Downloaded from SAE International by Imperial College London Alumni, Wednesday, June 11, 2025

SAE TECHNICAL 2016-36-0164

PAPER SERIES E

LAP TIME SIMULATION OF FORMULA SAE VEHICLE WITH

QUASI-STEADY STATE MODEL
RODRIGO PASIANI COSTA
ROBERTO BORTOLUSSI
 Downloaded from SAE International by Imperial College London Alumni, Wednesday, June 11, 2025

2016-36-0164

Lap Time Simulation of Formula SAE Vehicle With Quasi-steady State

Model

Rodrigo Pasiani Costa, Roberto Bortolussi

Centro Universitário FEI

Copyright © 2016 SAE International

Abstract processing power, offers several vehicles and tracks models for
download and is a free software.
Lap time simulator (LTS) is a simulation tool used by racing teams to
estimate lap time for a given vehicle setup, such as in Formula 1, Most commercially available LTS packages use quasi-steady state
Indy and Le Mans. This work employs a Formula SAE (FSAE) method, as RaceWare (Vehicle Dynamics Performance Ltd),
oriented LTS called OptimumLap, which uses a quasi-steady state Dynamic Response (Pressplay Ltd), LTS (Milliken Research
method, to simulate prototype cars from Centro Universitário FEI. A Associates), PiSim (Pi Corporation) and RaceSim (DATAS). Most of
MATLAB code has been created to simulate acceleration race and non-commercial LTS packages developed by universities also adopt
compare this result to OptimumLap. The latter has also been used to quasi-steady state method, as developed in Michigan University
simulate skid pad and autocross events. RS8 simulation results have (USA) and University of Brescia (Italy) [3].
been compared to Michigan 2014 FSAE competition elapsed times
for model validation. After validation of RS8 model, the new Steady state method, extensively discussed in [4], considers vehicle
prototype RS9 was simulated in order to predict its performance; through a constant radius corner at constant longitudinal speed.
finally the results and the behavior of these two vehicles were Introduction of braking and acceleration while cornering defines the
compared. method as quasi-steady state [5].

Introduction Model Definition

Circuit racing cars are designed to traverse laps of a circuit in “OptimumLap utilizes a quasi-steady state point mass vehicle model.
minimum time [1]. However, lap time varies according to driver’s […] The vehicle is able to accelerate and corner simultaneously as
skills and vehicle setup. well as decelerate and corner simultaneously” [5].

With modern computers, it is possible to model and predict the The track is divided into small segments, so maximum speed is
behavior of a vehicle without having to build a prototype [2]. A lap determined for each segment individually. As shown in Figure 1,
time simulator package can be used to find the fastest lap time for a LTS calculates the maximum achievable speed in each segment in
vehicle with a given setup. The setup can then be varied trying to three major steps: calculating corner speeds, then speed accelerating
reduce this time [3]. LTS also provides a better understanding of how out of the corners and at last the distance needed to decelerate for the
the vehicle reacts to parameter changes, allowing engineers to make corners. A smooth transition is applied in order to avoid unrealistic
decisions based on data rather than experience only [2]. leaps in vehicle speed [5].

RS8 is the prototype of Centro Universitário FEI that competed in

Formula SAE category for the past two years, this year succeeded by
RS9. Main difference is the aerodynamic package, including a size
reduction of the rear wing imposed by current regulation.

Quasi-steady State Method

“LTS packages tend to fall into two different categories, either those
that are commercially available or those that have been developed by
academic institutions or individuals for personal use (non- Figure 1. Vehicle speed calculation [5].
commercial). All but a few of these packages use a quasi-static
simulation approach [same as quasi-steady state]” [3]. Model Limitations

The program used in this paper calculates the lap time by quasi- For any simulation, it is essential to know the limitations of the
steady state method. OptimumLap (OptimumG ®) was chosen model. Any mathematical model represents reality under certain
because it was developed specifically for FSAE, requires little conditions – the boundary conditions – within a tolerance range.
According to [5], OptimumLap main limitations are:
Page 1 of 7
 Downloaded from SAE International by Imperial College London Alumni, Wednesday, June 11, 2025

