NISSAN 2022

No.

TECHNICAL 88
REVIEW

Nissan Intelligent Factory Starting from ARIYA／Next-generation flagship: ARIYA

NISSAN TECHNICAL REVIEW

2022 No. 88

“NISSAN TECHNICAL REVIEW” WEBSITE

https://www.nissan-global.com/EN/TECHNICALREVIEW/
 NISSAN TECHINICAL REVIEW 2022 No.88

Contents
Published in September, 2022

Preface
Preface Nissan s DNA develops Super evolution a production system in harmony with Blue Earth ………………………………… 1
Hideyuki Sakamoto

Special Feature 1 Nissan Intelligent Factory Starting from ARIYA

1. Nissan Intelligent Factory Starting from ARIYA…………………………………………………………………………………………………… 7
Hidetoshi Okaguchi

2. Application of Next-Generation Underfloor Process Accommodating Various Powertrains………………………………………… 13

Hiromasa Tsuchikane Atsushi Ugaji Yoshihisa Naito

3. Manufacturing Innovation in a New Motor Line through Digitalization and Internet of Things (IoT)………………………………… 19
Satoru Iizuka

4. Process Innovation Through the Development of a Simultaneous Body/Bumper Painting System………………………………… 23

Takeshi Goto Tomoyuki Okamoto Chie Michiura

5. Creation of a Quality Gate through Automated Technology of Appearance and Specification Inspection……………………… 29
Teruyuki Ishiwata Mitsuru Hirayama Daisuke Tanaka Takayuki Moriya
Yasuo Yamada Makoto Yamada Takeshi Honda

6. Improvement of the Equipment Operation Rate and Process Capacity by

Implementing Predictive/Preventive Maintenance and Remote Maintenance Technology…………………………………………… 37
Katahiro Ogawa Satoru Sakurai Koichi Makino Takuya Tsumoto
Susumu Abe Masayoshi Hayakawa Kazunari Seo Minoru Kono

7. Development of Die-Cast Technology for Realizing Innovative Thin-Walled Light-Weight Cylinder He…………………………… 43
Yuuta Sugiyama Shinichi Tsuchiya Hiromichi Kume Takahiro Nishi Machi Tabuchi

Special Feature 2 Next-generation flagship: ARIYA

1. ARIYA: new generation of flagship EV…………………………………………………………………………………………………………… 49
Eiichi Akashi Hikaru Nakajima

2. Development of EV-specific platform CMF-EV for high packaging efficiency…………………………………………………………… 53

Masahiro Oonishi

3. Integrated interface display………………………………………………………………………………………………………………………… 57

Tetsu Obata Takashi Eguchi Yasutomo Sasanuma

4. Development of new NissanConnect……………………………………………………………………………………………………………… 61

Toshiro Muramatsu Kazuhiro Funai Masaru Tasaki Shintaroh Murakami Hiroyuki Kamishima

5. High-capacity lithium-ion battery for ARIYA……………………………………………………………………………………………………… 65

Junji Katamura Kentaro Hatta Keisuke Wakabayashi Masanori Takagi Naoto Todoroki

6. Newly developed motor to realize ARIYA s performance……………………………………………………………………………………… 69

Takahito Okubo Hiroki Wada

7. Nissan s e-4ORCE : Proposition for evolution of electric AWDs

Aiming for a more exhilarating and confident driving experience……………………………………………………………………… 73
Hiroyuki Togashi Yutoku Miyagoshi Takeji Katakura
Technical papers
Study about Workload Reduction by Multiple Modal Interface Using Gaze and Voice Recognition…………………………………… 77
Yusuke Morita Sou Tamura

The Development of Bearings for the Multi-link Crank Mechanism to Achieve Variable Compression Ratio Engine……………… 85
Katsutoshi Nakamura Katsuya Moteki Satoru Ookuma

Surrogate Model Development for Prediction of Car Aerodynamics Using Machine Learning………………………………………… 93
Kei Akasaka Fangge Chen Takehito Teraguchi

A Study on Pitch Characteristic to Reduce Line Trace Deviation in Small Steering Angle……………………………………………… 101
Mitsunori Tao Naoya Machida Yutaka Hayashi Ken Nagasao

Technical Awards
An Approach to Exploring Vehicle Motion to Enhance Ride Quality of Passenger………………………………………………………… 109
Mitsuhiro Makita Akihiro Matsushita Yoshinori Kusayanagi Masahiro Miura

Development of driving assistance system in multi-lane highway……………………………………………………………………………… 117

Yohei Taniguchi Katsuhiko Degawa Seiji Tokunaga Shimpei Nagae Koji Sasaki

Strength evaluation of CFRP tank based on meso-scopic zooming analysis ……………………………………………………………… 119
Shinichiro Takemoto
 Preface

Nissan s DNA develops Super

evolution a production system in
harmony with Blue Earth
Executive Officer, EVP Hideyuki Sakamoto

solutions to realize new functions and achieve high

1. Introduction performance that will resonate with customers and
society, demonstrating the value of technological
Nissan Motor is a car manufacturing company where evolution that has been inherited from Nissan s founder
automotive engineers play a central role in manufacturing to today s engineers.
cars. Nissan's history began in 1933 with the development
and mass production of domestically manufactured 2. Societal Changes and the Automotive
automobiles based on the spirit of its founder, Yoshisuke Industry
Ayukawa, to "do what others will not do." This positioned
Nissan Motor as a company in the early days of the Nissan Motor envisions four missions for car
industry; presently, Nissan is a leading player in Japan s manufacturing and the next generation of the automotive
modern automobile industry. industry.
The domestic production of automobiles marked the The first is a shift from a labor-intensive business
beginning of the evolution of automobile technologies. model. The method of mass production for the Ford Model
Since then, Nissan has followed the values and vision of its T was introduced by Ford 100 years ago. It was a
founder at the company s core and expanded their business revolutionary method and involved many people who
internationally through exports and overseas production. worked together to efficiently produce cars. As
Nissan has experienced many waves of challenges, modernization has progressed, various types of
including global social issues such as environmental and automation have been incorporated into automobile
air pollution, traffic accidents, oil shocks, fuel economy manufacturing processes, but many manually intensive
regulations, and trade friction, among others. However, by methods are still used presently. However, given the
using technology created by Nissan s innovative and recent shortage of automobile factory workers in Japan
visionary engineers, the company has continued to and other countries around the world, which has been
successfully overcome new challenges in the industry, exacerbated by the COVID-19 pandemic, it is becoming
even where other companies have faltered. Nissan s increasingly difficult to ensure work by many people to
storied history is an excellent demonstration of how new produce cars. Therefore, it is necessary to move away
innovations in automobiles and manufacturing can be from the current production model and establish a
introduced to the market for the first time. streamlined automated production system with a
reformed working environment.
Connected, automated, shared/service, and electric Secondly, this trend is expected to worsen, given the
vehicles (EVs) collectively referred to as the "CASE" declining birthrate and aging population, which make it
areas and carbon neutrality have been widely covered increasingly difficult to provide young people with the
in the media as once-in-a-century issues for the global advanced training that is necessary for challenging car
automotive industry. The development of future manufacturing tasks.
automobiles will aim to realize visions of a technologically Thirdly, additional flexibility in responding to
advanced future and will rely on the large-scale and unexpected situations is also needed in production
high-level development of new technologies that interact systems. Various factors such as the COVID-19 pandemic,
with complex social systems. Automotive engineers play the shortage of semiconductors, and major disruptions in
a critical role in developing these technologies, and they the logistics network have become extremely serious
will continue to do so in the future. The focus of issues that can cause fluctuations in automobile
technological development is on the value it offers to production, making it crucial to develop systems that can
customers and society. With this in mind, automotive deal with manufacturing disruptions. In addition,
engineers should fully understand and apply market and batteries, e-components, and their raw materials will
technological trends to create the best possible technical also face issues with unstable supply in the future, which
solutions that perform key functions for customers and will undoubtedly cause major fluctuations in production
society. These values have been engrained in Nissan's that must be managed.
work since its inception and will remain as a guide for the Finally, the automotive industry should take concrete
company in the future. actions to combat climate change and contribute to
The special feature in this issue of Nissan Technical achieving carbon neutrality in conjunction with efforts
Review includes technical reports on the best available that are currently underway around the world.

NISSAN TECHNICAL REVIEW No.88 (2022) 1

 Preface Nissan s DNA develops Super evolution a production system in harmony with Blue Earth

Mission Minimize labor intensity in manufacturing Mission Reform factory working environments promoting the commercialization of EVs.
1 2
Men Women
In addition, efforts have been made to reduce CO2
emissions by reducing the amount of energy used to
produce these vehicles. Further innovations will be made
1913: Ford Model T enters
production
Today Declining birthrate and
aging population
in the future to make this a reality.
Source:

Mission Mission
Increase flexibility in responding to Address climate change through concrete Low-cost battery
3 4 EV - Cobalt not required
supply chain disruptions actions EV Leadership - All-solid-state battery Cost-competitive
electric vehicles
Renewable energy use
and the creation of
(1st generation - Cell to pack for downsizing social value
Promote carbon LEAF- ) System
- Improvement of inverter efficiency
neutrality
All new models to be launched in

Power generation engine

major markets will be electric
vehicles from the early 2030s.
Nissan's unique
Innovations in battery technologies for the development of more
competitive and efficient EVs HEV HEV system - Peak thermal efficiency: 50%
Development of a battery eco- System
Development of e-POWER with system that will contribute to
greater energy efficiency distributed power generation by - Improvement of inverter efficiency
renewable energy

2005 2010 2020 2030 2040 2050

Improving energy and resource use
efficiency through innovations in production
100%
Pandemics and natural disasters
technology

Respond to environmental concerns electrification of

vehicles
Battery
ecosystem - Reusable batteries (4R)
- V2H Carbon neutral
Four priorities for the future of automotive manufacturing
Distributed power generation
- Emergency power
supply (Blue switch)
life cycle

Innovations in car manufacturing

Manufacturing - Common powertrain mounting
- Multimaterials

High energy efficiency

How should production activities be developed for the

- Improvement of energy mix
- Minimization of energy loss

new car of the future with these four conditions met? Initiatives for life-cycle carbon neutrality
Cars and their production will become highly sophisticated
in the coming years as Electrified , Intelligent , and
Connected become the industry standard. Similarly, Under the Nissan Green Program, the CO2 emissions
manufacturing is becoming increasingly demanding as from car production were successfully reduced by 30%
guaranteeing the functionality of various systems per car from 2005 to 2019. The current energy used in
becomes more complex. For example, modern manufacturing is composed of 71% electricity and 29%
manufacturing requires a large amount of software to be gas and fossil fuel. Finding an approach to reduce this
installed simultaneously, many sensors to be calibrated, gas and fossil fuel usage to 0% will be a priority in the
and advanced self- cddiagnostic systems and fail-safes forthcoming years, and future innovations will be
to be verified, which poses challenges for manufacturing. invaluable in making carbon-neutral manufacturing and
As an additional layer of complexity, a wide range of vehicle life-cycles a reality.
customizations is required in car manufacturing,
including large batteries with long-range and standard ■ Shift from dependence on gas and fossil fuels toward carbon
neutrality in production and strengthen responses in Japan, China,
types for EVs, highly automated driving technology that and North America
enables hands-off driving, and standard driver assistance
systems for automated driving intelligence technologies.
0.75
[Global CO2 emissions in FY19>
0.514 tons-CO2 per car (2.2 million tons per year)

Resilient production facilities and technologies to

0.70
GLOBAL CO2 EMISSIONS PER

manufacture such complex and advanced vehicles are

0.65
Others
Gas and fuel
[T-CO2/CAR]

19%

critical to Nissan's future success.

Japan
CAR

0.60
29% Europe 35%
By energy type 6% By region
0.55

North America
0.50
21%
Electrified Intelligent Connected China
Electricity 71% 19%
0.45

Reduction of CO2 emissions during production

Meeting the needs of increasingly advanced products 3.2 Initiatives for Carbon Neutrality in Production
A roadmap to achieving carbon neutrality in Nissan's
3. Initiatives to Address Climate Change and plants by 2050 has already been determined with ongoing
a Carbon Neutral Life-Cycle technology development. Current plants operate with
electricity and high-pressure air from compressors as
3.1 Nissan Green Program power sources, as well as steam, heat, and electricity
Countries around the world are now committing to converted from gas and fossil fuel. Energy consumption is
achieving carbon neutrality by 2050 and implementing planned to be drastically reduced by 2030. Moreover,
various policies to tackle climate change. solid-oxide fuel cells (SOFCs) using clean alternative
Nissan has also committed to achieving carbon fuels will be used to generate carbon-neutral power,
neutrality by 2050 by electrifying vehicles in major expanding the trial of clean energy generation in
markets in the early 2030s and by promoting innovation combination with renewable energy sources, such as
in production technology to achieve carbon neutrality in solar radiation and wind. All production equipment
car manufacturing processes. Nissan was among the first running in the plants will be electrified by 2050, and all
companies to take action in reducing CO2 emissions energy sources will be replaced by renewable and clean
through the Nissan Green Program starting in 2005. energy generation.
Under the program, the company began the mass
production of its first-generation EV, "LEAF," in 2010,
offering it to the mainstream market as an EV for daily
use. Subsequently, e-POWER was launched with the aim
of helping customers feel the joy of driving an EV and

2 No.88 (2022) NISSAN TECHNICAL REVIEW

 Preface Nissan s DNA develops Super evolution a production system in harmony with Blue Earth

2019-2030 2030-2050
■ Alternative fuel and power generation
- Energy reduction - Electrification of some equipment and tools
Applying the experience of developing SOFCs for automotive to stationary systems
- Complete electrification of plant equipment
- Introduction of renewable energy
- Partial implementation of alternative energy generation
- Full application of renewable and alternative energy generation - FY22: Small-scale trial (5 kW or less, 60% efficiency)
- FY24-25: Scale-up trial (30 kW or more, 70% efficiency)
2019 2030 2050

SOFC (fuel cell) Power generation

- Bioethanol
- Hydrogen
Electricity Electricity - Artificial methane
Compressor Compressor
Air
Air (Zero CO2)

Cogeneration
Alternative
LB
Power
generation by Alternative
LB
Power
generation by
■ Renewable energy installation
fuel SOFC fuel SOFC

Gas and fuel Gas and fuel

Boiler Boiler
Steam Steam (Zero CO2)

Roadmap for reducing CO2 emissions from production

Nissan UK Nissan México Nissan UK
Wind power generation: 6.6 MW Biomass power generation: Solar power generation (planned): 20 MW
Solar: 5.5 MW 2.7 MW

One of the most difficult production processes to Alternative fuel and power generation strategies
electrify is the casting process, especially for carbon steel
commonly used for drivetrains, where coke is used to
carburize steel while it is melted. Further technological Nissan is aiming to reduce CO2 emissions by 41% by
developments are required to use high-frequency 2030 by drastically reducing the energy required to
furnaces for the electrification of this process. This manufacture cars at Nissan's plants and by trialing
technology will also be valuable in the efforts to completely carbon-neutral power generation strategies. Carbon-
electrify the melting process of aluminum alloys neutral power generation and renewable energy
commonly used for powertrains. Furthermore, the tools procurement will enable 100% carbon neutrality by 2050.
used in factories will also be replaced by electric tools The R&D and production technology departments are
instead of air tools. already collaborating to develop the necessary
It is important to note that the trends in the technologies to achieve this ambitious but critical goal.
development of advanced technologies in the energy
industry need to be closely monitored, as the future 0.514 t/car

development of new alternative clean fuels capable of

(2.22 million t/year)

CO2 (1) Innovative

directly providing thermal energy as an alternative to

Natural gas -41%
and fossil technologies
(3) Energy saving (1) Innovative
fuels
CO2 emissions [t/car]

technologies

conventional fuels (light oil, natural gas, etc.) will require

(3) Energy saving
(2) Electrification
Natural gas
and fossil fuels

us to pivot our technological development.

CO2
(2) Electrification -100% Alternative
Electricity energy
Electricity

Renewable
(4) Power generation (4) Power generation energy
from alternative from alternative
energy energy
2019 2030 2050
■ Steel casting ■ Aluminum melting ladle heating Technology
development
Technology
development Technology Technology Technology
for casting for painting Development Development Development
Molten metal outlet
Furnace lid electrification electrification
(1) Innovative
technologies Applied to plants in Japan and Complete
(2) Electrification
overseas successively electrification

Coil
(3) Energy saving 100% electrification of
Yoke Add 2%/year LED100% air conditioning
Furnace wall
electrode
(4) Power
Coke generation from Keep introducing
renewable energy SOFC trial Applied
successively
100% carbon-free
electricity
alternative energy

Now (cupola) Electrification

(high frequency furnace)

Numerical targets for CO2 emissions reduction during

■ Tool production

Gas direct heating Electrification

(far infrared heater)

3.3 Approaches to Coexistence between Factories

Pneumatic tool Electric tool
and the Local Environment
生産 排出削減の主な取組み Coexistence with the local environment by minimizing
Ongoing efforts to reduce CO2 emissions during production
through electrification environmental impacts is also important for production
activities. The area surrounding the Tochigi plant, which
implemented the Nissan Intelligent Factory initiative, is
Relying solely on external sources for clean energy full of greenery and a variety of ecosystems. In an attempt
poses challenges to business continuity due to unstable to protect the natural environment, environmental
supply and the inability to control costs. Therefore, plans dependency was minimized, and load during production
are being developed to test and implement clean energy was maintained at sustainable levels. Furthermore, the
power generation in Nissan s plants. Carbon-neutral plant acts as an environmentally friendly manufacturing
power generation from bioethanol using an SOFC will be site by protecting the ecosystem, managing water quality,
tested at the Tochigi Plant with plans to implement this maintaining the landscape, and utilizing renewable
strategy worldwide. Nissan's Sunderland plant in the UK energy. For example, the new painting booth at Nissan
has already begun to use solar and wind power and will Intelligent Factory does not use water to collect paint
further incorporate large-scale clean energy supplies and mist. Eliminating wastewater from the painting booth
solar power generation via microgrids in the future. made it possible to eliminate the need for energy for
wastewater treatment. Wastewater from other processes
is still thoroughly treated to control the water quality.
The small stream in the factory that receives treated

NISSAN TECHNICAL REVIEW No.88 (2022) 3

 Preface Nissan s DNA develops Super evolution a production system in harmony with Blue Earth

wastewater is inhabited by fireflies, which give off insufficient to realize the mass production of CFRPs.
beautiful light in the summer. However, Nissan introduced a mass-production
technology for CFRPs in 2020 through the application of
■Minimize environmental dependence and energy demands during a resin flow simulation, making it feasible to achieve a
production to protect natural environment around the plant cycle time of two minutes by applying the compression
■ Protecting ecosystems ■ Using renewable energy
resin transfer molding (C-RTM) method to mass
production. A rate of 2 minutes per car implies that the
Protecting the Protecting Controlling water Controlling water process can maintain a standard production speed of 30
jobs per hour (30 JPH), revolutionizing mass production
habitat of beetles and bamboo quality in the Iso quality in the
stag beetles forests River (east side of Munase River
the plant) (west side of the

for this material.

plant) V2X + reusable LEAF batteries Reborn light
+ solar panels
■ Thorough control of water quality ■ Landscaping inside and outside
the plant
◼ Simulation accuracy improved
◼ Development of parts that satisfy performance requirements by revising the shape of
Biomonitoring equipment
the gate and resin bead using simulation
Management of plant
wastewater by <Results of experiments and simulations> <Part development using simulation>
thorough water Greening along the Keeping the plant
treatment national road by clean
Rearing fireflies with treated industrial wastewater employees
Experimental result

Coexistence with the local environment Changed the

resin bead shape
Changed the ‘R’ shape
of the product corners

Resin bead
Simulation results
A Upper part

3.4 Mass-producible Weight Reduction Technologies Actual

part
A'
Lower part

Necessary for Electrification Non-impregnated area

While EVs are becoming increasingly used worldwide,

Before the After the
development development
Center pillar
the technologies required for their widespread use extend reinforcement Cross section A-A'

beyond EV powertrain technologies, such as motors,

inverters, battery materials, and other structures, to Electrification and vehicle weight
include critical mass-producible technologies for reducing
vehicle weight. EVs are becoming heavier with larger
batteries to ensure a longer range, demanding a drastic 4. ARIYA: A Flagship Vehicle from the
revision of the vehicle body structure and innovative Nissan Intelligent Factory
body weight reduction strategies.
It is common to use ultra-high-tensile materials for ARIYA features a design that embodies the new Nissan
parts that require structural strength and aluminum and Nissan Intelligent Mobility, equipped with a lineup
alloys for cover parts to reduce weight; however, the of four powertrains to meet a wide range of customer
extent to which other materials can be substituted for needs, comfortable and responsive driving performance,
steel is limited by material strength. The joining of a new generation of EV powertrains, state-of-the-art
aluminum alloys is conventionally performed with rivets driver assistance technologies, and advanced connection
(SPR), which increases the weight of the vehicle in technologies.
accordance with the number of rivets. To overcome this Nissan's EV innovation began with the first mass-
challenge, welding and innovative technologies are produced EV, LEAF," in 2010.
needed for joining different materials, such as aluminum
alloys, steel, and resin. LEAF ARIYA

◼ Further weight reduction is needed to meet the need for higher Innovation
fuel efficiency in engine vehicles. Various factors will increase
the weight of vehicle components due to electrification
Engine vehicles EV ⚫ Nissan's first mass-produced EV ⚫ Nissan's first crossover EV
Strengthening of steel
and structural Replace with (since 2010) ⚫ Range of up to 610 km (WLTC mode)
improvements aluminum
Suspension
brake.,etc
⚫ Range of up to 458 km (WLTC mode)
Replace with
LEAF ARIYA
Vehicle weight (kg)

(1) Compliance with multimaterial

enhanced crash test
Increase due to CFRP
Replace with
the shift to EV Intelligence Number of ECUs*: 36 2x Number of ECUs*: 78 *Electrical Control Unit
aluminum Battery
Number of CAN** signals: 3,600 2.5x Number of CAN** signals: 9,500 **Controller Area
(2) Tighter regulations on
exterior noise Network

Compliance Electrification Two types of batteries: 60/40 kWh 1.5x Two types of batteries: 91/66 kWh
with
performance Energy weight ratio Energy weight ratio
1.5x
improvement Engine T/M Inverter (Battery capacity/vehicle weight) (Battery capacity/vehicle weight)
regulations motor 32Wh/kg / 22Wh/kg 41Wh/kg / 34Wh/kg
EV-specific
components

Nissan EV innovation
Current model Next generation model

Electrification and vehicle weight

LEAF has been repeatedly improved during its model
life and now has a range of 458 km. In contrast, ARIYA
Carbon-fiber-reinforced plastics (CFRPs) have long has a range of up to 610 km and has been significantly
attracted attention for their potential to reduce the advanced in terms of electrification and intelligence with
weight of car bodies. Though several decades have passed twice the number of ECUs installed, 2.5 times the number
since CFRPs were first launched, the establishment of of signals used for communication and 1.5 times the
mass-production technology has been a perennial issue capacity of the two types of batteries as well as the
in terms of cost and cycle time. Furthermore, battery capacity per weight compared to those of LEAF.
breakthroughs in the production technology were The four pillars of the Nissan Intelligent Factory were

4 No.88 (2022) NISSAN TECHNICAL REVIEW

 Preface Nissan s DNA develops Super evolution a production system in harmony with Blue Earth

created to efficiently produce the complex and highly perspective of how their hard work and approaches have
advanced ARIYA. After many years of developing basic paved the way for the times. It is important to
and applied technologies, sophisticated product is now understand this history and pass on the lessons from
being realized. past visionaries to the future.
Although it takes time for a technology to mature,
developing the ability to understand its purpose will
enable individuals to ask what this technology is useful
for, to take action towards achieving future goals, to
1
Technologies to create
2
Advanced robots with
3
Coexistence of humans Zero-emission production
4 improve quality, and to promote steady evolution of the
the cars of the future expert craftsmanship and robots system tools and products of the industry. These efforts are
essential for creating the car of the future through
technological innovation alongside a production system
that is resilient in the face of change, instability, and new
global challenges. Furthermore, we hope that these
Promoting carbon neutrality

All new models to be launched in

major markets will be electric
vehicles from the early 2030s.
Innovations in battery technologies for the

efforts will contribute to the creation of a human-friendly

development of more competitive and efficient EVs
Development of a battery eco-system
Development of e-POWER with that will contribute to distributed
greater energy efficiency power generation by renewable
energy

Improving energy and resource use efficiency

through innovations in production technology

workplace where people can take pride in their work.

2050: Achieve life-cycle carbon neutrality for cars

Addressing CASE Highest quality mass production Comfortable workplace Addressing carbon neutrality

Nissan's automotive engineers are responsible for the

Four pillars of the Nissan Intelligent Factory future of Nissan, and their enthusiasm and efforts will
create new ideas that would shape the industry.
Nissan's automotive engineers have always taken
While there have been many challenges and successes action in the spirit of improving and advancing the
in recent years, an innovative automation line has now Japanese automotive industry. The products and new
been completed through the efforts of many Nissan technologies developed by Nissan's automotive engineers
automotive engineers. It was truly a moment of great joy have improved people's lives by enabling them to do
when the line began operating fully automatically, and things they could not or had given up doing before. This
this happiness and excitement was shared by everyone technical review will hopefully encourage individuals to
on-site for this momentous occasion. The Nissan take pride in your work, believe in their own abilities,
Intelligent Factory will be rolled out globally, with the innovate, and continue to change the industry by solving
Tochigi plant as the pilot model. the challenges of the future.

5. Summary
5.1 Future Challenges in Automobile Manufacturing
While the future remains uncertain, automotive
engineers and the broader Nissan team will undoubtedly
be able to make further breakthroughs by innovating
together. The following are some of the challenges in
automobile manufacturing that will need to be tackled in
the years to come:
- A new pandemic or friction between the US and
China that would wreak havoc on the global supply
chain
- Continued efforts are needed to avoid further weight
increases in response to electrification, with a
requirement for innovative car body weight reduction
strategies relating to the use of multi-materials,
joining of different materials, and innovative
structures, among others
- Market cost impact and increase in risk level due to
raw material procurement risk
- Intensifying competition for the acquisition of
employees with strong AI and systems skills due to
the emergence of new competitors amid a shortage of
new talent
- Worsening global hiring environments for automobile
production plants

5.2 A Message to Nissan's Automotive Engineers

Hopefully, all the Nissan engineers who are to
overcome the challenges of the future will revisit the
history and achievements of Nissan's past automotive
engineers. These individuals have worked passionately
and diligently to develop new technologies with the

NISSAN TECHNICAL REVIEW No.88 (2022) 5

Preface Nissan s DNA develops Super evolution a production system in harmony with Blue Earth

6 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1

Nissan Intelligent Factory Starting from ARIYA

1. Nissan Intelligent Factory Starting from ARIYA

Hidetoshi Okaguchi*

Mission Minimize labor intensity in manufacturing Mission Reform factory working environments
1 2
1. Trends in the Automotive Industry Men Women

The automotive industry is experiencing a once-in-a-

century revolution with rapid changes in vehicle 1913: Ford Model T enters production Today
Declining birthrate and
aging population

technology related to the connected, autonomous, shared/

Source:

Mission Mission
Increase flexibility in responding to supply Address climate change through concrete actions
service, and electric (CASE) areas. These changes have 3 chain disruptions 4

led to the increasing intelligence and electrification of Promote carbon neutrality

cars, resulting in more advanced vehicles. Thus, it is now

All new models to be launched in
major markets will be electric
vehicles from the early 2030s.
Innovations in battery technologies for the development of more
competitive and efficient EVs
Development of a battery eco-
Development of e-POWER with system that will contribute to

necessary to manufacture advanced vehicles with the

greater energy efficiency distributed power generation by
renewable energy

Improving energy and resource use

efficiency through innovations in production

Pandemics and natural disasters

technology

Respond to environmental concerns

same quality standards and cost-effectiveness as those of

conventional vehicle manufacturing.
Fig. 1 Trends and priorities in the automotive manufacturing
In addition to the advances in automotive technologies, industry
the automotive industry is expected to shift in response
to a shrinking workforce resulting from declining
birthrates and an aging population. This demographic 2. Future Direction of Nissan's Production
shift will cause the current manufacturing system, which Plants
relies on gathering many workers in the same place at
one time, to become unsustainable. In conjunction with Nissan Motor's production plants have been
changes in worker attitudes and critical job skills, these modernizing and modifying their production lines under
workforce changes will necessitate the creation of the Nissan Production Way/Alliance Production Way
workplaces where more individuals can work as well as a while globalization has accelerated. Since establishment,
shift away from labor-intensive automobile production Nissan has taken advantage of their technology to
systems. overcome new challenges. This approach led Nissan to
Finally, the automotive industry and other industries address the environmental impact of automobiles by
are facing several major new challenges. With the becoming the first company to mass-produce electric
increasing frequency of pandemics, natural disasters, vehicles. The efforts to address industry challenges with
and problems with semiconductor supply, many new technology have accelerated, leading to the
industries must become more flexible in response to introduction of ARIYA, a flagship Nissan Intelligent
unexpected situations. Furthermore, there is also a Mobility product, at Nissan's production plant.
growing sense of international urgency regarding climate Flexibility and responsiveness to accelerating industry
change; thus, there is an urgent need to implement changes are critical for Nissan s ongoing evolution. With
concrete measures to achieve carbon neutrality in the the advancement and diversification of both hardware
automotive industry (Fig. 1). and software in vehicles alongside the acceleration of
intelligence and electrification (Fig. 2), the mainstream
vehicle manufacturing process must shift from that of
internal combustion engine vehicles to that of electric
vehicles. The processes of mounting many components
and checking functions that have never been needed
before are also necessary for connecting components for
autonomous vehicle functions.

*Corporate Vice President **Nissan Product Development Department No.1

NISSAN TECHNICAL REVIEW No.88 (2022) 7

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 1. Nissan Intelligent Factory Starting from ARIYA

Eliminating roof molding to

reduce air resistance and
ensure plant competitiveness and the ability to profitably
create Nissan Intelligent Mobility products for decades.
improve the appearance
Superbly quiet and
lounge-like interior space

3. Four Pillars of the Nissan Intelligent

Highly intelligent Luxurious
Factory
electrical control unit appearance

Four options for The four pillars of the Nissan Intelligent Factory were
established based on the background presented in
powertrain
(Two types of batteries x
two types of drive systems)

A variety of color
Sections 1 and 2 to define the challenges and objectives to
combinations
be addressed by the factory (Fig. 3). The Nissan Intelligent
Factory consists of the latest technology for each process
Fig. 2 Features of the ARIYA, a Nissan Intelligent Mobility
flagship product based on these concepts. The core production technology
is supplemented by cross-sectional systems necessary for
maintenance and quality control
However, production plants that rely on human power
will not be attractive to workers in a shrinking workforce.
Thus, there is a need to eliminate jobs that tend to be
avoided, such as those requiring physical effort or
repetition of simple tasks, and to provide an attractive
1
Technologies to create
2
Advanced robots with
3
Coexistence of humans Zero-emission production
4
workplace environment. Notably, moving away from
the cars of the future expert craftsmanship and robots system

labor-intensive work will require workers to have skills

related to the use of advanced technologies, which will
create new opportunities and challenges for employees. Promote carbon neutrality

Furthermore, unexpected events such as pandemics,

All new models to be launched in
major markets will be electric
vehicles from the early 2030s.
Innovations in battery technologies for the
development of more competitive and efficient EVs
Development of a battery eco-system
Development of e-POWER with that will contribute to distributed
greater energy efficiency power generation by renewable
energy

natural disasters, and supply chain issues require

Improving energy and resource use efficiency
through innovations in production technology

2050: Achieve life-cycle carbon neutrality for cars

Addressing CASE Highest quality mass production Comfortable workplace Respond to carbon neutrality

production sites to respond to temporary changes in

production volume and to rearrange the workforce in Fig. 3 The four pillars of the Nissan Intelligent Factory
order to compensate for variable worker availability.
This poses an extra burden for both workers and the
factory that is unnecessary for normal production, such 3.1 Technologies to Create the Cars of the Future
as having multiple workers learn the same task for In recent years, production plants have been required
redundancy and ensuring that workers have developed to adapt in order to keep up with the advancement and
necessary skills from their training to allow for electrification of vehicles owing to accelerating changes
redeployment if needed. in the CASE areas. The car of the future, that is, a car
On a global scale, initiatives are accelerating with a wide variety of specifications to meet the needs of
internationally to protect the environment, demanding the growing diversity of users who desire advanced
that corporations implement more environmentally technologies in their products, is represented by Nissan
friendly initiatives and practices than in the past. Intelligent Mobility. For example, it is important to have
Consumer sentiment is also changing to expect companies a variety of powertrains to meet the diverse needs of
to proactively address environmental concerns and customers; thus, the ARIYA has four powertrain
portray a green image through their practices and alternatives that allow customers to select between two
products. types of batteries and two types of drive systems.
Given the variety of challenges facing the automotive Determining the common and customizable elements of
industry, it is important for the Nissan Intelligent the product structure and designing the product, process,
Factory to have the flexibility to respond to changes in and equipment to minimize the changes required for
the workforce and customer sentiments and to enhance customizable elements enhance the ability to easily adapt
the resilience of the manufacturing process against a to current and future production changes.
variety of challenges. This must be achieved without ARIYA is equipped with 78 electrical control units
excessively decreasing profits by minimizing costs that (ECUs) that provide the intelligence control for the car,
would not have been necessary in normal production which is more than twice that of the conventional Nissan
processes. It is also critical to ensure that manufacturing electric vehicle, LEAF. This advancement required the
quality is not affected by changes made to the production development of systems for high-speed data
process. Viewing flexibility and resilience against communication and expansion of the data-writing
challenges as key features in evolving manufacturing, process, which can be manufactured at the production
the Nissan Intelligent Factory introduced equipment and plant.
machinery that have been established through the The diversification of available cars leads to the
development of basic and applied technologies. These diversification of potential incidents in the process of
digital technologies were then connected as a single production, shipping, and quality assurance. The
production system operated by highly skilled workers. visualization of product quality and equipment
The Nissan Intelligent Factory is an initiative that will abnormality data using digital technologies to accurately

8 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 1. Nissan Intelligent Factory Starting from ARIYA

identify these complicated incidents, alongside the quality is difficult to assess automatically, and performing
development of a system that provides workers with data specialized work regarded as challenging for machines.
on all manufacturing equipment, has contributed to This combination of robotics and skilled craftsmanship
seamless production processes. To this end, the following leads to the production of vehicles of the highest quality.
technologies and processes have been integrated into the The work of expert craftsmen (Takumi) was repeatedly
Nissan Intelligent Factory: analyzed to quantify the domains of sensing, information
- Batch mounting system for the powertrain (Fig. 4) processing, and actuation. In addition to quantification,
- Digital technologies to improve worker learning the process and outcomes of intuitive decisions made by
- New joining process for dimple welding craftsmen using their vast experience were translated
- Establishment of electric systems for automobile into logic. This strategy, in conjunction with development
intelligence of basic and applied technologies that combine state-of-
- Automation of the magnet-free field motor winding the-art high-performance sensors, cameras, and
(Fig. 5) image-processing technologies, made it possible for robots
- Automatic inspection of paint job appearance to demonstrate expert craftsmanship in assembly and
- Integrated automatic inspection of specifications quality inspection work.
and scratches Robots can perform precise, repetitive operations and
- Assured quality control using a quality gate work at a consistent level to achieve quality beyond that
- Equipment failure diagnostic system and tools for of craftsmen in certain areas. Furthermore, if an
predictive and preventive maintenance of equipment unexpected event causes a temporary change in
production volume, robots can quickly start and stop
again without incurring extra costs, which increases
their resilience in the face of production changes or
disruptions. To this end, technologies for the following
processes have been integrated into the Nissan Intelligent
Factory:
- Automatic tightening and alignment of suspension
links (Fig. 6)
- Automatic headlining assembly
- Automatic cockpit module assembly
- Automatic inspection of painting appearance
- Integrated automatic inspection of specifications
Fig. 4 Batch mounting system for powertrain and scratches (Fig. 7)

Fig. 6 Automatic tightening and alignment of suspension links

Fig. 5 Automation of magnet-free field motor winding

3.2 Advanced Robots with Expert Craftsmanship

Expert craftsmanship refers to work that makes full
use of visual, tactile, and auditory perception, based on
the accumulated experience and training of a skilled
worker. This is an essential element for the production of
high-quality cars. Passing on the skills of expert
craftsmen to robotic operations and machine control
circuits will allow workers to focus only on improving
work sites where their skills offer greater added value,
Fig. 7 Integrated automatic inspection of specifications and
conducting inspections of components where perceived scratches

NISSAN TECHNICAL REVIEW No.88 (2022) 9

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 1. Nissan Intelligent Factory Starting from ARIYA

3.3 Coexistence of Humans and Robots adopts innovative technologies, improving the energy
Physical and mental stress on workers has been efficiency and reducing the material footprint through
reduced by partially automating tasks that are physically recycling.
demanding, such as heavy lifting, upward and bending For example, a water-based paint that cures at low
postures, and repetitive work; processes that are stressful temperatures was developed for the new painting process,
but production-critical; and operations that require making it possible to lower the baking temperature from
considerable memory. The Nissan Intelligent Factory 140 C to 85 C. Consequently, energy use was reduced by
environment has been established to reduce the burden 25%, representing a significant contribution to carbon
on workers by capitalizing on the advantages of neutrality. The following processes and technologies are
automated processing through digitization, data related to the goal of achieving carbon neutrality at the
collection, and communication. However, tasks that Nissan Intelligent Factory:
require manual calibration each time will continue to be - Integrated painting of body and bumper (Fig. 9)
carried out by workers as robots are poorly suited to - Dry booth for painting
these tasks. Thus, a work environment where humans In-house power generation using alternative fuels.
and robots can coexist has been achieved by optimizing
the assignment of tasks to integrate work for which
robots are well suited with specialized or rewarding tasks
that are performed manually. This combination allows
both humans and robots to perform at their best in their
respective areas of expertise and creates a more inclusive
and encouraging working environment for employees.
This integration is well represented in the following
processes:
- Batch mounting system for the powertrain
Automatic tightening and alignment of suspension
links
Automatic headlining assembly (Fig. 8) Fig. 9 Integrated painting of body and bumper

4. Nissan Intelligent Factory Value

Proposition

The creation of production lines compatible with the

electrification, intelligence, and diversification of
automobiles through the integration of advanced robotics
with human workers offers the opportunity to create a
more inclusive working environment that produces high-
quality modern vehicles. Importantly, these production
lines are also more adaptable and resilient to the various
Fig. 8 Automatic headlining assembly challenges facing the automotive industry. For example,
even when unexpected events such as pandemics or
semiconductor supply shortages occur, highly automated
3.4 Zero-emission Production System processes can be promptly stopped and restarted flexibly
Nissan Motor Company has been actively engaged in to meet the demand without incurring extra costs. This
the company-wide adoption of wind and solar power, will also contribute to increased business profits by
cogeneration, and other initiatives as part of the Nissan reducing the cost of production through the labor-saving
Green Program. effects of automation and digitalization, as well as the
To achieve carbon neutrality across the entire life cycle reduction of energy costs through changes in
of cars including corporate business activities by 2050, manufacturing processes such as integrated body and
the company will further increase the use of wind and bumper painting.
solar power generation and electrify the power sources Importantly, the implementation of the Nissan
for plant equipment by 2030. The company will also work Intelligent Factory will make a significant contribution
on the adoption of alternative fuels and power to achieving carbon neutrality in the automotive industry
generation such as bioethanol, hydrogen, artificial and will allow plants to evolve in parallel with the
methane, and solid oxide fuel cells (SOFCs) and is ongoing evolution of vehicles and the automotive industry
planning to start a demonstration experiment at its more broadly. Together, the advantages of the Nissan
Tochigi plant regarding the use of SOFCs in its production Intelligent Factory will enable automakers to continue
facilities in 2022. their business and evolve to meet customer demands for
The Nissan Intelligent Factory will contribute to upcoming decades.
carbon neutrality by electrifying plant equipment as it

10 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 1. Nissan Intelligent Factory Starting from ARIYA

5. Summary

The Nissan Intelligent Factory was first implemented

at the Tochigi Plant, where the ARIYA, a flagship Nissan
Intelligent Mobility product, is being produced. The
Nissan Intelligent Factory concept announced in 2019
was presented at the Tochigi Plant in October 2021 for
internal and external individuals for the company,
revealing the actual product. This presentation offered a
unique opportunity to communicate the value of the
Nissan Intelligent Factory, which will contribute to the
enhancement of the Nissan Motors brand and customer
confidence. The development of manufacturing technology
will continue in earnest to keep pace with the rapid
evolution of automobile technology and will be
implemented in both Japanese and overseas plants as
appropriate, in line with facility renewals and the launch
of new vehicle models.

Fig. 10 Concept announcement in 2019

Nissan Intelligent Factory

Expands Globally with the Tochigi Plant
as the Pilot Model

1
Technologies to
2
Advanced robots
3
Coexistence of
4
Zero-emission
create the cars of the with expert humans and robots production system
future craftsmanship

Fig. 11 Presentation at the Tochigi Plant in 2021

NISSAN TECHNICAL REVIEW No.88 (2022) 11

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 1. Nissan Intelligent Factory Starting from ARIYA

Authors

Hidetoshi Okaguchi

12 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1

Nissan Intelligent Factory Starting from ARIYA

2. Application of Next-Generation Underfloor Process

Accommodating Various Powertrains

Hiromasa Tsuchikane* Atsushi Ugaji** Yoshihisa Naito***

2.1 Flexible production process

1. Introduction To realize a decarbonized society, the prevalence of
electrified vehicles, which have batteries and motors,
As global activities for carbon neutrality have been such as EVs and e-Power, must increase rapidly (Fig.2).
accelerating, Nissan Motor has been aiming to realize Convectional vehicles, which have internal combustion
carbon neutrality throughout life cycles, such as by engines (ICEs), will also be produced for some time.
accelerating the implementation of electric vehicles (EVs,
e-Power) at an unprecedented level.
Meanwhile, production sites are experiencing
prospective reductions in the numbers of highly skilled
technicians and workers as the birthrate declines and the
population ages in Japan, Europe, and Asia. Therefore, it
is urgently necessary for vehicle assembly plants, as
labor-intensive facilities, to take measures against this
issue (Fig.1).

(ten thousand people)

Fig.2 Estimation of electrified new car sales

Thus, production sites are required to operate existing

production lines efficiently to produce various old and
new powertrains, including EVs, e-Power, and ICEs.
https://www.bcg.com/ja-jp/publications/2021/why-evs-need-to-accelerate-their-
market-penetration
However, the arrangements of parts and structures are
widely different between EVs and conventional vehicles
(Table.1). Thus, various models also have different
processes and operations; accordingly, mixed production (year)

in an existing production line does not enable workers to

75 y/o or older 65 ~ 74 y/o 15 ~ 64 y/o 0 ~ 14 y/o
Rate of 65 y/o or older (right axis) Rate of 75 y/o or older (right axis)

Fig.1 Transition of population by age and population aging rate perform the same operations in the same process for
Sources: Statistics Bureau, Ministry of Internal Affairs and Communications, “Census” (in and before 2010); Statistics Bureau, Ministry of Internal Affairs
and Communications, “Population Estimates” as of October 1, 2013 (in 2013); National Institute of Population and Social Security Research, The Results of
Projection According to the Medium-Fertility Assumption and Medium-Mortality Assumption from “Population Projections for Japan (estimated in January
2012)” (in and after 2015)
cyclic operations, resulting in considerable losses of
(Note) The total number from 1970 to 2010 includes uncertain ages. For the calculation of population aging rate, uncertain ages were excluded from the

quality and working time.

denominator.

As described above, the social environment is currently

changing quite rapidly, and vehicles are also changing
https://www.mhlw.go.jp/wp/hakusyo/kousei/14/backdata/1-2-1-07.html Table.1 Comparison of structural differences between ICEs and EVs
rapidly with electrification. To respond to this evolution
flexibly, the Nissan Intelligent Factory (NIF) at Tochigi
Plant applied a next-generation underfloor process, as
introduced here by Production Engineering.

2. Process required for next-generation

production site

If vehicle electrification, the declining birthrate, and

the aging population mentioned above are considered,
automotive production sites are required to take two
main actions.

*Nissan Product Development Department No.1

NISSAN TECHNICAL REVIEW No.88 (2022) 13

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 2. Application of Next-Generation Underfloor Process Accommodating Various Powertrains

Particularly for EVs, further technological development 3.1 Accommodation to various powertrains
and evolution are expected in the future, and the trend of Previously, in multiple processes, dedicated lifters
EV structures is fluid; consequently, production sites are have been used to install powertrain parts, such as
required to handle such electrified vehicle structures engines, transmissions, and large EV batteries in the
flexibly. Therefore, in this study, the objective was to underfloor areas of vehicles. (Fig. 4 Before). In addition,
achieve efficient and flexible production of various supports for parts fitted on lifters had to be replaced
powertrains by utilizing an existing production line and manually according to the part specifications. Mixed
applying a new method. production of ICEs and EVs, which have different
components, in such processes caused their work times to
2.2 Production process accessible to everyone differ because of the differences between the operation
At present, in vehicle assembly plants, processes for items (tooling change/mounting operations) in those
assembling the underfloor parts of vehicles include many processes. Then, workers were distributed according to
operations performed by highly skilled technicians, such the most time-consuming case, resulting in a loss of
as heavy work and elevating work (Fig.3). waiting in some models.
This time, the process design was revised: the
operations affected by differences in the powertrain
specifications were moved to sub-lines, the combination
of the operations could absorb time differences, and a
newly developed pallet system was applied so that the
operations of mounting parts assembled in the sub-lines
of the vehicle could be integrated into a single process to
achieve batch-mounting (Fig.4 After).

Part 1

Sub-line 1
Part 2
Part 3
Part 1 Part 3 Part 5

Assembly main line Assembly main line

Fig.3 Hard elevating work

Part 2 Part 4 Part 6 Part 6

Part 5
In addition, the globally declining birthrate and aging
Part 4
population will cause the number of workers to decrease. Before After
In order to retain workers, it is necessary to reduce the
high-impact operations that workers must perform, such
as heavy work and elevating work, through appropriate
planning processes that make these tasks accessible to
everyone. Before
Vehicle electrification will also lead to large and Batch-mounted on the main Assembled through
heavier batteries, which will increase the number of line the sub-line

tightening points (Table.1). Accordingly, underfloor Powertrain unit mounting

ICEs

operations, where many elevating actions are

concentrated, will impose more burdens on workers than e-Power

ever, requiring further measures to be taken to improve

the work environment. Consequently, in this study, the
objective was to realize a production process that is
EVs

accessible to everyone by reducing overburdened Lifter Six-axis robot

operations for workers. After
To achieve these two next-generation processes, the
Fig.4 Before and after introducing the underfloor process
NIF at Tochigi applied two systems: the Underfloor
parts batch-mounting system and Suspension link
automatic tightening system, which are introduced in In addition, full automation of batch mounting was
Chapters 3 and 4, respectively. achieved by measuring the dimensional tolerances of
each vehicle between the vehicle body and underfloor
3. Underfloor part batch-mounting system parts to develop a logic for automated positioning.

As described in the preceding chapter, to realize a 3.1.1 Pallet system

production process that is flexible and accessible to The conventional mounting process employs a method
everyone, an underfloor parts batch-mounting system to integrate supports for parts with lifters. Meanwhile,
was introduced, as reported in this chapter. an upper/lower two-layer pallet system was developed in
this study (Fig.5).

14 No.88 (2022) NISSAN TECHNICAL REVIEW

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 2. Application of Next-Generation Underfloor Process Accommodating Various Powertrains

The lower layer has a fixed shape to function as a Specifically, rear shock absorbers are installed on a
conveyer, and the upper layer has a structure vehicle body in the preceding process (Fig.7 [1]). Following
exchangeable according to the vehicle model and this, a vision system is used to measure the assembly
powertrain as a section touching the vehicle body and accuracy of the rear suspension on the pallet in a sub-line
parts. (Fig.7 [2]).
The vision is also utilized to measure the vehicle
Upper layer position (Fig.7 [3]), the measurement error effect is
dedicated to each model minimized by photographing multiple points, the two-
axis robots on the lifter are used to adjust the position of
Front
the rear pallet, and the rear suspension is moved to a
Center
fitting position (Fig.7 [4]).
Rear The six-axis arm robots hold the rear shock absorbers
and correct them to the positions of the rear suspension
link arms based on the measured assembly accuracy of
the rear suspension (Fig.7 [5]). The lifter moves upward
to complete the connection (Fig.7 [6]).
Lower layer
common to all models
[1] Shock absorber
Fig.5 Two-layer pallet installation
(Main line)

An ordinary pallet system has complicated mechanical

adjustment functions to absorb the effects of various [2] Link arm position
measurement
accuracy variations of the part supports and parts to be (Sub-line)
mounted. The functions must be maintained and
accurately controlled. In this study, a two-axis robot-type
positioning function was provided for the lifter. A method
[3] Vehicle body position
of absorbing the part accuracy variations was adopted, measurement
where each upper pallet was aligned with a conveyed
vehicle body to simplify the pallet accuracy control.
A system to feed bolts to a tightening robot directly Two-axis robots
[4] Pallet aligned with
without a pallet as an intermediate was developed to vehicle body by two-axis
eliminate bolt jigs on the pallet that were previously robots
necessary for feeding bolts to a tightening unit and
reducing the pallet cost.
[5] Shock absorber aligned
3.1.2 Positioning logic with link arm with six-
axis robot
The rear shock absorbers installed on a vehicle body
and rear suspension link arms mounted on the rear pallet
cannot be aligned and connected without any adjustment
by lifting the latter (Fig.6). Therefore, the position data of
[6] Connecting
a vehicle and the position data of the rear suspension link
arms were used to correct both positions, and six-axis
robots were employed to position the shock absorbers to
Fig.7 Rear shock absorber alignment process
achieve alignment.

3.2 Automation of tightening process

In the conventional underfloor process, workers elevate
their work for assembly. The parts handled in the
underfloor process are important and require reliable
tightening; therefore, large and heavy tightening tools
were applied. Because bolts differ in size and tightening
torque depending on their tightening points, tools are
exchanged frequently, resulting in overburdened work.
To realize a production process that is accessible to
everyone, as mentioned in Section 2 of this chapter, it is
Fig.6 Alignment of rear shock absorbers necessary to resolve these issues. This process is
introduced in the following subsections.

NISSAN TECHNICAL REVIEW No.88 (2022) 15

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 2. Application of Next-Generation Underfloor Process Accommodating Various Powertrains

3.2.1 Realization of in-line automatic tightening that pass through those parts (Fig.9).
To eliminate high-load tightening operations in the
underfloor process, tightening automation is promoted. Axle module (in dashed-line area)
Rounded-tip bolts, which are robust against misalignment Motor unit
Brake caliper
and suitable for auto-tightening, were applied. Positioning
mechanisms advantageous for auto-tightening were also
incorporated into the vehicle parts. In addition, it is
necessary to reverse the bolt direction because downward-
pointing bolts are conveyed and supplied to the robots
Hub Suspension
because of the structure of the bolt feeder, and they are member
tightened upward. Accordingly, a unique bolt-reversing Brake rotor Links
system was developed for automatic feeding to reverse
Link bolt tightening point
each bolt and align them in the tightening direction of
the vehicle body. A flexible socket system was employed Fig.9 Suspension unit link bolt tightening mechanism
to change the socket sizes with a single action to tighten
different bolts, from M8 to M14.
To place a bolt in an underfloor tightening point, In this link bolt tightening process, similar to an
approaching through a narrow gap between obstacles underfloor process, workers had to grab and use different,
such as a pallet and conveyor for the pallet, a tool more heavy, high-torque tools for different tightening positions
than 1.4 m long was utilized, and a vision system for of various models within a single process in the past
measuring and correcting the position of a tightening (Fig.10).
point that ensures accuracy even over a long range was
employed. The utilization of these items enabled
elimination of the effects of variations in the on-pallet
part, mounting, conveyance equipment production, and
stoppage accuracies and ensured a bolt position correction
accuracy of +/- 0.5 mm, leading to the realization of full
tightening automation (Fig.8).

