Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Design of Racing and

High-Performance Engines
2004-2013
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Other SAE books of interest:

We Were The Ramchargers

By David Rockwell
(Product Code: R-384)

Engine Failure Analysis

By Ernst Greuter and Stefan Zima
(Product Code: R-320)

Engine Design Concepts for World Championship Grand Prix Motorcycles

By Alberto Boretti
(Product Code: PT-155)

For more information or to order a book, contact SAE International at

400 Commonwealth Drive, Warrendale, PA 15096-0001, USA

phone 877-606-7323 (U.S. and Canada) or 724-776-4970 (outside U.S. and Canada);
fax 724-776-0790; e-mail CustomerService@[Link]; website [Link]
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Design of Racing and

High-Performance Engines
2004-2013
Edited by Douglas R. Fehan

Warrendale, Pennsylvania, USA

Copyright © 2013 SAE International. eISBN 978-0-7680-7986-9

 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

400 Commonwealth Drive

Warrendale, PA 15096-0001 USA

E-mail: CustomerService@[Link]
Phone: 877-606-7323 (inside USA and Canada)
724-776-4970 (outside USA)
Fax: 724-776-0790

Copyright © 2013 SAE International. All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system, distributed, or transmitted, in
any form or by any means without the prior written permission of SAE International. For permission and
licensing requests, contact SAE Permissions, 400 Commonwealth Drive, Warrendale, PA 15096-0001 USA;
e-mail: copyright@[Link]; phone: 724-772-4028; fax : 724-772-9765.

ISBN 978-0-7680-7983-8
Library of Congress Catalog Number 2012955673
SAE Order Number PT-157
DOI 10.4271/PT-157

Information contained in this work has been obtained by SAE International from sources believed to be
reliable. However, neither SAE International nor its authors guarantee the accuracy or completeness of
any information published herein and neither SAE International nor its authors shall be responsible for
any errors, omissions, or damages arising out of use of this information. This work is published with the
understanding that SAE International and its authors are supplying information, but are not attempting
to render engineering or other professional services. If such services are required, the assistance of an
appropriate professional should be sought.

To purchase bulk quantities, please contact

SAE Customer Service
e-mail: CustomerService@[Link]
phone: 877-606-7323 (inside USA and Canada)
724-776-4970 (outside USA)
fax: 724-776-0790

Visit the SAE Bookstore at

[Link]
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Table of Contents
Introduction
Douglas R. Fehan………………………………………………………………………………………… vii

List of Papers

Motorsports Industry Knowledge Exchange (MiKE): Oxymoron or Holy Grail

(An Imperative for Sustaining Regional, National and Global Competitiveness) (2006-01-3610)
Mike Meechan…………………………………………………………………………………………… 1

Hardware-in-the-Loop Testing in Racing Applications (2004-01-3502)

Peter Waeltermann, Thomas Michalsky, Johannes Held…………………………………………… 19

Coolant Flow Optimization in a Racing Cylinder Block and Head Using CFD Analysis
and Testing (2004-01-3542)
Jian Ye, Jim Covey, Daniel D. Agnew………………………………………………………………… 33

Improvement of a High-Performance CNG-Engine Based on an Innovative Virtual

Development Process (2011-24-0140)
Marco Chiodi, Alessandro Ferrari, Oliver Mack, Michael Bargende, Donatus Wichelhaus………… 43

The Application of Energy-Based Fuel Formulae to Increase the Efficiency Relevance

and Reduce the CO2 Emissions of Motor Sport (2008-01-2953)
J.W.G. Turner, R.J. Pearson…………………………………………………………………………… 57

High Power Density Motor for Racing Use (2011-39-7221)

Tamotsu Kawamura, Hirofumi Atarashi, Takehiro Miyoshi…………………………………………… 69

The Development of a Low Viscosity, Highly Efficient Lubricant for Sport Motorcycle
Applications (2011-32-0513)
Gianluigi Zoli, Matthew Symonds, May Turner, Ieuan Adams, Nick Solomon,
Mark Leonard, Cliff Newman………………………………………………………………………… 75

An Improvement of a Small Displacement Engine’s Efficiency with a Super Charging

System (2011-32-0571)
Changjoo Ahn, Takashi Suzuki, Yasuhumi Oguri, Hiroki Toshitani, Tatsuyoshi Nakahuku,
Yusuke Nakano………………………………………………………………………………………… 81

The Effects of Intake Plenum Volume on the Performance of a Small Normally Aspirated
Restricted Engine (2008-01-3007)
L.J. Hamilton, J.E. Lee………………………………………………………………………………… 89

Design and Development of a Turbocharged E85 Engine for Formula SAE Racing (2008-01-1774)
Jason P. Moscetti, Bryan Gilroy Smith, Volker Sick…………………………………………………… 97

v
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Optimizing the Design of the Air Flow Orifice or Restrictor for Race Car Applications
(2007-01-3553)
Harry C. Watson, Andrew Gauci, Faeez Yousuff, Alberto Boretti…………………………………… 103

A Theoretical and Experimental Study of Resonance in a High Performance Engine

Intake System: Part I (2006-01-3653)
S. Brennan, R.J. Kee, R.G. Kenny, R. Fleck, J.A. Gaynor, B. Fleck……………………………………… 111

Horsepower Retention by ISF (Isotropic Superfinishing) of Automotive Racing Components

(2004-01-3511)
William P. Nebiolo…………………………………………………………………………………… 125

About the Editor……………………………………………………………………………………… 137

vi
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Introduction

I sincerely believe we have entered the most exciting era in automotive engineering
since the advent of the assembly line. Now, more than ever before, real science and
mathematics have assumed a very essential position in the design and development
of our vehicles and specifically their associated power plants. Universities have joined
with design firms and the automotive manufacturers themselves to deal with the
rapidly advancing technologies that will create our engines of the future. As the
following compilation of papers will demonstrate, racing engines are one of the
primary tools being utilized to speed development. I think it will be clear that these
papers reflect my career spent in motorsports. To that end, I hope they will serve to
inspire engineers of all generations to embrace racing engine development and view it
as a very important tool in the future of transportation. The classifications listed below
are important areas of race engine design and are representative of the papers found
in this publication.

History
Interestingly, for over 100 years, the primary source of propulsion has been limited
to petroleum-based fuels combined with the internal combustion engine. And, until
recently, even the basic concept of the vehicles themselves has really not changed that
much. Now, a new day has dawned and racing and racing engines are playing an even
more vital role in that transition.

Electrification
When we look at today’s hybrid electric vehicles and KERS (kinetic energy recovery
system) systems, we are actually seeing what inventors envisioned in the early 1900s.
First employed in trams and trains of that era, the technology was, for the most part,
almost forgotten until relatively recent years. Racers resurrected their version in 2009
in F-1 racing.

Materials/Processes
The automotive industry has long admired the aircraft industry’s utilization of
lightweight metals, advanced finishing processes, and composites. The use of these
materials and processes has served to reduce overall mass and in turn vastly improve
speed, performance, and reliability of race engines. Their initial high cost was a limiting
factor in being able to integrate them into mass produced vehicles. With racing
leading the way, those limitations have been overcome, and vehicles today feature
some amazing adaptations of those processes and materials.

vii
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Race Engine Technology: The Pathway to the Future

Engine power, efficiency, durability, reliability, and more recently emissions have
never ceased being of primary importance to the automotive world. The use of
electrification, bio-fuels, compressed natural gas, high-pressure fuel delivery systems,
combustion air management, turbocharging, supercharging, and low-viscosity
lubricants has been expanding in the race engine development universe, and those
technologies are now finding their way into dealer showrooms.

Conclusion
As engineers develop more new and exciting materials and processes, in some cases
using very advanced mathematics, the production world will continue to include
them. It should not be forgotten that there are also good examples of production-
based advances within the racing industry.

The papers contained in this publication were selected for two reasons. One, because
they clearly demonstrate the leadership role the racing industry plays in the future
of automotive engineering and design as it relates to engines. And two, I believe that
they will be of interest to everyone who may currently be involved in racing and to all
those who may want to plan a future within the racing industry.

My primary objective is that each of you has as much fun and interest reading these
papers as I did in assembling them. They represent some truly great work by your
fellow engineers and hopefully will serve to inspire even more impressive work by the
SAE International membership. As the adage states…the future is now!

Douglas Fehan
Corvette Racing Program Manager

viii
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

2006-01-3610

Motorsports Industry Knowledge Exchange (MiKE):

Oxymoron or Holy Grail (An Imperative for Sustaining
Regional, National and Global Competitiveness)
Mike Meechan
Oxford Brookes University, England UK

Copyright © 2006 SAE International

ABSTRACT The Objective of this paper is to:

Of all high-performance engineering industries, x identify key enablers and inhibiters, perceived and
motorsports perhaps exemplifies best the unique actual, to successful motorsports industry-academe
combination of key engineering and business elements interaction;
vital for swift, industry-focused, successful, high- x investigate inhibiters, to offer theoretical and
technology product advancement. This heady mix practical insights into the reasons why key
includes: complex market analysis; sophisticated stakeholders in the process believe such barriers
research, design and development capability; rapid exist, indeed persist, and in some ways are
product innovation, prototyping and development; just- perpetuated;
in-time manufacturing using state-of-the-art processes; x provide instances of successes and failures in
and regular mandated exhibition of company industry-academe technology and knowledge
competitiveness, involving (at the highest environs) transfer;
demonstration of both personnel and new product x detail a number of mechanisms to aid the catalyzing
capability and reliability, on an unforgiving world stage, of vital industry-academe technology and knowledge
invariably to ensure continuing investor (sponsor) transfer.
confidence. Such examples of motorsport’s disparate
business model elements in many ways demonstrate INTRODUCTION
fundamentally the industry’s absolute reliance on what in
corporate speak is now termed ‘brains trust’ or ‘human OVERVIEW
capital’.
The terms ‘motorsports’ and ‘knowledge’ are bound
Invariably, however, funding for academic faculties inextricably, and their presence in the same compound
(certainly in the UK) is now predicated largely on noun yields a relationship that is easy to discern,
speculative - and in many respects unrealistic – future appreciate, understand, and empathize with. The
student intake numbers, driven by technological pursuit of ‘faster’, ‘stronger’, ‘safer’,
institution/Government recruitment targets, which are ‘lighter’, ‘more enduring’ and recently, ‘cheaper’ (a term
rigorously-policed. The corollary is that institutions can that some years ago would perhaps have been regarded
now no longer simply be considered as seats of learning as heresy), all in a quest – regularly played out on a
– instead, they increasingly reside, sometimes world stage, before a global weekly audience
uncomfortably, in the aggressive business domain of numbering hundreds of millions – for domination in a
skills, education, training and research provision, where variety of different competitions, formats and formulae,
the marketplace is open, competitive and formidably- requires the very best brains, talent, commitment and a
discerning, and majors on informed, industry-driven certain je ne sais quoi to create robust, state-of-the-art
service provision. Thus, it is surprising, but perhaps technological solutions.
understandable, that the arranged marriage of the
academia-based skills/expertise purveyor, and the Add one word - or concept - to the above, however, and
ostensibly-eager industry recipient of such services, the status quo changes. For many, the extended
remains largely unconsummated. combination of terms - ‘motorsports’, ‘knowledge’, and
‘exchange’ - presents an altogether less easy alliance, to

1
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

the extent now that learned motorsports commentators impediments to - knowledge and technology transfer
have described motorsports knowledge exchange as an (with the premise that solutions are intractable unless
oxymoron 1 (Bamber, 2005; Anon, 2005). This paper the problem is properly understood).
roundly challenges this assertion.

METHODOLOGY OF THE PAPER THE UK MOTORSPORTS INDUSTRY

In this paper, I look first at the motorsports industry, and Centred on the Oxfordshire/Northamptonshire county
compare and contrast the manner in which its various border, and subtending a 100-mile or thereabouts
global communities actually do share knowledge, and in crescent or arc that stretches from Derby to
doing so, far from losing competitive edge, actually Southampton (an area that has been termed ‘Motorsport
consolidate, strengthen and improve stakeholder Valley’), the UK motorsports industry presents some
advantage. (I use the term ‘stakeholder’ here because it fascinating statistics:
is not just companies – industry - which benefit from
knowledge exchange, but all within the so-called x It employs around 70,000 people, full- and part-time
‘knowledge community’ who participate in the process: x 38,500 of its employees are engineers
nations, regions, Government, industry and academia, x Seven of the 11 Formula 1 teams reside in the UK,
amongst a host of others.) six within a 30-mile radius of Oxford, and the other
within a 60-mile radius
Secondly, based largely on academic and other x World-class automotive/motorsport consultancy
documentary evidence (and if this appears somewhat companies such as Menard, MIRA, Prodrive and
sparse, it is because worthwhile papers seem authored Xtrac all reside centrally within Motorsport Valley
by a small number of the same very-highly-regarded x It comprises some 4000 companies, 90% of which
sources, who look predominantly only at the UK are ‘SMEs’ (Small-to-Medium-sized Enterprises), the
motorsports industry, and significantly cross-reference), majority of which are ‘small’ (5-50 employees) or
the paper investigates some of the mechanisms by ‘micro’ (less than five employees)
which academia and industry – and industry and x It contributes around $US9billion to the UK economy
industry, i.e. through inter-company relationships – (more than 1% of GDP, which is around four times
knowledge exchange and technology transfer. greater than the widely-acclaimed, respected, and
lauded British music industry)
Next, based on real, practical experience gathered over
x It spends around $US1billion on R&D, which as a
the duration of the Motorsports Knowledge Exchange
percentage of turnover is much greater than the UK
project which I manage, (and the 3-year period
national average for the engineering sector
immediately preceding this), this paper presents a
number of perceived impediments to successful
Henry, Angus, Jenkins et al combine and summarize the
academia-academia, academia-industry, and industry-
above statistics well in their statement: ’the UK
industry knowledge exchange and technology transfer.
motorsport cluster (‘Motorsport Valley’)… can be
unambiguously rated as world-class and dominant…
Finally, the paper provides a number of scenarios and
The strength, scale and diversity of the industry are not
suggestions in an effort to avoid some of the
yet replicated elsewhere in the world… There is no
impediments and barriers to knowledge exchange and
doubt that Motorsport Valley is not only a functioning
technology transfer that have been encountered in
cluster but a leading international example [which]… sits
practice thus far.
at the pinnacle of an expanding global motorsport
production system’. 2
THE THEORY, PRACTICE, PROBLEMS AND
BENEFITS OF – AND FUTURE STRATEGY FOR - Thus, it is not difficult to contend that the UK motorsports
KNOWLEDGE EXCHANGE AND TECHNOLOGY industry’s credentials as a candidate worthy of real
TRANSFER IN THE MOTORSPORTS INDUSTRY scrutiny, to determine what it is that actually makes this
element of the county’s high performance engineering
sector so competitive globally - particularly through
INTRODUCTION AND OVERVIEW knowledge exchange and technology transfer – are
impeccable. (That this should be so is evidenced by the
In this section, I provide first a brief overview of the UK number of research endeavors – academic,
motorsport industry and thus establish its credentials as Government, and other - which have looked closely at
world-class and worthy of real scrutiny, if one is to the UK motorsports industry to investigate and
understand that industrial competitiveness in the 21st understand how and why – firstly - it became so globally
century is rooted upon the transfer of technology and the competitive (in many ways, ‘punching well above its
exchange of knowledge. Next, I look at the academic weight’ as some commentators have commented) and –
conventions of the constituent parts of our ‘motorsport’, secondly - how it can sustain and improve its
‘knowledge’ and ‘exchange’, ‘equation’, to understand competitiveness, despite fierce global onslaught.
better some of the deep-rooted prejudices about - and

