ISSN (Online) 2192-9076

WORLDWIDE

06 June 2019 | Volume 121

CHASSIS

Integration from AWD to Steering

DUAL-CLUTCH TRANSMISSIONS SURFACES NVH DEMANDS

with Flexible Oil Management for an Innovative Interior on Wheel Hub Motors
for Variable Service of Future Passenger Cars of Compact Electric Vehicles

/// INTERVIE W Ralf Göttel [Benteler] /// GUEST COMMENTARY Christoph Bittner [Porsche]
 www.itk-karriere.de/en

10x in Germany | 4x worldwide

ADRENALINE OFF
THE RACETRACK
Hard deadlines, strict regulations and complex
functionality sound good to you?

In this case, you will fit right in! Our engineers at ITK are involved from the beginning of
the development lifecycle for race cars until the very end, extracting the last millisecond out
of the vehicle. We are looking for motivated and talented developers with the expertise to
support our team in the development of individual software functions.

For application and further information visit our website www.itk-karriere.de/en or

www.itk-engineering.de/en/industries/motorsports

Testing Takes Priority
Dear Reader,
The question is not whether automated
driving is coming, but when and where.
This was one of the key themes of the
ATZlive Conference “Automated Driving
– From Driver Assistance to Autonomous
Driving” which was held on April 2 and 3,
2019 in Wiesbaden (Germany). Last year’s
event focused primarily on automated
driving on SAE level 3, while this year
participants looked in more detail at
levels 4 and 5 and discussed what will
happen in the field of automated driving
in 2025 and beyond. The commercial
vehicle sector seems to be the main driv-
ing force behind developments because
“automated driving is primarily an issue
for commercial vehicles,” as Dr. Thomas
Dieckmann, Head of Pre-development at
Wabco, explained in his keynote speech.
He believes that driverless technology
will be used selectively during the first
half of the 2020s, in particular in China.
Hermann Meyer, Vice President Smart
City Solutions at Continental, also high-
lighted the fact that China is playing
the central role in developing intermodal
transport and automated driving.
The increasing automation of driving
functions presents new challenges for
the chassis. At the same time, completely
new vehicle concepts will be needed for
the mobility of the future, in particular
in urban areas. This is the main theme
of the new international ATZlive Confer-
ence “Vehicles of Tomorrow” which takes
place on November 19 and 20, 2019 in
Frankfurt/Main (Germany). The intelli-
gent chassis must be connected, electri-
fied, and automated in order to be fit for
the future. These subjects will be dis-
cussed by participants at the 10th Inter-
national Munich Chassis Symposium
“chassis.tech plus” on June 25 and 26.
They will also consider the question of
whether the chassis will continue to play
a key strategic role after 2025 or whether it
will become just an ordinary component.
ATZ worldwide 06|2019
EDITORIAL
Many experts are unwilling to hazard
a guess as to when SAE level-4 and -5
functions will become widely available,
but at the Automated Driving Conference
German and Japanese carmakers agreed
on the timeframe for the market launch
of level-3 systems: Masayuki Soga, Chief
Professional Engineer in the Vehicle
Dynamics Control division at Toyota,
predicted in his closing keynote speech
that level 2+ would be available from
2020 onward. This corresponded with
the views expressed by experts from
German companies. A well-informed
participant on the fringes of the confer-
ence told us that “there will definitely
not be a rush to start production of auto-
mated driving functions in Germany or
the rest of Europe – testing takes priority.”
This is where the chassis and the auto-
mated driving functions come together
at a central point.
I hope you enjoy reading
this issue of ATZ.
DATALOGGER
FOR THE HIGHEST REQUIREMENTS
Dr. Alexander Heintzel
Editor in Chief
▸ Datalogger for fleet testing and
endurance testing worldwide
▸ Ruggedized hardware for
extreme conditions
▸ Maximum reliability of
data acquisition
▸ Proven use in hybrid and
electric vehicles
SENSOR+TEST, Nuremberg
25.06. - 27.06.2019
Hall 1, Stand 415
3
www.IPETRONIK.com
C ONTENTS IS SUE 0 6 | 2019

r
le
n te
Be
©

CHASSIS

Integration from AWD to Steering

Electric powertrains and automated
driving are bringing about changes in 14 Introduction
Michael Reichenbach
chassis design. The role of the chassis
is to integrate the functions of the power-
train, bodywork, and steering safely and 16 The Electric All-wheel Drive in the Audi e-tron
Michael Wein, Christian Graf, Sebastian Strasser, Stephan Poltersdorf
reliably. Electric all-wheel drive systems [Audi]

allow the torque to be distributed between

the axles more dynamically than has
INTERVIEW
been the case in the past. At the same
© Uwe Lewandowski | ATZ

time, for the purposes of lane guidance 22 safety

“With our crash expertise we can bring
features into the chassis”
during automated driving it must be clear Ralf Göttel [Benteler]

whether the car or the driver is in control

of the steering wheel.

26 Steering Functions for Automated Driving

Robert Fuchs, Tsutomu Tamura, Maxime Moreillon [Jtekt]

4
 © Magna

© YFAI

36 Variable maintenance intervals allow for cost
and weight reductions. Magna Powertrain and BMW
have designed a flexible lubricant management sys-
tem for automatic transmissions with this in mind.
48 Kostal and YFAI have cooperated to develop
joint high-tech solutions for smart surfaces which
combine the HMI and interior design and exploit
all the available synergies.
© OVGU
66 Wheel hub motors are demanding in terms
of acoustics. The University of Magdeburg is
using optical vibration analysis to improve the
NVH characteristics of external rotor motors.

CON T E N T S DE V E L OP MEN T R E S E A RCH

3 Editorial TRANSMISSIONS COMFORT | NVH

36 Flexible Oil Management 66 NVH in Electric Mobility –
for Dual-clutch Transmissions Vibration Analysis Using a Derotator
NEWS Martin Bahne, Andreas Wiesebrock Fabian Duvigneau, Christian Daniel,
6 People + Companies [Magna Powertrain] Sebastian Koch, Elmar Woschke
7 Impulses [University of Magdeburg]
42 High-speed Electric Drive Unit
for the Next Generation of Vehicles ELECTRIC VEHICLES
IN T HE S PO T L IGH T Mathias Deiml, Tommie Eriksson, 72 Battery Pack Housing for
Matthias Schneck, Antoine Tan-Kim [AVL] Electric Vehicles Made by
8 Robo-cars – Breach of a Taboo Laser Beam Welding
Andreas Burkert HMI Achim Kampker, Georg Bergweiler,
48 Surface Innovations for Falko Fiedler, Ansgar Hollah
Interiors of Future Vehicles [RWTH Aachen University]
CO V E R S T ORY Dirk Blomeyer [YFAI],
Anna-Lena Schulte-Gehrmann [Kostal]
CHASSIS CONF ERE NCE RE POR T
14 Integration from AWD to Steering ASSISTANCE SYSTEMS
52 Exploring Corner Cases for 78 Automated Driving – The Big Picture
Camera-based Driving Functions – Mathias Heerwagen
NEWS High-fidelity Visualization in
32 Products Simulation
Martin Herrmann [IPG Automotive], S ER V ICE
Chris Hoyle [rFpro]
80 Imprint, Scientific Advisory Board
FILTERS 81 Preview
58 Air Quality of Vehicle Interiors
in Highly Polluted Cities
Ralf Mönkemöller, Sarah Ristau [Paragon]
GUEST
COMMENTARY
82 Mastering
Complexity
Christoph Bittner
© Porsche

[Porsche]

COVER FIGURE © Audi

ATZ worldwide 06|2019 5

 NE WS PEOPL E + C OMPANIES

Daimler Trucks | Majority Stake in Torc Robotics Aquired

Daimler Trucks and Torc Robotics, a pioneer in the field of autonomous
driving solutions, are entering into a partnership to commercialize highly
automated trucks (SAE level 4) in the USA. Daimler Trucks and Buses Hold-
ing Inc., a subsidiary of Daimler AG, is acquiring a majority stake in Torc
Robotics. The two companies plan to establish a connection that goes far
beyond the normal business relationship between a manufacturer and a
supplier. “Torc is not a start-up, but one of the world’s most experienced
© Daimler

companies in vehicle automation,” said Roger Nielsen, CEO of Daimler

Trucks North America. Under the agreement reached between the two
Michael Fleming (l.), CEO of Torc Robotics, and Roger Nielsen, companies, Torc will remain a separate entity and will retain its name,
CEO of Daimler Trucks North America team, existing customers, and facilities in Blacksburg, Virginia (USA).

Etas | Joint Venture with National Instruments

Etas GmbH and National Instruments have signed an agreement to jointly design, build, and
service pre-integrated Hardware-in-the-Loop (HiL) systems. By pooling their resources, the
two companies aim to improve the testing and validation of software in automotive electron-
ics, including Electronic Control Units (ECUs) and sensors, in order to fulfill customers’ cur-
rent and future requirements. Both partners have decades of experience in the automotive
industry. The objective of the partnership is to meet customers’ needs in the rapidly evolving
automotive sector, which is driven by electrification and driver assistance systems. Each

© Etas
company will own 50 % of the joint venture, which will have its headquarters in Stuttgart
(Germany). The transaction is subject to the usual conditions, including regulatory and anti- Leandro Fonseca (l.) is Managing Director, Hans-Peter Duerr
trust approvals, and is expected to be completed in June 2019. from Etas is the Technical Director of the joint venture

Akasol | New Headquarters and Test Center

Following a period of ongoing growth, the battery
technology company Akasol is building new head-
quarters in the southwest of the German city
of Darmstadt. The site, which covers an area of
20,000 m², will be home to a new 7000-m² central
administration building by mid 2020, as well as a
© Akasol
tests to be carried out on everything from an indi-
vidual cell to a complete battery system. In addi-
tion, a two-story production, assembly, and logis-
tics building will be constructed with a floor area
of 15,000 m². The company plans to invest a total
of around 50 million euros in the new facilities. The

state-of-the-art testing and inspection center. This production capacity of the new site will be based
will be equipped with the very latest testing equip- on future customer orders, which are currently
Akasol‘s new headquarters (Darmstadt/Germany) ment to allow mechanical, electronic, and electrical under negotiation.

ZF | Wabco | Takeover Agreement Signed

ZF has signed a binding agreement to acquire Wabco. The planned acquisition has
been approved by ZF’s management board and supervisory board and Wabco’s board
of directors. The two organizations will come together to form an integrated systems
provider for commercial vehicles. The combined company will have a revenue of
approximately 40 billion euros. Wabco supplies integrated braking, stability control,
air suspension, and transmission automation systems, together with aerodynamics,
telematics, and fleet management solutions. In 2018 the company reported a reve-
© ZF

nue of 3.3 billion euros. It employs around 16,000 people, including 2600 engineers,
in 40 countries around the world. The ZF headquarters in Friedrichshafen (Germany)

6
 Bosch | Changes to the
Board of Management
Dr Dirk Hoheisel, member of the board of manage-
ment of Robert Bosch GmbH, is retiring on June
30, 2019. He has been a Member of the Board of
Management since 2012, with responsibility for
systems integration within the mobility solutions
business unit, for the chassis systems control,
© Bosch
car multimedia, automotive electronics, and
automotive steering divisions, and for the two-
Dirk Hoheisel wheeler and powersports business unit. Hoheisel
has worked for Bosch for nearly 29 years. Harald
Kröger will be joining the board of management
of Robert Bosch GmbH on July 1, 2019, where he
will succeed Hoheisel. The 51-year-old electrical
engineer is currently president of the automotive
electronics division. Within the division’s manage-
© Bosch
ment team, he is responsible for body electronics,
e-bikes, innovation management, IoT, software,
Harald Kröger and quality.
Daimler | Geely | Electric Smart
Production Moves to China
Daimler and Geely have announced the formation of a 50:50 globally focused
joint venture to develop Smart as a leading electric vehicle brand. Under the joint
venture agreement, the next generation of Smart electric models will be assembled
at a new purpose-built electric car factory in China. The board of the new Smart
joint venture will be made up of six directors with equal representation from both
parties. The partners have agreed that the new generation of Smart vehicles will be
styled by the worldwide Mercedes-Benz Design network and developed by the Geely
global engineering centers in China. Until the market launch of the new models
in 2022, Daimler will continue to produce the current generation of Smart vehicles
at its plants in Hambach in France (Smart EQ fortwo) and Novo Mesto in Slovenia
(Smart EQ forfour).
© Daimler
Li Shufu (l.), Chairman of Geely Holding, and Dieter Zetsche,
CEO of Daimler, signing the agreement
ATZ worldwide 06|2019
IMPULSES
Dr. Johannes Liebl
Editor in Charge
ATZ | MTZ | ATZelektronik
Parallel Approaches
The ATZlive events “The Powertrain of Tomorrow”
in Frankfurt/Main and “The International Engine
Congress” in Baden-Baden in January and February
this year demonstrated that the mobility solutions
of the future will be based on two approaches:
electric powertrains and combustion engines run-
ning on carbon-neutral fuels. Electricity generated
from renewable sources and stored in batteries or
combined with CO2 to produce e-fuels is the key to
tomorrow’s mobility.
Sales of electric vehicles are growing and manu-
facturers are investing billions of euros in new
models. The market will only really take off when
the vehicles and the infrastructure are in perfect
harmony. The ATZlive conference “Grid Integration
of Electric Mobility” made it clear that we urgently
need to speed up the expansion of our power grids
and charging infrastructure. Given the current method
of measuring CO2 emissions, electric vehicles are the
solution to achieving the targets set for the transport
sector. However, our aim must be to introduce a
cradle-to-grave assessment method.
The manufacturing process for synthetic fuels is
carbon-neutral. They can be tailor-made for com-
bustion engines and they burn cleanly and without
producing soot. Adding them to fossil fuels makes
it possible to use them in existing vehicles. With
over 56 million vehicles on the roads in Germany
and more than 290 million in Europe, this approach
provides enormous leverage when compared with
the potential offered by new registrations of zero-
emission electric vehicles. Since 2016 the engine
congress has been highlighting the fact that the
combination of combustion engines and carbon-
neutral fuels represents a valuable parallel solution
to electric mobility.
It is high time for us to give equal emphasis to
both approaches and to synchronize our efforts
in these areas.
7
IN THE SP OTLIGHT
BUSINESS MODEL FOR ROBO - CARS
This business model is something that
needs to be explained. Who will be
earning from autonomous cars in future?
How much money is involved? How will
the revenue be generated? It is hard to
get a firm answer from established car-
makers and that is not the only problem.
Many people doubt that the business
model of a robo-car can even be suc-
cessful. One of them is Thomas Sedran,
Member of the Volkswagen Board of
Management. Even before the doors of
the Geneva Motor Show had opened in
March 2019, he made it clear that the
high costs involved meant that the busi-
ness model simply would not work. Now,
as CEO of Volkswagen Commercial Vehi-
cles, Sedran was talking about trucks.
8
More and more companies are expe-
rimenting with the move to driverless
commercial vehicles. In Colorado,
Nevada, California, and Texas, robo-
trucks like those developed by Daimler
and the Uber subsidiary Otto are a
common sight on highways in these US
states, as they are on the A9 in Germany
between Nuremberg and Munich. There
the logistics service provider Schenker
has been testing self-driving trucks in
connected platoons since 2018. Two
trucks with trailers travel autonomously
only 15 m apart in normal and real traf-
fic. However, the cost of developing safe
systems that comply with SAE J3016
level 5 and satisfy all the functional
safety requirements is enormous. On an
industry-wide scale they amount to sev-
eral billion euros. Who will pay for this?
DEVELOPMENT COSTS ARE
TOO HIGH TO BE VIABLE
The market regards success in the con-
text of commercial vehicles as being “the
complete absence of a driver,” as Tim
Schuldt, Analyst at the investment bank
Pareto Securities, explains. And this will
be the case only under certain circum-
stances, such as journeys through the
empty deserts of Texas and southern
California. There will be no driverless
delivery service whose vehicles can keep
up with the flow of traffic in the German
capital, for example, any time soon. Any-
one hoping for cheap travel or a lift in a
robo-car will be disappointed after read-
ing the analysis [1] of the Massachusetts
Institute of Technology (MIT) given by
Ashley Nunes and Kristen Hernandez in
 © Hyundai

Robo-cars –
Breach of a Taboo
Against the perspective of the more than 125-year history of the
automobile, it seems to be only a matter of moments until drivers
will no longer be needed. Robo-shuttles will soon be arriving outside
our doors on request and they will provide a safer and more punctual
service than any human could. The focus will no longer be on the
car as a product, but on intelligent mobility services. Robo-shuttle
buses will represent a historic opportunity. But for whom?

the Harvard Business Review (HBR).
The high development costs in this
new vehicle segment make it unlikely
that individual transport will become
cheaper in the near future.
But despite the poor market prospects,
General Motors is still experimenting
with its robo-taxi. The residents of Cali-
fornia will soon have their own robo-
taxi service, which will cost around
0.55 euros per mile. This is a significant
saving over a conventional taxi where
the charge is around three dollars
because of the driver’s wages. However,
Nunes and Hernandez do not believe
that the service will achieve financial
success quickly. The two authors of
the HBR study, who were previously
researchers at the MIT, are of the opinion
that the project would only be profitable
ATZ worldwide 06|2019
if General Motors “charged three times
the amount.” Alongside the high pur-
chase price, the increased repair and
maintenance costs of this type of vehicle
also have to be included in the calcula-
tions. But nevertheless autonomous cars
represent an opportunity.
ROBO-TA XIS RESOLVE THE
URBAN ECOSYSTEM PROBLEM
This could be the case at least in
megacities where the pace of urbani-
zation is rapid, the traffic density is
unbearable, and the infrastructure is
outdated. Almost 15 % of the 1.3 billion
vehicles in the world are clogging up
our city streets and thus the urban
ecosystem. “This leads to costs of more
than 300 billion US dollars as a result
of congestion and produces almost 20 %
of the greenhouse gas emissions from
transport,” explains Franck Leveque,
Partner and Mobility Business Unit
Leader at Frost & Sullivan. In his view,
the problem can only be resolved by
means of strategic cooperation between
public and private stakeholders with
regard to the operating models, vehicle
usage, multimodal journey planning,
and payment options. In other words,
by innovative mobility models.
Autonomous cars are the key to Mobil-
ity-as-a-Service (MaaS) initiatives in big
cities. According to Leveque, this would
increase the occupancy level of cars,
which is currently around 35 to 40 %,
to 85 %. It would “not only reduce the
number of kilometers traveled on the
roads, but also the traffic congestion.
9
IN THE SP OTLIGHT
The Cube of the supplier Continental is a delivery vehicle with no driver but with robot dogs for
parcel distribution
In addition, it would free up 20 % of the
road space currently used for parking.”
The study [2] indicates that autonomous
vehicles could contribute to an overall
saving of 4 % of the gross domestic
product. The analysts also expect a
reduction of up to 30 % in travel costs
by MaaS. However, many people believe
the claim made by the authors of the
study that “at the same time autonomous
cars would reduce the number of road
accidents to zero” is unrealistic.
FULLY AUTONOMOUS IN
ACCORDANCE WITH SAE J3016
“I believe the assumption that removing
the driver would prevent the majority of
accidents is misleading,” warns Prof.
Hans-Hermann Braess. In his opinion,
drivers’ intuition can help them to
overcome even the trickiest situations,
which results in some accidents being
avoided. Braess, previously researcher
and developer at Porsche and BMW,
played a significant role in the Euro-
pean research project called Pro-
metheus (Program for European Traffic
with Highest Efficiency and Unprece-
dented Safety). This was launched in
1986 and can be seen as the blueprint
for autonomous driving.
The fact that two generations of
researchers later there is doubt about the
vision of accident-free driving indicates
10
level 5 or full automation the system
will operate the vehicle dynamically
on every type of road and in all envi-
ronmental conditions as if a person
were in the driving seat. The funda-
mental problem with the control cir-
cuit remains unchanged.
THE RISK OF AN OPEN
CONTROL CIRCUIT
The rules of classic control engineering
were already familiar in Braess’ time:
“Only closed systems in which all the
system states can be taken into consid-
eration are suitable for automation.”
© Continental
But road traffic is a totally open system
where an almost unlimited number of
different situations can occur. At this
point, Keilhoff explains that level 5
“must function under all conditions.
The engineers will still be working on
this for decades.” It will not be possible
to overcome the challenge of driving
in fully automated mode at night in fog
the vast complexity of the plan. Like or rain on a road through a forest in the
Braess, Dr. Dan Keilhoff from the Chair near future.
in Automotive Mechatronics at the Insti- This is a task that “no company in
tute for Internal Combustion Engines the world can resolve single-handedly,”
and Automotive Engineering (IVK) of as Alejandro Vukotich explains to
the University of Stuttgart believes that ATZ in the interview. In 2019 Vukotich
there will always be a risk of accidents was appointed Senior Vice President
happening, despite the fact that the of Driver Assistance and Autonomous
SAE J3016 guideline specifies that for Driving Development at BMW and
When a robo-taxi is introduced, the existing means of transport will be ...
48 % 47 %
28 %
26 %
25 %
23 % 22 % 22 % 23 %
20 %
12 %
5%
© ATZ
... replaced. ... rather replaced. ... rather supplemented. ... supplemented.
In response to the question of whether the existing means of transport will be replaced or supplemented
when a robo-taxi is used, the IAO RoboCab acceptance study determined the following: Chinese customers
(red) see the RoboCab as an opportunity to substitute other modes of transport; German customers (yellow)
see the concept more as “complementary” to other offers; US american customers (blue) are rather indecisive
(source: Fraunhofer IAO)

© Dan Keilhoff
Dr. Dan Keilhoff
Project Head for Automated Driving at the Chair
in Automotive Mechatronics of the Institute for
Internal Combustion Engines and Automotive
Engineering (IVK) of the University of Stuttgart
believes that the company’s coopera-
tion with Daimler is the logical next
step. Together the two organizations
want to “develop one reliable shared
system instead of stand-alone solutions,”
as Ola Källenius, Member of the Daimler
Board of Management, made clear. The
robo-car project is ambitious: The plan
is that cars will be able to drive autono-
mously on the highway in only seven
years and the aim is to gain ground on
the Google subsidiary Waymo, which
is already some way ahead. But where
exactly are we today?
WORKING
FOR A SAFER
AND LIGHTER CAR
ATZ worldwide 06|2019
OVER 40,000 EMPLOYEES
2 QUESTIONS FOR …
ATZ _ When will the age of mobility
without drivers begin?
KEILHOFF _ These vehicles have been used
in driverless transport systems within
factory buildings for years. They demon-
strate that highly automated driving is
not a problem when the boundary condi-
tions are clearly defined. In cities, where
the number of influencing factors is near
infinite, it will be several decades before
level-5 vehicles come into operation.
HIGHLY AUTOMATED FUNCTIONS
ARE ON THEIR WAY
Leading carmakers are already promis-
ing level-3 functions by 2021, including
BMW with the automated electric car
iNext for example. Work on level-4 func-
tions is said to go ahead at full steam.
The emphasis is on environment detec-
tion, which represents a huge challenge
for the entire industry. Cars must be able
to reliably identify people who suddenly
step out into the road. This is already
feasible today at slow speeds, but who
MORE THAN 100 PRODUCTION PLANTS IN 22 COUNTRIES
Is this because the control algorithms
need to allow for some breaches of the
rules to ensure that vehicles do not have
to travel at walking pace through the
big and megacity as result of the strict
German road traffic regulations?
I prefer not to call this a breach of
the rules, but instead the humanization
of the control algorithms. We need to
be able to detect and resolve situations
of this kind.
wants to ride in a robo-car that travels at
walking pace through the city? The solu-
tion lies in high-performance computers.
Because future autonomous functions
require an enormous amount of comput-
ing power, Nvidia in particular has found
itself in a strong position within the
industry over recent years. One example
of its products is the Drive AutoPilot com-
puter, which it describes as “the world’s
first commercially available level-2+
automated driving system.” This is help-
ing many carmakers who want to launch
the first partially autonomous cars very
11
www.gestamp.com
IN THE SP OTLIGHT
In Bad Birnbach, the first autonomous bus in Germany has been
on public service since October 2017 – a project by the district
of Rottal-Inn and the community together with developers from
Deutsche Bahn, Regionalbus Ostbayern, and EasyMile
soon. The computer is based on Xavier
System-on-a-Chip (SoC) processors and
the specially developed Drive software,
which are said to allow full self-driving
autopilot functions, including highway
merge, lane change, lane splits, etc. to be
programmed in hardly any time at all.
TESLA IS ONCE AGAIN AHEAD
OF THE MARKET
The automotive industry supplier ZF
is also making use of the capabilities
of the Nvidia processors in its ProAI
RoboThink system, which according to
Wolf-Henning Scheider, the CEO of ZF,
is “the most powerful AI-capable super-
computer in the automotive field.” He
praised the work of his developers who
have succeeded in analyzing even highly
complex traffic situations in realtime
using a conglomerate of sensors. In order
to travel safely on public roads, autono-
mous cars from level 4 upward will have
to perform up to 600 trillion computing
operations per second (600 TeraOPS).
Tesla is forging ahead with the devel-
opment of autonomous driving func-
tions. It has created “the best chip in the
world,” as Elon Musk explained at Tesla’s
Autonomy Investor Day in April 2019.
The company claims that the 6 billion
transistors on a 260-mm² piece of silicon
provide 21 times the performance of the
Nvidia chips which it was using before.
This is good news for those supporters
of level 4 who already have the first
12
© Eva Stranzinger
© Bosch
The driverless, electric shuttle from Bosch with integrated services is designed to bring a
new kind of mobility to life
people-movers in pole position. These sharing vehicle that comes to pick up
include the German e.GO Mover, “the passengers on request and transports
first production-ready vehicle featuring them to their destination.
ZF systems which provides an autono-
mous mobility concept for cities,” as
UBER – HUGE SUMS AND
Scheider pointed out. However, the EZ10
HUGE UNCERTAINT Y
from French car manufacturer EasyMile
is also aiming to take over our cities. One company with a quite different
business model stands out from the
technology giants, carmakers, and
ROBO -SHUTTLES REPRESENT
start-ups in the field of autonomous
A HISTORIC OPPORTUNIT Y
mobility: Uber, the ride-hailing service
Equipped with on-board cameras, lidar, from California (USA). At first glance,
GPS tracking system, and the Simulta- Uber may not seem to be the ideal
neous localization and mapping (Slam) candidate for developing autonomous
technology, the EZ10 vehicle from cars, but the company has already
EasyMile has been operating as an “made the industry sit up and take
autonomous bus in Germany – more notice with a series of investments and
precisely in the lower Bavarian village acquisitions, together with a large-scale
Bad Birnbach. In Switzerland, by con- headhunting campaign in the academic
trast, the Arma vehicle developed by world” according to Clemens Wasner,
the French company Navya is being used CEO of Enlite.AI. There seems to be a
as part of the Mobility Lab Sion-Valais. lot of money at stake.
The Olli from Local Motors is similar in This is clear from Uber’s IPO filing
design and function. (known as S-1), which the company
The Association of German Transport submitted to the U. S. Securities and
Companies (VDV) is particularly pleased Exchange Commission (SEC) on April
about the progress being made in this 11, 2019. It explains how the Califor-
area, because vehicles like this can be nians see the future of mobility. From a
incorporated into the public transport potential market of almost three billion
system. In a position paper already US dollars, Uber is currently skimming
published in November 2015, the VDV off around 1 %. In order to see sustain-
explains that, as part of a fleet, “robo- able growth in its market share, the
shuttles represent a historic opportunity. company needs autonomous vehicles
They are the ideal complement to a high- that can travel on a large number of
performance local public transport sys- routes and in particular those where
tem.” In other words, this is perfect car- driverless robo-taxis can operate with-
 Discover
out the need to pay a driver’s wages. The company’s
planned stock exchange flotation also gives an initial
insight into the huge sums of money required to
remain competitive in the field of autonomous mobil-
ebm-papst.
ity. The business model needs to be designed with
great care.
New ideas for the automobile industry at
discover.ebmpapst.com/automotive
Andreas Burkert

