Rally Dynamics Blog
Rally Dynamics Blog
Rally Dynamics Blog
For the last two months what has been dominating my radar screen has been adapting ChassisSim
for WRC. The tarmac bit was easy but the tough bit has been adapting ChassisSim to run on dirt
and ice. This has been a job that has been challenging yet incredibly informative all at the same
time. The challenging bit has been resolving why you have to run well into the post stalled region of
the tyre and then resolving how to stay there. This is what we'll be discussing in this article.
Let me state from the get go right now I am not pretending to be an expert on this. If truth be told
I'm actually writing this article more for me than you at this point so I can start to get some things
straight in my head. That being said I've learnt a lot on the way so if you are involved with rallying
or have any interest as to what happens when a car goes sideways and then read on. Hopefully we
can all learn something in the process.
So here is the question why do you want to go sideways in a rally car? For all of us that have been
involved in tarmac/circuit racing this is a cardinal sin. It looks incredibly impressive but when it
comes to tarmac/open wheel racing we all know it's a guaranteed way to kill your speed. The
answer to this question lies in what the tyre is doing.
The answer as to why we want to go sideways on dirt and ice comes down to the slip characteristics
of the tyres. To really hammer home the point let's illustrate this graphically. A typical force vs slip
angle characteristic for a road racing tyre is shown below,
When you are on dirt and ice the tyre characteristics are a totally different ball game. When I was
doing my literature search I came across two excellent thesis. These where Michael Croft-White's
thesis from Cranfield University "Measurement and Analysis of Rally Car Dynamics at High
Attitude Angles" and a Thesis from Stanford university entitled - "Dynamics and control of drifting
in automobiles" by Rami Yusef Hindiyeh. The upshot from both of these thesis is that when you are
post stalled the grip doesn't drop off that much. In particular White's thesis did some basic tyre
modelling from a beta sensor he had developed. This is presented in Fig - 2
The key to note is what is happening in the post stalled region. Looking at slip angles well in excess
of 20 deg the grip has only dropped of by 10%. This is significant because there is grip to be had in
the post stalled region.
If the drop in post slip grip is mild the reason there is grip is because of what happens with the
longitudinal forces of the tyre at large slip angles. This is illustrated in Fig-3 along with the
equations of motion with the car,
Fig-3 - Free Body Diagram of the forces acting on the race car.
All tyre forces are applied along the slip angle line. FXF and FXR are the sum of all the longitudinal
forces at the front and rear respectively. Longitudinally this will not have a huge impact. As we'll
see shortly it has big ramifications laterally. This is particularly apparent at large slip angles.
Also just a note about small angle assumptions. Strictly speaking they are only apply to about +/-
100. However for practical calculation purposes we can stretch this to about 200. Let me illustrate
what I mean. In radians 200 is 0.349. The sin of 200 is 0.342. The cosine of 200 is 0.94. Yes we
sacrifice a little bit of accuracy longitudinally but the sine of the angles are still very close.
Consequently the equations we are about to present still workout. The other option is to include the
sine and cosine terms. While it is fully accurate, the problem is you start to lose any perspective
about what the maths is telling you. Also in rallying it is rare to see a side slip greater than 30 deg.
While this is not ideal we are certainly not in fantasy land.
Also to simplify things we have also lumped in the lateral forces here as well. Using small angle
assumptions it may be concluded;
where;
Vy sideways velocity.
Vx forward velocity.
r yaw rate.
Resolving forces and moments from Fig 1, the differential equations of the race car become;
2 4
mt (V x′ + V y r ) = F XR + F XF − ∑ (δ + α i ) ⋅Fyi − ∑ α i ⋅Fyi − 0.5ρV 2 CD A (7)
i =1 i =3
α1 + α 2 ⎛ α +α4 ⎞
mt (Vy′ + Vx r ) = FYF + FYR + (δ + ) ⋅ FXF + ⎜ 3 ⎟ ⋅ FXR (8)
2 ⎝ 2 ⎠
⎛ α +α2 ⎞ ⎛α +α4 ⎞
I z ⋅ r ′ = a ⋅ ( FYF + ⎜ δ + 1 ⎟ ⋅ FXF ) − b ⋅ ( FYF + ⎜ 3 ⎟ ⋅ FXR ) (9)
⎝ 2 ⎠ ⎝ 2 ⎠
Here,
mt = Total mass of the car.
Iz = the rotational inertia of the car
Equations (3) – (9) describe everything on how the race car will behave. The thing to note here is
the longitudinal forces. To reiterate they are applied on the slip angle line of the tyre. So at this
point you might be thinking so what? The key lies in the lateral components of the longitudinal
forces. We have,
α1 + α 2
FYF _ FXF = (δ + ) ⋅ FXF
2
(10)
⎛α +α4 ⎞
FYR _ FXR =⎜ 3 ⎟ ⋅ FXR
⎝ 2 ⎠
Here we have FYF_FXF is the lateral force at the front induced by the front longitudinal forces and
FYR_FXR is the lateral force induced by the rear longitudinal forces.