2
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

DEFINING ‘MOTORSPORTS’ AND ‘KNOWLEDGE’ We move now to consider ‘knowledge’. Blackler 8 (1995)
codified [knowledge] as types that are ‘embrained,
At this point, and as noted in the introduction, I feel it is embodied, encultured, embedded and encoded.
important to characterize – and define – in a formal
(academic) sense the various elements of the equation x ‘Embrained knowledge’ – ‘knowledge that’, which is
that constitute motorsports knowledge exchange (MKE). dependant on conceptual skills and cognitive
Such characterization – in many ways, a deconstruction, abilities.
or differentiation (in the mathematical sense) – of the x ‘Embodied knowledge’ – ‘knowledge how’, which is
various elements of the ‘MKE’ equation makes for far action orientated and is likely to be only partly
easier investigation of exactly where failures exist. This explicit.
characterization also allows examination of the reasons x ‘Encultured knowledge’ – the process of achieving
why such failures happen, what causes such failures to shared understandings.
persist, and the manner in which such failures perhaps x ‘Embedded knowledge’ – resides in systematic
can be remedied by intervention. routines in the relationship between technologies,
roles, formal procedures and emergent routines.
What actually constitutes ‘motorsports’, as far as a x ‘Encoded knowledge’ – conveyed by signs and
‘dictionary’ or ‘academic’ definition is concerned? symbols.’
Different commentators naturally describe it in various -
albeit (generally) complementary - ways. As The But how does this relate to the argument about
Financial Times (1990) notes, ‘motorsport[s] is a Motorsport Knowledge Exchange. Marotti 9 states,
disarmingly simple term for what has become, ‘knowledge is not primarily about data and information,
particularly through the 80s, a multi-facetted but about socialization processes… Recognizing
international industry’. 3 knowledge as a dynamic and social accomplishment has
important implications for how knowledge sharing and
Henry, Angus, Jenkins et al characterize what they term capabilities can be conceived… Knowledge creation is
‘the motorsports cluster value chain’ as comprising 10 by definition a dynamic process of interaction and one
key elements: ‘constructors; constructor suppliers; that involves dissimilarly-endowed partners.’ Thus,
entrants; events; event suppliers; media; consumption; within the Motorsports Cluster, we have all of the
supporting service industries; regulation and the ingredients and requirements for knowledge exchange.
regulatory and fiscal environment for business’. 4 These
elements reside in what the authors term ‘a complex Similarly, Henry and Pinch 10 note that, ‘Motorsport
system of interdependence’. Kirkpatrick and Faulkner Valley can best be conceptualized as a knowledge
(2003) are somewhat more poetic: ‘... rather like beauty community; a socially and spatially embedded economic
and quality, …[performance engineering] appears that it system which has developed the capacity to generate
is in the gift of personal perspective… [although] there and rapidly disseminate knowledge about superior ways
were common features… focus, time and timing of of designing and manufacturing racing cars’.
solution… creativity, innovation, quality, limiting
compromise, [and] a competitive holistic application of THE ROLE OF KNOWLEDGE EXCHANGE IN
engineering knowledge and skills.’ 5 CLUSTER DEVELOPMENT
Others (Stancombe, 2002) look to positioning the cluster Overview
within a wider sector. ‘[Whilst motorsports] originally
perceived position, [was] as a sub-sector of the Whilst a number of authorities 11 DTI (2001) have looked
Automotive industry… [recent research] places it at the to the clustering effect of the UK’s motorsports and other
centre of a new community of advanced engineering and industries, and the impact of this on its competitiveness,
services industries… In this position, the [motorsports] it is perhaps worth recapping upon what clustering
industry will act as a stimulating business catalyst (formally) constitutes, before looking formally then at
between the other industries, increasing collaboration, what ‘knowledge’ constitutes, and the argument about
and transferring competitive advantages in high added how and why knowledge exchange contributes to cluster
value, knowledge based development; teamwork; rapid growth and competitiveness, and vice versa.
innovation; technology and best practice, across all
sectors’. 6 Introduction
Also, as Kirkpatrick and Faulkner (Ibid, 2003) note,
The growth of the UK motorsports industry, from World
‘there are three vital components to an understanding of
War II onwards, and its dominance since the 1960s
the [UK] Motorsport Industry, as it exists today:
(when it overtook Italy in world competitiveness) are well
x ‘Understanding the overall ‘cluster’ concept and the documented. 12 (Pinch, N., 1999). I look here to
dynamics of the Motorsport Valley ‘cluster’ understand and describe the knowledge exchange and
x ‘The relationship between the sport and the industry technology transfer mechanisms which have been
x ‘That the industry is made up, almost exclusively, of instrumental in this growth and dominance. The
small to medium sized enterprises (SMEs)’ 7 following list details the ‘Criteria for Success in Cluster

3
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Development’, in descending order of importance, Likewise, ECOTEC 17 , in a report compiled on behalf of

according to the literature analysis of the cited reference: the UK Department of Trade and Industry, postulates the
following:

‘A number of common factors can be identified that

underpin successful clusters. These provide a clear
guide to the relative importance of different factors.
There is a strong degree of consistency in these results:

x ‘The ‘softer’ aspects of cluster development

particularly human and social capital, networking
and partnership are the dominant factors in the
development of successful:
o ‘Human and social capital cover a broad range
of factors such as skills, availability, trust, inter-
relationships and a willingness to co-operate.
o ‘Partnerships and networks (formal and
informal) supporting collaboration and co-
Source - ‘A Second Draft Report for the DTI – Success operation act as vehicles for developing the
Criteria for Success in Cluster Development’ 13 vitality and strength of firms within the cluster
x ‘Innovation, technology, product development and
It would be remise of any paper purporting to understand intellectual capital are key elements in the
the motorsports industry to ignore ‘cluster theory’ and its development of successful clusters, following the
impact on the industry’s competitiveness. Porter 14 softer dimensions identified above in importance.
(1990) defines a cluster as ‘a geographically proximate x ‘The presence of large firms can be a strong force
group of interconnected companies and associate driving cluster development, although the integration
institutions in a particular field, linked by commonalities into the local economy can sometimes be
and complementarities’. OECD 15 (2001) expands further challenging.
the definition: ‘clusters can be characterized as being x ‘Adequate physical infrastructure, relating to both
networks of production’. It is evident from this definition, sites and premises and accessibility, is an important
and several of the elements cited within the paper – requirement for cluster development
‘networks’, ‘strongly interdependent firms’, ‘knowledge x ‘A supportive policy environment is a positive factor
producing agents’ and ‘bridging institutions’, that in cluster development
‘clustering’, and its beneficial impact on industry’s ability x ‘Traditional business support functions are less often
to begin then sustain competitiveness, is predicated on identified as critical factors for the development of
knowledge exchange and technology transfer, and so for successful clusters’.
the motorsport industry in the UK to attain its present
global dominance. The interdependence of ‘Human or Social Capital’ and
‘Networking’ also has been noted. As Ecotec (Ibid)
The concept of ‘competitiveness’, and the inextricable describe, ‘social capital… is a key aspect of successful
link between competitiveness and clusters, also is partnership and networking between people,
important. As Porter and Ketels 16 (2003) note: organizations and institutions. Trust, inter-relationships
and a willingness to co-operate have all been found to
‘Competitiveness remains a concept that is not well have a strong positive impact on cluster development.
understood, despite widespread acceptance of its For example:
importance. To understand competitiveness, the starting
point must be the sources of a nation’s prosperity. A x Personal and professional relationships,
nation’s standard of living is determined by the interpersonal trust and reputation, [and] access to
productivity of its economy, which is measured by the reliable information, all reduce uncertainty on
value of goods and services produced per unit of the projects (Database Human Capital 2002).
nation’s human, capital and natural resources. x Trust in group processes, sharing of information and
Productivity depends both on the value of a nation’s openness of communications are critical to form
products and services, measured by the prices they can working clusters. Investing in the mechanisms that
command in open markets, and the efficiency with which promote the invisible parts of the community and
they can be produced. True competitiveness then, is integration are important. (Database Human Capital
measured by productivity. Productivity allows a nation to 2002).
support high wages, a strong currency and attractive x Firms, or people, should be encouraged to
returns to capital, and with them a high standard of appreciate and develop the willingness to share
living.’ information with competitors. Furthermore they
should avoid the “what’s in it for me” culture and

4
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

instead focus on what’s in it for me and the region often see no good reason (or any reason) to work with
(Practitioners Observations 2002). academic institutions that they view as competitors. This
situation prevails at a number of academic levels but is
Research has shown that cluster failure or decline perhaps most prevalent at the undergraduate level,
happens predominantly because of a number of where remuneration per student is lowest, but student
happenstances. Ecotec 18 (Ibid) reported the following numbers greatest; it is also often the area where
key reasons: greatest institutional reputations are to be made.

x poor networks At post-graduate level, remuneration-per-student is

x a lack of high-level skills highest (particularly for overseas students) although
x low levels of R&D student numbers are lower (perhaps by an order of
x Lack of leadership and entrepreneurship magnitude) compared to undergraduate levels.
However, successful postgraduate provision depends
ECOtec (Ibid) also note that ‘poor links with educational largely on individual academic prowess in niche (and
institutions were alluded to quite frequency in the context often lucrative) markets (such as ‘business’); institutions
of cluster decline’, and reinforce this by stating that thus wont protect academic IP to retain valuable
‘…poor links lead to a lack of information transfer from market share and inevitably shun collaboration.
universities to industry’. (Practitioner 2002) (However, a worthy example of a hugely successful
inter-university collaboration is the Cambridge-MIT
KNOWLEDGE EXCHANGE AND TECHNOLOGY Institute (CMI) whose aim is ‘enhancing the
TRANSFER IN THE UK MOTORSPORT KNOWLEDGE competitiveness, productivity and entrepreneurship of
– WHEN AND WHY KNOWLEDGE DOESN’T GET the UK economy by improving university-industry
EXCHANGED IN THE MOTORSPORT INDUSTRY knowledge exchange’. 19

Overview Finally, there is a definite ‘pecking order’ where

collaboration is concerned. ‘Ivy League’ universities
traditionally ally with other such universities on a national
Having cited mechanisms and exemplars of good
or international basis (e.g. the MIT-Cambridge University
practice for knowledge exchange and technology
alliance); where such institutions collaborate with
transfer within (predominantly) the UK motorsports
universities other than those occupying the same
industry, I look now to some of the notable examples
‘league’, it is most definitely on a subordinate basis,
where Knowledge Exchange and Technology Transfer
which has some attractions but perhaps many more
have not happened and investigate the reasons for this.
disadvantages also.
The mismatches are classified under three headings:
Academia-Academia, Industry-Academia, and Industry-
Industry, although the first and last categories are only Industry-Academia
briefly summarized, again based on practical
experience. As Ecotec 20 note, ‘direct university links to businesses
are relatively weak in the UK. [This is in stark contrast to
Academia-academia the US, where, for example] MIT aggressively markets
the potential of research partnerships between
The academia-academia relationship, as this relates academics and industry’. The authors of this paper state
particularly to collaboration, is a complicated one with that this is ‘due to levels of funding and the structure of
many elements and a certain edginess and peculiar the RAE, which does not provide major incentives for
dynamic. HEIs to engage with business’. This is worthy of some
expansion.
Any collaboration – personal, institutional, business, a
combination of these – most likely occurs between The UK RAE (Research Assessment Exercise) concerns
individuals or entities (as noted above) having some kind itself fundamentally with gauging an institution’s
of shared interest, and common ground/agenda. ‘research’ standing and capability, on a regional,
Unfortunately, in an academic environment, such national and international basis, via scrutiny of ‘research’
reasons for collaborating also reside at the heart of the output – journal papers (scored according to
reasoning not to collaborate. Again we return to the international standing of the journal in which the paper is
situation of the migration of academic institutions from published), patents etc. – and research income.
being ‘seats of learning’ to now being ‘in the business of ‘Research’ – which invariably resides in closer proximity
providing skills, education and training’. Universities and to ‘pure’ and ‘blue skies’ than to ‘applied’ and ‘near-
colleges providing similar academic offerings - and thus market’ – is considered differently – and infinitely
having apparently real reason to work together for the superior – to consultancy, which invariably is market-led
common good – also are in open competition: and instigated through contracts with third parties
(normally industry). Thus, to achieve a high RAE rating
(and the high levels of commensurate research council
Universities are now (more visibly than in past, perhaps)
funding that such RAE ratings attract), universities are
businesses – and in some cases, big businesses – and
compelled to pursue ‘research’ rather than ‘consultancy’,

5
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

1
despite the fact that the latter invariably provides better the UK would term ‘bums on seats’) with majority effect
industry-academe collaboration and consequent market- from departments invariably residing in the domain of
driven and market-focused solutions. undergraduate recruitment. This pressure is felt acutely
in Science, Engineering and Technology-based
However, it is felt through practical experience (gained faculties, where costs per student head are much higher
particularly during work of the Oxford Brookes-led than, say, Arts and Humanities, but departments attract
Motorsports Knowledge Exchange project to date) that only a derisory increased amount of funding from
this is just one of a large number of reasons why the Government.
industry-academia interface/collaboration sometimes
fails; further reasons are detailed below: Parents, almost without exception, now accompany
prospective ‘freshers’ to university open days and
x Academia responsiveness perhaps are more inquisitive (and informed) than
x Lack of market focus/industrial relevance possible-future students about reputation, infrastructure,
x Poor industrial content of courses facilities, campus, teaching and research prowess,
x Risk aversion of HEIs pastoral care etc. of the visited institution; such vigilance
x HEIs are typically reactive rather than proactive perhaps has been exacerbated by the rapid ramp-up of
UK university fees, which means that many parents now
x Staff availability
provide wholly for, or subsidize largely, their offspring’s
x Staff interest in engaging with industry
university experience, and so have become far more
x Timetabling issues (impact of terms/semesters) discerning as a consequence. The relationship between
x Academic pace student and academic institution has changed
x Poor or non-marketing of expertise profoundly with top-up fees (in the UK), and a university
x Standard of equipment and facilities now has become an expensive service provider to a
x Low visibility of HEIs on industry ‘radar’ demanding and highly critical customer expectant of
x Pressures of race calendar exemplary quality.
x Industry staff too busy and too important to train or
upskill Academics thus are charged principally with pursuits
x Industry staff too busy and too important to advise pertaining to ‘Teaching’ and ‘Research’, rather than so-
academia called ‘third-stream’ activities.
x Motorsports industry demographics
x Form or delivery format of academic courses Lack of Market Focus/ Industry Relevance
x IP ownership conflicts
x IP leakage concerns There is understandable pressure, from both Higher
x Contract and other legal issues Education Institution (HEI) Quality Assurance and
x Cost of service provision – research, consultancy, Accreditation (QAA) committees and learned bodies, on
contract testing – to third parties demonstration of academic rigor from the university’s
degree-awarding body, so that courses can be fully
I look now at the individual impediments - as detailed in validated and accredited, and high quality assured; there
the above list - in greater detail. is also pressure now to include non-engineering but
nevertheless vital skills development in ‘soft’ and
Academia responsiveness ‘management’-style subjects. With an academic
timetable that has very definite and also finite time-
bound constraints, this means that other subjects,
This is related to the slowness (or complete absence) of
particularly those with high engineering practical
response of academia in relation to industry
emphasis, often shrink or disappear completely from
communication/ service request.
timetables. This situation is highly at odds with what the
motorsports industry requires of its new graduate or
Whilst the original deep-rooted purpose and principal
technician entrants: as Faulkner and Kirkpatrick (2003)
concept of universities were as ‘seats of learning’, and
note, ‘…with the speed of technology change, it is often
this situation has prevailed for many hundreds of years,
difficult to match training need to educational lead times.
an increasingly competitive, aggressive (and
As a result, a need may not be satisfied within the
demanding) marketplace means now that institutions are
required timescale, or to the appropriate level of
in the business of teaching. Levels of finance - for
qualification.’ 21
teaching purposes - placed into departments from higher
education funding councils are predicated directly on
Also, academia - typically - has proven in practice to be
numbers of students – predominantly at undergraduate
reactive rather than proactive to external factors and
level – with penalties imposed for both under- and over-
influences, which means that courses – even where
recruitment of students, as predicted/forecast in
these have evolved to meet a perceived industry need at
departmental strategic plans. Thus, there is huge
a particular point in time – are, due to course
emphasis placed on filling course places, (what we in
1
The author recognizes that this phrase has rather
different connotations to a North American audience!

