DE VELO PMENT MOTOR SP ORT

Simulation-based Development
Process for Motor Sport and
Series Production

© Porsche

AUTHORS The Porsche development centre in Weissach (Germany) is home to

both of the vehicle series production development as well as the
motor racing application and development departments. We can see
clear parallels when we look at the tasks and challenges involved in
the development of road vehicles and racing cars. Both areas only
Dipl.-Ing. Kai Fritzsche
is Development Engineer
have short development periods and limited resources available.
Performance LMP1 at Porsche
in Weissach (Germany).
These overlaps suggest there ought to be close collaboration in the
areas of method and tool development.

RETURNING TO THE FIGURE 1, has an electric motor at the

TOP TIER AT LE MANS front axle, giving it a temporary all-
Dipl.-Ing. (FH) Florian Strecker wheel drive with an additional power of
is Senior Engineer System In addition to making a major commit- 300 kW. However, there are upper limits
Dynamics at Porsche ment to GT and road racing vehicles, on the amount of electric and fossil-fuel
in Weissach (Germany).
Porsche Motorsport has been participat- energy released per lap. To ensure racing
ing in the Le Mans-Prototype-1 (LMP1) success, the focus is therefore on devel-
class of the FIA WEC racing series with oping overall vehicle efficiency.
the Porsche 919 Hybrid since 2014. The
WEC includes nine races per year: eight
VEHICLE DEVELOPMENT 2.0
6 h races and the 24 Hours of Le Mans.
The manufacturer vehicles in the LMP1 Systems are becoming increasingly com-
Dr.-Ing. Malte Huneke
category are hybrid vehicles for which plex both in road vehicles and racing
is Manager Performance LMP1 at
Porsche in Weissach (Germany). the regulations allow a great degree of cars. Due to the growing influence of soft-
technical freedom. The 919 Hybrid, ware on vehicles in both fields, the issue
60
of software development and application motor sport is the use of a driving simula- times can be evaluated with greater accu-
is gaining importance. However, this does tor with a motion platform. Simulator racy through offline simulations as the
not eliminate the need for comprehensive specific hardware and software are not variance of a human driver does not
hardware development. Chassis and drive regarded as core competencies and are apply in this case.
control systems help tap into all the therefore procured from a system supplier In-house development of tools has sev-
potential offered by the overall system. (VI-grade). However, the same vehicle eral advantages:
In the LMP1 racing car, new software model developed in-house and used in – Models and evaluation software serve
versions can also be developed and the remaining simulation tools is used as as an executable knowledge database
introduced between races. In contrast the vehicle model and integrated into the and communication platform for all
to hardware development, this brings simulator environment. The use of the users, allowing for a lively exchange
improvements to the vehicle without simulation models can therefore also be between all the departments involved
involving major lead times. For road extended to the area of subjective evalua- as well as between motor sport and
vehicles, it is also already technically tion by the driver. The driver is essen- series production developers. Custom-
feasible to retrospectively load new soft- tially serving as an evaluation tool, ised integration into the processes
ware versions into the customer vehicle. whereby subjective impressions are used ensures that they can be used effi-
However, new functions can only be in addition to objective key figures in ciently and with ease.
tested using model-based variant studies order to evaluate handling. The vehicle – In-house development significantly
due to the high number of vehicle, model model delivers reproducible, deterministic improves understanding of both tools
and software variants. behaviour in a safe but still sufficiently and vehicle handling. It provides the
realistic environment. This permits rapid opportunity for quantitative evaluation
development and meaningful preliminary and creation of objective and authori-
SIMULATION AS AN
synchronisation of connected functions tative key figures.
EXCHANGE PLATFORM
that are relevant for handling. A positive – Using methods, models and evaluation
One of the core competencies at Porsche side effect of using the same vehicle programmes jointly across projects and
during vehicle development is objective model for this type of application is the departments automatically creates the
evaluation of vehicle dynamics. Although increase in model quality across the mod- need for close collaboration and com-
commercial software packages are availa- el’s entire application area. Evaluating lap munication and therefore ensures a
ble for this task, Porsche prefers to use times on the driving simulator is particu- constant exchange of knowledge
tools developed in-house for simulation larly useful for issues relating to vehicle between the different areas in the
of vehicle dynamics and objectification stability limits. In contrast to offline sim- development centre. It also results in
of vehicle handling. All models and pro- ulation, a human driver cannot tap into common metrics and a common lan-
grammes developed in-house are pro- the entire theoretical potential of a vehi- guage for describing driving physics. In
vided centrally for motor sport and series cle. Vehicle handling close to the individ- addition, tool extensions and improve-
production applications. In areas where ual stability limit has a significant influ- ments will benefit all users, FIGURE 2.
competition is less relevant, it makes ence on the lap times a human driver – In-house models and evaluation tools
more sense from a strategic perspective can achieve. On the other hand, there make it possible to work in a very quick,
to use commercially available products. are questions that are better answered flexible and efficient way and to quickly
An enhancement for tools developed through offline simulation. The influence adapt to new requirements, different
in-house that is particularly helpful in of changed driving resistance on lap vehicle concepts and new systems.

