Vous êtes sur la page 1sur 24

Lateral dynamics and stability – 3

Influence of traction/braking

Prof. R.G. Longoria


Spring 2016

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Model for wheel/tire dynamics under both driving
torque and traction force loading
The dynamics of a torque driven wheel with slip has two dynamic states,

hɺw = I wωɺ w = ∑ Tw = + Td + ∓Tt − Tb − T f (ωw )


   
torque supplied torque due to torque applied axle friction
by drivetrain traction/braking by braking system (speed dependent)

pɺ x = mvɺx = ∑ Fx = ± Ft − ∑ Road loads



force due to
traction/braking z

I I
ωw
r Traction ‘R’ Td
Td
ωw vx
ɺɺ vx
1 T 0 1
x
rw
Tb Ft
Tb
Brake R R
Ft
Under these types of conditions, the traction force depends on a dynamic slip condition.
ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering
Vehicle System Dynamics and Control The University of Texas at Austin
Dynamics of a braking tire
The rotational dynamics of a wheel in traction or braking,
I wωɺ w = Td − Tloss − Tt − Tb
= Td − Tloss − r ⋅ Ftx − Tb

Brake application

z
 ax = Ftx − ∑ Road loads
pɺ x = m
Tb ωw W
g

vx Control problem: forward velocity control


W by controlling Ftx and/or Tb
x
Td For a vehicle, need to consider vertical
Ftx loads and the pitch equation for load
Fz transfer, possibly influence of suspension.
Why is there no rolling resistance
included here? Refer to Wong from Eqs. 3.47 and 3.48.

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Bond graph for traction/braking (optional)
I wωɺ w = Td − Tloss − Tt − Tb
= Td − Tloss − r ⋅ Ftx − Tb
  ax = Ftx − ∑ Road loads
pɺ x = m
Brake application W
g

I For accelerating or I
braking,

Td − Tloss r ⋅ Ftx r
..
1 T 0 1

Tb Ftx Fres
Brake
Pa R
G
In this scenario, we set Tb although in reality this Fz
may be actuated (e.g., hydraulics) and/or feedback- 0.7

0.6

controlled (ABS). 0.5

0.4

0.3

0.2

0.1

Modern passenger vehicles also have intervention -0.1


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

by TCS and AEB.

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Wheel braking simulation

• Study transient dynamics of a the single wheel after a


brake torque is applied.
• Assume wheel is initially rolling without slip
• Slip is determined by the longitudinal velocity and
angular velocity
• The friction coefficient µ is related to value of slip.
0.7

0.6

0.5 c=[-68.593, 238.216,-324.819,219.283,-75.58, 12.088, -0.0068];


0.4

0.3
slip = 0:0.1:1.0;
0.2
N = length(slip);
0.1 µ-slip curve % mu - slip curve
0

-0.1
mu=c*[slip.^6;slip.^5;slip.^4;slip.^3;slip.^2;slip;ones(1,N)];
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Setting up a simple wheel braking simulation

• A main simulation routine sets up the wheel


parameters, initial conditions, etc.
• After initiating the simulation, braking is applied and
velocities are monitored to:
– Compute slip and find µ.
– Watch for when the wheel ‘stops’.
– Watch to see if the wheel locks up.
• The simulation should detect stopping. This can be
done with basic logic, or in Matlab ode functions you
can utilize the ‘events’ option (appendix).

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
simple_brake simulation code
This is a Matlab m-file that sets up the simulation.

clear all This is the ode function file called by the solver.
global mv g theta Iw Rw Set up parameters
global c Td Tb mu slip function xdot=simple_braking(t,x);
global mv g theta Iw Rw
% Define the simulation parameters: global c Td Tb mu slip
mv = 1400/4; % total mass, kg (1/4 of vehicle mass taken)
Rw = 0.31; % wheel radius, m V=x(1); omega=x(2); Wheel states
Iw = 2.65; % wheel moment of inertia, kg-m^2
g = 9.81; % gravitational accel., m/s^2 % determine slip state
theta = 0.0; % slope, radians if V >= Rw*omega,
Tb = 3200/2; Td = 0.0; Define µ-slip data s=(Rw*omega-V)/V;
else Find slip
slip = (0:.05:1.0); % this creates a vector of slip values 0 to 1, stepping 0.05 s=(Rw*omega-V)/(Rw*omega);
mu = 0.7*[0 .4 .8 .97 1.0 .98 .96 .94 .92 .9 .88 .855 .83 .81 .79 .77 .75 .73 .72 .71 end;
.7]; s1 = abs(s);
if s1 > 1.0, s1 = 1.0; end; % keep slip value below 1.0
Vo = 40; Initial conditions mu1=sign(s)*interp1(slip,mu,s1,'linear'); Determine µ
omegao = Vo/Rw;
to=0.0; tf=15.0; xo=[Vo;omegao;0]; % Define the torque input T = Td - Tb;
[t,x]=ode15s(@simple_braking,[to tf],xo); Call solver T = Td - Tb;