6
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

development/QA/accreditation timeframes, out of date The attraction of such a system is perhaps greatest for
and lacking relevance by the time they are actually non-performing institutions, where attainment of
delivered (because the motorsport industry moves so kitemarking becomes a key aspiration; for others (high-
fast, and the skills churn is somewhat unpredictable). performing institutions) who are measured by industry
Lack of market emphasis coupled to a reduction in the directly and regularly on their ability to actually deliver,
hands-on, practical capability of new graduates does not the process becomes an unwelcome, bureaucratic, and
inspire the confidence of the motorsports industry. This worthless intrusion.
lack of confidence is exacerbated by the number of
institutions now offering (ostensibly) ‘motorsports’ Poor industrial content
courses, but where motorsports industry applicability is,
in many commentators’ opinions, quite frankly Much of this problem stems from Lack of Market Focus/
execrable. Industry Relevance, as noted previously, and can be
particularly relevant and applicable to those motorsports
This situation is in many respects directly related to the courses delivered by HEIs far removed – geographically
‘students-on-seats/quality of course’ dichotomy: people’s – from the motorsports cluster, where, because of
perception of ‘motorsports’ is as a sexy, attractive, global distance, such institutions will be severely
phenomenon. Putting ‘motorsports’ as a preface to any disadvantaged, finding it difficult (if not impossible) to
engineering course title virtually guarantees increased engage with the motorsports industry in any meaningful
student enrolment; where Mechanical Engineering way (particularly with respect to pursuits such as student
degree recruitment is on the decline, so the converse is placements or Knowledge Transfer Partnerships).
true of ‘motorsports’ courses, with the UK now offering
something like 22 different motorsports courses, where However, before considering the problem with a view to
10 years ago there were none. This increases both offering improvement or solution, this (geographic)
industry and student confusion, since it is sometimes element must be considered on a nation-by-nation (UK-
difficult to discriminate ‘wheat’ from ‘chaff’, other than by versus-US) basis:
quality of past graduates, and reputation; unfortunately,
prospective undergraduate entrants onto courses can be x The UK, as a total geographic land-mass, is not
served only by the second element, which can be dissimilar to that of a typical single state within the
impeded because of the ineffectiveness of some HEI US.
marketing departments. x Comparatively, (to the US) the UK motorsports
industry centers on a very small area as a
Whilst the Report of the DTI Motorsport Competitiveness percentage of the overall land-mass of the country.
Panel 22 , specifically the section derived from the input of x The US considers its concentration of expertise in
the Skills, Education and Training Working Group motorsports on a state-by-state basis; the UK does
recommended ‘signposting… a service… to influence so on a county-by-county basis, where England and
provision [and] raise standards’ - in essence, some form Wales comprise some 52 counties (with the
of course ‘kitemarking’ (at some cost to the institution of majority of motorsports activity confined to just two
such service) – it is the case that academic institutions of these 52 counties).
already are over-burdened by quality audits. These x What is local to the US is regional to the UK; what is
include those by their own (internal) QAA bodies, as well regional (state-wide) to the US is national to the UK.
as those of external academic and professional body
accrediting and course-ratifying authorities, all of whose Furthermore, whilst at the highest levels, collaborations
predations, while provided free of charge, invariably, are between academia-industry happen at an institution-
considered with extreme trepidation by the schools, company level, in reality, the old maxim - that business
departments or faculties under scrutiny. A further is done between individuals - applies equally well in the
(perhaps unnecessary) level of accreditation presents a academe-industry relationship. This dynamic presents a
number of fundamental hurdles to academic institutions: number of strengths and weaknesses:
x Cost – why should institutions bother with another x Key staff have certain tendency to migrate to new
audit that costs money and places heavy demands positions, either within or without their organizations:
on staff resource for its duration loss of personal contacts has profound implications
x Value – if a course is accredited for Chartered to continued successful academia–industry
Engineer status (Professional Engineer in the US), relationships
this in many ways is in itself a ‘gold standard’ and of x Industrial ‘gatekeepers’ within companies, whilst
itself evidence enough of the engineering value of usually senior, are not sufficiently senior either to
the course engender policy change, or to commit real resource
x Resource – accreditations mean the deployment of x Academic staff, whilst perhaps senior, are not
staff resource, both during the actual audit process, always sufficiently empowered to make key ad-hoc
and in the creation, management and provision of strategic decisions
the audit trails necessary to satisfy accrediting
x Academic institutions (certainly in the UK) work
bodies
apparently on the basis of short (1-2-year) strategic

7
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

plans, and Finance Directorates certainly expect HEIs are typically reactive rather than proactive
payback (financial equilibrium, or equilibrium plus
surplus) within such timeframes: the corollary to this HEIs (typically) develop courses in response to a market
institutional ‘short-termism’ is an aversion to commit need, and motorsports courses have evolved no
to programs - teaching, research, or other - which differently. However, the pace of the industry
take longer than the strategic plan period to deliver discriminates it from others, and this causes problems
results and bottom line revenue and surplus for academia as course development must always
x Surpluses (generated by academic successes) happen in a strictly sequential manner:
invariably cannot easily be protected or vired across
into new financial years, to protect against the 1. Recognize a need
cyclical nature of business interaction and what has 2. Develop a course
been termed the fundamental famine-feast 3. Validate the course
syndrome which characterizes the industry. As 4. Deliver the course
Aylett notes in the DTI paper, 23 ‘businesses
operating in performance motorsport - by its very And, perhaps, ultimately:
nature - have a high "death, attrition and growth
rate" because the demand for technical innovation is 5. Accredit the course with a professional institution
so high’. This situation impacts profoundly on
faculties’ abilities to attract or justify funding for From ‘recognizing the need’ to ‘delivering the course’
projects or programs where outputs are speculative can take many months – even years for some programs
and payback periods long-term (by academic – by which time the need has changed, as the
standards, which equate perhaps to 2-5 years). technological environment has evolved: institutions thus
forever play ‘catch-up’ with this modus operandi, with
In essence, schools and faculties operate largely as graduates equipped for the skills shortages of several
stand-alone business units operating in a competitive years past. In practice, very few courses are developed
and risky environment. However, not only do they incur from scratch, being gradual iterations and evolutions of
(often-punitive) central overheads over which they have other courses, but the timeframes still apply.
neither influence nor control, but also schools and
faculties are penalized as much for over-achievement Industry staff too busy and too important to free up from
(over-recruitment x , or profit, in the business sense) as company to train or upskill
under-achievement (under-recruitment). In an external
business environment, very few would place venture The motorsports industry in the UK, as noted before, is
capital or other investment into such a conceptually- characterized by a very large number of SMEs (some
flawed business model. 90% of the 4000 companies) with most being ‘small’ (5-
50 employees) or ‘micro (less than five employees). In
HEIs are risk averse such instances, all staff is key, multi-disciplinary and –
perhaps most importantly – eternally busy.
As highly-visible beneficiaries of large-scale public
sector funding (even a ‘small’ UK University likely will Industry staff too busy and too important to advise
have turnover in the region of hundreds of millions of academia/steer content/mentor students
dollars), universities (rightly) must ensure transparency,
particularly where this relates to effective institutional A criticism often leveled at academia by industry is that
governance and financial propriety and probity. academia doesn’t deliver what it requires: academia
However, the prospect of harsh external scrutiny, ripostes acidly by saying that industry is very poor at
relating particularly to fiscal spends and speculative communicating its needs to academia in any meaningful
strategic investment, sits uneasily alongside the way. Unfortunately, the very people within a company
prospect of ‘risk’ in the view of many Finance Directors. equipped to provide necessary strategic steer, critical
This predicates a risk-averse business model that evaluation, help with course development, and some
perhaps must perversely exist and transact successfully foresight into the future of the industry, reside in senior
in a fundamentally risky environment – such a situation management or on the board: this seriously limits their
makes for a very uneasy marriage and provides severe availability to provide the necessary advisory and other
impediment to long-term, sustained commitment of an services to academia.
institution to meaningful engagement with such a risky,
external environment (i.e. the motorsports industry and Timetabling issues
its other stakeholders).
Academic timetables, particularly in those institutions
which have adopted semesterization, place immense
x
Existing numbers of staff will be expected to sustain pressures on academics to perform core duties and
teaching of numbers of students in excess of that produce academic results within the semester period
predicted for some years before resources are made (and these often involve long days perhaps coupled to
available from senior management to recruit new staff some weekend working). Such a working environment,
and alleviate pressure on staffing levels which prevails for perhaps 75% of the year, is not

8
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

conducive to the active pursuit of so-called ‘third-stream’ obsolete. However, ‘obsolete’ to F1 is perhaps in most
activities, i.e. non-teaching and non-research pursuits. cases ‘state-of-the-art’ for a decade or so for a typical
(under-funded) HEI: such a scenario makes it very
Academic pace difficult for academic institutions to demonstrate any kind
of real credibility to the high performance engineering
This problem is coupled in many ways to staffing levels, community (at this kind of level anyway), insofar as
timetabling, and speed of response. ability to undertake contract research, consultancy or
testing is concerned.
Where an industry, which relies on invention, rapid
innovation, rapid prototyping and rapid product Low Visibility of Academic Institutions on Industry
development for its sustenance and competitiveness, ‘Radar’
has a problem, it generally seeks to have this problem
solved in days if not hours. By and large, academic Low visibility of higher and further education institutions
institutions are not geared toward providing such is another impediment to industry-academia
customer focus or quick response, since such activity collaboration that stems ostensibly from a combination
has to be timetabled into what (during term-time) is an of themes previously detailed, including ‘response time’,
already full daily/weekly agenda. Also, it is the author’s ‘market focus’, ‘poor marketing’, and ‘standard of
experience – generally – that the majority of academics, equipment/facilities’.
at some point, are poor at returning communications,
even when the request is urgent. This trait, again, does It is the author’s experience (drawn from his 17 years of
not endear academia to industry. employment in industry, albeit not in the motorsports
industry) that many world-leading companies – in the
Marketing main – do not seem to engage at all, or in any
meaningful way, with academia. Furthermore, the
Much of the problem of poor or non-marketing of smaller the company, the more this situation seems to
expertise again stems in part, the author believes, from prevail. Part of this derives, I believe, from that the fact
the migration of academic institutions from being purely that world-class companies have a sometimes-
‘seats of learning’ to now being ‘in the business of complacent perception that they occupy such pedestals
providing skills, education and training’. In the UK, it is of competitiveness because they understand themselves
perhaps exemplified best by what are termed now ‘the to possess all of the skills and expertise in-house to
new universities’ (which became universities out of the sustain and improve competitive advantage and thus
old polytechnic system). The old polytechnics, in the have no need of external support. There can also prove
main, majored on teaching rather than research, and so an embedded-ness of culture that makes companies
had little, if any, real experience of research, either pure strongly resistant to change (akin to the ‘not-invented-
or applied; such institutions now are competing, in some here’ syndrome).
respects and in particular sectors, against institutional
‘brands’ which have long-established high reputations This belief prevails not just in ‘technology’, but in
and – critically - marketing budgets commensurate with ‘management’ also. This is to the extent that the
such high reputations. There is also unwillingness, in Executive Director of Engineering at an F1 company
certain academic quarters, to aggressively market local to Oxford (a company which by any standard or
academic prowess other than to academic peers, metric is incredibly successful, particularly over the last
through the established mechanisms of publication, two seasons) commented that out of a 400-strong
patent or professorship. workforce of highly skilled and motivated
(predominantly) engineering or technical staff, he was
Another factor which compounds this is academics’ the only staff member (at the time of the interview) to
understanding of industry and the market, and their possess any formal management qualification (and that
making the connection between what they (and the had been gained only in the recent past). This was
institution) can offer academically to industry, and what despite the fact that within a 30-mile radius of the team
industry might (and invariably does) require of headquarters, there reside three universities having
academia. Business Schools of high international standing; if we
extend this radius to 50 miles (to include London), the
Standard of equipment/ facilities number would treble.