Where things get really interesting is what happens when the slip angles go up. Let's illustrate this
with some numbers. Let's consider a typical AWD rally car that weighs in at 1300kg. Some
performance numbers are illustrated below,
Parameter Value
Car mass 1300kg
Cornering g 1
Peak slip angle - tarmac 60
Peak slip angle - dirt 160
CdA 1.1
Cornering speed 108 km/h
I realise the cornering g on dirt will be less than tarmac but let's keep these the same for the time
being. I want you to get a feel for the magnitude of the numbers.
So balancing the speeds and assuming front and rear slip angles to be the same we have
1
FXT = ρ ⋅ V 2 ⋅ C D A + α p ⋅ mt ⋅ a y (11)
2
Here we have,
Fxt = Total longitudinal force applied (N)
p = air density (kg/m3)
V = Car speed (m/s)
ap = Peak slip angle in radians
mt = total car mass
ay = Lateral acceleration in m/s2
I realise this is not strictly accurate but it is in the ball park and I'm doing this so you get a feel for
the numbers. Crunching the numbers for the tarmac and dirt modes the results are presented in
Table - 2
Table - 2 - Numbers for the balanced longitudinal forces in tarmac and dirt mode.
So in Tarmac mode we have about 20.7 kgf of lateral force produced by the applied longitudinal
force. In dirt mode this jumps to 118.7 kgf. While the analysis is incredibly over simplified it rams
home the rally observation that on dirt the engine force is a significant part of your corner grip.
So how do we determine that this is viable or not? Enter what I will term the Drift feasibility
equation. Let's illustrate this situation graphically. This is shown in Fig -4,
As can be seen from Fig-4 we have two equal forces acting laterally and longitudinally. I will term
this force R(α). Both of these components will have lateral components. Let me set R(α) such that,
So that we are clear I am slaving the force R(α) to the force vs slip angle equation that we all know
and love. However I'm still keeping it in traction circle limits so that we don't enter Fantasy Land.
So the optimum slip angle will be given by deriving equation (14) as a function of slip angle. Using
the product differential rule it is found that the optimum slip angle that will produce the most lateral
grip will be given by,
∂Fy ∂R
= 0 = R(α ) ⋅ (cos(α ) − sin (α )) + ⋅ (cos(α ) + sin (α )) (14)
∂α ∂α
Equation (14) is the drift feasibility equation. This won't necessarily tell you the optimum slip angle
you need to be at for drifting. However it will tell you if your tyre can actually do it. As a case in
point consider Fig 5a which is a road course tyre and Fig 5b which is a rally tyre. Evaluating
equation (14) for both of these curves show that Fig 5a has an optimum slip angle of 6.20 and Fig 5b
has an optimum slip angle of 160. Also a tip for young players here. Do this numerically. List out
R(α) and the subsequent derivatives. If you try and do it analytically you'll drive yourself nuts. This
is the first step in seeing if it is worth your while to go sideways or not.
So now that we have established if it's viable or not to go sideways we know need to nail down at
what angle we need to go sideways at. Remember we are drifting on dirt and ice not just because it
looks impressive but we are doing this to get grip. The answer lies in the lateral grip front and rear.
Let's put some maths to this. To simplify things a little bit let's use the bicycle equations of motion
for the front and rear slip angles. This is presented in equation (15)
a ⋅ r + Vy
αF = δ −
Vx
a ⋅ r + Vy
δ = αF + (15)
Vx
b ⋅ r − Vy
αR =
Vx
Here αf and αR are the front and rear slip angles. The front and rear lateral forces taking into
account both the forces due to slip angle of the tyre and the longitudinal forces are,
FYF = C F (α ) ⋅ (Fm1 + Fm2 ) + (δ + α F ) ⋅ FXF
⎛ a ⋅ r + Vy ⎞
= C F (α ) ⋅ (Fm1 + Fm2 ) + ⎜⎜ 2 ⋅ α F + ⎟⎟ ⋅ FXF (16)
⎝ Vx ⎠
FYR = C R (α ) ⋅ (Fm3 + Fm4 ) + (α R ) ⋅ FXR
Where things get really interesting is taking the derivative with respect to slip angle of equation(16).
Doing the differentiation we see,
∂ (FYF ) ∂ (C F (α F ))
= ⋅ (Fm1 + Fm2 ) + 2 ⋅ FXF
∂α F ∂α
(17)
∂ (FYR ) ∂ (C R (α R ))
= ⋅ (Fm3 + Fm 4 ) + FXR
∂α R ∂α
In order to be worth your while to drift The differential of the front and rear force curves must be
greater than zero. This is where the grip is and the reason the grip is there is as the slip angle
increases you will actually be producing force you can use. It's the reason that you see Sprint cars
on an oval hanging the tail out because that is where the grip is. If your car is rear wheel drive, the
last bit of equation (17) applies. If your car is front wheel drive the first bit of equation (17) applies.
If you are all wheel drive then both come into play. For rallying, equation (17) outlines the appeal
of all wheel drive.