One-touch socket

Reversed
bolt feed

Vision camera
Fig.10 Different tightening tools to be used for different points

Long tool
Furthermore, as tightening a bolt-nut system requires
placing a tightening tool on one side and an anti-spin
Fig.8 Bolt feeding system and automatic tightening robot wrench on the other side, workers had to perform a two-
handed operation in which it was required to place a
heavy tool in a tightening point with the non-dominant
hand.
4. Suspension link automatic tightening To automate such physically overburdened two-
system handed link bolt tightening operations, this equipment
applies a new concept regarding the combination of
As mentioned above, the issues existing in the robots and tightening tools.
underfloor operations of the main line are also present in
the sub-lines, so similar activities are performed. As an 4.2 Tightening action by two opposite robots
example, this chapter introduces a system to tighten link The automatic tightening system consisting of robots
bolts in a multi-link suspension unit automatically. and tightening tools has three main issues.
The first issue is the variation of the bolt position
4.1 Link bolt tightening mechanism caused by the aggregated dimensional accuracies of a
A suspension unit consists of a suspension member conveying trolley, jig, and workpiece. This variation does
with a motor unit mounted, axle modules with hubs on not allow a robot to play back a taught position or to
which tires rotate, brake rotors, brake clippers, and links establish the alignment of a bolt and tightening tool, so
that connect the suspension member and axle modules. that the bolt and nut cannot be rotated.
The links and axle modules as well as the links and The second issue is that the different shapes of the link
suspension member are tightened by bolt-nut systems bolt tightening sections include peripheral parts, as

16 No.88 (2022) NISSAN TECHNICAL REVIEW

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 2. Application of Next-Generation Underfloor Process Accommodating Various Powertrains

shown in Fig. 9, requiring a tightening tool to approach remain less than half the number of tightening points
the required point according to shape and placement. was realized.
The third issue is that the use of a dedicated tool whose
tightening-reaction support and anti-spin part are
designed according to each shape causes the number of
tightening tools and robots to increase in proportion to
the number of tightening points. In this case, multiple
points cannot be tightened from a single location at which
a worker can handle the operations; therefore, a larger
equipment installation area would be required.
To resolve these issues, the proposed system employs a
couple of robots with a vision system and tool-changing
mechanisms and placed on both sides of a workpiece to
Wrench Tightening tool
allow them to reproduce worker operations of using both Fig.12 Bolt and nut tightened by two opposite robots
hands to handle general-purpose tools and tighten
multiple points (Fig.11).
In addition to the resolution of the issues, the work
method of a highly skilled technician was incorporated
Left-side robot into the robot control to achieve an operating speed equal
Workpiece to that of the worker and the robustness of reliable tool
placement.
In bolt-nut opposite tightening, a tool on the tightening
side rotates a socket and presses it into a nut
simultaneously, which allows the hexagons of the nut
and the socket to align with each other and the nut to
enter the socket. On the other hand, on the anti-spin side,
Right-side robot simply pressing a wrench against a bolt does not enable
Conveying it to fit the bolt.
trolley/AGV
To resolve the above issue, focusing on the worker
Fig.11 Link bolt tightening with a couple of robots actions performed to fit a wrench into the hexagon of a
bolt while turning the wrench on the bolt axis, it was
proposed to make a robot reproduce this action. However,
Firstly, to handle variations in the bolt position, a if a multi-joint robot reproduces this action, the robot
vision system developed for NIF was applied similarly to must continually change its entire posture to hold and
the underfloor process, introducing a 3-D position- move the wrench. Because the movable range of a robot is
correcting system composed of low-cost general-purpose limited in link bolt tightening and the approached spot is
equipment. One of the robots has a camera installed, small, moving a wrench while changing posture is
which detects the position of the object workpiece. The difficult.
detected position data are shared with the other robot so Consequently, a wrench-holding hand whose rotational
that the two can correct their motions based on the same axis (J6 axis) is aligned with the center of the wrench was
data and handle bolt misalignment. designed (Fig.13). This structure enables a robot to turn
Secondly, to adapt to the different shapes of each a wrench without changing its posture and to place the
tightening point and to achieve the function in a minimum wrench on the bolt quickly and without interfering with
area, a system employing two robots to operate generally the surroundings.
shaped tools independently for tightening was developed.
Each of these robots has one anti-spin wrench and one
tightening tool (nut runner) to turn a nut on the
tightening side and to control the tightening torque
installed. The two robots operate independently to J6 rotational axis
tighten the tightening and anti-spin sides at opposite
positions. Because the tightening tools are not dedicated
to models or tightening points but have simple general Center of wrench
shapes, the same tools can approach multiple points for
tightening (Fig.12).
In addition, connecting a robot and tightening tool via
a tool-changing mechanism enables one tool to be Fig.13 Wrench hand
exchanged for another tool according to the tightening
points, which permits four robots and four tools to tighten
10 tightening points in total. In this way, compact As mentioned above, this system was realized by
equipment in which the numbers of robots and tools combining integrating general-purpose robots, low-cost

NISSAN TECHNICAL REVIEW No.88 (2022) 17

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 2. Application of Next-Generation Underfloor Process Accommodating Various Powertrains

vision systems, and simple tightening tools with motions

reproducing those of workers.

5. Conclusion

Two production systems were applied in the NIF at

Tochigi, realizing a next-generation underfloor process
that accommodates various powertrains and suspension
units flexibly. In addition, the equipment has separately
designed changeable and fixed sections and employs
software-controlled robots, enabling hardware
modification to be minimized and helping reduce facility
investments and the time required to prepare for
production of a newly introduced model.
In addition, heavy and elevating work, which is
difficult for workers, could be reduced by approximately
15%. Furthermore, the tightening quality ensured by the
equipment reduces the physical and mental burdens on
workers in important safety-related processes, realizing
a production process that is accessible to everyone.
The technology currently applied in the NIF should be
developed further to respond rapidly and flexibly to
continuously progressing technological innovation and
globally and continuously changing social environments.
The subsequent challenge will be to distribute the results
to the Production Engineering Departments at bases
worldwide.

6. References
1) Annual Health, Labour and Welfare Report 2014,
website of Ministry of Health, Labour and Welfare
2) Why Electric Cars Can t Come Fast Enough , Boston
Consulting Group, Inc. (April 2021)

Authors

Hiromasa Tsuchikane Atsushi Ugaji Yoshihisa Naito

18 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1

Nissan Intelligent Factory Starting from ARIYA

3. Manufacturing Innovation in a New Motor Line through

Digitalization and Internet of Things (IoT)

Satoru Iizuka*

the overall equipment effectiveness (OEE), which is an

1. Introduction index of production efficiency, to 95%. The new motor line
at Tochigi has introduced many digital systems to achieve
The term digital transformation (DX) is being these targets by exploiting digital technologies (Fig. 1).
recognized by individuals as well as production plants. This chapter introduces issues that need to be resolved
During the recent COVID-19 calamity, whereas in these systems and provides outlines of the systems.
manufacturing staff have been required to work at
plants, technical staff have been steadily working at 2. Current situation and issues in application
home. Therefore, it is essential to know the production of digital technology
status of plants in real time and to support production,
even in a telecommuting environment. To realize these Various issues have been identified to achieve the
objectives, the utilization of digital technologies, such as aforementioned targets, many of which are expected to be
the Internet of Things, is essential. The powertrain resolved by digital technology. This chapter introduces
production engineering division, to which the present two major issues.
writer belongs, is promoting activities to eliminate any
quality defects, stabilizing quality (not only retaining 2.1 Stabilization of product quality at high level
product quality and production conditions within the Various manufacturers seem to concentrate on the
control ranges, but also targeting the centers of their quality control of products, and Nissan also recognizes
standard ranges (the nominal values)), and increasing this aspect as the most important issue. The Quality

Digital technology introduction into new motor assembly line at Tochigi Plant

Real-time visualization of Real-time visualization of

Drive recorder for quality control information defect information Powertrain Production
equipment Engineering Division

Reinforcement of
network security

Digitalization of production
management data

Smart watch notice of

equipment abnormality

Operation recording Production L/T (stock)

camera visualization
MR work assist tool

Fig. 1 Digital systems of new motor assembly line

Alliance Internal

*Powertrain Planning Department

NISSAN TECHNICAL REVIEW No.88 (2022) 19

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 3. Manufacturing Innovation in a New Motor Line through Digitalization and Internet of Things (IoT)

Assurance Depa a leader, the Production Engineering equipment and utilizes Excel to analyze or visualize
Department, and the Manufacturing Department changing points. Therefore, action is taken only after a
cooperate in seeking to ensure high quality. Whereas the potential defect occurs. To further reduce quality defects
lineup of products sold by Nissan is shifting from internal and the accompanying errors, it is desirable to detect
combustion engine vehicles to e-power hybrid electric them before they occur.
vehicles and electric vehicles, quality control items to be Accordingly, the new line introduced a system that
controlled in new model projects are increasing year by collects and visualizes production equipment information
year. Manufacturing sites perform quality trend control in real time. The quality measurement results, as well as
through daily quick response quality control. However, the production conditions of the production equipment,
hundreds of items must be checked manually, causing which are regarded as factors, are visualized in real time.
delays in noticing trend changes and potentially delaying In addition, if a set threshold is exceeded, a notification is
the implementation of appropriate countermeasures. In sent to a person in charge to enable this person to notice
order to maintain and improve quality, it is important to the condition before a quality defect occurs, accelerating
notice any signs of an impending decrease in quality the feedback cycles of quality improvement activities.
before the quality trend worsens and to feed this The technical staff support the site remotely based on
information back to the production conditions of the that information, even in a telecommuting state.
equipment as soon as possible. How the current daily Such systems have been previously applied in some
checks can enable close to real-time control step by step, cases, many of which are based on packaged software. If
such as in hours or minutes, is an issue. the standard functions of a package are insufficient,
customization is usually performed. Nissan often
2.2 Highly efficient production outsources such customization from an external
Another issue at manufacturing sites is improving cooperative company as a form of entrusted development.
production efficiency. If production equipment has a An issue in this case is the lead time until software
short stop, its recovery work is left to the production site, development is completed. For outsourcing, a waterfall
and the recovery period varies depending on the members method is often applied as a development method from
and groups at the production site. For highly efficient the perspective of deliverable assurance, and at least
production, recovery period variations must be minimized. several months can elapse before delivery of the
In the current situation, however, the actual ability of requirements, specification decisions, system
each group cannot be determined quantitatively and in a development, and testing.
timely manner. In the complicated operations of However, manufacturing sites can benefit from a much
manufacturing sites, reducing the burden on supervisors faster process. If a line has a new potential defect, it is
and workers to organize the environment for efficient desired to install a necessary sensor immediately that
production is an issue. can produce a graph on the next day, even if the graph is
Another issue is implementing more efficient worker not implemented in the system. The conventional system
training performed by staff at manufacturing sites. development method cannot handle this task, so that a
Particularly in new model projects, considerable amounts USB flash drive is used to extract data and Excel is used
of time are allowed for training to improve work quality. to draw a graph. When such a system is finished, quality
Further efficiency improvement using only on-the-job effect factors will have been identified and equipment
training and/or educational content in the form of measures will have already been taken, rendering the
conventional videos and paper materials is difficult, so system meaningless.
the development of a system that enables workers to For outsourcing, because an agile method, which does
perform self-learning easily is an issue. not assure deliverables, is difficult to apply, the new line
introduced a system for developing in-house visualization
3. Introduction of systems applied to new by designating a development team in-house. As a base
motor line system, software including collection, accumulation, and
data visualization (graphing, etc.) functions was selected.
This chapter introduces systems applied to the new In-house staff can perform customization using this
motor line at Tochigi as countermeasures against the approach, such as creating a screen and changing the
issues mentioned above. data flow. An agile method is applied for development,
where the Plan-Do-Check-Action is cycled in a small unit.
3.1 Real-time data visualization and internally In the case in which a graph should be shown as soon as
produced agile development possible, similar to the occurrence of a new potential
As mentioned in Subsection 2.1 of this chapter, in product defect, the first edition with minimum necessary
order to better stabilize product quality, it is necessary to functions is released rapidly and a notification and other
notice any signs of an impending decrease in quality functions are added later, enabling the manufacturing
before the quality trend worsens and to feed it back to the sites to be treated immediately and flexibly. Fig. 2 shows
production conditions of the equipment as soon as an example of an in-house-developed screen. The main
possible. However, when a potential quality defect occurs, functions of this screen are to search for two types of data
a person in charge of Production Engineering uses a USB to be compared and to display the graph. Even the volume
flash drive to extract necessary data from the production of this functional level requires a development period of

20 No.88 (2022) NISSAN TECHNICAL REVIEW

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 3. Manufacturing Innovation in a New Motor Line through Digitalization and Internet of Things (IoT)

approximately three months for outsourcing. However,

Table 1 Number of abnormality stops and action time by
the period could be shortened to one week by using
Abnormality
in-house development. Working
group
occurrence
Average time to
start
Remarks
times
A-group 333 3:29 Average-level instructor, workers

B-group 227 2:20 Highly skilled instructor, workers

C-group 335 3:56 Highly skilled instructor, seasonal workers

In addition, a function to predict short stops of

production equipment will be incorporated into this
system to shorten recovery periods further in the future. Alliance Internal

3.3 Work mastery with mixed reality (MR)

To achieve highly efficient production as described in
Subsection 2.2, the period necessary for a supervisor to
educate a new worker should be shortened. A new model
Fig. 2 Data visualization system screen Alliance Internal
project experiences several trial lots before the start of
mass production. If any changes occur, such as in the
design, working order, or work content, education must
3.2 Notification of equipment defect to smart watch be performed from the beginning. Hence, education man-
As described in Subsection 2.2, for highly efficient hours should be reduced using digital technology to
production, the recovery period after a short stop occurs shorten the mastery periods of workers.
must be minimized. Previously, if a short stop occurred in For the new line, a work mastery system that exploits
production equipment, a site worker who noticed the MR, a cross reality (xR) technology, was developed. MR
abnormality by looking at an andon or rotating light goggles are used in the system. A worker wearing these
handled it. For the new line, a system to send a notice to goggles can master a work order and work points through
a smart watch (Fig. 3) worn by a site worker in case of an holograms (Fig. 4). Even information such as an assembly
equipment abnormality was developed and introduced. point, which is difficult to understand through the
Because a smart watch enables the name of the affected existing 2-D information forms such as paper sheets and
piece of equipment and abnormal contents to be checked, videos, can be indicated through holograms, enabling
recovery work can be started immediately. In addition, if anyone to understand the contents easily and thereby
a worker taking action presses the action start button on reducing the mastery period.
the smart watch, other workers can see who is taking As this system enables workers to study by themselves
action and concentrate on their own work. The without supervisor attendance, the number of supervisor
introduction effect of the system is estimated to increase man-hours for education can also be reduced. In the new
the OEE by 2% based on past results. Production lines motor line, the numbers of supervisor man-hours for
take action to improve daily and accumulate improvement education and worker mastery hours could be reduced by
measures in units of decimal percentage, making a 2% 90% and 50%, respectively.
improvement in this system remarkable. In addition, after worker training, images from the
Meanwhile, for this system, the contents of equipment eyes of the workers are retained as records, which the
abnormalities and the times of related occurrence/action supervisor can check to determine and feed back whether
start/recovery have been accumulated. Nissan exploits the worker understood the work correctly, and the
these data to analyze differences in action time between mastery level of the worker.
work orders or group members to discover improvement
points such as working group arrangement, personnel 4. Conclusion
distribution, and work movements (Table 1).
Three systems utilizing digital technology were
introduced. These were developed not for specific lines,
Details of but rather as general-purpose systems, so that they can
Smart equipment fault be deployed globally within Nissan. It is desired to apply
watch these systems to achieve an OEE of 95% with no quality
defects and to target nominal values in the future.

Fig. 3 Smart watch to notice equipment defects Alliance Internal

NISSAN TECHNICAL REVIEW No.88 (2022) 21

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 3. Manufacturing Innovation in a New Motor Line through Digitalization and Internet of Things (IoT)

However, these measures are not sufficient to utilize

Operation systems for the maximum effect. For example, the
instruction MR shortage of human resources in maintenance departments
through goggles
holograms
needs to be resolved in the future. If a system does not
work correctly because of difficulties after the system is
introduced, maintenance staff must handle this problem.
Although maintenance departments have human
resources with digital knowledge and skills, the number
Alliance Internal
is insufficient at present.
As digital technology is a tool, people must fully utilize
it to benefit from its advantages. The most important
aspect of digital technology application is to employ and
Assembly point
indication fuse this technology into operations at manufacturing
through sites to enable the manufacturing site to realize its
holograms
advantages. In addition, to maintain and control the
situation, sufficient human resources should be ensured,
and training should be implemented along with system
Fig. 4 Mastery work with holograms
Alliance Internal
development and introduction. Nissan intends to address
these issues urgently.

Meanwhile, the difficulties of the steps involved in

exploiting digital technology increase in the following
order: data acquisition, visualization, automatic analysis/
prediction, and automatic control. Nissan is at the stage
of starting a range of visualization and automatic
analyses. Some issues must be overcome for wider
implementation of these technologies. One of them is to
ensure sufficient human resources. It is said that Japan
is behind Western countries and Asian countries such as
China, Korea, and Singapore in terms of digitalization.
One of the main factors is the shortage of human
resources. Although highly skilled people, such as data
scientists, must be ensured, people who utilize digital
systems, or people at manufacturing sites, will gradually
be required to improve their digital literacy.
Using a tablet terminal to conduct start checks, which
were previously performed only with paper sheets, causes
manufacturing sites to experience significant stress. In a
previous incident, a manufacturing site avoided such
systems due to the notion that they were difficult to deal
with. Nissan has implemented measures to address this
issue, including providing educational opportunities and
consultation services that can be used by manufacturing-
site workers in need. It has also promoted the
incorporation of site operations that utilize systems into
standard operation sheets.

Authors

Satoru Iizuka

22 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1

Nissan Intelligent Factory Starting from ARIYA

4. Process Innovation Through the Development of a

Simultaneous Body/Bumper Painting System

Takeshi Goto* Tomoyuki Okamoto** Chie Michiura*

integration of two plants:

1. Introduction 1. Reduction in initial costs (15%)
2. Reduction in running costs (25%)
In 2015, Nissan s main painting plants in Japan 図１
3. Reduction in CO2 emissions (25%)
became obsolete, 35 years or more after their construction; As detailed in this paper, significant benefits can be
hence, there was an urgent need to upgrade the facilities. obtained from the new method.
The painting plant mentioned here is actually a large-
scale plant in which a combination of processing steps Body coating: Baking at 140℃ Coating body and resin parts at the same time means
CO2 emission is reduced (because of baking at 85℃)
including pretreatment/electrodeposition processes, a
sealing process, an intermediate coat process, a top coat
process, and a finishing process is adopted. The cost of
upgrading these facilities is expected to exceed 10 billion
yen, including the cost of upgrading the buildings. TdC is reduced by abolishing
Therefore, upgrading is usually carried out on a 25- to Bumper/resin part coating: Baking at 85℃ the bumper coating line
Developed process (body–bumper integral
30-year cycle. Conventional process
coating at lower temperature)
These circumstances led to the commencement of
planning for upgrading Nissan s Tochigi Painting Plant, Fig.1 Conceptual diagram of body–bumper integral coating
method
which later developed into what is now known as the
Nissan Intelligent Factory initiative (NIF).
The key concepts were established as follows: Furthermore, producing the body and bumper in the
The production facility should have QCT same area completely eradicates the color-matching
competitiveness, even after 20 years. problem that occurs at the painting site. Therefore, this
Energy-saving facilities shall be established, with the novel technology can exert a significant effect in terms of
aim of realizing carbon neutrality. improving quality.
Emission of volatile organic compounds (VOCs),
substances that pollute the environment, should be 3. Development of integral coating
regulated.
Plans were formulated considering these target The development of the body–bumper integral coating
concepts. Integral coating of the body and bumper, which technology is described in detail in this section. In the
is described in this report, is a novel technology that was conventional, so-called separate coating method, the
developed as a method for realizing the abovementioned target materials to be coated are significantly different:
target concepts. This technology is introduced forthwith. metals and resins. To satisfy the coating performance
requirements for the coated films, different coating
2. What is body–bumper integral coating materials are used for each target material. Therefore,
technology? different performance requirements are set for the paints
for the metallic vehicle body and the resin bumper, and
Body–bumper integral coating technology is technology different baking temperatures are used for these coatings.
that enables the simultaneous coating of the metallic The main difference between the performance
vehicle body and the bumper, which is composed of resins. requirements is that the paint for the vehicle body needs
Conventionally, these are produced in separate plants or to be more rigid, i.e., it needs to have a higher strength.
at separate locations in the same plant. This technology Therefore, only materials baked at a high temperature
is schematically illustrated in Fig.1. (140 C for 20 min), which ensures that this performance
The main advantages of this new technology are the requirement is met, are used. By contrast, the resin paint
following rationalization effects stemming from the materials need to be flexible, and the baking process

*Paint and Plastic Engineering Department **Body Engineering Department

NISSAN TECHNICAL REVIEW No.88 (2022) 23

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 4. Process Innovation Through the Development of a Simultaneous Body/Bumper Painting System

needs to be carried out at a relatively low temperature coat (for which there are only a small number of available
(85 C) to avoid the resin deformation caused by melting colors) as well as for the clear coat, two-pack paints are
at high temperatures. In other words, the two types of not a viable option for the base coat, because the diversity
Bumpe have completely different properties (see Table 1
paint of available colors would mean that the cost would be
for rdetails). prohibitively high.
Therefore, a conventional single-pack paint was used
Table.1 Performance requirements of body and bumper paints for the base coat. To enable the low-temperature curing
Common performance criteria Unique performance criteria
Physical property of of this paint, the method of impregnating NCO from the
coating film

・ Weather resistance ・ Chipping resistance

undercoat (intermediate coat) and topcoat (clear coat),
Body coating

Coating film that has rigidity

(Surface of steel sheets with low energy
hereafter referred to as the sandwich method, was
- XWON, natural exposure, etc.
(140℃)

・ Scratch resistance and high strength

absorbency)
(Bearing capacity against
adopted. In other words, the method of utilizing NCO in
- Car washer, wiping with dry cloth, etc.
・ Window glass adhesion
・ Adhesion property input)
- Hot water, humidity resistance, thermal (Bearing capacity against high separation force)

layers other than the base coat, which allowed the NCO
cycle, etc.
・ Atmospheric deposition
・ Resistance to light collision
Bumper coating

resistance Coating film that has flexibility

- Acid rain, bird droppings, pollen, etc. (Ability to follow the deformation of the resin
and high elongation rate in the other layers to impregnate the base coat after it
(85ºC)

・ Chemical resistance substrate)

(Elongation that allows for
had been applied and a film had formed, was invented
- Gasoline, alkaline substances, wax, etc. ・ Resistance to chipping and cracking at the rear
following the substrate)
・ Compatibility with indirect (Ability to follow the deformation caused by
materials stone thrown up by own vehicle)
- Double-sided tape, PGF, etc.
during this development project (Fig. 3).
図３
Therefore, with the aim of developing a coating that ◼ To solidify the intermediate and top coats at 80℃-90℃, we developed a 2K (two-
can simultaneously be applied to both the body and the pack) cured intermediate coat and clear coat that use a new curing agent.
◼ Impregnation of curing agent from the intermediate and clear coats allows curing of
bumper, a material that can be baked at a temperature of the 1K (single-pack) base coat so that the material costs and capital investment can
be minimized.
２ 85 C and also satisfies the characteristic performance Body (140℃ × 20 min) Bumper (85℃ × 20 min)
Current technology
requirements of both substrate materials was designed
Clear coat (2K) Clear coat (2K)
Coating separately Base coat (Water-based 1K) Base coat (Solvent 1K)

(Fig.2).
Body: Water-based 3-
Intermediate coat (Water-based 1K)
Wet @140℃ Bumper (PP),
Electrodeposition Resin parts (PP, ABS)
Bumper: Solvent @85℃
Steel, Aluminum

“Materials development target:

Body (85℃ × 20 min)
high strength and high elongation rate” Developed technology New Clear coat (2K)
Bumper
Clear coat (2K)

Integral coating New Base coat (Water-based 1K) Impregnation Base coat (Water-based 1K)

Body & Bumper: New Intermediate coat Intermediate coat

Water-based 3-Wet @85℃ (Water-based 2K) (Water-based 2K)
Paint for body Electrodeposition Bumper (PP),

Development Steel, Aluminum Resin parts (PP, ABS)

High target
Fig.3 Schematic comparing conventional coating method with 4

the new common material composition (sandwich method)

Strength
To obtain a coating film with high strength, flexibility,
and elongation, using the sandwich method, the basic
Paint for bumper
図４ composition of the material was selected as detailed in
Low Fig.4.

Low Elongation High 1. Film strength is secured by reducing the molecule sizes in the resins for all of the
intermediate, clear, and base coats, thereby increasing the reaction points.
2. The crosslinks possess flexible spring character, and hence flexibility will be
imparted.
Fig.2 Conceptual diagram showing development target of 3. The curing agent has a three-dimensional network structure, and hence the agent
coating material for both body and bumper can easily impregnate the base.
Crosslinking
in the coating film
Body
4. Details of material development Rigid
If the baking temperature for the body coating is Rigid
crosslinking
lowered to 140–85 C, the coating film will not possess
Developed material
sufficient strength, a property that is conventionally
conferred by thermosetting. Therefore, the development
of a material based on a two-pack paint containing
isocyanate (NCO), which can be cured even at low
temperatures, was set as a goal. At the time of Bumper
Flexible
development, no other water-based, low-temperature, crosslinking
3-wet car coatings had been developed. Therefore, this
attempt was a first for this industry. Flexible
When designing the material, the fact that two-pack
paints are expensive and that it would be necessary to Fig.4 Conceptual diagram of the design of the coating film for
the two materials
invest in mixing equipment for each paint machine was
considered. Thus, although it could be possible to use
two-pack paints in the production line for the intermediate

24 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 4. Process Innovation Through the Development of a Simultaneous Body/Bumper Painting System

図６
After fabricating the paint based on the concepts With the conventional oven, increasing the temperature
summarized in Fig.4, testing to determine the basic on the inner surface of the door, especially around the door hinge,
performance of the coated film was carried out; to was difficult.
evaluate the feasibility of its use in an actual production
line, the workability was concurrently evaluated.
Conventionally, such evaluations are performed in a
certain order. However, the material under development
was completely new, and hence, there was a strong
possibility that past experience was not applicable.
Therefore, the evaluations were conducted concurrently
with the aim of minimizing the loss due to rework. While
these evaluations were being conducted, the variable Temperature distribution: entire body outer panel

factors in the production line, such as 図７ the baking

Inner surface of panel, around the door hinge
temperature, coating thickness, and amount of curing
agent, as well as the ranges over which they varied, were Fig.6 Example of a poorly baked section: around the door hinge
of the vehicle body
identified. In addition, accurate information regarding
the margin for the ultimate performance limit (cliff of
quality) was obtained to ensure that the developed
material would be considerably robust. The relationships * Using the conventional oven, warming up requires at least 30 min.
The target for this time is 15 min.
between these factors are shown in Fig. 5.
図5 Temperature (℃)
Temperature upper limit: 90℃

◼ We selected a material (curing agent) that exhibits robust performance regardless of

variations during mass production in baking temperature, coating thickness, and
amount of curing agent. Outer panel
Inner panel (Target)
◼ We have checked each performance to the ultimate limit and obtained information
about the safety margin with respect to the current control range.
Inner panel (Initial)
Baking temperature (℃)
Baking temperature (℃)

Time required to warm up the inner

Mass-production Mass-
production surface of the body panel: 30 min.
control range control
range
Time
Not Not
acceptable acceptable

Standard
Coating thickness (μm) Amount of curing agent (%)

Baking temperature - Baking temperature - Fig.7 Target and initial (measured) temperature increase curves
Coating thickness Amount of curing agent for the inner surface of the panel
[Defect items in case of deviation from the control range]
Body: Whitening (Water resistance test, Humidity resistance test), Adhesion failure (DG test, Chipping resistance test)
Bumper: Adhesion failure (Chipping resistance test), Cracking (Bending resistance test)
6 Therefore, to resolve this problem, a flow analysis was
Fig.5 Relationships between baking temperature and coating conducted using the higher-order total variation
thickness (left) or amount of curing agent (right)
diminishing method to calculate the appropriate air flow
velocity and volume. Based on the results of this analysis,
5. Development of supporting technologies it was deemed appropriate to design a new nozzle shape
for integral coating for use in the baking oven. Fig.8 shows the results of
simulations that visualize the change in temperature of
5.1 Development of low-temperature oven for the hot air volume flowing into the vehicle body. In the
integral coating conventional oven, the hot air failed to reach inside the
To simultaneously bake the resin parts and the vehicle vehicle body, and heat exchange via convective flow
body, which is constructed from steel sheets, the blowout occurred only at the outer surface of the panel. However,
temperature of the drying oven needed to be fixed below when the air flow volume and velocity were increased,
the thermal deformation threshold temperature of the hot air entered the vehicle body and convective flow
resin (95 C). When the blowout temperature was occurred at the inner surface of the panel, as shown in
controlled in this manner, problems occurred in certain Fig.9. In this manner, it was possible to ensure that, in
sections, such as the inner surface of the vehicle body an actual production line, the target temperature
door (Fig.6). Compared to operating the oven with the increase curve could be realized, even for the inner
conventional blowout temperature, more than twice the surface of the panel.
time was required for the temperature to increase to the
specified value in this area (Fig.7). Hence, long machine
design were required for the baking and drying oven.

NISSAN TECHNICAL REVIEW No.88 (2022) 25

 図９
Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 4. Process Innovation Through the Development of a Simultaneous Body/Bumper Painting System

all the materials, leading to an acceptable quality of color

matching.

When integral coating is started from When integral coating is started from
somewhere other the grounding point the grounding point

Visual Not Visual Acceptable

evaluation acceptable evaluation

White White
Not Not
acceptable acceptable

Variation in the color

difference is suppressed
Acceptable Acceptable

Air temperature decreased such that convective flow

did not occur inside the vehicle body.
Not Not
acceptable acceptable

Black
Black Angle Angle

Fig.8 Temperature simulation for the conventional drying oven

Fig.10 Color as a function of angle: comparison between
coating from the grounding point and coating from somewhere
other than the grounding point

Even if complete matching of the spectral reflectance is

achieved for all the angles, the color of different parts will
appear to vary if there is a change in the angle at the
interface of the parts.
When using the integral coating method, if this
phenomenon occurs at the production site, there will be
no room for adjustments because color toning cannot be
performed for each material. Therefore, for each part
that undergoes color matching, it is important to
Fig.9 Temperature simulation for the drying oven used for
integral coating determine in advance the permissible angle deviation
and to conduct surface angle evaluations during the
actual color matching evaluation. In this development
5.2 Color matching under integral coating conditions project, a digital color-matching evaluation system was
In automotive coating lines, electrostatic coating is designed to prevent the occurrence of this problem, as
performed to improve the material yield. Because the detailed in Fig. 10. The system, used in combination with
vehicle body, which is constructed from steel sheets, and a 3D scanner, will be used to resolve such problems and
the resin parts have different electrical conductivities, it prevent their occurrence in the future.
is not possible to obtain the same color for these parts図10

simply by coating them simultaneously. Equipment configuration

To solve this problem, all the following conditions need Conversion software
(Lab→XYZ)
Reflectance

to be the same. Color ・ Output of color

measurement data

1. Coating environment (temperature, humidity, and

・ Addition of hue
Wavelength
fluctuation
Spectral colorimeter records

downflow air velocity)

spectral reflectance
Light Shape
source

2. Paint material
3D scanner measures vehicle shape Reproduction of hue in the rendering software

3. Coating conditions (discharge volume, rotation speed,

Relative strength

shaping air flow rate, and voltage) Wavelength

4. Coating locus (distance between objects to be coated,

Spectroradiometer records ambient
light spectrum

coating pitch, and speed)

The following items of information were used as inputs for the rendering software for
5. Coating thickness color matching evaluation:
In particular, considering point 5, it is important to 1) Spectral reflectance vs. material angle: multi-angle spectral colorimeter
2) Ambient light source spectral calibration: spectroradiometer
ensure that the electric conductivity of each workpiece is 3) Surface shapes of parts: 3D scanner

identical for achieving an equivalent electrostatic coating

Fig.11 Digital color-matching system
effect. In other words, the electrical conductivity of all the
materials must be the same.
Therefore, in this development project, water-based
paints were selected to ensure electrical conductivity,
and coating was performed from the grounding point of
each part such that the same electrical conductivity
would have been achieved regardless of the material. In
this manner, it was possible to achieve the same color for

26 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 4. Process Innovation Through the Development of a Simultaneous Body/Bumper Painting System
Figur
e 11
Color checker Actual vehicle
Photograph

Enlarged
view
Simulation

Enlarged
view

Fig.12 Color tone evaluation: actual photograph vs. Simulation

6. Value and future prospects of integral

coating method

By establishing the new body–bumper integral coating

technology, 25% reductions in energy consumption and
CO2 emissions, compared to the conventional method,
were achieved. Henceforth, this method will be regarded
as a new standard method and will increasingly be
adopted in newly established production lines and
contribute toward the realization of carbon neutrality.
By adopting this technology, it was confirmed that it is
possible to abolish the necessity for color matching
between the body and bumper of the vehicle, which
typically requires long periods of time at the painting site
for optimization. Apart from affording the possibility of
shorter lead times for the fabrication of new vehicle
models and improved stability of quality at launch, this
method should lead to the adoption of highly stylish
paints (for example, paints with high gradation), which
could not be achieved via conventional coating methods.
Therefore, it should be possible for manufacturers to offer
more attractive colors schemes.
By introducing this novel method, it is also expected
that the novel technology can be used to offer customers
higher-quality coatings.

NISSAN TECHNICAL REVIEW No.88 (2022) 27

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 4. Process Innovation Through the Development of a Simultaneous Body/Bumper Painting System

Authors

Takeshi Goto Tomoyuki Okamoto Chie Michiura

28 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1

Nissan Intelligent Factory Starting from ARIYA

5. Creation of a Quality Gate through Automated

Technology of Appearance and Specification Inspection

Teruyuki Ishiwata* Mitsuru Hirayama* Daisuke Tanaka* Takayuki Moriya**

Yasuo Yamada** Makoto Yamada* Takeshi Honda***

skills for difficult-to-judge inspections and various

1. Introduction specifications.
Further quality improvement should be attained by
To produce better cars with high quality, the ideas and exploiting automated inspection data. By controlling
knowledge regarding craftsmanship accumulated so far automation-provided inspection results for all vehicles
have been organized into the alliance production way sent to the Internet of Things server, a system in which
(APW). Accordingly, Synchronization of quality: a transition to the next process requires passing quality
Thoroughly making goods with quality meeting standards should be established. In addition, exploiting
customers expectations through processes have been recorded traceable data can lead to early identification
designed so that sub-par goods are not accepted, produced, of potential defects.
or released publicly.
To achieve this objective, a system of Necessary 2.3 Application of automated inspection technology
quality is defined and assured at a quality assurance In order to realize automated inspections by quality
stage set in each manufacturing process (from heading to gates, the following issues at the NIF must be considered:
assembly) to complete the assurance within the process paintwork defects, which greatly affect vehicle design,
was defined as a quality gate. This chapter will introduce and specification errors, which ruin customer trust. Here,
the development of automated inspection technology for two application cases will be introduced for 100%
achieving quality gates at the Nissan Intelligent Factory assurance inspection equipment.
(NIF). In the first case, total assurance of color, clarity, and
film thickness in paint processes is achieved (Section 3
2. Process assurance by quality gate of this chapter). Minute paint faults, which can only be
detected by experts, become detectable. In addition,
2.1 Necessity of automated inspection by quality gate 100% measurements of the color, clarity, and film
Training and maintaining skilled inspectors will thickness become possible.
become an issue due to the workforce decrease as the In the second case, assurance of vehicle paintwork
birthrate declines and the population ages. Meanwhile, secondary defects, glass defects, and specifications, is
all interior/exterior types and paint colors resulting from achieved (Section 4 of this chapter).
the diversification of customer needs must be ensured. To Regarding appearance inspections such as those
respond to such environmental changes, promoting the conducted to identify defects on the glass, which are
automation of inspections is an effective way to produce difficult to notice, and inspections of various specifications,
quality cars. In addition, utilizing inspection data, whose stable assurance independent of worker skills can be
acquisition will be possible with automation, will lead to achieved.
further prevention of defects.
To promote these aspects, it is necessary to exploit
automated inspection technology at a quality gate.

2.2 Expected effects of automated inspections

The following items should be achieved for in-process
assurance, promoting automated inspections by quality
gates.
The quality assurance level should be improved through
Glass pattern automated inspections. Completely
automatic inspections enable stable inspection/
judgment abilities to be kept independent of inspector

*Production Engineering Research and Development Center **Vehicle Quality Engineering Department
***Paint and Plastic Engineering Department

NISSAN TECHNICAL REVIEW No.88 (2022) 29

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 5. Creation of a Quality Gate through Automated Technology of Appearance and Specification Inspection

Measurement method In shooting

3. Paint appearance automated inspection

3.1 System composition and process outline

Multiple robots with inspection/measurement units
(Fig. 1) were utilized in the final paint appearance
inspection process, and a 100% automated inspection
area for bodies and bumpers was established.
The layout is shown in Fig. 2. The upstream side Shoot one period several times at the same point
Sensor: Monochrome camera
detects paint defects such as dust and seeding, and the Lighting equipment: LED monitor while shifting the fringe

downstream side measures color, clarity, and film

thickness.
Max. brightness image Min. brightness image Difference image

Defect

Highlight-
ing

Calculate a difference from the

Extract the max. and min. brightness
max. and min. brightness
images from taken images
images

Fig.1 Composition of inspection robot Fig.3 Fringe projection-type image recognition technique

The issue of fringe projection-type image recognition

technology is that detection rates vary depending on the
brightness levels (L-value) of paint colors.

<Issue-resolving-method details>
To eliminate the detection rate differences due to paint
color differences (differences in paint color brightness
(L-value)), camera-side gain adjustment by paint color
Fig.2 Layout of 100% paint appearance automated inspection (Table 1) was applied to stabilize the highlighting
area treatment.
Consequently, even paint defects at a level of 0.3 mm
3.2 Detection of paint defects and slight denting crater defects, which could be detected
<Conventional technologies and issues> only by humans previously, became automatically
Some widely used paint defect inspection units employ detectable (Fig.4, Fig.5).
fringe projection-type image recognition techniques. As
the fringe projection is moved, multiple images are Table.1 Performance requirements of body and bumper paints
captured by the camera. Then, highlighting the difference Paint color
Brightness Gain set value

between the maximum and minimum brightness images

(L-value)
1 1.5 2 2.5 3 3.5 4 5 6

enables the detection of micro-sized unevenness defects

K23 (silver) 126 85 100 100 100 60 10 15 10 10
QAB (white) 107 70 100 100 100 80 15 0 5 0
(Fig. 3*1)). EAU (orange) 83 70 75 100 100 100 75 60 20 0
RBK (blue) 56 70 80 100 100 100 100 100 65 40
NAH (red) 40 65 60 100 100 100 100 100 100 70
KH3 (black) 5 70 75 100 100 100 100 100 100 70

Defect type Dirt Popping Creater Thinner splash Water splash

0.5 mm 0.5 mm 1.0 mm 0.5 mm 0.5 mm

Defect shape

Cross-section shape Cross-section shape Cross-section shape

Top view Top view

Defect observation
photo

Fig.4 List of detectable paint defects

Fig.3 Fringe projection-type image recognition technique

30 No.88 (2022) NISSAN TECHNICAL REVIEW

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 5. Creation of a Quality Gate through Automated Technology of Appearance and Specification Inspection

<After the countermeasure logic was added>

Fig.6 Difference in judgment sizes of special defect shapes

between inspection equipment and a human and the effects of
Fig.5 Difference in detectable defective shape between implementing a countermeasure
conventional and new systems

In addition, placing an imaging unit on the robot and

In addition, for defective dirt shapes as in the two changing the fringe direction in the monitor to prevent
cases mentioned below, a range recognized as a defect by character lines on the vehicle surface from making the
a person becomes so wide that the size judged by an fringe unclear (Fig.7) enabled 98% of the outer surface to
inspector device becomes narrower than the size be detected even if the vehicle body shape was complicated.
measured by a human. Inspecting 562 areas per vehicle and checking 11
Case 1: A clear-layer swell makes its foot wider. resultant images per area for any defect yielded 6,182
Case 2: A defect is surrounded by a groove. imaging times per vehicle.
To prevent the defect judgment size from narrowing, a
logic in which both composite images and raw images
(close to human vision) were used for judgment was
added, leading to a narrower deviation from human
vision (Fig.6).

<Before the countermeasure logic was added>

Fig.7 Fringe image differences due to combinations of character

line and fringe pattern

Based on the above, a previous detection rate of 95%–

100% was achieved.

3.3 Measurements of color, clarity, and film

thickness
<Conventional technologies and issues>
In previous color, clarity, and film thickness
measurements, it was necessary to bring the head of a
measuring instrument into contact with the vehicle
surface to be measured, and a person adjusted the
position of the measuring head. To automate the position
adjustment in a contact state, it is necessary to control

NISSAN TECHNICAL REVIEW No.88 (2022) 31

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 5. Creation of a Quality Gate through Automated Technology of Appearance and Specification Inspection

the position accuracy of the measuring head with respect

to the surface to enable measurement within 1 mm.
Therefore, the stop position accuracy of a robot having a
measuring head must be improved.
This issue has made it difficult to realize 100%

Green Green
automated measurements in automotive production
lines.

<Issue-resolving-method details>
By utilizing a vision system, high-accuracy positioning
by a robot and 100% automated inspections were
achieved.
Fig.8 Layout of robots in the inspection stage
3.4 System features
In this system, if either
Incomplete correction completion registration for a
defect detected by paint defect inspection equipment
occurs or
The color, clarity, or film thickness measurement
results do not satisfy quality standards,
then the vehicle will not be conveyed to the next
process. This control system was established to realize
in-process assurance.

3.5 Advantages
The effect of establishing automated dust inspection is
Fig.9 Actual inspection scene
that a four-person inspection of man-hours is not required
and that the stable inspection execution causes the
shipment quality assurance level to increase. The quality
data of all vehicles measured in the final paint process by 4.2 Specification inspection
this system and the database of various in-production <Conventional technologies and issues>
conditions monitored by the painting equipment can (1) Paint color judgment
minimize the time required for potential for potential A judgment method utilizing Euclidean distances
defects to be identified. and help establish equipment (color differences) in the L*a*b* color space is available
conditions to maintain a more stable quality level (source for image processing in color specification inspection.
measurement). Vehicle paint colors, however, include similar colors such
as solid white and pearl white, whose differences cannot
4. Integrated automated inspections be recognized without placing them side by side and
(specifications and defect inspections) whose distance, as described above, is close. In addition,
the relation between a light and an inspection surface
4.1 System composition and outline may cause the reflection rate to vary, resulting in
This equipment performs two types of inspections: overlapped distributions and leading to misjudgment,
specification inspection and defect inspection. which is recognized as an issue (Fig. 10).
Specification inspection utilizes information from the
production management system to determine whether Solid white
object parts were installed according to the instructions.
Pearl white
Then, the defect inspection assesses not only the painted
surfaces, but also the glass surfaces. Because these White
inspections are performed in the same stage, the robots
have lights and vision sensors separately installed for Overlapped
distributions
specification inspection and defect inspection (Fig.8, Yellow
Fig.9).
Green

Red

Blue

Black
Fig.10 L*a*b* color space

32 No.88 (2022) NISSAN TECHNICAL REVIEW

Indirect light
Sliding
mechanism
Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 5. Creation of a Quality Gate through Automated Technology of Appearance and Specification Inspection

(2) Ensuring judgment accuracy for various inspection Indirect light

objects and surface conditions Sliding
In the object inspections performed by this equipment, mechanism
existence/absence, shape, position, posture, and color are
applied to pass–fail judgment. Every judgment other
than color is based on the outline of the part from image
processing. However, the light reflection rates of parts
with different surface conditions, such as painted, mirror-
finished, unpainted, and glass surfaces, vary, so that
some lighting methods may cause the outline of a part to
disappear. An inspection object part and a background of
the same color cause no difference in contrast, leading to Ring light
the disappearance of the outline. The combination of
multiple lights and vision sensors can serve as a Fig.11 Components of the hand section
countermeasure to overcome this issue. This (lights and vision sensors)
countermeasure, however, causes the number of
attachments to a robot to increase, making the avoidance In addition, while minimizing unexpected background
of interference with other robots an issue and causing appearances within an image, contrast difference is
posture changes to require so long that a target cycle ensured by switching not only the parameters of the
time cannot be achieved. lights and vision sensors automatically for each paint
color and part, but also the imaging postures.
<Issue-resolving-method details> Consequently, a defective-goods outflow rate of 0% and a
(1) Paint color judgment fine-goods detection error rate of 0.001% or less were
Lights are placed at positions such that they provide achieved (Fig.12).
indirect lighting so that the reflection rate can be
controlled, and the parameters of the lights and vision Painted surface Unpainted surface
sensors can be automatically switched by color group
based on information from the production management
system. In addition, the detection area is selected so that Unpainted surface
its position is hardly affected by the reflection rate and
two positions of an inspection object part and an adjacent
part can be within the vision, and a judgment logic is
utilized in which the color-space distances between the
two positions described above are compared. Charge port lid License plate
Consequently, 100% judgment accuracy can be achieved.
Additionally, because inspections can be performed Mirror-finished
without providing a judgment logic and threshold for surface Glass surface
each paint color, maintainability is also greatly improved.

(2) Ensuring the judgment accuracy for various inspection

objects and surface conditions Emblem
With an understanding of the equipment objects, Door glass
inspection types and part characteristics, scattered light
from indirect lighting was utilized for painted and mirror- Fig.12 Examples of detection images by surface condition
finished parts to control the reflection intensity, and
regular reflection light from a ring lighting device was
utilized for unpainted parts and on-glass printed letters 4.3 Defect inspection
to increase the reflection intensity. For slight differences <Conventional technologies and issues>
in the reflection rate between parts, the optimal reflection (1) Defect detectionPrinted pattern Glass Dirt
Part edge ID paint
intensity is ensured by providing a sliding mechanism for This equipment on glass the same fringewire
employs hot projection-
(glass surface
a lighting device and controlling its illuminating position. type defect detection sensors utilized for paint inspection,
The sliding mechanism can also store the lighting section as mentioned above. Most scratches, which are the
during non-shooting to achieve a small hand section. This Antenna
inspection items, have long and narrow shapes with
approach minimizes the effect on the cycle time (Fig.11). widths of 0.1 mm or less and have characteristics
completely different from those of paint defects. In
addition, it is necessary to detect defects even when there
Roof section is a difference in the light reflection rate between the
painted surface and glass surface.

NISSAN TECHNICAL REVIEW No.88 (2022) 33

 Charge port lid License plate PaintedPainted
surface surface

Mirror-finished
Special Feature 1 : Nissan
surface Painted Glass
Intelligent
surface surface
Factory Starting from
Unpainted ARIYA - 5. Creation of a Quality Gate through Automated Technology of Appearance and Specification Inspection
surface
Painted surface Unpainted surface ③ ③
Unpainted surface ② ④ ①
②② ④
① ②
(2) Detection error control processing①to extract areas ① based on the characteristics
Unpainted surface ①such items ①
② and automatically ② masking them
In defect detection, a defect is determined by extracting
Emblem (patterns) of
an area from the image Door where
glass the pixel brightness (Fig.15).
Glass surface
Charge port lid License plate
suddenly changes. Because the inspection is performed Glass surface
Mirror-finished
in the final assembly process, inGlass contrast
Charge port lid surface surface to inspection
only for painted bodies, License plateof the
the edges installed parts, Glass hot wire
③ ③ area
② ④ ② ④
on-glass identification paint, printed patterns, hot wires,
Mirror-finished
① ①
dirt, surface
and/or wiping
Printed GlassGlass
patternmarks surface
made duringDirtwork exist
edge ID paint Emblem Glass pattern
(Fig.13). These elements may wire
on glass hot Door glass
cause sudden changes in
(glass surface) ③ ③ area
② ④ ② ④
brightness and may be detected as defects by mistake;
① ①
therefore, this issue must be controlled.
Antenna Emblem Glass hot wire

Door glass Fig.15 Auto-masked areas (solid in yellow)

area

Printed pattern Glass Dirt Glass pattern

Part edge ID paint hot wire area
on glass (glass surface)
section
It is difficult for the current defect judgment method
Dirt Defect Glass hot wire
Antenna to handle dirt or wiping marks (distinction between area dirt
and defects), which occur in random positions and do not
Printed pattern Glass Dirt Glass pattern
edge ID paint
Roof section hot wire have unique characteristics such as shapes, leading area
to
on glass (glass surface)
potential misdetection. Dirt This issue canDefect
be addressed by
Fig.13 Examples of detection error factors considering a new method.
Antenna In addition, the inspection results (Fig.16), including
those of the specification inspection, are stored
<Issue-resolving-method details> simultaneously and accumulated.
f section (1) Defect detection
The characteristics of the defects that can occur on a Dirt Defect
vehicle and tuning the fringe width/cycle of lighting,
number of imaging times, imaging parameters, and others
factors for each of the painted surfaces and glass surfaces
enabled clearer defect images to be obtained (Fig.14).
Fig.16 Example of defect inspection results
Painted surface

4.4 Advantages
This equipment enabled the automation of the
following inspections based on certain criteria
independent of inspector skill in an assembly line for the
mixed production of various vehicle models and
specifications: complicated specification inspection and
① ② ① ② defect inspection for painted surfaces and glass surfaces
with minute and various characteristics. This approach
provides a more stable and higher quality assurance level
Glass surface than the previous visual inspections and will contribute
to achieving in-process assurance, preventing potentially
defective vehicles from being released to customers. This
method also provides a worker-friendly environment in
which further quality defects do not occur and are
prevented by utilizing accumulated inspection data to
identify defect causes and to design processes for fewer
③ ③ defects.
② ④ ② ④
① ① 5. Conclusion

Fig. 14 Images of imitated defect detection on painted surfaces This report introduced the concept of a quality gate
and glass surfaces and the technical development cases of automated
(left: Camera images; right: Processed images)
inspections to improve the quality assurance level of the
quality gate.
(2) Detection error control At present, the aim is to promote automation of
To control misdetection caused by part edges, on-glass inspections
Glass hot wirewithin manufacturing processes. In the
patterns, hot wires, and other factors, a logic is proposed future, the objectives the plan to be expanded the range
area
in which defect misdetection is prevented by using image to complete vehicle inspection processes and to promote
Glass pattern
34 No.88 (2022) NISSAN TECHNICAL REVIEW
area
Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 5. Creation of a Quality Gate through Automated Technology of Appearance and Specification Inspection

the application of automated inspections to the entire

assurance process.
These achievements would contribute to improving the
brand value of Nissan by improving vehicle quality
through high quality assurance level and promoting
quality improvement based on inspection results to
deliver high-quality products that meet customer
expectations.

[References]
*1) x3Projects Co., Ltd.:
Surface inspection equipment, utility model registration
number 3197766, (2015)

NISSAN TECHNICAL REVIEW No.88 (2022) 35

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 5. Creation of a Quality Gate through Automated Technology of Appearance and Specification Inspection

Authors

Teruyuki Ishiwata Mitsuru Hirayama Daisuke Tanaka Takayuki Moriya Yasuo Yamada Makoto Yamada

Takeshi Honda

36 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1

Nissan Intelligent Factory Starting from ARIYA

6. Improvement of the Equipment Operation Rate and

Process Capacity by Implementing Predictive/Preventive
Maintenance and Remote Maintenance Technology
Katahiro Ogawa* Satoru Sakurai** Koichi Makino** Takuya Tsumoto***
Susumu Abe*** Masayoshi Hayakawa* Kazunari Seo* Minoru Kono*

An investigation of the equipment failure history

1. Introduction revealed that three vehicle plants in Japan (Tochigi,
Kyushu, and Oppama) had 371 MBDs that could delay
Nissan Motor must stably manufacture vehicles of vehicle deliveries over the past five years, and 132 of
high functionality and high quality, such as the ARIYA, these cases (34%, and the greatest portion) were caused
to deliver them to customers on time. For this purpose, by equipment defects that were difficult to find through
advanced equipment with multiple functions and high regular inspections. In other words, it implies that
performance were introduced into the Nissan Intelligent conventional Time Based Maintenance is not sufficient to
Factory (NIF). However, there are also a lot of aged eliminate MBDs, and Condition Based Maintenance
equipment at plants. Consequently, the level of (CBM) is necessary.
maintenance required for smooth production is higher Accordingly, Nissan Motor decided to exploit IoT
than ever. technology to introduce a predictive/preventive
The maintenance department has stated the following maintenance system that constantly monitors data
three principles for some time in order to improve quality, indicating equipment defects. This chapter introduces
cost, and time (QCT): allow no equipment failure ; important elemental technologies comprising this
allow no failure recurrence ; and fix a failure predictive/preventive maintenance system.
immediately. Under these three principles, the
maintenance operations has been managed with a large 2.1 Measurement technology
number of skilled maintenance staff. However, the staff In order to monitor production equipment conditions
structure of the maintenance department has changed. continuously, it is necessary to measure and collect data
One reason is that as the number of skilled maintenance indicating equipment defects constantly. Data indicating
staff have decreased, it has become impossible to continue equipment defects can be classified into the following
the maintenance operations that have heavily relied on two types: data originally monitored by production
their skills and intuition. Another reason is that as the equipment itself for controlling the equipment and/or
total number of maintenance staff members tends to quality control, and data acquired by sensors retrofitted
decrease, it becomes more difficult to handle both sudden into production equipment, which were not originally
equipment failures and regular inspections, which are monitored by the production equipment. Whereas the
normal tasks. cost of acquiring the former data is low, acquiring the
In order to resolve these issues, the Maintenance latter data requires costs related to the following devices
Department of Nissan Motor is promoting various work and their installation: [1] sensors for measuring specific
innovations with digital technology, such as the latest physical quantities, [2] amplifiers for amplifying weak
Internet of Things (IoT). As one such innovation, this signals from sensors, [3] devices for converting analog
report introduces the predictive/preventive maintenance data from amplifiers into digital data, [4] devices for
and remote maintenance applied to NIF. transmitting digitally converted data into wired/wireless
networks, and [5] network wiring for wired transmission.
2. Predictive/preventive maintenance Installation of [1]–[5] at hundreds of in-plant locations
needs high cost and long construction period, resulting in
As a middle- and long-term strategy for addressing the the difficulty in increasing a number of equipment
goals of allow no equipment failure and allow no monitored. In addition, even for installing a measurement
failure recurrence, which are two of the three Nissan sensor in a movable point of equipment, the necessity of
Motor maintenance principles mentioned in the previous wiring for data transmission may compel such sensor
chapter, Nissan Motor has stated a target value, zero installation to be abandoned.
major breakdown (MBD) originating in equipment in Accordingly, Nissan Motor jointly developed a
FY30. diagnosis sensor HUB with Fuji Electric Co., Ltd. to

*Plant Maintenance Engineering Department **Powertrain Production Engineering Department

***Production Engineering Reserch and Development

NISSAN TECHNICAL REVIEW No.88 (2022) 37

 SpecialFeature1:NissanIntelligentFactoryStartingfromARIYA-6.ImprovementoftheEquipmentOperationRateandProcessCapacitybyImplementingPredictive/PreventiveMaintenanceandRemoteMaintenanceTechnology

reduce the installation cost of the retrofit sensor by retrofit sensors into data compatible with the OPC-UA
significantly and to measure the movable points easily standard before they are sent to IoT networks.
(Fig.1). The device has the following built-in components: Accordingly, Nissan Motor developed field storage, which
[1] a three-axis vibration sensor, [2] an amplifier, [3] an is a tool for converting data of various structures acquired
analog/digital converter, and [4] a wireless data from production equipment into data compatible with the
transmission system. It also has a four-channel external OPC-UA standard. The field storage converts data
input terminal and can be connected to various measuring received from production equipment so that they become
sensors to realize wireless communication of various compatible with the OPC-UA standard. The converted
measurement data at low cost. Furthermore, the data are transmitted to the predictive/preventive
dimensions of the diagnosis sensor HUB can be as small maintenance platform explained in Subsection 2.3 and/or
as 10 cm 10 cm 3 cm, enabling it to be installed easily the database for the quality control system (Fig.2).
even in complicated, quickly moving production Nissan Motor deploys the same type of production
equipment such as a robot (Fig. 1). equipment in many plants. On the predictive/preventive
The NIF has 189 diagnosis sensor HUBs installed for maintenance platform, in order to apply the same
constant production robot monitoring. Three-axis 800 Hz diagnosis logic programs to the same type of equipment
vibration data per cycle are transmitted to an IoT network at multiple plants, data acquired from the same type of
via the field storage system introduced in Subsection 2.2, equipment must be stored in a database having a common
and the predictive/preventive maintenance platform data structure at all plants. Accordingly, Nissan Motor
introduced in Subsection 2.3 diagnoses the data prepared a standardized data structure called data
automatically and constantly to identify any model . Field storage converts data so that they
abnormalities. conform to the data model. Thus, data acquired from the
same type of equipment can be transmitted to and
accumulated in the predictive/preventive maintenance
platform through the common data structure and the
common communication standard at every plant.