In an industry environment where at the highest In similar vein, in an interview with the Managing
echelons of the sport (Formula 1), the product (a racing Director of a world-leading automotive/motorsports
car) will evolve over an 18-race, typical 9-month racing design consultancy local to – and a long-term supporter
season to such an extent that perhaps only 10 to 15% of of – Oxford Brookes University, the MD’s response to
the product presented to the grid on the inaugural race the question, ‘What can the University do to improve
of the season is subsequently present in the closing your competitiveness?’ was, ‘… Nothing… Except
race, rapid product innovation is key. Thus, the perhaps provide me with more undergraduates like
bewildering array of Research, Design, Development (person x)’. Despite long-term involvement (10-years-
and Test equipment used by F1 quickly becomes plus) with the University, and a position (albeit on a

9
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

‘company’ rather than ‘personal’ basis) on the done, ask a busy person’ applies profoundly in the
Oxfordshire Motorsport Industry Forum, (a taskforce motorsports industry.
charged with remedying loss of the UK motorsport
industry’s international competitiveness, through In the UK, which as mentioned, numbers some 4000
enhanced and more market-focused skills, education motorsports companies having around 55,000 full-time
and training provision), the company could see no role equivalent employees, there number perhaps 10-20
for the University other than as a provider of industrial senior industry personnel (from recent personal
placement students and competent new graduates. experience) who engage regularly (and publicly) in
Such occurrences provide salutary lessons. initiatives – such as so-called ‘think tanks’, steering
groups, advisory committees etc. – which benefits
Pressures of race calendar immensely the academia-industry interface. It is thus the
case that such a valuable and committed – but finite –
In motorsports, it is invariably the case that teams industry resource is in high demand (and woefully short
(particularly engineers) are either thinking about how supply). Also, regardless of the real high value of the
they’re going to do it (pre-race prep.), doing it (racing voices of such a low number of people, 10-20 personnel
and competing), or analyzing the result to see if they can from the same number of companies can in no way be
do it better next time around (race post-mortem). fully representative of complex and diverse industry. The
Between race seasons, time pressures (somewhat Workforce Development Plan project 24 undertaken by
paradoxically, at least to the layperson) actually Kirkpatrick and Faulkner of the MIA suffered in similar
increase, as teams quest to make the next version of the respects in that of the 4000 or so questionnaires sent to
car competitive and ready for the new season. Thus, it is companies, (and which were to be used to inform the
notoriously difficult for academia to engage key players analysis of the industry situation at that time) only a very
within companies, i.e. those staff with key insight into small percentage (5.2%) were returned, and only a
and perception of: proportion of those returned contained worthwhile data.

x Evolving technologies Demographics of UK motorsports industry

x Emerging Skills, Education and Training shortages
x The present and possible future of the motorsport The UK motorsports is comprised largely (90%) of SMEs
environment, from a holistic perspective (Small-to-Medium-sized enterprises, i.e. companies
x Impediments to success having less than 250 employees); the majority of such
UK motorsports companies are ‘small’ (<50 employees)
This situation prevails particularly in larger companies. or ‘micro’ (<5 employees).
With these, HR departments most usually do not have
the same pressures exerted directly by the race With smaller companies (and the UK numbers some
calendar (unless a poor race performance necessitates 3600 of its 4000 motorsports companies in the ‘small’ -
lay-offs and redundancy)! However, the HR 5-50 employees - and ‘micro’ - less than five employees
department’s perception of a company and its skills - category), the problem is that most employees –
requirements provides a somewhat different – but because of the small number in a typical UK motorsports
nonetheless valuable – insight into a company’s human company – are absolutely key, being multi-functional
resource, particularly when compared to the intelligence and multi-tasking. In a 5-person company team,
yielded when an engineer (or engineering manager or occupying one hour of an employee’s time in a typical
director) is interviewed. With the proviso that HR 37.5-hour UK working week loses more than 0.5% of
departments invariably rule with an iron fist company that employees’ productive time. This percentage figure
training budgets, the extrapolation of both HR and negates of course any ‘productivity weighting’ accorded
engineering perspectives has to date proven the most an employee senior in the company hierarchy. One hour
fruitful and successful method of determining a lost of a ‘Finance’, ‘Managing’ or ‘Engineering’ Director’s
company’s Skills, Education and Training abilities, time – and these are the very people with whom HEIs
shortfalls and requirements. need to liaise to realize worth - might multiply the 0.5%
figure stated by one or two orders of magnitude. Thus, it
Industry staff availability is easy to understand why small companies are perhaps
reluctant to engage in any ‘extra-curricular’ activity that -
This factor in many ways echoes – and is inextricably on the face of it and in the short term – has impact
inter-related to – the ‘pressures of race calendar’ factor neither on core business nor on financial bottom line.
related above, particularly the section pertaining to SME
engagement. Insightful and worthwhile views into Courses aren’t available in accessible form or format
companies invariably are produced by senior personnel
having operational, tactical and strategic management It is the motorsports industry’s perception that courses
responsibilities, both in ‘business’ and in ‘engineering’ aren’t available in a form or format that allows
disciplines. The time of such people comes neither employees remote (typically work-based) to engage with
cheap or easily. It is also the case that, in motorsport, as learning materials.
in other disciplines, the old maxim of ‘if you want the job

10
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Industry’s ability to access academia’s course offerings, - one can understand quickly that company investment
particularly at post-graduate level – CPD, short course, in the generation of IP/subsequent product development,
MS, and post-graduate certificates and diplomas – is to gain competitive advantage, particularly given the
severely limited by a number of factors, as detailed high monetary gains to be had by attaining and retaining
previously. These include perceived industry technical supremacy, is not carelessly undertaken.
applicability, apparent worth (as this relates to employee
development/impact on company competitiveness), Within the strictly-controlled and policed confines of a
knowledge of courses, and perhaps most importantly, typical high-tech motorsports company, security of IP is
ability to engage with such programs. As the work of the assured (as much as is able) through a rigid framework
Technology and Innovation Working Group of the DTI of physical and legal mechanisms. Where IP
Motorsport Competitiveness Panel noted 25 , ‘the development is devolved – in whole or in part - to a third
development of a ‘post-graduation’ motorsport[s] party, companies lose the (physical) ability to control
academy’ is perceived as a better way of proceeding leakage of vital IP and thus are sometimes
rather than specialist long-term courses on motorsport, (understandably) jittery about entering into such bi- or
which tend to duplicate parts of existing courses… the tripartite agreements.
group recommends that a series of supplementary post-
graduate courses be funded to accommodate the needs Contract and other legal issues
of individuals wanting to become involved in motorsport’.
This sentiment was echoed in the report of the Skills,
Contract and other legal issues here relate specifically to
Education and Training Working Group of the Panel,
timeframe of turnaround of legal agreements.
which determined that ‘surveys found a need for… short,
effective courses delivered flexibly’. At present, there are
This element, like others, is related to – and completely
few – if any – dedicated motorsports courses at post-
inter-dependent upon - other elements, as previously
graduate level anywhere in the world that combine real
detailed, perhaps most noticeably, IP ownership. Many
industry worth with flexible delivery mode (although a
university departments will be required to invoke the
small number of courses meeting this first criterion do
services of an internal Legal Services Department,
exist), and the MKE is impacting profoundly on this
without option for outsourcing of the legal services that
situation.
such departments provide; thus, departments invoking
IP ownership conflicts such support services are hidebound by what Legal
Services decree| and must operate within a time and
The motorsports industry relies fundamentally on the responsiveness framework decreed not by the client (the
regular development of new and cutting-edge IP to university department) but by the service provider (legal
retain, sustain and improve competitive advantage. Any services), in direct contravention of the conventional
collaboration with a third-party agency – such as a model for service provision. The freedom to take custom
university – which seeks to retain a proportion of the elsewhere, as other small business might and would be
rights (even a small proportion of the rights) – to develop likely to do – particularly where service provision was
such IP is fraught. Within the UK university structure, slow, unhelpful, needlessly obstructive, and expensive –
there exist complicated and often-contradictory is noticeable by its absence.
university-specific regimes relating to ownership of IP
generated as a result of academic research/consultancy Such an arranged marriage, with a punitive pre-nuptial
endeavors: agreement attached, as detailed, is not without its
x Third party retains all IPR problems:
x Third party/university splits IPR
x University retains all IPR x Legal Service departments, like the majority of other
x University/individual academic(s) split IPR university elements, abide by a typical ‘public sector’
x Academic concerned retains all IPR timeframe, as this relates to action and reaction,
request and service, which often presents a highly
x A combination of the above
frustrating model for interaction, as far as industry is
Once one factors in an industry intent on maintaining concerned.
absolute secrecy of product/process development, it x As servants to the public sector, and recipients of
makes for a heady and explosive cocktail. Ecotec 26 public sector funding, universities have in place
reinforce this: ‘poor links with educational institutions labyrinthine processes and procedures in place to
were alluded to quite frequently… there were some ensure there exist appropriate checks and measures
familiar criticisms of educational institutions… relating to to satisfy funding councils and other interested
their ability to exploit intellectual property…’ stakeholders.
x By and large, universities are hugely risk averse,
IP leakage concerns and this is reflected in the policies adopted by legal
Given the above - succinctly represented perhaps as the service departments, and subsequently the
equation: contracts that they draw up.
Superior knowledge + fast product development = win

11
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

If one were in subversive mood, one might contend that as well as suggesting measures to address, ameliorate
many of the legal and financial processes and or reduce them, where at all possible.
procedures embedded within university culture are
dedicated toward thwarting worthwhile
engagement/collaboration with third parties… As Bell et
al (1994) concur, the first dilemma that Industrial Liaison Problem Solution
Officers face is between the marketing function of
offering a university’s technology and expertise to 1 Academe Employ empowered, dedicated
external users, contrasted with the regulatory function to response time industry professional having
ensure that the university only enters into appropriate knowledge of key stakeholders,
contracts. The two key issues on ‘appropriateness of network management etc., and
contracts’ are usually price and legal liability. 27 place as a single key point of
contact
Cost
2 Lack of As above – postholder charged
UK universities are bound now by Higher Education market focus also with regular industry
funding councils to adopt TRAC (TRansparent Approach engagement for intelligence
to Costing) and fEC (Full Economic Costing) purposes
methodologies when costing proposals (for research
grants, consultancy, and other work with third-parties). 3 Poor industrial As (2) above, plus regular
These methodologies take little notice of market content convening of an Industrial
expectations within specific sectors, or the concept of Advisory Panel/Steering Group to
‘loss-leading’ initially to demonstrate competence and allow industry-academe dialogue
perhaps gain or grow market share. This is of particular
significance when universities have to compete with the 4 Academic Recruit specific academic staff
private sector in respect of service provision to industry, staffing charged with involving themselves
most specifically – for the instance of this paper – difficulties in industry-focused
research, consultancy and contract testing. teaching/research pursuits.
Create an inter-disciplinary, multi-
Departments often struggle to compete with commercial institutional R&D facility
providers of the services detailed above, as a result of
punitive central overheads imposed when any external 5 Timetabling As (3) above, plus supplement
engagement results in some kind of service provision. with Hourly-Paid-Lecturer support
Levies imposed by Central Finance Directorates range to allow successful prosecution of
from 40 to 70%, which, in comparison to small-staffed, high value opportunities
agile companies residing in low-cost or subsidized
(perhaps company start-up) premises contrive in mant
6 Academic As above, but personnel must
instances to make HEIs a wholly unattractive proposition
pace have sufficient authority or the
as far as any industrial collaboration/service provision is
devolved authority of a senior
concerned. This is despite the high worth that
person, to exercise pressure
engagement with the individual department’s concerned
would realize from such external engagement.
7 Poor x Increased advertising
marketing x Visible presence at
Summary – Impediments to Industry-Academe
Knowledge Exchange networking events, trade
shows, trade missions etc.
The figure below seeks to encapsulate some of the x Better web presence,
major impediments to industry-academe engagement, particularly on search engines
and the interdependencies between the various x Employ sector-specific
elements. In doing so, it is hoped that some of the marketing professionals
complexities of industry-academe collaboration – within x Employ dedicated personnel
the motorsports industry environment – become more from industry
readily apparent. 8 Equipment x Seek funding from
standard ‘unconventional’ sources
ADDRESSING IMPEDIMENTS TO KNOWLEDGE x Pursue industry for equipment
EXCHANGE AND TECHNOLOGY TRANSFER pledges
x Sell institution to industry
The Theory based on staff excellence, not
The following table summarizes impediments to facilities
Knowledge Exchange and Technology Transfer, from an 9 Low industry As for (1), (4) and (7) above
Academe-Industry perspective, as detailed previously, visibility

12
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

10 Pressures of x Develop online courses for graduate and postgraduate courses aimed at
race calendar Work-based Learning industry usually provides long term benefits.
x Consider ‘boot-camp’ style 10. Special provision for SMEs - SME clubs, associate
intensive teaching sessions schemes and other initiatives.
11 Availability of No real solution, other than 11. Service laboratory - services based around
industry staff cultivating good relationships with departmental expertise/ facilities.
industry, and delivery of what is 12. Hosting conferences, etc. - usually a successful way
promised, so that value is evident, to link the department with an area of research
and interaction is seen as a expertise, and to increase its profile.
necessity 13. Promotional activities - targeted mail- shots and
open days, and visits to industry.
12 IP Ownership Create simple boilerplate template 14. Participation in local scene - particularly activities
relating to a specific research area.
13 IP Leakage Evolve facilities to include 15. Industrial advisors - engaging advisors from industry
demonstrable research ‘cell’ to ensure research relevance and increase industrial
security commitment.
16. Exploiting alumni network - ex-students can provide
14 Contract/ legal Develop boilerplate template become industrial customers.
issues 17. Providing incubator facilities - helps to capitalize on
previous research with additional positive feedback
15 Cost No real solution here, other than in terms of contracts/consultancy.
high-level policy change 18. Auditing technology - external technology audits is a
proven technique for identifying opportunities to
exploit expertise/facilities.’
A further DTI report 28 , (DTI, 1995) undertaken by Oxford
Innovation, detailed specifically the various ways that Similarly, the leverage that key policy interventions have
academia and industry could collaborate for mutual on cluster competitiveness also has been investigated
growth: by Government. Many of these policy interventions
center on the fostering of relationships and networks, to
provide mechanisms for knowledge exchange and
‘Eighteen key techniques are employed to promote and
facilitate collaboration: these are listed below in the technology transfer, as follows. (Ecotec, 2002) 29 :
order that these were utilized:
x Establishing links or communication structures with
1. Using UK/ Euro networks - especially conferences/ other companies and supply chains e.g. The
networks that offer opportunities to make linkages Oxfordshire Motorsport Forum
with potential industrial clients x Collaborating with the voluntary sector and
2. Narrow specialization - creating a unique research Universities to encourage technology transfer, for
resource in an industrial niche. instance, or creating R&D support agencies
3. Assigning responsibilities - giving one or more x Increasing the depth (more suppliers) or breadth
people in a department some specific responsibility (attracting more industries) of the cluster
for building links with industry. x Creating informal and formal networks with trade
4. Reputation for excellence - developing a reputation associations, unions or professional bodies, large
for excellence in research and in industrial companies and SMEs
collaboration is a good recipe for growth. x Establishing strategic alliances/trade agreements for
5. Abolishing sectional barriers - notably initiatives example establishing trade agreements between
which facilitate work that falls across traditional large companies and also their suppliers to promote
departments and improves marketing/ liaison. skills and technology transfer
6. Formation of clubs - includes local and industry- x Raising the profile of a cluster through centers of
specific clubs, many of which do not generate excellence and sharing good practice or
income directly through membership. demonstration projects
7. Student/ graduate placements - includes x Stimulating international networks and networks
undergraduate placements (often 12 months) and beyond the cluster, or forming links with companies
programs such as Teaching Company Scheme over seas
(TCS) [Now known as KTPs or Knowledge Transfer x Developing facilities or initiatives for collaboration
Partnerships] (e.g. the construction of research centers or the
8. Recruitment/ exchanges - recruiting staff from implementation of technology or knowledge transfer
industry and two- way exchanges helps to ensure or R&D initiatives)
the relevance of research to industry.
9. Industry- related courses - developing short, The same report notes that, ‘…the role of leadership in
the development of clusters remains open to debate. It is
interesting that this factor is cited strongly by