1. Does not account for weight transfer (lateral or Accelerating Segment

longitudinal):
Longitudinal acceleration is limited by engine power or tires
a. No suspension effects; adherence [4]. Neglecting resistive forces, Eq. (1) presents Newton’s
b. No inertia; 2nd Law, and Eq. (2) describes the available tractive force on the
wheels.
c. Tire grip is a linear function.
2. Does not utilize a real tire model: 𝐹𝑥 = 𝑀𝑎𝑝 𝑎𝑥 (1)
a. The effects of camber, slip ratio and
𝑇𝑒 𝑁𝑒𝑓 𝜂𝑒𝑓
𝐹𝑥 = (2)
slip angle not taken into account; 𝑟𝑑

b. The effects of temperature and

Apparent mass of the vehicle (Map) is the combination of vehicle
pressures not taken into account; mass and equivalent mass of the rotating components. According to
[4], apparent mass of a typical small passenger car is 120% of the
3. Does not account for vehicle yaw:
vehicle mass in second gear. Chosen LTS does not take into account
a. Since there is no center of gravity (CG) rotating components, so vehicle mass (M) will be employed from
now on rather than Map.
or wheelbase, the vehicle does not have
the capacity to oversteer or understeer. Including resistive forces, movement is described by Eq. (3), while
4. Does not account for banking or grade on the tire traction limit can be determined by Eq. (4).

track: 𝑀 𝑎𝑥 = 𝐹𝑥 − 𝑅𝑟 − 𝐷𝐴 − (𝑀 𝑔 sin 𝜃) (3)

a. No increase or decrease in traction due
𝐹𝑥 = 𝜇𝑥 𝑁𝑇 (4)
to centrifugal forces or added weight
transfer; Not only for accelerating segment, but also for all other segments,
b. Does not take into account transient drag force is defined as in Eq. (5).

effects (i.e. damping or inerter). 𝐷𝐴 = 0.5𝜌 𝑣𝑥2 𝐶𝐷 𝐴𝑝𝑟𝑜𝑗 (5)

The effect of adding weight (from downforce only) and over Neglecting any grade – true condition for most FSAE tracks –
saturating the tire can be taken into account by utilizing the ‘Tire longitudinal acceleration is described by:
Load Sensitivity’ parameters in the vehicle model. Yaw inertia can be
neglected for autocross races according to [1]. 𝐹𝑥 − 𝑅𝑟 − 𝐷𝐴
𝑎𝑥 = (6)
𝑀

Software utilizes both vehicle and driver at 100% of their

capabilities, resulting in calculated times faster than real times. Considering uniformly accelerated motion, speed at the end of the
Method is useful, though, to determine which parameters are most segment is calculated by Eq. (7). At last, time to go through the
sensitive to lap time, guiding project modifications. [5] states an segment can be determined by Eq. (8).
accuracy of 10%, not rare within 5%, for lap times.
𝑣𝑥𝑇 = 𝑎𝑥 𝑇 + 𝑣𝑥0 (7)
Mathematical Vehicle Model
T = ∆𝑥⁄𝑣𝑥0 (8)
OptimumLap simplifies the vehicle model as shown in Figure 2.
Equations (1) to (16) have been taken from [5]. Downforce and Braking Segment
rolling resistance equations are not revealed by [5], and were
determined in MATLAB Simulation. Newton’s 2nd Law applied to braking ideally results in Eq. (9). Tires
also limit braking performance; however, all four tires generate
braking forces unlike the acceleration, as in Eq. (10).

M 𝑎𝑥 = −𝐹𝑥 − 𝑅𝑟 − 𝐷𝐴 − (𝑀 𝑔 sin 𝜃) (9)

𝐹𝑥 = 𝜇𝑥 𝑁 (10)

Neglecting grade and solving Newton’s 2nd Law for acceleration

yields:

−(𝐹𝑥 + 𝑅𝑟 + 𝐷𝐴 )
𝑎𝑥 = (11)
𝑀

Analogously to acceleration segment, speed at the end of the segment

and time to go through the segment are respectively determined by
Figure 2. OptimumLap vehicle model [5]. equations (7) and (8).