For many years we have been in demand as engineering

REFERENCES
[1] Nunes, A.; Hernandez, K.: The Cost of Self-Driving Cars Will Be partners to manufacturers and system suppliers in the
the Biggest Barrier to Their Adoption. Published January 31, 2019.
automobile and commercial vehicle fields. Our innovative
Online: https://hbr.org/2019/01/the-cost-of-self-driving-cars-will-be-
the-biggest-barrier-to-their-adoption, access: May 3, 2019 ventilation and drive solutions contribute to making driving
[2] Frost & Sullivan: Frost & Sullivan präsentiert Top Cities, die weltweit
führend bei Smart Mobility Solutions sind und deren Strategien. Press
constantly more convenient, stress-free and safe.
release, April 16, 2019 You can’t see it. But you can feel it!
[3] Kurverwaltung Bad Birnbach: Erster autonomer Bus in Deutsch-
land. Press release, April 5, 2018. Online: https://www.badbirnbach.
de/presse/erster-autonomer-bus-in-deutschland, access: May 3, 2019

WHAT DO WE THINK?
“The biggest threat to the success of
robo-cars is our expectations. Driving
scenes like those in the movie “I, Robot”
will not be seen on our roads for decades
to come. But in specific areas with
well-defined conditions, fully automated
transport vehicles can be successful.
Highly automated cars will only achieve
significant market penetration when the
number of cars in private ownership falls
dramatically. Then we will have to answer
the question of which business model
the established automotive industry will
use to earn its revenue.

Dipl.-Ing. (FH) Andreas Burkert

is Correspondent of ATZ/MTZ.

ATZ worldwide 06|2019 13

C OVER STORY INTRODUC TION

CHASSIS

Integration from
AWD to Steering

16 The Electric All-wheel Drive

in the Audi e-tron
22 we
“With our crash expertise
can bring safety features
26 Steering Functions for
Automated Driving
Michael Wein, Christian Graf, into the chassis” Robert Fuchs, Tsutomu Tamura,
Sebastian Strasser, Stephan Poltersdorf Interview with Ralf Göttel [Benteler] Maxime Moreillon [Jtekt]
[Audi]

14
 Over recent decades fuzzy logic and neural networks
have paved the way for the breakthrough that has now
been made possible by Artificial Intelligence (AI). It
is only now that hardware components have become
available which can process all the AI computing oper-
ations in real time to enable cars and trucks to be steered
and braked effectively and safely. Many automotive
industry suppliers and carmakers are no longer using
CPUs in their ECUs for highly automated driving. Instead
they have moved to faster GPUs which were originally
developed for the gaming industry. They allow anima-
tions in video games to move more smoothly and to be
displayed in 3-D. “The things that have surprised me
about developments over the last 20 years, which we did
not predict back then, are the exponential growth in
computing power, the huge volumes of data, and the
importance of sensor technology,” said the philosopher
of science Klaus Mainzer in an interview in issue 43 of
the Springer Nature magazine “Spektrum – Die Woche”
as early as 2017.

Chassis development is still heavily influenced by the

trend for electric driving. A quite different approach to
controlling driving dynamics is needed in electric cars,
because energy recuperation absorbs all or part of the
braking power when an electric car is decelerating. This
combination of activating the mechanical service brakes
and regenerative braking via the electric motor needs to
be fine-tuned and carefully coordinated for all driving
and traffic situations. If electric all-wheel drive is added
to the mix, the engineer’s job becomes even more
difficult.

The steering system, which is the physical link between

the driver and the car, plays a key role in these interac-
tions, alongside the brakes. For the transition between
© Benteler
automated and manual driving to function correctly, it
is essential for the system to know whether or not the
driver’s hands are on the steering wheel. The develop-
ment of steering systems has recently been given a sig-
nificant boost by improved hardware (fault-tolerant
systems) and software (functions), when compared with
brake development.

Michael Reichenbach

ATZ worldwide 06|2019 15

C OVER STORY CHA S SIS

The Electric All-wheel Drive

in the Audi e-tron
In the development of sporty electric cars, emphasis has so far been placed on powerful and
spontaneous acceleration in the longitudinal direction of the vehicle. As the inventor of the original
quattro drive 40 years ago, Audi now designs its first electric all-wheel drive for the electric SUV
e-tron in such a way that a driving-active system is created through a dynamic torque distribution
between the axles – with 300 kW propulsion power and 220 kW recuperation power.

16
 © Audi
AUTHORS
Dr.-Ing. Michael Wein
is Project Manager Vehicle Function
quattro with e-tron Technology at
Audi AG in Ingolstadt (Germany).
Dr.-Ing. Christian Graf
is Project Manager Software Func-
tion quattro with e-tron Technology
at Audi AG in Ingolstadt (Germany).
Dr.-Ing. Sebastian Strasser
is Head of Development All-wheel,
Deceleration and Brake System
Functions at Audi AG in Ingolstadt
(Germany).
Dipl.-Ing. Stephan Poltersdorf
is Head of Development Chassis
Functions Longitudinal and Lateral
Dynamics at Audi AG in Ingolstadt
(Germany).
FROM THE ORIGINAL QUATTRO
TO AN ELECTRIC SUV
When Audi introduced the mechanical
all-wheel drive for passenger cars four
decades ago, vehicles with four driven
wheels were already out there. Specific
development of an all-wheel drive for a
sporty vehicle such as the original quattro
was, however, something quite new.
These days, with the electric quattro
drive in the electrical SUV e-tron, Audi
took up the challenge. Today also, elec-
tric all-wheel drive vehicles are already
in existence; their sportiness does not,
however, originate from dynamic torque
ATZ worldwide 06|2019
distribution between the axles, but
rather from the spontaneous perfor-
mance of the motors. The development
objective for the first electric quattro was
to implement the most capable, in terms
of vehicle dynamics, electric all-wheel
drive in a series production vehicle. It
should first and foremost control the
vehicle dynamics and performance of
the electric drives, even with no vehicle
stability and traction slip control system,
and provide, in puristic terms, an analog
driving experience in a digital vehicle.
While the all-wheel drive in conven-
tional vehicles is used to improve trac-
tion and driving dynamics, in electric
vehicles, the importance of recuperation
supervenes. In the Audi e-tron, a recuper-
ation power of up to 220 kW [1] is incor-
porated. Such a high regenerative braking
power can only be safely transmitted to
the roadway via distribution to both drive
axles. The correct torque distribution thus
makes a considerable contribution to the
electric range of the vehicle.
Drivetrain torque is distributed in both
driven and recuperation mode according
to the situation and dependent on the
driving condition. This article focuses
on describing the theory of operation
in the driven mode.
SYSTEM DESIGN
In contrast to the mechanical all-wheel
drive, an electric all-wheel drive system
is not an auxiliary system incorporated
in the drivetrain for four-wheel drive
purposes only. The entire electric drive-
train, as well as the networking and
functional architecture of the vehicle,
must take into account the all-wheel
functionality. In the Audi e-tron, the
quattro drive is therefore a key element
of the drive system.
The sensitivity of torque distribution
to the front and rear axle is directly
dependent on the electric machines fitted
to both drive axles. The design of what
is referred to as the distribution area in
FIGURE 1 reflects the base characteristic
of the electric all-wheel drive. Here, the
potential distribution of drive torque
between the axles versus the total drive
torque is shown [2]. The basic design of
the Audi e-tron all-wheel drive system
is, similar to that of the mechanical
quattro, slightly biased towards the rear;
this delivers sporty handling with high
vehicle stability and good traction.
17
C OVER STORY CHA S SIS
PmaxFA
100 %
Front Axle
(FA)
100 %
Rear Axle
(RA) 0
FIGURE 1 Distribution area – torque distribution between front and rear axle for an
electric all-wheel drive for continuous and boost power [3] (© Audi)
The dynamic torque, and therefore
current, distribution capacity places
demanding requirements on the
high-voltage vehicle electrical system
hardware and the powertrain software.
The components must support very
rapid current transfer, while avoiding
the generation of inadmissible overloads
caused by current peaks. This challenge
does not arise with traction control sys-
tems that limit, but do not distribute,
drive torque. With this limitation or
18
PboostFA
Continuous
All-wheel
Boost
power
reserve
100 Battery
PboostRA power [%]
PmaxRA
reduction, the high-voltage vehicle elec-
trical system is impacted by the reduc-
tion in current only. Current distribution
changes through the synchronous shift-
ing of torque to the second axle requires,
conversely, a significantly more complex
design in a system affected by communi-
cation latencies. Of particular impor-
tance is the redistribution at the battery
capacity limits. Here, no additional cur-
rent can be obtained; however, any per-
formance boost on one drive always
requires a parallel performance reduc-
tion on the other drive.
The effect chain to regulate differen-
tial speeds between both drive axles
presents the most time-critical require-
ment in the architecture, FIGURE 2. The
speed sensors in the wheels and elec-
tric machines are the entry point to this
effect chain. If a control deviation is
determined, the drive torque at the
machine or axle with the speed that
is too high must be reduced, and in
parallel, the drive torque at the other
machine increased by the same amount.
This effect chain can be optimized
through timing of the communication
between the functions and control units
involved. Communication via a deter-
ministic FlexRay bus system is particu-
larly suited to this. One such bus system
is therefore used in the Audi e-tron for
communication between the relevant
drive functions.
CONTROL CONCEPT
The objective for the electric all-wheel
drive system was both provision of opti-
mal traction and significant improvement
of vehicle dynamics. For this reason,
wheel-selective torque control, achieved
via braking interventions, was integrated
in the functional software of the electric
drive torque distribution. As these brak-
ing interventions generate yaw moment
directly, a reproducible manipulated vari-
able also becomes available for the con-
FIGURE 2 Control units within the system of
the distributed function quattro with e-tron
technology [3] (© Audi)

trol range limits of the electric torque
distribution, thus providing consistently
predictable driving behavior as well as
simplified application potential.
The drive torque effect on the wheels
can be divided into two fundamental
effects: The entire drive torque is either
transmitted fully to the road, and the
tire operates in the steady-state (linear)
range; or, parts of the drive torque can-
not be entirely converted into propulsion
and cause increased wheel slip in the
non-linear range of tire performance.
In both cases, the yaw moment of the
torque distribution results from the dif-
ference in cornering forces between the
two axles. Reduced cornering forces on
the front axle, in comparison to the rear
axle, create an inwards turning (agiliz-
ing) yaw moment; increased cornering
forces on the front axle, in comparison to
the rear axle, create an outwards turning
(stabilizing) yaw moment, FIGURE 3.
The developed control concept consists
of the following sub-functions: feedfor-
ward and feedback control of the agiliz-
ing brake torque, feedforward-control of
drive torque distribution, differential slip
control, actuator-driven virtual damping
function to reduce drivetrain vibrations,
and a monitoring function to ensure func-
tional safety. With the exception of the
virtual damping function, all sub-func-
tions are integrated in the Audi-developed
electronic chassis platform (Elektronische
Fahrwerkplattform, EFP).
For predictive reduction of power
understeering, the feedforward controlled
part of the all-wheel control system trans-
fers a higher proportion of drive torque to
the rear axle in accordance with the driv-
er’s desired torque. At high accelerator
pedal gradients, this effect is even further
intensified in order to increase the spon-
Driver torque
demand
FIGURE 4 Feed-forward control of the brake torque-vectoring function (© Audi)
ATZ worldwide 06|2019
Fy [N]
5°
4°
3°
2°
1°
Fx,decelerate [N]
-5 % -3 % -1 %
FIGURE 3 Correlation of circumferential and lateral forces at combined slip angle [°] and wheel slip [%]
in the diagram according to Krempel [4] (© Audi)
taneity of the vehicle response in a per-
ceptible manner. At the performance limit
of the drives, the distribution is, however,
limited, FIGURE 1, and requires compen-
sation via additional braking torque.
With the additional activation of the
hydraulic brake of the inner wheels, the
steer tendency – characterized by the
self-steering gradients – can be adapted
and improved independently of drive
torque availability. In addition, the cre-
ation of a specific slip difference between
both drive axles to influence the steer
tendency is not necessary. From the all-
wheel control perspective, this positively
impacts traction.
From the dynamic redistribution of
wheel contact forces during acceleration,
an equivalent change in sideslip stiffness
ax,des
Fx,accelerate [N]
1% 3% 5%
arises at both axles. This can result in
vehicle understeering or what is referred
to as power understeering, FIGURE 4. This
change in self-steering gradients in com-
parison to the stationary vehicle can be
compensated via braking torques on the
inside wheels. These braking torques
are, however, only applied, where the
forces transmitted to the road, in relation
to transferable forces, exceed a certain
amount. The non-linear behavior of the
sideslip stiffness in the vehicle limit
range are further balanced with a PI
understeering controller. This is, how-
ever, limited in such a way that over-
steering by the driver does not cause
the vehicle to oversteer.
If the accelerator pedal position
retracts suddenly when cornering at
Application
parameter
19
 C OVER STORY CHA S SIS
Magnitude [dB]
All-wheel
control
Virtual
damping
control
Frequency [Hz]
FIGURE 5 Activity of the functions all-wheel
control and virtual damping control in the
frequency domain, schematic (© Audi)
the stability limit (load change), a power
understeer-compensating effect occurs.
This significantly reduces the wheel con-
tact forces on the rear axle, thus invok-
ing an inward yaw moment. To attenuate
this load change response, drive torque
can be specifically distributed to the
front axle.
Even in an oversteering vehicle, the
cornering forces on the rear axle can be
increased via the all-wheel control by
reducing the longitudinal forces, that
means, by transferring the drive torques
to the front axle. At the performance
limit of the drives, the all-wheel control
has to reduce the overall drive torque
versus the driver’s desired torque to
enable the necessary front-biased drive
torque distribution. The ability to ensure
the ratio of drive torques via axle-selec-
tive torque limitation is a significant
advantage of electric all-wheel drives
over all-wheel systems with clutches
or locks. Their control range limits are
inherently defined by the ratio of axle
speeds and cannot be changed by the
control.
DIFFERENTIAL SLIP CONTROL
The differential slip control, in parallel
with the feedforward-controlled torque
distribution, ensures that the resulting
traction slip between both axles remains
within the specified limits. Effects of the
different turning radii of the axles when
cornering and differences in rolling cir-
cumference of the tires are, however, not
compensated here, to avoid subjecting
the drivetrain to virtual torsion.
The limits of the permitted differential
slip are calculated directly from the feed-
forward-controlled torque distribution,
scaled by a linear equivalent longitudi-
nal stiffness of the tires. This transforms
all the relevant characteristics automati-
cally into the differential slip. So that the
differential slip control is not permanently
20
active, activation thresholds, symmetrical
to the limits of the differential slip, are
also taken into account. The differential
slip control is implemented as a PID con-
troller whose three principal gains are
adapted to the respective drive torque.
The bandwidth of the closed-loop sys-
tem is limited on the one hand by the
natural frequency of the drivetrain and,
on the other hand, by the FlexRay con-
nectivity transfer speed.
What is challenging with electric all-
wheel drives is the low level of mechan-
ical damping in the drivetrain. Rapid
changes in drive torque distribution
invariably lead to an excitation of oscil-
lations at the resonant frequency of the
drivetrain. So that the control is not
required to absorb speed losses, an actu-
ator-based virtual damping function is
superimposed. The virtual damping is
achieved by a P controller, between mea-
sured wheel and rotor speeds, FIGURE 5.
OPTIMIZATION OF
VEHICLE DYNAMICS AND
FUNCTIONAL SAFET Y
In addition to the functions to optimize
vehicle dynamics and traction, efficient
torque distribution as well as functional
safety are also considered in the control-
ler concept. With no additional mea-
sures, the all-wheel control would apply
drive torque to both axles constantly,
thus causing a disadvantage in terms of
efficiency. As a result, the requirement
for all-wheel distribution is continuously
assessed based on traction and vehicle
dynamics reasons; where the require-
ment is low, the most efficient operating
strategy is selected. Here, the function
concept builds on the logic already used
FIGURE 6 Advantages of the all-wheel controller with an integral wheel-selective torque control while
snowy-road handling (© Audi)
in the quattro with ultra technology.
This allows for both maximum efficiency
and the best possible traction and vehicle
dynamics to be achieved.
The concept for functional safety of
the all-wheel control is based on a sim-
plified slip controller that keeps the dif-
ferential slip within a safe range depend-
ing on the driving situation. The all-
wheel function thus ensures the safe
operation of the two decoupled axle
drives. The intrinsically safe operation
of the wheel-selective torque control is
ensured in every driving situation via
downstream yaw rate monitoring at the
vehicle level.
RESULTS IN THE VEHICLE
IN COMPARISON
The advantages of an all-wheel controller
with an integral wheel-selective torque
control can be seen during snowy-road
handling, FIGURE 6. Therefore, the driv-
ing dynamics-related properties of the
electric all-wheel drive of the Audi e-tron
were compared to the clutch-based all-
wheel drive in the Audi R8 V10 sports
car, FIGURE 7 [3]. This comparison was
selected, as both vehicles are technical
implementations of the quattro drive
with an actively controlled system,
and because vehicle handling is more
strongly defined by the all-wheel drive
than by the vehicle concept when on
the snowy-road handling track.
Both vehicles show a very natural
handling, with a very low tendency to
over or understeer. The Audi e-tron, in
comparison to the Audi R8, has a slightly
increased tendency to oversteer at high
drive torques. This is caused by the
configuration of the wheel-selective
 iStock © ollo
Audi e-tron Audi R8
1 1
Oversteer Oversteer
0.9 Understeer 0.9 Understeer f a s t f o r wa r d s o l u t i o n s
0.8 0.8

Relative vehicle dynamics [-]

0.7 0.7

0.6 0.6

0.5 0.5

0.4 0.4

0.3 0.3

FIGURE 7 Comparison of
vehicle dynamics between
quattro with e-tron technol-
0.2

0.1
0.2

ogy and the clutch-based all-
wheel drive system in an Audi
R8 V10 quattro on a snow
0
0 0.5
handling track [3] (© Audi)
torque control at the stability limit. The
vehicle follows the steering even when
this results in slightly oversteered behav-
ior that is easy for the driver to handle.
However, the R8, as a sports car, is in
contrast designed for high speeds and
therefore exhibits more neutral handling
in this driving situation. The advantages
of the R8 with hang-on front axle come
into effect most of all at a high friction
coefficient and on racetracks.
The R8 already shows a very low
understeering tendency on a snowy-
road handling track. The fully variable
all-wheel drive of the e-tron can, how-
ever, further reduce this across the entire
drive torque range. The combination of
electric all-wheel drive with wheel-selec-
tive torque control reduces the under-
steering tendency even in driving situa-
tions with low traction slip, as handling
can be reliably improved with minimal
brake interventions here.
0.1
Automotive
1
0
0 0.5 1
Solutions for
Rel. torque request [-] Rel. torque request [-]
Future Mobility
With the Audi e-tron, the successful
| Clean Air Solutions
history of the quattro drive is developed
further and transferred into the electric
| Drivetrain Solutions
age. The approach involving an integral | Thermal Management
wheel-selective torque control as a cen-
tral element of the drive system allows
an electric all-wheel drive system to be
implemented for the first time that con-
tinues to fulfill its function even with the
traction control system disabled and the
motors at full capacity.
The prerequisites for reaching the
objectives defined at the start of the proj-
ect were – in addition to the right design
of the electric drivetrain – especially the
integral function approach as well as the
concept of the software functions. In
addition to these technical boundary con-
ditions, the close collaboration across all
chassis and drivetrain functions in the
development team made a critical contri-
bution to achieving the goals.

CONCLUSION REFERENCES
[1] Geuss, M.; Wein, M.; Hörter, M.; Strasser, S.;
Many battery-powered electric vehicles Schwarz, R.: High Efficient Regeneration in the
offer very good vehicle dynamics at a Audi e-tron – Audi SW Function Combined with a
high friction coefficient. The reasons for By-Wire Brake System. 27 th Aachen Colloquium,
October 8 to 10, 2018
this are attributable to the spontaneous [2] Wein, M.: Konzeption eines elektrifizierten All-
power output of the electric drives and radantriebs für ein Plug-In Hybridfahrzeug. Aachen:
the low center of gravity created by the ika-Schriftenreihe Automobiltechnik, 2015
[3] Graf, C.; Wein, M.; Baur, M.; Strasser, S.;
underbody battery. Like the original
Schwarz, R.: quattro with e-tron Technology –
quattro for conventional vehicles, the Electric All-Wheel Drive System with Wheel
Audi e-tron gives the electric vehicle for Selective Torque ControI. 27 th Aachen Colloquium,
the first time a previously unheard of October 8 to 10, 2018
[4] Heißing, B.; Ersoy, M.; Gies, S.: Fahrwerkhand-
sporty driving experience and the best
buch – Grundlagen, Fahrdynamik, Komponenten,
possible traction properties at all friction Systeme, Mechatronik, Perspektiven. 3 rd edition,
coefficients. Wiesbaden: Vieweg Teubner, 2011

ATZ worldwide 06|2019 21

w w w. b u e h l e r m o t o r.c o m
C OVER STORY INTERVIE W

© Uwe Lewandowski | ATZ

“With our crash expertise we can
bring safety features into the chassis”
Electric drive systems are having a major influence on chassis design, new requirements are being added.
Batteries must be installed where they are protected from collisions. Benteler is responding to the fluctua-
tions in the sales of electric cars by developing scalable, modular platforms for battery-powered electric
cars. In the ATZ interview, Ralf Göttel, CEO of Benteler, describes the advantages of this system, known as
the Benteler Electric Drive System, BEDS for short, how the Chinese can ensure that the sky over their larg-
est cities is blue, and highlights the connection between artificial intelligence and agile partner networks.

ATZ _ Ralf Göttel, you have significantly
increased the size of your portfolio of
electric mobility. When will we see more
electric cars than conventional models
being sold in German showrooms?
22
GÖTTEL _ We believe that electrification
is definitely on its way. Experts have
forecast that by 2030 a quarter of the
cars that are sold will have fully electric
or hybrid drives. Legislation will have a
major impact on the proportion of elec-
tric vehicles. If today’s plans remain
unchanged, electric mobility will play
a much bigger role, because we cannot
achieve our CO2 targets in any other way.
 Ralf Göttel (born in 1966) has been CEO of Benteler
International AG in Salzburg (Austria) since April 2017.
His current responsibilities as CEO include strategy and
M&A and group compliance and committee activities,
together with the automotive, steel/tube, and distri-
bution divisions. Since August 2018 he has also been
CEO of the Benteler Automotive division, which is based
in Paderborn (Germany). He had already headed this
division from 2014 to 2017 as Chairman of Benteler
Automobiltechnik GmbH. He has held a range of
management positions within the Benteler Group since
2010. He studied mechanical engineering at RWTH
Aachen University and began his career in 1991 as a
development engineer with a major car manufacturer.