13
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

practitioners but given less significance in the evaluation fax machines, Apple Mac computer, VHS domestic
literature. Industry leaders, it is argued, can be crucial Video Recorder)
for removing obstacles, assisting in enhancing
collaboration, developing a vision and acting as However, Rothwell also notes that the inadequacies of
‘champions’ for the future strategy of the cluster. Strong such models can be resolved in more complex models,
civic leadership can help to foster a ‘collaborative which show the innovation process to be more iterative
advantage’ too by raising mutual awareness of local and overlapping, with a blurring of the boundaries of the
strengths and a shared vision for business growth. various stages, and the sometimes cyclical nature of the
Leaders are typically committed to a local area, process which allows adaptation and evolution of the
perceived as having a high degree of influence, and able basic product: ‘…linear technology-push and need-pull
to cultivate interactions between cluster stakeholders. innovation began increasingly to be regarded as over-
Strong leadership can of course assist with human and simplified, extreme and atypical examples of a more
social capital (trust and motivation) as well as general process of coupling between science,
partnership and networking (by encouraging technology and the marketplace’:
collaboration and co-operation)… For practitioners, the
choice of leaders can be fundamental to the success of x The fourth generation – integrated – model mirrors
the cluster’. actual practice more closely by representing
innovation not as a simple sequential process but as
However, with respect specifically to motorsports, a more complex parallel process which can involve
several authors challenge the views presented in the simultaneous consideration of R&D, design,
‘evaluation literature’ cited and noted above. The reports manufacturing, marketing etc. This model
of two of the working groups of the DTI Motorsport emphasizes the importance of integration, both
Competitiveness Panel (Technology and Innovation, and between the various departments within a firm
Skills, Education and Training Working Groups) noted (R&D, manufacturing, marketing) and also outside,
specifically the critical importance of management, ‘upstream’ with suppliers and ‘downstream’ with
leadership, business development and entrepreneurial customers. Integrated – innovation with integration
qualities of the very best practitioners within the between marketing, design and production and a
motorsports field. parallel product development process (e.g.
Japanese automobiles, Landrover Discovery 4WD
Evolution of the Innovation Model vehicle, Sony Discman)
x The fifth generation model emphasises much closer
It is recognized by a number of authorities that the strategic integration between collaborating
process of innovation (innovation models), and that of companies (including e.g. co-development of new
achieving competitive advantage, has evolved since the products with suppliers). It also draws attention to
advent of the Industrial Revolution. the electronification of innovation via, for example,
linked supplier/user CAD systems and electronic
Rothwell 30 , renowned for seminal works about product design/manufacturing links leading to
innovation, product development, and competitive improvements in Design For Manufacture (DFM).
advantage, considers that there exist five (generational) Thus, the Fifth Generation Systems Integration and
models of the innovation process, which have evolved Networking model embraces innovation involving
over perhaps 100 years; the first three (simplistic) inter-company networking and advanced electronic
models of innovation are detailed below. aids for design, development and communications
(e.g. Boeing 777, Hitachi and other Japanese
1. Technology push – innovation stimulated by a new corporations adopting such models for the
idea, invention or technical possibility i.e. basic development of advanced electronic products)
science/ design and engineering/ manufacturing/
marketing/ sales (examples include: phonograph, jet Competitive innovating firms have migrated from
engine, CNC machine tools, laser, hovercraft, simplistic first, second and third generation innovation
Wankel rotary engine) models toward the Fourth (so-called ‘phase
2. Market pull – innovation driven by market need, overlapping’ or ‘concurrent’, or ‘integrated’) and Fifth
demand or opportunity i.e. market need/ Generation (Systems Integration and Networking)
development/ manufacture/ sale (examples include: models. The Fourth Generation model considers
car air bags, AZT drug treatment for HIV, safety innovation as a parallel process simultaneously involving
razor, car seat belts, ABS brakes, aircraft pilot elements of R&D and prototype development and
ejector seat, contraceptive pill) manufacturing and so on. During the 1980s, increased
3. Coupling/interactive – innovation through matching emphasis was placed on integration across the
technological opportunity to market needs or R&D/manufacturing interface (e.g. linking CAD and
opportunity, i.e. combines science, technology and FMS) and on closer collaboration with suppliers and
the marketplace, and has feedback loops but is leading edge customers. These developments were
essentially sequential (examples include: electric largely based on lessons from Japanese manufacturers
lamp, photocopier, solar cells, Sony Walkman, office and included the concept of lean manufacturing. Also,
horizontal collaboration (joint ventures and strategic

14
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

alliances) increased dramatically. As Rothwell (ibid) initiatives that Oxford Brookes University Dept. of
notes, ‘while the ‘Integrated’ model represents a Mechanical Engineering (initially) and its Motorsport
convergence between conceptualization and practice, Knowledge Exchange project (latterly) have engaged in,
SIN represents a model of the future in which over a period of some 10 years, to contribute to
conceptualization leads practice. Essentially SIN successful knowledge exchange and technology
represents what might be achieved when companies transfer, where the ultimate aim is the support,
catch up with and take on board leading edge sustenance, and improvement of the UK motorsports
developments in the management and implementation of clusters’ competitiveness. I then map such (practical)
innovation: while many companies are still striving to initiatives across to their (theoretical) counterparts from
master the fourth generation model, leading innovators literature, to demonstrate the veracity and real practical
are already introducing elements of the fifth generation worth of the various suggestions postulated in literature.
model into their innovatory practises.’
Theory into Practice – What Oxford Brookes University
It is key to the overall argument of Rothwell’s paper that has done over the last 10 years to enable and facilitate
the ‘fifth generation’ model is founded fundamentally on, motorsports industry knowledge exchange and
as he terms it, ‘systems integration and networking’. technology transfer
(SIN) As he notes, ‘while the integrate model represents
a convergence between conceptualization and practice, As in the ‘Theory’ section detailed previously, I present
SIN represents a model of the future in which now in tabulated form a number of literature-derived
conceptualization leads practice. Essentially SIN theoretical remedies (taken from the DTI list specified
represents what might be achieved when companies and numbered previously) and map these across to
catch up with and take on board leading edge practical, fully-realized, solutions, as adopted by Oxford
developments in the management and implementation of Brookes University, and its Motorsports Knowledge
innovation: while many companies are still striving to Exchange.
master the fourth generation model, leading innovators
are already introducing elements of the fifth generation Theory (from Putting the Theory into Practice
model into their innovatory practises.’ DTI, 1995) at Oxford Brookes University

In essence, the evolution of this model represents a 1 - Using UK/ Attendance at MIA Educators’ and
realization that, whilst local and regional firms at certain Euro networks Employers’ Forum; attendance at
levels compete (sometimes aggressively), to remain MIA and other networking events;
competitive overall, particularly against larger, predatory attendance at PRI and Autosport
organizations or conglomerations residing outside the International; being part of
immediate cluster, there is a real need for cluster-based international inward missions; part
firms to work together to counter such attacks. Overall, of Faraday Advance partnership;
the rationale is to: Motorsport Knowledge Exchange
partnership growing to include
x Compete as normal, to raise standards international partners
x Collaborate where it is expedient to do so (usually to
counter an attack by a larger and better-equipped 2 - Narrow Specialist research areas including
company predator). specialization Advanced Vehicle Engineering,
x Recognize that the ‘knowledge community’ is Joining Technology, Stress and
disparate and spread across a cluster and doesn’t Materials Analysis, Innovation
reside within a single company, and so networking Management, and Advanced
and communication have long-term benefits. Materials

The Practice 3 - Assigning Dedicated Industrial Liaison Officer

responsibilities who deals with industrial
Overview placements; dedicated Motorsport
Project Officer; Motorsport
Academic and other theory is all very well, but the litmus Knowledge Exchange (MKE)
test comes when attempts are made to put theory into project team
practice, and ultimately, company bottom lines become
exposed. 4 - Reputation for Part of the Oxfordshire Motorsport
excellence Industry Forum, from which in 1996
Discussion has been had in this paper of the various the world’s first market-focused and
impediments to knowledge exchange and technology industry-led motorsport degree was
transfer, as perceived both through practical involvement created
in the process, and academic conjecture.
5 - Abolishing Development of ‘Business for
I seek here to end on an optimistic note and reflect now Motorsports’ by School of
on the numerous successful theory-into-practice

15
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

sectional barriers Technology in conjunction with the in local scene 10 years

Business School, development of
Data Acquisition online course 15 - Industrial School overall has an Industrial
between Dept. of Mech. Eng., Dept. advisors Advisory Group; Motorsports
of Elec. Eng. and the Open Knowledge Exchange has had over
University 250 days of senior staff time
committed to it for strategic steer,
6 - Formation of MKE is looking to form an Industry critical review and help with course
clubs Club wherein industry leaders development and delivery; a
speak to their supply chain, and number of research projects (PhDs)
disseminate best practice are currently being undertaken with
local motorsport companies,
7 - Recruitment/ Three staff recently drawn from the including F1
exchanges highest echelons of motorsports;
majority of academic staff have an 16 - Exploiting 70% of Oxford Brookes’ students
industry background; most alumni network go directly into the auto. industry
Motorsports Knowledge Exchange including 14% into F1 and 43% into
staff have come from industry; other motorsports, so alumni
School engages in the DTI’s network is extensive, and given the
Knowledge Transfer Partnership time that the programs have been
(KTP) scheme. in existence, many alumni now
occupy senior positions (Adrian
9 - Industry- Motorsport degree program Reynard was one of the key
related courses incorporates vocational courses influencers behind the Oxfordshire
delivered by college partners as Motorsports Industry Forum).
well as BS, BEng, MEng, MS and
PhD programs. See below also. 17 – Incubator No activity here yet
10 - Special Short courses developed by the facilities
provision for MKE and delivered through on-line
SMEs and blended learning environments 18 - Auditing Technology audits for the School
will allow engagement by SMEs technology are regularly undertaken
and busy managers

11 - Service Facilities have been developed CONCLUSION

laboratory around departmental teaching and
research expertise (e.g. Multimatic Through a combination of: academic discourse, (the
industrial standard 4-post rig for the theory); real experience of engagement by HEIs with
Chassis Dynamics team, Particle industry (to endeavor transfer of both knowledge and
Analyzer and Drop-tube furnace for technology, or the converse) – in essence, the practice;
the Combustion and Engine and a number of propositions to highlight difficulties
Research group) (perceived and actual), this paper has presented a
number of solutions to the impediment of knowledge and
12 - Hosting Oxford Brookes University was technology transfer in the motorsports industry, and
conferences, etc. asked by Racecar magazine to host examples of good practice.
the World Motorsports Symposium;
MKE is developing a series of ACKNOWLEDGMENTS
‘Motorsports Masterclasses’
(evening lectures/gust I am indebted to a number of motorsports industry
presentations given by industry personnel, academics, and other key stakeholders, who,
experts) during the course of the Motorsports Knowledge
Exchange project (and before), have given freely of
13 - Promotional Adverts and articles placed their time and in doing so have helped inform the
activities strategically in various trade content of this paper. They include:
journals at magazines at key points
throughout the year; dedicated x Paul Atkin, General Manager, MIRA
industry-focused research and
x Chris Aylett, Chief Executive, Motorsport Industry
consultancy brochure created; local
Association (MIA)
industry has been very generous
x Charlie Bamber, Managing Director, Menard
Engineering Limited (MEL)
14 - Participation Participation in local motorsports
x Guy Faulkner, Westminster University
scene has been ongoing for over
x Jane Gilham, HR Manager, Xtrac Limited

16
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

x Geoff Goddard, former Chief Engine Designer, 12 Pinch, N., 1999, In Pole Position: Motorsports
Cosworth Engines, former Head of Engine Design, Successes in Britain and its lessons for the World’s
TWR/Arrows Motor Industry, Euromotors Reports, London
x Jon Hilton, Head, Engine Research, RenaultF1 13 ‘A Second Draft Report for the DTI – Success
x John Kirkpatrick, JK Consulting, owner, Jim Russell Criteria for Success in Cluster Development’, p19
Racing, Founding Director, MIA 14 Porter, M., (1990) ‘The Competitive Advantage of
x Brian O’Rourke, Chief Composites Engineer, Nations’, The Free Press, New York
Williams Grand Prix Engineering 15 OECD (2001) ‘Innovative Clusters: Drivers of
x Adrian Reynard, Principal, Auto Research Center National Innovation Systems’, OECD, Paris
x Sue Scarf, HR Director, Prodrive Limited 16 Porter, M., Ketels, H. M. K., (2003) DTI Economics
x Pat Symonds, Executive Director of Engineering, Paper No. 3, UK Competitiveness: moving to the
RenaultF1 next stage, DTI, London
x Peter Van Manning, Managing Director, McLaren 17 Ecotec (2002) ‘A Second Draft Report for the
Electronics Department of Trade and Industry: Criteria for
Success in Cluster Development ‘, p2
18 Ecotec (2002) ‘A Second Draft Report for the
I am also privileged to have been invited to be a Core
Department of Trade and Industry: Criteria for
Member of the Skills, Education and Training Working
Success in Cluster Development ‘, p37
Group of the Her Majesty’s Government’s DTI
19 [Link]
Motorsport Competitiveness Panel process. Much of the
exchange of information, discussion of personal 20 Ecotec (2002) ‘A Second Draft Report for the
perspectives, and presentation hard data gained during Department of Trade and Industry: Criteria for
presence on the Working Group has helped inform the Success in Cluster Development ‘, p44
content of this paper. 21 Kirkpatrick, J., and Faulkner, G., (2003) ‘Workforce
Development Plan’, p31, MIA Warwickshire
22 Anon (2003) ‘Report of the Motorsport
REFERENCES Competitiveness Panel’, presented to the Secretary
of State for Trade and Industry, July 2003, p14,
1 Bamber, C., (2005); Anon, (2005), private [Link]
conversation [Link]
2 Henry N., Angus T., Jenkins M. et al (2002) ‘A 23 DTI/SMMT (1999) ‘Motorsports - a sector case
Study into the UK Motorsport and Performance study where the UK leads’,
Engineering Cluster’, Motorsport Research [Link]
Associates,, MIA, Warwickshire 99_1?template=template%[Link], DTI,
3 Financial Times (1990) ‘Search for quality backing’, London
26th January 1990, London 24 Faulkner, G., Kirkpatrick, J. (2003) ‘Workforce
4 Henry N., Angus T., Jenkins M. et al (2002) ‘A Development Plan’, MIA, Warwickshire
Study into the UK Motorsport and Performance 25 Anon (2003) ‘Report of Technology and Innovation
Engineering Cluster’, Motorsport Research Group to the Motorsport Competitiveness Panel’,
Associates,, MIA, Warwickshire p9
5 Kirkpatrick, J., and Faulkner, G., (2003) ‘Workforce 26 Ecotec (2002) ‘A Second Draft Report for the
Development Plan’, MIA Warwickshire Department of Trade and Industry: Criteria for
6 Stancombe, J., (2002) ‘The UK Motorsport Success in Cluster Development ‘, p
Industry’, Proceedings of the 10th Welsh Economic 27 Bell, E.R.J. Stott, M.A., and Kingham, D.R. (1994)
Research Unit Conference, ‘Flexing Welsh Muscle: ‘Oxfordshire Firms and Technology Transfer,
Sport and Economic Development’, p7, Cardiff Interactions between small and medium sized
7 Kirkpatrick, J., and Faulkner, G., (2003) ‘Workforce enterprise (SMEs) and the local research base’,
Development Plan’, MIA Warwickshire The Oxford Trust
8 Blacker, F., (1995) ‘Knowledge, Knowledge Work 28 DTI, Oxford Innovation (1995) ‘Best Practices in
and Organizations: An Overview and Industry-Academe Collaboration’, DTI, London
Interpretation’, Organization Studies 16/6 pp. 1021- 29 Ecotec (2002) ‘A Second Draft Report for the
1046 Department of Trade and Industry: Criteria for
9 Marotti, F., (2001) ‘Learning to share knowledge in Success in Cluster Development ‘, p46
the British and Italian motorsport industries’, p.4-5, 30 Rothwell, R. (1992) Successful industrial
Cardiff innovation: critical factors for the 1990s’ R&D
10 Henry, N., and Pinch, S., (2000) ‘Spatialising Management vol. 22, no. 3 pp221-239
knowledge: placing the knowledge community of
Moltor Sport Valley’, Geoforum, Vol 31 No. 2,
p.1911-208 CONTACT
11 DTI (2001) Business Clusters in the UK – A First Mike Meechan interrupted two years of formal university
Assessment, DTI, London engineering education in 1985 to join the BBC World