figure(1) % Define the state equations with basic logic


subplot(2,1,1), plot(t,x(:,1),t,Rw*x(:,2)); title('Vehicle Braking'); if V < 0.0, V = 0.0; end; % don't let local values drop below 0.0
axis([0 tf -5 Vo+5]) if Rw*omega < 0.0, omega = 0.0; end;
xlabel('Time (sec)'); ylabel('Velocity, m/s');
legend('V','Rw*omega') Plot results Vdot = (mu1*mv*g-mv*g*sin(theta))/mv;
subplot(2,1,2), plot(t,x(:,3)), title('Stopping Distance') omegadot = (-Rw*mu1*mv*g+T)/Iw;
xlabel('Time (sec)'); ylabel('Displacement, m'); xdot=[Vdot;omegadot]; Return state
if V <= 0.0, xdot(1)=0.0; end; derivatives
Note detection of ‘stopping’. if omega <= 0.0, xdot(2)=0.0; end;

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Typical results from wheel braking, with lock up

Vehicle Braking
Goal: Adjust the
V
40
Rw*omega
braking torque to
Vehicle translational velocity
30 prevent lock-up but
Velocity, m/s

20
minimize stopping
Wheel lock-up
distance.
10

0 5 10 15
Time (sec)

Stopping Distance As we know, knowing


200
whether lock-up occurs is
important, especially for
150
Displacement, m

assessing stability of a
100
vehicle.

50 There are common


methods for analyzing
0
0 5 10 15
two-axle brake lock up –
Time (sec) see appendix.
ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering
Vehicle System Dynamics and Control The University of Texas at Austin
6 cases with locked-wheel braking

Brake Off
Brake On

1. Four wheels locked


2. Two rear wheels locked
Bradley & Wood (1931) 3. Two front wheels locked
4. Right-hand front and left-hand rear wheels locked
5. Right-hand front and both rear wheels locked
6. Both front and right-hand rear wheels locked
ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering
Vehicle System Dynamics and Control The University of Texas at Austin
Is there something missing?
• What happens to the ‘traction’ R force when
slip goes to zero?
• What if we parked on a slope? What is the
‘holding force’ that keeps the vehicle from
coming down?
• How would we change our model(s) to allow us
to predict a vehicle, say, riding up a slope and
braking to a stop? Does the traction R model
this situation?
ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering
Vehicle System Dynamics and Control The University of Texas at Austin
References
1. Den Hartog, J.P., Mechanics, Dover edition
2. Dixon, J.C., Tires, Suspension and Handling (2nd ed.), SAE, Warrendale, PA,
1996
3. Gillespie, T.D., Fundamentals of Vehicle Dynamics, SAE, Warrendale, PA,
1992
4. Wong, J.Y., Theory of Ground Vehicles, John Wiley and Sons, Inc., New
York, 1993 (or any subsequent edition)

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Example: Buggy Wheel
In a two-wheel horse-drawn buggy, the square
axle does not rotate but has round ends that fit
into the bearings of each wheel. Each end is
assumed to be loaded by half the weight of the
loaded buggy, W, less the wheel weight, w.

Den Hartog (see handout) describes how to find


the force P which must be supplied by the horse.

This example can be used to understand how we


can represent not only the loads put on any
vehicle by a wheel but also by friction in the
axle.

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Two-Axle Vehicle Braking

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
APPENDIX: Vehicle brake lockup model
Ref. Wong, Ch. 3
The conditions for lockup for a two-axle vehicle can be determined from the basic
longitudinal dynamics model. First, Reverse directions
pɺ x = Fbx + f rW from previous model

= Fbxf + Fbxr + f rW
Find axle load relations (see Wong):
1
W f = Wl2 + h ( Fbx + f rW ) 
L
1
Wr = Wl2 − h ( Fbx + f rW ) 
L Fbr Fbf
3 equations, 5 unknown:
Reverse directions and re-label
as braking forces.
pɺ ,W f , Wr , Fbxf , Fbxr
This formulation is for deceleration, so a positive change
in momentum describes deceleration.
ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering
Vehicle System Dynamics and Control The University of Texas at Austin
Vehicle brake conditions

Assume by brake proportioning: Fbxf = K bxf ⋅ Fbx


K bxr ≡ 1 − K bxf
Fbxr = K bxr ⋅ Fbx
So, only Fbx is not known. Now, for impending lockup, add one
of the additional relations, F = µ W
bxf p f

Fbxr = µ pWr

Unknowns (4): pɺ ,W f , Wr , Fbxf , or Fbxr with 4 equations, including,


 
Fbxf = Kbxf ⋅ Fbx = Kbxf  pɺ − f W 
 r 
Wa g 
ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering
Vehicle System Dynamics and Control The University of Texas at Austin
Basic brake actuation system