下
At present in NIF, 32,464 types of data are constantly
stored in the database for the predictive/preventive
maintenance of 1,229 pieces of equipment. As the number
of pieces of equipment requiring predictive/preventive
(Source: Website of Fuji Electric Co., Ltd.) maintenance increases in the future, the number of data
types will increase as well.

IoT network

NICS注：画面内は英訳不要 Field storage

CSV, etc.
Fig.1 Diagnosis sensor HUB installed into a vehicle-body robot

Equipment,
2.2 Data collection and utilized technology sensors
Nissan Motor declares collect once, use many as a Fig.2 Field storage
concept and is promoting the in-house standardization of
data communication standards and data structures so
that a huge amount of data acquired from production 2.3 Systemization technology
lines can easily be used for not only predictive/preventive As mentioned at the beginning of Section 2 in this
maintenance, but also various purposes such as quality chapter, Nissan Motor must implement predictive/
monitoring and tracing. preventive maintenance over all bases rapidly to achieve
zero major breakdown (MBD) originating in equipment
As a unified data communication standard, Nissan in FY30 and monitor the operation conditions in the
Motor selected OPC Unified Architecture (OPC-UA), future. As a predictive/preventive maintenance platform
which is spreading in Europe, where Industry 4.0 is for this purpose, Nissan Motor has developed MASTER
advanced. However, a lot of production equipment and CBM .
measurement sensors are not yet compatible with The greatest feature of MASTER CBM is that one
OPC-UA; therefore, it is necessary to convert data platform has the following three main functions
acquired from production equipment and data measured implemented as a single package: [1] a diagnosis logic

38 No.88 (2022) NISSAN TECHNICAL REVIEW

 SpecialFeature1:NissanIntelligentFactoryStartingfromARIYA-6.ImprovementoftheEquipmentOperationRateandProcessCapacitybyImplementingPredictive/PreventiveMaintenanceandRemoteMaintenanceTechnology

preparation function, [2] an online automated diagnosis

function, and [3] a diagnosis result dashboard function.
The diagnosis logic preparation function of [1] enables
diagnosis logic to be prepared in a non-programming
development environment (Fig.3). For the NIF, 29 types
of diagnostic logic were prepared. The prepared diagnosis
logics can also be shared with all plant of Nissan Motor a) View of in-building layout b) View of line layout

3 through the diagnosis model management system

developed in this study, and the operation status can be
monitored remotely.

図5
c) View of stage layout d) View of sensor layout

Fig.4 Drilled-down dashboard

Fig.3 Non-programming diagnosis logic development

environment

The online automated diagnosis function of [2] is a Fig.5 Data indication

function that executes automatically the diagnosis logic
created by [1] to judge the data measured and transmitted
constantly. Considering the market, few systems have Furthermore, for one of the three maintenance
NICS注：画面内は英訳不要
both functions to develop diagnosis logics and functions principles, allow no failure recurrence, MASTER CBM
to auto-diagnose online. Nissan Motor did not prepare has a function to attribute each measured datum for
these functions in separate systems, but rather intended which a diagnosis logic gave an alarm to the investigation
to use both functions on the same platform. This MASTER results of the object equipment as a part of the function of
CBM has both functions implemented. The results of [3]. It is desired to enable more accurate estimation of
auto-diagnosing measured data for every cycle are failure points and equipment stop times using measured
indicated in the plant layout view in real time. data from this information in the future.
The diagnosis result dashboard function of [3] indicates
the equipment that a diagnosis logic judges abnormal 2.4 Diagnosis logic development
with a flashing red indicator (Fig.4). Clicking the red The hardware aspects, such as devices and systems,
indicator repeatedly enables the following four layers to were explained in Subsections 2.1–2.3. Another important
be drilled down and finally leads to the equipment factor is the diagnosis logic in the software, which Nissan
component causing an alarm: a) in-building layout, b) Motor has also made an effort to develop.
line layout, c) stage layout, and d) sensor layout (Fig.4). It In the diagnosis logic development, after the object
also has a function to indicate an error point similar to equipment is selected, failure mode and effects analysis
the layout view if a defect occurs in a measurement (FMEA) is performed to extract parts to be deteriorated
function itself, such as a sensor. In addition, it can and their failure modes based on the components,
indicate measured raw data as well as the abnormality working mechanisms, and defect history information of
judging thresholds and feature quantities (statistical the equipment. Then, the optimal maintenance method is
values, etc.) used in diagnosis logics at any time scale selected based on the degree of influence, occurrence
(Fig.5). frequencies, detectability, and data to be measured,
judgment methods and other factors are examined for
failure modes that require constant monitoring.
For data to be measured, physical quantities likely to
indicate abnormality are selected mainly from electric
current, torque, vibration, pressure, flow rate, size,
temperature, etc. The pretreatment, feature quantity
settings, and criteria of data necessary for detecting
abnormality are tentatively fixed by investigating actual
data variations in the normal state. The raw values and/

NISSAN TECHNICAL REVIEW No.88 (2022) 39

SpecialFeature1:NissanIntelligentFactoryStartingfromARIYA-6.ImprovementoftheEquipmentOperationRateandProcessCapacitybyImplementingPredictive/PreventiveMaintenanceandRemoteMaintenanceTechnology

or waveforms of the measured data can be used for

judgment as they are in some cases; however, processing In the third case, the correlation between a change in
is required in most cases. Pretreatment such as noise the measured data and the equipment deterioration
elimination through a band-pass filter is performed, and condition is unknown. Because such measured data
the data are processed into feature quantities, such as change irregularly, quantification of the change is
mean values per cycle or per unit time, medians, difficult. For such cases, Nissan Motor fully utilized a
maximum/minimum values, and deviations, and statistical method to develop diagnosis logics and
verifications are repeated whether methods such as constructed a diagnosis logic as an example in which
upper/lower limit diagnosis, guard band diagnosis, many feature quantities were extracted from the
frequency analysis, and trend analysis can be used to waveforms of the measured data to digitalize the
judge the status as normal/cautious (signs of abnormality)/ waveform changes (Fig.7). However, this case also
abnormal. requires reviewing feature quantities, thresholds, and
Among these physical quantities, pretreatments, other aspects based on a huge amount of abnormality
feature quantities, and judgment methods, the optimal data in order to construct a diagnostic logic with high
combination must be selected, and the most important judgment accuracy, and the actual equipment
key point is the setting of certain thresholds for normal/ deterioration levels must be regularly investigated to
cautious/abnormal levels. Threshold is an important accumulate information about the correlation with

図7
factor in determining the diagnosis accuracy. An improper changes in measured data. Nissan Motor will use big
threshold causes an incorrect report, where a normal data to be accumulated by the predictive/preventive
state is misjudged abnormal, or, on the contrary, report maintenance platform explained in Subsection 2.3 of this
loss, where even an abnormal state cannot be judged chapter to improve the diagnosis accuracy continuously
abnormal and an alarm report is missed, leading to a in the future.
state in which it is impossible to attempt to prevent
abnormality.
The difficulty levels in the threshold settings were
roughly classified into three types. The first is a case in
which the absolute value of an abnormal judgment
threshold is predetermined based on standards,
equipment specifications, and other factors, so that the
threshold can easily be set.
In the second case, although no threshold is determined
based on standards etc., the correlation between a change
in the measured data and the equipment deterioration
condition is clear. If many samples of data during
equipment abnormality exist, thresholds can be
tentatively set by comparing the data in an abnormal Fig.7 Digitalization of linear-axis torque waveform in machining
state with data in a normal state (Fig.6). However, data center
in an abnormal state do not exist in many cases. In such
NICS注：画面内は英訳不要
cases, thresholds are tentatively set by statistically
processing the data in a normal state. In any case, the 2.5 Technology for visualizing tool exchange
thresholds must be verified by each equipment unit and/ schedules
or section, which requires considerable time and effort. Next, the technology for visualizing tool exchange
In addition, abnormal data samples are fundamentally schedules in the powertrain machining line is introduced.
few, therefore, thresholds must be continuously fine- To prevent a stop, stop factors must be predicted, so
tuned, and the occurrence statuses of incorrect reports that product quality information (measured values,
and report losses must be monitored even after thresholds trends), machining condition information (machining
are tentatively set. times, cutting tool use counts), and equipment status
information (main shaft vibration values, cutting fluid
Abnormal data
(before replacement of
Normal data
(after replacement of
flow rate/pressure, etc.) can be collected to monitor the
reduction gears) reduction gears) status. Here, the technology for visualizing tool exchange
Max. value in normal state schedules will be explained.
Threshold (tentative) For the exchanges of cutting tools used in machining
Acceleration (G)

Acceleration (G)

lines, Nissan Motor previously utilized signal towers and

Max. value in abnormal state

in-line large-sized displays to indicate tool exchange

notices to provide information to in-line operators (Fig.8).
Therefore, actions were taken each time a tool was
exchanged. If multiple pieces of equipment required tool
Time (s) Time (s)
exchanges, delayed actions could cause equipment stop
loss.
Fig.6 Vibration data before and after replacement of abnormal
robot reduction gears The system developed in this study collects information

40 No.88 (2022) NISSAN TECHNICAL REVIEW

 SpecialFeature1:NissanIntelligentFactoryStartingfromARIYA-6.ImprovementoftheEquipmentOperationRateandProcessCapacitybyImplementingPredictive/PreventiveMaintenanceandRemoteMaintenanceTechnology

from tool counters on the equipment in real time, which contained information such as maintenance manuals,
permits tool exchange hours on a particular day to be equipment design drawings, and descriptions of previous

Fig. 10
predicted (Fig.9). The visualization of tool exchange problems aggregated in an integrated fashion. This
schedules enables minimization of equipment stop loss system ensures that after an equipment abnormality
caused by exchanging tools through the batch exchanges occurs, skilled maintenance staff can understand the
of multiple tools and/or operator relief actions. situation and analyze the factors until the site
maintenance staff reach the site.
The system explained here is a small part of the IoT
system. System updates should be promoted to minimize
the production loss by the managing equipment for
predictive/preventive maintenance and visualization of
the line operation status.

Tool counter
Signal tower
9
Notice by
flashing
Fig.10 Centralized monitoring room

Fig.8 Existing tool exchange notice

As digital tools for connecting in real time between
skilled maintenance staff in the centralized monitoring
room and all site maintenance staff, 64 sets of wearable
cameras, headsets, and tablet terminals were provided,
enabling skilled maintenance staff and site maintenance

Fig. 11
Fig.9 IoT tool exchange schedule screen
staff to share site images, communicate by voice, and
share maintenance manuals and other information in
real time, providing a system to compensate for the
insufficient knowledge of site maintenance staff (Fig.11).
This remote maintenance system with a centralized
3. Remote maintenance monitoring room can expect to reduce the failure recovery
NICS注：画面内は英訳不要 period by an average of 30%.
In order to realize the last of the three maintenance
principles defined by Nissan Motor, fix a failure Site maintenance
staff
IoT
immediately, in an environment in which the number of
Centralized control
room network

skilled maintenance staff members is decreasing, Nissan

Motor established a centralized monitoring room in the
NIF to which skilled maintenance staff members were Skilled staff

assigned and introduced a remote maintenance system in

which digital tools are used to connect the centralized
Operation Failure Maintenance
Web camera Equipment
monitoring prediction control

monitoring room and maintenance sites.

In the background, as the result of analyzing the Manuals/ Equipment
Control

previous MBDs, the main factors in long recovery periods

Control screen circuit know-hows record

include insufficient grasp of the situation and analysis Fig.11 Remote maintenance
of factors and recovery work errors , which are related
to insufficient investigation time and insufficient
knowledge/skill of maintenance staff . The objective of 4. Conclusion
introducing remote maintenance is to resolve these two
factors. This report described the ability of the NIF to deliver
The centralized monitoring room had 27 large monitors high quality vehicles to customers on time by using
for displaying theequipment operation monitors, digital technologies to establish predictive/preventive
predictive/preventive maintenance platform, and images maintenance and remote maintenance systems to reduce
of production site cameras (243 units) for skilled equipment stop frequencies and hours in a strict
maintenance staff to grasp the operation status of the environment in which the number of skilled maintenance
plant (Fig.10). This approach enabled skilled maintenance staff is decreasing, and advanced maintenance skills are
staff to identify production equipment abnormalities in required due to the appearance of advanced equipment
real time. In addition, the centralized monitoring room and equipment aging.

NISSAN TECHNICAL REVIEW No.88 (2022) 41

SpecialFeature1:NissanIntelligentFactoryStartingfromARIYA-6.ImprovementoftheEquipmentOperationRateandProcessCapacitybyImplementingPredictive/PreventiveMaintenanceandRemoteMaintenanceTechnology

Nissan Motor will introduce the systems described in

this report into all production bases worldwide. In
addition, the remaining life prediction and automated
troubleshooting assistance of production equipment will
be achieved by learning data to be accumulated in the
future. Moreover, trouble-proof equipment with high
quality will be designed by feeding back knowledge to
accumulate in production equipment design. As a result,
the targets of the NIF are tominimize the equipment stop
frequencies and recovery time, and to deliver advanced
quality vehicles to customers worldwide with no delay.
Working style diversification and extension of the scope
of workers (working-style reforms) in an aging society
with a declining birthrate should be advanced by
promoting digitalization and innovating into maintenance
operations with lower reliance on individual skills.

Authors

Katahiro Ogawa Satoru Sakurai Koichi Makino Takuya Tsumoto Susumu Abe Masayoshi
Hayakawa

Kazunari Seo Minoru Kono

42 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1

Nissan Intelligent Factory Starting from ARIYA

7. Development of Die-Cast Technology for Realizing

Innovative Thin-Walled Light-Weight Cylinder Head

Yuuta Sugiyama* Shinichi Tsuchiya** Hiromichi Kume*

Takahiro Nishi** Machi Tabuchi**

minute internal cooling structure and thin-walled

1. Introduction products.

Global-scale activities for the reduction of CO2

emissions are being conducted at an accelerating rate. In
Japan, the declaration of aiming to achieve carbon
neutrality (a decarbonized society) by 2050 was presented
at the 203rd extraordinary Diet session in October 2020.
The automotive industry is required to reduce CO2
emissions by increasing production of electrified vehicles
as well as by achieving more efficient conventional Fig.1 Cylinder head
internal combustion engines and lighter vehicle weights.
It must reduce CO2 emissions from production plants, or Conventional part Developed part
vehicle production processes.
To achieve more efficient combustion in internal
combustion engines, a higher compression ratio in the
engine cylinder by a supercharger is required. To realize
this characteristic, a high cooling performance for
exhaust ports is necessary, so that the internal cooling
circuit spreading through an engine cylinder head must
be thinner and more minute. Therefore, the structure of Fig.2 Core
a water jacket in a cylindrical head is complex.
Simultaneously, a thin-walled cylinder head is required
for vehicle weight reduction to achieve better fuel 2. Outline of alliance standard line
economy. The development of a die-cast method that can
balance these two issues at a high level is urgently needed To achieve thin-walled lightweight cylinder heads, the
(Fig.1). required technologies were integrated and standardized
Also regarding production processes, a cylinder head as the alliance standard line.
raw material is produced through an aluminum die-cast To improve engine competitiveness, a cylinder head is
method called a low-pressure die-cast method, where a an important functional part because it manages intake
considerable amount of energy is used in processes for air, combustion, and exhaust. In order to realize these
melting aluminum and heat treatment processes for functions, an aluminum die-cast process, which produces
controlling strength. Consequently, these processes are cylinder heads, uses dies and cores, which form internal
the third highest CO2 emitting processes, following the hollow areas (Fig.2) and pours melted aluminum into the
paint and iron processes, among the production processes dies to form a part.
of Nissan Motor as of 2019. Therefore, CO2 emission To manufacture complex cooling paths, integrated parts,
reduction of the entire process is urgently needed. and thin-walled shapes with high quality, a new method
Towards this objective, a new cylinder head production and various types of automation for minimizing variations
line was developed along with the development and were incorporated into the alliance standard line.
global deployment of new-generation variable The flow of this process is shown in Fig.3. Above all,
compression ratio gasoline engines. the main new technologies for thin-wall applications are
This report will explain the details of the vacuum low- the vacuum low-pressure die-cast method, inorganic
pressure die-cast (V-LPDC) process and related cores, rotary heat-treatment furnaces, and automated
technologies developed for the compatibility of a complex inspections with computerized tomography (CT). In these

*Manufacturing and Engineering Department No.2 **Materials Process Engineering Department

NISSAN TECHNICAL REVIEW No.88 (2022) 43

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 7. Development of Die-Cast Technology for Realizing Innovative Thin-Walled Light-Weight Cylinder Head

processes, one piece is fed at a time, and error/defect As shown in Fig.5, in a cylinder head die cast, the
occurrence is eliminated by improving the automation shape is cast by pouring melted aluminum into a cavity
level from manufacturing to transport. in the die to replace the air with aluminum. Then, the air
From the perspective of CO2 reduction, the application exits structurally through gaps between the parting line
of inorganic cores, Independent melting furnace for each of the die. To prevent the liquid aluminum from flowing,
casting machine, and rotary heat treatment furnaces the back pressure in the air exit points must be decreased.
contributes to carbon neutrality. If the walls are made as thin as possible, this back
Improvement of the worker environment was also pressure causes the aluminum to solidify before the air is
taken into account. A conventional foundry is dark, foul- replaced with aluminum, and a product shape defect
smelling, and dirty due to smoke, tar, and dust and also occurs. Therefore, technology is necessary to remove the
requires considerable handwork, which cannot be air from between the die parting line to reduce the back
regarded as a good factory from an environmental pressure (Fig.6).
perspective. This new line employs technologies such as In this case, balancing the aluminum feeding pressure
inorganic cores to improve the above aspects remarkably. and pressure inside of cavity is important. Accordingly, a
sensor was used to monitor the pressure inside of cavity
Die-cast
machine
constantly (Fig.7), and a thermocouple was used to
Melting acquire the data of the aluminum flow inside of cavity.
Using die-cast computer-aided engineering (CAE) (Fig.8)
and these data to set the optimal suction conditions,
technology that allows cylinder head die-casting with a
minimum wall thickness of 2 mm was developed. In mass
Product cooling Finish
Core Core (sand removal, cutting)
conveyance
production, a weight approximately 4% lighter than that
in the conventional model was achieved, while the engine
AGV conveyance strength requirements were taken into account.
Inline CT Heat treatment

Aluminum filling pressure

Pressure, kPa

Visual appearance Transport to machining

inspection Palletization line

Fig.3 Process flow (outline)

Time, s
pressure, kPa
Vacuumed

In-mold suction
pressure

Fig.6 In-die suction conditions

Fig.4 Layout (air view)

Cavity pressure
3. Development contents
Aluminum filling

3.1 V-LPDC method

The main technology for realizing an innovative thin-
Pressure, kPa

図walled cylinder head is the V-LPDC method.

Time, s
Fig.7 Cavity pressure

Pressure

Fig.5 V-LPDC method

44 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 7. Development of Die-Cast Technology for Realizing Innovative Thin-Walled Light-Weight Cylinder Head
図

Fig.8 Die-cast CAE

Burring finish robot
In addition, this vacuum low-pressure die-cast method
has an issue in which the shell mold method or a core
method using a conventional organic bond causes the Core removal
Shooting head
suction paths to be blocked by smoke generated during
die-casting and/or tar accumulated on the die.
Consequently, the V-LPDC method employs the inorganic
Core box
core technique, in which no smoke or tar is generated
(Fig.9). This method uses a liquid glass-based inorganic
binder for core making. In this study, the alliance Fig.10 Inorganic core making machine
standard line employed these technologies to combine the
vacuum low-pressure die-cast method and inorganic
cores to realize the mass production of thin-walled Inorganic core making (Fig.10) is a method in which
lightweight cylinder heads. mixed sand is hardened by removing water. Therefore,
changes in the temperature and humidity of the
atmosphere around the core making machine affect the
hardening speed of the sand in the shooting head. The
inside of the shooting head was water-cooled to prevent
the sand in the shooting head from drying. However, a
change in the dew point due to temperature and/or
humidity causes the water to condense inside and outside
the shooting head; hence, too much water is supplied to
the sand, degrading the core quality. Therefore, the
following dew-condensation preventive measures were
taken: measuring the temperature and humidity around
the shooting head constantly, calculating and monitoring
the dew point constantly, and controlling the temperature
of the coolant.
Fig.9 Core
Because an inorganic core is soft immediately after
being molded, grasping it is difficult. The cores were
3.2 Full automation of core and die-cast processes図 handled by scooping them up from underneath, removal
As mentioned above, to reduce variations in quality, from the mold, and loading on the conveyance rack
improve the work environment, etc., the processes of (Fig.11). This approach permitted soft cores to be
melting, core making, die-casting, finishing, and automatically handled immediately after being molded,
conveyance in this line were automated as much as without breaking.
possible.
For core making, in order to automate the inorganic
core making method fully, the following issues existed:
natural sand hardening in the shooting head, handling in
a soft state just after shooting, and burring finish by the
robot.

Fig.11 Core grasping tool

NISSAN TECHNICAL REVIEW No.88 (2022) 45

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 7. Development of Die-Cast Technology for Realizing Innovative Thin-Walled Light-Weight Cylinder Head

図
A robot is used to burring-finish a core after it is
molded. However, owing to the fragility of an inorganic
core and its thinness, the core is easily broken during
burring-finishing. Accordingly, measures against core
breakage are taken by revising the positioning block
shape according to the core on the support during burring-
finishing and clearance settings to minimize core
movement during burring-finishing.
The finished cores are loaded onto the conveyance
rack, which is towed by an automated conveyance vehicle
(Fig.12) and conveyed to the core inspection area. If a
floor condition causes even a small impact to be applied
to a core being conveyed, the core can break easily; hence,
springs are placed under the conveyance rack and absorb
Fig.13 Core placement jig
impacts, enabling thin and fragile cores to be auto-
conveyed even across an uneven floor.
図

Fig.12 Automated core conveyance vehicle

The die-cast process uses a robot to automate the work

of putting the cores in a die and removing a product from Fig.14 Photograph of die-cast machine during core placement
a die.
Because cores are fragile and thin, there are concerns 3.3 Energy loss reduction in melting and heat
about dropped sand particles causing sand-inclusion treatment processes
defects and die-casting with a broken core placed in a die In the aluminum die-casting process, melting and heat
causing core breakage. These defects were minimized by treatment processes account for 35% of CO2 emissions.
designing the core print to the optimal shape to handle As activities for carbon neutrality are being conducted at
and grasp the part (Fig.13). A robot was also used to an accelerating rate, designing production equipment
automate the work of air blowing to remove any sand considering energy savings leading to CO2 reduction is
particles after the cores were placed to allow core required. This line also employs small modularized
placement in the die without breakage, and any residual equipment and superior heat efficiency to achieve energy
sand particles on the die surface after core placement savings. The relevant processes are introduced below:
were removed, which allowed the die-cast process to be
automated (Fig.14).
An automated inspection method with CT is applied to
inspect hollows composed by cores and assures all
complicated and minute internal shapes.

46 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 7. Development of Die-Cast Technology for Realizing Innovative Thin-Walled Light-Weight Cylinder Head

<Melting process> basket is also deformed by repeated heating and rapid

A conventional melting process is a system in which a quenching; thus, periodic maintenance is essential. To
large melting furnace with a melting capacity of 3–4 resolve these issues, a rotary-type rotation heat-
ton/h is placed in each plant, and a forklift with a ladle is treatment furnace (Fig.16) was applied. All processes
used to convey and distribute liquid aluminum to many from workpiece feeding to taking-out were automated
low-pressure die-cast machines in order. The melting using a robot to handle workpieces between processes
furnace is so large that the molten metal must be (solution treatment, water quenching, aging, and air
maintained with a burner regardless of its operation cooling). By feeding workpieces one by one, it is possible
level, which causes energy loss and safety risks in molten to reduce the variation in residual stress inside the
metal conveyance/distribution. This line employed a product that occurs in the heat treatment process, which
system in which a small melting furnace with a minimum contributes to the thin wall thickness of the product. Its
necessary capacity for one low-pressure die-cast machine production capacity was set to a small module equivalent
(Fig.15) was placed next to each low-pressure die-cast to two low-pressure die-cast machines.
machine, and a robot was used to distribute the molten Heat treatment without baskets eliminated heat loss,
metal to the holding furnace of the low-pressure die-cast and thus the heat efficiency was improved by
machine. Workers must simply put the aluminum ingot approximately 44%. The application of the small module
in place; hence, process automation from material feeding enabled the operation to be turned on and off according to
to melting/molten metal distribution, which is difficult the operation level, greatly reducing the energy loss.
for a conventional large furnace, is realized. 図 This
automation eliminates the risk associated with workers
conveying molten metal, thereby improving safety and
the environment. In addition, the melting furnace can be
stopped according to the operation level of the low-
pressure die-cast machine, reducing energy loss.
図

Fig.16 Heat treatment furnace appearance

4. Conclusion

A new die-cast technique called the V-LPDC method

Fig.15 Appearance of Independent melting furnace for each was developed and applied to a new project, which
casting machine
contributed to a cylinder-head product weight reduction
of 4%. In addition, the new alliance standard line, which
<Heat treatment process> realizes more attractive products, could be established as
Cylinder head die-casting involves a heat treatment a production plant line, whose application to actual
process to improve the mechanical properties of the production reduced CO2 emissions by 24%.
products. Conventionally, a continuous-type heat By continuing further development as well as
treatment furnace is applied, and similar to the melting spreading these technologies globally to maximize their
process mentioned above, several large heat treatment effect, die-cast technologies will provide highly functional
furnaces are placed in each plant; a stainless-steel basket, and high-quality parts in the future to deliver attractive
called a heat treatment basket, is filled with workpieces; products to customers.
and the basket is conveyed through a solution treatment
process (approximately 500 C), water quenching process
(approximately 80 C), and aging process (approximately
200 C). The basket used for conveyance in this continuous-
type heat treatment furnace is approximately twice the
weight of the workpiece. The workpieces and basket used
for conveyance are heated to similar temperatures,
causing the heat loss to be enormous. At a low operation
level, as the number of workpieces filling the basket
decreases, a lower heat efficiency is inevitable. The

NISSAN TECHNICAL REVIEW No.88 (2022) 47

Special Feature 1 : Nissan Intelligent Factory Starting from ARIYA - 7. Development of Die-Cast Technology for Realizing Innovative Thin-Walled Light-Weight Cylinder Head

Authors

Yuuta Sugiyama Shinichi Tsuchiya Hiromichi Kume Takahiro Nishi Machi Tabuchi

48 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2

Next-generation flagship: ARIYA

1. ARIYA: new generation of flagship EV

Eiichi Akashi* Hikaru Nakajima**

performance. In addition, ProPILOT 2.0 was enhanced

1. Introduction and a new remote-auto-park function was introduced.
The advancement of each of these technologies is
Nissan Motor Co., Ltd. commenced mass production of explained in detail in the next chapter and brief
EVs in 2010, globally ahead of all other companies. Based explanations are provided below.
on over 10 years of market experience, Nissan has
developed a new generation of EV: the ARIYA (Fig. 1). 3. Innovative packaging with EV-specific
ARIYA embodies the new Nissan Automotive and Nissan platform
Intelligent Mobility with the aim of being a new
generation of flagship EV that feels like the car of the In most conventional EVs, the platform designs are
future. With ARIYA, Nissan s goal is to provide based on internal combustion engine (ICE) vehicles. As
customers a blissful experience that only an EV can such, they do not fully benefit from the features of small
deliver. powertrains that are unique to EVs. To overcome this
problem, a new EV-specific package was developed
through a high-level combination of an EV-specific
powertrain and mounting technologies to increase the
interior space, reduce overhang at the front, widen the
wheel base, reduce the minimum turning radius, achieve
more stable control, improve driving comfort, extend the
driving range, and achieve a quieter interior.
The means of achieving these individual aims are often
conflicting. Hence, technological development of complex
systems was achieved by integrating the following high-
level elements.

3.1 Basic structure, interior space, and styling

Fig.1 ARIYA To increase the height of the interior space, a holistic
review of the underfloor structure was performed for
battery storage. It was determined that the cross
2. Toward a new generation of flagship EV members be embedded in the structure such that the
necessary battery capacity (explained later) is secured
ARIYA was developed to satisfy customers through while providing temperature control system without
integration of the latest technologies while fully utilizing
the features of an EV. To simultaneously realize many
values in ARIYA, new technological elements were
developed and their integration was achieved at a high
level using innovative system technology.
To start with, an innovative package was introduced
through a new EV-specific platform to realize a spacious
and noiseless interior, sufficient driving range, and
stylish design. The infotainment technologies were
revolutionized to offer abundant services and amazing
usability, along with integrated vehicle control Fig. 2 Subfloor structure
Fig.2 Subfloor structure
technologies to achieve high-power and stable driving

*Corporate Vice President **Nissan Product Development Department No.1

NISSAN TECHNICAL REVIEW No.88 (2022) 49

 Special Feature2 : Next-generation flagship: ARIYA - 1. ARIYA: new generation of flagship EV

compromising the strength and rigidity of the vehicle 3.2.1 Battery capacity
body (Fig. 2). The new structure enabled development of The driving ranges were set to B6- and B9-grade cars
a body-integrated battery package with high rigidity, to allow a breadth of selection according to customer
which allowed the mounting of an ultrathin large- usage. The B6 grade is targeted for daily use, including
capacity battery. Furthermore, the torsional rigidity of commuting and shopping, whereas the B9 grade is
the entire vehicle body was improved by connecting the targeted for weekend leisure activities, such as golf and
high-rigidity package to the front and rear parts of the camping. The battery capacities required for the expected
body. Because the center of gravity was lowered by the driving ranges were estimated from data obtained during
subfloor position of the battery with the weight technological development, as explained below, and
distribution being 50:50 between the front and rear of the determined to be 66 and 91 kWh for the B6 and B9
vehicle, handling and ride comfort were optimized grades, respectively.
simultaneously.
Based on the reduced size of the power train, an air- 3.2.2 Reduced air resistance
conditioning system (HVAC) was placed in the motor To reduce the air-resistance coefficient (Cd value), the
room instead of the inside of the cabin. Because HVACs roof shape was streamlined and an air curtain (a type of
have large internal spaces in their structures, when duct) was placed over the front mask to control airflow. In
placed in a motor room, they act as shock absorbers in the addition, a flat and full undercover was placed underfloor
event of a crash. This enabled increasing the interior to control airflow (Fig. 4). Thus, the Cd value was reduced
space while maintaining crash performance standards to 0.3 or less, which is outstanding aerodynamic
(Fig. 3). The additional interior space was used to increase performance for an SUV.
the tire turning angle and thereby achieve a minimum
turning radius equivalent to that of a B-segment car,
despite the large cabin space of that of a D-segment car.

Fig. 3 Layout of power-train and HVAC

Fig.3 Layout of powertrain and HVAC
Fig. 4 Air flow
Fig.4 Air flow

The interior cabin length was maximized by introducing

an electric center console and minimizing the installment 3.2.3 Quick charging performance
panel. The front and rear leg rooms were expanded owing Considering the quick charging conditions at the
to the flattened floor. This platform enabled increasing parking bays of highways, quick charging performance of
the effective length of the cabin interior to that of the 30 min was achieved for recovery from 10 to 80%.
D-segment level, whereas the overall body length was in Charging speed is strongly dependent on the internal
the C-segment. Because of this platform, a short overhang temperature of the battery. Therefore, a liquid battery
and long wheelbase layout was possible for realization of temperature-control system was developed to achieve
a radical yet attractive design. The revolutionary interior stable charging performance using thin batteries without
and appearance are the major attractions of ARIYA that the effects of air temperature and driving mode (Fig. 5).
were enabled by the new EV-specific platform. Based on expected future requirements, the allowable
input power of the on-board charging device was set to
3.2 Long driving range 130 kW during development.
To motivate more customers to purchase ARIYA and to
increase EV usage, the basic performance of the EV
including driving range, battery charge performance,
power consumption for heating/cooling, and usability in
real situations must be more advanced.
In particular, advancements were made to reduce the
general driving resistance and air resistance, which is
the major cause of driving resistance, in addition to
enhancing quick charging performance that is critical to
highway driving in summer, heating performance, and
battery low-temperature tolerance in winter.

50 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2 : Next-generation flagship: ARIYA - 1. ARIYA: new generation of flagship EV

Fig. 6 Display swipe

Fig.6 Display swipe

To ensure flexibility and rapid development of the

connected service system underpinning these functions,
Fig. 5 Temperature
Fig.5 Temperature controlcontrol
system system the service server was separated from the system server
and key function-integrating software was developed
in-house. Simultaneously, firmware-over-the-air (FOTA)
3.2.4 Heating performance technology was introduced to update the on-board
Next to driving the most energy consumed by an EV is firmware wirelessly to enable continued offering of
for heating in winter, when the air temperature difference abundant services in a user-friendly and timely supply.
between the inside and outside is larger than that in
summer. Hence, efforts were made to increase driving 5. e-4ORCE: Innovation for integrated
range in winter by adopting highly efficient heat pump vehicle control
system to reduce energy consumption.
The highly desired 4WD was installed in ARIYA to
3.3 Quietness offer additional value beyond ordinary 4WD functionality
EVs do not generate engine noise. The main sources of by integrating vehicle and powertrain controls. High-
noise in an EV are the road noise between the road and level integration of the following two advanced
tires, wind noise during high-speed driving, and motor technologies was achieved in ARIYA. Namely, 4WD
noise. Road noise was reduced by using special tires control technology used in GTR and ATTESA that
containing noise-absorption materials, adapting double controls torque distribution between the front and rear
antivibration subframes, and thoroughly strengthening axles in accordance with driving conditions, and the
the rigidities of the body frame members. Wind noise was motor-torque control technology used in LEAF for fast
reduced by eliminating uneven surfaces and gaps to and high-precision control of the motor torque in
avoid generating turbulence. External noise was reduced milliseconds. This combination enables stable vehicle
by adapting sound-insulation glasses. At low speeds, behavior under a variety of road conditions (snow, ice,
when the load noise is small, a slight rotational motion or rain, and dry), thereby providing a confident driving
radial vibration of motor may be noticeable. A newly- experience to customers.
developed motor with excellent vibrational properties at The two motors placed at the front and back seamlessly
low speeds enables realization of a quiet interior space produce flexible torque distribution between the front
that is appropriate for a flagship EV.(*) and rear by integrating driving information, such as
acceleration, yaw rate, tire slip recognition, and
4. Innovative infotainment technologies conventional driving-force control. Vehicle behaviors can
be controlled with high precision by integrating brake
When using conventional on-board operation display control of the four wheels to optimize the load weight on
devices and connected systems, the operations and each wheel. For example, winding roads can be navigated
display become complicated with increasing numbers of smoothly by independent control of the front and rear
services and functions, and the menu selections become torque. At the entrance of a curve, more torque is
wider and deeper. The addition of new functions in allocated to the rear, similar to an Front engine and Rear
complicated systems is time consuming, which makes it drive vehicle, to improve the turn-in ability. In the middle
difficult to provide timely services. To overcome this of the curve, the torque is distributed in a 50:50 ratio
challenge and minimize driver distraction, a new input between the front and rear to maintain smooth movement.
system with high operability was introduced in ARIYA to At the curve exit, more torque is allocated to the front,
allow the integrated use of two large displays comprising similar to an Front engine and Front drive vehicle (Fig.
an IVI panel and a meter panel while reducing the 7). In addition, the front and rear distributions of the
operational hierarchy. Additionally, a new interface was regenerative brakes were optimized by the motors to
developed for easy operation of the display device, where provide a smooth and comfortable ride. Hence, e-4ORCE
the two display panels could be used together via swipe is an outstanding system that offers driver secure feel,
actions (Fig. 6). Furthermore, a human-machine interface comfort, and pleasure, in addition to excellent acceleration
( HMI ) with high usability was realized through a voice performance.
recognition system that allows control of NAVI or air
conditioning settings.

NISSAN TECHNICAL REVIEW No.88 (2022) 51

 図８ Special Feature2 : Next-generation flagship: ARIYA - 1. ARIYA: new generation of flagship EV

Car in an understeer
condition

Driving force

Braking force

Friction circle
Tire-grip force

Fig.7 Operation of e-4ORCE

Fig. 7 Operation of e-4ORCE

6. ProPILOT 2.0: Remote Auto Park

ProPILOT 2.0, which is installed in Skyline, has

garnered much praise as a novel human–vehicle
relationship that offers convenience, secure feel and
comfort, along with being a buddy or partner. Its Fig.8 Remote Auto Park
functions were further enhanced and the vehicle location
accuracy was increased using signals from the quasi-
zenith satellite system. 7. Summary
In the past, there were system limitations in
determining the current lane of the car on multilane Although ARIYA is a C-segment SUV, the newly
highways. When using quasi-zenith satellite signals, the developed innovative system technologies offer enough
driving lanes can be precisely located by detecting vehicle interior space as a roomy D-segment vehicle and noiseless
positions within an accuracy of 50 cm. This significantly comfort as an E-segment vehicle, which is appropriate for
increases navigation accuracy, thus enabling smooth and a flagship EV. The maximum driving range of ARIYA
accurate guidance for selecting and changing driving was 610 km in the WLTC mode. ARIYA offers user-
lanes at highway junctions. To further assist customer friendly infotainment and abundant services through an
driving, a new technology called Remote Auto Park (RAP) integrated interface display and a connected service
was made available. Using this function, ARIYA can be system that allows flexible information manipulation.
remotely navigated forward and backward using a Owing to e-4ORCE, ARIYA has outstanding drivability
designated remote-control key (Fig. 8). Hence, it is for all occasions, from normal driving to snowy conditions.
possible to park ARIYA in a narrow space by first moving The following chapters discuss several technological
it forward using the RAP function for securing enough challenges that were not explained in this chapter.
space to open the doors. ARIYA is equipped with the latest technologies of
Nissan and is a icon of Nissan Intelligent Mobility.
ARIYA is the next-generation flagship EV that represents
the cars of the future. Nissan wishes that our passion will
resonate with all users of ARIYA

Where required by law, the EV may be additionally

equipped with an alert system that generates the
Authors necessary sounds and signals so that pedestrians can
detect and recognize the vehicle.

Eiichi Akashi Hikaru Nakajima

52 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2

Next-generation flagship: ARIYA

2. Development of EV-specific platform CMF-EV for high

packaging efficiency

Masahiro Oonishi*

1. Introduction

The Nissan LEAF, which is a full-scale mass-produced

EV released in 2010, has been accepted by customers for
its high reliability, safety features, and comfortable
driving experience that is typically offered by EVs. LEAF
has pioneered the widespread use of EVs.
The high reliability and safety features developed for
LEAF have been inherited and further advanced in the
newly developed EV-specific platform (CMF-EV) for
ARIYA. The target issues of vehicle packaging, driving
range, and driving performance have been set to exceed Fig. 1 Comparison of overall vehicle and effective
Fig.1 Comparison of overall vehicle and effective interior
interior lengths between SUVs
Nissan Confidential C

the expectations of our customers. lengths between SUVs

This chapter describes the key points for realizing a
large interior space, mounting a large-capacity battery, In LEAF, it was not possible to meet the customers
and achieving high driving performance. demand for a completely flat floor because the brake pipe,
cooling water pipe, and high-voltage wiring harness were
2. Current situation of EV packaging placed within the center tunnel.

Our customers expectations for EV packaging can 3. Evolution of the novel EV-specific
largely be classified under the following two features: platform CMF-EV
[1] Interior space that is large in the front-rear direction
owing to the compact powertrain The CMF-EV adopted in ARIYA has a breakthrough
[2] Completely flat interior floor from eliminating the feature, wherein the air-conditioning unit is mounted in
exhaust system and propeller shaft the moter room and has a higher packaging density
However, developing an EV that meets both owing to functional integration. Thus, both a completely
expectations is more difficult than the customers can flat floor and space efficiency (in the front-rear direction)
imagine. exceeding those of the top benchmarked SUV ICE
A comparison of the overall vehicle length and effective vehicles were realized (Figs. 1 and 2).
interior length (distance from the accelerator pedal to hip
point of a passenger in the rear seat, which is a representative
index of the interior roominess in the front-rear direction)
in different commercially available SUVs is shown in Fig. 1.
It is seen that the benchmark lines of SUV EVs are inferior
in terms of packaging efficiency in the front-rear direction
compared to SUV ICE vehicles. If an EV is mounted with a
high-capacity and high-voltage battery to secure sufficient Flat floor
driving range, the vehicle mass may increase significantly. Flat floor
Therefore, compared to ICE vehicles, the required crush C-seg No.1
roominess
stroke is higher for damage control during low-speed
collisions and for securing passenger and high-voltage
safeties during high-speed collisions, thereby offsetting the Fig. 2 Realization of high space efficiency and
Fig.2 Realization of high space efficiency and completely flat
completely flat floorfloor
greater space allowed by the compact drive system.
Nissan Confident

*Nissan Product Development Department No.1

NISSAN TECHNICAL REVIEW No.88 (2022) 53

 Special Feature2 : Next-generation flagship: ARIYA - 2. Development of EV-specific platform CMF-EV for high packaging efficiency

The aforementioned functional integration were For the CMF-EV adopted in ARIYA, the air-conditioning
included under the premise of stiffness of the vehicle unit was built to be crushed so that the necessary crush
body frame structure from the beginning of the planning stroke could be secured, and the front-rear length of the
stage. Consequently, precise maneuvering performance, motor room was shortened (Fig. 5). The air-conditioning
comfortable riding experience, and mass reduction, which unit primarily comprises a heat exchanger, refrigerant
are the expectations of the customers from an EV, have pipes, blower fan, and duct. Because the components other
been realized simultaneously. than the heat exchanger and blower fan motor are made of
resin or have hollow shapes, there is some allowance for
4. Layout of the air-conditioning unit in the crushing. Therefore, the Force-Stroke characteristic for
motor room collision was assigned to the air-conditioning unit when
developing ARIYA by assuming that the air-conditioning
The realization methods are described in more detail unit body will be crushed to approximately 200 mm during
from this section forward. The position of the air- a full-frontal collision. Of the air-conditioning unit
conditioning unit, which was previously situated in the components, the air PTC heater, which requires high
instrument panel of the cabin, was relocated to the motor voltage, was placed within the cabin (non-crushable zone)
room to realize a thin instrument panel. This change was in consideration of high-voltage safety during collision.
introduced during the world premiere of ARIYA and has
been well known since then. Crush Non-crushable zone
However, changing only the mounting position of the stroke

unit hardly contributes to improving the ratio of effective

Crushable HVAC
interior to overall vehicle length (i.e., packaging efficiency
in the front-rear direction). This is attributed to the fact
that as the interior space is elongated, the motor room
housing the air-conditioning unit must also be extended.
The breakthrough realized in ARIYA is that in addition
to the positional change of the air-conditioning unit,
450 mm
measures are provided to the structure so that the air- Ref) LEAF 750 mm
Air PTC heater

conditioning unit absorbs energy on impact. As shown in

Fig. 3, the air-conditioning unit was previously mounted Short moter room

at the rear side of the dash panel (i.e., in a non-crushable Fig.5 Crush space assuming crushing of the air-conditioning unit
zone). If the air-conditioning unit is relocated to the Fig. 5 Crush space assuming crushing of the air- Nissan Con

motor room and crush stroke is maintained, extending conditioning unit

the motor room is necessary to accommodate the front- A method for predictably crushing the air-conditioning
rear length of the air-conditioning unit (Fig. 4). unit during a frontal collision is described herein. Fig. 6
shows the layout diagram of the CMF-EV motor room.
Crush Non-crushable zone The air conditioning unit is mounted on the rearmost
stroke
side of the motor room. The upper powertrain unit, which
is an assembly of the battery charger, junction box, and
DC-DC converter, is mounted on the front side.

Upper powertrain Cowl-top member HVAC

Radiator core
support
HVAC in cabin

Fig.3 Relationship between previous air-conditioning unit

placement and crush space
Side view
Fig. 3 Relationship between previous air-conditioning ←Front
Nissan Confidential C
unit placement
Crush andNon-crushable
crush space zone
stroke
Fig. Fig.6 Layout
6 Layout of the
of the moter roommotor room
Extended Nissan Confidential C

During a frontal collision, the upper powertrain unit

starts retracting from the intermediate stage of the
collision event, and the maximum amount of retraction
occurs at approximately the same time as when the air-
conditioning unit is completely crushed. To achieve
HVAC in moter
room predictable crushing of the air-conditioning unit under
Fig.4 Relocation of the air-conditioning unit and extension of various types of collisions, it is desirable for the upper
the motor room powertrain unit to retract while maintaining the initial
Fig. 4 Relocation of the air-conditioning unit and extension of the moter Nissan Confidential C
room 54 No.88 (2022) NISSAN TECHNICAL REVIEW
 Special Feature2 : Next-generation flagship: ARIYA - 2. Development of EV-specific platform CMF-EV for high packaging efficiency

mounting angle (seen from the side view of the upper fixtures, and strut tower bar. Previously, the upper
powertrain unit). The following measures were powertrain unit was mounted on the member connecting
implemented to help control motion. the side members, and the air-conditioning unit was
[1] Connect the front side of the upper powertrain unit to mounted on the steering member in the cabin, as shown
the vehicle body frame structure to stabilize the in Fig. 8. An additional strut tower bar was added
retraction motion during collision separately in vehicles requiring greater maneuverability.
[2] Detach the rear fixing bracket when the upper In the CMF-EV, the functions of fixing the upper
powertrain unit retracts to ensure that the retraction powertrain and air-conditioning units as well as the strut
is not obstructed tower bar are integrated into a new cowl-top member
As a structure that realizes the first measure above, made of aluminum. Thus, the space and mass necessary
the radiator core support member connected to the front for the fixtures were reduced while increasing the
side member was affixed to the front side of the upper structural stiffness around the strut tower (Fig. 8).
powertrain unit. In addition, the allowable bracket
deformation amount and deformation modes were specified. Previous structure

To realize the second measure, a slit was introduced at

Strut bar

the bolt fastening portion of the rear fixing point to enable

the bracket to slide and come away only upon retraction
input due to collision (Fig. 7). To ensure both secure ARIYA

fixation during normal driving and sliding during Powertrain mounting member Cowl-top member

collision, the bracket stiffness, surface treatment,

dimensional accuracy, and fastening torque are strictly
controlled. By implementing the aforementioned
measures, the space of the air-conditioning unit was Powertrain, HVAC mounting & STRUT BAR
utilized as a crushable zone. Thus, sufficient crush stroke
Steering member integration
Nissan Confidential C

was secured while realizing large interior space in the Fig.8 Functional integration using the cowl-top member
front-rear direction owing to the compact powertrain , as
expected by the customers. The second is a multifunctional frame for the high-
voltage battery. Previously, the main functions of the
Upper powertrain Upper powertrain
high-voltage battery frame were to support the battery
front bracket rear bracket
A
Cowl-top member body and protect it from inputs, such as collisions and
=Fixed =Slide while crush
interference with road surfaces. While more details are
described later in the chapter on development of the
high-voltage battery, the CMF-EV uses aluminum
Upper powertrain HVAC extruded material for the battery frame. Using this
manufacturing method, a water jacket for battery
temperature control and a cooling water pipe connected
Side view
←Front

to the rear motor inverter (for 4WD) were integrated in

the cross section of the frame (Fig. 9).
Additionally, the high-voltage wiring harness connected
Fig. 7 Unit movement during frontal collision and slit structure at the to the rear motor inverter was changed to a bus bar and

A
Nissan Confidential C

fastening point
Upper powertrain rerouted within the high-voltage battery pack, and the brake
rear fixing bolt Slit pipe was rerouted in the gap between the battery pack and
side sill. Thus, removal of the tunnel was realized;
additionally, a completely flat interior floor from eliminating
the exhaust system and propeller shaft was realized in
Upper powertrain
accordance with customer expectations from an EV.
Plan view Cowl-top member
←Front Front view
←Right

Fig.7 Unit behavior during frontal collision and slit structure at Floor cross member
the fastening point Battery frame
Nissan Confidential C

5. Space saving through functional Brake pipe

integration and vehicle body stiffness
Battery module

To mount a high-capacity battery, increase the cabin

space, and realize a flat floor and high driving performance, Side sill
functions were integrated to increase packaging density.
Two major case examples are presented here. Water pipe
The first is a cowl-top member that integrates the Water jacket for battery for rear powertrain
Nissan Confidential C

upper powertrain unit fixtures, air-conditioning unit Fig.9 Integration of water pipe with high-voltage battery frame

NISSAN TECHNICAL REVIEW No.88 (2022) 55

 Special Feature2 : Next-generation flagship: ARIYA - 2. Development of EV-specific platform CMF-EV for high packaging efficiency

The high-voltage battery frame is one of the main Previous structure

structural components, and its connections with the Battery is only connected to body.
ARIYA
vehicle body frame structure and suspension members
were strengthened. As shown in Fig. 10, the hot-stamped
materials of the vehicle body floor (cross members of the
floor and high-voltage battery) were rearranged in an
alternating configuration in the front-rear direction. The
Connected to rear suspension member

Additional body fixing points

cross members of the high-voltage battery are connected

to the vehicle body side sill via the side frame and side
rail. This construction helps protect the high-voltage Connected to front suspension member

battery during side collisions and increases the stiffness Strengthening connection between suspension and high-
Fig.11 Strengthening connection between suspension and high-
voltage battery
Nissan Confidential C

of the vehicle body frame structure. voltage battery

Floor cross member

6. Summary

The CMF-EV adopted in ARIYA was developed for

potential use in Renault vehicles. Renault s core
development members, who are permanently stationed
Cross member in Thin & flat in the office at the Nissan Technical Center (Atsugi City,
HV battery case battery case Kanagawa Prefecture), exchanged ideas with us on
Fig.10 Positional relationship between vehicle body floor and various technical issues and attended thorough
high-voltage battery frame discussions to formulate technical solutions and select
Fig. 10 Positional relationship between vehicle body the best options during the developmental activities.
floor
The and high-voltage
high-voltage battery
battery frame
frame was connected with During the development stage, the environment
both the rear fastening point of the front suspension surrounding EVs
Nissan Confidential C changed drastically on a daily basis.
member and the front fastening point of the rear However, there were no changes in our initial target of
suspension member via suspension pin stays. Thus, the inheriting and furthering the high reliability and safety
lateral stiffness of the suspension member fastening developed for LEAF as well as the target exceed the
point was increased and steering responsiveness was expectations of our customers further for vehicle
improved (Fig. 11). By utilizing the high-voltage battery packaging, driving range, and driving performance.
frame as the main structural component, the cross- Although not mentioned in this chapter, we succeeded
sectional area of the vehicle body cross member was in evolving the ARYIA S performance from LEAF in
reduced and a flat floor was realized; additionally, a terms of the driving range, charging performance, and
vehicle body stiffness approximately 1.9 times higher resistance to capacity decline of the high-voltage battery
than previous C-segment SUVs was secured. Thus, a low under actual driving scenarios, such as when using the
flat floor and dynamic performance, such as ride comfort air conditioner or under varying temperature conditions.
performance, were realized simultaneously. In addition, we improved the driving experience unique
to EVs (through e-4ORCE) as well as convenience using
connectivity. Based on the developmental activities, we
are confident that ARIYA will be a popular choice for
customers not only as an EV but also as an automobile.
We are looking forward to contributing to a more
sustainable mobility by attracting and encouraging a
wide variety of customers who have previously hesitated
in purchasing EVs.

Authors

Masahiro Oonishi

56 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2

Next-generation flagship: ARIYA

3. Integrated interface display

Tetsu Obata* Takashi Eguchi* Yasutomo Sasanuma*

In recent years, smartphones have become daily tools visibility, whereas the infotainment information is
for handling information and the volume of onboard displayed at a distance conducive to high operability and
infotainment information has increased steadily with within easy reach of the driver (Fig. 2). These two
changes in usage mode and user experience (UX). information panels are placed adjacent because the
Further, with the advancements in driver-assistance human visual field is approximately 1.5 times wider
technologies, the types of information, including sensing horizontally than vertically, and eye movements are
information from the surroundings and system conditions, easier in the horizontal direction (Fig. 3).
have been changing. To reduce the burden of driving
operations under such changes in the information Meter information
environment, a new human–machine interface (HMI) Infotainment information
was developed for easy access to the required information.
This chapter describes new technologies installed in
ARIYA, namely an integrated display package for
Comfortable
combined visibility and operability to support easy Comfortable viewing distance
access to information and a structured graphical user operating distance
interface (GUI) for easy and intuitive handling of a large
variety of content.

1. Interface for combined visibility and Fig.2 Layout of the information display
operability
Fig. 2 Layout of information display
Various approaches have been considered for the
integration of the infotainment and meter displays into a
single information display area. During the development
of ARIYA, the dead angle of the steering wheel was
considered for determining the display location and its
features (Fig. 1).
~ 1.5

A: Effective visual field: The area within which

information can be instantly received by quick
eye movement
B: Stable fixation visual field: The area within
which information can be readily received by eye
and head movements.
C: Auxiliary visual field: The area within which
a visual stimulus can be recognized.

Fig.3 Characteristics of the human visual field

Fig. 3 Characteristics of human visual field

The panels were placed along an S-shaped curved

surface, which was in harmony with the futuristic design
of the cockpit interior (Fig. 4). The S-shaped display was
Fig.1 Integrated interface display manufactured as follows. First, a cover glass with
multiple curvatures was fabricated via 3D hot forming
Fig. 1 Integrated interface display technology (Fig. 5). Then, a flat liquid-crystal glass was
Meter information required for driving is displayed in bonded to the cover glass without distortion through
front of the driver at a convenient distance for clear optical bonding technology (Fig. 6).

*Connected Car and Services Engineering Department

NISSAN TECHNICAL REVIEW No.88 (2022) 57

 Special Feature2 : Next-generation flagship: ARIYA - 3. Integrated interface display

content to their tastes and needs. The tile menu can be

intuitively rearranged using drag-and-drop operations
(Fig. 10).