Page 2 of 7
 Downloaded from SAE International by Imperial College London Alumni, Wednesday, June 11, 2025

Corner Segment RS9 chassis is derived from RS8, and main difference between them
is the aerodynamic package, as seen in Table 1.
Equation (12) determines maximum cornering force.
Table 1. Vehicle models.
𝐹𝑦 = 𝑁 𝜇𝑦 (12)
Feature RS8 RS9 Unit
Mass (with fuel and driver) 255 252 [kg]
Equation (13) describes the lateral acceleration to which the vehicle
is submitted, generating lateral forces as in Eq. (14). Drag Coefficient 2.240 1.500 [ ]
Downforce Coefficient 3.016 2.240 [ ]
2 ⁄
𝑎𝑦 = 𝑣𝑥0 𝑅 (13) Projected Frontal Area 1.134 1.000 [m²]
Tire Dynamic Radius 0.226 0.226 [m]
𝐹𝑦 = 𝑀 𝑎𝑦 (14)
Rolling Resistance 0.015 0.015 [ ]
Longitudinal Friction 1.750 1.750 [ ]
Combining equations (13) and (14), and solving for speed, results in
Eq. (15). Then, combining equations (12) and (15) yields Eq. (16). Longitudinal Load Sensitivity 0.006 0.006 [kgf -1]
Lateral Friction 1.750 1.750 [ ]
𝐹𝑦 𝑅
𝑣𝑥0 = √ (15) Lateral Load Sensitivity 0.006 0.006 [kgf -1]
𝑀

𝑁 𝜇𝑦 𝑅
𝑣𝑥0 = √ (16) Track Models
𝑀

Analogously to acceleration segment, time is determined by Eq. (8). Acceleration track was modeled as a flat straight line 75 meters long.
Skid pad track was modeled as a flat circle with diameter 18.25
meters. Autocross track, same used for endurance event, has an
Modeling extension of 1070 meters and several corners to both sides. Complete
autocross track model was downloaded from [8].
Prototypes RS8 and RS9 have been modeled, together with three
racetracks: straight line for acceleration event, circular path for skid
pad and the autocross circuit, all representing competition held in Model Validation
Michigan in 2014. Further details of models can be consulted in [6].
Before performing any simulation of prototype RS9, the model of the
Vehicle Models well-known prototype RS8 was validated by simulation. Validation
was ensured by two ways: comparing OptimumLap results with
MATLAB simulation and with physical test.
Vehicle models were created based on data from simulations and
tests. For example, the aerodynamic coefficients were obtained from
computational fluid dynamics (CFD), while tire parameters were MATLAB Simulation
obtained from the tire consortium which tested several FSAE tires on
a test bench [7]. MATLAB has been chosen to validate OptimumLap results due to
the possibility to generate an entirely new code. Rolling resistance
Vehicle models include tire sensitivity to vertical load for both and downforce, not provided in [5], were determined by trying
longitudinal and lateral coefficient of friction, which are independent different equations found in literature and comparing MATLAB to
parameters. Tires used by both vehicles are Hoosier LC0. OptimumLap results.

Both prototypes are powered by a single cylinder 449 cm³ Yamaha A hypothetical drag race from standing start to top speed is enough to
naturally aspirated internal combustion engine, running on ethanol find missing equations. MATLAB simulation is the reference, and
E100. Engine torque and power curves were obtained from the acceptable deviation is 2%, based in [3].
dynamometer test, and are represented in Figure 3.
It was found that both rolling resistance and downforce were those
from [4], here in equations (17) and (18), respectively.

𝑅𝑟 = 𝑓𝑟 𝑀 𝑔 (17)

𝐿𝐴 = 0.5𝜌 𝑣𝑥2 𝐶𝐿 𝐴𝑝𝑟𝑜𝑗 (18)

From MATLAB simulation, tractive forces per gear and resistive

forces as function of speed are in Figure 4. The top speed of 104.67
km/h is represented by the intersection point of the lines “gear 5” and
“resistive force”.

Figure 3. Engine torque and power output for RS8 and RS9.