© Uwe Lewandowski | ATZ

What is this assessment based on?
This assessment is based on our experi-
ence, sharing information with car man-
ufacturers, and monitoring the market.
Another very important source of infor-
mation is our discussions with employ-
ees at our various sites. They give us a
clear indication of the prevailing mood.
That’s what is happening in Germany.
How is it that China is investing so
heavily in electric cars?
There are four reasons for this. In China
there is no tradition of manufacturing
combustion engines dating back more
than one hundred years as there is in
Germany. The Chinese are attracted to
new technologies and are early adopters.
Driving habits there are more suited to
electric transmissions because many
journeys are short. Finally, environmen-
tal issues in China’s megacities play a
major role. These are not an ideology
or a question of belief in the same way
as they are in the EU. They are part of
everyday life in Shanghai, for example.
If the sky over the city is not blue, the
Chinese opt for electric cars with zero
local emissions. This is something that
is very real for them. China is leading
the way in the field of electric mobility,
which is why we have had a presence
in the country since 2002 and now have
15 plants there.
Like many other automotive industry
suppliers, you have designed your Benteler
Electric Drive System, known as BEDS,
as a modular system solution. What are
the benefits of modular, scalable platforms
of this kind?
Our concept can be used as the basis for
vehicles of different lengths and widths,
which means that it is suitable for use
in cars from all segments. The batteries
are available in a standard form or as a
tailor-made solution to meet customers’
individual requirements. They are also
scalable, depending on the range. The
chassis solution, which is one of Bentel-
er’s core products, allows for electric
all-wheel, front-wheel, or rear-wheel
“There are no good
or bad materials”
drive. Other options include an integrated
battery cooling system and an inductive
charging solution. This means that the
BEDS provides all the functions needed
for a rolling chassis. It is a series-ready
platform that brings together all our tech-
nical knowledge in one product. As a
global partner of the automotive industry,
Benteler offers first-class engineering and
metal processing competence. This covers
components and modules for the chassis
and body-in-white and also includes mod-
ular solutions for electric mobility.
What is the central feature of the system –
its crash safety?
Yes. The BEDS is not just a collection
of individual components. The chassis
includes an integrated crash function
to protect the battery. Passive safety is
the key consideration for us and our cus-
tomers. It is often neglected in electric
vehicles, but it is very important in the
event of a side impact or of the battery
being punctured by an object on the
road. With our crash expertise we can
bring safety features into the chassis.
How advanced is the BEDS in technical
terms? What has the initial response been
from the market?
We have developed and tested a series-
ready platform solution. It is currently
undergoing testing at a well-known tech-
nology company. The pilot project there
will evaluate its functionality and perfor-
mance under real-life conditions. We are
confident that parts of the solution or the
complete system will go into volume pro-
duction with other car manufacturers.
That is precisely the idea behind it: to
provide a technology platform in the form
of a complete system, an individual com-
ponent, or a subassembly. This is what
modular design is all about. We have
received positive feedback from several
customers which indicates that there is
a need for our BEDS on the market.
You won’t be allowed to give us names
of your customers? But is your system
being used in a car or a truck and in
which country?

ATZ worldwide 06|2019 23

 C OVER STORY INTERVIE W
© Uwe Lewandowski | ATZ
You will understand that discretion
is very important to us and our cus-
tomer, but I can tell you that the system
will be used in China in a C-segment
and D-segment car. As has been the
case with many new systems, electric
mobility is being introduced via pre-
mium models.
But cities with major air pollution problems
need small electric cars, not large models,
so that they can reduce exhaust emissions.
The situation is the same as with many
other technologies. New developments
first come onto the market in high-
priced premium cars and then move
into the mid-sized and small models.
This was also true of ABS, safety belts,
and airbags.
The BEDS has electric front and/or rear
axles which form a so-called e-chassis.
Do you design and manufacture the electric
motors and power electronics yourselves?
We work with strategic partners who
include automotive industry suppliers
and car manufacturers. This allows
us to make the best use of everyone’s
expertise to produce a complete system.
We select our partners based on the
individual customer needs. This means
that we can offer our customers tailor-
made electric mobility solutions from
individual components through to
complete series-ready systems. As a
24
result, we can help our customers save
several years of development time and
support fast growth acceleration.
Many of the presentations at our 10th chassis.
tech plus conference, which is taking place
at the end of June, will highlight the fact
that AI will soon be making its way into the
chassis. How does a traditional steel tube
manufacturer like Benteler develop the
necessary knowledge of algorithms?
Artificial intelligence is an important
subject for us and we have resolved
this problem by entering into agile
partnerships. We are heavily involved
“All-wheel drive will
not achieve full market
penetration”
with start-ups that provide us with sup-
port in the area of AI and algorithm
development. We work with them and
our customers to analyze the signals
from the variety of sensors used in cars
and to identify correlations with the
results of chassis behavior that we are
familiar with.
Which AI projects is Benteler currently
working on? How does the company stay
at the cutting edge in this area?
“The battery tray forms a
very stiff structure with the
flexframe, which shows an
optimal deformation behavior
in the event of a side impact,”
explains Göttel (left), in conver-
sation with Michael Reichenbach,
Deputy Editor in Chief of ATZ
Some of our employees are involved
in the Start-up Autobahn project and
also in other start-up scenes. We have
acquired some very interesting experi-
ence in the field of AI. A start-up net-
work has grown up in the region around
Paderborn in Germany, because of the
strong tradition of the university and the
influence of Siemens-Nixdorf, and this
is proving very useful to us. We get a
constant supply of new ideas that help
us to anticipate trends and increase our
ability to innovate.
The BEDS will be available in both steel
and aluminum. In what situations do the
Benteler engineers recommend a specific
material to the car manufacturer?
I always say that there are no good or
bad materials, there is just the wrong
and the right application. When it comes
to choosing the material, we follow
well-worn paths. If a lightweight struc-
ture is needed, then aluminum is the
first choice. If a low-cost solution is
required, then steel is the better option.
The third major criterion is the local
availability of the material. Depending
on the market, import duties, and other
restrictions on a raw material, we can
take a quite different approach to the
development process.
Does that mean that aluminum is now on
the back burner?
We see the situation differently. This
lightweight metal is used slightly less
often for two reasons. One is that steel
has caught up with it in terms of its tech-
nical performance. It is now possible to
produce lighter-weight structures in steel.
The other is that it is no longer essential
to make the entire vehicle lighter. Often
it is only the weight of the front-end or
rear-end that needs to be reduced.
Electric motors in cars make it relatively
simple to develop all-wheel drive systems.
Will they become much more widely used?
I don’t think that we will see full market
penetration for all-wheel drive systems.
Just because something is cheaper does
not necessarily mean that more end
customers want to buy it. We think that
there is a trend for rear-wheel drive in
electric cars. This improves the weight
distribution, because now the battery
ATZ worldwide 06|2019
has to be taken into consideration. How- Ralf Göttel, thank you for this
ever, electric all-wheel drive systems interesting conversation.
are much lighter in weight than their
mechanical equivalents. Our BEDS You can read more of the interview
makes it possible to have drives of the in German on the ATZ online portal
same size on the front and rear axle, at www.springerprofessional.de.
which allows for both front- and rear-
wheel drive.
Do you drive an electric car yourself?
I sometimes get behind the wheel of an
electric car. I like to book electric cars
from carsharing schemes, partly out
of curiosity. And my next company car
will be a plug-in hybrid. We also give
our employees the opportunity to use
electric vehicles. For example, we have
had an electric car in the company pool
at our Talle plant in Paderborn since
2017. Anyone can drive it and I am one
of the people who enjoys taking it out
on the road. INTERVIEW: Michael Reichenbach
Enjoy performance and driving pleasure
www.jtekt.co.jp/e
25
C OVER STORY CHA S SIS

Steering Functions
for Automated Driving
Automated driving functions can take over the longitudinal and the lateral guidance
of the vehicle. The steering system plays a major role in the transfer of responsibility
between automated and manual driving. The Japanese supplier Jtekt has developed
a hand condition detection system based on an estimate of the driver torque as well
as a seamless override function. Both use exclusively the sensors of the steering system.

26
 AUTHORS

Dr. Robert Fuchs

is Head of the
Mobility Control Research
Section at Jtekt Corporation
in Nara (Japan).

Tsutomu Tamura
is Manager in the
Mobility Control Research
Section at Jtekt Corporation
in Nara (Japan).

Maxime Moreillon
is Engineer in the
Mobility Control Research Section
at Jtekt Corporation in Nara (Japan).

INTRODUCTION

Safety and user acceptance of automated

vehicles, necessary for market penetra-
tion, are strongly related to reliable and
intuitive human-machine interactions.
The steering system plays one of the
main roles in these interactions as being
physically the central component con-
necting the driver to the vehicle. While
brakes have been the focus of major
technical improvements in the past
decades, Electric Power Steering (EPS)
systems are currently witnessing signifi-
cant hardware (failed operational sys-
tem) and software (functions) evolu-
tions. This article presents two basic
functions of EPS part of Jtekt technical
roadmap on power steering evolution
toward driving automatization. These
two functions are Hands-on/off Detec-
tion (HOD) based on the EPS sensors and
Control Authority Transfer between auto-
mated and manual driving. Both func-
© Jtekt

tions were developed internally by Jtekt

advance development.
ATZ worldwide 06|2019 27
C OVER STORY CHA S SIS
HAND STATE DETECTION
Hand state on the steering wheel is the
most reliable indicator for evaluating the
ability of the driver to control his vehi-
cle. The application of HOD is required
for ensuring safe operation depending on
the level of automation (SAE Standard
J3016). For SAE level 1 and 2 where the
driver is fully responsible of the driving
task, hands-off detection is required.
For example, the Lane Keeping Assist
(LKA) is required by regulation (UNECE
Regulation 79, Revision 3) to start warn-
ing the driver if not holding the steering
wheel for a duration no longer than 15 s.
For SAE level above 2, the driver can be
relaxed from the driving task in some
conditions. Hands-on detection becomes
necessary when the automated system
requires the driver to take over the con-
trol, for example, when the traffic condi-
tion is too complicated for the automated
driving algorithms. The transition from
automated to manual driving mode
should only be initiated after the driver
hands have been detected on the steer-
ing wheel (this is a necessary but, likely,
not sufficient condition).
Two technologies are available for
detecting the hand state: touch sensor
on the steering wheel (mostly capacitive
type) and EPS sensor based detection.
A capacitive or touch sensor provides
direct measurement of the driver hand
contact with the steering wheel. Actual
implementations are limited by sensitiv-
ity setting trade off: misdetection when
a hand is close but not in contact with
the wheel (high sensitivity) and wide
hand surface in contact required (low
sensitivity). Another limitation is that
the sensitivity has an inhomogeneous
distribution over the irregular surface
of the steering wheel (for example lower
sensitivity near the “poles”). Wheel cov-
ers and gloves also degrade the sensing
performance. EPS sensor based detec-
tion relies on the driver torque as the
relevant information. Early develop-
ments used the torque sensor signal
with some filtering, which resulted in
recurrent misdetection especially with
road input of rich frequency content.
The method proposed by Jtekt is based
on the estimation of the drivers hand
torque [1]. This results in a detection of
low sensitivity to road inputs. However,
the method needs low torque input from
the driver, even on straight driving con-
28
Driver not in control
Hand state
Hands-off
Touch sensor
Hands-off
EPS sensor
FIGURE 1 Comparison of detection performance between touch sensor and EPS sensor based HOD
(© Jtekt)
Steering
wheel inertia
FIGURE 2 Steering wheel model and block diagram of the driver torque estimator (© Jtekt)
dition. EPS based hands on/off detec-
tion has the advantages of being cost
effective and functional safe as relying
on Asil-D compliant sensors (torque and
resolver). While touch sensor based
detection provides intuitive informa-
tion (hand contact corresponds to the
hands-on state), EPS torque based
detection relates the hands-on state to
the driver action of steering, which bet-
ter reflects the actual state of the driver
as in control of the vehicle, FIGURE 1.
Because the torque sensor of an EPS
measures the driver torque in static con-
dition only, a Driver Torque Estimator
(DTE) has been developed for detection
robust to road and assist motor inputs
as well as to steering wheel inertia
effect. Using both the torque sensor
and the assist motor resolver, a simple
model of the steering wheel is suffi-
cient, FIGURE 2. The linear nature of the
model and the high signal to noise ratio
of the sensor outputs enable the applica-
tion of the Luenberger observer theory
with the advantages of low computing
weight and objective gain tuning abil-
ity. This means that the bandwidth of
the observer can be tuned by placing
Driver in control
Touch Grip S
Steer
Hands-on
Steering column
DTE
+
L
-
Model
the poles at a desired cut-off frequency.
In principle, for a driver torque estima-
tor, it is set to the bandwidth of the
driver input of 5 Hz. Physically, this
approach forces the model used in the
observer to follow the plant and extracts
the driver torque by compensating the
inertia effect of the steering wheel.
Hence, the DTE enables the rejection of
the effect of the steering wheel inertia
while providing similar sensitivity to
that of the torsion bar. A time lag is
observed between the torsion bar signal
and the DTE output. This is due to the
bandwidth of the observer, set to 5 Hz,
which yields a time constant of 0.1 s.
On the other hand, it is significantly
less sensitive to the inertia effect.
The structure of the HOD function is
summarized in FIGURE 3. It is based on
the driver torque estimator, which out-
puts a signal virtually insensitive to
the inertia effect of the steering wheel.
A torque threshold ensures distinction
between the hands-on and hands-off
states and thus defines the sensitivity
of the function. At last, a time window
prevents the HOD output from mistak-
enly switching to hands-off during the
 Driver torque Torque Time Hands on/off
estimation threshold window
Amplitude [dB]

Torque [Nm]

Torque [Nm]
5 [Hz] FIGURE 3 Structure of the
0 0 0 HOD function (indicated
values are examples
0.2 [Nm] 2 [s]
obtained from a proof of
concept made on a column
Frequency [Hz] Time [s] Time [s] assist EPS) (© Jtekt)

short instants when the driver torque steering mode, as illustrated in the Title
can happen to be under the threshold Figure where the steering wheel is shak-
in hands-on situation. ing. The data shows that the DTE pro-
The hands-on state is defined as when vides an output free from the road and
the driver torque reaches values above assist motor disturbance compared to
the threshold. This threshold is tuned so the EPS torque sensor outputs. The
as to ensure that statistically no misde- detection and the hand state prove to
tection occurs in hands-off condition. be reliable in the test conditions.
Practical implementation should also
include consideration on the parts vari-
CONTROL AUTHORIT Y TRANSFER
ances. Driver torque based detection out-
puts a hands-on state well-defined. How- From SAE level 2 to 4, the lateral control
ever, the hands-off state is ambiguous. of vehicles is shared between the driver
Driver torque near zero or between the and the automated steering system. Over-
threshold boundaries does not strictly ride, when using LKA, is straightforward
indicate hands-off. For example, this is as the control of the EPS for both manual
the case during a slalom maneuver. Con- assistance and ADAS function are torque
sequently, hands-off is detected by wait- based. Typically, if the driver opposes
ing that the driver torque remains under the LKA torque above a certain threshold,
the threshold for certain time after the function is disengaged automatically.
which it is statistically unlikely that Because of the low level of intrusion of
a driver holding the steering wheel is such ADAS, the torque threshold is set at
able to maintain such a small torque. a relatively low value enabling seamless
FIGURE 4 shows vehicle test data on transfer of authority. At high level of auto-
cobble stone road surface in automated mation (SAE level 3 and 4) the principle

Torsion bar output Torsion bar output

DTE output DTE output
Threshold Threshold

2 2
Torque [Nm]
Torque [Nm]

0 0

-2 -2
5 10 15 20 25 30 5 10 15 20
Time [s] Time [s]
Hand state Hand state
HOD output HOD output

on on
Hands

Hands

off off

5 10 15 20 25 30 5 10 15 20
Time [s] Time [s]

FIGURE 4 Response examples of the HOD function (left: automated steering (sine wave) on asphalt road,
right: straight driving on cobble stone road) (© Jtekt)

ATZ worldwide 06|2019 29

C OVER STORY CHA S SIS

tion, FIGURE 5 [2]. Here, the motor torque

Authority transfer consists of the weighted sum of the auto-
Assist map x mated and manual steering controller
torques. The weights sum up to 100 %
Manual steering component and have been chosen to vary with
respect to the difference between the
- +
1 EPS command angle and the steering column
Weighting + +
angle. The automated control weight goes
Trajectory from 100 to 0 % and the opposite for that
Angle control
specification + - + of manual assistance when the column
x
rotates by a set angle away from the auto-
Active load Limitation +
Automated mated target angle. As the driver gener-
compensation
steering component ates this angle error when overriding the
system, the reaction force felt is that of
FIGURE 5 Block diagram of the EPS control including the authority transfer function (© Jtekt) the angle controller. To ensure that the
reaction force does not exceed the maxi-
mum torque that can be applied by the
5 driver, the torque command computed
Steering
torque
[Nm]

0 by the angle controller is saturated. This

method is independent from any torque
-5
threshold or transition time constraint.
wheel jerk

2000
Steering

Rather, using an angular displacement as

[rad/s³]

0
trigger of the authority transfer is justified
-2000
as it corresponds physically to the appli-
20 Actual value cation of a torque (driver intention) over a
Steering
angle
[rad]

10 Target value certain period of time (robustness on the

0 driver intention). Therefore, the proposed
α
function is operation safe because robust
100
to accidental shock on the steering wheel
Weight

β
[%]

50
0 and to torque sensor offset. It enables
0 2 4 6
FIGURE 6 Example of vehicle test data showing seamless transition (© Jtekt)
of operation of the steering changes from
temporarily intervention to continuous
automation. Hands-off driving is then
allowed. LKA becomes lane centering,
which means that the EPS controls the
wheel angle. Angle control in automated
mode and assist control in manual mode
oppose each other. A driver input over the
angle control is seen as a disturbance and
rejected. This raises the question of how
to operate safe and reliable transfer of
authority between the driver and the
system.
Using the hand state from the HOD
function is not realistic as unsafe to
accidental touch of the steering wheel.
Moreover, it would refrain the driver
from holding the steering wheel during
automated driving. Using a torque
threshold above which the authority is
transferred to the driver present a com-
promise more complex than that for
LKA. Indeed, when using the LKA func-
tion the responsibility of the driving
8 10 12 14 16
Time [s]
task remain to the driver. Therefore,
low threshold or low level of intrusion
is desired. In automated mode, lane
centering is stiff (also called “hard”
angle control). Low torque threshold
results in seamless authority transfer
but presents the risk of undesired tran-
sitions caused by accidental touch of the
steering wheel or disturbances affecting
the torque sensor such as friction for
example. Higher torque threshold for
improving robustness to accidental
touch results in steering wheel jerk
during the transition. The transfer of
authority could be smoothened by oper-
ating the transition gradually over a
certain amount of time. However, a
time dependency prevents the revers-
ibility of the transition and could delay
the transfer of authority.
The approach proposed by Jtekt is to
combine the automated and manual
steering strategies into a single, nonexclu-
sive control mode using a weighting func-
bidirectional transition from automated
to manual and the opposite depending if
the angle error increases or decreases
ensuring driver comfort and acceptance.
FIGURE 6 illustrates the seamless authority
transfer obtained with this function.
SUMMARY
Hands on/off detection and control
authority transfer are two basic func-
tions of EPS required for intuitive and
safe operation of automated driving.
They are enabler technologies for consis-
tent human-machine interaction. Further
efforts from Jtekt focus on the coopera-
tion between driver and automation sys-
tem as presented at the 2019 ATZlive
Conference Automated Driving [3].
REFERENCES
[1] Moreillon, M.; Tamura, T.; Fuchs, R.: Detection
of the Driver’s Hand on and off the Steering Wheel
for ADAS and Autonomous Driving. ATZlive Confer-
ence chassis.tech plus, Munich, 2016
[2] Moreillon, M.; Tamura, T.; Fuchs, R.: Transfer
of Lateral Control Authority for Automated Vehicles.
JSAE Autumn Congress, Nagoya, 2018
[3] Shoji, N.; Yoshida, M.; Nakade, T.; Fuchs, R.:
Haptic Shared Control of Electric Power Steering.
ATZlive Conference Automated Driving, Wiesbaden,
2019

30
 Automotive Acoustics Conference
5th International ATZ Conference on Vehicle Acoustics

9 and 10 July 2019

Zurich/Rueschlikon | Switzerland

CALCULATION AND SIMULATION

Efficient approaches
for analysis and development

VEHICLE NVH
Increasing relevance due
to future mobility

NVH OF HEVS & BEVS

New approaches for
the drive train and components

/// KEYNOTE LECTURES

Dr. Ralf Bocksch, Daimler AG, Germany |

Léon Gavric, PSA Groupe, France |
Yoshiki Konishi, Toyota
Motor Corporation, Japan

/// KINDLY SUPPORTED BY

Phone +49 611 7878-131 PROGRAM AND REGISTRATION

Abraham-Lincoln-Straße 46
65189 Wiesbaden | Germany
Fax +49 611 7878-452
ATZlive@springernature.com www.ATZlive.com
ATZ worldwide 06|2019 31
NE WS PRODUCT S
dSpace |
Radar Test Benches
dSpace has supplied the first radar test benches
fitted with the latest dSpace Automotive Radar
Test System (Darts). They allow for over-the-air
simulation of the radar sensor in order to test the
entire chain of effects. The test benches incorpo-
rate all the software and hardware layers into the
tests, from detecting the signal at the front-end
Audi | Four-cylinder Turbos Replace Naturally Aspirated V8 Engines
© Audi
32
of the radar sensor to evaluating it in the radar
control unit. The main components of the test
benches include an anechoic chamber with
send and receive functionality, a calibration,
and an in-house Scalexio HiL simulator. During
the tests, the radar sensor is stimulated in the
anechoic chamber with up to five radar echoes.
© dSpace
The coherent echoes allow the radar control unit
to measure the distance and the velocity of the
radar objects reliably. The concentric movement
of the antennas provides information about their
azimuth angle. The configurable test benches can
be used with 24, 77, and 79 GHz radar sensors in
all modulation schemes.
Naturally aspirated V8 engines will no longer
be used this season in the German Touring Car
Masters (DTM). They have been replaced by
four-cylinder turbocharged powerplants with
a capacity of 2.0 l. With a charge pressure of
3.5 bar, Audi’s engine are claimed to produce
around 450 kW and has a rev limit of 9500 rpm.
The compact front SI engine is longitudinally
mounted and weighs 85 kg, which is around
half as much as the previous V8 engine. As a
result, the dry weight of the Audi RS 5 DTM
has been reduced to less than 1000 kg and
the weight-to-power ratio is now around 2.2 kg
per kW. Despite its high power output, the
engine has to last for a complete season,
which is the equivalent of around 6000 km.
The fuel flow rate is restricted by the regula-
tions to 95 kg/h, but the restrictor mechanism
can be bypassed for a short period, for example
for overtaking purposes, using a push-to-pass
function. This provides an additional 5 kg
of fuel per hour for a period of 5 s, which is
reported to give a power boost of around 22 kW.
 csi | AMC | Alba |
Lightweight Seat
Over a period of seven months, a virtual ultra-lightweight
seat and a hardware prototype have been developed as
a part of a project involving automotive development
company csi Entwicklungstechnik, Alba Tooling & Engi-
neering, and the automotive management consultancy
AMC, together with Robert Hofmann, 3D-Core, LBK,
and Covestro. The seat weighs in at around 10 kg, which
is claimed to make it 20 % lighter than competitors’ pro-
ducts. A wrapping process for fiber composite compo-
nents, which was developed by AMC, is responsible
for the majority of the weight reduction. This was

© csi
used by csi Entwicklungstechnik to create the
seat structure. The process is known as xFK in
3-D and produces robust components from a
continuous fiber impregnated with resin. In
addition to the weight aspect, the prototype has
also demonstrated the possibilities of seat develop-
ment beyond the boundaries of conventional modular
systems. The seat concept would be an ideal solution for
hypercars and the air taxis of the future, for example.

© Peugeot Motorcycles
© Porsche
© Daimler

Mercedes-Benz | Porsche | Peugeot

New Diesel Engines
for the Vito
The OM 651 engines in the Mercedes-Benz
Vito were replaced by OM 654 diesel engines.
The 2-l, four-cylinder engines are available
with three power ratings. The 114 CDI produ-
ces 100 kW and delivers a maximum torque
of 330 Nm. The 116 CDI puts out 120 kW
and 380 Nm. In the 119 CDI, the figures are
140 kW and 400 Nm. The innovations include
the combination of aluminum cylinders and
steel pistons, the stepped recess combus-
tion process, and the Nanoslide bore coa-
ting to reduce friction within the engine.
The Euro 6d Temp engines have one oxida-
tion and two SCR catalytic converters, and
a particulate filter with SCR catalytic con-
verter functionality.
Cayenne Coupé with
V6 or V8 Engine
Porsche is launching its new version of
the Cayenne Coupé with two Euro 6d Temp
engines. The basic model has a 3-l V6 turbo-
charged gasoline engine with peak power
of 250 kW between 5300 and 6400 rpm and
maximum torque of 450 Nm from 1340 to
5300 rpm. The dual turbocharged 4-l V8
gasoline engine in the Cayenne Turbo Coupé
puts out 404 kW from 5750 to 6000 rpm
and 770 Nm between 2000 and 4500 rpm.
Both versions have an eight-speed automatic
transmission and all-wheel drive as standard.
Motorcycles | Euro 4
Two-stroke Engine
Peugeot has introduced a new two-stroke
Euro 4 gasoline engine with a displacement
of 49.9 cm³ in almost all its scooters in the
50-cm³ class. The single cylinder engine
produces its maximum power of 3.4 kW at
7000 rpm and maximum torque of 4.6 Nm
at 6700 rpm. The new engine has an elec-
tronic injection system and Nikasil-coated
aluminum cylinders and graphite coated
pistons. This leads to a reduction in friction
losses and a drop of 20 % in fuel consump-
tion and almost 50 % in oil consumption.
Other new features include an increase in
the alternator output from 90 to 160 W and
a reinforced drive belt.