17
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Service as a Trainee Engineer. After two years at Bush $1.6m from SEEDA to fund the kitting out of the new
House in London, he moved to the World Service’s Oxford Brookes Motorsport Academy with state-of the-
Monitoring Station operation at Caversham Park, art teaching and research equipment, US$370k to
Reading, England where he remained until 1996 working develop market-driven, industry-focused, innovative,
as a senior engineer on state-of-the-art data distribution, distance learning motorsport CPD postgraduate courses
audio, video, RF, and satellite newsgathering systems. delivered through the online environment and achieved
US$700k of follow-on funding to continue the activities of
After a period as a consultant audio designer, freelance the MKE until July 2008.
author, and pro-audio equipment reviewer, he then
worked for the Electrical Impedance Tomography He graduated with BSc (First Class) and MEng degrees
Research Group at Oxford Brookes University, before from the Open University and is a Member of the
joining News Digital Systems (the high-technology arm Institute of Engineers and Technologists. He has
of Rupert Murdoch’s News International business) as authored over 75 articles and papers.
Technical Author in its MPEG-2 Compression, Transport
Stream Multiplexing, and Conditional Access mmeechan@[Link]
department. From 1998 until 2002 he worked for Encoda
Systems (formerly Drake Automation) as Documentation ADDITIONAL SOURCES
Manager, and whilst there, documented some of the
largest and most sophisticated digital multi-channel The majority of papers cited as references within this
broadcasting complexes in the world, including those at paper are available for download from the Research
Sky, Osterley, London (UK), UPC EuroHITS section of the Motorsport Knowledge Exchange web
(Amsterdam), and DIRECTV (Los Angeles). portal at: [Link].

From 2002 until 2005, he worked as Motorsport Project DEFINITIONS, ACRONYMS, ABBREVIATIONS
Officer at Oxford Brookes University, within the
university’s Research and Business Development Office. CPD: Continuous Professional Development
While there, key deliverables included: undertaking a
consultancy - Motorsport Academy Feasibility Study –
DTI: Department of Trade and Industry
for the Regional Development Agency, SEEDA;
representing the University on the Skills, Education and
EU: European Union
Training Working Group of the UK Government’s DTI
Motorsport Competitiveness Panel, and feeding the
consultancy results beneficially into the panel process; FE: Further Education (colleges)
writing the winning bid to SEEDA (the Regional
Development Agency for the South East of England) for HEI: Higher Education
a $US4m grant to establish a regional Motorsport
Academy, and leveraging over $US2m of in-kind support HEI: Higher Education Institution (University)
from industry for the proposal; and conceiving and
developing the winning proposal to establish a KTP: Knowledge Transfer Partnership
Motorsport Knowledge Exchange, which was to be
based at - and led by - Oxford Brookes. QAA: Quality Assurance and Accreditation

Since March 2005, he has led the Motorsport SEEDA: South East of England Development Agency
Knowledge Exchange, a 4-year, US$1.4m 5-person
project concerned with impacting on regional motorsport TCS: Teaching Company Scheme
industry competitiveness, particularly through
development of innovative industry-focused employee SME: Small-to-Medium Enterprise (company having less
upskilling initiatives, using the strengths of its than 250 employees
collaborating motorsports college and university
partners. These initiatives will help address what the DTI TCS: Teaching Company Scheme
Motorsport Competitiveness Panel Report cited as a key
requirement for ‘short, effective courses delivered
flexibly (i.e. at or close to the workplace, at times to suit
the company and at short notice) covering both specific
technical skills but also… complementary skills [decision
making, project management etc.]’. In the period since
the MKE project start in March 2005, he has leveraged
around US$2m for the School of Technology, including

18
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

2004-01-3502

Hardware-in-the-Loop Testing in Racing Applications

Peter Waeltermann, Thomas Michalsky, and Johannes Held
dSPACE GmbH, Germany

Copyright © 2004 SAE International

ABSTRACT In racing particularly, the high cost of the vehicles and

test benches means that frequent virtual test drives and
tests are run on a simulator. The objectives are to
Hardware-in-the-loop (HIL) simulation has established ensure the greatest possible hardware and software
itself as a standard method of testing electronics quality before going over to the test vehicle or test bench,
components in the automotive industry. HIL simulators to cut costs, and to minimize the danger to test drivers.
are now being used both by automotive suppliers and by
automobile manufacturers. In racing particularly, the high This paper presents applied hardware-in-the-loop
cost of the vehicles and test benches means that simulation with reference to the particular requirements
frequent virtual test drives and tests are run on a of motorsports (especially Formula One) and describes
simulator. The objectives are to ensure the greatest relevant customer solutions.
possible hardware and software quality before going over
to the test vehicle or test bench, to cut costs, and to HARDWARE-IN-THE-LOOP-SIMULATION
minimize the danger to test drivers.
Virtually all major formula 1 teams and some rally teams In hardware-in-the-loop simulation, one or more real
(WRC) use dSPACE simulators as HIL electronics test components (for example, electronic control units) are
systems, for example, for the widely used electronics studied in interaction with components simulated in real
platform from Magneti Marelli and also for their in-house time (models, for example, of the engine, transmission,
developments. The major focus is on ECU testing but powertrain and whole vehicle).
aspects of ECU calibration will also be discussed. The The interactions between real and simulated subsystems
paper describes this technology and presents solutions make tough real-time demands on these systems. High
prepared for racing teams worldwide. computing power (processor hardware for simulating the
models, and the actuators and sensors) and I/O
performance are therefore essential preconditions.
INTRODUCTION Frequently, special hardware is used, for items such as
fast engine signals and for generating and measuring
Hardware-in-the-loop (HIL) simulation has become a CAN signals.
standard method of testing electronics components in Using HIL simulation has the following advantages:
the automotive industry. HIL simulators are now being
used by both automotive suppliers (for component tests) • The hardware and software of electronic control
and automobile manufacturers (for network and units (ECUs) can be tested (automatically) at an
integration tests). early stage because there is no need for a real
engine or a real transmission to be present -
resulting in greater maturity before transfer to the
test bench or vehicle.
• Fewer test drives and test bench experiments
are needed, and also fewer vehicle prototypes or
vehicle components - resulting in cost savings.
• Safety-critical components (such as drive-by-
wire) can be tested without danger. Component
failures and associated emergency scenarios
Figure 1: Virtual test drive with HIL simulation can be tested reproducibly.
• Difficult ambient conditions (winter tests, rain,
high or low temperatures) can be achieved by
simple parameter changes.
• Test bench experiments are highly reproducible.

19
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

• The vehicle and its components can be simulation for emulating absent network nodes. This is
visualized by 3-D animation. This functionality particularly important if the chassis and the engine are
can be extended, right through to a real drive not developed on the same site (as is the case with most
simulator for man-in-the-loop applications. F1 teams).

HIL simulation has established itself as a standard in the TYPICAL ECU ARCHITECTURES AND DETAILS
development process for passenger and sport-utility
vehicles and is also growing in importance in The functionality of sensor systems in motorsports is
motorsports. hardly any different from that in production vehicles.
However, there is a considerably greater volume of
MOTORSPORTS REQUIREMENTS physical variables such as temperature, pressure, and
air/fuel ratio to be captured. Racing cars generally have
Clearly, many of the above advantages also apply to far more sensors (> 100). The sensors used are
motorsports. Compared with the development of optimized for racing purposes (weight, dynamics,
production vehicles, however, motorsports have their precision) and are often specially produced custom
own specific and even tougher requirements: components.

• Test drives are several times as expensive in

motorsports as for production vehicles, and
moreover, they are partially restricted to only a
few weeks a year (F1).
• When test drives are performed, all the relevant
data has to be logged via the central electronics,
as most racing cars cannot take additional
measuring equipment on board. Moreover, it is
not possible to sit in the passenger seat and
modify ECU variables via a calibration system to
fine-tune the system. This is only possible via
telemetrics, if at all.
• Software development is characterized by
weekly software releases and last-minute
modifications (short iteration loops). The ECU Figure 2: Typical sensors and actuators for racing
parameters are individually adjusted for each applications
racetrack.
As vehicle electronics are increasingly decentralized, the
REQUIREMENTS WITH REGARD TO HIL ECU network is spread across almost the whole vehicle.
SIMULATION: With this decentralized structure, the number and weight
of cable harnesses can be reduced. Fig. 3 shows a
HIL simulation requires enormous processing power for typical decentralized F1 vehicle topology with a few
the real-time simulation of the engine, powertrain, and selected components.
vehicle dynamics components. Because of the
dynamicity of the components and the high engine
speeds, simulation step sizes of approx. 50 – 250 us are
usually used, as compared with production vehicle
applications, where 500 us – 1 ms are more usual.
-1
And engine speeds of up to 20,000 min , angle-
synchronous generation of crankshaft and camshaft
signals, and angle-synchronous capture of injection and
ignition signals, require extremely fast I/O subsystems.
Special hardware is generally used for this.
In motorsports, vehicles frequently have special sensors
(LVDTs, thermocouples, linear lambda probes, etc.) and
actuators (Moog valves, fast injection and ignition
systems [1] that have only very limited use in production Figure 3: Typical decentralized F1 vehicle topology
systems for cost reasons (for example, LVDTs for with selected components
automated transmission). Mapping these for HIL
simulation thus often requires special hardware or
software solutions.
In contrast to production vehicles, the CAN busses in
racing cars are always run at the highest baud rate
(1Mbit/s). Typical tasks include capturing all the CAN
messages of the ECU network and CAN restbus

20
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Figure 4: Components of a hardware-in-the loop test system for racing

dSPACE SIMULATOR TECHNOLOGY

HARDWARE
Virtually all major F1 teams (e.g., Ferrari, Toyota [2]) and
some rally teams (WRC) use simulators from dSPACE in The core element is the generic real-time hardware,
Paderborn, Germany, for HIL electronic testing, for which comprises the processor board and I/O boards
example, for the widely used electronics platform from tailored to the actual application. The simulation
Magneti Marelli (Step9, Step10, from 2005 also Step11), requirements for testing the ECU and ECU’s networks
from TAG McLaren, and also for their in-house are therefore addressed by the solutions described
developments. below.
The simulators are used on the one hand for testing
components, for example, the engine electronics Processor Hardware
subsystem. On the other hand, the entire system
including all the essential elements of the ECU network Dynamic models of the engine, transmission, powertrain,
is tested. The simulator has to be able to cope with the and vehicle dynamics are executed on a real-time
resulting complexity of the ECU network, and its processor hardware, such as the dSPACE DS1006
hardware and software need to be scalable. Processor Board at sampling rates of 50 us to 1 ms. The
board has an AMD Opteron® processor with a clock rate
OVERVIEW: THE COMPONENTS OF A SIMULATOR of 2.2 GHz, allowing it to compute a mean-value engine
model, a brake hydraulics model, and a vehicle
Fig. 4 shows the typical subcomponents of an HIL dynamics model, including the entire I/O for the engine
simulator and their interactions. Other important and vehicle dynamics control, in less than 350 us. For
hardware components apart from the ECUs are the even more complex models, or to connect several
processor and I/O hardware, the signal conditioning, and simulators, the boards can be networked to form
the electrical failure simulation. On the software side, a distributed multiprocessor systems.
graphical user interface (incl. animation environment)
and dynamic models of the vehicle components are
required. A test automation environment is also needed
to automate test sequences. These components are
described in detail in the following chapters.

21
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

I/O Hardware Electrical Failure Simulation

The I/O capability of the simulators is directly derived As a rule, all manufacturers of racing ECUs specify
from the actual application. Engine simulation, for electrically protected ECU pins, but the racing teams still
instance, involves generating crankshaft, camshaft and perform validation against this specification. Electrical
knock signals synchronously to the engine angle, while failure simulation is an integral component of the
injection and ignition signals are measured simulator, and allows actions such as switching ECU
synchronously to the crank angle. Special hardware is pins against VBAT or ground and interrupting lines.
generally used for this task, for example, the DS2211
HIL I/O Board, which is in widespread use in the
automotive industry, for passenger cars and trucks as SOFTWARE
well as racing applications. The board provides the entire
I/O for an 8-cylinder engine including signal conditioning, The software for the test systems basically consists of
for example, and is cascadable to accommodate F1 two components, the software for model and I/O
10-cylinder engines either. Thanks to its modularity, definition and the user control software for the test
further I/O boards can be added, covering all analog, system.
digital and resistor channels. Regarding CAN nodes, the
simulator hardware can provide as many CAN controllers Software for Model and I/O Definition
as required.
With the dSPACE test systems, modeling and I/O
® ®
Mechanical set-up specification are based on MATLAB /Simulink , the
quasi standard for describing dynamic systems in block
The mechanical simulator setup follows a generic form. For defining the dynamic behavior of engine,
approach, thus having a standard rack encompassing transmission, and vehicle dynamics, Simulink and
®
the individual components. Among others, the cabinet Stateflow have established their market position.
comprises real-time hardware, signal conditioning, failure Extensive Simulink block libraries are available for
simulation boards, load cards, break-out-boxes (BOB), defining I/O in the same Simulink environment.
etc. dSPACE’s Real-Time Interface (RTI) is a Simulink block
library that not only allows analog and digital inputs and
outputs to be specified in Simulink. It also contains easy-
to-use blocks for the most complex situations (see
Fig.6).