Fig.7 Home widgets

Infotainment display Meter display
Fig.4 S-shaped display harmonized with the interior
Fig. 7 Home widgets

Fig. 4 S-shaped display harmonized with the interior

Fig.8 Eight-tile menu

Fig. 8 8-tile menu

Fig.5 S-shaped cover glass Age
Fig. 7 Home widgets
Rate of correct answers (%)

Fig. 5 S-shaped cover glass

Flat liquid-crystal panel

Curved cover glass

Fig.6 Bonding of curved cover glass Number of

tiles in the
menu (total)
Fig. 6 Bonding of curved cover glass
2. Simple and intuitive GUI Fig.9 Ease of menu selection as a function of the number of
menu items
2.1 Infotainment information
Fig. 9 Ease of menu selection as a function of the
Because humans have limited ability for processing number of menu items
increasing volumes of information, the number and
positions of the displayed menu items were analyzed.
The widget sizes (Fig. 7) and tile menus (Fig. 8) were
enlarged and their numbers were optimized based on the
recognizability and operability of the functions.
The arrangement of the menu items was optimized
based on an evaluation of the ease of menu selection as a
function of the number of menu items for different age Fig.10 Customization of home widgets
groups. It was observed that short-term memory of a user
decreased as the number of menu items increased. A
4-column 2-raw configuration with a total of eight menu 2.2 Meter information
Fig. 10 Customization of home widgets
items showed a correct answer rate of over 90% The meter display consists of three basic zones (Fig.
irrespective of age group and was thus selected as the 11) and uses a simple mode, where the volume of
optimum configuration (Fig. 9). information in each zone is restricted such that the
In addition to basic operability, the display desired information can be found easily. For example, the
configuration allows customers to personalize a variety of power gauge, instantaneous speed, and drivable distance

58 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2 : Next-generation flagship: ARIYA - 3. Integrated interface display

are digitally displayed in separate zones, thereby

allowing easy confirmation of the conditions of the vehicle
(Fig. 12).

Fig.14 Voice interaction display

Fig. 14 Voice interaction display
Fig. 14 Voice interaction display

Fig.11 Fig.
Conventional meter display
11 Conventional meter display
Fig. 11 Conventional meter display

Fig.15 Home screen with parallax effects

Fig. 15 Home screen with parallax effects
Fig. 15 Home screen with parallax effects

Fig.12 New simple-mode display adopted in ARIYA 3. Summary

Fig. 12 New simple-mode display adopted
Fig. 12 New simple-mode
since ARIYAdisplay adopted
since ARIYA
ARIYA has been equipped with an integrated display
2.3 Information communication between the package for combined visibility and operability to enable
displays easy access to information and a structured GUI easy
The infotainment and meter displays were and intuitive handling of a large variety of content. In
synchronized via Ethernet communication and addition, ARIYA offers great emotional value to customers
information could be transferred between the two through various applications and animations. In the
displays instantly using intuitive swipe actions (Fig. 13). future, the cockpit HMI can be further developed into a
This function supports various usage modes, such as drive concierge.
moving a map to the meter display while a passenger
enjoys other content on the infotainment display.

Infotainment system Meter

Data-integration process

Encode Decode
Map data for Map data for meter
infotainment display display

To display Recognition of swipe action To display

Shifting of
map location

Infotainment display Meter display

Fig.13 Information transfer between the two displays via

Ethernet
Fig. 13 Information transfer between the two
displays via Ethernet
2.4 User experience (UX) with emotional value
To provide emotional value to the customer along with
high operability of the basic functions, new functions and
interactive systems were developed with high interactivity
and spatiality. Voice-animation expressions were used
for voice recognition because of the excellent interactivity
to enhance information recognition (Fig. 14). Home
widgets offer new UXs, such as comfortable and enjoyable
operations, by providing adequate space to tile menus
and icons with parallax effects (Fig. 15).

NISSAN TECHNICAL REVIEW No.88 (2022) 59

 Special Feature2 : Next-generation flagship: ARIYA - 3. Integrated interface display

Authors

Tetsu Obata Takashi Eguchi Yasutomo Sasanuma

60 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2

Next-generation flagship: ARIYA

4. Development of new NissanConnect

Toshiro Muramatsu* Kazuhiro Funai** Masaru Tasaki**

Shintaroh Murakami* Hiroyuki Kamishima**

1. Introduction 2. NissanConnect system

Connected services, which connect the mobile phone An overview of the NissanConnect system is shown in
network to the navigation system to provide a wide range Fig. 2. The vehicle is equipped with an in-vehicle
of services, were started in the late 1990s. Nearly a infotainment (IVI) system comprising a display and
quarter century has passed since Nissan launched its graphical user interface (GUI) as well as a telematics
NissanConnect (formerly COMPASSLINK and control unit (TCU), which communicates with the cloud.
CARWINGS) . Connections to the other electronic control units (ECUs)
Owing to technological evolution in recent years and sensors of the vehicle are via LAN. The TCU
accompanied by the widespread use of the Internet, communicates with the cloud using a mobile phone
connecting things , such as automobiles and household network. The cloud system is accessible by call centers
electrical appliances, to the Internet for better utilization and third-party information providers that offer services
(Internet of Things, IoT) has become popular. In 2019, and information to users. In addition, a dedicated app on
Nissan updated its connected car system to the a user s smartphone collaborates with the cloud system.
NissanConnect service, which provides new experiences The IVI, display, IVC, cloud system, and apps comprise
to customers by connecting the vehicle and customer the NissanConnect platform.
terminal seamlessly. The number of users of the
NissanConnect service has increased rapidly (Fig. 1). The
technologies developed through the dedicated efforts of Call center
Third-party
system

the NissanConnect development team have been fully

adopted in ARIYA to provide new and exciting experiences
Display In-vehicle Telematics Cloud system Smartphone
GUI infotainment control unit w/SW&Map OTA DB app

to customers. NissanConnect platform

This chapter describes the strengthened NissanConnect

Sensors ECUs

platform and the new voice recognition technology

developed for the platform. Fig.2 Overview of NissanConnect system
Fig. 2 Overview of NissanConnect
system

The cloud system consists of three main components:

vehicle connection/vehicle information management,
Ratio of number of subscribers

service logic/provision management, and software/map

over-the-air (OTA) distribution management. The core
parts of these components are common globally; hence,
global deployment can be achieved quickly. The
components may have portions that meet the unique
needs of different regions. The system requires both
robustness to provide stable service and flexibility to
adapt to changing customer requests and peripheral
IoTs. To realize these requirements, joint development
Fig.1 Increase of NissanConnect subscribers was conducted with Microsoft Corporation to develop
components requiring greater robustness (vehicle
connection/vehicle information management and
Fig. 1 Changes in the number of NissanConnect
software/map OTA distribution management). In-house
subscribers
development was conducted for the service logic/provision

*Connected Car and Services Engineering Department **Connected Technology Development and Service Operation Department

NISSAN TECHNICAL REVIEW No.88 (2022) 61

 Special Feature2 : Next-generation flagship: ARIYA - 4. Development of new NissanConnect

management, which requires higher flexibility, for Hello Nissan, um, hey, I want
to go to Tokyo Skytree
efficiency. Voice signal processing

Fig. 3 shows the sequential overview of various services

(In-vehicle device)
Voice data

provided by this platform. For example, for the remote

Automatic language
processing
(Off-board)

function that enables vehicle operation from an app, the 1. Phoneme

recognition
2. Word/sentence
recognition
3. Intention estimation

user input to the app is used as the trigger to establish a

AD conversion and
frame division of voice
data

communication path between the cloud server and TCU;

Comparing with the database Comparing with Nissan’s
of words/statistical language original dictionary
model
Comparing with the

subsequently, the commands are sent to the TCU. The

Okay. Do you want to acoustic model
set Tokyo Skytree as the
destination?
TCU operates the vehicle according to the received Recognition results

command and sends the results to the app via the cloud
system. In addition, for services that coordinate with Fig.4 Flow of natural language recognition processing
Fig. 4 Flow of natural language recognition processing
third-party information when the vehicle is traveling,
user input to the IVI is used as the trigger to convert the
cloud system to a proxy, which subsequently connects to 3.1 Phoneme recognition
the third-party system for service provision. Hence, The voice signal acquired by the microphone undergoes
numerous services can be provided by configuring the AD conversion and is sequentially divided into frames of
functions on the NissanConnect platform. equal lengths. Then, each frame undergoes frequency
spectrum analysis to obtain the acoustic features.
Services
ECU/
Sensor
Meter
& GUI
IVI TCU
Cloud&
SW/Map DB
Call
center
Third party App Subsequently, phonemes are recognized by comparison of
Remote control
(e.g., Air-conditioning
Action
Input
Presentation
the obtained acoustic features with the acoustic model,
which is created by statistical processing of human voice.
Smart speaker
before getting in the
Voice input
vehicle)
Action
Voice output

Remote settings
(e.g., Door-to-door Set Input
navigation)
Vehicle condition Frame division
notification
(e.g., Warning lamp
Trigger Notify Signal
notification) strength
Third party Input
Proxy

(e.g., Cooperation with Service

Google, Amazon Alexa) Presentation Phoneme recognition
Time
Trigger Frequency analysis for each frame & Frame
Operator service Input
Comparing with the acoustic model Frame n
(e.g., SOS call) Service
Call Frame 2
Frame 1
Vehicle data Extract the acoustic
collection Data Send Collect
features of frame 1
(e.g., Probe data)

OTA updating Request

OTA
(e.g., Remote SW distribution
update, map data update) Update

Fig.3 Overview of the NissanConnect service sequence

Acoustic model
Frame 1 = “t”

3. Voice recognition technology supporting Frame 1-n = “t”

Frame n + 1 = “o”

CCS Fig.5 Phoneme recognition

Fig. 5 Phoneme recognition

Voice recognition technology is a method to reduce the

burden of operating an in-vehicle device when driving. To 3.2 Word/sentence recognition
date, the mainstream voice recognition method has been Comparing the chronological connections of phonemes
a standalone method, where a voice recognition engine is with the database of words, the words in the voice signal
installed in the in-vehicle device. For ARIYA, we are obtained. In addition, a statistical language model in
developed a dictionary database used exclusively by the form of parameters obtained by learning the
Nissan in addition to the standalone method. The off- probability of shifting from one word to another is used
board method was adopted for the dictionary database; for comparison when selecting the final candidate for
thus, the database resides on a network configured on the the most probable word combination result. For
platform. The off-board method was adopted because it is example, given the input I want to go to tou . . . kyou
capable of accurately understanding the daily utterances skytree, it is possible to recognize an accurate sentence
used while driving and the intentions of such utterances. I want to go to Tokyo Skytree.
Natural language speech is acquired using a
microphone, processed via voice signal processing in the
in-vehicle device, and further processed in the following
order. Phoneme recognition, word/sentence recognition,
and intention estimation (Fig. 4).

62 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2 : Next-generation flagship: ARIYA - 4. Development of new NissanConnect

Compare the chronological connection of

Phoneme recognition results phonemes with the database
“Toukyou” = “Tokyo”
“Tou” / “Kyou” = “Tower” / “Today”

Database of words
* Checks whether the recognized connection of words are probable
as a sentence that has a meaning

Transition High
High High
probability

Tokyo Skytree to want to

go
Statistical language model High
Sum of transition
probabilities:
Low
Transition probability: Low

want
Tower Today Sky Tree to to go
Want to go to Tokyo Skytree

Fig. 6 Word/sentence recognition

Fig.6 Word/sentence recognition

3.3 Intention estimation

To estimate the intention of an utterance made during
driving, it is necessary to process and remove unnecessary
words (e.g., Um..., Hey ) and extract only the necessary
words (e.g., Tokyo Skytree + want to go). In addition,
synonyms of want to go to (e.g., go to, set the
destination at ) need to be recognized. There may also be
cases where the uttered sentences are not grammatically
correct. To separate these factors and secure sufficient
intention estimation performance in the driving
environment, Nissan s original dictionary database was
developed. The database contains learned data from
approximately 3,000 utterance patterns collected in
actual driving environments. Using this dictionary
database and narrowing the database to only the
necessary words allowed a voice recognition rate
achievement of approximately 85% for ARIYA compared
to approximately 70% for systems manufactured by other
companies.

Intention estimation processing

[1] Import sentence “Um, I want to go to Tokyo Skytree” Intention estimation dictionary
Compare (exclusively used by Nissan)
[2] Compare with the
Tokyo Skytree to want to go to Destination
intention estimation Um...
dictionary and go
Remove
perform filtering (Example)
Results want to go

[3] Intention Tokyo Skytree want to go to

estimation
Destination: Tokyo Skytree Intention: Set destination Taught additional data collected by Nissan
(utterances typically made while driving)

Fig.7 Intention estimation

Fig. 7 Intention estimation

4. Summary

We are at the beginning of an era wherein automobiles

have become a part of the Internet, and we are creating
new value using the NissanConnect platform. The voice
recognition technology described herein has the potential
to offer new experiences, such as dialog-type services,
including the virtual personal assistant (VPA). Once
again, we deeply consider the meaning and significance
of our principles What is important is the value for the
customer, Is there competitiveness in that subject?
and Only the customer knows the correct answer to
actively develop leading services for the future.

NISSAN TECHNICAL REVIEW No.88 (2022) 63

 Special Feature2 : Next-generation flagship: ARIYA - 4. Development of new NissanConnect

Authors

Toshiro Muramatsu Kazuhiro Funai Masaru Tasaki Shintaroh Murakami Hiroyuki Kamishima

64 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2

Next-generation flagship: ARIYA

5. High-capacity lithium-ion battery for ARIYA

Junji Katamura* Kentaro Hatta* Keisuke Wakabayashi*

Masanori Takagi* Naoto Todoroki*

1. Introduction 2. Technology for low and flat floor in the

cabin
Electric vehicles (EVs) have been deployed globally in
recent years, and there is demand for a longer driving As mentioned above, the newly developed EV-specific
range. The driving range depends primarily on battery platform was adopted in ARIYA to realize a flat and large
performance. Therefore, it is important to increase the interior space. This required designing a thin and flat
amount of energy within specified battery size, it is battery structure. The main technologies developed
important to increase the energy density per unit volume. newly for this requirement are described herein.
A newly developed EV-specific platform is adopted in
ARIYA to realize a flat and large interior space; thus, the 2.1 Thin cooling mechanism
requirement for the battery layout design is to fit within Typically, a temperature control system that uses
a limited space. LLC(Long life coolant) is adopted in battery EVs (BEVs).
On the other hand, from the viewpoint of improving In this system, a plate is placed at the bottom of the
convenience, there is a high demand for quick charging battery pack. The LLC is cooled using a chiller, heated
performance to ensure that a greater amount of energy using a heater, and supplied to the plate to control battery
can be stored in a short period of time. To meet these temperature.
demands, a temperature control system that appropriately
controls battery temperature is essential. However, the
battery temperature control system significantly lowers
LLC researve tank
the volume efficiency of the battery pack, and the control LLC heater

system has a large impact, particularly in the thickness

(height) direction.
The newly developed battery pack has improved LLC circulation pump LLC chiller
energy density per battery pack thickness (approximately
2.3 times that compared to the LEAF E+), even though it
is equipped with a temperature control system. Fig.1 Diagram of the battery temperature control system
Consequently, top level volume energy density and quick Fig. 1 Diagram of the battery temperature control
charging performance (for EVs) have been realized while system
improving interior comfort. This article describes the To uniformly cool/heat the battery modules arranged
development of the high-capacity lithium-ion battery within the battery, the LLC must be supplied uniformly
mounted in ARIYA. across the entire of the lower plate. However, the space
within the battery must be watertight and separate from
the LLC flow channels. Therefore, three structures,
namely the lower plate, LLC flow channel mechanism,
and protection cover, are necessary, as shown in Fig. 2. In
ARIYA, these structures are integrated to realize a thin
battery pack (Fig. 3).

B6 grade (66 kWh) B9 grade (91 kWh)

*Powertrain and EV Energy Storage Technology Department

NISSAN TECHNICAL REVIEW No.88 (2022) 65

 Special Feature2 : Next-generation flagship: ARIYA - 5. High-capacity lithium-ion battery for ARIYA

side to increase the heat-insulating property and cooling

efficiency (Fig. 5).

Lower plate
(Hollow-side face)

Fig.2 Battery cooling structure of Fig.3 ARIYA cooling Fig.5 LLC flow channel cross section
other companies structure Fig. 5 LLC flow channel cross-section

To realize a thin and flat battery layout design, the

2.2 Battery floor manufactured by aluminum battery modules are arranged in a thin and LLCwide
to layout
extrusion (Fig. 6). Therefore, the space between the battery pack
chiller
Aluminum extrusion method was adopted to peripheral frame and battery modules is narrow, which
manufacture the battery casing to integrate its lower results in very little crush allowance during collisions. To
plate, LLC flow channel mechanism, and protection solve this problem, multiple cross members are placed
cover. A cross-sectional structure was established such within the battery pack to realize high resistance against
that the LLC flowed inside the extruded lower plate. In impact.
this manner, the lower plate and cooling plate were
integrated to realize a cooling mechanism with reduced Cross member (orange parts)
thickness (Fig. 4).

LLC heater

-section

LLC to
chiller
of the
nels
LLC from Battery module (red parts)
chiller
Fig.6 ARIYA battery layout

Fig. 6 ARIYA battery layout

3. Balanced evolution of the trade-off
performance
Battery
lower plate Although our previous battery development activities
were based on balancing high-level performances, we
have improved the balance further in ARIYA. The main
battery performances that provided value to customers
Battery
Module
were compared among ARIYA, LEAF, and other existing
lower technologies; the comparison results are summarized in
plate Battery
Table 1. Balanced evolution of the battery performance
lower plate
was realized in ARIYA, which is a trade-off relationship.
LLC flow channels The representative items are described here.
Fig.4 cross section of battery lower plate with LLC flow
channels
Fig. 4 Cross-section of battery lower plate with LLC flow
channels
The cross section in which the LLC flows is structured
such that the flow channel is in contact with the upper
plate on which the battery module is mounted. A hollow
space is formed between the flow channel and lower plate

66 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2 : Next-generation flagship: ARIYA - 5. High-capacity lithium-ion battery for ARIYA

Table 1 Comparison with existing technologies

Table 1 Comparison with existing technologies
Natural heat
Temperature Thin temperature Fig. 8 shows the quick charging performances of
Value for our Battery control by fluid controlling
customers performance
dissipation
(LEAF)
(Existing
technology)
mechanism
(ARIYA)
Nissan EVs assuming a situation where the vehicles are
Interior space Battery pack BASE △ 〇 driven for long durations and mainly on highways, with
thickness
Driving range Energy BASE × 〇
repeated high-speed traveling and quick charging. In
Mass BASE × △ LEAF E+, the battery temperature increased with the
Price Number of
components
BASE × △ number of quick charges, and the charging speed tended
Structure BASE ← 〇 to decline from the second quick charge onward. However,
expansibility
Convenience Consecutive BASE ◎ ◎
the battery temperature is controlled appropriately in
ARIYA; therefore, the charging speed did not decline
charging
Low-temperature BASE ◎ 〇
charging
even with repeated high-speed traveling and quick
charging, thereby enabling comfortable long-distance
3.1 Convenience improvement driving.
The thin cooling mechanism described in the previous
section was instrumental in realizing a flat and large
interior space unique to EV-specific platforms.
Additionally, it helped realize quick charging
performances in various types of environments. Fig. 7
shows the amount of energy stored in 30 min of quick
charging at ambient temperatures of 40 C and –5 C. It
was possible to achieve excellent and quick battery
charging performance in ARIYA under various types of
environments by appropriately controlling the battery
temperature (heating during winters and cooling during
summers).

Fig.7 Quick charging performance

Fig.8 Transition in the amount of energy charged with repeated high-speed traveling and quick charging

NISSAN TECHNICAL REVIEW No.88 (2022) 67

 Special Feature2 : Next-generation flagship: ARIYA - 5. High-capacity lithium-ion battery for ARIYA

3.2 Battery structure for high expansibility and fewer 4. Summary

components
The battery floor structure fabricated by aluminum The battery mounted in ARIYA has a newly developed
extrusion enables a thinner battery and realizes high thin structure and is equipped with a temperature control
expansibility to various types of vehicle specifications. system. Thus, a flat and large interior space that is
Because the battery casing is extruded in the front-rear unique to EV-specific platforms was realized. Moreover,
direction of the vehicle body, simply changing the improved energy density per volume as well as high quick
extrusion length enables the application of the battery charging performances under various types of
casing in vehicles with different wheelbases. Thus, the environments were realized.
number of specifications was reduced. Even in the two- To promote the widespread use of EVs, we will continue
story-structured B9 grade (91 kWh battery specification, to improve the chemistry of lithium-ion batteries while
in contrast to the one-story-structured B6 grade (66 maintaining their high performance and reliability,
kWh)), where the battery modules are arranged in the improve the energy density, and develop batteries with
space below the rear seat to obtain high energy capacity, greater competitiveness.
the basic structure of the first-story portion can be
standardized (Fig. 9). As noted previously, integrating
the lower plate and LLC plate of the battery casing
enables an approximately 60 % reduction in the number
of components. These technologies can promote high
competitiveness.

Other vehicle models

Possible to be used in different vehicle body
lengths by changing the extrusion length

ARIYA

ARIYA B6 grade (66 kWh)

Commonizing
the first-story
ARIYA B9 grade (91 kWh) basic structure

Fig.expansibility
Fig.9 High 9 High expansibility to other
to other vehicle vehicle models
models

Authors

Junji Katamura Kentaro Hatta Keisuke Wakabayashi Masanori Takagi Naoto Todoroki

68 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2

Next-generation flagship: ARIYA

6. Newly developed motor to realize ARIYA s performance

Takahito Okubo* Hiroki Wada**

ARIYA
1. Introduction

One of the concepts realized in ARIYA is providing the

feeling of being in a lounge. The key points of the lounge
concept are ensuring a high level of quietude as well as
smooth acceleration and deceleration without abrupt
shocks, in addition to the design and quality of the
interior, such as the seats. This chapter describes how
these requirements were met. The main focus here is on
the motor, which has been completely redesigned for
ARIYA.

2. Overview of the newly designed motor

In ARIYA, EESM *2 was adopted instead of the IPM

(SM) *1, which was adopted in LEAF, NOTE, and SERENA
E-POWER. Although the EESM is similar to the IPM in
that three-phase AC is supplied to the stator by PWM
control, the EESM does not use permanent magnets in Rotor assembly Stator assembly
the rotor and instead uses electromagnets operated by
DC. Thus, appropriate amounts of current can be selected
ARIYA for the rotor and stator based on the driving load, thereby LEAF
leading to improved efficiency as well as reduced vibration
and noise. However, the heat generated increases owing
Fig. 1 Mo
to the supply of electric current to the rotor, and securing
torque responsiveness is an issue (as described in the
next section).
An expanded view of the EESM used in ARIYA is
shown in Fig. 1. For comparison and reference, a similar
view of the IPM of LEAF is shown. An eight-pole rotor
structure is adopted in the EESM in ARIYA. The DC
generated in the PEB *3 is supplied to the rotors via a
brush system. The amounts of DC for the rotor and three-
phase AC for the stator are calculated in the PEB
according to the torque instruction values based on the
driving force requirements. The current values are
controlled such that an optimum combination can be
achieved. Direct oil cooling is used to cool the motor.

Rotor assembly Stator assembly Rotor assembly Stator assembly

Fig.1 Motors used in ARIYA and LEAF

Fig.**Powertrain
*Powertrain and EV Engineering Division
1 Motors used in ARIYA and LEAF
and EV Electrical Technology Department

NISSAN TECHNICAL REVIEW No.88 (2022) 69

Alliance Interna
 Special Feature2 : Next-generation flagship: ARIYA - 6. Newly developed motor to realize ARIYA s performance

suppressed while the required torque is exerted. Thus,

3. Value realized by ARIYA s motor the decline in efficiency is reduced even when driving at
high speeds (Fig. 3).
3.1 Realizing quietness
Compared with combustion engine and hybrid vehicles, 300 63% 76% 82% 86% 88% 89% 91% 92% 92%

there is no generation of engine noise in a BEV *4.

62% 76% 83% 86% 88% 90% 91% 92% 93%

63% 77% 83% 87% 89% 90% 91% 92% 93%

65% 78% 84% 87% 89% 90% 92% 93% 93% 94%

64% 78% 84% 87% 89% 91% 92% 93% 93% 94%

240
Therefore, the road and wind noises are more prominent.
65% 79% 84% 88% 90% 91% 92% 93% 93% 94%

67% 79% 85% 88% 90% 91% 92% 93% 94% 94% 95%

68% 80% 85% 88% 90% 91% 93% 93% 94% 95% 95%

67% 80% 86% 88% 91% 92% 93% 93% 94% 95% 95%

68% 81% 86% 89% 90% 92% 93% 94% 94% 95% 95% 95%

Moreover, there are instances where high-frequency

69% 81% 86% 89% 91% 92% 93% 94% 94% 95% 95% 95%

70% 82% 87% 89% 91% 92% 93% 94% 94% 95% 96% 95%

180
71% 82% 87% 89% 91% 92% 93% 94% 95% 95% 95% 95% 95%

71% 82% 87% 90% 91% 93% 93% 94% 95% 95% 96% 95% 95%

noise generated by the motor and inverter are issues that

71% 83% 88% 90% 92% 93% 94% 94% 95% 95% 96% 96% 95% 95%

72% 83% 88% 90% 92% 93% 94% 94% 95% 95% 96% 96% 95% 95%

73% 84% 88% 91% 92% 93% 94% 94% 95% 95% 96% 96% 96% 95% 95%

74% 84% 88% 91% 92% 93% 94% 95% 95% 95% 96% 96% 96% 96% 95%

are never observed in combustion engine vehicles. In

74% 85% 89% 91% 92% 93% 94% 95% 95% 95% 96% 96% 96% 96% 96% 95%

75% 85% 89% 91% 93% 94% 94% 95% 95% 96% 96% 96% 96% 96% 96% 95% 95%

120
76% 85% 89% 91% 93% 94% 94% 95% 95% 96% 96% 96% 96% 96% 96% 95% 95% 95%

76% 86% 89% 92% 93% 94% 94% 95% 95% 96% 96% 96% 96% 96% 96% 96% 95% 95% 95%

ARIYA, several measures are implemented in the vehicle

77% 86% 90% 92% 93% 94% 94% 95% 95% 96% 96% 96% 96% 96% 96% 96% 95% 95% 95% 95%

77% 86% 90% 92% 93% 94% 95% 95% 95% 96% 96% 96% 96% 96% 96% 96% 96% 95% 95% 95% 94%

78% 87% 90% 92% 93% 94% 95% 95% 95% 96% 96% 96% 96% 96% 96% 96% 96% 95% 95% 95% 95% 94%

78% 87% 90% 92% 93% 94% 95% 95% 95% 96% 96% 96% 96% 96% 96% 96% 96% 95% 95% 95% 95% 94%

body to suppress the road and wind noises heard by

79% 87% 90% 92% 93% 94% 95% 95% 95% 96% 96% 96% 96% 96% 96% 96% 96% 95% 95% 95% 95% 94%

60
79% 87% 90% 92% 93% 94% 94% 95% 95% 95% 96% 96% 96% 96% 96% 96% 96% 95% 95% 95% 95% 94%

80% 88% 91% 92% 93% 94% 94% 95% 95% 95% 96% 96% 96% 96% 96% 96% 95% 95% 95% 95% 95% 94%

80% 88% 91% 92% 93% 94% 94% 95% 95% 95% 95% 96% 96% 96% 96% 96% 95% 95% 95% 95% 94% 94%

80% 88% 91% 92% 93% 94% 94% 95% 95% 95% 95% 95% 96% 96% 96% 95% 95% 95% 95% 94% 94% 94%

passengers. Further, measures are provided to suppress

80% 88% 90% 92% 93% 93% 94% 94% 94% 95% 95% 95% 96% 95% 95% 95% 95% 94% 94% 94% 94% 93%

80% 87% 90% 92% 92% 93% 93% 94% 94% 94% 94% 94% 95% 95% 94% 94% 94% 93% 93% 93% 93% 92%

80% 87% 89% 91% 92% 92% 92% 93% 93% 93% 93% 93% 93% 94% 93% 93% 92% 92% 91% 91% 91% 90%

0
79% 86% 88% 90% 90% 91% 91% 91% 91% 91% 91% 91% 91% 91% 91% 90% 90% 89% 88% 88% 87% 86%

the high-frequency noise generated by the motor. By

75% 81% 84% 85% 85% 85% 85% 85% 85% 86% 85% 84% 85% 85% 84% 84% 82% 81% 81% 79% 78% 77% 76%

0 2000 4000 6000 8000 10000 12000 14000

implementing these measures, the overall quietness was Fig.3 Efficiency
improved.
The high-frequency noise generated by the motor,
Fig. 3 Efficiency
whose main component is the electromagnetic excitation 4. Characteristics of ARIYA s motor
force, changes according to the torque and number of
revolutions. The magnetic force of the rotor is constant in 4.1 Structure of the wound rotor
an IPM; therefore, the torque is controlled only by the EESM, as the name suggests, is structured to supply
current supplied to the stator. Thus, the electromagnetic DC to the rotor that acts as an electromagnet. Therefore,
excitation force is proportional to the supplied current. the rotor has a wire-wound structure. The rotor developed
However, in the motor used in ARIYA, the current is for ARIYA has eight wire-wound poles. Renault, which is
controlled for both the stator and rotor, the electromagnetic Nissan s alliance group, has adopted an EESM in ZOE
excitation force is reduced, particularly in the low-torque that has four wire-wound poles. To obtain higher outputs,
area. Further, a higher level of quietness is realized in the rotor volume was increased and the number of poles
scenarios where the driving force requirement is not was increased to eight to secure the necessary
large, such as during gentle acceleration and cruising electromagnetic force and torque responsiveness.
(Fig. 2). Moreover, the number of revolutions is higher in
ARIYA. To realize this feature, the aforementioned
measures have been implemented for increasing the
output and additional measures are provided in the rotor
Interior noise [dB-A]

structure. The cross section of the rotor and a magnified

view of a portion of the cross section are shown in Fig. 4.
Good

After winding the wire on each of the eight rotor poles,

Constant speed
ARIYA wedge-shaped slots were inserted between the poles to
on highway EV made by suppress wound-wire deformations due to centrifugal
Company T
Usual drive on city road
LEAF e+ forces. Therefore, the stability of the wound wire shape is
crucial. When commencing the motor production process
0 50 100 150 200 250 300 350 for ARIYA at the Nissan Intelligent Factory, a fully
Motor torque [Nm]
automatic wire-winding machine capable of winding
Fig.2 Comparison of interior noise eight rotors at once was introduced to ensure sufficient
production capacity. In addition, collaboration was
undertaken with the Production Engineering Division to
Fig. 2 Comparison of interior noise
3.2 Improving electrical energy consumption realize uniform and stable wire-winding shape between
The induction voltage generated between the rotor and rotors.
stator in the IPM increases with the increase in the
number of revolutions. In such cases, the output is
usually secured in the high-revolution regime by
measures, such as field-weakening control; however, this
may simultaneously worsen the efficiency. Owing to a
similar mechanism as that for high-frequency noise, the
magnetic force is constant for the IPM. Therefore, the
magnitudes of the induction voltage and field-weakening
current are determined uniquely according to the torque
and number of revolutions. In ARIYA s motor, the Alliance Internal

magnetic force of the rotor can be changed by changing

the supplied current. Accordingly, the rotor and stator
currents are controlled such that the induction voltage is

70 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2 : Next-generation flagship: ARIYA - 6. Newly developed motor to realize ARIYA s performance

Stator

Water cooling system

Heat exchanger
Rotor

Bearing Bearing
sealing sealing

Oil T-sensor

Pump Oil strainer

Fig.5
Fig.Oil-cooling circuit
5 Oil-cooling circuitofofthe
themotor
motor

Slot wedge Rotor coil

Rotor core
5. Summary

Electric drive technology is rapidly evolving along with

the acceleration of electrification trends. In addition to
IPMs, which have been widely used previously, an EESM
has been developed for ARIYA with sufficient performance
and mass-production quality. Thus, we are now able to
provide the technology most suited for meeting customer
demands regarding vehicle performance. Furthermore,
we are able to reduce the cost and supply risks of rare-
Fig.
Fig.44Structure
Structureof theof therotor
wound wound rotor earth materials.
Alliance Internal
In future, we intend to further utilize the characteristics
of EESMs. For example, in vehicles that have drive
4.2 Improved cooling performance via oil cooling motors at both the front and rear ends, during low loads
Because the EESM is structured to supply electrical (e.g., when cruising), the electric current supply to the
current to the rotor that acts as an electromagnet, the stator/rotor can be turned off for one of the motors to
heat generated by the rotor is high. Therefore, a direct effectively improve electrical energy consumption.
oil-cooling method was adopted in ARIYA. The
configuration of the oil-cooling circuit is shown in Fig. 5. Terminology
The oil that accumulates in the oil pan below the motor is *1 IPM (SM): Interior permanent magnet synchronous
pumped using an electric pump, transferred through the motor, which is an embedded-magnet-type
channels in the motor casing, and sprayed over the stator synchronous motor.
and rotor. For the stator, oil is supplied from the outer *2 EESM: Electrically excited synchronous motor, which
periphery primarily to cool the core. For the rotor, the is an external-field-type synchronous motor.
supplied oil cools the front and rear coil end portions, and *3 PEB: Power electric box that mainly integrates the
oil is simultaneously supplied to the bearings and oil inverter function with the peripheral circuits and
seals to ensure lubrication. power source branching function.
The oil circulates only inside the motor, and the *4 BEV: Battery electric vehicle, which is an electric
generated heat is discharged to the water-cooling circuit vehicle driven primarily by the energy stored in an
of the vehicle side through a heat exchanger attached to externally charged battery.
the side of the motor. The discharge quantity of the oil
pump can be varied according to the temperatures of the
stator and rotor. Thus, the electrical power consumed by
the pump is suppressed.
Additionally, the oil is used to cool the high-voltage
components. Therefore, oil that can achieve stable
performance from very low to high temperatures while
ensuring electrical insulation performance was selected.
In addition, a maintenance-free system that requires no
replacement or replenishing of the oil is realized.

NISSAN TECHNICAL REVIEW No.88 (2022) 71

 Special Feature2 : Next-generation flagship: ARIYA - 6. Newly developed motor to realize ARIYA s performance

Authors

Takahito Okubo Hiroki Wada

72 No.88 (2022) NISSAN TECHNICAL REVIEW

 Special Feature2

Next-generation flagship: ARIYA

7. Nissan s e-4ORCE : Proposition for evolution of

electric AWDs
Aiming for a more exhilarating and confident driving experience

Hiroyuki Togashi* Yutoku Miyagoshi** Takeji Katakura***

AWDs to ensure that drivers feel secure when operating

1. Introduction to e-4ORCE the vehicle regardless of the environment, and the vehicle
posture can be controlled such that all passengers can
Starting from ARIYA, Nissan will be adopting feel comfortable.
e-4ORCE, which is a novel technology for controlling
electric all-wheel drives (AWDs). The e in e-4ORCE
indicates the electric drive of EVs. Whereas, the 4ORCE
(read as force ), which combines the physical power and
Mechanica
energy with the 4 of four-wheel drives, indicates the
will to control the power or force of the tires of Brake Power supply cable distributor
four-wheelers.
In recent years, electric AWDs mounted with two high-
power motors, one each at the front and rear, have been Power supply cable
increasingly adopted by companies to meet the demands
for higher vehicle dynamics and environmental
performance. Compared to previous mechanical AWDs,
Nissan considers that electric AWDsMotor have very high
technological potential because of their good Fig.1 Diagram of e-4ORCE AWD system
acceleration and high ultimate performance levels ,
and because they can achieve better vehicular motion Fig. 1 Diagram of e-4ORCE AWD system
quality even in daily usage environments. The
e-4ORCE technology is an attempt to unleash the Engine
potential of electric AWDs and improve the quality of
Battery
vehicular motion. This section describes the e-4ORCE
technology.
Mechanical
Mechanical torque Torque is distributed
2. Overview of the e-4ORCE technology torque
Brake Power supply cable distributor distributor through the connected
The previous AWD system, which was mounted with
propeller shaft
combustion engines, realized power generation at the
Powercombustion
supply cable engine and its mechanical distribution to the
Motorshaft
front and rear wheels via the connected propeller
(Fig. 2). Therefore, there may be mechanical delays when
transferring power and mechanical limits to the Fig.2 Diagram of previous AWD system
resolution of power distribution. Unlike electric motors,
it is difficult for combustion engines to control their
outputs with high responsivity. Therefore, it is difficult to When a vehicle body is supported by four wheels, the
Fig. 1 Diagram of e-4ORCE
achieve AWD
total driving system
force control of the order of 0.1 s. load applied
Fig. 2toDiagram
each wheel changes constantly
of previous AWD systemowing to
The basic configuration of e-4ORCE is an electric the road surface and vehicular conditions. Additionally,
AWD. As electric AWDs are mounted with independent the frictional force limit of each tire changes according to
electric motors at the front and rear (Fig. 1), it is possible the load on the wheel (Fig. 3). The key criterion here is to
to distribute driving force to the front and rear tires with achieve control such that the load is distributed to all
high responsiveness and accuracy. The technological aim tires, with good balance including some margin, and
of e-4ORCE is to realize the full potential of electric within the limits of the tires.

*Vehicle Performance Engineering Department & Powertrain and EV Performance Engineering Department
**AD/ADAS and Chassis Control System Engineering Department ***Powertrain and EV Performance Engineering Department

NISSAN TECHNICAL REVIEW No.88 (2022) 73

Special Feature2 : Next-generation flagship: ARIYA - 7. Nissan s e-4ORCE : Proposition for evolution of electric AWDs Aiming for a more exhilarating and confident driving experience

~1990 2000 2010 2015 2020

Electrification
technology

Tama electric car Compact Li-ion Batt. LEAF e-POWER e-Pedal LEAF e+ e-4ORCE

Force to move forward

Increased number
of motors
Higher power

4WD control
technology

Integration
Resultant force
ATTESA ETS e-4WD GT-R ALL MODE e-POWER 4WD
4X4-i

Chassis control
technology

Friction circle: Force capacity

Force to turn HICAS Super
HICAS
Hydraulic
Active Susp.
Yaw Rate
F/B HICAS
4WAS Intelligent
Trace
Control
Intelligent
Ride
Control
DAS

Fig. 5 History of e-4ORCE

Fig.3 Tire friction circle and resultant force Fig.5 History of e-4ORCE

In front-wheel drive vehicles, the turning and 3. Customer experience offered by

acceleration/deceleration functions are integrated at the
Fig. 3 Tire friction circle and e-4ORCE
front tires; thus, there is little room for margin. In four-
resultant force
wheel drive vehicles, the turning and acceleration/ An exhilarating and confident driving experience
deceleration functions are optimally shared between the Given the premise of an electric AWD configuration,
front and rear wheels (Fig. 4) to achieve stability by the target of e-4ORCE is to fully unleash the high
preventing each tire from exceeding its limit. In the potential of the electric motor of a vehicle to a level that
ultimate limiting behavior, the load is distributed such has never been achieved by conventional combustion
that the maximum value (edge of the friction circle) of engine vehicles or electric AWDs manufactured by other
each tire is optimally used. companies. This includes excellence in handling
Below the limiting behavior, load distribution is maneuverability; sense of security regardless of the type
controlled loosely yet carefully within the limits of each of road surface; and comfortable riding experience for all
tire. Thus, smooth and easy-to-maneuver vehicle motions passengers.
that are impossible in mechanical AWDs can be realized.
In addition, the left/right brakes are controlled according 3.1 Excellence in handling performance
to the situation, and the braking force is utilized in In general, accelerating a vehicle along a curve causes
situations other than deceleration to improve turning the vehicle to deviate outward with respect to the normal
performance. line of travel. Therefore, the driver must operate the
steering wheel further or decrease the speed (understeer).
This phenomenon occurs because a portion of the limit
value of the friction circle determined by the load on the
wheel is used for acceleration, which weakens the turning
Complete margin usage
force. In accordance with the road surface and driving
conditions, the e-4ORCE technology distributes the
driving force to the front and rear wheels so that optimal
Resultant force
tire gripping is achieved based on the fluctuating loads
No margin on the wheels. In normal driving mode, the front/rear
Margin
torque distribution is close to 50:50. Moreover, the
distribution is changed automatically to an optimal ratio
within the range of 0:100 to 100:0 without driver effort
Some margin based on the road surface and driving conditions. In
addition, curve turning performance is improved during
Fig.4 Relationship between the friction circle and resultant force deceleration by combined control of regenerative braking
of the four wheels by the front-rear motor and the hydraulic brakes of the
Fig. 4 Relationship between the friction circle and four wheels (Fig. 6).
resultant force of the four wheels
These technologies have been created and developed
by crossbreeding, combining, and further advancing the
technological and empirical genes developed by Nissan
over the years. These technologies include the AWD
technologies, such as the Advanced Total Traction
Engineering System for All Electronic-Torque Split
(ATTESA E-TS) 4WD system, Intelligent 4 4 system,
and knowledge gained for electric drive control through
LEAF and E-POWER (Fig. 5).

74 No.88 (2022) NISSAN TECHNICAL REVIEW

Special Feature2 : Next-generation flagship: ARIYA - 7. Nissan s e-4ORCE : Proposition for evolution of electric AWDs Aiming for a more exhilarating and confident driving experience

Understeered High power 4WD EV reference [1]

vehicle
High power 4WD EV reference [2]
Previous TOP BM ICE 4WD reference

Fluctuation in steer characteristic

e-4ORCE

Driving force Acceleration from state of turning

Braking force Time [s]

Friction circle Fig.7 Turning acceleration performance on a snowy surface

Tire gripping force
Fig. 7 Turning acceleration performance in a sno
Fig.6 Comparison between e-4ORCE and understeering A high levelsurface
of sense of security can be experienced
Fig. 6 Comparison between e-4ORCE and understeering through this high stability, regardless of the situation,
road surface, and driver. In addition, this reliability may
This sophisticated control exerts its effects, particularly bring joy to skilled drivers when driving on snowy and
when turning a curve. Because the vehicular motion is frozen road surfaces during winter.
guided by the steering operation of the driver, it is
expected that steering correction can be minimized. We 3.3 Comfortable riding experience for all passengers
hope that the drivers can enjoy a smooth and highly Control of the regenerative brakes by the motor also
stable driving experience enabled by the e-4ORCE contributes to a comfortable riding experience. Even if
technology. the vehicle is an EV, applying regenerative brakes to
front-wheel drive vehicles can cause a substantial tilting
3.2 Sense of security on various road surfaces feeling to the passengers because only the front motor is
The technologies for independently controlling the used when braking; thus, the rear of the vehicle is raised
front/rear motors and brake control developed by Nissan when decelerating (referred to as pitch behavior) (Fig. 8).
over the years enable excellent driving performance even In the e-4ORCE technology, the regenerative brake
on constantly changing and slippery road surfaces with of the rear motor is used in addition to that of the front
snow, compacted snow, ice, ruts, or wetness. motor. Thus, there will be lesser vibration even when the
Fig. 7 shows the turning acceleration stability data as road is congested, thereby resulting in a comfortable
a case example of achieving a sense of security. This data riding experience. When accelerating and decelerating,
graphically compares fluctuations in the steer fluctuations in the vehicle posture are suppressed by the
characteristic when pressing the accelerator for 4 seconds optimal control of the motor (right side of Fig. 8). Because
up to full throttle after stabilizing the turning condition the vibrations of the vehicle body are minimized, a
to a constant 40R at a vehicle speed of 25 kph. The steer smooth and comfortable riding experience is available to
characteristic is an index of the degree of further steering both the driver and passengers in other seats.
required to align the vehicle with the travel line when the (Adopted for the first time in NOTE E-POWER 4WD)
vehicle deviates outward from the target line. If the road
surface is covered by snow, the frictional force will be low
FWD e-4ORCE
and surface will be uneven. Therefore, the size of the Change in posture due to regenerative
Change in vehicle posture (pitch behavior) is
suppressed by appropriately distributing the
friction circle fluctuates constantly, and the travel line braking force
front/rear regenerative braking forces
may deviate both outward and inward; thus, stabilizing
the steering condition will not be easy. Even on such
snowy road surfaces, ARIYA s e-4ORCE causes very
little fluctuation in the steer characteristic, thereby
enabling stable driving. Compared to the AWD SUVs of Fig.8 Tilting of the vehicle
competitors that have high-power motors (green and blue Fig. 8 Tilting of the vehicle

lines in Fig. 7), the fluctuation in steer characteristic was

approximately 1/9 for ARIYA. Moreover, the fluctuation 4. Summary
achieved was small and approximately 1/3 when
compared with the top benchmarked AWD SUV mounted The principle behind the development of e-4ORCE
with a conventional combustion engine (orange line). was to allow the system to achieve control naturally such
that the driver would not notice the operation of the
system. It is expected that anyone riding in a vehicle
equipped with e-4ORCE technology for the first time
would feel that the vehicle can be maneuvered naturally

NISSAN TECHNICAL REVIEW No.88 (2022) 75

Special Feature2 : Next-generation flagship: ARIYA - 7. Nissan s e-4ORCE : Proposition for evolution of electric AWDs Aiming for a more exhilarating and confident driving experience

and comfortably without great effort. References

For our long-time customers who have continued to N. Nakata, Special Feature 2: Evolving e-POWER 2.
choose Nissan vehicles, we hope that the Nissan way System Characteristics and Future Development ,
can occasionally be felt when driving the vehicle. We Nissan Technical Review, No. 88 (2020)
hope that the Nissan way, but redefined can be felt by
our customers. The e-4ORCE technology was developed R. Hiraku, E. Sakagami, T. Katakura, Special Feature 2:
with the aim that our customers will find it indispensable Evolving e-POWER 6. Value-Adding Electric All Wheel
in their vehicles once they have experienced it and that Drive (AWD) Technology, Nissan Technical Review, No.
they would become used to the sense of security it 88 (2020)
provides.
Because ARIYA provides stable steering operation Nissan Motor Co., Ltd., Intelligent Trace Control,
with very little fluctuation in the steer characteristic, the Nissan Global Site, https://www.nissan-global.com/JP/
vehicle can be handled confidently, even though it has TECHNOLOGY/OVERVIEW/cornering_stability_assist.
the power equivalent to a sports car. Moreover, the html
e-4ORCE technology has potential for further
development to enable enhanced performance for sports
cars.

Authors

Hiroyuki Togashi Yutoku Miyagoshi Takeji Katakura

76 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers

Study about Workload Reduction by Multiple Modal

Interface Using Gaze and Voice Recognition

Yusuke Morita* Sou Tamura*

Abstract Our aim is to improve an HMI that reduces the workload by distributing the different resources
(visual / auditory / verbal / spatial etc ) related with human s recognition. In this study, the effect on the driver s
workload is clarified by comparing two types of operations: the current voice operation as single modal, and
operation using gaze pointing and voice command as multimodal. With regard to gaze and voice control, there are
2 use cases that are more effective than the current voice operation. It was confirmed by subjective evaluation
(NASA-TLX) and objective evaluation (Prediction Error) at the driving simulator tests.

response, which represent the input and output of

1. Introduction information; the second consists of visual and
auditory, which are modalities; and the third includes
Over recent years, the demands for vehicle safety and spatial and verbal, which are forms of expressing
comfort have increased. In addition, vehicles are being information. In addition, visual is further divided into
mounted with increasingly new functions such as focal vision and ambient vision. Using different levels
autonomous driving, parking assistance, and providing in each of these four dimensions instead of focusing on
information using connected services. Owing to these just one resource is considered to contribute toward
trends, the operational workload related to setting, reducing the workload. Moreover, based on this theory, it
adjusting, and checking functions is increasing. Single- has been reported that the workload will be reduced on
modal interfaces such as touch selection (visually dispersing the workload across multiple resources.2) In
searching the position before touching the position) and automobiles, voice operation using voice recognition
voice input (command utterance) are typically used for technology is being employed to reduce the workload.
operations in current vehicles. However, in this method, only the verbal resource of
In the case of touch selection, for example, a hierarchical the information expressing form (third dimension) is
structure is generally adopted because the display area is utilized. The authors hypothesized that it may be possible
limited. The user must recognize and understand the to reduce the driver workload by dispersing the resources
operation logic and order beforehand to determine the into spatial and verbal, if the spatial information at
types of functions available at each location. In the case which the driver is gazing is utilized by employing gaze
of voice input, the desired operation can be determined detection technology, which has evolved over recent
directly. However, there are often difficulties when using years.
voice inputs for commands that accompany selection; this Therefore, this study focused on the visual recognition
is because the workload will be high when uttering the that is regularly performed during driving. In conjunction
necessary commands (such as the name of the device to with voice recognition, multimodal operating commands
be operated) followed by the specific choices (such as using gaze pointing and voice commands were created to
right-front side or lowermost ). disperse the workload on the resources, with the aim of
To reduce the workload in these single-modal reducing the workload in the overall operation. When
operations, the authors focused on multimodal operation utilizing the visual recognition activity as the trigger for
based on the multiple resource theory.1) According to an operation (for example, using gaze pointing and voice
Wickens et al., multiple resources are used according to commands to view a navigation map that is off-scale and
the type of information when a mental workload, make corrections or employing an auxiliary monitor to
including cognition and response, is imposed to execute visualize a direction that cannot be seen), monitoring the
an action. In theory, it is considered that a set of three gaze pointing information of the driver can be valuable
dichotomies of information processing resources are for estimating the operation target. Using this method,
present in humans. The first includes perception and the authors assumed that the driver can directly give

*Interior and Exterior Engineering Department

NISSAN TECHNICAL REVIEW No.88 (2022) 77

 Technical papers Study about Workload Reduction by Multiple Modal Interface Using Gaze and Voice Recognition

commands even if the operation target is not indicated,

thus reducing the workload (Figure 1). Vehicle
Among the multimodal interfaces, those utilizing gaze Driver Single-
Multi-modal I/F
pointing and voice commands, in particular, have been modal I/F
studied previously.3) 4) However, these studies were Behavior Thinking Voice Voice Gaze
limited to general information devices, and the operation

Time
object
Look
of the devices in these works was different from that of
in-vehicle devices, because the precondition for in-vehicle

(object Selection)
Trigger
devices is that the devices will be used while the driver is

Input
see…
Can’t
always alert to the surroundings. Furthermore, in a
study related to in-vehicle devices,5) the workload
reduction effect was not clearly indicated.

Want to
operate
Therefore, this study focused on the operations
triggered by visual recognition activities. Operation by
voice commands only was compared with operation by

to operate)
(decision
(device assignment +
decision to operate)
Speech
gaze pointing and voice commands in order to evaluate

Input
the effects of the multimodal interface.

Operation

Input
result
Look
Fig.2 Comparison with single and multi modal operation

Fig.1 Navi operation by gaze and voice According to Figure 2, when the driver attempts to
view an object, moves the line of sight, and becomes
aware that it is impossible to see, a desire to resolve the
2. Multimodal interface using gaze pointing feeling of discontent will arise.
and voice commands In the case of voice recognition, the driver will give
Alliance Internal
commands by uttering the device name or direction and
2.1 Outline the operation command. By contrast, in the case of the
To evaluate the effect of detecting the gaze of the driver operation using gaze pointing and voice commands, the
and using it for operation together with voice recognition, operation target device or direction can be estimated based
this approach was compared with the current operation, on the gaze detection function of the vehicle system when
i.e., that based on voice recognition alone (single-modal visual recognition, which serves as the trigger, is performed
operation). The details are presented in Figure 2. The by the driver. In this method, the driver does not need to
driver behaviors and thoughts are provided on the left, select the target and can perform the operation by uttering
while the input timing and details for different interfaces simple commands such as camera or expand as the
are presented on the right for comparison. input. Therefore, compared with the operation using voice
commands only, the driver only needs to utter simple
words, and supplemental information such as that
regarding direction and position is unnecessary.

2.2 Target operation

As described above, the tasks that can realize
operational workload reduction by using gaze pointing
have a common pattern. Specifically, the information
obtained from gaze pointing is utilized as the trigger to
proceed with the operation. Considering this pattern, the
case in which devices with the same name exist (example:
right camera/left camera/rear camera) in a vehicle, and
the case in which multiple devices may react to the same
command word (example: expand of the electronic
mirror image/ expand of the map), the three following
case examples were established as the tasks:
(i) Map expansion:
If two or more maps are presented on a large screen

78 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers Study about Workload Reduction by Multiple Modal Interface Using Gaze and Voice Recognition

that extends from the meter cluster to the center cluster

(which is increasingly adopted in vehicles in recent years)
or if the maps are shown on multiple screens, a command
of type of map+details of operation is needed for voice
recognition to operate a certain map. By contrast, in the
case of operation commands using gaze pointing and
voice commands, the map can be operated directly
without selecting it verbally. In this study, a map
expansion operation was utilized as the target task.
(ii) Electronic room mirror image expansion:
In this study, the image expansion operation of the Fig.4 Passing scene
rear-view camera served as the target task, assuming
that the vehicle or vehicle model at the rear side needs to
be checked in a scene where the vehicle at the rear is Operation by voice command alone and operation by
tailgating (a social problem in recent years), multiple both gaze pointing and voice command for the
vehicles are traveling in a group, etc. In the case of voice abovementioned three scenes were compared. The results
operation, the vehicle will not be able to recognize what are summarized in Table 1.
device the command is intended for if only expand is
uttered. By contrast, in the operation by gaze pointing Table1 Comparison with control of voice and gaze & voice
and voice command, the input of the expansion command
will be complete simply by uttering expand after
looking at the mirror. In this study, considering the field
of view of the rear-view mirror when an image is
expanded, the expanded image is partially overlaid
(Figure 3).