ATZ worldwide 06|2019 33

 NE WS PRODUCT S
Renault-Nissan-Mitsubishi |
Cloud Solution for Connected Services
© Renault
© Unsinn
© Audi
Unsinn |
Towing Dynamometer
for Road Trials
Unsinn has developed the Unbrake towing
dynamometer to allow road tests to be carried
out under specific loads and at different
speeds. It is fitted with a digitally controlled
permanent brake system with water cooling
and not only ensures that the test results are
reproducible but also records the effects on
the car’s cooling system, the thermal capacity
of the engine compartment, and the air condi-
tioning. The towing dynamometer can be used
to test cars, light commercial vehicles, and
trucks weighing up to 40 t. It is available with
a maximum tractive force of 8 to 15 kN and
a power range of 125 to 300 kW at speeds up
to 160 km/h.
34
© Volvo
Audi |
More Power for
the Natural Gas A3
Audi has given its A3 Sportback g-tron
model a more powerful engine. The 1.5-l,
four-cylinder engine with direct injection
has a modified cylinder head, turbochargers,
injection system, and catalytic converter to
allow it to run on natural gas. The SI engine
produces a maximum power of 96 kW and
200 Nm of torque. It generates 15 kW more
power than its predecessor, while the torque
figure remains unchanged. The front-wheel-
drive compact car comes as standard with
a seven-speed dual-clutch transmission.
Another new feature of the natural gas A3
is a third gas tank with a capacity of 4 kg,
which takes its maximum gas capacity to
17.3 kg. The gasoline tank has been reduced
in size accordingly from 50 to less than 10 l.
The Renault-Nissan-Mitsubishi Alliance
is developing new cloud solutions for con-
nected services. The first vehicles produ-
ced with Alliance Intelligent Cloud techno-
logy will be the new Renault Clio and the
Nissan Leaf. The system will use cloud,
artificial intelligence, and internet of
things technologies from Microsoft Azure.
It will be available in nearly all the 200
markets where the Alliance has a pre-
sence. The platform captures, manages,
and analyzes the vast volume of data
generated by connected vehicles and uses
the data to provide new services. These
include the car’s Internet connection,
connected navigation, accurate remote
diagnostics, ongoing over-the-air soft-
ware and firmware updates, and access
to personalized infotainment services.
Volvo |
Key Increases
Safety Levels
In the future a new key will enable Volvo
owners to set a maximum speed limit when
they lend their car to family members, friends,
and, most importantly, newly qualified drivers.
The Care Key will come as standard with all
Volvo cars from model year 2021. Volvo Cars
intends to send out a strong signal about
the dangers of speeding, after announcing
that it will limit the top speed on all its new
models to 180 km/h from mid 2020. In addi-
tion to the safety benefits, the speed limit
and the system could also bring financial
advantages for Volvo drivers if insurers
consider special premiums to drivers of
cars fitted with safety features of this kind.
Volvo expects to announce the first agree-
ments with insurance companies soon.
 Lightning Motorcycles | CONFERENCES THAT COULD INTEREST YOU:

More Affordable Electric Motorcycle

UPCOMING CONFERENCES

chassis.tech plus
June 25 and 26, 2019 | Munich (Germany)

Electrified Mobility |
Grid Integration of Electric Mobility
© Lightning Motorcycles

November 12 and 13, 2019 | Stuttgart (Germany)

CALL FOR PAPERS

The Powertrain of Tomorrow
Deadline for submission proposals:
June 13, 2019
The US manufacturer Lightning Motorcycles has launched the Strike, a new and more
affordable motorcycle with electric drive. It comes in three variants: Standard, Mid
Range, and Carbon Edition; these have different power outputs and battery sizes. The
electric motor in the Standard and Mid Range variants generates a maximum of 67 kW
and 244 Nm of torque. The Carbon Edition has a power output of 98 kW and 244 Nm of
torque. In the Standard model, the battery capacity is 10 kWh, while in the Mid Range it
is 15 and in the Carbon Edition 20 kWh. This top-of-the-range variant comes as standard
with a 6.6-kW charger. The other two versions are fitted with a 3.3-kW charger, but the MORE INFORMATION AT
more powerful charger is available as an optional extra. www.atzlive.com

Fraunhofer | Tires Made from Synthetic Rubber

Researchers from several Fraunhofer
institutes have developed and tested
a synthetic rubber product called Bisyka
with the aim of making more abrasion-
resistant tires. Firstly they identified
the organic components of natural rubber
that are important for abrasion behavior
and then they synthesized the Bisyka
rubber from the relevant biomolecules
and functionalized polyisoprene with a
high microstructural purity. Four car tires
were manufactured with a tread made
from Bisyka and they were then compared
with conventional tires in the test lab.
The results showed that tires made of
the synthetic rubber lost 30 % less mass
than equivalent tires made of natural
rubber. In addition, the synthetic tires
had only half the tread loss. The rolling
© Till Budde | Fraunhofer IAP

resistance of the synthetic rubber was

also better. The formula of the tread
compound for truck tires is currently
being adapted to the new rubber. The
researchers are now looking for partners
who will bring the product to the market.
ATZ worldwide 06|2019 35
DE VELO PMENT TR ANSMIS SIONS

© Magna

Flexible Oil Management for

Dual-clutch Transmissions WHY FLEXIBLE OIL CHANGE
INTERVALS?

Usual fixed-service intervals for auto-

matically shifted transmissions are
intended for loads [1] that most driv-
ers hardly ever come close to. This
approach ensures safety margins for
use cases like regular trailer opera-
Over the lifetime of a vehicle, variable service intervals are tion, high heat, humidity, demanding
topography, and very sporty drivers.
one key to decrease costs, lighten weight, and save resources.
The drawback is that the transmission
The flexible lubrication oil management by Magna Powertrain oil fill volume is over-dimensioned for
and BMW is created for automatically shifted transmissions. average drivers, thus increasing weight,
CO2 emissions and service costs.
It is based on real driver’s behavior and enables smaller oil On-demand oil management [2]
fill volumes as well as extended oil change intervals. offers advantages, especially for
automatic transmissions that use
wet clutches. Manual transmissions
for passenger cars, usually with dry
36

clutches, need just approximately
1.5 l oil for wheel set lubrication as
life-time operation. In automatic trans-
missions, life-time oil filling would
increase weight and consequently
fuel consumption. Another reason
why flexible oil intervals become
more relevant is that engine torques
have risen about 35 % through the
last 15 years. This effectively means
many car drivers are pushing the
load limits of their powertrains to
an even less extent.
A prerequisite for a flexible oil man-
agement is to know the lubrication oil
condition as precisely as possible. Such
a management, developed by Magna
Powertrain and BMW for dual-clutch
transmissions, is based on an advanced
load monitoring and damage prediction
ATZ worldwide 06|2019
FIGURE 1 Oil change intervals and oil fill volumes of the 7DCT300 compared to common ATs, CVTs and DCTs
(© Magna)
models. They enable reduced oil fill
volume and extended change intervals
for most car drivers.
EXTENDED OIL CHANGE INTERVALS
FIGURE 1 illustrates typical oil change
intervals as well as the advantages of
the flexible oil management. The con-
sideration includes the typical intervals
of transmissions types like Torque-con-
verter Automatic Transmissions (ATs),
Continuous Variable Transmissions
(CVTs) and Dual-clutch Transmissions
(DCTs) for a Front-wheel Drive (FWD).
The shaded upper part of the AT column
in FIGURE 1 (left) represents transmis-
sions with a claimed lifetime filling
for normal drivers, being assumed for
240,000 km in this illustration. At higher
AUTHORS
Dipl.-Ing. Martin Bahne
is Senior Director Functional
Integration & Testing at Magna
Powertrain Transmission Systems in
Untergruppenbach (Germany).
Dr.-Ing. Andreas Wiesebrock
is Technical Attribute Leader
Durability at Magna Powertrain
Transmission Systems in
Untergruppenbach (Germany).
loads (“hard” drivers), however, ATs
require an earlier oil change.
Regarding the range up to the next oil
change in relation to the oil fill volume,
the AT performs best, and the CVT and
DCT are second on a similar level. The
AT´s advantage is primarily based on
the hydrodynamic torque converter that
does not have to handle frictional power.
However, thanks to the flexible oil man-
agement and a specifically optimized
oil, the 7DCT300 by Magna Powertrain
is on par with the best ATs. FIGURE 1
(right) shows the required oil fill vol-
ume of these three types. The 7DCT300
needs significantly less oil fill volume
compared to the best ATs.
The flexible oil management has
been in series use since 2017. Typical
oil change intervals are between 60,000
37
DE VELO PMENT TR ANSMIS SIONS
and 125,000 km. BMW succeeded
in extending oil change intervals by
approximately 50 %, while reducing
the oil fill volume by 35 % compared to
mean values of DCTs. A large majority
of car owners profit from the flexible
oil management. Just a small minority
of “hard” drivers must accept shorter
change intervals like usual DCTs.
FINDINGS FROM
REAL-LIFE OPERATION
The potential of the flexible oil manage-
ment becomes obvious when looking at
real driving data. Driving profiles from
more than 2000 owners of a BMW C-seg-
ment passenger car show a wide range
of different loads, FIGURE 2. Here, the
transmission input power varies signifi-
cantly under different boundary condi-
tions like city or constant driving, rais-
ing slopes, sporty driving, etc. The aver-
age transmission input power was just
12.5 kW (including coasting phases).
Within one year, the average of drivers
FIGURE 3 Modular software model in the transmission control unit (© Magna)
38
reached just 10,000 km (6214 mi), with
a minority reaching smaller or larger
mileages. The average speed was just
37 km/h (23 mph). Only a few drivers
regularly drove at very high speeds
and accompanying high loads.
Loads within the torque path and clutch
system can vary by a factor of 10, depend-
ing on individual driving style and condi-
tions. The field data shows that fixed ser-
vice intervals inadequately represent the
behavior of most drivers. This explains
the huge potential for smaller oil fill vol-
umes and extended change intervals.
OPERATING PRINCIPLE OF
THE MONITORING SOF TWARE
A modular built-up software model
within the Transmission Control Unit
(TCU) records the transmission loads,
and processes and interfaces the data
usage, FIGURE 3. The data acquired from
the transmission are transferred in the
module model input (green) to the dam-
age model (orange) of the load monitor-
FIGURE 2 Driving profiles
from more than 2000
owners of a BMW
C-segment passenger
car (© Magna)
ing via a CAN interface. Here, they are
matched with a prediction and reaction
model. The boundary conditions control
when both models are activated each.
The next diagnostics module (blue)
is responsible for checking the results
for integrity and plausibility, and for
carrying out a so-called de-bouncing.
This function dampens the recording
of load peaks – primarily in early life-
time – to prevent premature data inter-
pretation. This can be compared to
the average fuel consumption displayed
shortly after fueling, which obviously
cannot be precise. However, over lon-
ger use, the software learns to classify
the transmission load data correctly,
making de-bouncing unnecessary.
After all, the driver interface (red)
is needed to provide the data for exter-
nal use on several channels. This can
be the warning light, the service dia-
gnosis interface (OBD) or an Over-the-
Air (OTA) interface for real-time data
transfer in communication processes
between customer, garage and OEM.
INTERACTION OF PREDICTION
MODEL AND REACTION MODEL
The prediction and the reaction model
in the damage model of the TCU can
complement each other usefully, FIGURE 4.
The prediction model’s operating princi-
ple is to match consolidated knowledge
regarding load and damage with the data
being delivered via CAN. These are for
example temperature, engine torque and
engine speed, as well as transmission
internal and component-specific data.
Today´s prediction model accuracy is
quite precise, because there is comprehen-
 FIGURE 4 Interaction of the prediction model and the reaction model (© Magna)
sive experience from development, calcu-
lation and real driving. However, predic-
tion models have no insight into the real
condition of components or lubricants; the
prediction is made solely by matching
knowledge and vehicle data that can be
used indirectly. For example, it is known
which load and consequently damage
engine torque will probably cause. This
can be cumulated in the model level.
The reaction model in contrast relies
on measured changes or degradation
of a component using dedicated sensors.
Therefore, it can take component toler-
ances into account or detect unusual
wear. This is why this model delivers
measurement data from the begin-
ning, whereas the prediction model,
as described earlier, starts to evaluate
lifetime prediction after a calibration
phase. Accuracy will increase with life-
time. Lifetime prediction become even
better by combining both models, and
FIGURE 5 The quality of the prediction increases with the accuracy of the damage model (© Magna)
ATZ worldwide 06|2019
thus allowing smaller safety factors
to be defined in the model.
The principal correlation between
loads and cumulated damage is exem-
plarily shown in FIGURE 5. The solid
blue line within the Stress-number (S-N)
diagram represents the maximal load
which can be cumulated before damage
occurs. The solid green line stands for
the load spectrum when operating an
exemplary vehicle; it lies well below
the maximum load. The dotted green
line represents a higher mileage and
shows how load converges to maximum
load over time based on load monitoring.
Before it meets the blue line, mainte-
nance should be carried out.
The more exact and filigree the
damage models are, the more pre-
cisely the right timing for oil change
can be predicted. The red line stands
for too roughly classified loads and
thus less precise prediction.
BENEFIT FOR FUTURE
APPLICATIONS
Weight optimization for CO2 reduction
is a continuous challenge [3]. As a part
of a “design-to-target,” Magna Power-
train consequently and regularly makes
damage calculations of its products.
Regarding transmissions, this applies
for gear wheel sets, clutches and oil,
etc. The more extensive the knowledge
of customer use load spectra is, the
more precise the prediction models can
be designed. In the next DCT generation,
the torque to weight ratio will be another
25 % better through improvements in all
these areas. The flexible oil management
is a major contributor to that.
Beyond that, the oil management is
a valuable element to reduce weight of
module-based and hybrid drivetrains [4].
One reason why Magna Powertrain
favors the P2.5 architecture is that the
39
DE VELO PMENT TR ANSMIS SIONS
FIGURE 6 The maximum-torque-to-weight ratio of the 48-V hybrid transmission 7HDT400 is close to an AT without hybridization
(© Magna)
small high-speed e-machine causes com-
paratively little weight and no installa-
tion length. In combination with flexible
oil management, the new hybrid trans-
mission 7HDT400 (48 V) offers a maxi-
mum-torque-to-weight ratio close to
an AT without hybridization, FIGURE 6.
When increasing electrification further,
the advantage becomes even more obvi-
ous: Within one vehicle platform, the
whole range from 48-V to high voltage
hybrid drives can be implemented with-
out having to modify the installation
position or the front axle design due to
significantly more weight.
POTENTIALS THROUGH
COMMUNICATION
Much of the benefits of the flexible oil
management can be opened with pres-
ent technical means. When oil change
is due, the car driver can be informed
by a classic display. Furthermore, tech-
nical methods are available that practi-
cally rely on real-time communication.
BMW has been offering tele services
for many years, which enable transfer-
ring maintenance data between driver
or vehicle and the OEM.
As described before, the monitor-
ing software offers interfaces that can
be used for any OTA service models.
While GSM is functionally restricted
40
in terms of data size and speed, newer
protocols like 4G or 5G in future allow
further improvements to communica-
tion quantity and quality. One addi-
tional benefit is to gain more and more
precise data from the field, thus being able
to further improve prediction accuracy.
An increasing number of users will
improve quality as additional benefit.
Another attractive option is offering
new service models to the car owners
that include their familiar communica-
tions end devices. For example, instead
of the simple message that a service is
due, the OEM could provide car drivers
with comprehensive feedback, how ser-
vice costs could be reduced. Besides,
service packages priced specifically to
customers can be offered which reflect
individual needs.
SUMMARY
Flexible oil management based on a
prediction and reaction damage model
allow BMW and Magna Powertrain for
reducing fill volumes for a transmission
up to 35 % as well as stretching change
intervals by approximately 50 %. The
reduced weight contributes to less CO2
emission, and longer intervals save aver-
age service costs for car owners.
The benefit of flexible oil management
grows with future communication solu-
tions. This will allow for further opti-
mization of the models, improve com-
munication between the OEM and end
customer, and make possible personal
service offerings to car drivers.
The next load monitoring software
generation will contain new subsys-
tems to use take advantage of these
opportunities to full effect. The more
detailed prediction and reaction model
will allow for even longer oil change
intervals, and one is prepared with flexi-
ble software interfaces for future OTA
communication concepts.
REFERENCES
[1] Krieger, H.: Alterung von Schmierstoffen
im Zahnradprüfstand und in Praxisgetrieben.
Munich, Technical University, dissertation, 2004
[2] Bahne, M.; Wiesebrock, A.: Flexible
Oil Change Management for Dual Clutch
Trans missions. 17 th CTI Symposium, Berlin,
December 3-6, 2018
[3] Bahne, M.; Frey, U.: Transmission Technology
Contribution to CO 2 Roadmap – a Benchmark.
39 th International Vienna Motor Symposium,
April 26 and 27, 2018
[4] Gindele, J.: A Comprehensive System Approach
for Future Powertrain Application. 12 th CTI Sym-
posium USA, Novi, May 14 to 17, 2018
 Two conferences
together under one roof
Electrified Mobility
Digitization and Electrification for Sustainable Mobility
2nd International ATZ Congress

/// KINDLY SUPPORTED BY

Grid Integration of Electric Mobility

Customer orientation - a key to success
4th International ATZ Conference

/// KINDLY SUPPORTED BY

12 and 13 November 2019 | Stuttgart | Germany

Phone +49 611 7878-131 PROGRAM AND REGISTRATION

Abraham-Lincoln-Straße 46
65189 Wiesbaden | Germany
ATZ worldwide 06|2019
Fax +49 611 7878-452
ATZlive@springernature.com www.ATZlive.com
41
DE VELO PMENT TR ANSMIS SIONS

© AVL

AUTHORS

Dipl.-Ing. Mathias Deiml
is Center Lead for Electrified Drive
Innovation at AVL Software &
Functions GmbH in Regensburg
(Germany).
Tommie Eriksson Matthias Schneck, M. Sc. Dr.-Ing. Antoine Tan-Kim
is System Engineer is System Engineer at is E-Machine Development
at AVL MTC AB AVL Software & Functions GmbH Engineer at AVL Software &
in Gothenburg (Sweden). in Regensburg (Germany). Functions GmbH
in Regensburg (Germany).

AN ALTERNATIVE TO ernment incentives, an ecological faculties in BEVs and Plug-in Hybrid

CONVENTIONALLY
POWERED VEHICLES
The Battery-electric Vehicle (BEV) is on
its way to greater production volumes
and enjoys more and more customer
acceptance. Established BEVs such as the
Tesla models or newly introduced mod-
els such as the Mercedes-Benz EQC or
the Jaguar I-Pace compete against com-
bustion-engine cars as a real alternative.
These vehicles have a high driving range
(> 400 km), good driving characteristics
and a good charging performance. Gov-
42
image in times of driving bans in
Germany, and customer demand for
latest high-technology are additional
volume drivers for BEVs in Europe
and worldwide.
Nevertheless, this generation of pre-
mium battery-powered vehicles will
probably be still too expensive to make
it into comparable volumes with conven-
tional technology. So, future electric
vehicle prices need to be more competi-
tive compared to combustion technology.
The reduction in cost is therefore the
development goal for all engineering
Electric Vehicles (PHEVs). This article
presents a way to reduce the cost of elec-
tric powertrains.
Integrated Electric Drive Units (EDUs)
are already widely in use. They integrate
the electric machine, the transmission,
the differential, and, in most cases, the
power inverter in one easy-to-install
unit. An EDU is characterized by its
torque capability at the wheel(s), power
output, and maximum vehicle speed.
For the development described, the
experts defined such data for a premium
passenger car of the upper mid class
 High-speed Electric Drive Unit
for the Next Generation of Vehicles

For developers of hybrid and electric vehicles, significant cost

reductions are among the most important development goals.
With its newly developed integrated electric drive unit, AVL has
found a way to cut the cost of electric powertrains.

