobian matrix is also quite large in dimensions. which effect a load transfer Fig.MULTI-BODY DYNAMICS IN FULL-VEHICLE HANDLING ANALYSIS 27 and the compliance functions (e. velocity and acceleration of all parts) and the Lagrange multipliers representing the joint reactions. lane changing and slalom motions.2).3 and 76. 4 FULL-VEHICLE SIMULATION STUDIES AND DISCUSSION An important measure of vehicle performance is its handling characteristics when subjected to a given steering input. The solution to the differential-algebraic set of equations is obtained in small variable time steps. A number of important parameters are investigated. including double-lane changes. spring forces. Typical vehicle handling simulations include transient cornering. the rear tyres carry approximately 54. The tyre forces at time t ˆ 0 correspond to the initial static equilibrium position. as it embodies appropriate coefficients for six equations of motion for all the vehicle parts listed in Table 1. during which body roll in the region 2±88 can occur. During the manoeuvre. and all the formulated constraint functions for the joints in the model (see Table 2 and Section 3. ëgT ˆ fFq g The Jacobian matrix is of the following form: s dK dK T dt dq ‡ dq ! [J] ˆ T R dC dq P ! !Q dC dë U U S [0] (20) (21) The Jacobian matrix contains many zero entries. Five hundred time steps of simulation were undertaken.5 per cent of the total weight. as the steering wheel input increases from 0 to 408. in matrix form the set of equations is represented by [J]fq. tyre forces. In fact fewer than 10 per cent of all the elements of the matrix are usually non-zero.8 g. dt. Thus. with or without braking. Test procedures have been specified by international standards in ISO 7401± 1988 or in the British Standard BS AU 230:1989 [21]. 6 Vehicle manoeuvre for 5 s of simulations K00199 # IMechE 1999 Proc Instn Mech Engrs Vol 213 Part K . thus being referred to as sparse. Due to the weight distribution of the vehicle. the outside wheels generate larger vertical tyre forces owing to the inertial forces. Figures 7a and b show the vertical tyre forces during the specified manoeuvre. These include the tyre forces.7 mm above the ground respectively. The vector of unknowns includes the system state variables (position. The specified manoeuvre should represent a realistic test of vehicle behaviour under severe conditions. after an initial static equilibrium analysis is carried out to ensure vehicle placement at the kerb height. The vertical excursion of the front and rear roll centre heights is also of interest as a large displacement of these can affect the vehicle stability. The current analysis is concerned with a transient cornering manoeuvre with a constant forward velocity. employing a predictor±corrector technique with a Newton ±Raphson method for the solution of a non-linear set of simultaneous equations. particularly on the inside wheels as the vertical force diminishes with increasing lateral acceleration. The corresponding front and rear roll centre heights are 70. Most modern vehicles can undergo cornering manoeuvres with lateral accelerations of up to 0. The stability of the vehicle under various specified manoeuvres can be investigated. bushing reactions) have to be solved in discrete small time steps. damper forces and bump and rebound stop forces.g.

the vertical tyre forces on these wheels decrease. corroborating this argument. The front and rear roll centres undergo vertical excursions towards the ground with downward articulation of the control arms (see Fig. The bump and rebound forces are shown in Figs 8a and b. The reduction in the vertical tyre force on the inside wheels is an indication of vehicle instability. Consequently. This trend can be observed in Figs 7a and b. Reference [21] outlines the standard test for the transient response of a vehicle subjected to a step input steering K00199 # IMechE 1999 .28 S HEGAZY. suspension articulation and generation of shock absorber reactions. 9a). H RAHNEJAT AND K HUSSAIN Fig. 7 Vertical forces for (a) front tyres and (b) rear tyres from right to left. It can be observed that for this simulation the inside wheels off-load rapidly before a steady condition is reached Proc Instn Mech Engrs Vol 213 Part K through body roll. The inside wheels are subjected to the opposite effect owing to rebound. It should also be noted that the transient response dies down quickly after a few oscillations. this being an indication of an acceptable response of the vehicle to the steering input. The inside wheel tyres still carry sufficient vertical forces to maintain a good contact with the road. Therefore. the vertical tyre forces on the outside wheels increase and generate a bump stop force in the opposite direction.

roll angle.39 s (see Fig.3 s. the peak response time and the lateral acceleration overshoot. reached after 0. Both the lateral acceleration and the roll angle rise with an increasing steering angle. 9b). optimal conditions relate to the minimization of all of these parameters. 0. Proc Instn Mech Engrs Vol 213 Part K .MULTI-BODY DYNAMICS IN FULL-VEHICLE HANDLING ANALYSIS 29 Fig. 8 (a) Bump stop forces for front tyres and (b) rebound stop forces for front tyres function. 5 and equation (19). The peak response time is also measured in the same time frame. Under this test. The response time for the lateral acceleration is. therefore. The maximum steering angle is 408. if the origin of the time base K00199 # IMechE 1999 is set at the point where half the maximum steering input has been accomplished. lateral acceleration and yaw rate) can be obtained as measures of vehicle performance.e. Figures 9b and c illustrate the lateral acceleration and roll angle change for the step steering input shown in Fig.56 s. The parameters of interest in this transient analysis are the response time. This is also shown in the same figure and is 0. In this standard transient response test the vehicle response time is defined as the period taken for the vehicle to reach 90 per cent of the maximum response value. 9c. These are shown in the figure. Clearly. The same parameters can also be measured from the variation in roll angle in Fig. the variations in the response (i.