-Proportioning
-Metering
-Pressure differential
Stockel, et al

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Finding lockup condition
Front lockup occurs when, Fbxf → µ pW f
a   l2 h a 
Fbxf = Kbxf W  − f r  = µ pW f = µ pW  + 
 g   L L g 
a  l h a
Kbxf  − f r  = µ p  2 + 
g  L L g
 h a  l2
 Kbxf − µ p    = K bxf f r + µ p
 L g  L

Front lockup: Rear lockup:


l2 l1
a K bxf f r + µ p
L  a  (1 − K bxf ) f r + µ p L
g = g =
 f h  r h
K bxf − µ p
L
(1 − Kbxf ) + µ p L

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Lockup condition

For a given vehicle, given proportioning and µ, you can estimate


that the front tires will lock first if,
This says that the rear wheels
a a can support a higher level of
g < g ‘deceleration’ than the front
  f  r – so the front will lock first.

So, if you are trying to decelerate at a level “a”, the determination


of impending lockup is based solely on:
a) geometry (l1, l2, L), and
b) material (µ) properties,
in combination with a specified distribution of braking effort
represented by Kbxf.
ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering
Vehicle System Dynamics and Control The University of Texas at Austin
Optimum lockup condition

The maximum deceleration potential is achieved when the front


and rear brakes lock simultaneously,
a a
g = g
  f  r
Plot each of these quantities versus the proportioning factor.
a a
g = g
  f  r
a
Highest deceleration level
g a a
possible at this intersection.
g > g
  f  r a a
g < g
  f  r Fbxf = K bxf ⋅ Fbx
rear
 
front total
front
braking braking
K bxf percent
ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering
Vehicle System Dynamics and Control The University of Texas at Austin
Example: Lockup Conditions – Light Truck
These graphs are also useful for
looking at the effect of having a
loaded vs. unloaded vehicle
(see Wong). Wong (Fig 3.50)

With an unloaded vehicle, you might


need to distribute braking more
unequally to lock the front wheels
first. Compromise
loaded
Distribution to achieve a certain
deceleration can be very different. unloaded

Contrast the example with a passenger


vehicle in Wong (Fig. 3.51), where the
difference is not as significant. K bxf

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Example: Problem 3.5 in Wong

A passenger car weighs 20.02 kN (4500 lb) and has a wheelbase of


279.4 cm (110 in.). The center of gravity is 127 cm (50 in.) behind
the front axle and 50.8 cm (20 in.) above ground level. In practice,
the vehicle encounters a variety of surfaces, with the coefficient of
road adhesion ranging from 0.2 to 0.8 and the coefficient of rolling
resistance of 0.015. With a view to avoiding the loss of directional
stability on surfaces with a low coefficient of adhesion under
emergency braking conditions, what would you recommend
regarding the braking effort distribution between the front and rear
axles?

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Example: Problem 3.5 in Wong (solution)

Parameters:
h = 50.8 cm
L = 279.4 cm
L1 = 127 cm
fr = 0.015

µ varies 0.2 to 0.8

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
Example: Problem 3.5 in Wong (Matlab)

Here is how you can solve for Kbf for different mu conditions.
% Solve for Intersections of Brake Lockup Curves
% function f = brake_lockup(kbf)
global mu L1 L2 L h fr global mu L1 L2 L h fr

L = 279.4; L1 = 127; L2 = L - L1; h = 50.8; % all dimensions in cm agf = (mu*L2/L + kbf*fr)/(kbf-mu*h/L);


agr = (mu*L1/L + (1-kbf)*fr)/(1-kbf+mu*h/L);
fr = 0.015;
f = agf - agr;
% Vary the surface coefficient of friction, mu

kbfo = 0.5; % guess


>> solve_brake_lockup
mu = 0.0;
sprintf('mu vs. kbf')
ans =
for i = 1:4,
mu = mu + 0.2;
mu vs. kbf
kbf = fzero('brake_lockup',kbfo);
disp([mu kbf])
0.2000 0.5845 To assure lockup
end
occurs on front first
0.4000 0.6209
for all of these
0.6000 0.6573 conditions, select
kbf for highest mu
0.8000 0.6936 value.

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin
References
• R. Bosch, Driving Safety Systems, SAE, 1999.
• T. Gillespie, Fundamentals of Vehicle Dynamics, SAE, 1992.
• H. Heisler, Vehicle and Vehicle Engine Technology, SAE, 1999.
• D.B. Maciuca, “Brake Modeling and Control”, Ch. 12 in
Intelligent Vehicle Technologies, SAE, 2001.
• Steeds, W., “Mechanics of Road Vehicles,” Illiffe and Sons,
Ltd., London, 1960.
• J.Y. Wong, Theory of Ground Vehicles, Wiley-Interscience,
2001.

ME 360/390 – Prof. R.G. Longoria Department of Mechanical Engineering


Vehicle System Dynamics and Control The University of Texas at Austin

Vous aimerez peut-être aussi