with a start of production in 2022. Scal-
ability to other vehicle classes was a
clear development target. AVL assumed
a 2-t rear wheel driven car with well
dimensioned 5000 Nm axle torque.
A trend for sophisticated BEVs, that
can be observed on the market, is torque
vectoring. Each wheel can be driven and
controlled separately to get a traction
advantage in curves. For this reason, the
EDU presented here consists of two drive
systems, with 150 kW and 2500 Nm on
each side. The maximum vehicle speed
was targeted to 240 km/h in order to
match usual specifications of premium
combustion engine vehicles.
A system voltage range around 800 V
was selected for this EDU. The higher
voltage offers the advantage of reducing
the weight of conducting materials on
the one hand and the chip area in the
inverter on the other. Both effects have
the potential to lower costs of inverter
and wiring [1]. The charging power can
be increased despite current limitations
on the charge connector.
The costs of the electric machine can
be decreased by reducing the size and
ATZ worldwide 06|2019
thus the material costs. To reach the
required power despite the smaller size
and as a result lower torque, the speed
has to be increased as shown in Eq. 1:
Eq. 1 Pmech = ω *  M;  M ~  d 2 * l;   
1 To fulfill high efficiency, the number
Volume ~  __
ω
d : airgap diameter;  l: lamination length
Provided that special or expensive tech-
nologies in machine and transmission
can be avoided, there is a substantial
cost saving in the machine. Weight and
size of the machine can be reduced
accordingly, which has an additional
beneficial effect on the vehicle. These
relations between speed and weight/
volume/cost are shown in FIGURE 1.
The AVL design stays left of the point in
speed from where the costs are increas-
ing again because of special bearings,
lamination material, windings, etc.
When increasing machine speed,
influence on transmission and inverter
must be observed. Transmission design
should be simple and highly efficient.
Shiftable two-speed transmissions offer
the possibility to select the most efficient
operation points of the electric drive,
but they also increase complexity,
weight and costs. Using a single-speed
transmission, the high-speed e-machine
achieves both the desired wheel torque
and vehicle top velocity.
of stages was limited to two, resulting
in a transmission ratio of about 16:1.
With this ratio the e-machine reaches
about 30,000 rpm at vehicle top speed
of 240 km/h. Advantages in the most
important properties efficiency, simplic-
ity and costs lead to the use of a two-
stage lay shaft transmission.
ELECTRIC MACHINE
To reach a high power density, a Per-
manent Magnet Synchronous Machine
(PMSM) was chosen. Magnets without
expensive heavy rare earth materials
such as Dysprosium (Dy) and Terbium
(Tb) are used to reduce the cost of the
machine. The magnets are arranged in
a simple V-shaped single layer structure,
FIGURE 2.
Regarding high-speed machines, one
of the first concerns is the mechanical
43
DE VELO PMENT TR ANSMIS SIONS
Weight/
volume/
costs
100 %
Target:
Find optimum for low weight,
low volume and minimal costs
FIGURE 1 Speed versus cost (© AVL)
strength of the rotor, which is linked to
the maximum peripheral speed of the
rotor. The current design reaches a maxi-
mum speed of 155 m/s, exceeding most
automotive e-machines. Based on the
mechanical stress and the elastic defor-
mation, the rotor geometry was opti-
mized to achieve a robust design without
using costly materials such as cobalt-
based steel sheets, FIGURE 3.
A high rotation speed results in
high losses in the different parts of
the machine. The first step to mitigate
these high-frequency losses is the
appropriate choice of the number of
pole pairs. A three-pole-pair machine
was chosen as the best compromise
between an optimal magnetic circuit
and a moderate maximum electrical
frequency of 1500 Hz.
FIGURE 2 Flux density in the machine at the corner point
(peak operation) (© AVL)
44
Increasing costs
due to expensive technology
E-machine
max. speed
To reduce the high-frequency losses in
the windings, a distributed round wire
winding was preferred to a hairpin wind-
ing. Thin wires are used to prevent skin
effect even at maximum speed. However,
the influence of copper wires on one
another and flux leakages across the slot
lead to an inhomogeneous current density
distribution in the wires and result in
increased Joule losses [2]. A closed stator
slots helps reducing these losses by cana-
lizing the magnetic flux and thus reduc-
ing the flux crossing the slot. The wires
are also pushed away from the airgap,
that is towards the outer diameter, where
they are less affected by flux leakage.
FIGURE 3 shows the FEM simulated cur-
rent density in the conductors.
Iron losses are the sum of hysteresis
losses, which increase linearly with the
FIGURE 3 Mechanical simulation (von Mises stress) at spinning
speed 36,000 rpm (left), current density in the conductors of
one slot at the maximum rotation speed (right) (© AVL)
frequency, and eddy current losses,
which increase quadratically with the
frequency. They are predominant in
an electrical machine at high rotation
speeds. In order to reduce them, a thin
0.2 mm lamination was selected to offer
the best compromise between cost and
manufacturability.
Like the laminations, permanent
magnets are also prone to eddy current
losses. An axial segmentation of magnets
was implemented to significantly reduce
them.
Mechanical losses come from bearings
and windage losses. Both were reduced
thanks to a specific bearing design and a
smooth rotor outer surface to reduce fric-
tion in the airgap.
For efficient heat dissipation, a direct
oil cooling system such as the one inte-
grated in [1] was implemented, but in a
more advanced version. The coolant is
pumped through the stator slots in the
axial direction to directly cool the wind-
ings. The oil is kept inside the stator
thanks to closed slots. This solution
spares the cost of an additional airgap
tube and, at the same time, significantly
reduces the airgap width, the cogging
torque, and the torque ripple.
FIGURE 4 presents the complete
efficiency map of the machine. The
10-s-peak and the continuous torque-
speed curves show the thermal limits
of the machine. The torque of the high-
speed machine reaches 160 Nm (peak)
and 75 Nm (continuous) up to 9000 rpm,
FIGURE 4 High-speed e-machine
efficiency map at 80 °C copper
and magnet temperature (© AVL)
corresponding to a power of 150 kW
(peak) and 75 kW (continuous). The
limits of this machine could of course
be scaled for another application
depending on the cooling system, the
available battery voltage, and the maxi-
mum inverter current.
At the maximum speed of 30,000 rpm,
the simulations show a maximum avail-
able power of 125 kW, which still rep-
resents almost 85 % of the peak power.
Furthermore, thanks to the applied mea-
sures to reduce the high-frequency losses,
ATZ worldwide 06|2019
the machine offers a large high-efficiency
operating range above 95 %.
INVERTER
The integrated dual inverter has a com-
mon DC-link capacitor, Silicone Carbide
(SiC) modules and is capable of a rms
phase current up to 525 A to fit the e-ma-
chines and a DC voltage up to 820 V.
Thanks to the low switching losses of
the SiC power modules, it can use vari-
able switching frequencies up to 30 kHz.
FIGURE 5 High-speed EDU with integrated
dual inverter (© AVL)
The high switching frequency enables
a high fundamental frequency and a
decreased DC-link capacity. In addition,
the dual inverter utilizes an interleaving
strategy making it possible to reduce
DC-link capacitor size, emissions, and
voltage ripples further. With this strat-
egy, it is possible to reduce the DC-link
capacitor size down to 65 %. FIGURE 5
shows the dual inverter integrated into
the EDU.
Further advantages of the SiC technol-
ogy are improved efficiency, good coola-
bility due to higher thermal conductivity,
higher operating temperatures, and a
smaller package size. The higher costs
for SiC modules are reimbursed on sys-
tem level by enabling a very compact
inverter package.
To reach the power at the highest
speed, the e-machine design requires
a slightly oversized inverter in terms
of voltage and current. By reducing the
power requirement at the maximum
e-machine speed for other applications,
the inverter current and/or voltage
requirements can be decreased. On the
other side, due to the chosen design,
the back Electromotive Force (EMF) is
low, so that an active short circuit is not
needed as a safety measure in the sys-
tem. The dual inverter can control the
two e-machines independently from
each other and thus enables torque
vectoring.
45
DE VELO PMENT TR ANSMIS SIONS
TRANSMISSION
FIGURE 6 shows the two single-speed,
two-stage layshaft transmissions inte-
grated in one housing. For torque vector-
ing, the two transmissions can be oper-
ated separately.
Lubrication and cooling are key param-
eters for high-speed helical gears. As
the pitch line velocity increases, these
requirements need to be met for proper
and reliable operation of the gears. Forced
lubrication was chosen because of the
advantage to increase control over the oil
flow rate to each component and higher
efficiency compared to splash lubrication.
Two of the key parameters for a forced
lubrication are gear design and the posi-
tioning and type of the nozzle.
Gear design parameters include pitch
line velocity, helix angle, axial meshing
velocity, face width, pressure angle, and
surface roughness. Pitch line velocity is
the linear speed at the pitch diameter of
a gear. The helix angle has a direct effect
on the transmission noise level.
The overlap/face contact ratio depends
on the helix angle and the gear width.
A large helix angle is used because it
decreases the axial meshing velocity
and enables a smaller face width. Air
and oil trapped in the gear mesh will
be pumped out at lower speeds, which
means less friction heating and lower
temperature of gear tooth. However,
high helix angles generate higher
axial forces that must be supported
by a bearing.
Lubrication nozzles can be placed in
such a way that they spray oil into the
in-mesh or out-mesh or both. The general
theory is that spraying oil into the in-
mesh is good for gear lubrication and
spraying oil into the out-mesh improves
46
the gears cooling. A high flow rate and
pressure might be needed for gears
with high rotational speed to make
sure that oil can reach the gear
tooth face. In order to decide on
the nozzle size and the oil
flow rate, it is necessary to
calculate the heat produced
by the gears. It is important
that oil covers the gear tooth
surface for a short time
interval before it is being
slung off.
BEARINGS FOR
HIGH SPEED
SKF was chosen as the bearing part-
ner. For current SKF bearings, the
speed factor (ndm) is around 0.7 Mndm.
The speed factor is the multiplication of
the bearing mean diameter by the rota- FIGURE 6 Layout of the two single-speed,
tional speed. To withstand the speeds up two-stage transmissions (© AVL)
to 1.6 Mndm, which are necessary in
high-speed applications, SKF has devel-
oped a new design for their sealed bear-
ings. The cage has been redesigned and REFERENCES
lubrication optimized. [1] Engstle, A.; Deiml, M.; Angermaier, A.; Schelter,
W.: 800 Volt for Electric Vehicles – Voltage Level
Suitable for Calibration. In: ATZworldwide 115
(2013), No. 9, pp. 39-43
SUMMARY AND OUTLOOK
[2] Popescu, M.; Dorrell, D. G.: Skin effect and
The high speed EDU was designed for proximity losses in high speed brush-less perma-
nent magnet motors. IEEE Energy Conversion Con-
a maximum speed of 30,000 rpm and gress and Exposition, Denver, 2013
160 Nm in the e-machine. Roller bear-
ings and standard dynamo sheet metal
were used to operate at high speed.
The magnets use no heavy rare earth
materials. Its small size lowers costs
in the machine and supports integra-
tion into various vehicles. Based on
the same technology, EDUs with lower
THANKS
requirements can be derived. These The authors would like to thank Peyman Jafarian,
can help to increase the market share Mario Brunner (both AVL) and SKF partner Anthony
of EVs further. Simonin for contributing to this paper.
 chassis tech . plus

10th International Munich Chassis Symposium

25 and 26 June 2019 | Munich | Germany

NEW CHASSIS SOLUTIONS

Urban mobility
drives innovations

AUTOMATED DRIVING
AI conquers the chassis

STEERING TECHNOLOGIES
Steer-by-wire offers
new opportunities

/// PLENARY LECTURES

Bosch Engineering GmbH |

Honda R&D Co., Ltd. | IHS Global SRL |
Dr. Ing. h.c. F. Porsche AG |
TRE GmbH | Volkswagen AG |
WiTech Engineering GmbH

/// ONE FOR ALL

Four congresses
in one event

/// PARTNER /// SPONSORS

Phone +49 611 7878-131 PROGRAM AND REGISTRATION

Abraham-Lincoln-Straße 46
65189 Wiesbaden | Germany
Fax +49 611 7878-452
ATZlive@springer.com www.ATZlive.com
ATZ worldwide 06|2019 47
DE VELO PMENT HMI
© YFAI
CHANGES IN VEHICLE INTERIORS
Over the next ten years, automobile interi-
ors will change more than in the previous
five decades. A paradigm change is taking
place from self-driving and use of time to
overcome distances, to a partially or fully
automated driving experience. Electrified
drive concepts also create flexibility in
the interiors of future vehicle generations,
both for the design of individual mobility
as well as for implementation of innova-
tive shared mobility concepts. Though in
future, vehicles and passengers will also
be able to remain in constant contact with
the digital world.
Emotional ties to the vehicle will be
governed more by passengers’ experience
during its use than by the vehicle as an
object in itself. This so-called user experi-
ence includes all aspects of the user’s
interaction with the vehicle as a product.
INITIAL SITUATION
Yanfeng Automotive Interiors (YFAI)
and Kostal have set themselves the goal
of further developing vehicle interior
surfaces, which up to now had a purely
decorative character, into seamlessly inte-
grated mechatronic user interfaces. These
intelligent surfaces or smart surfaces pro-
vide new flexibility for the design of auto-
motive experience and can be adapted to
OEM-specific control concepts.
48
Understanding of the interaction of
functionality, appearance, feel, struc-
ture, electronics and software as well
as seamless integration of interior sur-
faces and personalized human-machine
interfaces pave the way for the automo-
bile interior experience of the future.
Responding to this trend, YFAI and
Kostal launched a strategic partnership
in 2016 [1]. Collaboration by complemen-
tary suppliers is a proven accelerator of
research and development as well as for
the production engineering implementa-
tion of innovative solutions in the auto-
motive industry.
YFAI and Kostal pool their know-how
in this strategic partnership: The interior
expert YFAI contributes its experience
in surface technologies and integration
expertise for automotive interiors. The
mechatronic specialist Kostal provides
its strength as a leading supplier of new
switching technologies such as capac-
itive touch sensors and electronics as well
as its Human Machine Interface (HMI)
competence. Besides the potentials of
technological integration, this combina-
tion of the expertise of both companies
provides automobile manufacturers sig-
nificant advantages due to the reduction
of organizational interfaces along the
value chain of smart surface products.
After less than one year of the partner-
ship, both companies presented their
joint concept, the so-called Yanfeng-
Kostal module, at the IAA 2017. Since
then, cooperative teams from both com-
panies have been working on various
pre-development projects for the automo-
tive industry in an interactive develop-
ment process.
The following article provides insights
into the seamless integration of func-
tions into surfaces with both hard and
soft decor.
SEAMLESS INTEGRATION OF
FUNCTIONS INTO SURFACES
For seamless integration of surfaces and
HMI technologies, product and produc-
tion engineering know-how from both
a design and mechatronic point of view
is essential.
Regarding large surfaces in vehicles,
a differentiation can be made between
hard and soft decor. These days, with
decreasing importance in terms of quan-
tity, molded in color, painted, foil cov-
ered (usually back- or over-molded films)
as well as authentic wood and aluminum
decor are used for hard decorative ele-
ments. In the upper and middle vehicle
segment, other, differentiating alterna-
tives such as glass will increasingly be
used. For the usually larger, soft decora-
tive surfaces of the instrument panel and
door trims of the vehicle, depending on
the different segments, leather and artifi-
cial leather are wrapped, formed PVC,
 AUTHORS

Surface Innovations for

Interiors of Future Vehicles Dr.-Ing. Dirk Blomeyer
is Director Business Incubator –
Smart Interior Surfaces at Yanfeng
Automotive Interiors (YFAI)
in Neuss (Germany).

A cooperation between Kostal and YFAI is intended to synergetically

combine the HMI and interior technology areas in order to offer
market-ready solutions for intelligent surfaces in the short term and,
at the same time, to reduce the organisational complexity for OEM
Dr.-Ing. Anna-Lena
customers. Instead of electronic modules that are installed in Schulte-Gehrmann
is Head of Cooperation
the vehicle interior and then covered, the partners are developing Management in the Product Line
Switch Panels Sensoric at Kostal in
new products from a single source. Lüdenscheid (Germany).

PU or TPU skins are foamed in place, or
PVC, TPO, TPU foam-foils or textiles are
laminated on substrates.
In the past, such decorative surfaces
were visually and functionally clearly
separated from the control and display
elements of the vehicle. This separation
was ultimately due to the use of different
design and material languages. In future
interior concepts, this separation will
become increasingly blurred, that is to
say control and display elements will be
seamlessly integrated into decorative sur-
faces. This symbiosis is taking place to
the extent that HMI elements such as
switches or functional and ambient light-
ing are integrated by the use of secret-till-
lit technologies and can disappear behind
the surfaces until they are needed by the
driver. With this, almost all soft or hard
surfaces in vehicle interiors can be trans-
formed into intuitive, digital interaction
surfaces. In addition to capacitive switch-
ing technologies and displays, actuators
for acoustic and/or tactile feedback are
now used to functionalize surfaces in
order to make the control experience visi-
ble, audible, and palpable.
A special variant of this is the linkage
of the analog and digital world by use of
classic, iconic switching elements, which
are seamlessly inserted into translucent
surfaces and which interact with the sur-
rounding functionalized surfaces in an
entirely new way.
ATZ worldwide 06|2019
FUNCTIONAL INTEGRATION
INTO HARD SURFACES
In the first step, elimination of the bound-
aries between the interior and the HMI
results in the functionalization of hard
surfaces, as hard decorative elements are
established materials both in interiors as
well as for control elements. Kostal and
YFAI provide an outlook for design possi-
bilities in several application examples.
As their first joint project, YFAI and
Kostal have developed the so-called
Yanfeng-Kostal module [2]. For this center
console module, real glass was selected as
the surface. This material enables a high
quality appearance with great brilliance
and a high quality feel with a cool-touch
effect. As a special design feature, the
glass is curved in two axes. Behind the
glass surface, there is a black decorative
film, which creates a high-class, glossy
appearance. If no information is accessed,
the integrated display screen and control
surfaces remain invisible as a purely
decorative glass surface. In this way, driv-
ers are not distracted by unnecessary
information.
On the other hand, if the driver needs
information or wishes to operate func-
tions, an LCD and the so-called button
field, a capacitive switching surface, illu-
minate. Above the surface of the glass,
the button island floats as a design icon,
a slim, curved capacitive switch panel
with active tactile feedback to control the
air conditioning. This tactile feedback
enables the driver not to turn his eyes
off the road.
Up to now, this three-dimensional
shaping of large glass surfaces in all axes
and the precise positioning of the corre-
sponding symbols over the electronic
function elements is unique. The 3-D
forming process for glass provides design-
ers with complete freedom to arrange
switches and control elements, FIGURE 1.
In a functionalized real-wood decora-
tive door component, Kostal and YFAI
have combined decorative lighting ele-
ments, capacitive switching technolo-
gies, and tactile actuation with force
sensors under an authentic material.
By means of laser perforations in the
wood, control and display functions can
be implemented using the non-translu-
cent material. The capacitive switching
functions for the seat memory and seat
heater control only become visible when
the search lighting is activated. Tactile
actuation gives users palpable feedback
when the switching function is actuated
and supplements this with visual feed-
back in the form of a change of color of
the function lighting, FIGURE 2. Further
design freedom is opened up by the func-
tionalization of over-molded films with
almost any kind of decor. In a smart
decorative component for the instru-
ment panel, a 1-m-long over-molded
49
DE VELO PMENT HMI

wood-effect decorative film was enhanced

with capacitive slider functions, search
and functional lighting, and tactile actu-
ation. The latter presents a special chal-
lenge, especially with large components,
and can only be satisfactorily imple-
mented in an integrated development
process. The result is that the impulse
can only be felt in the fingers, while the
surrounding parts do not vibrate.

EX AMPLES OF
FUNCTIONAL INTEGRATION
INTO SOF T SURFACES

Integration of functions into soft surfaces

provides a significant extension of the
options for innovative control concepts.
Here, the innovation or experience effect
FIGURE 1 The freely designable 3-D glass surface seamlessly combined with new HMI technologies
is even greater than with hard surfaces in the Yanfeng-Kostal module enables the implementation of innovative operating concepts for tomorrow‘s
as, on the one hand, traditionally, control mobility in demanding design (© YFAI | Kostal)
and interaction is normally carried out
with hard switches and displays, and
passengers are surprised when these are
not necessary. On the other hand, due to reaches beyond local, discrete interaction With this concept, the old design princi-
their size, soft surfaces offer greater lati- with the surface. The entire surface ple “form follows function” can be exe-
tude for positioning control functions becomes a smart, decorative interface. cuted in completely new ways. Applica-
and for large-scale interaction of passen- In the concept study produced, functions tion demonstrators were implemented for
gers with the vehicle. Kostal and YFAI were integrated into a textile surface. The the insert area of a door as well as for an
demonstrate the possibilities of soft special feature of this technology is that extended, textile-covered armrest of a
smart surfaces in two different applica- soft, fabric-covered surfaces are now able tunnel console and were integrated into
tion examples. to interact through large areas with the the vehicle control and lighting concept.
In a smart door bolster, high quality, occupants of the vehicle and provide
soft translucent artificial leather was visual information, FIGURE 4.
DEVELOPMENT PROCESS
combined with capacitive sensor tech- With a vague, gesture-like contact
FOR INTELLIGENT SURFACES
nology and an illuminated iconic seat with the surface a functional area or
adjustment switch. The capacitive seat an ambient, decorative, or dynamic The integrated development process
memory function is concealed behind light is activated as necessary. Behind for intelligent surfaces differs from the
the artificial leather. The switching func- the water and dirt repellent textile, development of purely Electrical/Elec-
tion only becomes visible when it is a capacitive sensor film is connected tronic (EE) control elements as well
approached (secret-till-lit). By touching to a flexible LED matrix via a control as from the traditional, interior-driven
the translucent high-resolution symbols, unit. Active Skin provides a totally new development process. Typically, E/E
the switching process is initiated in the degree of design freedom for the interior. development starts earlier than interior
same way as with traditional switches.
Switching is confirmed by a change of
color and flashing symbols. When the
hand approaches, corona lighting is acti-
vated around the seamlessly mounted
seat adjustment switch. When the seat
adjustment switch is actuated, the
corona illumination dynamically dis-
plays the seat adjustment process. This
increases the impression and highlights
the clean, seamless functional integra-
tion. When it is inactive, the surface sim-
ply has a puristic design and form and
does not distract the driver with unnec-
essary function displays, FIGURE 3.
YFAI presented the Active Skin tech- FIGURE 2 The door panel trim spear demonstrates the combination of seamlessly integrated capacitive
nology at the IAA 2017 [3]. This concept switching technology and active haptics with dynamic ambient lighting (© YFAI)

50

FIGURE 3 When the hand approaches, lighting is activated around the seamlessly mounted
seat adjustment switch (© YFAI)
development, as due to platform stra-
tegies the E/E architecture is usually
initiated at an earlier stage, and the
corresponding suppliers are nomi-
nated earlier. Also normally, func-
tional demonstrators have to be pre-
sented prior to nomination. For inte-
riors, corresponding vehicle-specific
prototypes prior to the selection of
suppliers are not usual, as the expense
for representative off-tool parts is very
time- and cost-intensive. Also, the inte-
rior design has not been finalized at
this point, so that corresponding defini-
tive prototypes for design verification,
for example with regard to crash behav-
ior, material characteristics, vibration
behavior or the necessary assembly pro-
cedures and protection against misuse,
FIGURE 4 With the Active Skin concept, Yanfeng Automotive Interiors presents its vision of integrating
intuitive controls into vehicle interiors (© YFAI)
ATZ worldwide 06|2019
are not possible. A possible solution
consists in mechatronic simulation
models for smart surface components.
In contrast, from the E/E development
point of view, the critical path for the
development process is definitely ori-
ented to the tool production times for
large intelligent surfaces. This is espe-
cially true if additional design verifi-
cation using near-series prototypes is
required. Here too, verified simulation
models offer a possible method to elimi-
nate the corresponding prototypes and
perform design verification with series
production tools.
Many multi-dimensional simulation
models are necessary for the develop-
ment of intelligent surfaces. These
include structural mechanics, vibration
characteristics, electronics, software,
light and thermal behavior. Through
integration of the interior and E/E CAE
approaches, this is driven forward in the
partnership between YFAI and Kostal.
Both of the timelines for E/E and the
interior must be taken into account and
combined for the development of smart
surfaces. For this, close cooperation
between specialist departments is neces-
sary – not only by the supplier, but also
by the car manufacturer. YFAI and
Kostal support this by showing custom-
ers possible technical and scheduled
solutions in a single step and thus reduc-
ing coordination effort for the customer.
CONCLUSION
Through the cooperation of Kostal and
YFAI, the know-how in the fields of HMI
technology and interiors is synergetically
combined. This creates the possibility of
offering market-ready solutions for intelli-
gent surfaces at short notice while at the
same time reducing organizational com-
plexity for OEM customers. Instead of
electronic modules which are installed in
the interior of the vehicle and then cov-
ered, YFAI and Kostal develop innovative
products as a single source seamlessly.
The electronic platform and the interior
design are taken into account and inte-
grated accordingly right from the start of
the development process, so that tedious
interface management is eliminated for
the customer. As a result, intelligent sur-
faces for the interiors of future vehicle
generations can be better tailored to the
needs of future users from the very start.
REFERENCES
[1] Yanfeng Automotive Interiors: Yanfeng Automo-
tive Interiors and Kostal cooperate on innovative
smart interior surfaces. Press release, September
27, 2016. Online: https://www.yfai.com/en/yanfeng -
automotive-interiors-and-kostal-cooperate-innova-
tive-smart-interior-surfaces, access: January 11,
2019
[2] Kostal: Yanfeng Automotive Interiors and Kostal
present smart surfaces in 3D glass. Press release,
September 12, 2017. Online: https://www.kostal.
com/en-gb/news/aktuelles/2017, access: January
16, 2019
[3] Yanfeng Automotive Interiors: Yanfeng Automo-
tive Interiors brings surface sensitivity to the interi-
ors of future vehicles. Press release, September 14,
2017. Online: https://www.yfai.com/en/press-
releases, access: January 11, 2019
51
DE VELO PMENT A S SISTANCE SYSTEMS
© rFpro
AUTHORS
Martin Herrmann
is Business Development Manager
ADAS & Automated Driving at
IPG Automotive GmbH
in Karlsruhe (Germany).
Chris Hoyle
is Technical Director at rFpro
in Romsey (UK).
52
INTRODUCTION
Situations which exclusively arise
when many different environmental
conditions simultaneously occur and
parameters are at extreme levels that
are outside or at the limit of normal
operating conditions are generally
referred to as “corner cases.”
In particular, difficult weather
conditions such as low sun elevations
or heavy rain represent corner cases
for camera-based systems, FIGURE 1.
Unusual textures or material properties,
for instance dirt buildup on vehicle sur-
faces may also contribute to an incorrect
identification of the respective vehicles
by a camera. Tests of these systems are
especially necessary in complex traffic
situations as well as at the limits of
vehicle dynamics. Due to the fact that
these situations arise very rarely in the
real world, that they are not reproduc-
ible, and that there are myriad possible
parameter variations, virtual test driv-
ing lends itself as a potential solution.
REQUIREMENTS
In order to transfer tests of camera-based
systems into the simulation, it is neces-
sary to virtually model the real world as
realistically and completely as possible.
This applies not only to the generation of
image data but also to infrastructure
modeling, the modeling of other road
users as well as a highly accurate model-
ing of vehicle dynamics and the virtual
prototype. The virtual prototype trans-
lates the calculated control signals of the
function under test into a corresponding
response of the vehicle, which in turn
has a significant impact on the sensor
data, for example when the vehicle body
pitches during hard braking.
For the generation of realistic image
data, the rendering of the virtual 3-D
environment as well as the modeling
of the camera, which includes both
optical (lenses) and electronic (image
processor) components, are basic pre-
requisites. Sufficiently detailed 3-D
models, high-resolution textures and
physically precise material properties
are fundamental for rendering. In addi-
tion, detailed lighting effects with sev-
eral light sources and the representa-
tion of weather conditions are required.
Relevant effects that need to be
factored into the modeling of optical
effects of the camera are lens distor-
tion, vignetting and any potential
dirt buildup on the lens, for instance,
FIGURE 2. The image sensor properties
need to be taken into account in terms
of the image refresh rate and resolution
as well as the modeling of different color
filters, for instance the Bayer filter.
Exploring Corner Cases for
Camera-based Driving Functions –
High-fidelity Visualization in Simulation

When testing driver assistance systems and automated driving functions, one of
the challenges is to evaluate scenarios whose probability of occurrence is extremely
low. IPG Automotive and rFpro have developed a simulation environment in which
such scenarios can be reproduced.