Figure 5: General structure of a dSPACE Simulator

Full-Size

Signal Conditioning

Although the signal conditioning needed for many

automobile-specific signals is provided by the DS2211
HIL I/O Board mentioned above, some signals still
require special signal preparation, for example, for
broadband lambda probes, LVDT sensors, Figure 6: Simulink block library for I/O signal
thermocouples, etc. specification

22
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

One example is the angular processing unit (APU) of the The signals can either be generated synthetically (e.g.,
DS2211 HIL I/O Board. This intelligent I/O subsystem sinus function for variable reluctance sensors) or come
performs all crankshaft angle based functions. The APU from actual measurement. It is also possible to switch
receives the current engine speed from the real-time between different wave tables at run time, for example,
simulation and integrates it with the current crankshaft to simulate a faulty toothed wheel. With analog sensors,
angle, at an update rate of 4 MHz and a precision of the amplitude of the signal can also be adjusted to
0.01°. Digital or analog crankshaft and camshaft signals, engine speed behavior (Uamplitude = f(rpm)). This is
and knock signals, are then generated at the same particularly important for racing systems from Magneti
precision. Angle-synchronous capture of injection and Marelli.
ignition signals (incl. multiple injection and ignition) is
based on the APU, with the stated time and angle For defining CAN messages to read from or write to the
resolution. CAN bus, RTI also provides block libraries that allow
CAN data (DBC format) to be read in so that messages
from the test system can be coded and decoded with the
correct signal scaling.

In addition to the component models, ECU models are

also frequently implemented (“soft ECUs”) or are already
available from function development. The soft ECUs are
used in offline simulation or if real ECU prototypes are
not yet available. Central switches in the I/O model can
be used to switch back and forth between the simulated
and the real ECU. The soft ECUs can also be employed
for dynamic CAN restbus simulation, making it possible
to mix real and simulated CAN network nodes, for
example, if not all CAN nodes are available in real form.
As already mentioned, because of the high clock rates
and associated processing power requirement,
Figure 7: Crankshaft and Camshaft signal of a typical
® multiprocessor (MP) systems are often used. Here too, it
racing application (MATLAB plot)
is important to specify the entire MP system, including
interprocessor communication, from one Simulink model.
Although these hardware-related functions are RTI blocks with appropriate sensor and actuator scaling,
processed on an FPGA, the ignition sequence, and the and dynamic model behavior, therefore allow seamless
® ®
crankshaft and camshaft signals, are completely specification based on one MATLAB /Simulink model.
®
specified in Simulink, by means of wave tables (MAT The MATLAB Real-Time-Workshop (RTW) is then used
files). to generate the C code for the real-time hardware from
this model. Thus, it is generally not necessary to have a
knowledge of C programming to generate the simulator
code.

Figure 8: Graphical User Interface based on ControlDesk Layouts [2]

23
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Interactive Operation There are generally two approaches to engine models:

a) Extended mean value models in which the whole
The ControlDesk program is used for interactive dynamic engine process is regarded as a
operation of the test system. ControlDesk allows users to continuous energy flow [5]. Sampling rates of 250
generate convenient, graphical user interfaces (layouts) – 500 us are usually sufficient to simulate
with a great variety of control elements, from simple GUI Formula One engines with this approach.
elements (push-buttons, displays, radio buttons, etc.) to b) Models based on the Vibé curve, in which each
complex plotters and photorealistic graphics. Fig. 8 cylinder-selective combustion process (engine
shows examples of application-specific layouts. torque behavior) is described separately and
computed on the basis of individual injection and
In addition, ControlDesk is used to control the simulation ignition signals, and of the intake manifold
and electrical failure simulation, and to manage pressures. For common Formula One engine
interactive experiments. A macro recorder and speed sampling rates, between 50 and 100 us
programmable switch elements for convenient are needed to describe the engine torque
maneuvering between layouts round off the functionality. behavior precise enough for ECU testing.
Apart from this interactive work on the simulator using
ControlDesk, other important aspects of working with the Test Motivation in Racing
HIL test system are automated testing and (for vehicle
dynamics particularly) the animation of motion data. The requirements for the test scenario are in large part
the same in racing as for testing production vehicle
Real-Time Models ECUs. However, there are some specific aspects to be
taken into consideration.
To map the real world in the simulator, real-time models
for all the relevant vehicle components are implemented. Component Testing
For example, the engine, transmission, powertrain, and The ECU software is constantly being further developed
engine dynamics (incl. wheels) are modeled and then and modified. There are new releases every week, and
computed on the real-time hardware. The models are these have to be tested before they can be put in the
generally provided by the simulator customers, but racing vehicle. Component testing also includes tests on
sometimes commercial models are integrated, such as submodules of an ECU network before they are
ve-DYNA/en-DYNA [3] from TESIS, Munich, Germany, integrated into the overall system.
and CarSim [4] from MSC (Mechanical Simulation Network Testing
Corporation), Ann Arbor, MI, USA. Network tests are derived directly from the component
The dynamics of the engine, transmission and tests. Sufficient validation is required that the individual
powertrain must be represented precisely in order to test ECU network components add up to a functioning overall
dynamic control behavior. For example, if the gearshift system, i.e., whether communication is via CAN busses
strategy needs to be checked or optimized, the following or ArcNet functions.
items are investigated: Regression Testing
Any time one element in the ECU network is modified, all
• Environment model (track, ground surface, functions have to be tested again to ensure that
irregularities, etc.) complete functionality has been preserved.
Limp Home Testing
• Vehicle model, incl. aerodynamics (ground
The ECUs must guarantee that even if sensors and
effect) to determine wheel forces
actuators fail, the overall system will still go on working.
• Tire model (to compute slip) Suitable limp-home strategies allow the racing car to
• Powertrain model (for torque distribution in the continue moving, as much as possible, despite functional
differential) faults. And before safety-relevant functions can be
• Transmission model (gearshift dynamics and integrated into the vehicle, the possibility of function
clutch) failure has to be excluded under all circumstances. This
type of test includes investigating whether drive-by-wire
• Engine model (engine torque and speed
systems react to the failure of specific sensors (e.g., the
computed from injection and ignition during
throttle valve sensor) in the manner expected.
gearshift)
Test Automation
These are all tested in interaction with the real ECUs.
When one considers that the entire gearshift operation
usually takes only between 20 ms and 40 ms, it is Further development of the extremely complex ECU
obvious that the quality of both the model and the functionality, and very short times between releases,
simulator hardware has to fulfill very tough requirements. make it essential to automate tests. Automation also
makes it easy to implement 24/7 operation, which can
dramatically shorten development times. The racing
teams implement tests in different ways. The following
methods are used (Fig. 9):

24
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

ControlDesk Stimulus Editor

Graphical description of the test sequence using the
signal patterns, that determine, for example, the behavior
of a sensor. The Stimulus Editor provides access to
models of all sizes. It is also possible to specify the
control flow (e.g., If-Then-Else) within the test sequence.
Textual Description
An integrated script language is used to describe the test
procedure.
Test Description in Simulink
The test sequences are implemented in the Simulink
model. Items such as an ECU input are stimulated by a
test signal via standard Simulink blocks.

AutomationDesk
Tests must be systematically automated, appropriately
structured, and reused, in order to achieve the targets
regarding test depth and test width, and therefore also
quality. In the recent past, test automation applications
were frequently developed in programming languages
and were customer-specific. Unlike previous solutions,
AutomationDesk allows tests to be constructed in
graphical form, without knowledge in programming.
Further advantages include integrated test management, Figure 9: Different approaches to defining ECU tests:
a powerful mechanism for creating custom test libraries,
a) ControlDesk Stimulus Editor,
result management and integrated report generation
b) Scripting, c) Test definition in the model
(Fig. 10).

Figure 10: AutomationDesk, an integrated environment for automated testing [6]

25
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

APPLICATIONS

SYSTEM DESIGN IN RACING

There is no such thing as “the” standard racing

simulator. For component testing, the functional scope of
the simulator, and therefore also the I/O capacity, can
be much smaller. For example, a component testing
system (Fig. 11) can do without items such as ignition
and injection channels if tests are to be performed on the
transmission or powertrain.

Simulators for motorsports are precisely tailored to the

ECUs to be tested. Their size depends on factors such
as the following:

• The number of ECU pins

• Whether an ECU network is involved (e.g.,
Magneti Marelli Step10/11, McLaren MCU
300/TAG-300) or a single ECU (e.g., Magneti
Marelli Step9)
• How many additional components have to be
integrated into the overall setup
• The scope of the tests (component tests, system
tests, etc.)
• The processing requirement of the model, i.e., Figure 11: Simulator for ECU component testing
whether a multiprocessor system is needed

Figure 12: Laboratory setup of the hardware-in-the-loop system [2, 7]

26
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

SENSORS AND ACTUATORS Thermocouples

Thermocouples have a very low output voltage, which is
The sensors and actuators are implemented in off-the- provided with great precision by a special mV module.
shelf signal conditioning modules for racing simulators. The control range, e.g., 0 - 50mV, is specified in the
The simulator hardware is precisely tailored to the model.
scaling in the Simulink model. In contrast to production-
vehicle components, sensor characteristics in racing MODEL STRUCTURE
applications are captured individually. The I/O model is
parameterized via a start-up script. The racing teams To fulfill motorsports requirements with regard to
generate the major part of the Simulink model processing time, the simulators are mainly
automatically, so that they can perform actions such as multiprocessor systems. Typically, the model is split into
integrating new sensors into the simulation in a matter of the I/O part and the dynamic model, each of which has
seconds. its own processor board.

These are examples of components installed in

motorsports simulators (by analogy to the components
described in chapter dSPACE Simulator Technology):

Linear Lambda Probe

Broadband lambda probes are mostly used in
motorsports to operate the racing engines at optimum
performance (lambda appr. 0.9). The simulator is
equipped with a lambda probe module that sets the
pump current according to the lambda value provided by
the model and if necessary also varies the internal
resistance of the Nernst cell with reference to the
exhaust temperature. The characteristics are
parameterized directly in Simulink, for example, by
inserting the 1-D look-up Ip=f(lambda). In an ideal case,
the manufacturer’s specification or the developers’ own
measurement series can be transferred to the model for Figure 13: Simulink model structure for a multi-
this. processor system incl. interprocessor
communication
LVDT Sensor
To detect the positions of the clutch or to capture the
brake disk wear, racing vehicle developers frequently IGNITION AND INJECTION
use (linear variable differential transformer (LVDT)
sensors. To simulate the sensor, the LVDT module in the The DS2211 always provides angle-based ignition and
simulator reads the excitation signal (primary side) and injection pulse capture and flags the absence of pulses.
makes the output signal (one or two secondary coils) It is extremely important to ensure that any cut-off is
available to the ECU in the correct phases. The detected, because when the ECU cuts off the ignition or
deflection is placed directly in the model as the function injection, the associated status information must be
UDAC = f(s). generated as soon as possible in the simulator. The
pulse status for each cylinder is then used within the
Active Wheel Speed Sensor model, to reflect the torque drop. When there is no pulse,
Active wheel speed sensors generally have a current the DS2211 generates a pulse state of 0. Typically this
interface whose level is provided by the simulator. The information is included in the computation of the engine
frequency to be output is scaled in Simulink according to torque. (Fig. 14)
the sensor data (teeth/wheel circumference) and
converted, for example into a 7/14 mA level.

Moog Valves
The current-controlled Moog valves are read in indirectly
via voltage measurement at a shunt. In the model, the
voltage is converted into control current, which in turn is
the input for the dynamic Moog model. Usually, the
effective deflection of the actuator is computed via the
integration of the coil current (which corresponds to a
specific control speed). Here too, the transfer behavior
from coil current to control speed is included in the model Figure 14: The torque computation must take
as a look-up table. ignition cut-off into account

27
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

The resolution of the engine variables depends directly

on the sampling rate at which the Simulink model is
computed. For the best possible resolution, for example,
of the pressure behavior in the cylinder, the sampling
rate must be as low as possible.

An example: Revolutions of 18000 rpm correspond to a

change speed of the crankshaft angle of 108°/ms or
0.108°/us. If the dynamic model is computed at 1 kHz,
the resolution is only 108°. A crankshaft angle of 21.6°
can be achieved at 5 kHz, and of 5.4° at 20 kHz.
Figure 15: ControlDesk Layout showing the effective
Obviously, the real-time hardware must compute the
computing time in seconds for individual simulator
model as fast as possible to achieve optimum resolution.
Especially for engine models that compute the tasks by means of numeric displays
combustion behavior in the cylinder (Vibé curves), model
clock rates of 50 us are desirable (Fig. 15).
REAL PARTS
REAL-TIME PROFILING
There are many different reasons for integrated real
In all cases, the application as implemented must include parts into the simulator. For example, if there are
the model’s behavior over time. In addition to the submodels that are not (or not yet) available in Simulink,
specified sampling rate, the effective computing time the control loop concerned can still be closed by
required for each task is also represented. If a violation including the real part. Some components can be very
of real time is detected, it may be necessary to increase difficult to model, so engineers fall back on integrating
the sampling rate. The timing is profiled with the help of the real parts, for example, a real throttle valve, to close
ControlDesk, giving access to vital simulation time data the control loop for simulation.
such as sample rate, task turn-around time, task call
counts, task overruns, etc. It is crucial to check the With racing car components that are controlled by the
application’s behavior over time, i.e., how much driver, such as the steering wheel, pedals, or console
headroom actually is available taking into account time (rally cars), the aim is to make the simulation as realistic
consumption peaks. as possible.

Figure 16: Simulator for rally (WRC) applications with original customer components

28
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Junction Boxes USER INTERFACES

The connections in racing vehicles are generally military Specific ECU functions have their own control layouts, in
industrial cylindrical connectors (MIL). To connect the addition to the general layouts for system control (such
ECU network to the simulator, specific junction boxes as voltage supply). For example, different layouts are
are used. As a rule, the boxes are made by the teams, available for low-level functionalities such as ignition and
and they can be located inside or outside the simulator. injection and higher-level control functions such as
This makes it possible to use the original vehicle cable launch control and traction control.
harness, or parts of it, in the simulation environment.

GATEWAY OPERATION

The simulator is able to corrupt specific signals that are

exchanged between the ECUs and other peripherals in
the vehicle. This method of intervention (gateway) is
based on the principle of the simulator fetching all the
relevant signals, corrupting them in real time, and
outputting them to their recipients. It does not matter
whether the gateway is for discrete signals or for CAN
bus signals.

Figure 18: F1 simulator layout for dynamic model

data analysis
For pure channel stimulation, stimulus layouts are also
used. When a test value has been set in a layout, its
correct implementation in the ECU can immediately be
checked by a calibration tool. Fig. 19 shows a test layout
from a rally application, in which all speed signals can be
specified on one layout. These specifications can of
course also be completely automated.

Figure 17: F1 steering wheel as an example of

gateway operation

For the steering wheel, for example, analog signals from

position sensors are read in on ADC channels. These
signals can then be output again unchanged (correct
behavior), or their values can be corrupted by a test
value on the operator’s PC and then sent to the ECU via
the appropriate analog output port (incorrect behavior).
CAN data are also manipulated on the same principle.
Signals received from both the ECU and the steering
wheel are available as physical values according to CAN Figure 19: Test layout in a rally application for
scaling (offset and gain). These values can be corrupted stimulating individual channels
on the basis of physical units and then put on the bus by
the simulator. Besides changes to individual CAN signals
(such as checksums), entire messages (missing, wrong
timing), and even the complete absence of a bus node
can be simulated and their effect on the rest of the
network investigated (see [8]).