Fig.3 Sample picture of expand view

In operations (i)–(iii), the driver always visually

(iii) Checking the side camera images: confirms the device (i, ii) or the attention direction (iii) to
When initiating the camera that assists in viewing the check the situation before executing the operation. By
blind spot in a vehicle equipped with multiple cameras, having the system estimate the intention of the driver
the vehicle will not be able to recognize which camera to from the gaze pointing direction, it was assumed that the
start up if only camera is uttered. In particular, when voice command workload concerning the device and
passing another vehicle on a narrow road (Figure 4), direction could be reduced.
contact with the wall or the oncoming vehicle must be
avoided. Therefore, various types of information must be 3. Evaluation test
obtained, judgments must be made, and the vehicle must A
be operated within a short period of time, which may 3.1 Evaluation system
cause impatience. In the case of operation by gaze An outline of the system used to evaluate the load
pointing and voice command, the operation target device reduction for the operations set in Subsection 2.2 is
can be estimated based on the gaze pointing direction, depicted in Figure 5. The system comprises an infrared
and thus, the input for starting up will be complete (IR) camera, a microphone, a personal computer (PC)
simply by uttering camera. that processes the data from them, and a display
employed as an output device.
The face images obtained by the IR camera were
analyzed using a PC. Based on the gaze pointing
information of the driver and the mockup layout
information from the camera, the data of which device/
area the driver is looking at were created (Figure 6).
Smart Eye CDK v.0.9.0, manufactured by Smart Eye®
AB, was used as the gaze detection system, including the
camera. In addition, the voice information obtained from
the microphone was processed using a voice recognition

NISSAN TECHNICAL REVIEW No.88 (2022) 79

 Technical papers Study about Workload Reduction by Multiple Modal Interface Using Gaze and Voice Recognition

system and converted into text data. Google Cloud™

Speech-to-Text API was utilized as the voice recognition
system.
In the operation command processing, the gaze point
information and voice command information were
integrated, and the image to be output was determined.

IR Camera Microphone
Get image of Get speech
face/eye ball information

PC

Gaze detection Voice recognition

software software Fig7. Test condition of Driving Simulator

Output gaze Output text from

An 11-inch liquid-crystal display (center layout)+8-
position from image speech
inch liquid-crystal display (meter layout)+7-inch
liquid-crystal display (room mirror layout) were used as
the on-board devices which show results .
Operation command The tests, including the driving tasks, were conducted
processing using a driving simulator manufactured by Mitsubishi
Precision Co., Ltd. The evaluation scene is illustrated in
Create operation commands based on Figure 7.
information of gaze and text
3.2 Evaluation method
NASA-TLX,6) which has been adopted in previous
Output device studies, was employed for the subjective evaluation. In
addition, for comparisons using quantitative values, the
Receive device prediction error of the steering angle was utilized. The
operation command prediction error of the steering angle is the basis for
calculating the steering entropy,7) which represents the
Fig.5 System outline and function flow
smoothness of the steering operation of the driver.
The prediction error of the steering angle was derived
by a time-series calculation using the difference with
past data. The prediction error, and, therefore, the
e Mirror area Front difference, will be larger if sudden steering is necessary.
windshield The distribution of the values will be wider and the peak
value will be lower if the driver is preoccupied with tasks
Outside of apart from driving, such as looking aside or becoming
absorbed in thought.
front left area
Meter Display Tests were conducted using seven men and women
Center Display area who had their driver s licenses. Their ages ranged from
area 20 to 69 years. After the drivers underwent sufficient
training for each operation, the tests were conducted
three times, and the results were evaluated each time the
driving test was conducted. The driving and operating
conditions are listed in Table 2.
Steering Alliance Internal
Fig.6 Image of recognizable area
Table2 Test condition

80 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers Study about Workload Reduction by Multiple Modal Interface Using Gaze and Voice Recognition

The conditions for driving on the highway were as The results under the current test conditions merely
follows. Individually driving in the center lane of a suggest a trend of load reduction.
highway section (i.e., with no other vehicle and no lane The initiation of camera operation while passing
changing) with a continuous gentle curve (minimum another vehicle on a narrow road was analyzed based on
radius of approximately 1,000 m). Half of the curve the aforementioned four dimensions of the multiple
consisted of a right-hand curve, whereas the other half resource theory.1) It was assumed that the driver first
consisted of a left-hand curve. Therefore, it is considered had to check the clearance between the vehicle and
that there was no bias in the curve conditions. The surrounding obstacles. Therefore, it was necessary for
conditions for driving on the city road were as follows: the driver to concentrate on the focal vision resource
driving on a road (with a width of approximately 5 m) among the focal and ambient visions (dimension 4) and
that has no center line while avoiding a car parked on the then operate the vehicle to avoid contact with the
passenger side. surrounding obstacles. Thus, the driver was required to
For both driving on the highway and on city roads, concentrate on the response resource among perception
tests were conducted with the following three patterns: and response (dimension 1). In addition, both the voice
no load (driving operation alone, without any other task), command operation and the operation using gaze pointing
operation by voice command, and operation by gaze and voice commands use the focal vision (checking the
pointing and voice command. Considering the order device) and response (operating the device) resources,
effect, the tests were conducted in a random order for leading to interference. Therefore, if focal vision is used
each participant. frequently and if the driving operation workload is high,
The experiments performed in this study were there is little difference between the voice command
examined and approved by the Ethics Committee of operation and the operation using both gaze pointing and
Nissan Motor Co., Ltd. The tests were conducted after voice commands. Furthermore, people with low driving
obtaining informed consent from the participants. skills may be more affected by the driving operation
when passing another vehicle on a narrow road, which
3.3 Subjective evaluation results may also affect the results.
The results of the subjective evaluation by the NASA- On the other hand, when driving on a highway, the
TLX are presented in Figure 8. To assess whether the driver workload is assumed to be relatively low because
workload reduction by using gaze pointing and voice the vehicle remains in the same lane. Therefore, there
commands, as compared to that when using voice were likely sufficient resources to naturally perform
commands alone, was significant, the evaluation results activities such as looking at the map or in the mirror, and
of each modality of target tasks (i)–(iii) were paired. The uttering simple commands also probably reduced the
probability associated with the student's paired t-test workload.
(with a two-tailed distribution) was calculated to derive
the p-value. 3.4 Quantitative evaluation results
Examples of the frequency distributions of the
prediction error in the steering angle, calculated as
quantitative evaluation values of a participant (a female
in her 20s), are shown in Figures 9–11. The time elapsed
from giving instructions to start the task to its completion
was approximately 10 s. Therefore, data for 10 s from the
start of the task were used for the evaluation. For in
Figure 9–11, the 90th percentile value of the no-load
distribution was used to represent the differences in
driving characteristics among individuals.8)
As mentioned above, in the frequency distribution, the
sharpness near the center will increase if the driving is
smoother. In the current example, the peak gradually
Fig.8 Result of NASA-TLX * p<0.05 p<0.1 decreases and the distribution spreads out in the order of
no load operation by gaze pointing and voice
command operation by voice command, which
Compared to the voice command operation, the suggests that the deviation in steering becomes larger
weighted workload (WWL) score for the operation by and sudden operations increase in this order.
gaze pointing and voice command was statistically
significant for (i) map expansion and of low statistical
significance for (ii) mirror expansion. The results suggest
that, on incorporating gaze pointing, the workload is
decreased, as compared to that with the current voice
command input.
On the other hand, the camera starting operation
method does not have a statistically significant effect.

NISSAN TECHNICAL REVIEW No.88 (2022) 81

 Technical papers Study about Workload Reduction by Multiple Modal Interface Using Gaze and Voice Recognition

Fig.9 Frequency distribution Example Map expansion Fig.12 Comparison of value * p<0.05 p<0.1

Similar to the subjective evaluation, there exists a

statistically significant difference or a trend of statistical
significance for (i) map expansion and (ii) mirror
expansion. Therefore, the workload reduction was also
expressed quantitatively.
For (iii) camera start up, the value (which shows
that the prediction error in the steering angle is large)
tends to be larger than those for (i) and (ii). This result
suggests that the driver workload is high for (iii), which
is a trend similar to that obtained in the subjective
evaluation. No significant difference is observed between
the two surgical modalities with increasing visual
Fig.10 Frequency distribution Example eMirror expansion recognition load. This trend is in line with the subjective
evaluation results; thus, the trend is considered
reasonable.

4. Summary

In this study, a multimodal interface using gaze

pointing and voice commands was investigated, focusing
on the operations triggered by visual recognition
activities. This approach was compared with the
conventional voice command operation to evaluate the
effects of the multimodal interface. The evaluation
confirmed that the workload could be reduced, as
compared to that during operation using voice commands
alone (single modality).
Fig.11 Frequency distribution Example Camera However, operation by gaze pointing may be difficult
in some situations depending on the driving environment
or task. Therefore, it is considered that operation by gaze
To understand the overall trend of the participants, pointing will not completely replace other operation
the 90th percentile values, , of the frequency distributions methods such as voice command operations and steering
of the steering angle prediction errors were compared. A switches. Nevertheless, it would be appropriate to design
graph of the comparison results is depicted in Figure 12. the settings such that both operation methods are
Similar to the subjective evaluation described in available and that the user can choose the one that is
Subsection 3.3, a t-test was conducted to compare the most convenient.
different cases. In addition, as described above, the driver workload
will be higher if information is concentrated on a certain
resource ( focal vision and response in this study),
making it more difficult for the driver to respond in the
same manner as that during low workloads. For
situations where the driver drives on a narrow road and
the workload is high, as in the test conditions of this
study, it may be necessary to devise an appropriate

82 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers Study about Workload Reduction by Multiple Modal Interface Using Gaze and Voice Recognition

method such as automatically showing the necessary

camera image.
In future studies, the test candidates will be expanded
beyond the three tasks used in this study, with the aim of
extracting and realizing tasks with stronger influences.
Evaluations must also be conducted under actual driving
environments, and it is necessary to assess whether the
trend remains the same.

References
1) Christopher Wickens: Multiple Resources and Mental
Workload, Human Factors, Vol.50, No.3, p.449-455,
(2008)
2) Takahiro Iwata, Tetsuo Yamabe, Tatsuo Nakajima:
Workload Evaluation of Information Cognition in a
Multi-Task Environment, IPSJ SIG Technical Report,
Vol.2009-UBI-22, No.8, (2009)
3) Tomoya Okada, Daisuke Sakamoto, Tetsuo Ono:
Method of Presenting Translations When Reading
English Language by Combining Gaze Pointing and
Voice Information, Interaction Papers 2019, IPSJ
Symposium, p.747-751, (2019)
4) Chihiro Nakagawa, Toshikazu Matsubara, Martin J.
Durst: Proposal and Implementation of the Web
Browsing Interface That Combines Gaze Pointing
Information and Voice Information, Proceedings of
the 78th National Convention of IPSJ, p.369-370,
(2016)
5) Abdul Rafey Aftab: Multimodal Driver Interaction
with Gesture, Gaze and Speech, ICMI '19, p.487-492,
(2019)
6) Shigeru Haga, Naoki Mizukami: Japanese Version of
NASA Task Load Index: Sensitivity of Its Workload
Score to Difficulty of Three Different Laboratory
Tasks, Japanese Journal of Ergonomics, Vol.32, No.2,
p.71-79 (1996)
7) Okihiko Nakayama, Tooru Futami, Tomokazu
Nakamura, Erwin R. Boer: Development of a Steering
Entropy Method for Evaluating Driver Workload,
Proceedings of the JSAE Annual Congress, No.45-99,
p.5-8 (1999)
8) Toshiyuki Ito, Okihiko Nakayama, Erwin R. Boer:
Establishment of JAMA's Safety Guideline for
In-vehicle Information System and Study of a Steering
Entropy Method for Evaluating Driver Workload,
IATSS Review, Vol.26, No.4, p.243-250, (2001)

Source

Vol.51, No.5. 20204473

NISSAN TECHNICAL REVIEW No.88 (2022) 83

 Technical papers Study about Workload Reduction by Multiple Modal Interface Using Gaze and Voice Recognition

Authors

Yusuke Morita Sou Tamura

84 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers

The Development of Bearings for the Multi-link Crank

Mechanism to Achieve Variable Compression Ratio Engine

Katsutoshi Nakamura* Katsuya Moteki** Satoru Ookuma*

Abstract To achieve variable compression ratio, the multi-link system which has secondary merit as good
quietness performance and low mechanical loss was selected. Under these circumstances, it was very important
how to design the journal bearings and bushes which is added as important elements to constitute the multi-link
system. In this paper, it is described that how to design the journal bearings and bushes for multi-link system
based on each peculiar load characteristic and sliding characteristic. At that time, each dynamic characteristic and
cooling effect by the engine oil by the added oil feeding path was considered.

conventional mechanism. Accordingly, it was necessary

1. Introduction to ensure a high load capacity and to supply lubricating
oil actively to the sliding portions. However, implementing
In 2018, the manufacturing of a mass-production type the usual countermeasures such as increasing the shaft
variable compression ratio engine was commenced. The diameter hinders the size reduction of the link
variable compression ratio mechanism used in this type components, which is necessary to ensure the
of engine can alter the compression ratio according to the mountability of the engine. Satisfying these contradictory
driving scene to achieve both low fuel consumption and requirements was considered in this research.
high output, which are contradictory characteristics.
Among the various variable compression ratio 3. Components of multi-link mechanism and
mechanisms proposed through previous research, the bearing elements
multi-link system adopted in this engine exhibits unique
value-adding characteristics such as quietness and 3.1 Components of multi-link mechanism
friction loss suppression, in addition to the main function The components of the variable compression ratio
of altering the compression ratio.(1) Bearing elements are (VCR) system, presupposing the use of a multi-link
important functional components that constitute the system, are presented in Figure 1. Compared to the
multi-link system, which realizes the abovementioned conventional mechanism (Figure 2), which comprises a
features. Regarding the bearing elements, this study piston, connecting rod, and crank shaft, the VCR system
focused on improving the load capacity according to the uses a multi-link crank mechanism that consists of the
peculiar load characteristics and slide characteristics connection of an upper link, a lower link, and a control
and also on the foreign matter discharging effect and link (hereinafter referred to as U-Link, L-Link, and
cooling effect on feeding engine oil to the sliding portions. C-Link, respectively), which are utilized instead of the
As a result of this research, functional reliability of the connecting rod. Below the crank shaft, the C-Link big end
bearings was ensured, without impairing the additional is connected to the control shaft eccentric journal and
values of this mechanism. fixed using the bearing cap. This control shaft is connected
At the same time, the friction loss equivalent for to the actuator link (referred to as an A-Link in this
engines with the conventional mechanism was realized, report) via the connecting portion, and the A-Link is
even though the number of bearing elements with small connected to the actuator (referred to as an ACTR in this
shaft diameters was increased. paper) installed on the side of the engine.

2. Development concept

Due to the lever structure of a multi-link system, the

load input to the bearing elements described herein tends
to be larger than the load input to bearings with the

*Powertrain and EV Mechanical system Technology Department **Powertrain and EV Advanced Technology Department

NISSAN TECHNICAL REVIEW No.88 (2022) 85

 Technical papers The Development of Bearings for the Multi-link Crank Mechanism to Achieve Variable Compression Ratio Engine

Piston
hand, the U-pin bush, C-pin bush, and C-Link big end
U-link bearing (which were newly established for the multi-link
(Upper link)
mechanism) as well as the piston pin bush (equivalent to
L-link
(Lower link) the conventional mechanism) function as the oscillating
Crank shaft
bearing. For the clearance in diameter, all bearings have
C-link almost the same clearance as the general main bearings
(Control link)
and connecting rod bearings.

Control Shaft ACTR Assy

Fig.1 Table1. Specification of bearings No.1
Fig.2
Fig. 1 Components of VCR Fig. 2 Components of L-Link bearing Main bearing
system conventional Upper Lower

3.2 Components of bearing elements in the multi- Diameter(mm) 42 52

link mechanism Width(mm) 25 17
In the multi-link mechanism, bearing elements are Width of groove(mm) - 3.15
The rate of increase
established to form connections between the of load
1.9 1.9
aforementioned piston, links, and shafts. The components Sliding property Rotating Rotating
of the bearings are illustrated in Figure 3. The bearing U-pin bush C-pin bush

elements for this mechanism are at the following locations.

There are five types of bearing element unique to this
mechanism: the U-pin bush, which connects the U-Link Diameter(mm) 23 21
and the upper pin press-fitted into the L-Link; the L-Link Width(mm) 16.6 18.5
bearing, which connects the L-Link and the crank pin; the Width of groove(mm) 2.29 2.29
C-pin bush, which connects the control pin press-fitted The rate of increase
1 0.9
of load
into the L-Link and the C-Link; the C-Link big end
Sliding property Oscillating Oscillating
bearing, which connects the C-Link and the control shaft
eccentric journal; and the control shaft bearing, which Table2. Specification of bearings No.2
connects the control shaft main journal and the bearing
C-Link big end bearing Controlshaft bearing
cap. In addition, there are two types of bearing element
that are identical to those in the conventional mechanism:
the main bearings that connect the crank shaft main
journal and the bearing cap, as well as the piston bush. Diameter(mm) 35 40
Width(mm) 17 17
Width of groove(mm) 2.58 3.15
The rate of increase
0.9 0.9
of load
Sliding property Oscillating Non-sliding

Piston pin bush

U-pin bush L-link bearing 4.1 Designing L-Link bearing

Three issues arose when designing the L-Link bearing.
Main bearing C-pin bush
The first issue originates from the characteristic structure
of the L-Link. As described by Tanabe,(2) compared to the
C-link big end conventional structure, the L-Link has a wider structure
bearing in the engine shaft direction than the connecting rod and
it also has a slit structure at the center, which covers half
Contshaft bearing of the overall width. Therefore, in the conventional
mechanism, the main pressure-receiving area has lower
Fig.3
Fig. 3 Bearing components of Multi-link system rigidity than the connecting rod (refer to Figure 4). In
addition, the shape of the L-Link is asymmetrical with
respect to the center of the L-Link bearing for the U-pin
4. Designing bearing elements and C-pin sides (refer to Figure 5). Owing to the above-
discussed characteristics, the housing side (link side)
The aforementioned bearing elements were designed tends to slip relative to the bearing side during service.
considering their individual load characteristics and The second issue is that the input load increases by up to
usage environments. The appearance, main dimensions, (L2+L4)/L2 = 1.9 times due to the lever structure of the
load properties, and sliding properties of each bearing link mechanism, as shown in Figure 6. The third and final
are listed in Tables 1 and 2. The L-Link bearing and main issue is that the inertia load, which tends to increase at
bearing function are the rotating bearings (same as the high engine speeds, is input in the direction of the plain
bearings for the conventional mechanism). On the other bearing mating surface, as shown in Figure 7. For this

86 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers The Development of Bearings for the Multi-link Crank Mechanism to Achieve Variable Compression Ratio Engine

bearing, the cylinder pressure load direction was

established to be almost the same as the L-Link bolt axis
direction (inclined at 90 with respect to the plain bearing
mating surface). However, owing to the oscillating
operation of the L-Link, an inertial load is generated
during service at approximately 90 with respect to the
cylinder pressure load direction. Consequently, this input
of the inertial load acts in the mating surface direction,
which has a low load capacity. In this research, the
following countermeasures were implemented to overcome
the three issues described above and secure the bearing
function. To address the first issue, i.e., the relative
slipping at the bearing back, the shaft diameter was
reduced as compared to that of the connecting rod bearing
in the conventional mechanism. In addition, as the L-Link
width was increased, compared to that of the connecting Fig. 6 Principle of lever by multi-link
rod bearing in the conventional mechanism, the bearing
width was increased. Using these methods, the total
Inertia load
adhesion at the bearing back was increased by a factor of
Combustion load Matching surface
approximately 1.5. The equation for calculating the total
adhesion and the specifications of the bearings are listed
in Table 3. On increasing the adhesion to the L-Link inner
diameter surface, followability was improved. Thus, it was
possible to address the issue of relative slipping. In
addition, the shape and arrangement of the claw structure U-pin side
C-pin side
at the mating surface were determined. As a result of the
above countermeasures, the increased input load (the
second issue) was leveled to the average surface pressure,
equivalent to that of the connecting rod bearing in the Fig. 7 Load of L-Link bearing
conventional mechanism. Furthermore, to address the
third issue, the inertia load in the direction of the bearing
mating surface and the height of the relief portion of the Table3. Specification of L-link(conrod) bearing
crush relief shape was reduced, compared to the height for Multi link Conventional
the general connecting rod bearing, to address to this issue D(mm) 45 47
by securing the pressure-receiving area (refer to Figure 8). t(mm) 1.5 1.5 D
L(mm) 25 16 ｔ
Fr ratio※ 1.56 1 L
Conrod L-Link
Total adhesion ： Fr = 2(1-t/D)(t/D) σB L D π (N)
σB: Stress in bearing circumferentional direction
※ Fr ratio = Fr value of each model / Fr value of conventional

Conventional Multi link system

6.5

20 32
2

Fig. 4 Width of L-Link

Height of crash relief
Fig. 8 The height of crush relief
Reaction force High rigidity part

4.2 Designing main bearing

As previously reported,
Fig.8
(3)
the load input into the crank
shaft main bearing increases by (L2+L4)/L2 = 1.9 times
Input force
due to the lever structure of the multi-link mechanism,
Low rigidity part similar to the aforementioned L-Link bearing (refer to
Figure 6). However, this lever structure also increases
Reaction force the piston stroke. Therefore, the crank radius can be
Fig. 5 Mechanical asymmetry of L-Link reduced by approximately 50%, as compared to that in

NISSAN TECHNICAL REVIEW No.88 (2022) 87

 Technical papers The Development of Bearings for the Multi-link Crank Mechanism to Achieve Variable Compression Ratio Engine

the conventional mechanism; thus, the opening

deformation of the crank shaft can be reduced (refer to
Figure 9). The results of these advantages and
270
disadvantages are presented in Figure 10. Owing to the
°
multi-link mechanism, the tendency of uneven contact at Rr-side 180
the bearing is significantly mitigated. Consequently, the 0
°
load capacity of the bearing is improved, and thus, the
Load Fr-side 90° °
bearing lubrication property is improved, although the
maximum load is increased to approximately twice that Main bearing
in the conventional mechanism. The effect of the local
Fig. 12 Positional relation of main bearing
surface pressure reduction is depicted in Figure 11. This
effect is shown as the elastohydrodynamic lubrication
(EHL) results of EXCITETM Power Unit, which is a multi- 4.3 Designing C-Link big end bearing, U-pin bush,
body dynamics analysis software. It is evident
Large from the and C-pin bush
deformation
results that the local surface pressure decreases, even The C-Link big end bearing, U-pin bush, and C-pin
compared to engines with smaller outputs. bush, which are the three elements newly established as
3DUWLDOFRQWDFW
the bearing elements for the multi-link mechanism, are
Conventional system all used in the oscillating conditions (identical to the
Load Load piston pin bush). Figure 13 shows the PV values for each
bearing element. The PV value represents the abrasion of
Stroke of VC-T the sliding portion, which is a problem in oscillating
bearings. For each bearing element, the PV value is
80 90 100 110 120
higher than that of the piston pin bush with the
Piston stroke (mm)
conventional mechanism. Therefore, as in the case of the
Large piston pin bush, the copper-alloy bush material intended
deformation Small for high surface pressures, which features excellent
deformation abrasion resistance, was adopted for each bearing
3DUWLDOFRQWDFW element. Another problem with these bearing elements is
Conventional system Multi-link system that, because their sliding is based on oscillation, the oil
Fig. 9 Comparison of bending deformation flow in the circumferential direction of the bearing, which
Load
is induced especially via continuous one-way sliding, is
weak compared to that in the bearings used under
300 rotating conditions. Therefore, these bearing elements
Bending deformation(μm)

120
tend to accumulate 㻢㻜 contaminants, which degrade the
250
cooling performance 㻢㻡 of the lubricating oil at the sliding
㻡㻜
200 surface; this, in turn, may increase the risk of seizures.
Small To address this problem, a groove structure was
150 deformation established on the bearing surface for the entire
100 circumference. In this manner, the oil flow shortage in
Multi-link system the bearing circumferential direction is compensated for
50 by the oil flow in the bearing width direction (refer to
Table 1). In addition, oil passages leading to the sliding
0
surface were established in the components (refer to
60 80 100 120
Figure 14). Thus, oil passages that induce the supply and
Piston stroke (mm) discharge of oil were
㻝㻣㻟 established. However, there was a
Fig. 10 Bending deformation of crankshaft concern regarding the effect of the increase in the surface
pressure due to the decrease in the pressure-receiving
area, resulting from the establishment of the groove
270° 270° along the entire circumference. Therefore, this effect was
assessed through EHL calculations, as well as by using
analyses and actual measurements of the temperature at
Rr-side the sliding portion. First, EHL calculations were
180° 180°
203(MPa) performed for the C-Link big end bearing for the cases
Fr-side 0° 338(MPa) with and without the groove along the entire
0°
circumference. These calculations considered the
90° 90° deformation of the housing portion under a combustion
Multi link system Conventional pressure load, and the results are presented in Figures
(L4 Turbo 2.0L) (L4 2.5L) 15 and 16. Owing to the deformation of the housing, the
center of the bearing width has a larger oil film thickness
Fig. 11 Local pressure of main bearing
than the edges; thus, it is not the main load-bearing

88 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers The Development of Bearings for the Multi-link Crank Mechanism to Achieve Variable Compression Ratio Engine

portion. Therefore, it can be concluded that the function (MPa)

of the pressure-receiving area is limited for the center Engine rear side Without groove

portion. This suggests that the pressure-receiving 124MPa

function of the bearing will not be impaired even if an oil 0

groove is established at the center of the bearing width.

Engine frront side
With groove 270 90
The results of the evaluations of the effects of establishing 180MPa
180
an oil passage obtained using actual equipment are
presented in Figure 17. The temperature of the sliding 0 180 360

portion was measured, and the results indicate that the Fig. 15 Local pressure of C-link big end bearing
temperature decreases. Therefore, the actual
measurements indicate a temperature reduction effect
caused by establishing the oil passages as well as the oil (mm)
groove along the entire circumference. Consequently, the Engine rear side Without groove
effect of improving the function of the bearing is confirmed
2.9μm
via both analytical results and actual measurements.
The U-pin bush and C-pin bush were also evaluated in Engine frront side
With groove
the same manner. Examples of the EHL calculations for 1.0μm
the U-pin bush and C-pin bush are presented in Figures
18, 19, 21, and 22. Similar to the results for the C-Link 0 180 360

big end bearing, these results confirm that the pressure- Fig. 16 Oil film thickness of C-link big end bearing
receiving function will not be impaired on establishing
the oil groove along the entire circumference. It is also
evident that the temperature decreases in the sliding
Temperature rising(℃)

portion. Furthermore, it is considered that the 16

establishment of the oil passages as well as the oil groove
along the entire bush circumference contributes toward 12 Without groove
the constant supply of oil to the sliding portion, which is
similar to the results for the C-Link big end bearing (refer 8
to Figures 20 and 23). Meanwhile, hydrogen-free diamond
like carbon (DLC) material was adopted as the surface 4
coating film for the U-pin and C-pin shaft portions (the With groove
0
shaft side) to withstand high PV value condition. In this
2800 3600 4400 5200 6000
manner, the value was decreased and the hardness
was increased at the sliding portion, such that the seizure Engine speed(rpm)
resistance performance and abrasion resistance Fig. 17 Thermal measurements of C-link big end bearing
performance were improved.

250
(MPa)
Engine rear side Without groove
200
PV(MPa m/s)

150 180
66MPa
100 90 270
Engine frront side
With groove
50 0

0
Piston pin Piston pin bush U-Pin Bush C-Pin Bush C-Link big-end 92MPa
bush(conventional) Bearing 0 360
180

Fig. 13 PV value of bearings for multi-link

Fig. 18 Local pressure of U-pin bush

Main bearing cap

(mm) Engine rear side Without groove
C-pin bush

5.5μm
U-Link
C-Link L-Link
Engine frront side
With groove

5.5μm
Main journal Eccentric journal

Crank pin 0 180 360

Controlshaft Oil splash from
crank pin oil hole Fig. 19 Oil film thickness of U-pin bush
Oil from main bearing cap

Fig. 14 Oil passage to bearings

NISSAN TECHNICAL REVIEW No.88 (2022) 89

 Technical papers The Development of Bearings for the Multi-link Crank Mechanism to Achieve Variable Compression Ratio Engine

Temperature rising(℃)
20 lubricating oil owing to the sliding motion cannot be
expected for this bearing. Therefore, there are concerns
that the abrasion will increase due to direct contact. To
15
Without groove address this problem, an oil groove was established for
the entire circumference, as in the case of the
10
aforementioned oscillating bearings, to ensure lubrication
via oil flow in the bearing width direction. In addition,
5 With groove and oil passage because the flexural rigidity of the control shaft is
significantly higher than that of the crank shaft, the
0 tendency of uneven contact at the bearing can be
2800 3600 4400 5200 6000 mitigated significantly, as compared to that at the main
Engine speed(rpm) bearing of the crank shaft. For the bearing material, the
Fig. 20 Thermal measurements of U-pin bush copper-alloy bush material intended for high surface
pressures, which was used for the oscillating bearing
elements, was adopted at this location to ensure adequate
(MPa) Engine rear side Without groove abrasion resistance.
300MPa

4.5 Designing piston pin bush

180 Compared to the conventional mechanism, the multi-
Engine frront side
282MPa
90 270 link mechanism tends to have a lower PV value at the
With groove
piston pin bush (refer to Figure 24). This is attributed to
0
the decrease in the sliding velocity of the piston pin bush,
272MPa

which occurs because the U-Link (corresponding to the

0 180
300MPa
360
connecting rod in the conventional engine), connected
directly below the piston, maintains a vertical orientation
Fig. 21 Local pressure of C-pin bush
compared to conventional connecting rod when the piston
is lowered during after top dead center (ATDC), the
(mm) Engine rear side Without groove expansion stroke during which the combustion load is
received (refer to Figure 25). This tendency suggests the
6μm ease of reducing the piston pin diameter when designing
the piston for the multi-link mechanism. Therefore, when
Engine frront side considered together with the side thrust force reduction
With groove
effect of this mechanism, this tendency also suggests the
possibility of reducing the piston size and mass.
4.3μm

0 180 360 250

Pmax
Fig. 22 Oil film thickness of C-pin bush 200
Conventional
PV(MPa m/s)

150
40
Temperature rising(℃)

100
Mluti link
30 Without groove 50

20 0
0 90 180 270 360 450 540 630 720
With groove and oil passage Crank angle(deg)
10
Fig. 24 PV of piston pin bush

0 3
2800 3600 4400 5200 6000 Pmax
Engine speed(rpm) 2 Conventional

Fig. 23 Thermal measurements of C-pin bush 1

m//ss))

0 Multi link
VV((m

0 90 180 270 360 450 540 630 720

4.4 Designing control shaft bearing -1
Compared to the aforementioned oscillating bearing
-2
elements, the control shaft bearing is used in the non-
sliding case where no rotating motion or oscillating -3
Crankangle(deg)
Angle(deg)
motion occurs, except for when the compression ratio is
Crank
changed. Owing to this characteristic, the flow of Fig. 25 Sliding velocity of piston pin bush

90 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers The Development of Bearings for the Multi-link Crank Mechanism to Achieve Variable Compression Ratio Engine

5. Organizing characteristics of bearing

elements of multi-link mechanism
Piston pin bush Group 1
The bearing elements of the multi-link mechanism can Group 2 Load : 1.9 times
Load : 0.9～1.0 times Slinding property : Rotating
be classified into three groups according to the usage Slinding property : Oscillating
L-link bearing
environment and characteristics of each bearing. The U-pin bush Main bearing
L-Link bearing and main bearing, classified under group
C-pin bush
1 in Figure 26, are required to function as the rotating
bearing, similar to the connecting rod bearing and main C-link big end
bearing
bearing in the conventional mechanism. Meanwhile, the
input load increases by up to 1.9 times due to the lever
structure of the multi-link mechanism. To accommodate Contshaft bearing
this increased load, the width of the L-Link bearing was Group 3
Load : 0.9 times
increased as a countermeasure against the relative Slinding property : Non-sliding
slipping at the bearing back, which was likely to occur
because of the mechanism and structure. The Fig. 26 The grouping of bearings
countermeasure also led to a reduction in the surface
pressure. The functional reliability of the main bearing 6. Evaluation of friction loss
was ensured by increasing the rigidity of the crank shaft
to prevent uneven contact. The C-Link big end bearing, Bearing elements have been newly added to the multi-
U-pin bush, and C-pin bush, classified into group 2, are link mechanism, which would increase friction loss.
bearing elements that were newly established for the However, each of the added bearing elements was used in
multi-link mechanism. These bearing elements function the oscillating operation, and they all had small
as oscillating bearings, similar to the piston pin bush in diameters. Therefore, the increase in friction loss was
the conventional mechanism. However, with regard to limited. By contrast, owing to the reduction in the piston
the PV value, an index that represents abrasion (refer to thrust force in the multi-link mechanism, caused by the
Figure 13), these bearing elements tend to have larger vertical orientation of the U-Link during ATDC, and
PV values than the piston pin bush. This tendency leads because it was possible to reduce the crank shaft diameter
to a degradation in the seizure resistance. Therefore, to via the aforementioned correspondence, it was possible to
improve the seizure resistance performance, oil passages reduce the friction loss, as compared to that in naturally
were established to draw engine oil into the sliding aspirated engines with the same displacement, as shown
portion. In addition, a groove structure was established in Figure 27.
on the bearing itself along the entire circumference, to
ensure that the cooling effect and foreign matter
discharging effect are enforced and that the seizure 7
resistance performance is secured. The oil groove
6
established for the entire circumference was arranged at Balancer
the center of the bearing width, which is not the main 5 shaft
area subjected to the load. Thus, it was possible to achieve
Mechanical Friction

Pisto
4
the cooling function and foreign matter discharge function
Link pins
without impairing the load capacity. The control shaft 3
bearing, classified into group 3, often functions as a non- Crank main
sliding bearing. For this location, an oil groove was 2 journal
established along the entire circumference, in the same
1
manner as in group 2, to realize the foreign matter Crank pin
discharging function. In addition, the tendency of uneven 0
contact was significantly suppressed by the control shaft, Multi1 link Conventional
2

L4 2.0L L4 2.0L
which had high flexural rigidity. The bush material was
adopted as the bearing material, in the same manner as Fig. 27 Comparison of mechanical friction
in group 2, to ensure adequate abrasion resistance
performance. Although the specific countermeasures
against the increasing load and decreasing oil flow 7. Conclusion
(caused by an inability to supply sufficient oil) may be
different for each group, the functional reliability of each The conclusions of this work are listed below.
group was secured by establishing oil passages to secure It was possible to ensure the functional reliability of
oil flow in the bearing width direction, while preventing each bearing of a multi-link mechanism by implementing
increases in the component sizes and ensuring the load the following countermeasures according to the
capacity at the bearing by focusing on the rigidity characteristics of each bearing.
characteristics. - With respect to the bearings around the crank shaft,
which are present in the conventional mechanism, the

NISSAN TECHNICAL REVIEW No.88 (2022) 91

 Technical papers The Development of Bearings for the Multi-link Crank Mechanism to Achieve Variable Compression Ratio Engine

load applied to the bearings will increase owing to the Source

lever structure. However, it is possible to accommodate to
the tendency of internal rotation and the local increase in
the surface pressure by increasing the rigidity of the Vol.51, No.4 20204316
bearing itself or that of the component parts, such as the
crank shaft.
- Regarding the oscillating bearings and non-sliding
bearings that were newly established for the multi-link
mechanism, the copper-alloy bush material, intended for
high surface pressures, was adopted for improving the
abrasion resistance performance.
- Due to the high hardness of the above material, there
was a tendency for degradation in the seizure resistance
performance owing to foreign matter. In addition, because
the shaft rotation was limited, the oil flow was also
limited, resulting in degradation in the foreign matter
discharging performance. As a countermeasure for these
problems, grooves were established along the entire
bearing circumference and oil passages to ensure an oil
supply path to the grooves; this also increased the oil flow
in the bearing width direction.

Acknowledgements
For this research and development, we received
significant support from the Strategic Development of
Energy Use Rationalization Technology, a NEDO
commissioned project. The authors would like to thank
everyone who helped us with our research.

References
(1) Katsuya Moteki: A study of a Variable Compression
Ratio System with a Multi-Link Mechanism, SAE
Paper 2003-01-0921, (2003).
(2) Takashi Tanabe: The Structure of the Multi-link
Parts for the Variable Compression Ratio Engine
VC-T That Aims at Leveling the Bearing Surface
Pressure, Proceedings of the 2018 JSAE Annual
Congress (Spring), 20185421, (2018)
(3) Katsutoshi Nakamura: The Development of the
Multi-link Crank Mechanism for the Variable
Compression Ratio Engine VC-T Utilizing the
Dynamic Behavior Analysis, Proceedings of the 2018
JSAE Annual Congress (Spring), 20185420, (2018)

Authors

Katsutoshi Nakamura Katsuya Moteki Satoru Ookuma

92 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers

Surrogate Model Development for Prediction of Car

Aerodynamics Using Machine Learning

Kei Akasaka* Fangge Chen** Takehito Teraguchi**

Abstract In the evaluation of car aerodynamics, Computational Fluid Dynamics (CFD) are frequently used
as well as a wind-tunnel. However, the CFD simulations consume a lot of resources and time. In this study, a
surrogate model using machine learning was developed to reduce the amount of resource and time needed for
CFD. In the proposed model, the relation between car shapes and CFD results was learned for rapid prediction of
pressure, velocity, and coefficient of drag for aerodynamics. In this paper, we introduce the proposed model, the
training dataset, the accuracy, and the computational time.

constructed a machine learning model to learn and

1. Introduction estimate the quantified features and characteristics
evaluated via CFD (flow velocity, pressure, and CD) for
In recent years, it has become necessary to further the car shape by using the Gaussian process to estimate
reduce aerodynamic drag in order to improve the fuel the CD values and flow fields of the cars. They trained the
efficiency competitiveness of cars and comply with new model with approximately 800 examples of car shapes
environmental regulations. Computational fluid and CFD results. They reported that the CD and flow field
dynamics (CFD) is often used as well as wind tunnel estimation results matched the CFD calculation results,
experiments to evaluate the aerodynamic performance of which served as reference information for our research.
cars (1,2). However, CFD requires a significant amount of However, it is difficult to learn complex shapes of actual
time and resources. In particular, the resources required cars using their method. Specifically, this method uses a
to perform CFD computations are increasing every year polyhedron called PolyCube and projects the constituent
owing to the necessity of pre-calculations prior to points of the PolyCube onto a car shape within a range in
experiments, the increase in the number of evaluation which the topology does not vary. The coordinates and
specifications, and the demand for improved calculation heights of each projected constituent point are used as
accuracies. Therefore, this study attempted to develop a features. It is necessary to prepare a PolyCube with a
surrogate model that estimates the CFD results (flow topology similar to the shape of the car to be learned.
velocity, pressure, and drag coefficient CD) using machine However, it is difficult to use the same PolyCube for all
learning method, by learning the relationships between car shapes because actual shapes have different
the car shape and the quantities evaluated using CFD. topologies; for instance, minivans, hatchbacks, and
Using this approach, aerodynamic analyses can be sedans have 1-, 2-, and 3-BOX topologies, respectively.
replaced with the surrogate model, such that the amount Hence, a PolyCube suitable for each car model needs to
of CFD computations and time consumption, as well as be prepared for individual learning, which requires time
the corresponding costs, can be reduced. This paper and effort. In addition, the actual cars have complicated
outlines the proposed method, the dataset, and the shapes such as engine room parts, tires, floor components,
validation results.*Received on November 10, 2020. and door mirrors, and it is difficult to prepare a PolyCube
Presented at the Technical Sessions of the Autumn with a topology that can express such complicated shape
Congress of the Society of Automotive Engineers of Japan elements. Therefore, the results obtained by Umetani et
on October 23, 2020. al. can only be applied to relatively simple car shapes.
By contrast, Guo et al. (4) used a distance function that
2. Machine learning model represents the distance from an object as input data and
constructed a machine learning model that estimates
2.1 Previous Studies & Related Studies two- and three-dimensional flow fields (corresponding to
Several studies similar to the present investigation the flow velocity vector and pressure) based on a
have been reported. For instance, Umetani et al. (3) convolutional neural network (CNN) (5). Because the

*Integrated CAE PLM Department **Nissan Research Center Mobility & AI Laboratory

NISSAN TECHNICAL REVIEW No.88 (2022) 93

 Technical papers Surrogate Model Development for Prediction of Car Aerodynamics Using Machine Learning

distance function can be applied to complicated shapes

and can also be extended to three dimensions relatively
easily, the model structure proposed by Guo et al. was
adopted for the flow field (i.e., flow velocity and pressure)
estimation.
However, because CD cannot be estimated using the
model of Guo et al., a separate surrogate model for
estimating CD is required. Therefore, we proposed a
surrogate model that estimates CD (6) based on the input
of voxel data. Because the estimation accuracy of CD is
satisfactory, the surrogate model structure of CD was
adopted in this study as well.
In related studies, a single model that can learn both
the flow field and CD has not been proposed thus far.
Therefore, in this research, two different machine
learning models based on related studies were developed Fig.2 Structure of proposed model for flow field estimation
to estimate the flow field and CD.

2.2 Outline of estimation model for estimating flow

field
The basic configuration of the flow field estimation
model adopts the network structure proposed by Guo et
al., as shown in Fig. 1. This model consists of an encoder
and a decoder connected by a fully connected layer. When
applying the model practically, such as in this study, it is
necessary to increase the size of the distance function to
be input according to the degree of complexity of the
shape. In addition, the intermediate layers between the
encoder and decoder need to be multilayered to improve
the estimation performance. However, the use of
multilayered intermediate layers easily causes gradient
disappearance or gradient explosion problems during
training, making it difficult to proceed with training.
Therefore, to solve the gradient disappearance and
explosion problems, in this study, residual blocks (7) were
used instead of the fully connected layer in the model
structure of Guo et al. The model proposed herein is
illustrated in Figs. 2 and 3. This model inputs a distance
function and outputs (a) the three components of the flow
velocity vector and (b) the pressure. For the encoder and
decoder, batch normalization (8) and instance
normalization are used to suppress the disappearance
(9)
Fig.3 Details of encoder, residual blocks and decoder of
of the gradient. proposed model for flow field estimation

2.3 Outline of estimation model for estimating CD

The estimation model for estimating CD is described in
this section. We previously proposed an estimation model
that estimates CD based on the input of voxel data (6).
Although the model is based on Voxnet, proposed by
Maturana (10), global average pooling, instead of a fully
connected layer, is used according to the research of Lin
et al. (11) to reduce the amount of calculation when
converting the intermediate layer into one dimension.
Because the estimation accuracy of CD is satisfactory, the
estimation model structure based on our previous study
was adopted in this study as well.
Fig.1 CNN structure of prediction model for flow filed proposed
by Guo 4

94 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers Surrogate Model Development for Prediction of Car Aerodynamics Using Machine Learning

Table 1 Parameter of distance function

Table 2 Optimizer, loss function and hyper-parameters

Fig.4 Structure of proposed model for coefficient of

aerodynamic drag CD estimation

However, in terms of convenience, it is desirable to

avoid the use of different input data when estimating the
flow field and CD. Therefore, the input data for the CD
estimation model are converted from the voxel into the 3. Dataset and validation results
distance function to unify the flow field and CD estimation
model input data into the distance function. The network 3.1 Dataset
structure in which the input data are unified into a The dataset is shown in Fig. 6. One case includes six
distance function is show in Fig. 4. pieces of information: the distance function, three
components of the flow velocity vector, pressure, and CD
2.4 Distance function and hyper-parameters value. Only the distance function is input into the
First, the distance function to be input into the estimation models, and the flow fields and CD value are
estimation model can be described as follows. The the outputs of the models respectively. Training was
distance field of the three-dimensional distance function performed such that these outputs matched the flow
used in this study is illustrated in Fig. 5. The colors velocity, pressure, and CD value of the training case. For
indicate the distances from the car shape at each grid this study, 1,123 cases were prepared and divided in a
point. The distance function was created as follows. A ratio of 10:1 into training and testing datasets. The
bounding box was set around the car shape, and datasets included six types of cars: sedans, coupes, SUVs,
orthogonal equal-interval grid points were arranged in hatchbacks, pickup trucks, and light cars. The proportions
the bounding box. The shortest distance from the car of car types included in the training and test datasets
shape was calculated for each grid point to create the were approximately the same. The flow velocity vectors,
relevant distance field. The sizes of the bounding boxes, pressures, and CD values used for training and test were
grid pitches, and numbers of grids are provided in Table calculated using commercial CFD software. The
1. For the distance function used to predict CD, a smaller calculations were performed under straight running
grid pitch is employed to ensure that even a slight shape conditions at a car speed of 33.3 m/s and Reynolds number
change can be captured. of 1.0 107. Because a grid unrelated to the distance
The optimizers, loss functions, and hyper-parameters function was used in the CFD calculations, the CFD
used in this study are listed in Table 2. calculation results were mapped to the grid for the
distance function.

Fig.5 3-dimensional distance function around car shape

NISSAN TECHNICAL REVIEW No.88 (2022) 95

 Technical papers Surrogate Model Development for Prediction of Car Aerodynamics Using Machine Learning

Fig.7 Histogram of mean absolute error in x-component of

velocity vector on test cases

The mean absolute error (MAE) of the x-direction

component of the velocity vector is presented in Fig. 7.
The mean error of the x-component is 0.2–1.1 m/s, which
is within the error range of the car speed wind (0.6%–
3.3%). In addition, the mean errors of the y- and
Fig.6 Distance function, flow fields and aerodynamic drag CD z-components were smaller than that of the x-component.
in dataset

In addition, because the distance function and CFD 20

results are dimensional information, normalization was
performed before using them for learning and testing in 15
this study.
10
A NVIDIA DGX-1™ (1 GPU) was used as a
computational resource for learning, and Tensorflow 2.0 5
was employed as a library for machine learning. The time
required for learning the flow field was approximately 4 0
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
days, whereas the time for learning CD was approximately
0.5 days. In addition, by calculating the number of epochs
shown in Table 2, it was confirmed that the values of the
Fig.8 Histogram of mean absolute error in pressure on test
final loss function of the three components of the velocity cases
vector, pressure, and CD value are reduced to the extent
shown in Table 3.
The MAE results for the pressure estimated by the
Table 3 Errors of loss functions in velocity vectors, pressure and proposed model and evaluated at 120,000 points in the
CD at last epoch of proposed model same manner as the velocity vector are presented as a
histogram in Fig. 8. The MAE of the pressure is 2.0–9.0
Pa, which is within the error range of approximately
0.3%–1.3% of the dynamic pressure based on the car
speed wind. There were no significant differences in the
error tendencies among the car types in terms of both
3.2 Flow field validation results velocity and pressure.
Next, the flow field estimated by this method was The pressure and velocity magnitude distributions
considered as follows. Each test case had a flow velocity obtained from the CFD calculations and proposed model
vector of 120,000 points (100 40 30). The mean error are compared in Figs. 9 and 10, respectively. Cases A–D
of those 120,000 points was considered using the mean are typical examples of the test cases, and the velocity
absolute error (MAE) shown in equation (1): and pressure errors in each case are shown in Fig. 7 and
Fig. 8, respectively. In both Fig. 9 and Fig. 10, the upper
rows show the CFD calculation results, whereas the
lower rows indicate the estimation results of the proposed
model to be compared for each case.

where ai and yi represent the CFD results and the

estimation results of the proposed model, respectively. In
addition, n represents the number of grid points, which
was 120,000 in this case.

96 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers Surrogate Model Development for Prediction of Car Aerodynamics Using Machine Learning

3.3 Effect of the residual blocks

The network structure proposed by Guo et al., in which
a fully connected layer was adopted for the connection
between the encoder and decoder, and the network
structure proposed in this study, in which residual blocks
were adopted instead of a fully connected layer, are
compared in this section. The velocity magnitudes at the
central cross-section of the car are shown in Fig. 11. The
results of the CFD calculations and training using the
residual blocks are presented in Figs. 11 (a) and (b),
respectively. Here, the results of the proposed model are
similar to the CFD calculation results. By contrast, in the
results of training using the fully connected layer, shown
in Fig. 11 (c), the tendency of the velocity magnitude is
different from that in the CFD results. Owing to the
gradient explosion, the weight optimization failed during
training, and the value of the loss function of the velocity
Fig.9 Comparison of pressure between CFD and proposed vector was 0.1. This value is larger than that of the loss
model in central cross-section function of the proposed method, as shown in Table 3,
and it did not decrease even when the number of epochs
was increased. According to these results, the residual
blocks were considered to be effective.

Fig.11 Comparison of velocity magnitude between Guo s model

and proposed model in central cross-section

Fig.10 Comparison of velocity magnitude between CFD and

proposed model in central cross-section
3.4 CD validation results
The errors in the CD values estimated using the
The pressure distribution at the central cross-section proposed model are depicted in Fig. 12; the errors
of the car is shown in Fig. 9, where the pressure is obtained in the training and testing cases are shown in
converted into a dimensionless value by using the (a) and (b), respectively. Table 4 shows the results of
dynamic pressure. In each case, the pressure distribution evaluating the error using the mean absolute percentage
estimated by the proposed model reproduces the tendency error (MAPE), described in equation (2), and the standard
of the CFD results. deviation:
The velocity magnitudes in the central cross-section of
the car are shown in Fig. 10. The velocity was made
dimensionless by using the car speed. The proposed
model qualitatively reproduces the CFD results in terms
of the size of the rear stream region behind the car and
the tendency of the velocity magnitude, which are
important, especially in aerodynamic evaluations.
According to the comparison results shown in Fig. 9
and Fig. 10, it was confirmed that the proposed model can
reproduce not only the qualitative tendency of the flow
field even in case D, in which the error between pressure
and velocity is relatively large, but also the characteristics
of the flow fields of different car types. Fig.12 Histograms of error in CD

NISSAN TECHNICAL REVIEW No.88 (2022) 97

 Technical papers Surrogate Model Development for Prediction of Car Aerodynamics Using Machine Learning

Table 4 Error in CD of train and test

The CD values and CFD results of the test cases
estimated by the proposed model are compared using a
scatter plot, as shown in Fig. 14. The vertical axis
represents the CD values estimated by the proposed
model, and the horizontal axis represents the values
calculated using CFD. The testing case includes six car
types, and the CD values estimated by the proposed model
for each car type are generally within 0.012 of the CFD
results.
In the initial phase of car development, aerodynamic
evaluation is performed on multiple car design proposals;
In equation (2), ai and yi represent the CFD result and however, various car shapes have been proposed, and CD
the estimation result of the proposed model, respectively, also varies considerably more than the error of the
and n represents the number of training cases or the proposed model. Therefore, in the initial phase, the
number of testing cases. proposed model can be used to evaluate the superiority or
Because the standard deviation and MAPE of the inferiority of CD for each proposed design, and it is
training case (a) shown in Fig. 12 and Table 4 are expected that the frequency of CFD usage and the amount
sufficiently small, it is evident that the dataset has been of calculations can be reduced.
trained appropriately. For the testing case (b), however,
the error is larger than that for the training case, and 3.5 Estimated time required for machine learning
there exists a problem with the generalization model
performance. The times required to predict the velocity, pressure,
and CD in Case A described above when using the
proposed model are shown in Table 5 and compared with
the times required when the commercial CFD software is
used.

Table 5 Comparison of computational time between CFD and

proposed model

Fig.13 Error history in CD with increase of train cases

In the case of CFD, approximately 24 h was required

The error trend with an increase in training cases is for mesh preparation and flow calculation (1). This
shown in Fig. 13. The horizontal axis represents the calculation time was recorded when using Intel® Xeon®
number of training cases, and the vertical axis represents (approximately 80 million meshes and 256 CPU cores).
the error (standard deviation). As the number of training For the proposed model, a notebook PC (CPU: Intel®
cases increases, the error of the testing cases decreases. Core-i5™) was used to compute the distance function;
Therefore, it is considered effective to increase the this required approximately 11 minutes, whereas the
number of training cases in order to improve the predictions of the velocity, pressure, and CD only required
generalization performance. 3 seconds. Therefore, the proposed model is deemed
useful for checking the qualitative flow field and CD for
various shape designs within a short period of time.

4. Summary and conclusion

In this study, a practical surrogate model that can

predict the flow field and CD around a car with a
complicated three-dimensional shape was developed.
1 We developed a model using residual blocks, instead
of the fully connected layers used in the network
structure of Guo et al., to effectively avoid gradient
explosion and accurately predict the flow field around
the car shape.
2 It was confirmed that the flow field estimated by the
Fig.14 Comparison of coefficient of aerodynamic drag CD
between CFD and proposed model proposed model could qualitatively reproduce the

98 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers Surrogate Model Development for Prediction of Car Aerodynamics Using Machine Learning

CFD results.
3 The error (MAPE) in the CD value estimated by the
proposed model was 1.4%.
4 The proposed model is a useful surrogate model that
can replace CFD because the flow fields and CD values
for various shape designs can be estimated within a
significantly short period of time. Using the proposed
model, it is expected that the resources required for
aerodynamic analyses can be reduced considerably.

References
(1) K. Akasaka, et al.: Simultaneous Estimation of
Aerodynamic and Thermal Performances Using
CFD, Proceedings of SAE Seminar, No.96-05, pp.11–
14 (2005).
(2) M. Arai, et al.: Development of the Aerodynamics of
the New Nissan Murano, SAE Technical Paper,
2015-01-1542 (2015).
(3) N. Umetani, et al.: Learning Three-Dimensional
Flow for Interactive Aerodynamic Design,ACM
Trans. Graph., Vol. 37, No. 4, Article 89 (2018).
(4) A. Guo, et al: Convolutional Neural Networks for
Steady Flow Approximation, Proceedings of the
22nd ACM SIGKDD International Conference on
Knowledge Discovery and Data Mining, pp.481–490
(2016).
(5) Y. LeCun, et al.: Gradient-based learning applied to
document recognition, Proceedings of the IEEE,
86(11), pp.2278–2324 (1998).
(6) K. Akasaka, et al.: Development of A Tool for
Interactive Prediction of Car Drag Coefficient Using
Machine Learning, at The 34th Annual Conference
of the Japanese Society for Artificial Intelligence,
2O6-GS-13-02 (2020).
(7) K. He, et al: Deep residual learning for image
recognition, Proceedings of the IEEE conference on
computer vision and pattern recognition, pp. 770-
778 (2016).
(8) Ioffe, et al.: Batch normalization: Accelerating deep
network training by reducing internal covariate
shift, arXiv preprint arXiv:1502.03167 (2015).
(9) Ulyanov, et al. :Instance normalization: The missing
ingredient for fast stylization, arXiv preprint
arXiv:1607.08022 (2016).
(10) D. Maturana, et al.: VoxNet: A 3D Convolutional
Neural Network for real-time object recognition,
IEEE/RSJ International Conference on Intelligent
Robots and Systems (IROS), Hamburg, pp. 922-928
(2015).
(11) M. Lin, et al.: Network in network, arXiv:1312.4400
(2013).