So what is the procedure to determine the slip angle that you should be drifting to satisfy equation
(17). Firstly you start by choosing a corner speed and looking at the peak curvature you want to
corner at. You then nominate the factor of grip you want to maintain at the rear. The crux of this is
we want to maintain equilibrium both laterally and longitudinally. Keeping the slip angles the same
front and rear we have the following,
⎛1 ⎞
FXR = tsp R ⋅ ⎜ ⋅ ρ ⋅ V 2 ⋅ C D A + α R mt ⋅ V X ⋅ iR ⎟
2
(19)
⎝2 ⎠
Here tspR is the torque split at the rear and FXFR is the factor of rear longitudinal tyre force we want
to contribute to the lateral grip. Putting equation (19) into (18) yields the following relation for the
rear slip angle we are after,
0.5 ⋅ ρ ⋅ C D A F
αR2 + ⋅ α R − XFR = 0 (20)
mt ⋅ iR tsp R
The solution of equation (20) will give you a reference check. You are then going to go through an
iterative process to see if this makes sense. In particular if it is achievable through the slip angle
curves you have. The other thing to check is the load transfer so you have the traction. The
limitation will be the inside rear tyre loads. You will then check equation (17) and if all this adds up
you have equilibrium. When this all checks out you have determined the rear slip angle and side slip
angle you should be drifting at.
What all these equations tell you is that drifting to improve vehicle grip is only viable in low grip
situations. Let's re inspect equation (17) but this time let's do it through the lens of load transfer. As
a rough rule of thumb your tyre loads for a given aero load and load transfer are given by,
wdf ⋅ mt ⋅ g awf ⋅ C L ⋅ A ⋅ 2 ⋅ ρ ⋅ V
1 2
prr ⋅ mt ⋅ a y ⋅ h
TL1 = + +
2 2 tm
wdf ⋅ mt ⋅ g awf ⋅ C L ⋅ A ⋅ 2 ⋅ ρ ⋅ V
1 2
prr ⋅ mt ⋅ a y ⋅ h
TL2 = + −
2 2 tm
(21)
wdr ⋅ mt ⋅ g awr ⋅ C L ⋅ A ⋅ 2 ⋅ ρ ⋅ V
1 2
(1 − prr ) ⋅ mt ⋅ a y ⋅ h
TL3 = + +
2 2 tm
wdr ⋅ mt ⋅ g awr ⋅ C L ⋅ A ⋅ 2 ⋅ ρ ⋅ V
1 2
(1 − prr ) ⋅ mt ⋅ a y ⋅ h
TL4 = + −
2 2 tm
Where the downforce is not significant what will limit you will be the inside front and rear tyres
unloading. Consequently your ability to apply the longitudinal forces you need to ensure equation
(17) is greater than or equal to zero will be limited. Strictly speaking you could channel all the
longitudinal force to the outside rear wheel but you will have a destabilising moment due to the
tractive force trying to destabilise the car.
The last topic to touch upon is what does the racecar stability look like in the post stalled drift zone.
As discussed in some of my previous articles an excellent tool to look at this which is the stability
index. This can be written as,
∂FYF ∂F ⎛ ∂F ∂F ⎞ r
a⋅ − b ⋅ YR − ⎜ a 2 ⋅ YF + b 2 ⋅ YR ⎟⋅
∂α ∂α ⎝ ∂α ∂α ⎠ VX
SI = (22)
⎛ ∂FYF ∂FYR ⎞
⎜ + ⎟ ⋅ wb
⎝ ∂α ∂α ⎠
Inspecting equation (17) and putting it into equation (22) we will have still have some measure of
stability. However it will be much more marginal. This is because the slope of the force vs slip
angle curves are much smaller. The applied longitudinal forces are the dominant terms. The
combination of equations (17) and (22) mean that if you are sliding in a rear wheel drive car you
have no option but to keep the power applied. This was also confirmed in Hindiyeh(2) where his
controller uses engine force as an integral part of his drift controller.
Lastly to show this isn't just theory the beginnings of this have been incorporated into ChassisSim.
Find an example of a predictive rally simulation done in real time controlling the car in the post
stalled zone of the tyre,
Fig - 6: Predictive rally sim from ChassisSim controlling the car in the post stalled zone.
The first trace is speed, the second is steered angle and the third trace is throttle. However the real
traces are the fourth and fifth traces that show front and rear slip angles. The stall angle for this tyre
is 60 . The front slip angles are in the order of 4-50. However you can see the rear slip angles are in
the order of 100 and they are being controlled. I should add the car model needs a bit of work and
there is still some work to do. However this feature is on it's way.
In closing the vehicle dynamics of drifting are an exceptionally interesting field of vehicle
dynamics. The thing that dictates why you want to drift is what happens to the tyres on dirt and ice.
Here the slip angle curves drop of moderately in the post stalled region of the tyre. This makes it
viable to drift and we can readily calculate where we need to be drifting. While this is certainly not
the last word on the vehicle dynamics of rallying I trust what I have given you is the mathematical
framework to put some numbers to this.
References
1) "Measurement and Analysis of Rally Car Dynamics at High Attitude Angles" - Michael Croft-
White - Cranfield University 2005
2) "Dynamics and control of drifting in automobiles" - Rami Yusef Hindiyeh - Stanford University
2013
3) J.Y Wong, Theory of Ground Vehicles, Wiley and sons, 1978