Page 3 of 7
 Downloaded from SAE International by Imperial College London Alumni, Wednesday, June 11, 2025

Physical Test

The reference time for model validation is FSAE competition held in

Michigan International Speedway in the year of 2014, where
Brazilian team Formula FEI competed with prototype RS8. The team
ranked 10th in overall result among 120 participants. Official results
are available in [9].

[5] states 10% accuracy for lap time, not rare limited to 5%. “Any
autocross time in excess of 145% of the fastest time will receive no
‘performance’ points” [10]. According to [9], the top eleven cars
were within the range of 10% in autocross. Hence, this paper
Figure 4. Tractive and resistive forces as function of speed, from MATLAB, considered the acceptable tolerance of 10% for lap time.
for RS8.
Comparison between real lap times and those calculated by
Gear shifting speeds reflect software strategy for gear shifting, which
OptimumLap is presented in Table 4.
is mandatory for acceleration performance. MATLAB code considers
maximum tractive force on tires as parameter for gear shifting. Table
Table 4. Lap time validation for RS8.
2 compares speed for shifting; MATLAB is the reference for
deviation. Deviation is within the margin of 2% and therefore
Race Real Time Simulation Error [%]
acceptable. [s] Time [s]
Acceleration 4.448 4.520 1.6
Table 2. Speed for gear shifting (RS8).
Skid Pad 5.031 4.390 -12.7
Speed [km/h] Deviation Autocross (Enduro) 57.232 57.010 -0.4
Gear
[%]
MATLAB OptimumLap
1-2 45.15 44.90 -0.6
2-3 61.45 60.76 -1.1 Both acceleration and autocross races represent reality within a
margin of 2%. Skid pad, however, presents a deviation of nearly
3-4 78.85 77.73 -1.4
13%, therefore beyond acceptance. Yaw inertia, neglected by
4-5 98.55 97.13 -1.4 OptimumLap, plays an important role in skid pad races. Added to
that, steady state cornering is ideally applied to skid pad, which
means constant speed and constant steering – whereas the driver must
correct ground corrugations and car imbalances in reality.
Calculating the top speed with different resistive forces allows
analyzing each resistive force separately, including the influence of
downforce in rolling resistance. Top speed was a perfect match under Simulation results for the skid pad must be analyzed carefully. Since
all conditions, as seen in Table 3, confirming equations (17) and (18). the yaw inertia of RS8 and RS9 are similar, comparison between their
times may be valid. Comparison between simulated lap times and
Table 3. Top speed for different resistive forces (RS8). real times must be avoided due to the large discrepancy.

Resistive Force
Speed [km/h] Deviation Results
[%]
MATLAB OptimumLap
None 133.08 133.08 0.0 Once quasi-steady state models have been validated with MATLAB
and physical tests, RS8 and RS9 performances were simulated in
Rolling 133.08 133.08 0.0
OptimumLap. All results in this chapter are from OptimumLap
Aerodynamic 106.43 106.43 0.0 software.
Rolling and 105.38 105.38 0.0
aerodynamic (1) Table 5 presents overall results, showing a faster RS9 versus an RS8
Rolling and 104.67 104.67 0.0 better in cornering and braking thanks to its larger rear wing. Key
aerodynamic (2)
performance indicators (KPI) have been adopted for this comparison.
Notes: 1 – neglecting downforce; 2 – including downforce
Table 5. Overall results.
MATLAB simulation validated following parameters:
KPI RS8 RS9 Unit

a. speed (and consequently RPM) for gear shifting; Highest speed 104.67 121.56 [km/h]

b. rolling resistance; Traction limit (force) 2455 2358 [N]

Traction limit (speed) 52.78 54,58 [km/h]
c. aerodynamic drag;
Time to 0-100 km/h 6.99 4.56 [s]
d. downforce.
Time to 0-100 m 5.48 5.30 [s]
Distance to 100-0 km/h 17.30 18.70 [m]
Downforce participation in rolling resistance was confirmed,
Downforce at 100 km/h 1456 954 [N]
increasing normal force generated between tire and ground.
Drag at 100 km/h 1141 690 [N]