(b) lateral acceleration variation with time and (c) vehicle body roll angle variation with time Proc Instn Mech Engrs Vol 213 Part K K00199 # IMechE 1999 .30 S HEGAZY. 9 (a) Instantaneous roll centre height variation. H RAHNEJAT AND K HUSSAIN Fig.

and Willis. Kortum.. R. Machines and Mechanisms. Chace. W. The effect of tire cornering stiffness on vehicle linear handling performance. pp. British Standard Automobile Series. 28±48. paper C553/039/98.2. USA). Society of Automotive Engineers. Syst. Dynamics. 310±330. Kubler. The vehicle lateral acceleration overshoot is 0. 1971. Pacejka. Kortum. R. M. in handling of vehicles under emergency conditions. October 1987 (Mechanical Dynamic Inc. R. Proc. 1984. Pacejka. The simulation highlighted in the paper indicates that sufficient tyre forces are generated. B. The extent of lateral acceleration overshoot. Syst. 1968±9. 97±102. J. The Magic Formula tire model. Veh. Dynamics. Gear. W. An analysis of the dynamics of automobiles during simultaneous cornering and ride motions. Warrendale. Allen. (7). 1989 (British Standards Institution). Nato ASI Series. and Sharp. Rahnejat. C. Trans. 1998. Shear force development by pneumatic tyres in steady state condition: a review of modelling aspects. Computer Aided Analysis and Optimization of Mechanical System Dynamics. A tire model for use with vehicle dynamics simulations on pavement and off-road surfaces. S. 12 13 14 ACKNOWLEDGEMENT The authors wish to express their gratitude to the Egyptian Military Attache's Office for the financial support extended  to this research project. 1984 (Springer-Verlag). A. UK. A large amount of proprietary physical and geometrical data has been included in the paper which can be used by others who intend to carry out similar vehicle handling studies. 249±258. A. Magdaleno. Veh. J. X. Veh. and Rosenthal. 15 16 REFERENCES 1 Segel. Engng for Industry. N. ADAMS/Tyre User's GuideÐVersion 5. Instn Mech. The results of such an analysis can be used to measure the vehicle `responsiveness' in terms of given parameters specified in ISO and BS standards for non-steady conditions. Simplified analysis of steady state turning behaviour of motor vehicleÐPart 1: handling diagram of simple system. Multi-body Computer Codes È in Vehicle System Dynamics. R. Proc. Gear. 1956±7. K00199 # IMechE 1999 Proc Instn Mech Engrs Vol 213 Part K . 89±95. 49±78. H. Dynamics. Instn Mech. W. F. 27. W. and Calahan. Rosenthal. J. in turn can provide a good measure of vehicle stability. BS AU 230. and Chace. W. Veh. R. H. In Proceedings of NATO ASI on Computer Aided Analysis and Optimization of Mechanical System Dynamics. 229±263. L. particularly on the inside wheels. pp. 1997. Objective evaluation of the simulated handling characteristics. R. July 1998 (Professional Engineering Publishing. B. R. 95±115. SAE Technical paper 950312.. 3 Chace. Allen. 4 Orlandea. E. D. 2. 1987. pp. T. Theoretical prediction and experimental substantiation of the response of the automobile to steering control. Bury St Edmunds. 1995. W. 1977. 117±126. W. Engrs. T. and Szostak. Simultaneous numerical solution of differentialalgebraic equations. H. 1973 (Swets und Zeitlinger. 1995. Vehicle modular simulation in È system dynamics. SAE Technical paper 870495. F9. Circuit Theory. and Sharp. SAE Technical paper 770053. Pennsylvania. London.07 g and its response time to the steering input is adequately short. SAE Technical paper 950313. 21. J. ASME. 1993. The procedure highlighted in this paper can be employed for simulation of vehicle models under transient manoeuvres. R. and Steyn. N. 1985. Engrs. E. J. H.. W. Allen. Lateral transient response behaviour of passenger cars. L. A sparsity- 17 18 19 20 21 oriented approach to the dynamic analysis and design of mechanical systems. and Schiehlen. Klyde. A. in research in automobile stability and control and in tyre performance. 20. R. 1993. 1977. Pacejka. D. ensuring vehicle adherence to its path. Parts I and II. Amsterdam).). W. A non-linear dynamic analysis for a realistic multi-degrees-of-freedom vehicle model (having 94 degrees of freedom) subjected to a step steering function has been presented. This approach is particularly useful under transient conditions arising from the application of a steering function. Tire modelling requirements for vehicle dynamics simulations. Multi-body Dynamics: Vehicles. pp. B.. S.MULTI-BODY DYNAMICS IN FULL-VEHICLE HANDLING ANALYSIS 31 5 CONCLUSION 5 6 7 8 9 10 11 This paper has shown the use of multi-body dynamics in vehicle handling analysis. The change in the roll angle is within the specified limit for the type of vehicle under investigation. 773±784.. Syst. and Hogue. Christos. H. Simulation of a vehicle suspension with ADAMS computer program. J. Steady state and transient analysis of ground vehicle handling. T. and Schiehlen. T. 18. Methods and experience in computer aided design of large displacement mechanical systems. Dynamics. H. Rosenthal. Dynamics. General purpose vehicle È system dynamics software based on multi-body formalisms.. 1973. Veh. M. Xia. 1991. roll angle variation and generated vertical tyre forces. 2 McHenry. The numerical solution of problems which may have high frequency components. and Bakker. pp. C. 1±18. J. B. 121±176. pp. Syst. R. SAE Technical paper 930826. In Transactions of IMechE Conference on Multi-body Dynamics: New Techniques and Applications. A. Naude. 318± 321. 233±259. J. Orlandea. (13). N. 99. 161±172. A. Syst. 14. M.

Sign up to vote on this title
UsefulNot useful