These effects also must be modeled
dynamically as they may change during
a test drive. For example, during rain,
the optical properties of the glass in front
of a camera lens may vary continuously.
Each raindrop will cause distortion,
transforming as it flows or dissipates.
Between successive camera frames the
wipers may have cleared the glass caus-
ing a more dramatic change and occa-
sionally the camera’s exposure may coin-
cide with the passage of a wiper blade
and there may be full or partial occlu-
sion of the camera’s field of view. Other
examples would include the varying
density of fog along a route, the tempo-
rary occlusion of the sun by cloud cover,
roadside objects or passing vehicles.
EFFICIENCY ENHANCEMENT
One of the primary requirements is to
overcome the disadvantages of real test
driving and to configure the test environ-
ment as efficiently as possible. In com-
parison, virtual test driving offers a range
of advantages that may considerably accel-
erate and simplify the development of
advanced driver assistance systems. The
ability to fully control the environment
is the foremost advantage. Weather con-
ditions, the time of day and complex traf-
fic situations can be configured at will.
By separating the complete system
under test into its functional components,
a method called “functional decomposi-

FIGURE 1 Certain weather conditions or solar altitudes result in corner cases for the detection of lane markings (© rFpro)

ATZ worldwide 06|2019 53

DE VELO PMENT A S SISTANCE SYSTEMS

FIGURE 2 Typical output formats (YUV, left) and modeling of lens effects (vignetting, right) (© rFpro)

tion”, the required test effort can be sig-
nificantly reduced [1]. Initially, the relevant
components are validated independently
from each other as the individual systems
react to the same test parameter changes
with varying levels of sensitivity. Using
simulation, input data for the subsystems
can be generated which may be realistic
image data for perception testing or
object lists for maneuver planning, for
instance, FIGURE 3. When testing the
complete system, high-precision realistic
image data is vital.
Another major benefit affects safety.
Simulation also allows the possibility to
consistently reproduce tests that involve
a substantial risk of injury to the occu-
pants of the vehicle or to other road users.
Furthermore, tests that represent the con-
ditions in a variety of areas around the
world can be generated with little effort.
Thus, in comparison to real test driving,
the effort involved is significantly lower.
In addition, all tests are fully reproduc-
ible. Simulation opens up the possibility
of building a virtual prototype, a fully
parameterized vehicle model that corre-
sponds to its real-world counterpart in
all its components and subsystems. Early
availability during the development pro-
cess is a particular advantage.
When the systems under test make
use of Supervised Learning based AI sys-
tems, for example deep learning neural
networks, the simulation can be used to
generate training data sets, FIGURE 4.
Some supervised learning algorithms are
sensitive to the quality of synthetic train-
ing data. By making use of very detailed
digital twins, that means 3-D environ-
ment models that faithfully replicate the
real world, the simulation will be as sto-
chastic as the real world. This avoids the
repeating patterns and video artifacts
that afflict most computer-generated
synthetic models. The benefit of using
higher quality training data is that you
do not risk overfitting (that is over-train-
ing) the system, reducing your reliance
on training data acquired in the real
world, which is expensive to capture
and annotate. Training data produced
from simulation offers another benefit.
The annotation of the images is error-
free and automatic. Manual annotation
is time-consuming and suffers errors.
Simulation is expected to form part
of the regulatory framework for auto-
nomous vehicles. This means that the
simulation framework implemented to
accelerate the testing process can later
be repurposed to form part of an auto-
mated validation process. The require-

Automated driving function/ADAS function

Other Sensor data Decision Trajectory Control signal

Other sensor processed
data fusion making planning generation
sensor data
processing

Camera data Processed Environment Desired Planned

processing sensor data model maneuver trajectory

Other
sensor data
FIGURE 3 The CarMaker-
Camera data and-rFpro co-simulation
Processed Environment Control architecture enables the
sensor data model signals
generation and processing
Simulation of raw sensor signals in
environment addition to the input of
processed sensor data
CarMaker-and-rFpro co-simulation and the global environment
model (© IPG Automotive)

54

FIGURE 4 Generation of camera images and related training data with semantic segmentation (© rFpro)
ments described earlier that allow to
extend testing to corner cases have the
additional benefit of allowing the testing
to correlate with the real world, which
means that the testing can be used for
validation. It is expected that statistical
samples of these test libraries will be
reproduced at real-world test facilities
and their digital twins.
Physically modeling the real world
means abandoning simplified optical
effects that are adequate for computer
games and cinematography. It is, how-
ever, essential in ensuring correlation
between the real-world and virtual tests.
SCENARIO GENERATION
AND TEST AUTOMATION
In the following example, the combi-
nation of IPG Automotive’s open inte-
Test run
CarMaker
simulation
Road5
definition
CarMaker
FIGURE 5 Measured 3-D environment in rFpro as a basis and export of the CarMaker Road5 or generation of the CarMaker Road5 and conversion to rFpro
(© IPG Automotive)
ATZ worldwide 06|2019
gration and test platform CarMaker
with the simulation environment rFpro
is described.
Test scenarios can originate from dif-
ferent sources and contain the definition
of the static and dynamic environment.
It is possible to manually generate the
static environment in the scenario editor
in CarMaker as well as to import it from
map data. Subsequently, this static envi-
ronment can be transferred into the 3-D
format required by rFpro. This “single-
click” conversion process automates the
recreation of the 3-D environment model
ready for use in rFpro. Another option is
the measurement of real routes and the
generation of a highly precise 3-D model,
a digital-twin of the real world based on
the obtained measurement data. This
enables a targeted reproduction of corner
cases that occur at a specific position in
Scenario
editor
rFpro 3-D
location
Conversion
the real world. Examples are special
shadow effects, real objects blocked
from view, or a certain unevenness of
the road surface, etc. Models built in this
way are also particularly useful for the
creation of training data for supervised
learning systems.
Traffic objects can be modeled
either by simulating a realistic traffic
flow, through a co-simulation with PTV
Vissim [2], or by generating event and
maneuver-based behavior in CarMaker
that controls the traffic objects in rFpro,
FIGURE 6. This last option enables the
targeted variation of individual para-
meters, for instance the speed or dis-
tance. These parameter variations as
well as the variation of weather condi-
tions can be accomplished in automated
mode using the Test Manager integrated
in CarMaker, which enables the auto-
mated identification of corner cases by
testing various parameter variations.
IMPLEMENTATION OF THE
TEST PLATFORM IN THE
DEVELOPMENT PROCESS
Applying this development process
and identifying corner cases is possible
throughout the entire development pro-
cess. Already during the training of
AI-based algorithms, this approach
enables the automated logging of train-
ing data from the camera’s perspective
from the intended mounting position
in the vehicle including the correct reso-
lution, lens properties, etc. By comple-
menting the data set with these syntheti-
Image generation
Vehicle model
interface
rFpro
55
DE VELO PMENT A S SISTANCE SYSTEMS
cally generated training data, the results
of the training are improved. In addition,
a very high variance and the included
corner cases ensure that the system
functions reliably in as many situations
as possible. This approach allows for
training data to be generated even before
a physical camera is available. The com-
bination of this approach with paral-
lelized simulation and automated anno-
tation results in a significant reduction
of development time.
When testing the system, there are
two approaches depending on the level
of integration. Before a physical control
unit is available, the system can be inte-
grated as a model (model-in-the-loop) or
as software (software-in-the-loop). Here,
it is recommendable to use CarMaker as
an integration platform and to receive
the image input via the rFpro Simulink
toolbox or its C++ API. When following
a hardware-in-the-loop approach, rFpro
delivers the generated image data to
IPG Automotive’s Video Interface Box.
The camera’s image sensor is emulated
via the Video Interface Box and the image
signal is fed into the camera ECU.
The system’s output parameters such
as steering or braking requests are trans-
ferred to the component controllers and
subsequently to the vehicle dynamics
model.
The testing architecture described can
also be deployed in driving simulators
FIGURE 6 Freely configurable and reproducible generation of any desired rare traffic situation (© rFpro)
56
with a human driver in the loop. This
offers further benefits that accelerate the
testing process and also help to ensure
that the system ultimately achieves con-
sumer acceptance. The following use
cases describe the exploitation of a
driving simulator that is capable of faith-
fully reproducing the vehicle dynamics
of the vehicle under test. The use of
rFpro’s video pipeline reduces the
latency between the simulated experi-
ment and the human driver. This is criti-
cal to the success of driving simulation
as any lag between the vehicle’s simula-
tion and the user’s perception causes
motion sickness and restricts usefulness
in highly dynamic maneuvers such as
emergency situations.
The first use case is to ride as a pas-
senger or driver in the vehicle under
test. For SAE levels 4 and 5 the human
test subject is expected to be a passen-
ger. For ADAS system testing it is ex-
pected that the experiment will close
the control loop through the human
driver. This involves human test drivers
in both objective measurement and sub-
jective evaluation earlier in the develop-
ment process.
The second use case is to drive one
or more simulated conventional vehicles
in the same virtual world as a level 4 or 5
autonomous vehicle under test. Human
participants are slightly random, unpre-
dictable and make mistakes. There is no
better way known to provoke new failure
modes in automated driving and safety
systems. rFpro allows up to fifty driving
simulators to simultaneously connect to
the same experiment.
SUMMARY
Following the approach described
enables the development and testing of
critical test cases and corner cases in the
context of the whole vehicle before a real
prototype becomes available. An exem-
plary test environment was presented
which fulfils these requirements.
This test environment enables corner
cases to be generated as needed. When
adopting suitable search strategies, sys-
tem limits can be identified efficiently.
It thus becomes possible to bridge the
gap between simulation and reality for
camera-based advanced driver assis-
tance systems to a considerable extent
and to greatly accelerate the develop-
ment process.
REFERENCES
[1] Amersbach, C.; Winner, H.: Funktionale
Dekomposition – Ein Beitrag zur Überwindung
der Parameterraumexplosion bei der Validation
von höher automatisiertem Fahren. 12. Workshop
Fahrerassistenzsysteme und automatisiertes
Fahren, Walting, 2018
[2] Frings, A.: Der virtuelle Fahrversuch unterstützt
durch die mikroskopische Verkehrsflusssimulation.
27. PTV-Traffic-Anwenderseminar, Karlsruhe, 2017
 NEW: Top Magazine
for Lightweight Construction
www.lightweight-design.de

f r e e o f c harge:
a z i n e f o r 3 0 d ay s w e i g h t d esign
n o w t h e eM ag g e. c o m / l i g ht
Tes t - k n ow l e d
c i a l i z e d
-spe
w w w.my

The consistent application of innovative lightweight
construction concepts today plays a key role in
reducing weight and saving energy.
lightweight.design worldwide provides you with all
the practical information you need for implementing
lightweight construction principles in developing and
manufacturing new 06|2019
products across the entire value
ATZ worldwide
chain from the material to construction technology,
simulation and optimization procedures up to
production technology, quality assurance and
recycling.
Test now the eMagazine for 30 days free of charge!
www.my-specialized-knowledge.com/lightweightdesign
57
DE VELO PMENT FILTERS

Air Quality of Vehicle Interiors

in Highly Polluted Cities © Paragon

AUTHORS Particulate matter is a source of pollution in vehicle interiors

Dipl.-Ing. Ralf Mönkemöller
is Head of the Sensors Division
at Paragon GmbH & Co. KGaA
in Delbrück (Germany).
in many places. This is especially evident in China where the
particulate matter levels measured are extremely high despite
the very low share of diesel vehicles. Paragon has developed an
high voltage plasma particle filter that removes the particularly
hazardous ultrafine particles from indoor air of passenger cars.

EXPOSURE TO directly exposed to traffic emissions.

Sarah Ristau, B. Sc.
is Technical Project
Manager for Electric Filters
at Paragon GmbH & Co. KGaA
in Delbrück (Germany).
58
PARTICULATE MATTER
Despite the widespread introduction
of diesel particulate filters, particulate
matter continues to be one of the pri-
mary pollutants of air quality in many
parts of the world. In 2018, the annual
average particulate concentration with
particles of up to 2.5 µm diameter (PM2.5)
in Beijing was around 43.4 µg/m3 air [1],
which is more than four times the WHO’s
maximum average permissible limit of
10 µg/m3 [2]. In China, many people
therefore have devices for displaying
the Air Quality Index (AQI) [3]. Drivers
and passengers of motor vehicles are
especially affected because they are
According to measurements carried
out by Paragon, classic air filter tech-
nology is capable of reducing the pol-
lution in the vehicle interior from class
“unhealthy” (AQI over 150) to an accept-
able class (“moderate,” AQI less than 100
or “good,” AQI less than 50) at par ticle
sizes of PM2.5. However, this is not the
case for the Ultrafine Particles (UFPs).
CAUSES OF PARTICULATE MATTER
Public perception puts the blame for
particulate matter on traffic and espe-
cially cars and trucks that run on diesel
fuel. The causes of Particulate Matter
(PM2.5) are far more diverse, and they

FIGURE 1 Causes of particulate matter emissions in China (© Paragon)
are also very different in China (Beijing)
compared to Germany (Berlin).
In Berlin, around half (53 %) of the
particulate matter can be attributed to
traffic. There are various other emitters,
which can be categorized into house-
holds (wood combustion), commercial
operations and natural sources of partic-
ulate matter. A detailed breakdown of
FIGURE 2 Schematic representation of in-vehicle air quality management
(© Paragon)
ATZ worldwide 06|2019
the traffic-related causes of particulate
matter shows that combustion engine
exhaust emissions do not constitute even
half of this emissions group. More pre-
valent causes that also play a role with
electric vehicles include resuspension
due to traffic, tire wear and brake abra-
sion. When it comes to particulate mat-
ter, electric vehicles are far from emis-
sion free when on the road. In addition,
particulate matter must still be exam-
ined even in the age of electromobility.
The relatively largest share of parti-
culate emissions in Beijing (27 %),
FIGURE 1, is caused by the reaction of
already existing gaseous air pollutants,
for example through solar radiation or
interaction with each other (secondary
pollutants) [6]. Exhaust emissions from
traffic contribute directly to 17 % of the
particulate matter. Road and ground
dust resuspension is also an important
factor. Other causes include fossil fuel
combustion (not in motor vehicles), bio-
mass combustion and metal processing.
EFFECTS ON THE BODY
The health risks of particulate matter
are strongly linked to particle size. The
smaller the particles, the deeper they
can penetrate into the lungs. According
to current knowledge, larger particles
up to a diameter of 10 µm (PM10) can
lead to impaired lung function and trig-
ger asthma or allergies [7]. Particles with
diameters of up to 2.5 µm (PM2.5) are
capable of reaching the alveoli and are
suspected of stimulating tumor growth
and causing cardiovascular problems.
UFPs with diameters of 0 to 100 nm are
classified as carcinogenic because of
their ability to penetrate deep into the
lungs, travel through the cell membranes
and enter the bloodstream.
HOLISTIC AIR QUALIT Y
MANAGEMENT
The very different local and regional
quality of ambient air requires a holis-
tic approach to in-vehicle air quality
management, FIGURE 2. Particles with
a weight proportion of up to 2.5 µg/m3
(PM2.5) can be detected by means of
a particle sensor. If the climate control
system switches to recirculation mode
at high loads, this reduces the particle
load inside the vehicle and extends the
service life of traditional activated car-
bon filters. However, these sensors do
not measure the UFPs.
The Air Quality Sensor (AQS) is an
essential component of air quality man-
agement. It consists of two comb-like
interdigital structures that are coated
with metal oxide and heated to tempe-
ratures above 320 °C. The metal oxides
react to concentration changes of NOx,
59
DE VELO PMENT FILTERS
CO and, in three-gas sensors, of NH3 in
the intake air and provide information
to the climate control system at one-sec-
ond intervals. If the algorithm considers
this is necessary, the climate control unit
switches to recirculation mode. In real
operation, particles of all sizes often
occur together with NOx and CO. The
AQS therefore also indirectly reduces
the particle load inside the vehicle.
A CO2 sensor is another component that
can be used to increase air quality. While
such sensors are available for serial pro-
duction for automobile applications, they
were mainly developed for monitoring
air conditioning systems that use CO2
instead of R134a refrigerant. In metropoli-
tan areas with a high particle load, where
drivers are recirculating the air virtually
all the time, such a sensor can also be
used to measure the concentration of
CO2 inside the vehicle. Depending on seat
occupancy, the level of CO2 can increase
to 800 to 900 ppm after only 8 to 15 min
in recirculation mode, which causes
diminished concentration in the occu-
pants. Although calculation models are
already being used that rule out inadmis-
sible CO2 concentrations and, if necessary,
deactivate air recirculation, an additional
sensor could increase functional safety,
especially if it is already in the vehicle.
Other components of a holistic air qual-
ity management include ionizers and odor-
izers. The latter as a fragrance dispenser
60
is purely a matter of convenience. Ionizers
can improve the air inside the vehicle fol-
lowing the intake process by producing
negatively charged ions using a voltage of
6000 V by means of emitter needles and
releasing them into the air of the vehicle.
These ions attach themselves to the sur-
rounding dust particles. This creates larger
dust particles from the dangerous ultra-
fine dusts, which are harmless to health.
PARTICULATE MATTER
MEASUREMENTS IN BEIJING
Paragon carried out measurement runs
in Beijing to determine the particulate
matter load both outside and inside a
vehicle in China, including particles of
size classes PM2.5 and PM0.1. The values
for PM2.5 and PM0.1 were first measured
and recorded outside of the vehicle by
means of sensors. In addition, the resis-
tance signals of an AQS air quality sensor
for NOx and CO were recorded and con-
verted to ppm for the display. All other
data were calculated from the measure-
ment data by means of simulation. This
procedure is also used in order to evalu-
ate various boundary conditions in the
air quality management design such as
different vehicles, filters or switching
sensitivities of the air quality sensor.
A simple model was used for the vehi-
cle interior, with the air volume set at
3.0 m3 and the displacement volume of
FIGURE 3 Measurement run in Beijing:
particulate matter concentration in the
outside air and vehicle interior (© Paragon)
the air conditioner at 1.0 m3/min. In
addition, the simulation was calculated
with a CO separation rate of the combi-
nation filter of 0 %, an NOx separation rate
of 60 %, a PM2.5 separation rate of 80 %
and a UFP separation rate of 10 %.
FIGURE 3 shows the results of the simu-
lation with the parameters described for
fine particles with a PM2.5 size class.
This shows that the interior load also
decreases continuously under sustained
high load of the outside air with short-
term peak loads and permanently drops
below the critical value of 10 to 12 µg/m³
after around 7 min. As mentioned before,
the WHO guideline for fine particles
(PM2.5) is an annual maximum mean
of 10 µg/m³ air [2]. In China, the value
for an AQI of 50 (good) is a maximum
of 12 µg/m³. The mean value of the test
drive is 59.76 µg/m³ PM2.5 outside and
14.09 µg/m³ PM2.5 inside. This shows
that managing air quality with an AQS air
quality sensor and conventional combina-
tion filters for the fine particles (PM2.5)
already delivers significant improvements.
The results of the same measurement
run for fine particles of size class PM0.1
are shown in FIGURE 4. The external value
fluctuates significantly and reaches an
average value of 8.57 µg/m³ PM0.1. The
value inside the vehicle initially shows
only slight improvement, then increases
again over time and reaches a mean value
of 6.68 µg/m³ PM0.1. This demonstrates
FIGURE 4 Measurement run in Beijing:
ultrafine particle concentration in the
outside air and vehicle interior (© Paragon)
that a system with an AQS and conven-
tional combination filter has a minor
effect on UFPs.
FILTER EFFECT ON PARTICLE SIZE
The air filter is at the heart of air quality
management for vehicle interiors. Up to
now, one of the uses of combination fil-
ters with various separation mechanisms
has been to filter particles and harmful
ATZ worldwide 06|2019
gases from the outside air. However,
their effectiveness is greatly dependent
on particle size. FIGURE 5 shows the filter
effect of a new, conventional combina-
tion filter and that of a new, HV plasma
particle filter dependent on particle size
at an volume flow of 127 m3/h. This
clearly demonstrates the “filter gap” of
conventional air filters: Although they
are capable of filtering out all particles
from the outside air starting at a particle
size of larger than 10 µm, the filtration
effect especially for particularly hazard-
ous UFPs with particle matter less than
0.1 µm drops to less than half and in
some instances to even just one-tenth.
This clearly illustrates that a conven-
tional combination filter alone is not
enough, particularly in highly polluted
regions. Therefore, paragon has deve-
loped an HV plasma particle filter that
closes this filter gap.
FIGURE 5 Filter effect on particle size
(© Paragon)
61
DE VELO PMENT FILTERS
FIGURE 6 Structure and functional principle of the HV plasma particle filter (© Paragon)
HV PLASMA PARTICLE FILTER
The Dustprotect HV plasma particle filter
transfers a principle that is well estab-
lished in large industrial plants to auto-
motive applications. As FIGURE 6 shows,
the particle filter essentially consists of
three components:
– a prefilter for the separation of
airborne dust (for example leaves
and other larger particles that are
moved around)
– an Electrostatic Precipitator (ESP)
consisting of spray electrodes and
collecting plates
– an activated carbon filter to reduce
ozone and odors.
First, the outside air flows through the
prefilter, which filters out coarse dust
and heavy particles larger than 10 µm.
The smaller particles (PM10, PM 2.5,
PM1 and UFP) enter the ESP and pass
through the electron cloud between the
stainless steel electrodes charged to a
high voltage of 6000 to 10,000 V and
the lamellas of the collection plate on
the other side. The ions attach them-
62
selves to the particles, charge them
and move them in the direction of the
collection plate.
Since the ESP already filters out a
large portion of the particles, the subse-
quent activated carbon combination fil-
ter can have smaller dimensions. The
larger remaining or resulting particles
as well as odors and ozone are filtered
out here. However, some of the charged
ions also migrate to the activated car-
bon filter, which regenerates repeat-
edly due to the permanent electrostatic
charge. This means that it has to be
replaced much less frequently than
conventional combination filters
(up to 50,000 km servicing interval,
depending on the application).
The major advantages of the elec-
tric particle filter are the highly
effective filtering of UFPs, a stable
and high deposition rate over a lon-
ger period of time, the low installa-
tion space requirements, and lower
pressure loss because the activated
carbon combination filter can be
made with a thinner profile. This
allows a higher air volume to flow
through the filter than is the case
with conventional filters.
SUMMARY
Particulate matter continues to be
one of the greatest pollutants of air
quality in the metropolitan areas of
China as well as many other parts
of the world. The very small particle
classes with a size less than 0.1 µm
(UFPs) in particular are believed to
cause cancer or encourage its growth.
During measurement runs in Beijing,
paragon found that the hitherto widely
used air quality management for vehi-
cle interiors with an AQS and conven-
tional combination filter for the fine
particles (PM 2.5) results in quite good
improvements. However, it only has a
minor effect on UFPs. This filter gap
could also be substantiated in an exter-
nal laboratory. Paragon has therefore
developed an HV plasma particle filter
with a prefilter, electrostatic precipitator
and activated carbon filter. It is able to
remove UFPs from the outside air in
a very effective way and thereby protect
the health of vehicle occupants.
REFERENCES
[1] Berkeley Earth: Local Particulate Air
Pollution (PM 2.5), Beijing, China. Online:
http://berkeleyearth.lbl.gov/air-quality/local/
China/Beijing/Beijing, access: January 17, 2019
[2] Air quality guidelines – global update 2005
Online: https://www.who.int/phe/health_topics/
outdoorair/outdoorair_aqg/en/, access:
February 6, 2019
[3] International Air Quality. Online: https://
www.airnow.gov/index.cfm?action=airnow.
international, access: February 6, 2019
[4] Rauterberg-Wulff, A.; Lutz, M.; et al.: Luft-
reinhalteplan 2012 bis 2017 für Berlin. Senats-
verwaltung für Stadtentwicklung und Umwelt, 2013
[5] Bayerisches Staatsministerium für Umwelt und
Gesundheit: Daten, Fakten, Ziele Feinstaub. Dif-
fuser Staub, klares Handeln. 2010
[6] Yu, Lingda; et al.: Characterization and source
Apportionment of PM 2.5 in an Urban Environment
in Beijing. In: Int. Journal of Aerosol and Air Quality
Research (2013), No. 13, pp. 574-583
[7] Schadwinkel, A.: Stickoxide und Feinstaub –
Was die Wissenschaft über Abgase weiß – und
was nicht. In: Die Zeit, January 24, 2019. Online:
https://www.zeit.de/wissen/gesundheit/2019-01/
stickoxide-gesundheit-feinstaub-luftverschmutzung,
access: February 6, 2019
THE LOOK INTO THE FUTURE.

MTZ has successfully accompanied the development of engines and powertrains for the past
80 years. The internationally leading technical and scientific magazine for decision-makers in
powertrain development and production sees itself as an information platform on future power-
trains – conventional, electrified or purely electric. MTZ promotes the transfer of information
between manufacturers, suppliers, service providers and research and development centers all
over the world, in German and English, in print and digital editions.

GET YOUR FREE TRIAL NOW

© istockphoto.com/matdesoígn24

www.mtz-magazine.com
ATZ worldwide 06|2019 63
ON OUR OWN

On the Way to More Quiet and

Comfortable Vehicles
Acoustical engineering and NVH
are amongst the prime indica-
tors of vehicle quality and refine-
ment. These aspects enjoy an
increasing focus in automotive
development, at a time of
changed powertrain configura-
tions. Vehicle architectures,
powertrain systems, and design
paradigms are globally chal-
lenged by emissions legislation.

© Alexander Stertzik | ATZlive

Acoustics and NVH have always been an important factor of
vehicle quality. The increased variety and complexity of today’s
powertrains is further increasing the effort required for acous-
tics development. Noises that the internal combustion engine
has masked are now clearly emerging when an electric drive is
used. As a result, acoustic experts are faced with the challenge
of effectively reducing the noises also from e-motors, servomo-
tors and auxiliary units that are no longer superimposed.
Lightweight design also increases the complexity of striving for
a comfortable NVH level. Better measurement analyses and
faster calculation methods are important tools for this and
central topics in the experts’ discussion. The biennial Automo-
tive Acoustics Conference with Autoneum as cooperation part-
ner has established itself as the leading global conference in the
field of automotive acoustics, providing an up-to-date overview
of the latest trends and most up-to-date expertise. At the 2017
conference, more than 250 participants representing 119 com-
panies gathered at the Gottlieb Duttweiler Institute (GDI) in
Zurich/Ruschlikon (Switzerland). The 2019 edition will again
take place at this renowned Swiss think tank under the patron-
age of Prof. Dr. Paolo Ermanni of ETH Zurich. The conference
program includes presentations by OEMs, suppliers, engineering
service providers, and university research institutes. The event
is complemented by parallel workshops as well as an extensive
exhibition of various suppliers presenting innovations in the
field of acoustic management.