Source

Vol.52, No.3 20214248

NISSAN TECHNICAL REVIEW No.88 (2022) 99

 Technical papers Surrogate Model Development for Prediction of Car Aerodynamics Using Machine Learning

Authors

Kei Akasaka Fangge Chen Takehito Teraguchi

100 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers

A Study on Pitch Characteristic to Reduce Line Trace

Deviation in Small Steering Angle

Mitsunori Tao* Naoya Machida** Yutaka Hayashi*** Ken Nagasao****

Abstract When driving along a straight roadway, a driver regularly makes minute steering adjustments
in order to maintain the vehicle s position within the driving lane. In the research of the previous report, the
quantitative relationship of several steering characteristics in this scenario was clarified for accurate line tracing.
However, it was found that the deviation of the tracing line by trials was not small at the same time. This paper
clarifies one of the factors of the deviation and its improvement by adding minute pitching motion to the vehicle in
proportion to the steering angle.

variations can be reduced by adding a certain amount of

1. Introduction pitch motion to the steering angle in order to reduce
variations. This hypothesis was verified using a driving
For the daily driving scenarios encountered by typical simulator, and the quantitative effect of the variation
drivers, most driving is performed along straight reduction was extracted. Following our previous work,
trajectories, including operations performed for fine which clarified the vehicle characteristics that enable
adjustments in the travel path with small steering drivers to drive accurately, on average, along a target
angles. According to our survey conducted in Europe and driving line, this research focused on the vehicle
Japan, steering with angles of 5 or less, which is characteristics that enable accurate driving with little
considered as minute steering, accounts for more than variations between trials.
half (51.1%) of all steering across all vehicle speed ranges To explain this research, an outline of our previous
and approximately 80% (79.6%) in zones with higher work and the variation in the line trace deviations, which
speed limits of 80 km/h or more, which is supported by is the subject of the present study, is first provided. Next,
some data. The traceability of a target driving line with the driver steering mechanism, which is considered to
such small steering angles is an important aspect of cause variations in the minute steering angle range and
performance that is frequently encountered by many also defined as a hypothesis in the previous work, will be
drivers; however, its mechanism has not been fully described; the reasons for choosing the pitch as the
clarified thus far. Therefore, at actual development sites, vehicle behavior to be added to reduce variations are also
there are many cases where the development of this explained. Subsequently, the quantitative pitch
ability relies on tuning an actual vehicle. characteristics that can reduce the variation in the line
With regard to this issue, our previous paper 1) clarified trace deviations based on the driver sensitivity
the quantitative relationships between the steering force characteristics (related to the pitch motion obtained from
and yaw characteristics with respect to line traceability; experiments) are hypothesized, and the results of the
this was achieved by quantitatively determining the dead experimental verification are reported.
zone of the driver in the minute steering angle range and In this study, two experiments were conducted with 25
the dead band of the yaw characteristics of the vehicle male drivers in their 20s and 50s who drove their cars on
and then analyzing the relationship between the a daily basis. The tests were examined and approved by
evaluated line traceability defined in the study and the the Experimentation Ethics Committee of Nissan Motor
two dead zones. However, a new issue was also observed: Co., Ltd., and informed consent was obtained from the
the line trace deviations of the drivers varied considerably test participants.
between trials.
In this study, considering that one of the causes of this 2. Outline of and issues in our previous study
variation in the line tracing deviations originates from
the steering mechanism of the driver within the minute In our previous study, an evaluation course was
steering angle range, it was hypothesized that the established, consisting of an R1200 curve and its

*Nissan Research Center Mobility & AI Laboratory **Powertrain and EV Performance Engineering Department
***Customer Performance and Test Engineering Methodology Innovation Department ****Vehicle Performance Engineering Department

NISSAN TECHNICAL REVIEW No.88 (2022) 101

 Technical papers A Study on Pitch Characteristic to Reduce Line Trace Deviation in Small Steering Angle

relaxation section; this required a steering input of

approximately 5 from a straight line at a constant speed 0.4
Average : 0.21deg/s
of 80 km/h, as shown in Fig. 1, and the drivers drove

Yaw Rate [deg/s]

0.3
along the course with the target line. The actual
deviations in the travel trajectories during driving were 0.2
defined as line trace deviations, whose mechanism was
clarified. 0.1

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Driver No.
80kph Straight Fig.3 Driver s Yaw-Rate Sensitivity Experimental Result
Line
30m 0.8
Evaluation Section
0.6

Yaw rate [deg/s]

Fig.1 Evaluation Course
Fig.1 Evaluation Course 0.4
Dead
As shown in Fig. 2, the line trace deviations can be 0.21deg/s
Band
accurately indicated using two types of characteristics: 0.2
the steering reaction force characteristics, which are
highly sensitive to the minimum steering angle that the 0.0
driver recognizes as steering input to the vehicle, and 0.0 1.0 2.0 3.0 4.0
the yaw-rate dead band characteristic (shown in Figs. 3 Steering angle [deg]
and 4), which is highly sensitive to the steering angle
Fig.4 Yaw-Rate Characteristic in Small Steering Angle
generating the minimum yaw rate of 0.21 /s that enables
the driver to sense that the vehicle has moved. This
was clarified through the driver–vehicle model hypothesis
and the simulator-based verification conducted by the
Inside >>
Line Trace Deviation [m]

evaluators. The verification results are shown in Fig. 5. 0.15

For the specifications at the front of the graph, indicating 0.10
significant steering forces and large yaw-rate dead bands,
0.05
the line trace deviation is negative, causing the vehicle to
move outward with respect to the target line. Conversely, 0.00
<< Outside

under the specifications for weak forces and small dead -0.05
zones, the vehicle moves inward. Therefore, the target -0.10
line can be traced accurately by setting the steering force
-0.15
and dead band along the thick broken line.
0 -50
0.5 1 0 -25
1.5 2 25
50
16
14 Fig.5 Experimental Result of Line Trace Deviation

12
However, as a new issue, it was confirmed that the line
Steering Force [N]

10 trace deviations varied considerably between the trials

8 performed by individual drivers. The line trace deviations
for a certain driver across six trials under the specifications
6 of a steering reaction force of +25% and a yaw-rate dead
4 band of 1 , which allow the average of the line trace
Average Angle of deviations of 25 drivers to be almost zero, are shown in
2 Steering Input Recognition
by Drivers Fig. 6. On average, they appear to coincide with the thick,
0 dashed target line; however, certain variations exist.
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
Steering Angle [deg]
Fig.2 Steering Force & Driver s Input Recognition Angle

102 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers A Study on Pitch Characteristic to Reduce Line Trace Deviation in Small Steering Angle

band in the yaw rate such that the vehicle can initiate
Y[m] the yaw movement simultaneously, the vehicle movement
1.5 Target Line that matches the feeling of the driver can be realized. As
Vehicle Trace Line a result, on an average, the line trace deviations are
1
reduced. However, the driver can only sense the
0.5 movement due to this yaw rate as visual information
when the value exceeds 0.21 /s and the steering angle
0
reaches approximately 2 . In other words, in the section
-0.5 X[m] where the steering angle is 1 –2 , it can be considered
0 30 60 90 that the driver is performing blind steering without
obtaining feedback regarding vehicle movement in the
Fig.6 Line Trace Deviation by a Driver
visual sense and relying solely on the information from
the steering reaction force.
The variations in the line trace deviations of 25 drivers
based on their respective standard deviations under the
Driver recognize the Input
same specifications are depicted in Fig. 7, with the
average value of 0.047 m. This average is not sufficiently
small, even compared with the fluctuation in the line
trace deviations under the different specifications defined
for the average value of all the trials by all the drivers, as Blind Steer
shown in Fig. 5. with only Force Information

0.08 Driver detect the Motion

Line Trace Deviation

Average : 0.047m
0.21deg/s
0.06 Dead ↓ Undetectable Motion ↓
Band
[m]

0.04
1deg 2deg Steering Angle
0.02
Fig.8 Driver s Steering Process in Small Steering Angle
0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Driver No. It is clear that visual feedback is one of the most
important types of information for stable and accurate
Fig.7 Average of Line Trace Deviation by 25 Drivers
driving, and it is conceivable that even a small steering
angle range, such as 1 –2 , has an unexpectedly large
Following our previous work, which clarified the influence on stable and accurate driving. In particular,
vehicle characteristics that enable drivers to drive with regard to the line traceability at minute steering
accurately, on average, along a target driving line, this angles of less than 5 , as in this case, the effect is
research focused on the vehicle characteristics that considerably large and likely one of the major factors in
enable accurate driving with little variations between the variation of the line trace deviations.
trials. Therefore, visual information is fed back to the driver,
indicating that the vehicle begins to move in the yaw
3. Causes of variation in line trace deviations direction at a steering angle of approximately 1 , which
and reduction measures the driver can recognize as a steering input. When
movement increases with the steering angle, the driver
3.1 Steering mechanism for minute steering angles will be able to steer in a more accurate and stable manner;
and variation in line trace deviations consequently, variations in the line trace deviations will
Various factors can be considered as the causes of the be reduced.
variation in the line trace deviations for each trial of a
driver, such as differences in the line of sight, how the 3.2 Vehicle behaviors for providing visual
steering wheel is held, or the degree of concentration information to drivers
when driving. This work focused on the steering As mentioned earlier, the yaw movement itself does
mechanisms of the drivers with minute steering angles, not enable the driver to sense that the vehicle responds to
as hypothesized in the previous work, and conducted a the steering and begins to move in the yaw direction; this
detailed analysis. is because a yaw rate of 0.21 /s or higher is required for
As shown in Fig. 8, when steering from straight the driver to detect movement in the yaw direction. By
running toward the target line, the driver first recognizes contrast, if a higher yaw rate threshold for the driver to
that the steering input to the vehicle is commenced at a sense vehicle start movement is set, the yaw-direction
steering angle of approximately 1 based on the steering movement of the vehicle may become excessive and the
reaction force information. Therefore, by setting a dead vehicle will not travel accurately along the originally

NISSAN TECHNICAL REVIEW No.88 (2022) 103

 Technical papers A Study on Pitch Characteristic to Reduce Line Trace Deviation in Small Steering Angle

targeted driving line. and the evaluation comments confirmed that all the
In this study, based on the vehicle movements in the drivers could feel the pitch movement as soon as it was
six translational and rotational motion directions, the initiated; therefore, the data in Fig. 10 can be considered
movement to convey to the driver as visual information as the operational time differences of the drivers between
on behalf of the yaw movement of the vehicle that the the moment at which the pitch rate of 0.3 /s was generated
vehicle is responding immediately after the start of the and the time at which the button was pressed by the
yaw movement was examined. First, the roll, pitch, and individual drivers. The average value (0.51 s) of the
bounce motions were selected as candidates, excluding operation times for all the drivers was subtracted from
the plane motions in the front–back and left–right all the experimental data for correction.
directions, which directly affect the travel trace of the
vehicle, such as the yaw-direction motion. Subsequently, 0.7
referring to previous studies 2) where human sensitivity
0.6
was investigated, the motion in the pitch direction, to
which it was reported that humans have the same high 0.5
visual sensitivity as the yaw-direction motion, was finally 0.4

Pitch Angle [deg]

selected. 0.3
In previous literature 3), 4), it was reported that the 0.2
feeling of steering, or the equivalent characteristics,
0.1
could be improved by imparting a downward pitch to the
vehicle during steering, as compared to the cases where 0
no pitch motion was provided or an upward pitch was -0.1
provided. Therefore, in this study, downward pitch -0.2
motion was imparted to the vehicle during steering. 0 0.05 0.1 0.15 0.2 0.25 0.3
Pitch rate [deg/s]
4. Hypothesis for reducing variations in line Fig.10 Driver s Sensitivity for Pitch Motion
trace deviations using vehicle pitch motion

4.1 Experiments on driver sensitivity to pitch motion At low angular velocities with pitch rates of 0.05 /s or
To determine the minimum pitch motion that a driver less, all the drivers sensed the movement at a pitch angle
can detect, a sensitivity experiment was conducted using of 0.15 –0.2 (refer to the left side of the graph); the
a driving simulator. The experimental setup is depicted drivers could not sense pitch rates lower than 0.05 /s and
in Fig. 9. When a constant downward pitch was provided the movement was detected for the first time by the size
to the vehicle, the drivers were instructed to gaze at the of the integrated pitch angle. By contrast, at a high pitch
front view of the screen and to press the switch at hand rate of 0.3 /s, the drivers sensed the movement with
the moment they sensed the pitch movement. A total of almost no pitch angle (see the right side of the graph),
eight experiments with different pitch rates, ranging which is likely because the drivers could detect the speed
from an extremely low pitch rate of 0.0125 /s to a at this pitch rate, as described above. The central region
relatively high pitch rate of 0.3 /s, were conducted five between these two regions is considered as the section
times each, and the pitch angles sensed by the drivers where the drivers could sense the pitch rate and the pitch
were recorded for each trial. angle as a whole; hence, the sensitivity graph presents a
downward-sloping line connecting the two regions. Next,
using this sensitivity line, the pitch motion that enables
drivers to sense the initiation of yaw motion was
hypothesized.
Vehicle
Pitch Motion 4.2 Hypothesis of pitch characteristics to reduce
variation in line trace deviations
The pitch characteristics were hypothesized based on
the pitch motion required for drivers to recognize a
Fig.9 Pitch Sensitivity Experimental Scene steering input and sense the vehicle response near a
steering angle of 1 (at which the yaw motion commences
The measurement results are presented in Fig. 10. The and increases). Although various generated pitch motion
experimentally specified pitch rates (eight specifications) patterns can be considered, such as steering angle
are provided on the horizontal axis, whereas the pitch proportionality, steering wheel angle rate proportionality,
angles sensed by the driver at each pitch rate are denoted and the addition of non-linearity to each, this study
on the vertical axis. The measurement data obtained aimed to verify the mechanism in the simplest possible
from the 25 experimental evaluators (drivers) are manner and selected the pitch angle (downward)
indicated by the dots, and the average values are proportional to the steering angle, shown in Fig. 11, to
connected by the solid line. For the highest pitch rate of formulate a hypothesis of the required proportionality
0.3 /s, the results of the push-button switch operation constant. In addition, the experimental results of our

104 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers A Study on Pitch Characteristic to Reduce Line Trace Deviation in Small Steering Angle

previous research confirmed that drivers steer at an angle of 1 . However, when K = 0.08, the pitch motion has
angular velocity of approximately 3 /s in the target scene; already considerably exceeded the driver sensitivity line
hence, this steering wheel angle rate was added to the at a steering angle of 1 ; thus, it is considered that the
prerequisites. driver can sense the motion before the steering angle
reaches 1 . When K = 0.06, the pitch motion almost falls
on the driver sensitivity line. In other words, this value of
Steering Angle [deg] K enables the driver to sense Steering Angle
the pitch [deg]at a
motion
steering angle of 1 . K=0.00

Pitch Angle [deg]

Pitch Angle [deg]

(Downward)
(Downward)

K=0.02

K=0.04
As shown in equation (3), which uses the pitch angle
and steering angle, the visual information (pitch motion)
K=0.06
enables the driver to recognize a steering input and sense
K=0.08
the vehicle response near a steering angle of 1 (at which
Fig.11 Pitch Characteristics for Driver s Information Steering
the yaw motion of the vehicle Angleand
commences [deg]
increases).
Therefore, it is assumed that the variation in the steering
angles and the line trace deviations can be reduced by
adding visual feedback in the minute steering angle
range, wherein the drivers were deemed to have
performed blind steering. The abovementioned hypothesis
was verified as discussed in the following section.

5. Verification of hypothesis

The hypothesis was verified using the model of a

certain C-segment vehicle and the driving simulator that
reproduced the evaluation course shown in Fig. 1, as in
our previous research. The outline, method, and results
Based on the driver sensitivity line for the pitch motion of the experiment are as follows.
(Fig. 10), obtained through the experiment, and the pitch
characteristics during steering, defined in equations (1) 5.1. Overview of vehicle model and simulator used
and (2), the proportionality constant K, which enables in experiments
drivers to sense the pitch motion at a steering angle of 1 , The main parameters of the vehicle model used in the
was obtained. A graph of the pitch rates and pitch angles experiment are listed in Table 1.
generated at a steering angle of 1 when K was changed
from 0 to 0.08 in 0.02 increments, with respect to the
Table
Table11Vehicle ModelParameter
Vehicle Model Parameter
driver sensitivity line for the pitch motion, is shown in
Fig. 12. Parameter unit Value
Vehicle Mass kg 1670
0.20 Vehicle Yaw Inertia kgm2 2600
Vehicle Pitch Inertia kgm2 2300
0.15 Height of C.G. m 535
Front axle～C.G. m 1.08
Pitch Angle [deg]

0.10
Rear axle～C.G. m 1.62
K=0.08 Front Cornering Power kN/rad 62.5
K=0.06
0.05
Rear Cornering Power kN/rad 95.5
K=0.04
K=0.02 Steering Gear Ratio - 15.0
K=0
0.00
0 0.05 0.1 0.15 0.2 0.25 0.3
Pitch rate [deg/s] An external view of the driving simulator used in the
experiment is provided in Fig. 13. This simulator is
Fig.12 Pitch Motion by each K at Steering Angle 1deg
equipped with a cabin with a hexapod mounted on slide
rails (motion range of 22 6 m) and 64 linear motors that
When K is 0.04 or less, the pitch motion generated at drive the hexapod with high precision; thus, this
the steering angle of 1 does not reach the driver simulator can reproduce the actual vehicle motions under
sensitivity line, indicating that this setting does not many scenarios. 5)
enable the driver to sense the pitch motion at a steering

NISSAN TECHNICAL REVIEW No.88 (2022) 105

 Technical papers A Study on Pitch Characteristic to Reduce Line Trace Deviation in Small Steering Angle

Steering Angle [deg]

K=0.00

Pitch Angle [deg]

(Downward)
K=0.02
360°
Screen
Turn table K=0.04

K=0.06
K=0.08
Steering Angle [deg]
Fig.15 Pitch Characteristics

Fig.13 Driving Simulator 5.3 Experimental results

The measured line trace deviations for a certain driver
Next, the accuracy of this simulator was verified as with respect to different values of K of the pitch are
follows. The actual data measured using the C-segment presented in Fig. 16. The right-turn data were mirror-
vehicle and simulator for the dependence of the steering compensated in the left-turn direction. The variation in
force and yaw rate on the steering angle are presented in the line trace deviation from the target line decreases as
Fig. 14; notably, both exhibit high reproducibility. K increases. In addition, the numerical values in each
graph, which represent the calculated line trace
deviations, quantitatively confirm the tendency of the
variations.

Y[m] Y[m]
Target Line
1.5 Vehicle Trace Line
1.5
1
0.5
Dev.= 0.052
0.5
0
-0.5 X[m] -0.5
0 30 60 90
(a)K=0

Y[m] Y[m] Y[m]

Fig.14 Driving Simulator Accuracy Confirmation 1.5 Dev.= 0.044 1.5 Dev.= 0.037 1.5
1 1
0.5 0.5 0.5
5.2 Experimental method 0 0
The experiment was conducted using the same 25 -0.5 X[m] -0.5
X[m] X[m] -0.5
30 0 60 90 90 0 30 60
drivers as in the pitch rate sensitivity experiment. The
(b)K=0.02 (c)K=0.04
vehicle speed was fixed to 80 km/h, and the drivers were
instructed to operate the steering wheel to solely drive Y[m] Y[m]
Y[m]
the vehicle along the center of the lane. With regard to 1.5 Dev.= 0.025 1.5
1.5 Dev.= 0.050
1 1
the experimental specifications, the yaw-rate dead band
0.5 0.5
0.5
and steering reaction force were set to 1 and the standard 0 0
+25%, respectively, which afforded the smallest average X[m] -0.5 X[m] -0.5
X[m] -0.5 X[m]
value (almost zero) of the line trace deviations for all 0 30 60 90 0 30 60 90
trials in our previous research. Furthermore, K of the (d)K=0.06 (e)K=0.08
pitch angle to the steering angle was changed five steps
Fig.16 Line Trace Experimental Result
from 0 to 0.08 in increments of 0.02. For each step, the
experiments were conducted by performing a total of 30
randomly arranged trials, with three right steering and The variations in the line trace deviations under
three left steering experiments. The steering angle and different pitch specifications, calculated from the
the line trace deviations were measured. measurement results of the travel trajectories of 25
drivers, are depicted in Fig. 17, with the horizontal axis
representing K. It is evident that the variation in the line
trace deviations tends to decrease as K increases. The
thick solid line, which connects the average values
obtained under the different specifications, indicates that
the deviation is the smallest at K = 0.06 and then
increases again at K = 0.08. Thus, the hypothesis
established in this study was verified. Further, when K =

106 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical papers A Study on Pitch Characteristic to Reduce Line Trace Deviation in Small Steering Angle

0.06 was added, the variation in the line trace deviations

0.15
could be reduced by approximately 33%, as compared to

Line Trace Dev. Ave. [m]

that in the case where no pitch motion was added. 0.10
0.05
0.00
0.08
-0.05
Line Trace Deviation [m]

0.06 -0.10
-0.15
0 0.02 0.04 0.06 0.08
0.04
Proportional Constant K [-]
▲33%
0.02 Fig.18 Line Trace Deviation Average Experimental Result

0
0 0.02 0.04 0.06 0.08
6. Summary and conclusion
Proportional Constant K [-]
Considering that one of the causes of the variation in
Fig.17 Line Trace Deviation Experimental Result the line trace deviations originates from the blind
steering section in the minute steering angle range, it
Subsequently, the change in the line trace deviation was confirmed that this variation can be reduced by
with increasing K was evaluated. There was no significant adding a certain amount of pitch motion as visual
difference between the deviations for K = 0 and K = 0.02, information within the blind steering section.
because K = 0 does not generate any pitch motion; Based on investigations of driver sensitivity to the
furthermore, even when K = 0.02, the driver cannot pitch motion, the appropriate amount of pitch motion to
detect the movement before the steering angle reaches be added was hypothesized and experimentally verified,
approximately 3 and the yaw movement is near 2 . and the quantitative effect of reducing the variation in
Hence, it is likely that no visual information was provided the line trace deviations was extracted. In addition to the
within the blind steering section of 1 –2 . Similarly, when vehicle characteristics that enable drivers to drive
K = 0.04, the driver can detect the pitch movement at a appropriately, on average, in the minute steering angle
steering angle of approximately 1.5 ; therefore, even if it range, which were clarified in our previous research, the
is not possible to completely cover the blind section, as in other characteristics necessary for accurate driving were
the case of K = 0.06, it is possible to provide visual identified.
information to the driver for approximately half of this
section, thereby reducing the variation to a certain References
extent. However, when K = 0.08, it is conceivable that the (1) Mitsunori Tao et al., A Clarification of Relationship
pitch motion was detected by the driver before the of Steering Force and Yaw Characteristics to Line
steering angle reached approximately 1 . Generally, Traceability in Small Steering Angle , Transactions
when the driver recognized the steering input and the of the Society of Automotive Engineers of Japan
vehicle actually commenced the yaw movement, there Vol.51 No.3 pp.428-433 (2020)
was an increase in the variation, as compared to that in (2) Katsuhiko Fukui et al., Technology for Improving
the case of K = 0.06; this was attributed to the additional Roll Feeling Based on Visual Characteristics ,
visual information that hindered accurate operation by Transactions of the Society of Automotive Engineers
the driver. of Japan Vol.40 No.5 pp.1185-1190 (2009)
Finally, the effect of pitch motion on the average value (3) Hideki Sakai et al., Damping Force Control That
of the line trace deviations was studied. In this study, the Emphasizes Transient Turning Sensation ,
purpose of adding the pitch motion was to reduce the Transactions of the Society of Automotive Engineers
variation in the line trace deviations by providing drivers of Japan Vol.43 No.3 pp.709-716 (2012)
with visual information regarding the pitch motion, in (4) Makoto Yamakado, Study on 3-Dimensional Ideal
addition to the steering reaction force information, on Body Behavior by EV's Precise Braking and Driving
which drivers relied at the start of vehicle movement. Force Control , Journal of the Japan Society of
Therefore, it is considered that the pitch motion required Precision Engineering Vol.84 No.9 pp.765-768 (2018)
for this purpose does not affect the average value of the (5) Masayuki Imamura et al., Development of High-
line trace deviations. The average values of the line trace Performance Driving Simulator , Nissan Technical
deviations for the 25 drivers who participated in this Review (Dynamic Performance Technology) No.83
study are shown in Fig. 18. It is confirmed that the pp.60-65 (2018)
addition of the pitch motion does not significantly alter
the values.

NISSAN TECHNICAL REVIEW No.88 (2022) 107

 Technical papers A Study on Pitch Characteristic to Reduce Line Trace Deviation in Small Steering Angle

Source

Vol.52, No.2 20214137

Authors

Mitsunori Tao Naoya Machida Yutaka Hayashi Ken Nagasao

108 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical Awards

2021 JSAE Award The Outstanding Technical Paper Award

An Approach to Exploring Vehicle Motion to Enhance Ride

Quality of Passenger

Mitsuhiro Makita* Akihiro Matsushita**

Yoshinori Kusayanagi*** Masahiro Miura***

Abstract A passenger s passive motion, as a result of the vehicle s motion, impacts ride quality. So, we built
the methodology to explore desirable passenger body motion, under a fixed driving task, by employing inverse
dynamics analysis based on optimal control. First, we embedded a passenger body model in a vehicle model using
a set formula. Next, we developed a soft sensor model so that acceleration and jerk could be integrated into the
cost function for minimizing by optimal control. Then, inverse dynamics simulation was becoming executable to
derive vehicle motion meets to the aiming passenger s passive motion. Designing passenger motion will lead to the
understanding of human perception on vehicle.

increasingly advanced, setting the travel path is

1. Introduction important for controlling vehicle dynamics. Research has
focused on generating travel paths while considering the
This study aimed to establish a method for analyzing motion perception of the driver(1) and also on setting
the effects of vehicle dynamics on non-driving passengers travel paths that can reduce the occurrence of car sickness
from the perspective of ride quality. With regard to ride among passengers, which can degrade the ride quality.(2)
quality, ride comfort is normally considered, which is As a research approach that focuses on passenger
affected by the vehicle motion in the direction behavior, the use of a model that represents the behaviors
perpendicular to the road surface. In the future, when of the bodies of passengers can be considered. In cases
autonomous driving becomes common, driving tasks will where the vehicle motion is known, a method for analyzing
be conducted independently from that of humans. the behavior of passengers has been proposed,(3) and the
Therefore, to create a comfortable moving space, the impacts of the travel path have been investigated.(4)
vehicle dynamics in the direction parallel to the road Previous value-creating research includes analyses of
surface will also be an important research subject. the timing of the inertial force application to the bodies of
The body of a passenger is moved by an inertial force passengers and its effects on the autonomous posture
that is generated by the acceleration of the vehicle. control of passengers;(5) the creation of the vehicle
Therefore, the passenger may experience some discomfort acceleration that stimulates the motion perception of
due to the body movement. Although the application of passengers, enables passengers to recognize vehicle
an inertial force to the body of the driver, which is the motion, and assists passengers in preparing their posture
result of his or her own driving style, might be a positive for motion has also been evaluated.(6) The approach of
experience for some drivers, this force is usually a cause focusing on the senses and body movement mechanisms,
of discomfort for passengers. Hence, body movement including car sickness, is gaining importance as the
must be restrained by the seat and a restraint device if understanding of humans has become increasingly
high-acceleration motion, such as emergency evasion, is advanced.
necessary. However, within the range of normal driving, Considering the aforementioned preceding research,
it is desirable to enhance ride quality by devising the the establishment of a method for exploring vehicle
vehicle motion such that the effect of the inertia force on behaviors that suppress the body behavior of passengers
the passengers is reduced. and the amount of accompanying motion perception
In this regard, it is important to determine how vehicle under an assumed travel scene can be expected.
dynamics should be perceived to enhance the ride quality Therefore, to respond to such expectations in this study,
of passengers. The main factors related to vehicle the research that the authors have been conducting
acceleration are <path> and <speed>. Therefore, it is the exploration of vehicle motion that optimizes
possible to mitigate the inertial load acting on the the evaluation values set by applying optimal control
passenger by driving the vehicle along an appropriate was designated as an inverse vehicle dynamics
path. As autonomous driving technology is becoming analysis method,(7) and the scope of this method was

*Nissan Research Center Mobility & AI Laboratory **Customer Performance and Test Engineering Methodology Innovation Department
***AD/ADAS and Chassis Control System Engineering Department

NISSAN TECHNICAL REVIEW No.88 (2022) 109

Technical Awards : 2021 JSAE Award The Outstanding Technical Paper Award - An Approach to Exploring Vehicle Motion to Enhance Ride Quality of Passenger

expanded to include passengers. Table.1 Variables and functions for optimal control
If optimal control is employed, it is possible to derive
the desired vehicle motion and the control inputs for its
realization by converting the matter into deriving the
minimum value of the evaluation function. In other
words, the characteristics of vehicle motion that can
enhance ride quality can be extracted using the evaluation
function by using various values that are considered to
have impact on ride quality.
The following section of this paper describes the case
where inverse vehicle dynamics analysis is performed
using optimal control and indicates that the analysis
method is limited when applied for analyzing the behavior
of passengers. Subsequently, technical solutions for
overcoming these limitations are described. Lastly, the
method for exploring the ride quality of a passenger is
presented using transient motion as a case example and
by describing the calculations, actual vehicle testing, and
evaluations.

2. Ride quality analysis using inverse vehicle

dynamics

If the forward vehicle dynamics analysis is considered

as set the control inputs and derive the corresponding
vehicle motion, inverse vehicle dynamics analysis(7) can
be considered as reverse the cause-and-effect
relationship, specify the desired vehicle motion, and
derive the control inputs and vehicle motion that satisfy
the specified motion. For the inverse direction, the
calculation method will differ depending on what is
specified for the desired vehicle motion. In this study, the
travel scene was specified as the precondition, the motion Fig.1 Schematic flow of the inverse vehicle dynamics
comfort of the passengers was explored, and inverse methodology by optimal control: Ride quality is reflected
to stage cost to describe motion design.
vehicle dynamics analysis was performed using optimal
control (the schematic flow of the inverse vehicle
dynamics methodology is shown in Fig. 1, and the To evaluate the ride quality of passengers, in the
variable names are provided in Table 1). vehicle model, reproducing the three-dimensional
Optimal control uses the stage cost function , movements of the vehicle body in which the passengers
which is described using variables such as the input ride is necessary. In addition, when formulating the
variables and state variables that represent the optimal control using the ordinary variational principle,
state of the subject of motion . The objective is to the vehicle model (in the form of a state equation) is
derive the vehicle motion that yields the minimum value added to the stage cost function using the adjoint
of the function. From a different perspective, the vehicle variables. However, there are limitations in the model
motion is designed using the stage cost function. The description for the evaluation function, which are related
forms of functions that can actually be used for description to the time integration of the added values. Specifically,
are limited, such as the quadratic forms of the variables. in the description of the state equation, the differential
However, vehicle motion can be designed at least using quantities of the state variable needs to be
this function. described <Explicitly> using the state variable, as shown
The travel scene is specified by setting the initial and in equation (2), and not in the form of equation (1):
terminal values of the state variables and control inputs
in the vehicle motion task. In other words, the state of the
vehicle motion is fixed at both ends (the analysis becomes
a boundary value problem). Therefore, it is possible to
derive the control inputs that minimize the stage
cost function value. In a previous study,(7) the authors developed a vehicle
model to express the vehicle dynamics accurately,
including the load transfer among the wheels, while
observing the restrictions for describing the equation.
Table 2 and Fig. 2 present the vehicle model, including

110 No.88 (2022) NISSAN TECHNICAL REVIEW

Technical Awards : 2021 JSAE Award The Outstanding Technical Paper Award - An Approach to Exploring Vehicle Motion to Enhance Ride Quality of Passenger

the vehicle dynamics and state variables. The coordinate state variables, as shown in equation (2). If the inertia
system is based on the unsprung position. For the tire force of the passenger motion acts on the vehicle, it will
characteristics, the response characteristics using the not be possible to make <explicit> expressions using the
relaxation length were added to the simplified Magic state variables of the vehicle and passenger. Therefore,
Formula. Refer to the previous research of the authors(7) only the passive motion due to the inertia force of the
for details regarding the vehicle model using equation (2). vehicle motion is considered, and a model for the motion
of the upper body in the lateral direction (roll), without
Table.2 Variables and input functions of vehicle model
separating the head portion, is proposed.

Table.3 Variables and parameters of passenger model

Fig.2 Two track and body motion vehicle model conforms to

equation (2): Refer to author s previous research(7)
Fig.3 Passenger seating point configuration: Passenger upper
body (including head) is simplified to a single inverted
pendulum, then affected with vehicle motion.
Using the inverse vehicle dynamics analysis described
thus far, passenger ride quality enhancement is
discussed. First, the passenger body model must be As the first step, the inertia force of the vehicle motion,
placed inside the vehicle body model. When doing so, which is applied to the passenger, is expressed. The
passenger modeling is conducted using the form of passenger seating point in the fixed coordinate system of
equation (2), as in the case of the vehicle model. the vehicle body as well as the rigid upper body model that
Next, to deal with the ride quality in motion design, a has a pivot at that point are presented at the left and
value related to the passenger perception of the vehicle center of Fig. 3; the variables and constants that are
motion needs to be provided as a state variable . For related to the passenger model described below are listed
example, the vestibular organ recognizes the translational in Table 3. Owing to the vehicle motion, the three-axis
acceleration and angular velocity. Because, as shown in translational acceleration shown in equation (3) is applied
equation (1), acceleration is not a state variable but is to this pivot (terms that have little effect are omitted):
rather the differential of the state variable,
acceleration cannot be incorporated into the stage cost
function when creating the motion design policy. Thus, a
method of converting acceleration into a state variable is
required.
The abovementioned considerations suggest that two
issues need to be addressed to apply the inverse vehicle
dynamics analysis to passenger ride quality. The In the next step, the force acting on the center of
resolution of these issues is discussed in the subsequent gravity (CG) of the upper body is expressed. To this end,
section. it is necessary to derive the acceleration at the CG of the
upper body. Here, the acceleration is given by the second-
3. Expansion to passenger ride quality order differentiation of the upper body CG position (pivot
evaluation point + effect of upper body roll posture angle), which is
expressed using the coordinate system based on the
3.1 Passenger body behavior modeling ground. The force acting in the lateral and vertical
The passenger body motion model was created while directions, as shown on the right side of Fig. 3, are
observing the restriction of <explicitly> expressing the approximated as the acceleration by multiplying the

NISSAN TECHNICAL REVIEW No.88 (2022) 111

Technical Awards : 2021 JSAE Award The Outstanding Technical Paper Award - An Approach to Exploring Vehicle Motion to Enhance Ride Quality of Passenger

mass of the upper body and are expressed in equation (4): incorporated into the evaluation function.
The acceleration can be converted into a state variable
by differentiating the equation of motion, which is
expressed as a second-order differential equation, once
more to a third-order differential equation (and to the
fourth order for conversion to the jerk state variable).
However, this increases the number of equations that
need to be solved, leading to a calculation burden and
making this approach unsuitable for inverse vehicle
The equation of motion of the upper body about the dynamics calculations. To address this problem, a state
pivot in the roll direction, caused by the forces FyB and observer is used to derive the approximate state variables
FzB acting on the upper body CG, is expressed as equation that can be incorporated in the evaluation function.
(5). For the roll stiffness and damping, the effects of the
supporting force acting between the passenger and the
seat surface, friction, etc. are included as approximate
values:

In the last step, equations (3) and (4) are substituted Fig.4 Acceleration and jerk detection with soft sensor (physical
into the equation of motion of the upper body (equation image of state observer) : Acceleration and jerk at the
(5)) to determine the roll angle acceleration of the upper attached body are approximated to state variables.
body <explicitly>, as shown in equation (6):
To formulate optimal control by combining the vehicle
and passenger models, the state observer is used as a soft
sensor, as shown in Fig. 4. Specifically, the state equation
described in equation (7) is employed, and the acceleration
at the passenger body or at the measurement target
position of the vehicle body is used as input .
Utilizing this method, the translational acceleration, roll
acceleration, and jerk at the mounting position can be
observed as approximate state variables :

The differential quantities of the state variables of the The approximation accuracy of the state variables is
vehicle model, and , are included in the determined by setting the c and k values in equation (7).
derived passenger model, as shown in equation (6). Fig. 5 shows the characteristics corresponding to different
However, if equation (6) is converted from the equation of natural frequencies when the set values are altered. The
motion to the equation of state and if the converted set values are selected such that the gain, which
equation and the vehicle model are considered as a set of represents the degree of approximation, is close to 1; this
simultaneous equations, replacement with the state implies that the characteristics will be flat.
variables can be performed such that the equation will
converge to the form of equation (2). In this manner, the
formulation of the optimal control used for the inverse
vehicle dynamics analysis was realized.

3.2 Converting motion perceived by passengers into

state variables
The physical quantities that can be expected to be
related to the ride quality may include the translational
and angular accelerations applied to the passenger body.
However, as mentioned in the previous section, quantities Fig.5 Frequency response characteristic of state observer:
other than the state and input variables cannot be Four cases of translational and angular accelerations.

112 No.88 (2022) NISSAN TECHNICAL REVIEW

Technical Awards : 2021 JSAE Award The Outstanding Technical Paper Award - An Approach to Exploring Vehicle Motion to Enhance Ride Quality of Passenger

At this point, the effect of the approximation accuracy three phases:

of the state observer in the inverse vehicle dynamics (i) Set the travel scene to be evaluated (vehicle
analysis is reviewed. In the lane change behavior motion task).
analysis, discussed in the next section, the state observer (ii) Set the ride quality factors in the evaluation
was mounted at the vehicle body CG position, and function and calculate the optimal solution.
calculations were performed such that the lateral (iii) Reproduce the optimal solution as the vehicle
acceleration at the vehicle CG was minimized. The motion and conduct tests and evaluations.
calculation results are presented in Fig. 6. With regard to The characteristics that can be obtained as the optimal
the stage cost function values, it is evident that there solution are the control inputs and motion of the target
exists an appropriate value range within which the vehicle. Therefore, when reproducing the vehicle motion
approximation accuracy can be increased by increasing in the last phase, the control inputs are provided to the
the response frequency. In other words, if the response actual vehicle, or the motion is applied to the passenger
frequency is excessively high, the differential equation of using equipment that can reproduce the acceleration
the state observer stiffens the overall system to be solved. space (as shown in Fig. 7).
Thus, when calculating the values, it is necessary to
consider the trade-off between accuracy and the
accumulation of calculation errors.

Fig.7 Methodology to construct acceleration space

In the first phase, i.e., setting the travel scene for ride
quality evaluation, there will be no significant difference
in the results of the evaluation function related to motion
Fig.6 The degree of approximation of state observer (stage perception if the travel scene is the steady driving mode;
cost is quadratic value of lateral acceleration): There is
an appropriate value for the resonance frequency of however, the results may depend on the types of control
state observer. inputs (such as steering only and the addition of active
acceleration/deceleration using the brake/accelerator).
Using the state observer described above, it was Therefore, in this work, transient motion is utilized as
possible to approximate the values of the acceleration the travel scene example because it is easier to obtain the
and jerk (which are related to the motion perceived by difference in motion perception under this case than in
passengers) and to design the evaluation function by the other situations (Fig. 8).
incorporating the approximated values in the evaluation
function as state variables.

4. Exploring ride quality factors by designing

passive motion of passengers

4.1 Designing motion for ride quality evaluation

Here, the method of exploring the vehicle motion to
enhance the ride quality is specified. In this method, it is Fig.8 Example of vehicle motion task: Lane change as a
necessary to consider how to use the inverse vehicle typical transient vehicle motion scene.
dynamics to derive the minimum value of the evaluation
functions, which include the amount of motion perception Next, the design specifications for the vehicle motion
that occurs due to the body motion of the passengers. in this example are presented in Table 4. In the vehicle
An important aspect in the vehicle evaluation is to motion task, a lane change scene was set, in which the
maintain identity, that is, the condition that should be following terminal conditions were returned to straight-
kept the same when making comparisons. The logic of line travel (=0): the control inputs and the track angle
this matter has been presented in previous reports.(7) and yaw angle velocity, which are the state variables for
When solving the functions in the inverse vehicle plane motion on the road. Assuming passenger ride
dynamics analysis, identity is ensured by setting the quality enhancement, the following specifications were
state variables and input variables for the initial and set for the evaluation functions: the steering control
terminal motion conditions for the vehicle motion task input minimizing specification V0, assuming the ride
shown in Fig. 1. Using this method, a motion design quality of the driver, and the body roll angle acceleration
based on the evaluation functions is established. minimizing specification R1. For comparison, specification
The design flow can be classified into the following R2 was also set; R2 represents the body roll angle

NISSAN TECHNICAL REVIEW No.88 (2022) 113

Technical Awards : 2021 JSAE Award The Outstanding Technical Paper Award - An Approach to Exploring Vehicle Motion to Enhance Ride Quality of Passenger

acceleration minimization specification added with the

conditions for returning the roll angle acceleration and
angle velocity to zero at the terminal.

Table.4 Vehicle and passenger motion design specification:

Three condition to exemplify design methodology.

Fig.10 Steering pattern and vehicle C.G. path (with enlarged

view) to satisfy design specification of vehicle and
passenger motion: Inverse vehicle dynamics analysis
fixes both input and state variables simultaneously.

4.2 Calculating passenger acceleration space based The calculated optimal vehicle motion, lateral
on design acceleration of the vehicle CG, and roll posture angle of
The calculations were performed in accordance with the passenger body are depicted in Fig. 11. According
the motion design specifications. Here, the characteristics to the calculation results, the transition of V0 appears as
of the passenger body model need to be determined (the a sine function. Compared to V0, the peak value of the
settings of the vehicle characteristics are the same as passenger behavior is suppressed for R1 (in which case
those mentioned in a previous report by the authors(7)). the passenger behavior is optimized). In R2 (in which
Because the passenger model is simple, the accuracy case additional terminal constraints are implemented),
improvement effect will be small, even if the thoroughness the posture angle converges to a certain value at the end
of the model characteristics definition is improved. of the calculation.
Therefore, for the characteristic values of the shape and
mass, values close to those of the actual vehicle test
evaluator, as mentioned in subsection 4.3, are used. The
body stiffness, KB , near a lateral acceleration of 2 m/s2
was set by referring to the distribution results of the
measurements of the test subjects (n = 15). Examples of
the stiffness measurements are presented in the box-
and-whisker plot in Fig. 9. The value when the passenger
was facing down and reading text was used for the
calculations, and the passenger was considered to be
seated on the left side of the rear seat.

Fig.11 Vehicle and passenger motion to satisfy the design

specification: Calculated results.

4.3 Evaluation using the acceleration (vehicle) space

In the test method, a steering robot was mounted on a
vehicle to reproduce the calculated steering waveform
Fig.9 Passenger body roll stiffness (box-and-whisker plot) and such that acceleration was applied to the passenger
other characteristics: Passenger model parameters were seated on the rear seat. To correspond to the actual
fixed from subject measurement (n=15) and estimation.
environment, a term that represents the effects of slopes
was included in the vehicle model when performing the
The settings necessary for the calculation were calculations. The test values were compared with the
implemented. Thereafter, inverse vehicle dynamics calculated values for the last 4 seconds.
analysis was performed to calculate the control inputs The passenger has a body build close to the average of
and vehicle motion (Fig. 10 shows the path of the vehicle the measurement results of the passenger characteristics
CG) that minimizes the evaluation functions while and is an expert in vehicle dynamics evaluation. The
satisfying the terminal conditions. passenger can recognize the start of the vehicle motion
based on the inevitable driving noise of the steering
robot. However, it should not be easy for the passenger to

114 No.88 (2022) NISSAN TECHNICAL REVIEW

Technical Awards : 2021 JSAE Award The Outstanding Technical Paper Award - An Approach to Exploring Vehicle Motion to Enhance Ride Quality of Passenger

predict the vehicle motion because the passenger can

only view the external environment through the 5. Considerations for exploring ride quality
surrounding view. Therefore, the test environment
should be similar to that in the passive motion evaluation. In the passenger body model, increasing the degree of
The transitions in the passenger posture were estimated freedom (such as by separately defining the head portion)
using the movement detected by the displacement sensor is possible. However, this will make it difficult to
attached to the rear of the neck. determine the appropriate model characteristic values,
The test results are presented in Fig. 12 in a manner such as the stiffness at the neck portion, even if the scope
that corresponds to the calculation examples (Fig. 11). It of the model is limited to passive motion. It is assumed
can be considered that the reproducibility of the that passengers will perform various activities and be in
acceleration space was ensured. The deviation between various states in the vehicle. Therefore, in addition to
the passenger posture transition and the calculated value addressing the complexity of the passenger model,
(the results in the figure were multiplied with a coefficient determining the settings of the characteristic values of
of 2.0) is attributed to the differences between individual the model needs to be resolved in future research.
evaluators, as the passenger model characteristic values However, the motion perceived by the passenger
in Fig. 9 were used for the calculation (the values differ cannot be mechanically calculated based on the body
by a factor of approximately 2 for some data within the model. Thus, it may be important to add a model that
quartile range). It is judged that the relative relationship represents the human inner sensory characteristics such
between the specifications is in-line with the predictions that the evaluation function is configured using the
based on the calculations. actual perception values. Fig. 13 shows the body behavior
(calculated values) under the specifications, where the
body roll angle acceleration is used as the evaluation
function. R1 does not include terminal constraints,
whereas R2 has terminal constraints. Compared to
specification R1, specification R2 causes more lateral
movements during the intermediate test time. It is clear
that the body roll movement must occur early to control
the body behavior at the terminal. Although the roll
motion is suppressed, the translational acceleration is
increased. Therefore, the mechanism of the perceptual
recognition of translational and rotational motions must
be considered.

Fig.12 Vehicle and passenger motion of actual vehicle test:

Every data curve was averaged (n=5).

In the subjective evaluation, it seems that the motion

is smooth while the body posture transition is large with
specification V0, whereas it seems that the body roll
posture transition is small with specification R1 (in
which the body roll angle acceleration is suppressed). In
other words, the result that the body seems to have
moved along with the vehicle motion was obtained.
However, in the case of specification R2 (in which
additional terminal constraints for the roll posture are
implemented), the evaluation movement at the early
stage of the motion was large compared to V0 was
obtained. In addition, the evaluation that while the
high-frequency motion was somewhat noticeable, it was
possible to feel the load reduction due to the suppressed Fig.13 Passenger upper body posture transition: Earlier motion
generation is required for terminal posture control, then
body motion was obtained for the body roll suppression needs quick initial motion, too.
terminal condition, with the objective of maintaining
body posture. In summary, it was possible to parameterize
the ride quality in terms of both the measurements and The degree of freedom of the motion design depends on
subjective evaluations using the passenger passive the limitations in the form of the evaluation function.
motion design. The variables that can be incorporated are typically
limited to a combination of quadratic forms. In addition,
in the travel scene expanded to multiple control inputs,

NISSAN TECHNICAL REVIEW No.88 (2022) 115

Technical Awards : 2021 JSAE Award The Outstanding Technical Paper Award - An Approach to Exploring Vehicle Motion to Enhance Ride Quality of Passenger

the terms corresponding to each control input must be set conducted after obtaining informed consent from the
in the evaluation functions. Otherwise, simultaneous participants.
optimization is not possible (ensuring orthogonality of
the terms). To advance research on ride quality, the References
flexibility of expressing the evaluation functions needs to (1) Tatsuya Yoshimoto, Takanori Fukao, Yasuyoshi
be improved. Yokokohji, Hiroshi Inou: Path Generation of Lane
Change in Dynamic Environments Considering
6. Conclusion Characteristics of Driver s Operation, Transactions of
the JSAE, Vol.47, No.2, p.633-638 (2016)
In this study, the effects of vehicle dynamics on (2) Takahiro Wada: On Motion Sickness in Highly
passengers were considered from the perspective of ride Automated Vehicles, Proceedings of the JSAE Annual
quality, and a <methodology for testing/evaluation> for Congress (2016), 20165263
improving ride quality was presented herein. Passenger (3) Katsunori Yamada, Harutoshi Motojima, Yuichi
ride quality can be improved by determining the vehicle Kitagawa, Tsuyoshi Yasuki: Investigation of Relations
motion if it is assumed that the ride quality depends on between Occupant Kinematics and Supporting by the
the passive motion load applied to the body and on the Seat in Lane Change Maneuvers, Proceedings of the
perception caused by the vehicle motion within the JSAE Annual Congress (2016), 20165176
acceleration space. (4) Makoto Yamamoto, Tadashi Yamamoto, Shuji
Therefore, the desirable motion was set as the Nishiyama: Proposal for a Transition Curve with
evaluation function, and motion design was attempted Smoothly Changing Curvature Using Multiple
using inverse vehicle dynamics analysis, where the Clothoid Curve, Transactions of the JSME, Vol.83,
vehicle motion was derived from the optimal control. An No.852 (2017)
observer was established to incorporate a passenger body (5) Tohru Yoshioka, Masato Abe, Makoto Yamakado,
model which estimate the load, and acceleration which Yoshio Kano, Yusaku Takeda, Kazuhiro Takemura,
used to estimate the motion perceived by passengers into Takatoshi Tsukano, Fuminori Kato, Daisuke Umetsu:
the evaluation function. In this manner, an inverse Analyses on Vehicle Dynamics Improvements from
dynamics analysis that combines the passenger and the Human Aspects with G-Vectoring Control,
vehicle was realized. Proceedings of the JSAE Annual Congress (2016),
It is necessary to subjectively evaluate whether the 20165250
form of the evaluation function successfully indicates the (6) Alexander Lange, Martin Albert, Karl-Heinz
improvements in the ride quality of the designed vehicle Siedersberger, Klaus Bengler: Ergonomic Design of
motion and to also investigate the passengers perception the Vehicle Motion in an Automated Driving Car, 6th
of vehicle motion. For these purposes, this paper presents International Conference on Applied Human Factors
an evaluation flow using a specific example. In the first and Ergonomics (AHFE) (2015)
phase, a travel scene is assumed. In the next phase, the (7) Mitsuhiro Makita: A Solution to Apply Load-Transfer
vehicle dynamics are designed using a candidate Including Vehicle Model to Optimal Control
evaluation function setting. In the last phase, evaluation Formulation, Transactions of the JSME, Vol.81,
is performed by placing a passenger in the designed No.826 (2015)
acceleration space (the actual vehicle in this study).
This research will be extended such that the vehicle Source
motion gentle to the passenger can be incorporated into
the technological developments in autonomous driving
via control logics and artificial intelligence, among others. Vol.50, No.5 20194687

(Addendum): The tests conducted in this study were

examined and approved by the Experimentation Ethics
Committee of Nissan Motor Co., Ltd. The tests were

Recipient

Mitsuhiro Makita Akihiro Matsushita Yoshinori Kusayanagi Masahiro Miura

116 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical Awards

2021 JSAE Award The Outstanding Technical Paper Award

Development of driving assistance system in multi-lane

highway

Yohei Taniguchi* Katsuhiko Degawa* Seiji Tokunaga*

Shimpei Nagae** Koji Sasaki**

vehicle) and determines the correct position on the road.