Page 4 of 7
 Downloaded from SAE International by Imperial College London Alumni, Wednesday, June 11, 2025

In the acceleration event, the lighter and more streamlined (less drag)
RS9 defeated its predecessor, racing 0.11 seconds faster after a
standing start, reaching higher speed and slightly higher longitudinal Due to the greater braking capability of RS8, it spent more time
acceleration. Results can be seen in Table 6. accelerating and less time braking in comparison to its successor.
However, RS9 maximum speed was 12.23 km/h higher, and fuel
Table 6. Acceleration event results. consumption was lower. Achieving higher speeds in straights and
having less cornering power demanded more gear shifts: 60 for RS9
KPI RS8 RS9 Unit versus 54 for RS8.
Lap time 4,52 4,41 [s]
Highest speed 92,34 98,70 [km/h] Figure 5 shows RS9 speed around the autocross track for one lap,
after rolling start. Black arrowhead indicates finish line and direction
Gear shifts 3 4 [ ]
of travel.
Maximum longitudinal 0,860 0,862 [g]
acceleration

After a rolling start, RS8 took advantage of its larger rear wing to win
the skid pad contest by 0.05 seconds, as detailed in Table 7. Once
both vehicles have similar chassis and same tires, greater downforce
of RS8 was decisive in this match, allowing RS8 to develop higher
lateral acceleration. Both RS8 and RS9 ran in second gear.

Simulation presented the same value for lowest, average and highest
speed for each car, revealing that the full lap was done in steady state
maneuver. In Reality, driver must turn the steering wheel in order to
correct any asymmetry of the vehicle, as well as road corrugations, in
a way that the steady-state maneuver does not occur. Figure 5. RS9 performance on autocross track.

Table 7. Skid pad results for one lap with rolling start. Figure 6 allows one to compare the speed of both prototypes. RS9 is
faster in high speed sections, while RS8 excels in some corners.
KPI RS8 RS9 Unit There is a case, around 170 m, were RS9 needs to brake and regain
Lap time 4.39 4.44 [s] speed, while RS8 continues to gain speed. The driver must pay
Lowest speed 47.03 46.46 [km/h]
attention to this behavior required by new vehicle. The driver should
also notice the need to anticipate braking. In those cases where speed
Average speed 47.03 46.46 [km/h] exceeds 100 km/h for RS9, braking should be anticipated as much as
Highest speed 47.03 46.46 [km/h] five meters. Additionally, the team must pay attention to brake wear.
Maximum lateral 1.91 1.86 [g]
acceleration
Engaged gear 2nd 2nd [ ]

RS9 performed better than RS8 in autocross track, completing one

lap 0.27 seconds faster after a rolling start. An extended comparison
between the two prototypes is given in Table 8.

Table 8. Autocross results for one lap with rolling start.

KPI RS8 RS9 Unit

Lap time 57.01 56.74 [s]
% accelerating 72.21 69.12 [%]
% braking 27.37 30.23 [%]
Figure 6. Speed comparison on autocross track.
% coasting 4.36 3.94 [%]
Average speed 72.29 73.63 [km/h] Figure 7 shows RS8 bearing higher lateral accelerations than RS9
Highest speed 100.50 112.73 [km/h] around the corners of autocross track. This is true for peak lateral
acceleration, but not for all corners.
Fuel consumption 0.17 0.15 [kg]
Gear shifts 54 60 [ ]
Maximum lateral 2.38 2.27 [g]
acceleration
Maximum longit. 0.85 0.86 [g]
acceleration
Maximum longit. 2.79 2.53 [g]
deceleration

Page 5 of 7
 Downloaded from SAE International by Imperial College London Alumni, Wednesday, June 11, 2025

2. “OptimumLap Documentation,” Last modified 2012.