Automotive Acoustics Conference – 5th International ATZ Conference on Vehicle Acoustics

Date: July 9 and 10, 2019

Location: Gottlieb Duttweiler Institute (GDI), Zurich/Ruschlikon (Switzerland)
Scientific Director: Dr. Davide Caprioli, Autoneum Management AG
Cooperation Partner: Autoneum
Conference Fee: € 1,345.00 (without VAT)
Registration: www.atzlive.de/en/events/automotive-acoustics-conference/

64
 ON OUR OWN

Heavy-Duty, On- and

Off-Highway Engines
Challenge: CO2 limits

14th International MTZ Conference

on Heavy-Duty Engines

26 and 27 November 2019

Friedrichshafen | Germany

NEW DIESEL, GAS, AND DUAL-FUEL ENGINES

Heavy-duty engines, new developments

REDUCING CO 2 EMISSIONS
Powertrain concepts and engine approaches

POLLUTANT EMISSIONS
Legal requirements and solutions

/// SCIENTIFIC DIRECTOR

Prof. Dr. Peter Eilts, TU Braunschweig

/// FACTORY TOUR

Rolls-Royce's MTU plant

/// PARTNER OF THE CONFERENCE

Phone +49 611 7878-131 PROGRAM AND REGISTRATION

Abraham-Lincoln-Straße 46
65189 Wiesbaden | Germany
Fax +49 611 7878-452
ATZlive@springer.com www.ATZlive.com
ATZ worldwide 06|2019 65
 RESE ARCH C OMFORT | N V H

AUTHORS
NVH in Electric Mobility –
Vibration Analysis Using a Derotator

Wheel hub motors favor package, but are demanding in terms of NVH due to
Dr.-Ing.
Fabian Duvigneau their installation position. A working group of the Otto von Guericke University
is Postdoctoral
Researcher at the Institute Magdeburg uses an optical vibration measurement system to analyze the
of Mechanics of the Otto
von Guericke University
vibrations behavior of external rotor electric motors in order to optimize the
Magdeburg (Germany). acoustics of wheel hub motors.

Dr.-Ing. Christian Daniel

is Postdoctoral
Researcher at the Institute
of Mechanics of the Otto
von Guericke University
Magdeburg (Germany).

Dipl.-Ing.
Sebastian Koch
is Researcher at the
Institute of Mechanics
of the Otto von Guericke
University Magdeburg
(Germany).

Jun.-Prof. Dr.-Ing.
Elmar Woschke
is Junior Professor at the
Institute of Mechanics
of the Otto von Guericke
University Magdeburg
(Germany).

© OVGU

66
1 RESE ARCH ON CLOSE-TO -WHEEL ELECTRIC DRIVES
2 INVESTIG ATION OF THE STRUCTURE WITH A SCANNING VIBROME TER
3 ME ASURING PRINCIPLE AND E XPERIMENTAL SE TUP
4 RESULTS
5 SUMMARY AND OUTLO OK
1 RESEARCH ON CLOSE-TO-WHEEL ELECTRIC DRIVES
At the Otto von Guericke University Magdeburg, the field of elec-
tric mobility has increasingly been in the focus of research since
2011. The working group “Editha” developed several electric drive
concepts using a Smart fortwo MC450 as a base. The aim was to
develop a street-legal prototype, which uses close-to-wheel driving
concepts to create space for the traction batteries.
FIGURE 1 Wheel hub motor implemented in the research vehicle Editha 3
(© OVGU)
ATZ worldwide 06|2019
In 2012, the concept vehicle Editha 1 was completed and got an
official approval and homologation for road service. The heart of the
driving concept are two DC motors, which are mounted rigidly by a
single-stage planetary gear directly on the rear axle. As a further
development of this concept, Editha 2 has been initiated to replace
the DC motors by permanent-magnet synchronous motors and to
improve the drivetrain. With Editha 3 the step has been taken
towards wheel hub motors, FIGURE 1. This wheel hub motor was
developed at the Institute of Mobile Systems. The aim is to save the
entire drive train and the planetary gear in the interior of the vehicle
and thus to obtain even more space for batteries. Another advan-
tage is that each wheel can be controlled separately, realizing a
more intelligent vehicle dynamics control. The developed wheel hub
motor is an external rotor motor. By combining an air gap winding
with an additional slot winding, a high power density is achieved.
Due to their decentralized driving concept, all three series of
Editha vehicles have an increased unsprung mass. The effects of
this additional masses in the wheels are among others the subject
of experimental and numerical investigations at the Institute of
Mechanics [1]. In addition to changed vertical dynamics, the
sound radiation of the wheel hub motor is an important issue, as
the motor is placed in the rim, which can lead to increased noise
emission due to the large flat components.
Although electric motors are generally considered to be less
noisy compared to internal combustion engines, it is necessary to
evaluate the acoustic behavior of electric vehicles. On the one
hand, electric machines emit tonal high-frequency noises that are
perceived as particularly annoying by human hearing even at low
volumes. On the other hand, other sources of noise are no longer
masked. In the case of a wheel hub motor the acoustic shielding
of the car body is missing and the application of passive measures
is not easily possible. The aim of the experimental analyses is the
generation of a validation basis for complex simulation models,
which allow a holistic view of the entire chain of effects, that
means take into account all relevant sources of excitation as well
as transmission phenomena. Using the validated numerical mod-
els, computer-aided optimizations will then be carried out in order
to improve the acoustic behavior of the wheel hub prototypes.
2 INVESTIGATION OF THE STRUCTURE
WITH A SCANNING VIBROMETER
After generating suitable simulation models for the complex overall
system, these were validated using conventional experimental vibra-
tion analyses in the stationary system. This was done at component
as well as assembly level. For vibration analysis, a one-dimensional
laser scanning Doppler vibrometer from Polytec was used, FIGURE 2
(left). For the measurements, the test object was attached to a
frame made of aluminum profiles via synthetic polymer threads,
since this so-called free-free mounting is particularly suitable for
being able to compare numerical and experimental vibration anal-
yses in a simple manner. FIGURE 2 (left) shows the entire assembled
rotor as an exemplary test object. Other boundary conditions, such
as clamping or jointed mounting, cannot be easily realized in a finite
element model identical to the experiment [2]. The uncertainties
due to the influence of the boundary conditions can require
time-consuming model updating processes in order to achieve a
good match between the simulation model and the experiment.
In order to avoid an undefined boundary condition by the cou-
pling of the excitation to the structure, the vibration excitation was
67
RESE ARCH C OMFORT | N V H

FIGURE 2 Experimental setup used for validation measurement (left) and results of experimental and numerical vibration analysis (right) (© OVGU)

realized with the aid of an impact hammer. Thus, free-free bound-
ary conditions are maintained. The excitation must be reproduc-
ible since it must be repeated for each individual scan point of the
measurement grid and the corresponding number of averages. For
this reason, the head of the impact hammer was mounted on an
electrodynamic shaker. This combination is highlighted in FIGURE 2
(left) by the blue ellipse.
FIGURE 2 (right) shows the comparison of the experimentally
measured (first row) with the numerically calculated eigenmodes
(middle row) for five conspicuous natural frequencies. The com-
parison of the eigenmodes is executed for the measured surface
of the laser vibrometer, that means the outer side surface of the
rotor. It can be seen that the simulation model is able to predict
the resulting vibration behavior of the complex overall system very
well, although the assembly has several joints. In this case, both
the natural frequencies and the natural vibration modes have a
very good match.
In the third row in FIGURE 2 (right), the eigenmodes of the entire
rotor are additionally illustrated in order to show that the numeri-
cal analysis offers the possibility to get a better impression of the
vibration behavior of the whole system compared to the laser
vibrometer measurement. The first and third columns demonstrate
that the investigation of the side surface in these cases is not rep-
resentative for the eigenmode of the overall system and the critical
vibration regions are not detected. This shows the value of a
numerical vibration analysis that provides information about all

FIGURE 3 Experimental setup for the vibration analysis of the running wheel hub motor (© OVGU)

68

areas of the examined structure. Nevertheless, validation mea-
surements are always recommended to prove the predictive capa-
bility of the simulation models.
3 MEASURING PRINCIPLE AND EXPERIMENTAL SETUP
For an assessment of the vibroacoustic behavior of an engine, the
behavior of the overall system during operation is particularly
important. Therefore, experimental vibration analyses were car-
ried out for different stationary operating points of the electric
wheel hub motor in order to obtain a broad validation basis for
the holistic simulation method developed in [3, 4]. A one-dimen-
sional laser vibrometer is used to measure the operating vibrations
of the wheel hub motor. The excitation of the motor structure is
achieved by the electrical commutation, whereby the motor must
rotate for the operating vibration measurement. Since the wheel
hub motor is designed as an external rotor motor, the rotor forms
the motor housing, which rotates together with the rim. The mea-
surement of local out of plane vibrations on the rotor is not pos-
sible with a conventional vibrometer.
Therefore, a rotating glass prism is attached in front of the
vibrometer. As a result, the rotating surface can be measured in
accordance with the defined measuring point grid by means of a
laser scanning vibrometer as if it would not rotate. This so-called
Polytec derotator is used for the scanning beam of the scanning
vibrometer and another fixed-beam vibrometer, which acts as a
reference channel. The reference is required to obtain a correct
phase reference for each point that is measured by the scanning
vibrometer. The experimental setup is shown in FIGURE 3. The elec-
tric brake, see left, was used to apply different loads and thus to
realize different stationary operating points. At the prototype stage,
the power electrics of the wheel hub motor was installed outside
the stator during the test bench measurements. For later use, the
power electronics will be integrated into the wheel hub motor.
The angular velocity of the glass prism of the derotator must be
synchronized with the angular velocity of the rotating test object.
ATZ worldwide 06|2019
FIGURE 4 Vibration amplitude
spectra of different operating
points averaged over the mea-
surement surface (© OVGU)
This was done in the present case with an incremental encoder
with 1024 increments. The mechanical connection of the incre-
mental encoder to the measurement object must be torsionally
rigid, otherwise the angular position of the test object cannot be
detected correctly and the prism cannot be driven with the correct
speed. The consequence would be a temporally variable rotation
of the measurement object in relation to the defined measurement
grid of the scanning vibrometer, whereby the location reference
would no longer be correct. The derotator, laser scanning vibrom-
eter and reference laser system is mounted on an adjustable base
and must be precisely aligned with the axis of rotation of the tar-
get to obtain reliable results.
4 RESULTS
In this section, some results of the optical vibration analysis of the
operating system are presented exemplarily. FIGURE 4 shows the
frequency response functions of the vibration amplitudes averaged
over all points of the measurement grid for different load and
speed variations. As expected, both higher loads and higher
speeds lead to a more acoustically noticeable behavior. The typi-
cal tonal frequency content for electric machines can be clearly
seen. In addition, it can be seen that the frequency range of 3.7
to 4 kHz has particularly high amplitudes.
FIGURE 5 shows an exemplary result for a stationary idling oper-
ating point. Here again the averaged frequency spectrum is shown.
In addition, the most conspicuous vibration modes are illustrated.
It is noticeable that both symmetrical and asymmetric vibration
modes occur. Due to the symmetrical rotor design, only symmet-
rical operating vibration modes were expected, as observed in the
measurements in the stationary system, FIGURE 2. The asymmetric
modes are a clear indication of a non-symmetrical electrical exci-
tation or a non-symmetric boundary condition.
For comparison, FIGURE 6 shows the result for an operating
point with comparable speed and a defined torque. In contrast to
FIGURE 5, it shows no symmetrical vibration modes. As the elec-
69
RESE ARCH C OMFORT | N V H

FIGURE 5 Spectrum
of the vibration ampli-
tudes averaged over
the measurement
grid with conspicuous
vibration modes during
idling (© OVGU)

trical excitation forces are significantly higher in the case of the
operating point shown in FIGURE 6, it can be concluded that the
asymmetrical vibration modes of the operating engine are proba-
bly caused by a spatially non-uniform electrical excitation.
Due to the optical measurement of the wheel hub motor during
operation, possible explanations for an acoustically conspicuous
behavior could be found in the present case. Further steps are still
pending. The experimental operating vibration analysis is repeated
as soon as the problems are solved. The results obtained should
then serve as the first validation basis for the holistic simulation
methodology.
5 SUMMARY AND OUTLOOK
In this article of the Otto von Guericke University Magdeburg, Ger-
many, an optical vibration analysis of an innovative wheel hub
motor, as a special case of an electric drive unit, during operation
was presented. It has been shown that optical vibration measure-
ment can be used beneficial for both vibroacoustic characteriza-
tion and problem cause analysis. Currently, different prototypes
are manufactured that differ significantly from the engine mea-
sured in this paper. The differences range from the basic design
of the magnetic circuit (Hallbach array), through the design of the

FIGURE 6 Spectrum
of the vibration ampli-
tudes averaged over
the measurement
grid with conspicuous
vibration modes at a
stationary operating
point (© OVGU)

70
outer geometry, to the material selection of the main engine com-
ponents (aluminum, aluminum foams, fiber reinforced compos-
ites). The subsequent experimental analysis of the new prototypes
will provide a broad validation basis for holistic simulation meth-
odology. The aim is to use the qualified methodology to develop a
comprehensive understanding of the system and to carry out
numerical optimizations in order to resolve the trade-offs between
performance, lightweight design and acoustics in the future.

REFERENCES
[1] Daniel, C.; Nitzschke, S.; Woschke, E.; Strackeljan, J.: Konstruktion,
Berechnung und experimentelle Belastungsmessung des Antriebsstranges
von “Editha”. Proceedings 11. Magdeburger Maschinenbau-Tage, 2013
[2] Duvigneau, F.; Koch, S.; Orszulik, R.; Woschke, E.; Gabbert, U.:
About the Vibration Modes of Square Plate-like Structures. In:
Technische Mechanik 36 (2016), No. 3, pp. 180-189, https://doi.org/
10.24352/UB.OVGU-2017-004
[3] Duvigneau, F.; Liefold, S.; Höchstetter, M.; Verhey, J. L.; Gabbert, U.:
Analysis of simulated engine sounds using a psychoacoustic model.
In: Journal of Sound and Vibration 366 (2016), pp. 544-555,
https://doi.org/10.1016/j.jsv.2015.11.034

STAINLESS
[4] Duvigneau, F.; Nitzschke, S.; Woschke, E.; Gabbert, U.: A holistic approach
for the vibration and acoustic analysis of combustion engines including hydro-
dynamic interactions. In: Archive of Applied Mechanics 86 (2016), No. 11,
pp. 1887-1900, https://doi.org/10.1007/s00419-016-1153-5
IS STANDARD
Our rings provide the same fit and function as
stamped rings, but are easier to assemble and
remove with no special tools. Standard parts
available in stainless (302 & 316) and carbon steel.
THANKS
The presented work is part of the joint project “Kompetenzzentrum eMobility Standard or custom, we’ll provide you with the right
(KeM),” which is financially supported by the European Union through the ring, in the right material, for your application.
European Funds for Regional Development (EFRE) as well as the German State
of Saxony-Anhalt. This support is gratefully acknowledged.

Electrical Coupler Gear Assembly

THE ENGINEER’S CHOICE

FREE SAMPLES:
Call +1 (847) 719-5900 or visit
ATZ worldwide 06|2019 71 expert.smalley.com/ATZ/Rings
 RESE ARCH EL ECTRIC V EHICL ES

AUTHORS
Battery Pack Housing for Electric Vehicles
Made by Laser Beam Welding

As part of current research activities at the Chair for Production Engineering

Prof. Dr.-Ing.
Achim Kampker of E-Mobility Components (PEM) at RWTH Aachen University, a production-
is Head of the Chair for
Production Engineering of appropriate low-cost housing for battery packs is being further developed and
E-Mobility Components
(PEM) at RWTH Aachen
examined on the basis of welding tests and prototypes. The approach has the
University (Germany). potential to reduce high costs faced by established car manufacturers, minimize
the costs of prototype vehicles, as well as to allow new start-ups to enter the
electric mobility market.

Dr.-Ing.
Georg Bergweiler
is Chief Engineer at the
Chair of Production
Engineering of E-Mobility
Components (PEM) of
RWTH Aachen University
(Germany).

Falko Fiedler, M. Sc.

is group leader body shop
at the Chair of Production
Engineering of E-Mobility
Components (PEM) of
RWTH Aachen University
(Germany).

Ansgar Hollah, M. Sc.

is Research Assistant at
the Chair of Production
Engineering of E-Mobility
Components (PEM) of
RWTH Aachen University
(Germany).

© RWTH Aachen University

72
1 ENERGY STOR AGE IN AN ELECTRIC CAR
2 REQUIREMENTS FOR BAT TERY SYSTEMS
3 C ONCEP T OF A LOW- C OST BAT TERY PACK HOUSING
4 INVESTIG ATION OF ALTERNATIVE PRODUCTION C ONCEP TS
5 SUMMARY AND OUTLO OK
1 ENERGY STORAGE IN AN ELECTRIC CAR
Due to the elimination of the internal combustion engine and all
dependent components, the advance in knowledge of the estab-
lished car manufacturers and thus the entry barriers for new com-
petitors have been reduced significantly. This is reinforced by the
fact that many components for the electrification of the powertrain
are developed by suppliers and freely available on the market.
These conditions have led to new start-ups such as StreetScooter,
e.GO Mobile, or Sono Motors being able to develop and offer elec-
tric vehicles suitable for series production on the market with sig-
nificantly lower capital investment [1]. The central element of an
electric car is the battery system, which consists of many battery
cells in the smallest unit. The battery modules are the next larger
unit of a battery system in which multiple battery cells are con-
nected. Several battery modules are then combined in a battery
pack housing to form a battery pack [2]. The battery system has
a particularly strong influence on the development and production
cost in the electric car. According to rough estimations, the bat-
tery system’s share of total vehicle costs can be estimated between
35 and 50 % [3], while the investment costs of a battery pack
assembly line can amount to approximately 57 million euros [4].
I. Battery module II. Battery pack housing
Tightness/protection against accidental contact 1,3
– Mechanical effects during driving1
– Mechanical impact in the event of a crash1
Thermal management 1,3
– Fire resistance
– Penetration into the passenger compartment 2
– Assembly in the vehicle
1 = [5]; 2 = [6, 7]; 3 = [10]
TABLE 1 Requirements for battery systems of an electric car (© RWTH Aachen University)
ATZ worldwide 06|2019
2 REQUIREMENTS FOR BATTERY SYSTEMS
The battery cell of an electric car is a sensitive component in the
overall body structure, which is considered a potential source of
danger for users and the environment due to the energy and chem-
icals it contains. In order to ensure safe and reliable operation, the
battery system is subject to legal requirements. The UNECE R100
Regulation is crucial and defines various test procedures, which
need to be conducted for the homologation of an electric car [5].
On the one hand, the battery pack must be sealed against the
ingress of moisture and foreign objects and, on the other hand,
the electrolyte must be secured from uncontrolled escape in the
event of damage. This also ensures that no electricity-leading parts
can be touched, which is especially important in the high-voltage
range above 60 V. Mechanical influences or extensively high cell
temperatures can trigger a short circuit in a battery cell and sud-
denly release energy. Even a heat source outside the battery pack,
such as a vehicle fire, must not lead to overheating. It is essential
to avoid that djacent battery cells overheat and a so-called ther-
mal runaway (thermal chain reaction) spreads through the entire
battery pack. In addition, the battery pack must not penetrate the
passenger compartment in the event of a crash and thereby endan-
ger passengers [6, 7]. The battery pack is a crucial component
that must be considered in the body structure of an electric car
due to its high mass and the required large installation space.
It has been shown that placing the battery pack in the vehicle floor
is advantageous for the use of space and, due to the low center of
gravity, for the driving dynamics [8]. Even with the two most com-
mon crash cases, the frontal and rear impact, the battery pack is
relatively well protected. In order to create the necessary installa-
tion space in the body, design changes are necessary compared
to a conventional car with an internal combustion engine [9]. The
usual requirements for the body structure must not be neglected,
such as the protection of other road users and the internal pas-
senger compartment. A production requirement is the final assem-
bly into the car, where the battery system usually has to be
installed in the body as an assembly part and must therefore be
pre-assembled within an assembly group. The requirements for
the battery system mentioned before are summarized in TABLE 1
and assigned to the applicable components.
A comparison of battery packs from several car manufacturers
showed that some requirements are redundant for different com-
ponents. Redundancies can be explained by the fact that the man-
ufacturer is free to carry out tests of the individual battery pack
III. Car body
–
–
1 –
–
73
RESE ARCH EL ECTRIC V EHICL ES

I. Battery module II. Battery pack housing III. Car body

Tightness/protection against accidental contact1,3 – –

– Mechanical influence during driving1 –

– – Mechanical impact by a crash1

Thermal management1,3 – –

– – Fire resistance1

– – Penetration into the passenger compartment 2

– Assembly into the car –

1 = [5]; 2 = [6, 7]; 3 = [10]

TABLE 2 Clearly assigned requirements for the battery system in the concept of the low-cost battery pack housing (© RWTH Aachen University)

or the entire vehicle. Tightness and protection against accidental
contact may be necessary for battery modules due to pre-assem-
bly work and are redundant in the battery pack housing due to
influences in driving operation. The mechanical influences during
driving are reduced both in the car body and by the battery pack
housing to minimize the load on the battery modules. In order to
withstand mechanical impacts in the event of a crash, the entire
battery pack is designed to be crash-proven, whereas the car body
already reduces an effect on the battery pack due to its crush-
collapsible zone. The function of thermal management is to dissi-
pate the resulting heat loss from the battery cells into the envi-
ronment and therefore affects the battery module and the battery
pack housing. Protection against penetration into the passenger
compartment is in turn, however provided independently by the
battery pack housing itself and the car body.
3 CONCEPT OF A LOW-COST BATTERY PACK HOUSING
There is extensive potential, especially for new entrants in the
electric mobility market and for manufactures of prototypes from
established OEMs, to reduce the development and production cost
of battery systems. The approach of clearly assigning the require-
ments for a battery system among the individual components and
thus avoiding unnecessary redundancy is investigated by the con-
cept of a low-cost battery pack housing.
The redundant requirements listed in TABLE 1 are attempted to
be resolved by using a supplier’s closed 48-V battery module. In
addition to the clear assignment of requirements for tightness and
protection against accidental contact, thermal management is also
guaranteed due to passive cooling in the battery module. More-
over, the purchase of the entire battery module allows for a signif-
icant reduction in development cost and time. The requirements
resulting from mechanical effects during the crash and penetration
into the passenger compartment are completely taken over by the
car body through appropriately designed body structures. Fire
resistance is ensured by a floor pan as part of the car body after
the final assembly of the battery pack. By clearly assigning require-
ments, the concept of the low-cost battery pack housing no longer
has redundancies to the respective components of the battery sys-
tem, TABLE 2.
The focus of the study is on optimizing the battery pack hous-
ing, since the battery modules are purchased parts and the car
body is manufacturer-specific. According to the breakdown in
TABLE 2, the only requirements placed on the battery pack housing
are those resulting from mechanical influences during driving and
general assembly in the vehicle. To meet these requirements, a
battery pack housing was manufactured from sheet metal compo-
nents, FIGURE 1. The battery pack housing mainly consists of two
floor pans connected by cross-beams. This creates eight com-
partments in which the battery modules can be mounted and
screwed into the battery pack housing via upper longitudinal car-
riers. For prototyping, these sheet metal components are manu-
factured using laser beam cutting, mechanical bending and man-
ually mounting with a total of 308 blind rivets. Experiments of

FIGURE 1 Low-cost
battery pack housing
as a rivet design
(© RWTH Aachen
University)