2.0-l V4 turbo petrol engine

Li-ion battery

Exhaust energy recovery system

Electric machine at the front axle

FIGURE 1 Technical details of
Porsche 919 Hybrid (© Porsche)

ATZ worldwide 09|2017 61

DE VELO PMENT MOTOR SP ORT

Generally speaking, simulation permits a WHEN ONE RACE ENDS, races, but within a shorter period of time.
high level of test coverage and thorough ANOTHER BEGINS In the case of functional software
screening of new system approaches updates, the steps are as follows:
prior to setting up actual prototypes. The preparation for each LMP1 race can – Analyse the previous race as well as
The key factors for good integration be seen as an individual development the upcoming race from the previous
into development activities include a process. Hardware changes between races season to identify potential for
sensible choice of model complexity and are only possible to a limited extent, but it improvement for the next race.
an established catalogue of objective test is possible to develop and introduce new – Develop potential for improvement iden-
criteria and key figures that make it software functions, FIGURE 3. This is why tified in the simulation. To that end, the
possible to compare measurement and similar steps are followed as during the control system software is linked to vari-
simulation. series production process between the ous simulation environments as Soft-
ware-in-the-Loop (SiL). On the one
hand, recorded track data can be loaded
into the software to analyse direct
changes to the previous version based
on track data. On the other hand, the
software can be linked to the dynamic
simulation model to evaluate variants in
the closed control loop based on individ-
ual manoeuvres or lap time simulations.
– Test the newly developed software ver-
sion on various levels. New functions
are always tested on the Hardware-in-
the-Loop (HiL) test bench. This ensures
that the functionality is also provided
on the target hardware in the compiled
state. In addition, innovations are
tested on the driving simulator as part
of the race preparations. This ensures
that the desired influences on handling
FIGURE 2 Central simulation environment as project platform (© Porsche) are working correctly in the closed con-
trol loop. As a side effect, the test on
the driving simulator in a realistic envi-
ronment also checks the functional
robustness. The driving simulator also
makes it possible to develop the appli-
cation of the software based on the
data gathered and subjective driver
evaluation in rapid iteration steps.
Updates, which to a large extent are
directly linked to the hardware, will
continue to be released on drive test
benches as well as during test drives on
the test track where necessary.

BOOST STRATEGY ON THE

DRIVING SIMULATOR

The 2016 season race in Mexico is a con-

crete example of the process and the inte-
gration of the development tools as part of
the race preparations. The WEC was held
in Mexico for the first time in 2016, which
meant that none of the competitors had
any experience on the race track. This
made the importance of the simulation
tools more evident than in other races.
A key issue as part of the race prepara-
tion is the development of the boost strat-
FIGURE 3 Steps during the software development process between races (© Porsche) egy. It defines the permitted amount of
62
electrical energy and the power level for
the boost in each sector of the track. 3 projectors,
120 Hz WUXGA,
However, the entire amount of energy 215° field of view,
that can be released for each lap is lim- screen height: 2.4 m,
radius: 4 m
ited; in Mexico it was limited to 3.92 MJ. Cockpit of Porsche 919 Hybrid,
ECUs as HiL,
The additional power of 300 kW could digital audio and intercom
therefore be applied for a duration of
13.07 s per lap. With a lap time of roughly
1:26 min, this corresponds to 15.2 % of
the lap. It is always the goal of strategic
energy distribution to minimise the lap
time. However, since the powerful tempo-
rary all-wheel drive also influences han-
dling, the influence on driving dynamics
must also be considered.
The development of the boost strategy Motion platform: 1 operator workstation,
for a new track begins with the optimisa- - VI-grade DiM 4 engineer workstations,
(Driver in Motion) same working process as
tion of the energy distribution using the - 9 degrees of freedom on the race track
vehicle model. The necessary prerequi-
sites for these lap time evaluations on FIGURE 4 Profile of the Porsche LMP1 Team driving simulator with 3 projectors, cockpit, motion platform
the track typically involve high-quality and 5 workplaces (© Porsche)
measurement data that can be used to
derive the vehicle trajectory.
01:40
As an alternative, a driving line can be
Lap times from 6 h of Mexico Collective Test,
recorded with a factory driver on the driv-
sorted in ascending order,
ing simulator, FIGURE 4, due to the highly 2 vehicles per manufacturer
01:38
precise track models. This driving line is
then used as a basis for the first simula-
tions to prepare for the race and therefore
01:36
also to optimise the boost strategy. A first
iteration variant of the boost strategy is
available as the result of the optimisation.
It is then used as a starting point for 01:34
Lap time [mm:ss]

another session in the driving simulator.

The goal of the session is to fine-tune the
boost strategy in terms of vehicle balance 01:32
and driving quality. This is done on the
basis of a subjective evaluation by one or
more factory drivers in the simulator. Fur- 01:30
thermore, variants are developed that, for Competitor 1
example, include a strategy for the start of Competitor 2
the race or different strategies to provide 01:28 Porsche
assistance with a passing manoeuvre. As
the original variant is used as a basis for
comparison, all variants generated on the 01:26
driving simulator are evaluated once more 0 5 10 15 20 25 30
during an offline simulation. At the end of Total number of laps driven per vehicle [-]
the process, relative lap time differences FIGURE 5 Lap times from the “6 h of Mexico” 2016 Collective Test (© Porsche)
compared to the original variant without
any driver influence are available.
The result of this race preparation pro-
cess is a high-quality base strategy. The tool, it has a positive side effect on new The results show that, thanks to the
time available during free practice can tracks in particular as the drivers are procedure described above, the lap times
be used effectively to fine-tune the familiar with the track layout and have by the Porsche vehicle were already more
remaining unknown variables. This may already driven the vehicle virtually on consistent and faster during the first free
include the response to weather condi- the new track. The FIGURE 5 illustrates practice than those of the competitors.
tions, the interaction between tyres and the lap times from the flying laps by the Porsche ends its LMP1 engagement with
asphalt or the response to traffic condi- three LMP1 manufacturers in the first the 2017 season, but will also use the
tions. Although the driving simulator is session of the “6 h of Mexico” race (col- simulator technology in future motor-
not typically used as a driver training lective test) in ascending order. sport projects.
ATZ worldwide 09|2017 63