1. Introduction The 3D high-definition map data include information
detailing the road structure, the number of lanes, and the
A new driving assistance system that guides a vehicle positions of merging points, branching points, and
until reaching the specified highway exit and enables intersections. These information allow system control
hands off under each certain condition has been that foresee the road structure ahead and realize smooth
implemented on a production vehicle. The system can driving as that of an experienced driver. A sensor
guide a vehicle through a multi-lane highway,working in mounting configuration and a 360 sensing image
tandem with the navigation-system-plannedroute. It also diagram are presented in Figures 2 and 3.
can operate in hands off mode, provided thedriver
consistently monitors the objects and events andresponds 4. Driving assistance on a multi-lane highway
immediately to operate the steering wheel inresponse to
the road, traffic and vehicle conditions. The details of this Acquiring information pertaining to the lanes ahead is
system are described in this paper. one challenge in realizing lane-change assistance.
At the start of the lane change, the camera may not be
2. Function overview able to collect comprehensive lane information for the
point where the lane change ends. Accordingly, if
This system offers navigation linked driving such as information is only received from the camera, the
merging on to a highway when destination is set in the appearance of a dividing line forbidding lane changes
navigation system. When the navigation linked driving is during the lane change (or other operations) may prevent
initiated, system supports driving through interchanges the lane change. Meanwhile, because the 3D high-
and assists overtaking another vehicle until reaching the definition map data include lane-level data (e.g.,
relevant exit. Provided that the driver always attends to curvatures and dividing lines), the acquisition of road
the road ahead and can operate the steering wheel with shapes beyond the camera s detection range allows lane-
certainty and immediately in response to the road, traffic, change opportunities to be judged by considering the
and vehicle conditions, the system operates with hands section of any lane between the start and end points.
off with in the same lane and provides wide range driving In this system, lane-level information included in the
maneuver support to the driver. In addition, using the 3D high-definition map data are used to create the route
navigation data and information of the surroundings plan such that the system can propose changing lanes at
from a 360 degrees sensing, the system determines an a suitable timing for the purpose of providing passing
appropriate lane-change starting point for branching or assistance and for proceeding along the route
overtaking whilst following the planned route, and recommended by the navigation system.
proposes the relevant actions to the driver. Lane-change
assistance is initiated by the driver when both hands are 5. Conclusion
placed on the steering wheel and the switch to approve
the proposal is operated. An operational image is shown In this development project, the navigation system,
in Figure 1. 360 sensing, and 3D high-definition map data were
combined to realize a driving assistance system for use
3. Sensor configuration on multi-lane highways. This technological development
helps advance driving assistance technology and also
This system integrates and uses seven cameras, five introduces possibilities for further developments.
radars, 12 vehicle-mounted sonars, a global navigation
satellite system, and 3D high-definition map data to
collect information from the surroundings (i.e., not only
from the front but also the right, left, and rear of the

*AD/ADAS Advanced Technology Engineering Department

**Customer Performance and Test Engineering Methodology Innovation Department

NISSAN TECHNICAL REVIEW No.88 (2022) 117

 Technical Awards 2021 JSAE Award The Technological Development Award - Development of driving assistance system in multi-lane highway

高速道路複数車線の運転支援システムの開発

System available
Highway
Start Local road

Start of Trip Destination

Hands-off single-lane Assists with Assists with lane changes

driving vehicle passing and highway exiting

End

Fig.1 System Overview

図１ System Overview
Front camera
Sonar
AVM camera
Side radar
Front radar
Driver monitoring camera

Sonar
Fig.3 360
図３ degree
360sensing
degree sensing
AVM camera
図 3 360 degree sensing
Side radar

なタイミングでの車線変更提案を実現した．
Source

5 │おわりに
Vol.75 No.06
本開発では，ナビゲーションシステム，360 度セ
ンシング，3D 高精度地図を組み合わせることによ
Fig.2 Onboard
Onboard technology
り，高速道路複数車線の運転支援システムを実現し
図２ technology
た．本技術開発により，運転支援技術の進化に貢献
で，追い越し支援やナビゲーションシステムの推奨 し，さらなる発展の可能性を示すことができたと考
Recipient
する経路に沿って進む目的で，システムによる適切 える．

Yohei Taniguchi Katsuhiko Degawa Seiji Tokunaga Shimpei Nagae Koji Sasaki

118 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical Awards

2021 JSAE Award The Outstanding Technical Paper Award

Strength evaluation of CFRP tank based on meso-scopic

zooming analysis

Shinichiro Takemoto*

Abstract For stringent cost reduction with sufficient safety of a high pressure hydrogen tank made by carbon
fiber reinforced plastic (CFRP) installed in a fuel cell vehicle, a method to precisely predict the burst pressure of
the tank has been intensely required. CFRP layers in the tank consist of complicated meso-structures with carbon
fiber bundles and matrix resin, in which crimps of the bundles are stacked by a filament winding process. We
propose a method to predict the burst pressure by zooming analysis of conventional axisymmetric continuum model
of the whole tank with the meso-scale model. The proposed method is validated through the burst test of tank with
accurate prediction of the burst pressure by means of the carbon fiber rupture in the meso-scale model.

structure of CFRP layers (Iida, 1978), the comparison

1. Introduction between laminate theory and FEM analysis of
axisymmetric shell model (Hirase et al., 1991), and the
To promote the widespread use of fuel cell vehicles eEvaluation of tank with uniform stress surface in
(Ohnaka, 2009), there are growing demands for reducing domepart (Liu et al., 1997). In each of these studies, the
the weight and price of high-pressure hydrogen fuel prediction accuracy for tank rigidity and strength was
tanks (Council for a Strategy for Hydrogen and Fuel improved within the framework of continuum mechanics.
Cells, 2019). To withstand a pressure of 70 MPa, a fuel In addition, the fracture mechanism has been clarified
tank employs a liner, which seals the hydrogen inside an experimentally (Arai et al., 2007). Although these
inner aluminum or plastic tank composed of carbon fiber- methodologies are effective for evaluating the macroscopic
reinforced plastic (CFRP) laminated using the filament deformation in tanks, it is difficult to accurately evaluate
winding (FW) method (Maki et al., 1970). One effective the fracture phenomenon caused by the increase in the
means of reducing the weight and price of these tanks is local stress and strain that occurs in a meso-structure
to minimize the amount of fiber used by optimizing the comprising carbon fiber bundles and resin. Meanwhile,
laminated structure; to this end, a highly accurate the evaluation for viscoelastic constitutive equation of
method for predicting the tank burst pressure is required. CFRP laminates using homogenization methodhave tried
The fracture strength of tanks laminated using the FW the combination of the micro-analysis of CFRP unit cells
method is governed by the local fracture phenomenon and the macro-analysis of the entire member (Arai et al.,
that occurs in the meso-structure composed of carbon 2008). However, thus far, fracture evaluations for actual-
fiber bundles and matrix resin. An approach equivalent size components, including tanks, have not been achieved.
to a conservative strength evaluation method (High In the authors previous study (Takemoto, Yoshikawa,
Pressure Gas Safety Association, 2016) based on macro- 2019), assuming that the CFRP tank would be destroyed by
stress criteria, such as that employed for steel tanks, has tensile forces acting in the circumferential and/or axial
not been established thus far; therefore, the safety and directions, a biaxial tensile test and finite element analysis
reliability are assessed by performing a series of tests were conducted using a test piece with crossed fiber bundles;
(Takehana, and Ota, 1997) or (United Nations, 2013) by furthermore, the factors governing the tensile strength
spending considerable amounts of time and money to development mechanism in the CFRP helical layer were
impose loads similar to actual loads on real tanks. clarified. This current report presents a method for
Although there exists a strong desire to replace actual predicting the burst pressure of a high-pressure hydrogen
tank tests with computer simulations, a method for tank designed for fuel cell vehicles by elucidating the factors
appropriately predicting the tank damage and fracturing that control the strength mechanism and using mesoscopic
has not been presented thus far. zooming analysis. Hydraulic burst tests were performed on
To date, various studies have focused on the strength Type 3 small pressure tanks designed for convenient
evaluations of CFRP tanks. The related reports include fracture evaluations, and it was demonstrated that this
the evaluation based on laminate theory in the macroscale method can predict the test results with high accuracy.

*Advanced Materials and Processing Laboratory

NISSAN TECHNICAL REVIEW No.88 (2022) 119

 ♧ࡍࡿ㸬◚⿣ホ౯ࡀᐜ᫆࡞ࢱ࢖ࣉ
♧ࡍࡿ㸬◚⿣ホ౯ࡀᐜ᫆࡞ࢱ࢖ࣉ 3 ᑠᆺᅽຊᐜჾࡢỈᅽ◚⿣ヨ㦂ࢆ㏻ࡌ࡚㸪ᮏᡭἲࡀヨ㦂⤖ᯝࢆ㧗⢭ᗘ࡟ண
3 ᑠᆺᅽຊᐜჾࡢỈᅽ◚⿣ヨ㦂ࢆ㏻ࡌ࡚㸪ᮏᡭἲࡀヨ㦂⤖ᯝࢆ㧗⢭ᗘ࡟ண࡛ ࡛
〇సࡋࡓ A6061-T6 ࢔࣑ࣝྜ㔠〇㸦ᅗ 1㸧࡜ࡋࡓ㸬ᐜჾ⬗㒊㸦ࢩࣜࣥ
ࡁࡿࡇ࡜ࢆᐇドࡍࡿ㸬
ࡁࡿࡇ࡜ࢆᐇドࡍࡿ㸬
ረᛶຍᕤ࡟ࡼࡾ㙾㒊㸦ࢻ࣮࣒㒊㸧⫗ཌࡀቑ኱ࡋ㙾㒊ࡢᙉᗘࡀྥୖࡋ࡚
  Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis
ࡾ CFRP ᙉᗘࡢホ౯ࡀ⾜࠸᫆࠸࡜⪃࠼ࡽࢀࡿ㸬ࡇࡢࣛ࢖ࢼ࣮࡟㸪FW

ࢱ࢖ࣉࢱ࢖ࣉ ࣇ࣮ࣉ✚ᒙ㸦ᅗ 2(b)㸧ࢆ᪋ࡋࡓᚋ㸪⇕◳໬ࡉࡏࡓ㸬
 ᑠᆺᅽຊᐜჾࡢỈᅽ◚⿣ヨ㦂
 ᑠᆺᅽຊᐜჾࡢỈᅽ◚⿣ヨ㦂
referred
㒊ᮦࡢᮦᩱᐃᩘࢆ⾲ to as the first,1second, third, and fourth layers,
࡟♧ࡍ㸬ࡇࢀࡽࡣᮦᩱ࣓࣮࣮࢝ࡀᥦ♧ࡋ࡚࠸
࣭
࣭ ᅽຊᐜჾࡢస〇
ᅽຊᐜჾࡢస〇 from the innermost layer on the liner side. The thickness
2. Hydraulic burst test on Type 3 small ཌࡉ 0.5 mm ࡢᮾࣞ T700SC-24000-50C㸪࣐ࢺࣜࢡࢫᶞ⬡ࡣ୺๣ࡢ࢚࣏
ᮏ◊✲࡛ᥦ᱌ࡍࡿᙉᗘホ౯ᡭἲࡢ᳨ドࡢࡓࡵ㸪 pressure tanks
ᮏ◊✲࡛ᥦ᱌ࡍࡿᙉᗘホ౯ᡭἲࡢ᳨ドࡢࡓࡵ㸪 FWFW of eachCFRP
ᕤἲࢆ⏝࠸࡚
ᕤἲࢆ⏝࠸࡚ CFRP
layer was
ࢆ✚ᒙࡋࡓࢱ࢖ࣉ
ࢆ✚ᒙࡋࡓࢱ࢖ࣉ 3 ᑠᆺᅽຊᐜჾࢆస
0.5 mm. The 3helical layer exhibited
ᑠᆺᅽຊᐜჾࢆస
〇ࡋỈᅽ◚⿣ヨ㦂ࢆ⾜ࡗࡓ㸬ᐜჾෆࡢỈ⣲ࢆᑒධࡍࡿᙺ┠ࢆᢸ࠺ࣛ࢖ࢼ࣮ࡣ㸪ᅽᘏ࠾ࡼࡧࢫࣆࢽࣥࢢຍᕤ࡟ࡼࡾ ❧໬ᡂ
fiber HN2200㸧࡜ಁ㐍๣㸦ⰼ⋤࣮࢝࢜ࣛ࢖ࢨ࣮No.
orientation angles of 30 with respect 20㸧ࢆ㉁㔞ẚ 10
to the central
〇ࡋỈᅽ◚⿣ヨ㦂ࢆ⾜ࡗࡓ㸬ᐜჾෆࡢỈ⣲ࢆᑒධࡍࡿᙺ┠ࢆᢸ࠺ࣛ࢖ࢼ࣮ࡣ㸪ᅽᘏ࠾ࡼࡧࢫࣆࢽࣥࢢຍᕤ࡟ࡼࡾ
2.1 Preparation of pressure tank axis of the
ྵ᭷⋡ࢆ Vf =tank. Tank No. 1 had a helical layer alone,
55 %࡜ࡋࡓ㸬✚ᒙᚋ㸪◳໬⅔࡟࡚ᐜჾࢆᶓ᪉ྥ࡟㓄⨨
〇సࡋࡓ
〇సࡋࡓ A6061-T6
A6061-T6 ࢔࣑ࣝྜ㔠〇㸦ᅗ
࢔࣑ࣝྜ㔠〇㸦ᅗ 1㸧࡜ࡋࡓ㸬ᐜჾ⬗㒊㸦ࢩࣜࣥࢲ࣮㒊㸧ࡢ⫗ཌ
1㸧࡜ࡋࡓ㸬ᐜჾ⬗㒊㸦ࢩࣜࣥࢲ࣮㒊㸧ࡢ⫗ཌ d =d 2=mm
2 mm ࡟ᑐࡋ㸪ᡂᙧ᫬ࡢ
࡟ᑐࡋ㸪ᡂᙧ᫬ࡢ
To verify the strength evaluation method developed in Tank No. 2 had a hoop layer with a fiber orientation
ࡕࡉࡽ࡟ 110 Υ࡛ 3 ᫬㛫ಖᣢࡍࡿࡇ࡜࡟ࡼࡾ⇕◳໬ࢆ⾜ࡗࡓ㸬
ረᛶຍᕤ࡟ࡼࡾ㙾㒊㸦ࢻ࣮࣒㒊㸧⫗ཌࡀቑ኱ࡋ㙾㒊ࡢᙉᗘࡀྥୖࡋ࡚࠸ࡿࡓࡵ㸪◚ቯ㉳Ⅼࡣ⬗㒊࡟Ⓨ⏕ࡋ᫆ࡃ࡞
this study, Type 3 small pressure tanks with CFRP
ረᛶຍᕤ࡟ࡼࡾ㙾㒊㸦ࢻ࣮࣒㒊㸧⫗ཌࡀቑ኱ࡋ㙾㒊ࡢᙉᗘࡀྥୖࡋ࡚࠸ࡿࡓࡵ㸪◚ቯ㉳Ⅼࡣ⬗㒊࡟Ⓨ⏕ࡋ᫆ࡃ࡞ angle of 90 inside the helical layer, Tank No. 3 had a
ᮏ◊✲࡛⏝࠸ࡓ 4 ✀㢮ࡢᐜჾࡢ CFRP ✚ᒙᵓᡂࢆ⾲ 4 had an
2 ࡟♧ࡍ㸬CF
ࡾlaminates
ࡾ CFRP
CFRP were prepared using the FW method and
ᙉᗘࡢホ౯ࡀ⾜࠸᫆࠸࡜⪃࠼ࡽࢀࡿ㸬ࡇࡢࣛ࢖ࢼ࣮࡟㸪FW
ᙉᗘࡢホ౯ࡀ⾜࠸᫆࠸࡜⪃࠼ࡽࢀࡿ㸬ࡇࡢࣛ࢖ࢼ࣮࡟㸪FW hoop layer on the outside, and Tank 2(a)㸧࠾ࡼࡧ
ᕤἲ࡟ࡼࡗ࡚࣊ࣜ࢝ࣝ✚ᒙ㸦ᅗ
ᕤἲ࡟ࡼࡗ࡚࣊ࣜ࢝ࣝ✚ᒙ㸦ᅗ No.
2(a)㸧࠾ࡼࡧ
subjected to a hydraulic burst test. The liner, designed to 2 3 4 0.5 mm
additional hoop layer on the outside. In addition, for each ࡛࠶ࡿ㸬࣊
ᒙ㸪➨ ᒙ㸪➨ ᒙ㸪➨ ᒙ࡜ࡍࡿ㸬1 ᒙࡢཌࡉࡣ
ࣇ࣮ࣉ✚ᒙ㸦ᅗ
ࣇ࣮ࣉ✚ᒙ㸦ᅗ 2(b)㸧ࢆ᪋ࡋࡓᚋ㸪⇕◳໬ࡉࡏࡓ㸬
2(b)㸧ࢆ᪋ࡋࡓᚋ㸪⇕◳໬ࡉࡏࡓ㸬
encapsulate hydrogen in the tank, was composed of of theᗘ࡜-30
ゅ+30 tanks, hoop
ᗘ࡜ࡋࡓ㸬winding reinforcement was
࣊ࣜ࢝ࣝᒙࡢࡳࡢ applied
No.1 ᐜჾ㸪 to⧄⥔㓄ྥゅࡀ
㒊ᮦࡢᮦᩱᐃᩘࢆ⾲
㒊ᮦࡢᮦᩱᐃᩘࢆ⾲
A6061-T6 aluminum alloy
1 processed
࡟♧ࡍ㸬ࡇࢀࡽࡣᮦᩱ࣓࣮࣮࢝ࡀᥦ♧ࡋ࡚࠸ࡿ௦⾲್࡛࠶ࡿ㸬Ⅳ⣲⧄⥔᮰ࡣᖜ
1 ࡟♧ࡍ㸬ࡇࢀࡽࡣᮦᩱ࣓࣮࣮࢝ࡀᥦ♧ࡋ࡚࠸ࡿ௦⾲್࡛࠶ࡿ㸬Ⅳ⣲⧄⥔᮰ࡣᖜ
through rolling and the bosses at both ends to prevent water leakage 10 10 mm㸪
mm㸪
from the
࠶ࡿ No.2 ᐜჾ㸪እഃ࡟࠶ࡿ No.3 ᐜჾ㸪ࡉࡽ࡟ࣇ࣮ࣉᒙࢆእഃ࡟㔜ࡡ
ཌࡉཌࡉ 0.50.5
spinningmmmm
(Fig.
ࡢᮾࣞ ࡢᮾࣞ T700SC-24000-50C㸪࣐ࢺࣜࢡࢫᶞ⬡ࡣ୺๣ࡢ࢚࣏࢟ࢩᶞ⬡㸦୕⳻໬Ꮫ
1). T700SC-24000-50C㸪࣐ࢺࣜࢡࢫᶞ⬡ࡣ୺๣ࡢ࢚࣏࢟ࢩᶞ⬡㸦୕⳻໬Ꮫ
The wall thickness of the tank body threads due to the increased water jER828㸧࡟◳໬๣㸦᪥
pressure. Photographs
jER828㸧࡟◳໬๣㸦᪥
(cylindrical part) was d = 2 mm, and the thickness of the ჾ࡟࠾࠸࡚ࡶ㸪Ỉᅽୖ᪼࡟ࡼࡿࢿࢪ㒊࠿ࡽࡢỈ₃ࢀࢆ㜵ࡄࡓࡵ㸪୧➃
of the tanks are provided in Fig. 3
❧໬ᡂ HN2200㸧࡜ಁ㐍๣㸦ⰼ⋤࣮࢝࢜ࣛ࢖ࢨ࣮No.
❧໬ᡂ HN2200㸧࡜ಁ㐍๣㸦ⰼ⋤࣮࢝࢜ࣛ࢖ࢨ࣮No. 20㸧ࢆ㉁㔞ẚ 100:10:1 ࡛㓄ྜࡋࡓࡶࡢ࡜ࡋ㸪Ⅳ⣲⧄⥔ࡢయ✚20㸧ࢆ㉁㔞ẚ 100:10:1 ࡛㓄ྜࡋࡓࡶࡢ࡜ࡋ㸪Ⅳ⣲⧄⥔ࡢయ✚
dome part was increased through the molding process to ࡽᐜჾࡢ෗┿ࢆᅗ 3 ࡟♧ࡍ㸬
ྵ᭷⋡ࢆ
ྵ᭷⋡ࢆ V V
= = 55
f 55 %࡜ࡋࡓ㸬✚ᒙᚋ㸪◳໬⅔࡟࡚ᐜჾࢆᶓ᪉ྥ࡟㓄⨨ࡋᅇ㌿ࡉࡏ࡞ࡀࡽ㸪80
%࡜ࡋࡓ㸬✚ᒙᚋ㸪◳໬⅔࡟࡚ᐜჾࢆᶓ᪉ྥ࡟㓄⨨ࡋᅇ㌿ࡉࡏ࡞ࡀࡽ㸪80 Υ࡛ Υ࡛ 2 ᫬㛫ಖᣢࡢࡢ
2 ᫬㛫ಖᣢࡢࡢ
enhance its strength, such that the burst could easily occur
f
500 mm
ࡕࡉࡽ࡟
in
ࡕࡉࡽ࡟ 110110
the cylindrical
Υ࡛ Υ࡛ 3 this
part; ᫬㛫ಖᣢࡍࡿࡇ࡜࡟ࡼࡾ⇕◳໬ࢆ⾜ࡗࡓ㸬
enabled a convenient evaluation
3 ᫬㛫ಖᣢࡍࡿࡇ࡜࡟ࡼࡾ⇕◳໬ࢆ⾜ࡗࡓ㸬 d = 2 mm
of the CFRP strength.
ᮏ◊✲࡛⏝࠸ࡓ
ᮏ◊✲࡛⏝࠸ࡓ 4A✀㢮ࡢᐜჾࡢ
helical layerCFRP
4 ✀㢮ࡢᐜჾࡢ (Fig.
CFRP2✚ᒙᵓᡂࢆ⾲
(a)) and a hoop2 ࡟♧ࡍ㸬CFRP
✚ᒙᵓᡂࢆ⾲ 2 ࡟♧ࡍ㸬CFRP ࡢ✚ᒙࡣ㸪ࣛ࢖ࢼ࣮ഃࡢ᭱ෆᒙ࠿ࡽ➨
ࡢ✚ᒙࡣ㸪ࣛ࢖ࢼ࣮ഃࡢ᭱ෆᒙ࠿ࡽ➨ 1 1
Volume: 7.5 L
layer (Fig. 2 (b)) were applied to this liner using the FW

ȭ160 mm
ᒙ㸪➨
ᒙ㸪➨ 2 ᒙ㸪➨
2 ᒙ㸪➨ 3 ᒙ㸪➨
3 ᒙ㸪➨ 4 ᒙ࡜ࡍࡿ㸬1
4 ᒙ࡜ࡍࡿ㸬1 ᒙࡢཌࡉࡣ
ᒙࡢཌࡉࡣ 0.50.5
mmmm ࡛࠶ࡿ㸬࣊ࣜ࢝ࣝᒙࡣᐜჾࡢ୰ᚰ㍈࡟ᑐࡋ࡚⧄⥔㓄ྥ
࡛࠶ࡿ㸬࣊ࣜ࢝ࣝᒙࡣᐜჾࡢ୰ᚰ㍈࡟ᑐࡋ࡚⧄⥔㓄ྥ
method, followed by the application of the thermoset. Material: A6061-T6
ゅ+30
ゅ+30 ᗘ࡜-30 ᗘ࡜ࡋࡓ㸬
ᗘ࡜ࡋࡓ㸬 ࣊ࣜ࢝ࣝᒙࡢࡳࡢ
࣊ࣜ࢝ࣝᒙࡢࡳࡢ No.1
No.1
The material constants of the members, as presented
ᗘ࡜-30 ᐜჾ㸪ᐜჾ㸪 ⧄⥔㓄ྥゅࡀ
⧄⥔㓄ྥゅࡀ 90 90 ᗘ࡛࠶ࡿࣇ࣮ࣉᒙࡀ࣊ࣜ࢝ࣝᒙࡢෆഃ࡟
ᗘ࡛࠶ࡿࣇ࣮ࣉᒙࡀ࣊ࣜ࢝ࣝᒙࡢෆഃ࡟
by
࠶ࡿ࠶ࡿ theNo.2
No.2 manufacturers,
ᐜჾ㸪እഃ࡟࠶ࡿ
ᐜჾ㸪እഃ࡟࠶ࡿ No.3
are listed
No.3 Table 1. The carbon
in ᐜჾ㸪ࡉࡽ࡟ࣇ࣮ࣉᒙࢆእഃ࡟㔜ࡡࡓ
ᐜჾ㸪ࡉࡽ࡟ࣇ࣮ࣉᒙࢆእഃ࡟㔜ࡡࡓ No.4
No.4
Dome part
ᐜჾࢆస〇ࡋࡓ㸬ࡲࡓ㸪࠸ࡎࢀࡢᐜ
Cylinder part
ᐜჾࢆస〇ࡋࡓ㸬ࡲࡓ㸪࠸ࡎࢀࡢᐜ
fiber bundle was T700SC-24000-50C from Toray, with a
ჾ࡟࠾࠸࡚ࡶ㸪Ỉᅽୖ᪼࡟ࡼࡿࢿࢪ㒊࠿ࡽࡢỈ₃ࢀࢆ㜵ࡄࡓࡵ㸪୧➃ࡢ࣎ࢫ㒊࡟ࣇ࣮ࣉᕳࡁ⿵ᙉࢆ᪋ࡋࡓ㸬ࡇࢀ
ჾ࡟࠾࠸࡚ࡶ㸪Ỉᅽୖ᪼࡟ࡼࡿࢿࢪ㒊࠿ࡽࡢỈ₃ࢀࢆ㜵ࡄࡓࡵ㸪୧➃ࡢ࣎ࢫ㒊࡟ࣇ࣮ࣉᕳࡁ⿵ᙉࢆ᪋ࡋࡓ㸬ࡇࢀ Fig. 1 Aluminum liner. (a) Heli
width of 10 mm and a thickness of 0.5 mm. The matrix
ࡽᐜჾࡢ෗┿ࢆᅗ
ࡽᐜჾࡢ෗┿ࢆᅗ 3 3 ࡟♧ࡍ㸬
࡟♧ࡍ㸬
resin was prepared by blending the main agent epoxy Fig. 1 Aluminum liner. Fig. 2 C
resin (jER828 from 500 Mitsubishi
mm Chemical) with a curing
500 mm
agent (HN2200 from Hitachi Kasei) 2=mm
d =dand 2 mm
an accelerator Table 1 Material properties.
(Kaorizer No. 20 from Kao) at a mass of 100:10:1
ratio 7.5
Volume:
Volume: 7.5 L L Carbon fiber Matrix resin
ȭ160 mm

and adjusting the volume content of the carbon fibers to 

ȭ160 mm

Material: A6061-T6 Young’s modulus [GPa] 230 3.75

Vf = 55%. After laminating, thermosettingMaterial:was
A6061-T6
performed
by arranging the tanks in the horizontal direction in a Poisson ratio 0.26 0.41
curing furnace 80 C for 2 h and
Dome part and
Dome part maintaining
Cylinder part them at
Cylinder part Elongation [%] 2.1 2.6
then at 110 C for 3 h while rotating.
(a) Helical layer (±30 degree). (b) Hoop layer.
The CFRP layer structures of the four tanks used in (a) Helical layer (±30 degree). (b) Hoop layer.
[DOI: 10.1299/transjsme.19-00338] © 20
Fig. 1 Aluminum
this study are described 2.liner.
in Tableliner.The CFRP layers are Fig. 2 2CFRP layer manufactured by filament winding.
Fig. 1 Aluminum Fig. 2Fig.
CFRP layerlayer
CFRP manufactured by filament
manufactured winding.
by filament winding.

Table 1 Material properties.

Table
Table11Material
Materialproperties.
properties.
Carbon fiber Matrix
.LNDL*LM\XWVXDQG7RN\R7UDQVDFWLRQVRIWKH-60( resin Aluminum liner Steel plug

 Carbon fiber Matrix resin LQ-DSDQHVH 9RO
Aluminum liner 1RSteel
[plug
.LNDL*LM\XWVXDQG7RN\R7UDQVDFWLRQVRIWKH-60(
Young’s modulus [GPa] 230 3.75LQ-DSDQHVH 9RO
1R [ 210
70.0 
Young’s modulus [GPa] 230 3.75 70.0 210
 Poisson ratio 0.26 0.41 0.33 0.30
Poisson ratio 0.26 0.41 0.33 0.30
 Elongation [%] 2.1 2.6 12 -
Elongation [%] 2.1 2.6 12
7DEOH9DULDWLRQRI&)53OD\HUVRIWDQNVIRUHYDOXDWLRQ -

7DQN1R 7DEOH9DULDWLRQRI&)53OD\HUVRIWDQNVIRUHYDOXDWLRQ
[DOI: 10.1299/transjsme.19-00338] &)53VWOD\HU &)53QGOD\HU
Table 2 Variation of CFRP © 201x
layers of tanks The Japan Society of Mechanical
&)53UGOD\HU
for evaluation. Engineers
&)53WKOD\HU
[DOI: 10.1299/transjsme.19-00338] © 201x The Japan Society of Mechanical Engineers

7DQN1R +HOLFDO“GHJUHH
&)53VWOD\HU &)53QGOD\HU 
&)53UGOD\HU 
&)53WKOD\HU
  +RRS
+HOLFDO“GHJUHH 
+HOLFDO“GHJUHH  
  +RRS +HOLFDO“GHJUHH+HOLFDO“GHJUHH+RRS  
  +HOLFDO“GHJUHH
+HOLFDO“GHJUHH +RRS
+RRS +RRS

 +HOLFDO“GHJUHH  +RRS +RRS
 
 
 
 
 
 
       D 7DQN1R           E 7DQN1R               F 7DQN1R              G 7DQN1R   

Fig. )LJ7\SHWDQNVIRUEXUVWWHVW
3 Type 3 tanks for burst test.
       D 7DQN1R           E 7DQN1R               F 7DQN1R              G 7DQN1R   

120
)LJ7\SHWDQNVIRUEXUVWWHVW No.88 (2022) NISSAN TECHNICAL REVIEW
࣭ Ỉᅽ◚⿣ヨ㦂

స〇ࡋࡓᐜჾࡢ◚⿣ヨ㦂ࢆ⾜ࡗࡓ㸬㟼ⓗ㈇Ⲵ࡜࡞ࡿࡼ࠺࡟タᐃᅽຊࢆ 03D ้ࡳ࡛ࢫࢸࢵࣉ≧࡟  ࢫࢸࢵ
࣭ Ỉᅽ◚⿣ヨ㦂
  +RRS +HOLFDO“GHJUHH 
 +HOLFDO“GHJUHH  
 +HOLFDO“GHJUHH +RRS 
 +RRS +HOLFDO“GHJUHH 
Technical Awards 2020  +HOLFDO“GHJUHH CFRP tank+RRS +RRS
 JSME Medal for Outstanding Paper - Strength evaluation of +RRS
+HOLFDO“GHJUHH based on meso-scopic
 zooming analysis

 +HOLFDO“GHJUHH  +RRS +RRS

 2.2 Hydraulic burst test The measured burst pressure was 12.0, 26.3, 24.8, and


 The burst test was performed on the prepared tanks as 35.3 MPa for Tanks No. 1, 2, 3, and 4, respectively. The
follows.
 The test pressure was applied as a static load in states of the broken tanks after the hydraulic burst test

steps of 0.75 MPa for 4 s or more per step, and the water are shown in Fig. 6. As indicated by Fig. 6 (a), both the
 
pressure was increased such that the preset pressure liner and helical layers of Tank No. 1 were fractured at

reached  0.05 MPa. The fiber-direction strain of the carbon the center of the body. For Tank No. 2, shown in Fig. 6
fiber  bundle on the tank surface was  also measured. The (b), the liner was broken at the center of the tank,
positions        D 7DQN1R           E 7DQN1R          whereas
at which the strain gauges were attached are the damage
    F 7DQN1R   on
  the
   outermost helical layer
    G 7DQN1R    was
depicted in Fig. 4. On one surface, seven gauges were limited. When the CFRP layer was disassembled, the
       D 7DQN1R           E 7DQN1R        )LJ7\SHWDQNVIRUEXUVWWHVW
       F 7DQN1R              G 7DQN1R   
attached at equal intervals on the geodesic at an inclination hoop layer, which was the innermost layer at the center
of 30  with respect to the central axis, which )LJ7\SHWDQNVIRUEXUVWWHVW
is the of the cylinder, was fractured along the axial direction, as
orientation
 ࣭ angle of the helical layer; on the other surface,
Ỉᅽ◚⿣ヨ㦂 presented in Fig. 6 (c). In Tanks No. 3 and No. 4, the hoop
seven࣭ gaugesỈᅽ◚⿣ヨ㦂were also attached at equal intervals on the layer, which was the outermost layer, was broken along
స〇ࡋࡓᐜჾࡢ◚⿣ヨ㦂ࢆ⾜ࡗࡓ㸬㟼ⓗ㈇Ⲵ࡜࡞ࡿࡼ࠺࡟タᐃᅽຊࢆ 03D ้ࡳ࡛ࢫࢸࢵࣉ≧࡟  ࢫࢸࢵ
meridian. The strain gauges
స〇ࡋࡓᐜჾࡢ◚⿣ヨ㦂ࢆ⾜ࡗࡓ㸬㟼ⓗ㈇Ⲵ࡜࡞ࡿࡼ࠺࡟タᐃᅽຊࢆ (KFG-5-120-C1-11 L5M3R, with the liners
03D at the centers of the cylinders.
้ࡳ࡛ࢫࢸࢵࣉ≧࡟  ࢫࢸࢵIn Tank
Kyowaࣉ࠶ࡓࡾ Electronic  ⛊௨ୖࢆ࠿ࡅ㸪タᐃᅽຊs03D
Instruments) were used for one-way ࡜࡞ࡿࡼ࠺࡟Ỉᅽࢆୖ᪼ࡉࡏࡓ㸬ᐜჾ⾲㠃ࡢⅣ⣲⧄⥔᮰ࡢ⧄⥔᪉
No. 4, the aluminum liner was considerably deformed as
ࣉ࠶ࡓࡾ  ⛊௨ୖࢆ࠿ࡅ㸪タᐃᅽຊs03D ࡜࡞ࡿࡼ࠺࡟Ỉᅽࢆୖ᪼ࡉࡏࡓ㸬ᐜჾ⾲㠃ࡢⅣ⣲⧄⥔᮰ࡢ⧄⥔᪉
measurements
ྥࡦࡎࡳࡶྜࢃࡏ࡚ィ and attached ࡋࡓ㸬ࡦࡎࡳࢤ࣮ࢪࡢ㈞௜఩⨨ࢆᅗ
in the direction of the the burst  ࡟♧ࡍ㸬୍᪉ࡢ㠃࡟࣊ࣜ࢝ࣝᒙࡢ㓄ྥゅ࡛࠶ࡿ୰
pressure increased.
ྥࡦࡎࡳࡶྜࢃࡏ࡚ィ
outermost fiber bundle. ࡋࡓ㸬ࡦࡎࡳࢤ࣮ࢪࡢ㈞௜఩⨨ࢆᅗ
For Tanks No.  ࡟♧ࡍ㸬୍᪉ࡢ㠃࡟࣊ࣜ࢝ࣝᒙࡢ㓄ྥゅ࡛࠶ࡿ୰
1 and No. 2, the  ࠿ᡤ㸪௚᪉ࡢ㠃࡟ࡣᏊ༗⥺ୖ࡟➼㛫㝸࡛
ᚰ㍈࡟ᑐࡋ࡚  ᗘࡢഴࡁࡢ ᆅ⥺ୖ࡟➼㛫㝸࡛  ࠿ᡤ㈞௜ࡋࡓ㸬ࡦ
gauges
ᚰ㍈࡟ᑐࡋ࡚ were attached
 ᗘࡢഴࡁࡢat an inclination
ᆅ⥺ୖ࡟➼㛫㝸࡛ of 30 with respect
࠿ᡤ㸪௚᪉ࡢ㠃࡟ࡣᏊ༗⥺ୖ࡟➼㛫㝸࡛ 3. Prediction of burst  ࠿ᡤ㈞௜ࡋࡓ㸬ࡦ
pressure using
ࡎࡳࢤ࣮ࢪࡣ୍᪉ྥィ ⏝㸦ඹ࿴㟁ᴗ .)*&/05㸧ࢆ⏝࠸㸪᭱እᒙࡢ⧄⥔᮰᪉ྥ࡟㈞௜ࡋࡓ㸬ࡍ࡞
to the central axis along the
ࡎࡳࢤ࣮ࢪࡣ୍᪉ྥィ ⏝㸦ඹ࿴㟁ᴗfiber bundle.)*&/05㸧ࢆ⏝࠸㸪᭱እᒙࡢ⧄⥔᮰᪉ྥ࡟㈞௜ࡋࡓ㸬ࡍ࡞
of the outermost axisymmetric continuum model
helical ࢃࡕ㸪1R
layer,ᐜჾ࠾ࡼࡧ ᐜჾ࠾ࡼࡧ
whereas for1R
Tanks 1R
No. 3ᐜჾࡣ᭱እᒙࡢ࣊ࣜ࢝ࣝᒙࡢ⧄⥔᮰࡟ἢࡗ࡚୰ᚰ㍈࡟ᑐࡋ࡚ഴࡁ
and No. 4, they were  ᗘ࡛㈞ࡾ௜ࡅ㸪
ࢃࡕ㸪1R ᐜჾࡣ᭱እᒙࡢ࣊ࣜ࢝ࣝᒙࡢ⧄⥔᮰࡟ἢࡗ࡚୰ᚰ㍈࡟ᑐࡋ࡚ഴࡁ  ᗘ࡛㈞ࡾ௜ࡅ㸪
attached 1R along
ᐜჾ࠾ࡼࡧ the outermost hoop layer in the
1R ᐜჾࡣ᭱እᒙࡢࣇ࣮ࣉᒙࡢ⧄⥔᮰࡟ἢࡗ࡚෇࿘᪉ྥࡢྥࡁ࡟㈞ࡾ௜ࡅࡓ㸬୍౛࡜ࡋ࡚㸪ࡦ 3.1 Axisymmetric continuum model
1R ᐜჾ࠾ࡼࡧ 1R ᐜჾࡣ᭱እᒙࡢࣇ࣮ࣉᒙࡢ⧄⥔᮰࡟ἢࡗ࡚෇࿘᪉ྥࡢྥࡁ࡟㈞ࡾ௜ࡅࡓ㸬୍౛࡜ࡋ࡚㸪ࡦ
circumferential
ࡎࡳࢤ࣮ࢪࢆ㈞ࡾࡘࡅࡓ≧ែࡢ 1R ᐜჾࡢ෗┿ࢆᅗ  ࡟♧ࡍ㸬 pressures of the tanks subjected to the burst
direction. A photograph of the strain gauges The burst
ࡎࡳࢤ࣮ࢪࢆ㈞ࡾࡘࡅࡓ≧ែࡢ
attached to Tank No. 4 is provided1R ᐜჾࡢ෗┿ࢆᅗ
in Fig. 5 as an example  ࡟♧ࡍ㸬 test were predicted via macroscale simulations through


 
 
 
 
 
 
 
 
 
 
Fig. 4 Locations of strain gages.                    )LJ6WUDLQJDJHVRQWDQN1RVHWXS
)LJ/RFDWLRQVRIVWUDLQJDJHV Fig. 5 Strain gages on tank No. 4 set up.
)LJ/RFDWLRQVRIVWUDLQJDJHV                    )LJ6WUDLQJDJHVRQWDQN1RVHWXS
 
◚⿣ᅽຊࡣ㸪1R ᐜჾ㸸03D㸪1R ᐜჾ㸸03D㸪1R ᐜჾ㸸03D㸪1R ᐜჾ㸸03D ࡛࠶ࡗ
◚⿣ᅽຊࡣ㸪1R ᐜჾ㸸03D㸪1R ᐜჾ㸸03D㸪1R ᐜჾ㸸03D㸪1R ᐜჾ㸸03D ࡛࠶ࡗ
ࡓ㸬Ỉᅽ◚⿣ヨ㦂ᐇ᪋ᚋࡢᐜჾ◚ቯ≧ែࢆᅗ  ࡟♧ࡍ㸬1R ᐜჾࡣ㸪ᅗ  D ࡟♧ࡍࡼ࠺࡟㸪ࣛ࢖ࢼ࣮㸪࣊ࣜ࢝ࣝ
ࡓ㸬Ỉᅽ◚⿣ヨ㦂ᐇ᪋ᚋࡢᐜჾ◚ቯ≧ែࢆᅗ  ࡟♧ࡍ㸬1R ᐜჾࡣ㸪ᅗ  D ࡟♧ࡍࡼ࠺࡟㸪ࣛ࢖ࢼ࣮㸪࣊ࣜ࢝ࣝ
ᒙ࡜ࡶ࡟ᐜჾ୰ኸ㒊࡛◚᩿ࡋࡓ㸬1R ᐜჾࡣ㸪ᅗ  E ࡟♧ࡍࡼ࠺࡟㸪ᐜჾ୰ኸ㒊࡛ࣛ࢖ࢼ࣮ࡀ◚᩿ࡋࡓࡶࡢࡢ᭱
ᒙ࡜ࡶ࡟ᐜჾ୰ኸ㒊࡛◚᩿ࡋࡓ㸬1R ᐜჾࡣ㸪ᅗ  E ࡟♧ࡍࡼ࠺࡟㸪ᐜჾ୰ኸ㒊࡛ࣛ࢖ࢼ࣮ࡀ◚᩿ࡋࡓࡶࡢࡢ᭱
እᒙࡢ࣊ࣜ࢝ࣝᒙࡢᦆയࡣࢃࡎ࠿࡛࠶ࡗࡓ㸬ࡑࡇ࡛ &)53 ᒙࢆศゎࡋࡓ࡜ࡇࢁ㸪ᅗ  F ࡟♧ࡍࡼ࠺࡟㸪ᐜჾ୰ኸ
እᒙࡢ࣊ࣜ࢝ࣝᒙࡢᦆയࡣࢃࡎ࠿࡛࠶ࡗࡓ㸬ࡑࡇ࡛ &)53 ᒙࢆศゎࡋࡓ࡜ࡇࢁ㸪ᅗ  F ࡟♧ࡍࡼ࠺࡟㸪ᐜჾ୰ኸ
㒊᭱ෆᒙࡢࣇ࣮ࣉᒙࡀ㍈᪉ྥ࡟ἢࡗ࡚◚᩿ࡋ࡚࠸ࡓ㸬1R ࠾ࡼࡧ 1R ᐜჾࡣ㸪ᐜჾ୰ኸ㒊࡛ࣛ࢖ࢼ࣮࡜࡜ࡶ࡟
㒊᭱ෆᒙࡢࣇ࣮ࣉᒙࡀ㍈᪉ྥ࡟ἢࡗ࡚◚᩿ࡋ࡚࠸ࡓ㸬1R ࠾ࡼࡧ 1R ᐜჾࡣ㸪ᐜჾ୰ኸ㒊࡛ࣛ࢖ࢼ࣮࡜࡜ࡶ࡟
᭱እᒙࡢࣇ࣮ࣉᒙࡀ◚᩿ࡋࡓ㸬1R ᐜჾࡣ◚⿣ᅽࡢୖ᪼࡟క࠸࢔࣑ࣝࣛ࢖ࢼ࣮ࡀ኱ࡁࡃረᛶኚᙧࡋࡓ㸬
᭱እᒙࡢࣇ࣮ࣉᒙࡀ◚᩿ࡋࡓ㸬1R ᐜჾࡣ◚⿣ᅽࡢୖ᪼࡟క࠸࢔࣑ࣝࣛ࢖ࢼ࣮ࡀ኱ࡁࡃረᛶኚᙧࡋࡓ㸬
>'2,WUDQVMVPH@                                      ‹[7KH-DSDQ6RFLHW\RI0HFKDQLFDO(QJLQHHUV

>'2,WUDQVMVPH@                                      ‹[7KH-DSDQ6RFLHW\RI0HFKDQLFDO(QJLQHHUV
 


Fig. 6 The tanks listed Table 2 after burst tests are shown. Both CFRP layers and aluminum liner are broken at center of
cylindrical part at all the tanks. The CFRP helical layer is broken at tank No.1. The CFRP hoop layer of inner side is
broken at tank No.2. The CFRP hoop layer of the most outside is broken at tank No.3 and No.4.

NISSAN TECHNICAL REVIEW No.88 (2022) 121

 Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis

the axisymmetric continuum modeling of the tanks (Kim respect to the central surface, and the displacement Uz in
et al., 2017). The CFRP layer was composed of an the z-direction was constrained along the central surface.
orthogonal anisotropic linear elastic material, and its The internal pressure of the tank was increased in
material constants were calculated based on those of the increments of 0.750 MPa. In Figs. 8 (a)–(d), the right side
carbon fiber and resin shown in Table 1, according to the presents a magnified view of the layers of the cylindrical
law of mixtures (Maki et al., 1970). The helical layer was part, from the inner aluminum liner to the CFRP first
modeled based on the laminate theory (Vasiliev and layer, to the second layer (a), to the third layer (b) and (c),
Morozov, 2001), assuming that the layers with fiber up to the fourth layer (d). The aluminum liner, hoop
orientation angles of +30.0 and -30.0 were laminated layer, and helical layer were all modeled with triangular
with the same thickness. The local increase in strain due primary elements. The CFRP helical layer increases in
to the crossing of the fiber bundles caused by the FW thickness from the equator to the boss in the dome
method could not be evaluated. The stress–strain section. This change was set according to a theoretical
Kikai, Gijyutsu and Tokyo, Transactions of the JSME (in Japanese), Vol.00, No.00 (201x)
constitutive law of the aluminum liner is based on the formula (Kim et al., 2017). Furthermore, because the
isotropic elastoplastic solid, as shown in Fig. 7. In fiber orientation angle changes continuously, different
addition, because the steel plug did not ࡟♧ࡋࡓ◚⿣㛤ጞⅬ࡟୍␒㏆࠸఩⨨ࡢ್ࢆ◚⿣ᅽࡲ࡛ࣉࣟࢵࢺࡋࡓ㸬ᐇ⥺ࡣ㍈ᑐ⛠㐃⥆యࣔࢹࣝࢆ⏝࠸ࡓゎ
break even after material constants were set for the orthogonal anisotropy
the burst test and possessed sufficient
Kikai, Gijyutsu strength andof the JSME
ᯝ࡛࠶ࡿ㸬㯮⥺ࡣ࢔࣑ࣝࣛ࢖ࢼ࣮ࡢ෇࿘᪉ྥࡦࡎࡳ㸪㟷⥺ࡣ
and Tokyo, Transactions of (in
each element.
Japanese), Vol.00,Tank (201x)1CFRP
No.00 No. 17,842 elements and
had ࣊ࣜ࢝ࣝᒙࡢ⧄⥔᪉ྥࡦࡎࡳ㸪㉥⥺ࡣ
rigidity, it was selected as an isotropic elastic body, whose
ࣇ࣮ࣉᒙࡢ⧄⥔᪉ྥࡦࡎࡳ࡛࠶ࡿ㸬 9,551 nodes, Tanks No. 2 and No. 3 both had 19,171
ᐜჾࡢ◚⿣఩⨨࡛࠶ࡿᐜჾ୰ኸ㒊௜㏆ࡢྛᒙࡢ᭱ෆ㠃࡛ࡦࡎࡳࡣ᭱኱࡜
material properties 1.
are listed in Tableᅗ୰ࡢᐇ⥺ࡣࡑࡢ್ࢆ♧ࡋ࡚࠸ࡿ㸬ࡲࡓ⾲
࡟♧ࡋࡓ◚⿣㛤ጞⅬ࡟୍␒㏆࠸఩⨨ࡢ್ࢆ◚⿣ᅽࡲ࡛ࣉࣟࢵࢺࡋࡓ㸬ᐇ⥺ࡣ㍈ᑐ⛠㐃⥆యࣔࢹࣝࢆ⏝࠸ࡓゎᯒ⤖ elements and1 10,216࠿ࡽ CFRP nodes; and Tank No. 4 had 20,444 2.10 %࡟㐩ࡍ
ࡢ⧄⥔᪉ྥࡦࡎࡳࡀ⧄⥔ࡢ◚᩿ࡦࡎࡳ
The axisymmetric finite element model used for the
ᯝ࡛࠶ࡿ㸬㯮⥺ࡣ࢔࣑ࣝࣛ࢖ࢼ࣮ࡢ෇࿘᪉ྥࡦࡎࡳ㸪㟷⥺ࡣ elements and 10,853 nodes.
CFRP ࣊ࣜ࢝ࣝᒙࡢ⧄⥔᪉ྥࡦࡎࡳ㸪㉥⥺ࡣ CFRP
analysis and its boundary conditions are depicted in Fig.
ࣇ࣮ࣉᒙࡢ⧄⥔᪉ྥࡦࡎࡳ࡛࠶ࡿ㸬 ᐜჾࡢ◚⿣఩⨨࡛࠶ࡿᐜჾ୰ኸ㒊௜㏆ࡢྛᒙࡢ᭱ෆ㠃࡛ࡦࡎࡳࡣ᭱኱࡜࡞ࡾ㸪
8; those relatedᅗ୰ࡢᐇ⥺ࡣࡑࡢ್ࢆ♧ࡋ࡚࠸ࡿ㸬ࡲࡓ⾲
to Tanks No. 1, 2, 3, and 4 are shown 1 ࠿ࡽ CFRP
in ࡢ⧄⥔᪉ྥࡦࡎࡳࡀ⧄⥔ࡢ◚᩿ࡦࡎࡳ 2.10 %࡟㐩ࡍࡿ࡜
Figs. 8 (a), (b), (c), and (d), respectively. Half of each tank
was modeled on the r-z plane, while considering symmetry
around the z-axis. At the boss part at the end of the liner,
the plug and CFRP hoop layer that reinforced the boss
part were completely connected, and the displacement Ur
in the radial direction (r-direction) of the tank was
constrained by the central axis of the plug. Moreover,
Fig. 7 Constitutive
Fig. Constitutive lawlaw
of aluminum liner.liner.
of aluminum
each tank was set to be vertically symmetrical with
Fig. 7 Constitutive law of aluminum liner.

Unit: mm Unit: mm Unit: mm Unit: mm

(a) Tank No.1. (b) Tank No.2.

(a) Tank No.1. (b) Tank No.2.

Unit: mm Unit: mm

(c) Tank No.3. (d) Tank No.4.

Fig. 8 Axisymmetric analysis models of tank. Unit: mm Unit: m
Fig. 8 Axisymmetric analysis models of tank.
(c) Tank No.3. (d) Tank No.4.
[DOI: 10.1299/transjsme.19-00338] © 201x The Japan Society of Mechanical Engineers
122 Fig. 8 Axisymmetric
No.88analysis
(2022) models
NISSANof tank.
TECHNICAL REVIEW

[DOI: 10.1299/transjsme.19-00338] © 201x The Japan Society of Mechanical Engineers

 Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis

3.2 Analysis result CFRP first layer (helical layer) at 17.2 MPa; Tank No. 2,
The analysis results for Tanks No. 1, 2, 3, and 4 are from the CFRP first layer (hoop layer) at 29.6 MPa; Tank
presented in Figs. 9 (a), (b), (c), and (d), respectively, No. 3, from the CFRP third layer (hoop layer) at 29.8
along with the experimental results from Section 2. The MPa; and Tank No. 4, from the CFRP third layer (hoop
horizontal axis represents the pressure inside the tank, layer) at 47.4 MPa.
and the vertical axis represents the strain. The green
squares indicate the strain values measured during the 3.3 Comparison of analysis and experimental results
tank burst test. Among the measurement positions For all the tanks, the measured values of the fiber-
shown in Fig. 4, the value at the position closest to the direction strain in the outermost CFRP layer, indicated
burst start point shown in Fig. 6 was plotted up to the by the green squares in Fig. 9, agree well with the
burst pressure. The solid line indicates the analysis analysis results in the micro-deformation region, where
results obtained using the axisymmetric continuum the strain was 0.5% or less. The strain values measured
model. The black, blue, and red lines represent the by the strain gauges do not accurately indicate the
Kikai, Gijyutsu and Tokyo, Transactions of the JSME (in Japanese), Vol.00, No.00 (201x)
circumferential strain of the aluminum liner, the fiber- increments in the local strain in the fiber bundle and the
direction strain of the CFRP helical layer, and the resin, although these gauges were attached to the fiber
fiber-direction strain of the CFRP hoop layer, respectively. bundle. To summarize, this is because the gauges simply
The strain is maximized at the innermost surface of each represent the average strains in the gauge-mounted
layer near the center of the tank, which is the burst areas; this is explained in detail based on the meso-scale
position; the solid line in the figure indicates this value. model analysis in subsequent sections. Therefore, if the
In addition, as shown in Table 1, when the fiber-direction average measured strain is low, the strain measurement
strain of the CFRP reached 2.10% of the fiber fracture results agree relatively well with the analysis results
strain, the relevant CFRP layer broke and led to tank obtained0 using the axisymmetric continuum model.
fracture. A solid line is drawn up to the burst pressure However, if the strain becomes high and the nonlinear
calculated. behavior of the resin becomes locally remarkable, the
The maximum strain of the aluminum liner did not analysis results would deviate from the strain
(a) Tank No.1.

exceed the fracture strain (12%) in any of the tanks, as measurement results because the influence of the
shown in Table 1, and it is considered that the aluminum nonlinear behavior was not considered.
was not the starting point of the fracture. Based on this For Tank No. 1, the analysis results agree with the
result, the respective tanks were predicted to burst at the measurement results until the plastic deformation of the
central part of the cylindrical part, sufficiently far from aluminum liner commences. However, when the plastic
the dome part, starting from the following layers at the deformation progresses beyond an internal pressure of 8
0
corresponding internal pressures: Tank No. 1, from the MPa, the nonlinear behavior of the resin in the CFRP
Kikai, Gijyutsu and Tokyo, Transactions of the JSME (in Japanese), Vol.00, No.00 (201x)

(b) Tank No.2.

0 0

(a) Tank No.1. (c) Tank No.3.

0
(b) Tank No.2.

(d) Tank No.4.

Fig. 9 Pressure-strain curves of axisymmetric analyses and experiments at the tanks are shown. The slopes are identical for experiment

Fig. 9 Pressure-strain curves of axisymmetric analyses andand analyses in small strain

experiments at the region before
tanks arenonlinear
shown. behavior of resin becomes
The slopes prominent.
are identical forThey deviate in large strain region
experiment and analyses in small strain region before nonlinear
[DOI: behavior of resin becomes prominent. They©deviate
10.1299/transjsme.19-00338] in large
201x The Japan strain
Society of Mechanical Engineers
region where nonlinear behavior of resin affects much, since the nonlinear behavior of resin is not taken into account in axisymmetric
analyses. Predicted bust pressure by the analyses exceed the experimental results by 30 % because of inadequate handling of
nonlinear behavior of resin.
0

NISSAN TECHNICAL REVIEW No.88 (2022) 123

(c) Tank No.3.
 Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis

layer becomes remarkable; hence, the analysis results volume content Vf of 95.0%, in order to ensure that Vf of
deviate from the measurement results. In Tank No. 2, the entire meso-scale model reaches 55.0%. The helical
the fiber-direction strain of the outermost helical layer is layer meso-scale model shown in Fig. 10 (b) has a length
relatively low; therefore, the analysis results agree well of 80.0 mm in the z-direction, width of 46.2 mm in the
with the measurement results over the entire pressure x-direction, and thickness of 1.00 mm in the y-direction.
range. In Tanks No. 3 and No. 4, the analysis results and In this region, four carbon fiber bundles were arranged,
the measurement results agree well in the low strain with the upper layer inclined at +30.0 and the lower
region; however, the analysis results gradually deviate layer inclined at –30.0 with respect to the z-axis, thus
from the measurement results as the inelastic enabling the bundles to intersect at the center of the
deformation of the resin progresses. z-direction length such that the bends appeared in steps.
The fracture positions predicted by the analysis, Only the carbon fiber bundles are shown in Fig. 10 (b).
including those in the thickness direction of the CFRP Similar to the helical layer meso-scale model, the hoop
layer, agree well with the experimental results. However, layer meso-scale model was arranged at equal intervals
the burst pressures of Tanks No. 1, 2, 3, and 4 predicted along the z-direction within a region with a length of 80.0
by the analysis, which are 17.2, 29.6, 29.8, and 47.4 MPa, mm in the z-direction, width of 46.2 mm in the x-direction,
respectively, are approximately 30% higher than the and thickness of 0.500 mm in the y-direction; eight fiber
corresponding experimental results of 12.0, 26.3, 24.8, bundles were included in the z-direction. Only the fiber
and 35.3 MPa. In other words, using the axisymmetric bundle is depicted in Fig. 10 (c).
continuum model affords predictions on the The meso-scale model used for the analysis was
overestimation side. If the results are to be utilized for constructed by combining the helical and hoop layer
design purposes, an appropriate safety margin should be models to appropriately represent the meso-structure of
set considering the differences. The methodology of the relevant part. The nodes were shared and bonded
repeating design, manufacturing, and burst tests to set between the layers and between the fiber bundle and
the safety margins according to an empirical rule is often resin. As the analysis target was the center of the
adopted; however, in many cases, a large safety margin is cylindrical part, which is the fracture position of the
set conservatively, resulting in an unnecessary increase tanks shown in Fig. 6, it is not necessary to consider the
in the tank weight. To address this and realize a rational effect of the bending deformation that occurs in the dome
minimum weight design, analyses based on a meso-scale part. Consequently, the macroscopic deformation field of
model are required. the internal tank pressure can be evaluated using the
membrane theory. According to the membrane theory,
4. Burst pressure predicted by zooming the curvature of the cylindrical part affects only the
analysis using meso-scale model equilibrium equations of the internal pressure and the
in-plane stress. In this analysis method, the displacement
4.1 Meso-scale model field that satisfies the equilibrium equation is evaluated
This work attempted to predict the burst pressure of through axisymmetric model-based analysis, and the
the pressure tanks with high accuracy by performing boundary conditions of the meso-scale model are
zooming analysis using a meso-scale model (Takemoto determined based on the results. Therefore, the influence
and Yoshikawa, 2019), which separates the carbon fiber of the curvature should be neglected.
bundles and matrix resin. According to the test results The meso-scale model of Tank No. 1 is depicted in Fig.
discussed in Section 2 and the analysis results described 11 (a). This model only has a helical layer, that is, the
in Section 3, the burst evaluation target was narrowed helical layer meso-scale model shown in Fig. 10 (b). It
down to the straight cylindrical portion near the center of was discretized with 10-node tetrahedral quadratic
the tank body, because this region was the starting point elements, resulting in 1,033,956 elements and 1,480,278
of the burst. A meso-scale model was set in the x-y-z nodes. A meso-scale model of Tank No. 2 is illustrated in
Cartesian coordinate system by approximating a part of Fig. 11 (b). This model has a two-layer structure
the side surface of the cylinder as a plane. The x-axis consisting of a hoop layer (lower layer in the y-direction)
coincides with the circumferential direction ( -direction) and a helical layer, with 1,530,251 elements and
of the cylinder, the y-axis coincides with the radial 2,149,325 nodes. A meso-scale model of Tank No. 3 is
direction (r-direction), and the z-axis coincides with the provided in Fig. 11 (c). In this model, the order of the
central axis direction. layers is the opposite of that in the model of Tank No. 2,
A meso-scale model of a CFRP bundle featuring a with the lower layer being the helical layer, followed by
columnar body with an elliptical cross-section, which is a the hoop layer. The numbers of elements and nodes were
single carbon fiber bundle (blue), and a strip-shaped the same as those of Tank No. 2. A meso-scale model of
region with a width of 10.0 mm and a thickness of 0.500 Tank No. 4 is depicted in Fig. 11 (d). This model has a
mm, which is the matrix resin (yellow) and serves as a three-layer structure with two hoop layers and one helical
basic unit, is shown in Fig. 10 (a). With the major axis of layer (lowest layer), comprising 2,034,346 elements and
the ellipse set to 3.84 mm and the minor axis set to 0.230 2,827,060 nodes.
mm, the effect of stress concentration at the major axis
end is minimized. The carbon fiber bundle was designed
to possess orthogonal anisotropy with a carbon fiber

124 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis

Even if cracks occur during resin fracture, the resin

does not disappear and the compressive force
perpendicular to the fracture surface is maintained.
In other words, the compressive stress is retained
after the fracture, although the tensile stress is not
retained (LSTC, 2015).
3) Anisotropic fracture law of carbon fiber bundle: The
fracture in the orthogonal direction of the fiber is
governed by the fracture strain of the resin and not
by the fracture strain of the fiber.
Even in the model that accurately reproduces the
meso-structure of the carbon fiber bundle, the stress field
generated in the thin plate test piece owing to the in-plane
tensile load is similar to the plane stress field, and the
locality in the test piece thickness direction is limited. By
contrast, the stress field generated near the boundary
between the helical and hoop layers of the tank changes
considerably along the thickness direction, and it is

Fig. 10 Meso-scale model for zooming analysis.