74

building five battery pack housings unveiled that a worker takes
an average of 38 min to manually rivet a battery pack housing.
4 INVESTIGATION OF ALTERNATIVE
PRODUCTION CONCEPTS
Due to the labor costs, the rivet design of the prototype is not suit-
able for scaling up to higher volumes in series production. As alter-
native production concepts, there are several competing joining pro-
cesses in semi-automated and fully automated production facilities
of the body shop under investigation, such as automated riveting pro-
cesses, Resistance Spot Welding (RSW), and laser beam welding.
The product architecture of the low-cost battery pack housing
as a laser-beamed bending component is similar to classic sheet
metal components from the body shop in many respects. Compo-
nent-integrated jig features is an alternative production concept
in the body shop for sheet metal [11]. Two or more components
are connected in such a way that at least one of the jig features
clamping, orienting, or fixing are fulfilled during production. This
can reduce jigs and thus initial, component specific investments
compared to conventional production concepts. The design with
blind rivets in pre-cut holes is a very simple example of this.
A flexible process chain for component-integrated jig features is
shown in FIGURE 2 [12].
For use in the low-cost battery pack housing, a component-
integrated jig feature is particularly suitable [11], with which
the functions of orientation and fixation for laser beam welding
FIGURE 3 Component-integrated jig feature by forming a T-shaped connecting element by breaking off a twisting table; according to Bergweiler [13]
(© RWTH Aachen University)
ATZ worldwide 06|2019
FIGURE 2 Flexible
process chain
for component-
integrated 
jig features
(© RWTH Aachen
University)
can be transferred into the component. The function can be seen
in FIGURE 3 and is generated by the formation of a T-shaped con-
necting element by manually braking off a twisting lug.
By using component-integrated jig features in combination with
laser beam welding, the design of the battery pack housing can
be greatly simplified. FIGURE 4 shows that the two floor pans were
only slightly modified compared to the riveted construction. The
cross beams are now inserted from above into the floor pans and
contain the connectors with twisted tabs at their bottom. By add-
ing more flaps at the top edge of the cross beams, the outer lon-
gitudinal beams can be left out. The inner longitudinal carriers
remain unchanged as they are only screwed together after the
battery modules have been inserted.
As part of the design and production planning of the laser-beam
welded battery pack housing, the mass was reduced by about
34 % from 12.4 to 8.2 kg compared to the riveted version and
resulted in a 42-% reduction from 26 to 15 components. Assem-
bly and welding of the components is possible in a single set-up
with one-sided accessibility. Connectors with component inte-
grated jig features achieve the orientation of the components rel-
ative to each other and the clamping function at the welding spots.
The manufacturing process was examined using a laser beam
welding cell with a double turntable, in which a battery pack hous-
ing can be welded inside while simultaneously the next battery
pack housing can be assembled outside the cell. First experiments
showed that the laser beam welding time of 30 connectors takes
about 3 min and the assembly can be fulfill in the same work cycle.
75
RESE ARCH EL ECTRIC V EHICL ES
The remaining 20 s are needed for the table rotation. The produc-
tion time of the laser-welded low-cost battery pack housing is
therefore 3 min and 20 s and has been reduced by 91 % com-
pared to the riveted version with 38 min.
5 SUMMARY AND OUTLOOK
A clear assignment of the battery system requirements to the
respective components in the entire car enables the resolution of
redundancies and enhances the production efficiency. As part of
this work, the concept of the low-cost battery pack housing was
further developed, the prototypical production process was ana-
lyzed and alternative production concepts were designed to be
scaled to series production.
In order to compare the alternative production concepts, the
plant layout and a conceptually automated robotic system with
RSW welding devices should be considered in addition to the pro-
duction time. Design rules as well as the process control of laser
beam welding need to be developed for component-integrated jig
features.
THANKS
This project is funded by the European Union and the Federal State of
North Rhine-Westphalia (Germany).
76
FIGURE 4 Low-cost battery
pack housing as a laser beam
welded design with component-
integrated jig features
(© RWTH Aachen University)
REFERENCES
[1] Ritz, J.: Elektromobilität: Was bremst den Erfolg elektrischer Fahrzeuge?
In: Ritz, J. (ed.): Mobilitätswende – autonome Autos erobern unsere Straßen:
Ressourcenverbrauch, Ökonomie und Sicherheit. Wiesbaden: Springer, 2018,
pp. 9-20
[2] Heimes, H.; Kampker, A.; Wessel, S.; Kehrer, M.; Michaelis, S.; Rahimzei, E.:
Montageprozess eines Batteriemoduls und -packs. 3 th revised edition. Aachen:
PEM der RWTH Aachen University, 2018
[3] Kampker, A.; Vallée, D.; Schnettler, A.: Elektromobilität. Berlin,
Heidelberg: Springer, 2013
[4] Kampker, A.: Elektromobilproduktion. Berlin: Springer Vieweg, 2014
[5] UNECE Reg. 100 – E/ECE/324/Rev.2/Add.99/Rev.2. Uniform provisions
concerning the approval of vehicles with regard to specific requirements
for the electric power train
[6] UNECE Reg. 94 – E/ECE/324/Rev.2/Add.93/Rev.3. Uniform provisions
concerning the approval of vehicles with regard to the protection of the
occupants in the event of a frontal collision
[7] UNECE Reg. 95 – E/ECE/324/Rev.2/Add.94/Rev.2. Uniform provisions
concerning the approval of vehicles with regard to the protection of the
occupants in the event of a lateral collision
[8] Arora, S.; Shen, W.; Kapoor, A.: Review of mechanical design and
strategic placement technique of a robust battery pack for electric vehicles.
In: Renewable and Sustainable Energy Reviews 60 (2016), No. 7, pp. 1319-1331
[9] Zentner, C.; Eckhardt, P.: Ampera-e-Opels – Weg der Fahrzeug-
elektrifizierung: Karosseriebautage Hamburg 2017, 15 th ATZ Conference,
Wiesbaden, 2017
[10] Frieß, B.; Soczka-Guth, T.; Hofbeck, F.; Nietfeld, F.: Challenges in Battery
Development – Basics and Optimization Strategy. In: ATZworldwide 2/2018,
pp. 48-53
[11] Kampker, A.; Bergweiler, G.; Hollah, A.; Fiedler, F.; Hansen, J. O.:
Produktionseffizienz in der Kleinserie (ProeK). In: ZWF 112 (2017),
No. 9, pp. 576-579
[12] Kampker, A.; Schleser, M.; Triebs, J.; Berg, A.; Hansen, J. O.;
Fiedler, F.: Bauteilintegrierte Vorrichtungsfunktionen im Karosseriebau.
In: ZWF 113 (2018), No. 11, pp. 767-770
[13] Bergweiler, G.: Verfahren zum Verbinden von Blechteilen und
Blechanordnungen. Aachen, Technische Hochschule, Patent, 2017
DISCOVER THE
WORLD’S LEADING
SPECIALIST
MAGAZINE FOR
THE AUTOMOTIVE
SECTOR!
TAKE A FR E E TE S T D R I V E
AT Z-MAG A Z I N E .C O M

For 120 years, ATZ – Automobiltechnische Zeitschrift has been

presenting cutting-edge solution concepts in automotive
development and the very latest information for the everyday
work of engineers relating to every aspect of the complete
vehicle – whether it is the chassis or body, lighting technology
or NVH, packaging or thermal management. atz-magazine.com
ATZ worldwide 06|2019 77
C ONFERENCE REP ORT

Automated Driving –
The Big Picture
More than 180 experts from
the fields of research and
industry took part in the 5th
ATZ Conference “Automated
Driving – From Driver Assistance
to Autonomous Driving” at April
2 and 3, 2019 in Wiesbaden
(Germany). During the course
of over 30 presentations, they
had the opportunity to find
out about the latest technical
© Uli Regenscheit | ATZlive

developments, improvements
in human-machine interaction,
and insurances issues relating
to autonomous driving.

LEVEL-3 FUNCTIONS – spondingly expensive. The investment cyclists, cars, buses, and traffic signs,
PLAYING THE WAITING GAME will have to pay for itself by replacing takes around 3.5 min per image with
working time. “The driver represents today’s technology. Given a typical train-
We are promised that the benefits of one third of the TCO (Total Cost of ing dataset of 100,000 images, around
automated driving will include the Ownership) of trucks,” explained Dieck- 737 days would be needed for annota-
opportunity to relax and read a book mann and added: “Even with level-4 tion. However, Pfeffer did admit that this
during the journey, check emails in a functions, using an automated system was an exaggerated example, because
traffic jam, or watch a movie. But the instead of a driver can be profitable.” in practice it is not necessary to process
car manufacturers are many years In his view the systems will primarily every single image.
away from true autonomous driving be used on the highway for journeys If the images are generated artificially
on level 5. The reality is that there is from hub to hub on precisely predefined using simulation software, a lot of time
not yet a single production model on routes. This will only be possible in a can be saved. In this context it is inter-
the road with highly automated SAE few years’ time. esting to note that the quality of the data
level-3 functions that allow the driver The engineers have a lot to do before obtained exclusively by means of simu-
to perform other activities under certain we reach that point. They are increas- lation is not particularly good, but com-
conditions. ingly using Artificial Intelligence (AI) bining it with real driving data can lead
The delays make one thing very clear: systems to help them in areas such to significant improvements. Even data-
highly automated and autonomous driv- as the validation of functions. Just like sets consisting of 5 % real driving data
ing is much more complex than the a person, an AI system must first be and 95 % simulation data provide very
marketing strategists would have us “trained” as accurately as possible. good results.
believe. This was a subject covered by Raphael Pfeffer of IPG Automotive
Dr. Thomas Dieckmann, Head of Pre- explained that although training data
A COMBINATION OF
development at the automotive industry play a decisive role in the performance
SIMULATION AND REAL DRIVING
supplier Wabco. The first automated of an AI system, the evaluation of the
vehicles will have high-tech sensors, a data can in theory take years. Manual Experts disagree on the number of test
huge computing capacity, and complex annotation, in other words, marking rel- kilometers needed to validate automated
software and will therefore be corre- evant information such as pedestrians, systems reliably: The figures range from
78

240 million to several billion kilome-
ters. Simon Tiedemann from Elektrobit
Automotive GmbH highlighted the fact
that this generates enormous volumes
of data which have to be processed and
evaluated. In the case of real driving
data, covering 240 million km would
take 25 years with 100 vehicles. How-
ever, simulating this with 100 comput-
ers would take five years. Using the
cloud, 240 million km could be simu-
lated in approximately two weeks. This
would speed up the development pro-
cess enormously, but it would also be
extremely expensive. In addition, the
huge amounts of data – and these could
be in the petabyte range (1 petabyte
equals 1024 terabytes) – would first
have to be uploaded to the cloud. Even
at fast upload speeds, this would take a
relatively long time.
DRIVERS DISTRACTED
BY OTHER ACTIVITIES
While the engineers are focusing on the
technical challenges, it is important that
we do not neglect the drivers and users
of the new functions. Will they respond
quickly enough after a long automated
journey? What other activities can they
carry out without risk? Which situations
could overtax drivers? Paula Laßmann
from the University of Stuttgart looked
into these and other questions. During
the courses of studies involving test
subjects in the simulator, she and her
team investigated how the response time
changes for different activities. She is
planning a real driving study during the
coming year.
© Uli Regenscheit | ATZlive
ATZ worldwide 06|2019
QUOTES
A QUESTION OF LIABILIT Y
Hermann Meyer,
In addition to the hardware and the
Conti Teves:
interaction between the driver and the “We are developing
vehicle, another area is becoming solutions for around 500
increasingly important. Disregarding for smart city projects that
are currently underway
a moment the technical considerations, it
in China. Approximately
is clear that automated driving will only 30 % of the traffic in
gain wide acceptance when the insur- cities involves people
ance issues have been resolved. Once looking for parking
spaces. We plan to
again this year, the insurance strand of
reduce the stress on
the conference, which ran in parallel drivers and the volume
with the technical presentations, proved of traffic with functions
to be very popular. Experts from insur- such as Automated
Parking.”
ance companies such as R+V, E+S and
Allianz, together with lawyers and uni-
versity lecturers, discussed liability
issues, payment of compensation in the
Dr. Thomas Dieckmann,
event of an insurance claim, and the rel-
Wabco:
evance of data from automated vehicles “Automated commercial
for the aftermarket. Verena Reuber, vehicles are likely to
Project Manager at R+V-Versicherung, come onto the market
first in China, where
described the company’s experiences
there is a great deal of
from the test operation of two autono- political interest.”
mous shuttles. Fabian Pütz from Cologne
University of Applied Sciences discussed
the possibility of liability costs being
shared between insurance companies Paula Laßmann,
and vehicle manufacturers. Universität Stuttgart:
Next year experts from the fields of “On the basis of our
studies, we can already
research, development, and insurance
say that it would be
will once again have the opportunity to best if drivers did not
find out about the latest technological take part in any visual
and insurance-related innovations in activities during a journey
in an automated car.”
the field of automated driving. The 6th
international ATZ Conference will take
place on March 31 and April 1, 2020
in the Rhein Main Congress Center
in Wiesbaden.
PHOTOS: Uli Regenscheit | ATZlive
Mathias Heerwagen
The parallel insurance strand of the conference
proved to be very popular; presentations were given
by experts from insurance companies such as R+V
Allgemeine Versicherung, E+S Rückversicherung,
and the Allianz Center for Technology
79
IMPRINT

W O RL DW ID E

www.ATZonline.com

Founded 1898 as „Der Motorwagen“ Organ of the VDI-Gesellschaft Fahrzeug- und Verkehrstechnik (FVT)
Organ of the Forschungsvereinigung Automobiltechnik e. V. (FAT) and of the Normenausschuss Automobiltechnik
(NAAutomobil), formerly known as FAKRA, in the DIN Deutsches Institut für Normung e. V.
Organ of the Wissenschaftliche Gesellschaft für Kraftfahrzeug- und Motorentechnik e. V. (WKM)

06|2019 _ Volume 121

Springer Vieweg | Springer Fachmedien Wiesbaden GmbH
P. O. Box 15 46 · 65173 Wiesbaden · Germany | Abraham-Lincoln-Straße 46 · 65189 Wiesbaden · Germany

Amtsgericht Wiesbaden, HRB 9754, USt-IdNr. DE811148419

Managing Directors Stefanie Burgmaier, Joachim Krieger, Juliane Ritt

Director Production Ulrike Drechsler ISSN (Online) 2192-9076

EDITORIAL STAFF ADVERTISING

SCIENTIFIC ADVISORY BOARD
EDITOR IN CHARGE HEAD OF MEDIA SALES YOUR HOTLINE
Dipl.-Ing. Dirk Adamczyk Dr. Johannes Liebl (responsible for advertising)
ZF Friedrichshafen AG EDITOR IN CHIEF Volker Hesedenz TO ATZ
(responsible for editorial content) phone +49 611 7878-269 · fax +49 611 7878-78269
Dipl.-Ing. Dietmar Bichler Dr. Alexander Heintzel volker.hesedenz@springernature.com Editorial Staff
Bertrandt AG phone +49 611 7878-342 · fax +49 611 7878-462
alexander.heintzel@springernature.com
SALES MANAGEMENT AUTOMOTIVE
Rouwen Bastian
☎ +49 611 7878-341
Dr.-Ing. Joachim Damasky Customer Service
VDA/FAT DEPUTY EDITOR IN CHIEF phone +49 611 7878-399 · fax +49 611 7878-78399
Dipl.-Ing. Michael Reichenbach rouwen.bastian@springernature.com ☎ +49 6123 9238-204
Dr. Thomas Dieckmann phone +49 611 7878-341 · fax +49 611 7878-462 Advertising
DISPLAY AD MANAGER
Wabco GmbH michael.reichenbach@springernature.com
Susanne Bretschneider ☎ +49 611 7878-399
MANAGING EDITOR phone +49 611 7878-153 · fax +49 611 7878-443
Prof. Dr.-Ing. Lutz Eckstein
Kirsten Beckmann M. A. susanne.bretschneider@springernature.com
RWTH Aachen phone +49 611 7878-343 · fax +49 611 7878-462
kirsten.beckmann@springernature.com AD PRICES
Dr.-Ing. Ulrich Eichhorn The articles published in ATZworldwide are produced
Advertising ratecard from October 2018.
EDITORS with the greatest possible care and attention. However,
Volkswagen AG
Dipl.-Journ. (FH) Mathias Heerwagen the editors do not accept any liability for the correctness,
SALES | MARKETING | OFFPRINT
Dipl.-Ing. Frank F. Esser phone +49 611 7878-188 · fax +49 611 7878-462 completeness and current accuracy of the content
mathias.heerwagen@springernature.com HEAD OF SALES + MARKETING printed. The content of advertisements is the sole
Ford Werke GmbH
Thomas Schneider M. A. Jens Fischer responsibility the advertiser.
Dipl.-Ing. Klaus Fröhlich phone +49 611 7878-121 · fax +49 611 7878-462
MARKETING
thomas.schneider@springernature.com Every annual subscription includes access to the
BMW AG Arzu Ugur
Dipl.-Ing. Thomas Siebel online archive at Springer Professional. Access
phone +49 611 7878-261 · fax +49 611 7878-462 phone +49 611 7878-511 · fax +49 611 7878-78511
Prof. Dr.-Ing. Uwe Dieter Grebe arzu.ugur@springernature.com
is available exclusively to the individual subscriber
thomas.siebel@springernature.com and is not transferable.
AVL List GmbH Robert Unseld OFFPRINTS
Prof. Dr.-Ing. Peter Gutzmer phone +49 611 7878-257 · fax +49 611 7878-462
Martin Leopold No liability can be accepted for the unsolicited sub-
robert.unseld@springernature.com
Schaeffler AG phone +49 2642 907-596 · fax +49 2642 907-597 mission of manuscripts, photographs or illustrations.
Marc Ziegler
phone +49 611 7878-120 · fax +49 611 7878-462 leopold@medien-kontor.de
Dr.-Ing. Markus Heyn
marc.ziegler@springernature.com
Robert Bosch GmbH PRODUCTION | LAYOUT
INTERNATIONAL AND Kerstin Brüderlin
Dipl.-Ing. (FH) Harald Keller DIGITAL MAGAZINES phone +49 611 7878-173 · fax +49 611 7878-78173
Edag Engineering GmbH Dipl.-Reg.-Wiss. Caroline Behle
kerstin.bruederlin@springernature.com
phone +49 611 7878-393 · fax +49 611 7878-462
Matthias Kratzsch caroline.behle@springernature.com
Kim Woods All persons specified may be contacted
IAV GmbH by post at the publisher’s address.
phone +49 611 7878-337 · fax +49 611 7878-462
Dipl.-Ing. Ralph Lauxmann kim.woods@springernature.com
CUSTOMER SERVICE | SUBSCRIPTIONS
Continental Teves AG & Co. oHG ENGLISH LANGUAGE CONSULTANT Customer Service Springer Fachmedien
Paul Willin 65341 Eltville, Germany
Dr.-Ing. Peter Mertens
FREELANCERS phone +49 6123 9238-204 · fax +49 6123 9238-244
Audi AG Dipl.-Ing. (FH) Richard Backhaus, Dipl.-Ing. (FH) Monday to Friday, 8 a.m. to 5 p.m.
Andreas Burkert, Dipl.-Ing. (FH) Andreas Fuchs, (according to German time)
Prof. Dr. Dipl.-Ing. Peter Pfeffer
Gernot Goppelt, Stefan Schlott worldwide-magazines@vuservice.de
Hochschule München
ONLINE | ELECTRONIC MEDIA www.my-specialized-knowledge.com/automotive
Dr.-Ing. Otmar Scharrer Channel Manager Automotive
Christiane Köllner ORDERING OPTIONS
Mahle International GmbH
phone +49 611 7878-136 · fax +49 611 7878-462 The eMagazine ATZworldwide appears 11 times
Dipl.-Ing. (FH) Dipl.-Wirtsch.-Ing. (FH) christiane.koellner@springernature.com a year for subscribers. Ordering options and details
Wolfgang Schwenk Patrick Schäfer on the subscription conditions can be found at
phone +49 611 7878-557 · fax +49 611 7878-462 www.my-specialized-knowledge.com/ATZworldwide.
Opel Automobile GmbH
patrick.schaefer@springernature.com
Dr. Michael Steiner All rights reserved.
CORPORATE SOLUTIONS
Dr. Ing. h.c. F. Porsche AG Head
Markus Bereszewski Reproduction: This magazine and all its contents,
Dipl.-Ing. (FH) Derrick Zechmair phone +49 611 7878-122 · fax +49 611 7878-78122 including all images, graphics and photographs,
Valeo GmbH markus.bereszewski@springernature.com are protected by copyright. Unless its usage is, in
exceptional cases, expressly permitted by copyright © Springer Vieweg |
PROJECT MANAGEMENT I ASSISTANCE
Dr. Frederik Zohm law, no part of this magazine may be reproduced Springer Fachmedien Wiesbaden GmbH,
Yeliz Konar
MAN Truck & Bus AG without the prior written consent of the publisher. Wiesbaden 2019
phone +49 611 7878-180 · fax +49 611 7878-462
This applies in particular to copies, reprints, Springer Vieweg is an imprint of Springer Fachmedien
yeliz.konar@springernature.com
adaptations, translations, microfilms, provision Wiesbaden GmbH, which is part of the publishing group
ADDRESS of public access and the storage and processing of Springer Nature.
Abraham-Lincoln-Straße 46 · 65189 Wiesbaden this magazine in whole or in part in databases and
P. O. Box 1546 · 65173 Wiesbaden, Germany other electronic systems and its distribution or usage
redaktion@ATZonline.de via electronic systems.

The next issue will be published on July 1, 2019
CO V E R S T ORY
Acoustics and NVH – Using Sound
Engineering to Track Down the Noise
© Daimler
Quiet electric cars present vehicle
acoustics specialists with new chal-
lenges, because the noise of the
combustion engine is no longer pres-
ent to mask other sounds. Impor-
tantly, one of the quality criteria
when buying a passenger car is the
sound it makes. The 5th International
ATZ Automotive Acoustics Confer-
ence, sponsored by Autoneum, which
is taking place on July 9 and 10,
2019, will highlight the effects of
these changes in powertrain technol-
ogy during its presentations. Experts
from the worlds of industry and
research will gather in Rüschlikon
near Zurich (Switzerland) to discuss
whether good NVH performance can
still be achieved in the near future
when lightweight structures, smaller
turbocharged gasoline engines, and
hybrid and electric drives are making
the job of acoustics engineers more
ATZ worldwide 06|2019
IS SUE 07- 0 8 | 2 019 PRE VIE W
DE V E L OP MEN T
Challenges in the Digital Transformation
of the Automotive Industry
Agile Development with Physical
Prototypes for a Better Project Planning
Effectiveness of Autonomous
Emergency Braking Systems and
Automated Driving Functions
Virtual Simulation of
an Electro-hydraulic Braking System
Tuning of Audio Systems with
Engine Simulation Tools
Artificial Intelligence for Vehicle Interiors
difficult. Thus, the Research Institute R E S E A RCH
of Automotive Engineering and Vehi-
cle Engines Stuttgart (FKFS) and Ethernet-based Network Architectures for
Daimler have developed a technique Future Realtime Systems in the Car
for reducing irritating noises from
the cooling fan. This requires an Networked Drivetrain Development in
improved methodology of the kind a Physical and Virtual Environment
also developed and used by Bosch
Engineering. The company is carrying
out investigations into how the sound
characteristics of a diesel engine
combustion process can be quanti-
fied without additional NVH measure-
ment systems and using only the cyl-
inder pressure profile. Klaus Genuit,
Founder and Managing Director of
Head acoustics, explains in the ATZ
interview why the artificial sound of
electric vehicles is counterproduc-
tively and whether the irritating whis-
tling noise emitted by the power elec-
tronics in Formula E racing cars will
soon be a thing of the past.
81
GUEST C OMMENTARY
Mastering Complexity
Requirements on mechatronic chassis systems in our cars
are continuously increasing. Just a few years ago, development
and validation was realized highly component-oriented with
only a few interfaces within the vehicles whereas the content
could be mastered well, using classic development methods.
The permanent reduction of energy demand, the fulfill-
ment of the highest safety and security goals as well as
modern connectivity functionalities meanwhile require
a high degree of interconnection and perfect coordination
of all systems. The same applies to the rapidly developing
functions for assisted and automated driving, which ideally
can incorporate the whole controller network into their work.
At the same time, there is the desire to offer customer an even
greater driving experience with more dynamics and a higher
level of comfort with each new vehicle model generation.
The examples mentioned outline only an excerpt of the
rapidly growing catalogue of requirements, but already give
an idea of the soaring challenges in the development of
mechatronic chassis systems. Understandably, development
resources such as budget, infrastructure and time cannot
grow to the same extent.
Therefore, we need an answer to the question of how we
can master this complexity under steady conditions. We have
82
Dr.-Ing. Christoph Bittner
Director Development Chassis
Systems at Dr. Ing. h. c. F.
Porsche AG
© Porsche
to break new ground with standardized yet flexible pro-
cesses. Methods and tools need a fundamental reorien-
tation. Our engineers are therefore working on different
approaches to achieve the required development efficiency,
while at the same time continuing to raise the quality level.
In order to achieve a high degree of maturity well before the
first physical prototypes are built, virtual development is con-
stantly being optimized and extended to new areas of appli-
cation. Latest methods allow a significantly increased quality
of validity with reduced modelling effort. In parallel, novel
test benches are being integrated deeper into the process as
a development and validation tool. The high level of connec-
tivity in the target vehicle is reflected in an equally intensive
connectivity of the test environment. Finally yet importantly,
we use agile development methods for strengthening inter-
disciplinary teamwork across department boundaries and
we intensify cooperation within the group and with our
suppliers and engineering partners.
In addition to the operational technical measures, how-
ever, we at Porsche are particularly interested in further
enhancements of our corporate culture in order to consoli-
date the spirit of the team and to sustainably achieve the
best possible result for our customers.
Get a Grip
on Success

For those who want to reach new heights, only the best will do. With adhesion, the international trade magazine
for industrial adhesives and sealing technology, you will benefit from valuable insider knowledge, practical
information and the latest trends and technologies. Get access to the industry‘s largest knowledge pool – essential
for gaining competitive advantage. Try our comprehensive service: four issues per year available as print and
digital editions, website and newsletters, now optimized for mobile devices. As a subscriber, you also gain access
to our comprehensive online archive with all contributions since 2003, plus the “Adhesives Technology Compendium”
- the reference book for adhesives practitioners.

For more information:

www.my-specialized-knowledge.com/adhesion
POLISH YOUR KNOWLEDGE
WITH IST – International Surface Technology

© creative republic 2015 / Foto: © shutterstock & creative republic

Success requires going beneath the surface. Enhance your industry knowledge with Germany’s leading magazine for surface
technology. International Surface Technology (IST), the “Best of JOT” in English takes you layer-by-layer, with articles that
are practical and user-oriented. Read 3 issues per year and also receive the digital edition, plus free access to the online archive
featuring expert articles.

Discover even more at: www.my-specialized-knowledge.com/ist