http://www.optimumg.com/software/optimumlap/optimumlap-
documentation/.
3. Siegler, B. “Lap Time Simulation for Racing Car Design” (PhD
diss., University of Leeds, 2002).
4. Gillespie, T. D, Fundamentals of Vehicle Dynamics,
(Warrendale: Society of Automotive Engineers, 1992).
5. OptimumG. OptimumLap v1.4: Vehicle Dynamics Simulation.
(2015).
6. Costa, R. P, “Simulação de Tempo de Volta de Veículo Fórmula
SAE com Modelo Quasi-steady State,” (MSc. diss., Centro
Universitário FEI, 2016).
7. Lotto, R. S. “Dinâmica Lateral: Construção e Validação de um
Modelo de Simulação Multicorpos de um Veículo FSAE
Usando Modelo Semi-empírico de Pneu,” (MSc. diss., Centro
Figure 7. Lateral acceleration comparison on autocross track. Universitário da FEI, 2014).
8. “OptimumLap: Track Database,” last modified 2015,
Figure 8 presents a comparison of lateral acceleration of both http://share.optimumg.com/tracks/?page=5.
vehicles. RS8 negotiates corners with up to 36 meters of radius faster 9. “Results and Awards,” last modified 2015,
than RS9. Newest vehicle is therefore faster around corners with http://students.sae.org/cds/formulaseries/results/.
radius greater than 36 meters. This seems to be the radius from which 10. “2015 Formula SAE® Rules,” last modified 2015.
the longitudinal performance becomes more relevant than lateral http://www.fsaeonline.com/content/2015-
when comparing these two prototypes. 16%20FSAE%20Rules%20revision%2091714%20kz.pdf.

Contact Information
[email protected]

[email protected]

Definitions/Abbreviations
Aproj projected frontal area [m²]

ax longitudinal acceleration [m s-2]

ay lateral acceleration [m s-2]

Figure 8. Lateral acceleration as a function of corner radius in autocross race.

CD drag coefficient [ ]
Summary/Conclusions
CL lift coefficient [ ]
OptimumLap, from OptimumG, has been chosen as LTS package. A
brief review of model definition and limitations was given. Vehicle DA aerodynamic drag force [N]
and track models were presented. A MATLAB model was created to
find two missing equations of motion and to validate longitudinal
Fx longitudinal force (acceleration if
dynamics. A physical validation was done with real lap times from
positive, braking if negative) [N]
Michigan, 2014, including acceleration, skid pad and autocross
events. Despite the limitations, quasi-steady state model proved to be
accurate enough for FSAE competition, except for the skid pad event, Fy lateral force [N]
where the yaw inertia and road corrugations play a major role.
Comparing former prototype vehicle of Centro Universitário FEI, fr coefficient of rolling resistance [ ]
RS8, with current one, RS9, the latter was faster than its predecessor
in acceleration and autocross events. On the other hand, RS8 has g acceleration of gravity [m s-2]
been faster in skid pad, although results from skid pad simulation
should be analyzed carefully for they present larger error. Lastly, LA aerodynamic lift force [N]
change in vehicle behavior was analyzed.
M vehicle mass [kg]
References
Map apparent mass of the vehicle [kg]
1. Casanova, D.; Sharp, R. S. and Symonds, P, “Minimum Time
Maneuvering: The Significance of Yaw Inertia,” International N normal force on four wheels [N]
Journal of Vehicle Mechanics and Mobility 34 (2000): 77-115,
doi: 10.1076/0042-3114(200008)34:2;1-G;FT077. Nef effective transmission ratio [ ]

Page 6 of 7
 Downloaded from SAE International by Imperial College London Alumni, Wednesday, June 11, 2025

NT normal force on tractive axle [N]

R corner radius [m]

rd dynamic radius of tractive tires [m]

Rr rolling resistance [N]

T Time [s]

Te engine torque [N m]

vx longitudinal speed [m s-1]

vx0 longitudinal speed at the beginning of

segment [m s-1]

vxT longitudinal speed at the end of

segment [m s-1]

ΔxA acceleration segment length [m]

ΔxB braking segment length [m]

ΔxC cornering segment length [m]

ηef effective powertrain efficiency [ ]

θ grade angle [rad]

μx longitudinal friction coefficient [ ]

μy lateral friction coefficient [ ]

ρ air density [kg m-3]

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or
by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE.
ISSN 0148-7191
Copyright © 2016 SAE International.
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The authors solely
responsible for the content of the paper.

Page 7 of 7