4.2 Strength model setting

Epoxy resins used as the CFRP matrix resins generally
exhibit strain-rate-dependent properties (Littell et al.,
2008). In our previous report (Takemoto and Yoshikawa,
2019), it was clarified that the following three factors are
appropriate for reflecting the dependence and
constructing the tensile strength expression mechanism
of the CFRP helical layer:
1) Dependence of resin tensile strength on strain rate:
When the resin and fibers break, elastic energy is
released in the resin near the fractured part, and
the strain rate increases locally. The dependence on
the strain rate of the resin increases the resin
strength to control the fracture.
2) Tension/compression asymmetry of resin strength: Fig. 11 Meso-scale model of tank.

NISSAN TECHNICAL REVIEW No.88 (2022) 125

 ­ 1
½
§ dV m ·c ° § H&m · P ° dV m
Kikai, Gijyutsu and Tokyo, Transactions of the JSME¨ (in ¸ ® ¨ ¸ ¾ No.00 (201x)
Japanese),
1  Vol.00,
© dH m ¹ ° © C ¹ ° dH m
Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP ¯ tank based
¿ on meso-scopic zooming analysis
௨ୖࡢᙉᗘࣔࢹࣝࢆලయ໬ࡋࡓ࣐ࢺࣜࢡࢫᶞ⬡ࡢᛂຊ㸫ࡦࡎࡳ⥺ᅗࢆᅗ 12 ࡟♧ࡍ㸬ᶓ㍈ࡣ┦ᙜࡦࡎࡳ㸪⦪㍈ࡣ
┦ᙜᛂຊ࡛࠶ࡿ㸬ᶞ⬡ࢆ➼᪉ᛶᙎረᛶయ࡜ࡋ㸪ᘬᙇᛂຊୗ࡛ࡣᶞ⬡ࡢ┦ᙜࡦࡎࡳࡀ⾲ ࡇࡇ࡛㸪V m ࡣᶞ⬡ࡢ┦ᙜᛂຊ㸪1H࠿ࡽ◚᩿ࡦࡎࡳ m ࡣᶞ⬡ࡢ┦ᙜࡦࡎࡳ㸪 2.60 %ࢆ H& ࡣᶞ⬡ࡢࡦࡎ
necessary to consider the influence of the normal stress (Takemoto, Yoshikawa, 2019) shown in Fig. 13 was set tom
in the thickness direction. To increase the generality of
㉸࠼ࡓ࡜ࡁ࡟ᶞ⬡ࡀ◚ቯࡋࡓ࡜ุ᩿ࡋᛂຊࢆゎᨺࡋࡓ㸬ᶞ⬡ࡀ◚ቯࡋ࡚ࡶ㸪◚ቯ࡟ࡼࡾⓎ⏕ࡍࡿᛂຊ෌㓄ศ࡛⏕ represent the fracture
≉ᛶ࡟ࡼࡿኚ໬㸪C ࠾ࡼࡧmodeP ࡣࢡ࣮ࣃ࣮࣭ࢩࣔࣥࢬᘧࡢᐃᩘ࡛࠶ࡿ㸬๓ሗ
due to the transverse cracks.
the strength model considering the abovementioned
ࡌࡿᅽ⦰ᛂຊࢆᢸ࠺ࡇ࡜ࡣ࡛ࡁࡿࡓࡵ㸪せ⣲ࡢ๐㝖ࡣ⾜ࢃࡎ㸪ᅽ⦰ᛂຊୗ࡛ࡣ㸪㝆అᛂຊ࡟㐩ࡋࡓࡢࡕࡣᅽ⦰ࡦ The vertical axis represents the stress, the horizontal
㦂∦ࡢືⓗᘬᙇヨ㦂ࡼࡾᚓࡽࢀࡓ⤖ᯝ࠿ࡽ㸪◳໬ ᗘ 80/110 Υࡢᶞ⬡࡟ᑐ
aspects,ࡎࡳࡀቑຍࡋ࡚ࡶᛂຊࡣ㝆అᛂຊࡢࡲࡲ࡛࠶ࡿ࡜ࡋࡓ㸬ᘬᙇ࡜ᅽ⦰ࡣయ✚ࡦࡎࡳࡢṇ㈇࡟ࡼࡗุ࡚ᐃࡋࡓ㸬㝆అ
the following two constituent factors were added ᅗaxis
12 represents
࡟㉥⥺࡛♧ࡍ㸪ࡦࡎࡳ㏿ᗘ the strain, the43.4
black
/sec line
ࡢ indicates the
ᐃ⤖ᯝࡲ࡛ࡋ࠿ᚓࡽࢀ࡚࠸࡞
to the aforementioned three constituent factors in order fiber direction, and the blue line shows the fiber
ᑐࡋ࡚ࡶࡑࡢᛂຊ㸫ࡦࡎࡳ㛵ಀࢆ㐺⏝ࡍࡿࡶࡢ࡜ࡋࡓ㸬
Ⅼࡢࡦࡎࡳ࠾ࡼࡧ◚᩿ࡦࡎࡳࡣ୙ኚ࡜ࡋ㸪 㝆అᛂຊࡢࡦࡎࡳ㏿ᗘ౫Ꮡᛶࢆ⾲ࡍࢡ࣮ࣃ࣮࣭ࢩࣔࣥࢬᘧ (Dietenberger
to create the strength model: orthogonal direction. The fracture࠸ࢃࡺࡿࢺࣛࣥࢫࣂ࣮ࢫࡁ⿣࡟ࡼࡿ◚
୍᪉㸪Ⅳ⣲⧄⥔᮰ࡢᙉᗘࣔࢹࣝࡣ㸪 strain of the carbon
et al., 2005) of࡟‽ࡌ࡚㸪ᘬᙇ࠿ࡽᅽ⦰ࡲ࡛ࡢᛂຊ඲ᇦ࡟࠾࠸࡚㸪௨ୗࡢࡦࡎࡳ㏿ᗘ౫Ꮡᛶ≉ᛶࢆ᭷ࡍࡿ࡜ࡋࡓ㸬
4) Dependence resin rigidity on strain rate: Young's fiber was set to 2.10% in the fiber direction, based on
ࡍ␗᪉ᛶ⥺ᙧᙎᛶయ◚ቯࣔࢹࣝ (➉ᮏ㸪ྜྷᕝ㸪2019) ࢆタᐃࡋࡓ㸬⦪㍈ࡣᛂ
modulus increases as the strain rate increases Table 1, whereas the fracture strain in the orthogonal
᪉ྥ㸪㟷⥺ࡣ⧄⥔┤஺᪉ྥ࡟㛵ࡍࡿࡶࡢ࡛࠶ࡿ㸬◚ቯつ‽ࡣ㸪⾲ 1 ࠿ࡽ㸪⧄
under compressive ­ and tensile
1
½ stresses. direction of the fiber, in which resin fracture seemed to be
5) Dependence § dV mof·cresin § H&m · P ° dV mstrength on strain
° compressive ࡎࡳ 2.10 %㸪⧄⥔┤஺᪉ྥ࡟㛵ࡋ࡚ࡣᶞ⬡ࡢ◚ቯࡀᨭ㓄ⓗ࡜⪃࠼◚᩿ࡦࡎࡳ
dominant, was set to 2.60%. It was assumed that
¨ 1
¸ ® ¨ ¸ ¾ (1) the
rate: The dH
© compressive
m ¹ C
°¯ © strength d H
¹ ° m increases as the fiber bundle did not break under compressive stress.
⧄⥔᮰ࡣ◚᩿ࡋ࡞࠸࡜ࡋࡓ㸬
¿
strain rate increases, under both tensile and
compressive stresses.
ࡇࡇ࡛㸪V m ࡣᶞ⬡ࡢ┦ᙜᛂຊ㸪 H m ࡣᶞ⬡ࡢ┦ᙜࡦࡎࡳ㸪 H&m ࡣᶞ⬡ࡢࡦࡎࡳ㏿ᗘ㸪( )’ࡣᶞ⬡ࡢࡦࡎࡳ㏿ᗘ౫Ꮡ
A stress–strain
≉ᛶ࡟ࡼࡿኚ໬㸪C diagram࠾ࡼࡧof the matrix resin that
P ࡣࢡ࣮ࣃ࣮࣭ࢩࣔࣥࢬᘧࡢᐃᩘ࡛࠶ࡿ㸬๓ሗ (➉ᮏ㸪ྜྷᕝ㸪2019) ࡛⾜ࡗࡓᶞ⬡ヨ
embodies the above strength model is provided in Fig. 12.
㦂∦ࡢືⓗᘬᙇヨ㦂ࡼࡾᚓࡽࢀࡓ⤖ᯝ࠿ࡽ㸪◳໬ ᗘ 80/110 Υࡢᶞ⬡࡟ᑐࡋ࡚ C = 24.1㸪P = 1.00 ࡜ྠᐃࡋࡓ㸬
The horizontal and vertical axes represent the equivalent
ᅗ 12 ࡟㉥⥺࡛♧ࡍ㸪ࡦࡎࡳ㏿ᗘ 43.4 /sec ࡢ ᐃ⤖ᯝࡲ࡛ࡋ࠿ᚓࡽࢀ࡚࠸࡞࠸ࡢ࡛㸪ࡑࢀࢆ㉸࠼ࡿࡦࡎࡳ㏿ᗘ࡟
strain and stress, respectively. The resin was formed into
ᑐࡋ࡚ࡶࡑࡢᛂຊ㸫ࡦࡎࡳ㛵ಀࢆ㐺⏝ࡍࡿࡶࡢ࡜ࡋࡓ㸬
an isotropic elastoplastic body, and when the equivalent
the resin exceeded 2.60% (based on
strain of ୍᪉㸪Ⅳ⣲⧄⥔᮰ࡢᙉᗘࣔࢹࣝࡣ㸪 Table 1) of
࠸ࢃࡺࡿࢺࣛࣥࢫࣂ࣮ࢫࡁ⿣࡟ࡼࡿ◚ቯ࣮ࣔࢻࢆᐇ⌧ࡍࡿࡓࡵ㸪ᅗ 13 ࡟♧
the fracture strain under tensile stress,
ࡍ␗᪉ᛶ⥺ᙧᙎᛶయ◚ቯࣔࢹࣝ it was judged ࢆタᐃࡋࡓ㸬⦪㍈ࡣᛂຊ㸪ᶓ㍈ࡣࡦࡎࡳ࡛࠶ࡾ㸪㯮⥺ࡣ⧄⥔
(➉ᮏ㸪ྜྷᕝ㸪2019)
that the resin had broken and that the stress had been
᪉ྥ㸪㟷⥺ࡣ⧄⥔┤஺᪉ྥ࡟㛵ࡍࡿࡶࡢ࡛࠶ࡿ㸬◚ቯつ‽ࡣ㸪⾲ 1 ࠿ࡽ㸪⧄⥔᪉ྥ࡟㛵ࡋ࡚ࡣⅣ⣲⧄⥔ࡢ◚᩿ࡦ
released. Even if the resin breaks, it can bear the
ࡎࡳ 2.10 %㸪⧄⥔┤஺᪉ྥ࡟㛵ࡋ࡚ࡣᶞ⬡ࡢ◚ቯࡀᨭ㓄ⓗ࡜⪃࠼◚᩿ࡦࡎࡳ 2.60 %࡜タᐃࡋࡓ㸬ᅽ⦰ᛂຊୗ࡛ࡣ
compressive stress generated during the stress
⧄⥔᮰ࡣ◚᩿ࡋ࡞࠸࡜ࡋࡓ㸬
redistribution caused by the facture; hence, it was
designed without removing any elements to enable the
yield stress to be maintained after being reached under
the compressive stress, regardless of the increase in Fig. 12 Asymmetric fracture model of matrix resin
for meso-scale analysis.
compressive strain. Tension or compression was judged
based on whether the volumetric strain was positive or Fig. 12 Asymmetric fracture model of matrix resin Fig. 13 An
negative. Based on the permanent strain for the yield
for meso-scale analysis. fib
point strain and the fracture strain, the following strain
rate dependence characteristics were applied over the
entire stress range from tension to compression, according
to the Cooper–Simmons formula (Dietenberger et al., 4࣭3 ࢬ࣮࣑ࣥࢢゎᯒ
2005), which expresses the strain rate dependence of the 3 ❶࡛⾜ࡗࡓ㍈ᑐ⛠㐃⥆యࣔࢹࣝ࡟ࡼࡿᐜჾࡢゎᯒ⤖ᯝ࠿ࡽᚓࡽࢀࡿ◚ቯ
yield stress: ሙࢆ㸪࣓ࢰࢫࢣ࣮ࣝࣔࢹࣝࡢቃ⏺᮲௳࡜ࡋ࡚タᐃࡋࢬ࣮࣑ࣥࢢゎᯒࢆ⾜ࡗࡓ
㈇Ⲵࢆ㍍ῶࡍࡿࡓࡵ௨ୗࡢ㏆ఝゎᯒᡭἲࢆ᥇ࡗࡓ㸬
[DOI: 10.1299/transjsme.19-00338] © 201x The

Fig. 13 Anisotropic fracture model of carbon

fiber bundle for meso-scale analysis.

Fig. 12 Asymmetric fracture model of matrix resin Fig. 13 Anisotropic fracture model of carbon
where is the equivalent stress of the resin, is 4.3 Zooming analysis
for meso-scale analysis. fiber bundle for meso-scale analysis.
the equivalent strain of the resin, is the strain rate of The displacement field near the center of the cylindrical
the resin, ()' indicates the changes in the resin due to part, which is the starting point of the fracture, as
strain-rate-dependent characteristics, and C and P are indicated by the analysis based on the axisymmetric
constants4࣭3
of theࢬ࣮࣑ࣥࢢゎᯒ
Cooper–Simmons formula. According to continuum model described in Section 3, was set as the
the results of the dynamic tensile test on the resin test
3 ❶࡛⾜ࡗࡓ㍈ᑐ⛠㐃⥆యࣔࢹࣝ࡟ࡼࡿᐜჾࡢゎᯒ⤖ᯝ࠿ࡽᚓࡽࢀࡿ◚ቯ㉳Ⅼ࡛࠶ࡿ෇⟄⬗㒊୰ኸ㏆ഐࡢኚ఩ boundary condition of the meso-scale model, and zooming
piece described in our previous report (Takemoto and analysis was conducted. Considering the issues associated
ሙࢆ㸪࣓ࢰࢫࢣ࣮ࣝࣔࢹࣝࡢቃ⏺᮲௳࡜ࡋ࡚タᐃࡋࢬ࣮࣑ࣥࢢゎᯒࢆ⾜ࡗࡓ㸬タィၥ㢟࡬ࡢ㐺⏝ࢆ᝿ᐃࡋ㸪ィ⟬
Yoshikawa, 2019), C = 24.1 and P = 1.00 were identified with the application to design, the following approximate
㈇Ⲵࢆ㍍ῶࡍࡿࡓࡵ௨ୗࡢ㏆ఝゎᯒᡭἲࢆ᥇ࡗࡓ㸬
for the resin with a curing temperature of 80/110 C. As analysis method was adopted to reduce the calculation
only measurement results corresponding to a strain rate
[DOI: 10.1299/transjsme.19-00338] load: © 201x The Japan Society of Mechanical Engineers
of 43.4/s, indicated by the red line in Fig. 12, were 1) In the axisymmetric continuum model analysis for
obtained, the stress–strain relationship was applied to the tank, the CFRP is considered as a linear elastic
the strain rates exceeding this value. body; however, in the meso-scale model, the
By contrast, for the strength model of the carbon fiber nonlinear behavior and fracture model of the resin
bundle, the anisotropic linear elastic fracture model and the orthogonal anisotropic linear elastic body

126 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis

fracture model of the carbon fiber bundle were local increase of approximately 2.1% in the fiber fracture
incorporated into the analysis. strain occurred at the crimp (bending) part of the helical
2) According to the homogenization method (Arai et layer intersection, and the fracture commenced from this
al., 2008), the nonlinear analysis results of the point. Mesoscopic zooming analysis enabled the prediction
meso-scale model should be fed back to the of the burst mechanism at the meso-scale, which was not
axisymmetric continuum model to proceed with the captured by the axisymmetric continuum model.
sequential analysis. However, this step was not According to the results for Tank No. 2 in Fig. 16, the
performed in this study; instead, a one-way increase in local strain due to the crossing of fiber bundles
procedure from the axisymmetric continuum model in the upper helical layer induces a fiber-direction strain
analysis to the meso-scale model zooming analysis similar to the fracture strain in the lower hoop layer fiber
was employed. bundles, which was the starting point of the fracture. The
Based on the analysis results in Section 3, the predicted fracture from the lower hoop layer matches the
displacement field corresponding to the average vertical experimental results presented in Section 2. According to
strain in the central axis direction and the average the results for Tank No. 3 in Fig. 17, although the layer
circumferential strain of each element with respect to the order is reversed compared to that of Tank No. 2, the
position of 40.0 mm on the central axis direction from the strain increases according to the same mechanism, and
8QLWPPof the0DWUL[UHVLQ
central surface tank body were obtained at each the fracture commences at the hoop layer fiber bundle
internal  pressure;
 furthermore, the prescribed near the intersection with the helical layer fiber bundle.
displacements Uz, -Uz, Ux, and -Ux with respect to the The predicted fracture starting position also agrees with
meso-scale model edge, shown in Fig. 14, were calculated. the experimental results.


To suppress the out-of-plane deformation caused by the The results for Tank No. 4 immediately before
structural asymmetry of the meso-scale model, the edges fracturing are presented in Figs. 18 (a), (b), and (c), and
&DUERQILEHUEXQGOH
that yielded the prescribed displacement were constrained those immediately after fracturing are depicted in Fig. 18
along the y-direction. As the increases in local stress or (d). The distribution maps for all fiber bundles from the
strain occurred near the edges, it was determined that no outermost layer side are provided in Figs. 18 (a) and (d);
fracture checks would be performed y in an area equivalent the internal view of the hoop layer fiber bundles is shown
y 8QLWPP
to the size of one fiber bundle from the edge (outside the
8QLWPP in Fig. 18 (b), and the view from the innermost layer is
z x
z x area, 60.0 mm in the z-direction and 34.7 mm in the provided in Fig. 18 (c). Fig. 18 (b) indicates that the local
㸦㸧
㸦㸧

x-direction, surrounded by the blue line shown in Fig. strain inside the hoop layer increases. Moreover, Fig. 18
14). The pressurization rate for the analysis was set to (c) shows that the local increase in strain due to the
1.50 10–4 s/MPa to shorten the calculation time as much crossed fiber bundle in the helical layer resulted in an
as possible within the range where there was no difference increase in the strain up to approximately the fracture
in the analysis results. strain in the hoop layer fiber bundle, which was the
starting point of the fracture shown in Fig. 18 (d), similar
to the cases of Tanks No. 2 and No. 3. These results are
1R IUDFWXUH DUHD also consistent with the fracture test results. The main
-Uz 3UHVFULEHGGLVSODFHPHQW
-Ux 3UHVFULEHGGLVSODFHPHQW Uy ŽŶƐƚƌĂŝŶƚ strain distribution of the matrix resin for Tank No. 4 is
Uy ŽŶƐƚƌĂŝŶƚ presented in Fig. 19. Internal views of the hoop layer are
provided, where (a) and (b) correspond to the first and
y
third principal strains, respectively. The maximum value
z x
of the first principal strain exceeds the static fracture
strain of the resin (2.6%), as shown in Table 1, and the
Ux 3UHVFULEHGGLVSODFHPHQW minimum value of the third principal strain is
Uz 3UHVFULEHGGLVSODFHPHQW Uy ŽŶƐƚƌĂŝŶƚ approximately –5%. If the proposed strength model that
Uy ŽŶƐƚƌĂŝŶƚ depends on the strain rate is not used, it is expected that
8QLWPP the fracture progress would be accelerated and the burst
Fig. 14 Boundary condition of meso-scale model pressure would be lower than that obtained in the burst
for zooming analysis. test. The prediction of a low burst pressure is safe and
desirable for ensuring tank reliability; however, a
conservative design is not economical. The proposed
4.4 Analysis results method enables more accurate predictions, affording
The first-principles strain distributions of the carbon both high reliability and economy.
fiber bundle surfaces, obtained from the analysis results
based on the meso-scale models of Tanks No. 1, 2, 3, and
4, are presented in Figs. 15, 16, 17, and 18, respectively.
The region depicted in each case is that surrounded by
the blue lines in Fig. 14. In Figs. 15–17, the distributions
immediately before and after fracture are shown in (a)
and (b), respectively.
According to the results for Tank No. 1 in Fig. 15, a

NISSAN TECHNICAL REVIEW No.88 (2022) 127

 Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis

7DNHPRWRDQG<RVKLNDZD7UDQVDFWLRQVRIWKH-60( LQ-DSDQHVH
7DNHPRWRDQG<RVKLNDZD7UDQVDFWLRQVRIWKH-60( LQ-DSDQHVH 9RO1R
9RO1R 


7DNHPRWRDQG<RVKLNDZD7UDQVDFWLRQVRIWKH-60( LQ-DSDQHVH 9RO1R 

VWSULQFLSDO
VWSULQFLSDO
6WUDLQFRQFHQWUDWLRQ
6WUDLQFRQFHQWUDWLRQ )UDFWXUHVWDUWLQJSRLQW
)UDFWXUHVWDUWLQJSRLQW
VWUDLQ>@
VWUDLQ>@
VWSULQFLSDO


6WUDLQFRQFHQWUDWLRQ )UDFWXUHVWDUWLQJSRLQW
VWUDLQ>@















            DD 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH
3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH     EE 3UHVVXUHRI
3UHVVXUHRI 03D
03DMXVWDIWHUIUDFWXUH
MXVWDIWHUIUDFWXUH
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
Fig. 15 1st principal strain distributions of carbon fiber bundles obtained by zooming analysis of tank No.1 are shown for just before
(a) and
   after  (b)
   fracture.
 DDDQGDIWHU  Local strain concentration occurs to raise the
D 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH   maximum
 first principal strain in
E 3UHVVXUHRI theMXVWDIWHUIUDFWXUH
03D vicinity of crimp part of the
DQGDIWHU EE IUDFWXUH/RFDOVWUDLQFRQFHQWUDWLRQRFFXUVWRUDLVHWKHPD[LPXVILUVWSULQFLSDOVWUDLQLQWKHYLFLQLW\RIFULPS
IUDFWXUH/RFDOVWUDLQFRQFHQWUDWLRQRFFXUVWRUDLVHWKHPD[LPXVILUVWSULQFLSDOVWUDLQLQWKHYLFLQLW\RIFULPS
helical layer. The fracture initiates at the point of the strain concentration, when it reaches to 2.1 %, that is, fracture elongation of
carbon fiber.
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
SDUWRIWKHKHOLFDOOD\HU7KHIUDFWXUHLQLWLDWHVDWWKHSRLQWRIWKHVWUDLQFRQFHQWUDWLRQZKHQLWUHDFKHVWRWKDWLVIUDFWXUH
SDUWRIWKHKHOLFDOOD\HU7KHIUDFWXUHLQLWLDWHVDWWKHSRLQWRIWKHVWUDLQFRQFHQWUDWLRQZKHQLWUHDFKHVWRWKDWLVIUDFWXUH
D DQGDIWHU E IUDFWXUH/RFDOVWUDLQFRQFHQWUDWLRQRFFXUVWRUDLVHWKHPD[LPXVILUVWSULQFLSDOVWUDLQLQWKHYLFLQLW\RIFULPS
HORQJDWLRQRIFDUERQILEHU
HORQJDWLRQRIFDUERQILEHU
SDUWRIWKHKHOLFDOOD\HU7KHIUDFWXUHLQLWLDWHVDWWKHSRLQWRIWKHVWUDLQFRQFHQWUDWLRQZKHQLWUHDFKHVWRWKDWLVIUDFWXUH
HORQJDWLRQRIFDUERQILEHU
VWSULQFLSDO
VWSULQFLSDO
6WUDLQFRQFHQWUDWLRQ
6WUDLQFRQFHQWUDWLRQ )UDFWXUHVWDUWLQJSRLQW
)UDFWXUHVWDUWLQJSRLQW
VWUDLQ>@
VWUDLQ>@
VWSULQFLSDO


6WUDLQFRQFHQWUDWLRQ )UDFWXUHVWDUWLQJSRLQW
VWUDLQ>@















                DD 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH
3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH       EE 3UHVVXUHRI
3UHVVXUHRI 03DMXVWDIWHUIUDFWXUH
03DMXVWDIWHUIUDFWXUH 
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
Fig. 16 1st principal strain distributions of carbon fiber bundles obtained by zooming analysis of tank No.2 are shown for just before
(a) and after  (b)
   fracture.
 DDDQGDIWHU  The D local strain concentration of the crimp part   in  the
  upper helical layer causes increases of the local
    EE IUDFWXUH7KHORFDOVWUDLQFRQFHQWUDWLRQRIWKHFULPSSDUWLQWKHXSSHUKHOLFDOOD\HUFDXVHVLQFUHDVHVRIWKHORFDO
3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH E 3UHVVXUHRI 03DMXVWDIWHUIUDFWXUH  strain
DQGDIWHU IUDFWXUH7KHORFDOVWUDLQFRQFHQWUDWLRQRIWKHFULPSSDUWLQWKHXSSHUKHOLFDOOD\HUFDXVHVLQFUHDVHVRIWKHORFDO
of the lower hoop layer. The strain reaches to 2.1 % at the moment of fracture.
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
VWUDLQRIWKHORZHUKRRSOD\HU7KHVWUDLQUHDFKHVWRDWWKHPRPHQWRIIUDFWXUH
VWUDLQRIWKHORZHUKRRSOD\HU7KHVWUDLQUHDFKHVWRDWWKHPRPHQWRIIUDFWXUH
D DQGDIWHU E IUDFWXUH7KHORFDOVWUDLQFRQFHQWUDWLRQRIWKHFULPSSDUWLQWKHXSSHUKHOLFDOOD\HUFDXVHVLQFUHDVHVRIWKHORFDO
VWUDLQRIWKHORZHUKRRSOD\HU7KHVWUDLQUHDFKHVWRDWWKHPRPHQWRIIUDFWXUH
VWSULQFLSDO
VWSULQFLSDO
VWUDLQ>@
VWUDLQ>@ 6WUDLQFRQFHQWUDWLRQ
6WUDLQFRQFHQWUDWLRQ )UDFWXUHVWDUWLQJSRLQW
)UDFWXUHVWDUWLQJSRLQW
VWSULQFLSDO


VWUDLQ>@

 6WUDLQFRQFHQWUDWLRQ )UDFWXUHVWDUWLQJSRLQW













                      DD 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH
3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH         EE 3UHVVXUHRI
3UHVVXUHRI 03DMXVWDIWHUIUDFWXUH
03DMXVWDIWHUIUDFWXUH 
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
       DDDQGDIWHU   D 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH WKHORFDOVWUDLQHQKDQFHPHQWDWWKHFULPSSDUWRIWKHORZHU
        EE IUDFWXUH%\WKHVLPLODUPHFKDQLVPRIWDQN1R
DQGDIWHU IUDFWXUH%\WKHVLPLODUPHFKDQLVPRIWDQN1R        E 3UHVVXUHRI 03DMXVWDIWHUIUDFWXUH 
WKHORFDOVWUDLQHQKDQFHPHQWDWWKHFULPSSDUWRIWKHORZHU
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
Fig. KHOLFDOOD\HUFDXVHVLQFUHDVHRIWKHORFDOVWUDLQRIWKHXSSHUKRRSOD\HU
17 1st principal strain distributions of carbon fiber bundles obtained by zooming analysis of tank No.3 are shown for just before
KHOLFDOOD\HUFDXVHVLQFUHDVHRIWKHORFDOVWUDLQRIWKHXSSHUKRRSOD\HU
(a) and after (b) fracture. By the similar mechanism of tank No.2, the local strain enhancement at the crimp part of the lower helical
D
layer causesDQGDIWHU
increaseEofIUDFWXUH%\WKHVLPLODUPHFKDQLVPRIWDQN1R
the local strain of the upper hoop layer. WKHORFDOVWUDLQHQKDQFHPHQWDWWKHFULPSSDUWRIWKHORZHU
KHOLFDOOD\HUFDXVHVLQFUHDVHRIWKHORFDOVWUDLQRIWKHXSSHUKRRSOD\HU
>'2,WUDQVMVPH@
>'2,WUDQVMVPH@ ‹7KH-DSDQ6RFLHW\RI0HFKDQLFDO(QJLQHHUV
‹7KH-DSDQ6RFLHW\RI0HFKDQLFDO(QJLQHHUV 


>'2,WUDQVMVPH@ 128‹7KH-DSDQ6RFLHW\RI0HFKDQLFDO(QJLQHHUV
No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis
7DNHPRWRDQG<RVKLNDZD7UDQVDFWLRQVRIWKH-60( LQ-DSDQHVH 9RO1R 

7DNHPRWRDQG<RVKLNDZD7UDQVDFWLRQVRIWKH-60( LQ-DSDQHVH 9RO1R 

ࡿୖ࡛ࡣᮃࡲࡋ࠸ࡀ㸪ಖᏲⓗࡍࡂࡿタィࡣ⤒῭ⓗ࡛ࡣ࡞ࡃ㸪ᮏ◊✲࡛♧ࡋࡓᡭἲ࡟ࡼࡾⓗ☜࡟◚⿣ᅽຊࢆண ࡍ
ࡿࡇ࡜࡛ᐜჾࡢಙ㢗ᛶ࡜⤒῭ᛶࢆ୧❧࡛ࡁࡿ㸬
ࡿୖ࡛ࡣᮃࡲࡋ࠸ࡀ㸪ಖᏲⓗࡍࡂࡿタィࡣ⤒῭ⓗ࡛ࡣ࡞ࡃ㸪ᮏ◊✲࡛♧ࡋࡓᡭἲ࡟ࡼࡾⓗ☜࡟◚⿣ᅽຊࢆண ࡍ
ࡿࡇ࡜࡛ᐜჾࡢಙ㢗ᛶ࡜⤒῭ᛶࢆ୧❧࡛ࡁࡿ㸬 6WUDLQFRQFHQWUDWLRQ
7DNHPRWRDQG<RVKLNDZD7UDQVDFWLRQVRIWKH-60( LQ-DSDQHVH 9RO1R 
VWSULQFLSDO 6WUDLQFRQFHQWUDWLRQ
VWUDLQ>@
ࡿୖ࡛ࡣᮃࡲࡋ࠸ࡀ㸪ಖᏲⓗࡍࡂࡿタィࡣ⤒῭ⓗ࡛ࡣ࡞ࡃ㸪ᮏ◊✲࡛♧ࡋࡓᡭἲ࡟ࡼࡾⓗ☜࡟◚⿣ᅽຊࢆண
VWSULQFLSDO ࡍ

VWUDLQ>@
ࡿࡇ࡜࡛ᐜჾࡢಙ㢗ᛶ࡜⤒῭ᛶࢆ୧❧࡛ࡁࡿ㸬



 6WUDLQFRQFHQWUDWLRQ


VWSULQFLSDO


VWUDLQ>@

 D 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH      E 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH 
   :KROHOD\HUVYLHZIURPRXWHUVLGH     +RRS OD\HUV YLHZ IURPLQQHU VLGH
D 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH      E 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH 

 :KROHOD\HUVYLHZIURPRXWHUVLGH    )UDFWXUHVWDUWLQJSRLQW
 +RRS OD\HUV YLHZ IURPLQQHU VLGH
6WUDLQFRQFHQWUDWLRQ
 )UDFWXUHVWDUWLQJSRLQW
 6WUDLQFRQFHQWUDWLRQ
D 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH      E 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH 
 :KROHOD\HUVYLHZIURPRXWHUVLGH     +RRS OD\HUV YLHZ IURPLQQHU VLGH

)UDFWXUHVWDUWLQJSRLQW
                  6WUDLQFRQFHQWUDWLRQ

F 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH         G 3UHVVXUHRI03DMXVWDIWHUIUDFWXUH

:KROHOD\HUVYLHZIURPLQQHUVLGH     :KROH OD\HUV YLHZ IURPRXWHU VLGH
F 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH         G 3UHVVXUHRI03DMXVWDIWHUIUDFWXUH
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
:KROHOD\HUVYLHZIURPLQQHUVLGH     :KROH OD\HUV YLHZ IURPRXWHU VLGH
D  E  F DQGDIWHU G IUDFWXUH%RWK D DQG G LOOXVWUDWHE\WKHYLHZIURPRXWHUVLGHIRUZKROHOD\HUV E LVLQQHUVLGHYLHZ
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
Fig. 18 1stIRUKRRSOD\HU
principal strainF distributions of carbon fiber bundles obtained by zooming analysis of tank No.4 are shown for just before
LVLQQHUVLGHYLHZIRUZKROHOD\HUV 7KHORFDOVWUDLQFRQFHQWUDWLRQRFFXUVLQWKHKRRSOD\HUFDXVHGE\WKH
(a), (b), (c)Dand  E after
 F (d) fracture.
DQGDIWHU G Both (a) and (d)
IUDFWXUH%RWK D illustrate
DQG G by the view from outer side for whole layers. (b) isE inner
LOOXVWUDWHE\WKHYLHZIURPRXWHUVLGHIRUZKROHOD\HUV side view for hoop
LVLQQHUVLGHYLHZ
layer. (c)
  is
 inner
    side     for
   view   whole
    layers. The local strain concentration occurs in the hoop layer caused by
FULPSLQKHOLFDOOD\HU7KHIUDFWXUHRIKRRSOD\HUILEHUEXQGOHVWDUWVZKHQLWVORFDOVWUDLQUHDFKHVWR  the crimp in helical
layer. TheIRUKRRSOD\HU
fracture of hoop layer fiber bundle starts when its local
F LVLQQHUVLGHYLHZIRUZKROHOD\HUV strain reaches to 2.1 %.
7KHORFDOVWUDLQFRQFHQWUDWLRQRFFXUVLQWKHKRRSOD\HUFDXVHGE\WKH
F 3UHVVXUHRI03DMXVWEHIRUHIUDFWXUH         G 3UHVVXUHRI03DMXVWDIWHUIUDFWXUH
FULPSLQKHOLFDOOD\HU7KHIUDFWXUHRIKRRSOD\HUILEHUEXQGOHVWDUWVZKHQLWVORFDOVWUDLQUHDFKHVWR
VWSULQFLSDO :KROHOD\HUVYLHZIURPLQQHUVLGH UG SULQFLSDO     :KROH OD\HUV YLHZ IURPRXWHU VLGH
)LJVWSULQFLSDOVWUDLQGLVWULEXWLRQVRIFDUERQILEHUEXQGOHVREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1RDUHVKRZQIRUMXVWEHIRUH
VWUDLQ>@ VWUDLQ>@
VWSULQFLSDO
D  E  F DQGDIWHU G IUDFWXUH%RWK D DQG G UG SULQFLSDO
LOOXVWUDWHE\WKHYLHZIURPRXWHUVLGHIRUZKROHOD\HUV E LVLQQHUVLGHYLHZ
 
VWUDLQ>@ VWUDLQ>@
 IRUKRRSOD\HU F LVLQQHUVLGHYLHZIRUZKROHOD\HUV 7KHORFDOVWUDLQFRQFHQWUDWLRQRFFXUVLQWKHKRRSOD\HUFDXVHGE\WKH


 FULPSLQKHOLFDOOD\HU7KHIUDFWXUHRIKRRSOD\HUILEHUEXQGOHVWDUWVZKHQLWVORFDOVWUDLQUHDFKHVWR
  
 
 
 

VWSULQFLSDO  SULQFLSDO
UG
 

VWUDLQ>@ 
VWUDLQ>@
 
 
 
 
D VWSULQFLSDOVWUDLQ                          
    E UGSULQFLSDOVWUDLQ

)LJ'LVWULEXWLRQVRIVWSULQFLSDOVWUDLQ D DQGUGSULQFLSDOVWUDLQ E RIPDWUL[UHVLQREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1R

 D VWSULQFLSDOVWUDLQ                               E UGSULQFLSDOVWUDLQ
DUHVKRZQE\WKHYLHZIURPLQQHUVLGHDWLQWHUQDOSUHVVXUHRI03DWKDWLVMXVWEHIRUHWKHIUDFWXUH
 7KHPD[LPXPYDOXH

)LJ'LVWULEXWLRQVRIVWSULQFLSDOVWUDLQ D DQGUGSULQFLSDOVWUDLQ E RIPDWUL[UHVLQREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1R
RIWKHVWSULQFLSDOVWUDLQH[FHHGVWKDWLVIUDFWXUHHORQJDWLRQRIPDWUL[UHVLQ7KHPLQLPXPYDOXHRIWKHUGSULQFLSDO


DUHVKRZQE\WKHYLHZIURPLQQHUVLGHDWLQWHUQDOSUHVVXUHRI03DWKDWLVMXVWEHIRUHWKHIUDFWXUH 7KHPD[LPXPYDOXH
VWUDLQLVDERXW
RIWKHVWSULQFLSDOVWUDLQH[FHHGVWKDWLVIUDFWXUHHORQJDWLRQRIPDWUL[UHVLQ7KHPLQLPXPYDOXHRIWKHUGSULQFLSDO

VWUDLQLVDERXW
࣭ ࢬ࣮࣑ࣥࢢゎᯒ⤖ᯝ࡜◚⿣ヨ㦂⤖ᯝ࡜ࡢẚ㍑
D VWSULQFLSDOVWUDLQ                               E UGSULQFLSDOVWUDLQ

)LJ'LVWULEXWLRQVRIVWSULQFLSDOVWUDLQ
Fig. 19 Distributions of 1st principal strain (a) andD DQGUGSULQFLSDOVWUDLQ
3rd principal strain (b) E
ofRIPDWUL[UHVLQREWDLQHGE\]RRPLQJDQDO\VLVRIWDQN1R
matrix resin obtained by zooming analysis of tank No.4 are
shown ࣭
by theࢬ࣮࣑ࣥࢢゎᯒ⤖ᯝ࡜◚⿣ヨ㦂⤖ᯝ࡜ࡢẚ㍑
view from inner side at internal pressure of 36.5 MPa, that
>'2,WUDQVMVPH@ is, just before the fracture. The maximum value of 
‹7KH-DSDQ6RFLHW\RI0HFKDQLFDO(QJLQHHUV the 1st
DUHVKRZQE\WKHYLHZIURPLQQHUVLGHDWLQWHUQDOSUHVVXUHRI03DWKDWLVMXVWEHIRUHWKHIUDFWXUH
principal strain exceeds 2.6 %, that is, fracture elongation of matrix resin. The minimum value of the 3rd principal 7KHPD[LPXPYDOXH
strain is about -5 %.
RIWKHVWSULQFLSDOVWUDLQH[FHHGVWKDWLVIUDFWXUHHORQJDWLRQRIPDWUL[UHVLQ7KHPLQLPXPYDOXHRIWKHUGSULQFLSDO
>'2,WUDQVMVPH@ ‹7KH-DSDQ6RFLHW\RI0HFKDQLFDO(QJLQHHUV 
VWUDLQLVDERXW

NISSAN TECHNICAL REVIEW No.88 (2022) 129
࣭ ࢬ࣮࣑ࣥࢢゎᯒ⤖ᯝ࡜◚⿣ヨ㦂⤖ᯝ࡜ࡢẚ㍑

>'2,WUDQVMVPH@ ‹7KH-DSDQ6RFLHW\RI0HFKDQLFDO(QJLQHHUV 
 

Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis
7DNHPRWRDQG<RVKLNDZD7UDQVDFWLRQVRIWKH-60( LQ-DSDQHVH 9RO1R 
E 7DQN1R

 

D 7DQN1R F 7DQN1R




E 7DQN1R G 7DQN1R
)LJ3UHVVXUHVWUDLQFXUYHVRIDQDO\VHVDQGH[SHULPHQWVIRUWKHWDQNVDUHVKRZQ7KH]RRPLQJDQDO\VHVSUHGLFWWKHEXUVWSUHVVXUH
Fig. 20 Pressure-strain curves of analyses and experiments for the tanks are shown. The zooming analyses predict the burst pressure
precisely. The analyses yield higher strain than axisymmetricSUHFLVHO\7KHDQDO\VHV\LHOGKLJKHUVWUDLQWKDQD[LV\PPHWULFDQDO\VHVIRUDOOWKHWDQNVEHFDXVHRISUHFLVHHYDOXDWLRQRIORFDO
analyses for all the tanks because of precise evaluation of local strain
enhancement caused by the mesoscopic structure of fiber bundles and resin. The strains of zooming analyses exceed the
experimental one at tank No.1 and 2, since effect of fiber bundle crimp in helical layer cannot be measured
>'2,WUDQVMVPH@ by strain gage. The strains
‹7KH-DSDQ6RFLHW\RI0HFKDQLFDO(QJLQHHUV 
of zooming analyses become closer to those of experimental measurements at tank No.3 and 4 since the strain enhancement caused
by nonlinear behavior of resin is taken into account in zooming analyses.


measurement result. For Tanks No. 3 and No. 4, whose
4.5 Comparison of zooming analysis and burst test outermost layers are the hoop layers, the measured fiber-
results
F 7DQN1R direction strains of the hoop layer are almost identical to
The relationship between the strain and internal the maximum strain values obtained from the zooming
pressure of the carbon fiber bundle obtained via zooming analysis. As there is no crossing of the fiber bundles in
analysis using the meso-scale model is presented in Figs. the hoop layer and the nonlinear behavior of the resin
20 (a), (b), (c), and (d) for Tanks No. 1, 2, 3, and 4, that occurs prior to fiber fracture is accurately expressed
respectively. The vertical and horizontal axes represent in the zooming analysis, the zooming analysis results are
the fiber-direction strain of the carbon fiber bundle and closer to the measurement results than the axisymmetric
the pressure inside the tank, respectively. The maximum analysis results.
values
 evaluated for each layer are indicated by the solid The burst pressures predicted by the zooming analysis
blue and red lines for the helical and hoop layers, were 10.8, 26.8, 26.8, and 36.8 MPa for Tanks No. 1, 2, 3,
respectively. The green squares indicate the fiber-
G 7DQN1R and 4, respectively, whereas the actual burst pressures
direction strain in the outermost CFRP layer measured obtained from the burst test were 12.0, 26.3, 24.8, and
3UHVVXUHVWUDLQFXUYHVRIDQDO\VHVDQGH[SHULPHQWVIRUWKHWDQNVDUHVKRZQ7KH]RRPLQJDQDO\VHVSUHGLFWWKHEXUVWSUHVVXUH

during the burst test, the blue dashed line represents the 35.3 MPa, respectively, indicating a difference of less
SUHFLVHO\7KHDQDO\VHV\LHOGKLJKHUVWUDLQWKDQD[LV\PPHWULFDQDO\VHVIRUDOOWKHWDQNVEHFDXVHRISUHFLVHHYDOXDWLRQRIORFDO

maximum value of the carbon fiber-direction

2,WUDQVMVPH@ strain in the
‹7KH-DSDQ6RFLHW\RI0HFKDQLFDO(QJLQHHUV than
10%. Thus, the zooming analysis achieved a
helical layer evaluated using the axisymmetric continuum significant improvement in prediction accuracy, as
model, and the red dashed line denotes the maximum compared to the axisymmetric continuum model
strain in the hoop layer. described in Section 3; the layers that triggered the
As described in Section 3.3, the values measured by bursts, as identified by the zooming analysis, were also
the strain gauges can be regarded as the average values consistent with the burst test results. Thus, the proposed
for the gauge-mounted areas; in the low strain area, method was validated.
these values agree well with the results of the
axisymmetric continuum model analysis. The maximum 5. Conclusion
strain obtained from the zooming analysis exceeds that
determined by the axisymmetric analysis, accurately The proposed method based on a meso-scale model
expressing the influence of the crossing of the fiber that divides the CFRP layer into carbon fiber bundles
bundles and the nonlinear deformation behavior of the and the matrix resin is effective for evaluating the burst
resin. For Tanks No. 1 and No. 2, whose outermost layers pressures of high-pressure hydrogen tanks manufactured
are the helical layers, the local strain increase caused by using the FW method. In constructing the meso-scale
the crossing of fiber bundles is remarkable. It is impossible model, the following five factors were selected to build the
to measure this local strain accurately by using strain strength model:
gauges, and the maximum strain in the helical layer 1) Dependence of resin tensile strength on strain the
obtained from the zooming analysis exceeds the gauge rate

130 No.88 (2022) NISSAN TECHNICAL REVIEW

 Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis

2) Tension/compression asymmetry of resin strength Optimization of dome shape for type3 accumulators,
3) Anisotropic fracture law of carbon fiber bundle Journal of High Pressure Institute of Japan, Vol. 55,
4) Dependence of resin rigidity on strain rate No.1 (2017), pp.3-11.
5) Dependence of resin compressive strength on strain Littell, J. D., Ruggeri, C. R., Goldberg, R. K., Roberts, G.
rate D., Arnold, W. A., and Binienda, W. K., Measurement
For the demonstration, four types of Type 3 small of epoxy resin tension, compression and shear stress-
pressure tanks with different CFRP layer configurations strain curves over a wide range using small test
were prototyped and subjected to burst tests. Results of specimens, Journal of Aerospace Engineering,
the burst pressure test predicted based on the Vol.21, No.3 (2008), pp.162-173.
conventional axisymmetric continuum model were Liu, J., Yamada, K., Hosomi, M., Ikeda, T., Hirano, T.,
approximately 30% higher than the actual results, even Structural design and analysis of fiber reinforced
when the layer structures were faithfully modeled; this plastic pressure vessels with load-carrying metallic
indicates the overestimation by the model. Meanwhile, liners, Transactions of the Japan Society of
the zooming analysis could accurately predict the burst Mechanical Engineers, Series A, Vol.63, No.612
pressures by representing a part of the central cylinder of (1997), pp.1758-1765.
the tank using a plate-shaped meso-scale model and by LSTC, LS-DYNA R8.0 keyword user's manual, Vol.2
setting the displacement field obtained from the (2015), pp.127-129.
axisymmetric continuum model analysis as the boundary Maki H, Shimamura S, Tabei K, and Oishi S, Filament
condition. Accurate predictions of the burst pressures by winding, The Nikkan Kogyo Shimbun Ltd. (1970).
using this approach are expected to help manufacturers Onaka, H., Development status of fuel cell, Journal of
in designing more reliable and economical products, High Temperature Society, Vol.35, No.5 (2009),
without being overly conservative. pp.231-238.
In this research, the zooming analysis was focused on Takehana, T., and Ota, H., Estimation of safety and
the cylindrical part of the tank; nevertheless, we believe standard for fiber reinforced plastic composite
that the strength model presented herein is a realistic cylinders, Journal of High Pressure Institute of
model that includes the tension and compression Japan, Vol.35, No.1 (1997), pp.15-20.
asymmetry. Therefore, we will continue to verify this Takemoto, T., and Yoshikawa, N., Advanced nonlinear
methodology by applying it to the CFRP layer in the material model of resin for meso-scale tensile
dome part of the tank, where complicated bending strength evaluation of CFRP structure of high
deformation occurs. pressure hydrogen tank, Transactions of the JSME,
Vol.85, No.869 (2019), DOI:10.1299/
References transjsme.18-00304.
Arai, M., Kaku, K., Fushimi, Y., and Matsuda, T., The Council for a Strategy for Hydrogen and Fuel Cells,
Evaluation for viscoelastic constitutive equation of Strategic Roadmap for Hydrogen and Fuel Cells,
CFRP laminates using homogenization method, 16th Meeting of the Council for a Strategy for
Journal of Japan Society for Computational Methods Hydrogen and Fuel Cells, METI (2019).
in Engineering, Vol.8, No.10-081128 (2008). United Nations, Global Technical Regulation on
Arai, Y., Otsuka, T., Suzuki, H., Shibata, S., Obuchi, T., Hydrogen and Fuel Cell Vehicle, Global Technical
and Iijima, T. Evaluation of fatigue fracture Regulations No.13 (2013).
mechanisms of Al alloy liner/FW-CFRP composite Vasiliev, V., and Morozov, V., Mechanics and Analysis of
pressure vessel, Transactions of the Japan Society of Composite Materials, Elsevier (2001), pp.236-237.
Mechanical Engineers, Series A, Vol.73, No.736
(2007), pp.1388-1394. Source
Dietenberger, M., Buyuk, M., and Kan, C. D., Development
of a high strain rate development vehicle model,
LS-DYNA Anwenderforum Bamberg 2005 (2005). Vol.86,No.8832
High Pressure Gas Safety Association, KHKS0220
Standards for ultra high pressure gas equipment
(2016).
Hirase, Y., Mori, T., Ishihara, T., Katoh, A., and
Kawahara, M., An investigation of stress analysis of
a FRP vessel by using the netting theory, laminar
theory, and FEM, Transactions of the Japan Society
of Mechanical Engineers, Series A, Vol.57, No.535
(1991), pp.550-555.
Iida, H., Mechanical characteristics of filament-wound
pressure vessel, Journal of the Japan Society for
Composite Materials, Vol.4, No.4 (1978),
pp.163-167.
Kim, S. W., Yoshikawa, N., Tojo, S., and Suzuki, J.,

NISSAN TECHNICAL REVIEW No.88 (2022) 131

 Technical Awards 2020 JSME Medal for Outstanding Paper - Strength evaluation of CFRP tank based on meso-scopic zooming analysis

Recipient

Shinichiro Takemoto

132 No.88 (2022) NISSAN TECHNICAL REVIEW

 Editorial Postscript

Thank you for regularly reading the Nissan Technical Review. Globally, there are large transitions
toward meeting the carbon neutrality goals of 2050; therefore, we highlight two topics that represent the
major transformations undertaken by Nissan as special features in No. 88. The first topic is on ARIYA,
Nissan’s next-generation EV, and the second is on the Nissan Intelligent Factory, which produces future
vehicles that have undergone significant evolutions triggered by ARIYA’s production.
The Nissan Intelligent Factory aims to achieve carbon-neutral production as well as allow flexibility and
strength for handling various environmental changes. This is an important topic along with promotion of
the widespread use of EVs. The other topic, ARIYA, is the next-generation EV model of Nissan, who
launched their first EV LEAF in December 2010. This special feature describes the renovated technologies
that we have confidence in, including the new EV-specific platform, newly developed powertrain and four-
wheeler chassis control technology, and vehicle interface and connectedness.
The contributors to these articles are experts working at Nissan in fields ranging from technological
development to production. I would like to take this opportunity to thank all the contributors and editors
who wish to make Nissan’s technologies known widely.
I sincerely hope that the advanced technologies developed in accordance with “Nissan Ambition 2030,”
the long-term vision announced by Nissan in November 2021, will be accessible to our readers.

Tomohiro Yamamura,
Research Planning Department,
Nissan Research Center

NISSAN TECHINICAL REVIEW 2022 No.88

Published September,2022

Publishing office NISSAN MOTOR CO., LTD.

Research Division Research Planning Department
Address: 1-1, Morinosatoaoyama, Atsugi-shi, Kanagawa 243-0123

Publisher Research Division Research Planning Department

General Manager TOMOHIRO YAMAMURA

Editorial office NISSAN CREATIVE SERVICES CO., LTD.

Development Information Service Division
560-2 Okatsu.Koku Atsugi Kanagawa 243-0192 C/O Nissan Technical Center
Nissan Motor Co., Ltd.