29
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

AUTOMATED STIMULUS TESTS • The vehicle must be embedded in a virtual, 3-D

landscape. For motorsports applications, the
The ECU is stimulated by the precise data recorded animation scene is given items such as a track
during an actual test drive or in test bench trials. Feeding definition that corresponds to real racetracks or
this data into the simulator brings the ECU to the the team’s real test track.
operating point near which a function fault, for example,
was detected, making it possible to investigate in detail in When these steps have been completed, the vehicle
the laboratory. dynamics can be analyzed in the virtual 3-D world in
If a Simulink model of the ECU functionality is available, MotionDesk in parallel to the simulation run. As a rule,
function tests that validate the software ECU against the real steering wheels are available on the simulator to
real hardware can also be automated. allow man-in-the-loop operation, i.e., the user can steer
the simulated vehicle, interacting with the ECU network
ANIMATION while at the same time receiving visual and acoustic
feedback (e.g., by sound generation to mimic the engine
Animating the vehicle is done with dSPACE MotionDesk, speed).
a tool which provides 3-D animation of mechanical
system in a virtual world. There are essentially three
steps involved in animation for HIL simulation:

• The motion data of the dynamic vehicle model

(roll/pitch/yaw, translation) must be linked to the
bodies to be animated (chassis, wheels, steering
wheel, etc.) via special Simulink blocks
(MotionDesk library and others) to create
kinematic chains.
• Geometries from the CAD systems of racing
teams are very complex and have a large
number of polygons (>50000). These large
polygon counts cannot be handled under real-
time conditions on a PC. Thus the original data
formats of these CAD systems must be
transferred to a data format (VRML2) that is
suited to animation.
Figure 21: 3-D visualization of a racing vehicle

Figure 20: Sequence of an automated stimulus test using track data

30
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

TEST BENCH INTEGRATION

Electric
In engine and powertrain development for motorsports,

Brake
the test benches used are generally highly dynamic and
allow transient processes during an acceleration phase
(incl. gearshifts) and during start-up to be mapped Test Bench Real Test
sufficiently well. The load definitions for such test Real Engine
Real
Differ- Bench
benches are based either on recorded track data or on Gearbox
ential Control
predefined excitation scenarios, quasi in an open loop.
However, to study the dynamic behavior of the

Electric
powertrain, including the engine and the steering

Brake
CAN
electronics, as they interact with the environment, the
highly dynamic test bench has to be connected to the Engine Transm. Diff.
Veh. Dyn.
...
Controller
real-time simulation of the environment. Controller Controller Controller
(Model)
Some examples of this are optimizing traction control in
F1 and rally vehicles, optimizing the gearshifts (up shift,

Test Bench Interface

Vehicle
down shift), and investigating the interaction between the Engine Transm. Diff.
Dynamics
Model Model Model
engine and the transmission if one of them is not Model
available. This can be the case, for example, with F1
teams that do not perform development at a single
Driver/Excitation Environment
location, unlike Toyota and Ferrari. Model Model
A transmission test bench with a real gearbox and
Real-Time Simulation
transmission ECU, a real or simulated engine controller,
and electric motors applying loads to the drive shafts of = Real = Real or Virtual = Virtual
the gearbox is an example of this (Fig. 22). The
behaviors of the engine and the vehicle are simulated in Figure 23: Test bench with real powertrain (engine,
real time, and the resulting variables, drive shaft torque, transmission, differential and corresponding con-
and speed are introduced via the electric drive and the trollers) in combination with a real-time simulation
electric brake. (HIL) environment

The test bench system and real-time simulation can be

Real Test connected via CAN bus, Interbus, or Profibus, or via
Electric Drive Electric Brake Bench
Gearbox
Control proprietary interfaces (such as a dSPACE AVL link).
This test bench configuration has the following
CAN
advantages:

•
Engine Diff. Veh. Dyn.
Controller
Transm.
Controller Controller ... All ECUs can be integrated with real
Controller
(Model) (Model) (Model)
components at an early stage. This is particularly
important if an ECU is available, but not the
component to be controlled (engine,
Test Bench Interface

Engine
Dyn.
Transm.
Differ-
ential
Vehicle
Dynamics
transmission, vehicle).
Model
Model Model Model
• The real testbed component does not need a
dynamic model, nor do the real ECUs.
Driver/Excitation
Model
Environment
Model • Tests on the interaction between real and
Real-Time Simulation
simulated components are highly reproducible.
= Real = Real or Virtual = Virtual
• Suitable parameter sets for the ECUs can be
generated even before the track test. This
makes it likely that there will be fewer problems
Figure 22: Test bench with real gearbox and trans- in transferring from the test bench to the
mission controller in combination with a real-time racetrack, and greater software quality is
achieved.
simulation (HIL) environment
Another example is shown in Fig. 23. Here the engine,
transmission and differential are mounted on the test
bench, while the vehicle dynamics, driver and the
environment are simulated on an HIL system in real time.
As in the first example above, the corresponding torque
and speed signals of the drive shafts are simulated and
applied to the powertrain by means of two electric
brakes.

31
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

SUMMARY AND OUTLOOK REFERENCES

This paper discusses the various options for using

[1] Wright, P.: Formula 1 Technology; SAE Society of
hardware-in-the-loop simulation in motorsports. Starting
Automotive Engineers, Warrendale, PA, USA, 2001.
with the necessary hardware and software components
of modern HIL test benches, concrete motorsports [2] Urban, P.; Wältermann, P.; Henking, B.: Toyota
requirements and implemented solutions are presented. Motorsport Racing Ahead with dSPACE (In German:
Finally, the integration of real-time simulation with Toyota Motorsport mit dSPACE auf der Überholspur.
concrete drive test benches and the resultant application Automotive Engineering Partners 6/2001, S. 40.
scenarios are described. However, due to obligations of English Version available via: [Link]
confidentiality, it was largely necessary to avoid detailed ftp/papers/dspace_AutEP_0111_e_f22.pdf
descriptions of concrete customer projects.
In the future, the authors see a clear trend towards a [3] TESIS DYNAware: enDYNA®, veDYNA®, Product
continued increase in the complexity of software and Description. Munich, Germany, [Link]
hardware systems in racing vehicles, as the basis for [4] Mechanical Simulation Corporation: CarSim®
further performance growth in motorsports. The Product Description. Ann Arbor, Mi, USA,
necessity of increased testing and more complex test [Link]
systems will also rise in direct proportion to this. In both
production and motorsports vehicles, determining [5] Hendricks, E. / Sorenson, S.: Mean Value Modelling
variables that cannot be measured via substitute models of Spark Ignition Engines. SAE Paper No. 900616,
in the ECUs will grow in importance. Observer systems Society of Automotive Engineers, Warrendale, PA,
such as those that are already in standard use for slip USA, 1990.
detection in vehicle dynamics control systems might also [6] Lamberg, K.; Richert, J.; Rasche, R.: A new
be used more in motorsports. Environment for Integrated Development and
As active intervention is subject to tough regulations Management of ECU Tests. Society of Automotive
(especial in Formula One), the focus is mainly on Engineers, SAE Paper No. 2003-01-1024.
existing actuators (especially the engine and
transmission). This is different in rallying, where far more [7] Urban, P.: Verification of a decentralized ECU
active intervention is allowed. system for a Formula 1 car by means if Hardware-in-
Future technical developments in Formula One remain the-Loop-Simulation (in German). Conference:
difficult to predict, as frequent new proposals on the „Hardware-in-the-Loop-Simulation in Vehicle
number of cylinders and displacement, and on the much- Development“, Haus der Technik, Essen, 2001.
discussed “standard electronics”, are made by the FIA. [8] Plöger, M.; Sauer, J.; Büdenbender, M.; Held, J.;
As regards the latter, it is still unclear whether only the Costanzo, F.; de Manes, M.; di Mare, G.; Ferrara, F.;
hardware, or the software functions too, or even software Montieri, A.: Testing Networked ECUs in a VirtualCar
parameterization, are to be standardized. It remains to Environment. Society of Automotive Engineers, SAE
be seen which way the discussion will go. Paper No. 2004-01-1724.

32
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

2004-01-3542

Coolant Flow Optimization in a Racing Cylinder Block and

Head Using CFD Analysis and Testing
Jian Ye
Managed Programs, LLC (MPI)

Jim Covey
GM Racing

Daniel D. Agnew
Managed Programs, LLC (MPI)

Copyright © 2004 SAE International

ABSTRACT realized that localized improvements and changes

helped cylinder head durability by reducing local hot
Coolant flow absorbs heat generated from engine areas. In this paper, a more recent work is reported as a
combustion, and continued efficient heat transfer is continued two-part development program conducted to
crucial in high performance racing engines to provide further optimize the coolant flow through an SB2
maximum engine power and durability. Any potentially Winston Cup engine.
stagnant flow may cause an overheated zone in the
coolant jacket resulting in structural damage. Critical The first part of this current project utilized CFD to study
areas can include the cylinder head around exhaust the coolant flow patterns through the engine [2]. Of
ports. Due to complexity of the geometries, flow particular interest were the impacts of parameters, such
patterns around those areas are not well understood. In as the location of the coolant input and the size and
this report, Computational Fluid Dynamics (CFD) location of the holes in the cylinder head gasket that had
analysis, combined with testing, is utilized to study a the greatest effect on the coolant flow through the
typical racing engine coolant jacket. The results include engine. This analysis provided the necessary insight to
coolant flow assessment in critical zones as it is affected determine the relevant influences on coolant flow, so
by cooling flow inlet, head gasket design, and flow outlet that later testing could focus on only a few very
geometries. important parameters.

INTRODUCTION Upon completion of the CFD analysis, another SB2

engine was instrumented with thermocouples, flow
The cooling system of a racing engine is an important meters, and pressure transducers. The instrumented
element of the race car system. On a vehicle level, it engine was evaluated on a dynamometer. A radiator
must manage the heat generated, primarily from engine and cooling tank were also provided so that the tests
combustion, by transferring it to heat exchangers. On could be conducted with a pressurized cooling system
the engine level, ideally, the cooling system should used to simulate the cooling system in the race car. The
remove enough heat to allow proper engine function data were collected at 7300 rpm (peak torque) during a
while evenly distributing the heat, primarily throughout “steady state” dynamometer test run with minimum
the water jacket in the block and cylinder head. This duration of ten seconds. To assure that the
optimization is often a delicate balance of engine temperatures of the components in the engine had
function (durability and performance) and vehicle stabilized, the data recorded between eight and ten
aerodynamics. The heat needs to be removed from the seconds during the steady state dynamometer test were
engine and transferred to the radiator, which in turn used for the analysis.
transfers the heat to incoming air flow. However,
incoming cooling air flow needs to be balanced against COOLING JACKET MODEL AND CFD ANALYSIS
vehicle dynamics.
Figure 1 gives the inner surface configuration of the SB2
In 1997 a project was undertaken to further optimize the coolant jacket. The main components include the inlets
coolant flow in an SB2 NASCAR racing engine. This (InB1, InB4, InB5, and InB6), gasket holes, heads and
optimization began in previous work [1] where it was outlets (Out1 L and Out6 R) as shown in Figure 1, 2, and

33
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

3. The commercially available CFD program FLUENT

was used to perform the analysis. A summary of the
boundary conditions of the various Cases 1 through 6
that were studied is listed in Table 1. Those cases
applied several different arrangements of inlet coolant
flow and cylinder head gasket designs.

INLET LOCATIONS
InB6
The total coolant flow rate into all inlets of the block was InB4
16.63 gpm (gallons per minute). The two configurations
of inlet combinations evaluated included “Side-Only” and
“Front and Side”. The “Side-Only” configuration Figure 2, SB2 Block Coolant Jacket Configuration
consisted of all of the coolant being fed into each side of
the block between the cylinders as InB4, InB5, and InB6
or just InB5 alone. The “Front and Side” configuration
incorporated feeding coolant through a 0.500 inch Cut 3 Cut 4
restrictor installed in each of the coolant pump legs InB1 Cut 1 Cut 2
(in Figures 1 and 2) and to each side of the block as
InB5.

Later calculations also included cutting sections in

pertinent areas of the coolant jacket and determining the
flow rate through these sections. This provided a means
to compare the subsequent cases. The section cuts in
the block jacket are shown in Figure 2 as “Cut 1, Cut 2,
Cut 3, and Cut 4”. These correspond to cylinders 1, 3,
5, and 7, respectively.

Similarly, section cuts in the cylinder head jacket are

shown in Figure 3 as “Cut 1, Cut 2, Cut 3, and Cut 4”. Figure 3, SB2 Cylinder Head Coolant Jacket
These correspond to the exhaust areas of cylinders 1, 3, Configuration
5, and 7, respectively.
Table 1, Summary of Boundary Conditions
Case # Case 1 Case 2 Case 3 Case 4 Case 5 Case 6
InB1 10.59 10.59
(Front) gpm gpm

InB4 5.54 gpm 5.54 gpm

(Side
Line)
InB5 6.06 gpm 5.54 gpm 5.54 gpm 16.63 16.63 6.06
(Side gpm gpm gpm
Line)
InB6 5.54 gpm 5.54 gpm
(Side
Line)
InB6 Gasket Fel Pro Fel Pro Gasket Gasket Gasket Gasket
holes 1034 1034 Design Design Design Design
Head Head #2 #3 #4 #4
InB4 Gasket Gasket

BASELINE (CASE 1 AND CASE 2)

Figure 1, SB2 Coolant Jacket Configuration, Block
and Head Initially, a Fel-Pro 1034 head gasket was used in the
CFD analysis as a baseline gasket design to evaluate
the coolant flow pattern through the engine block and
head. The location and size of the coolant holes in a
Fel-Pro 1034 gasket are shown in Figure 4. An
evaluation of the baseline gasket was conducted with
the coolant input through the “Front and Side”, Case 1,
and through the “Side Only”, Case 2.

34
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

As shown in Figures 5a and 5b, introducing the coolant

1034 Fel-Pro
into the block through the “Front and Side“, Case 1,
Coolant-In: Front & Side
results in poor distribution of the coolant through the
block. There appears to be good water flow at the front Block Coolant Jacket
of the block, as indicated in the figure as a dark area of
very high density velocity vectors. However, the water Stagnant Area
flow at the rear of the block, especially around the
inboard side of cylinders #5 and #7, is stagnant, as
indicated in the figure by very light area of low density
velocity vectors.
InB5
By feeding all of the coolant into the side of the block,
Case 2 in Figures 6a and 6b, the distribution of the
coolant flow around the cylinder bores, front and rear, InB1
was much improved. However, the coolant flow around
the inboard side of the cylinder bores is still relatively low
compared to the flow around the outboard side. In
addition, the coolant flow through the cylinder head Figure 5a, Flow Patterns in Block and Head, Case 1,
appears to be well-balanced front to rear. Tables 2 and Baseline Fel-Pro Gasket
3 list the calculated flow rates through the section cuts in
the head and block, respectively. These calculated flow
rates will be used as a basis of comparison for later 1034 Fel-Pro
analyses. Coolant-In: Front & Side
Further observation showed that the coolant flow in the Cylinder Head Jacket
water jacket could be short-circuited if the head gasket
holes are too close to the coolant inlets or outlets. This
confirmed that gasket configurations play a key role in
the distribution of flow inside coolant jacket head and
block.

Cylinder Head Bolt

Holes – 17 Places

Figure 5b, Flow Patterns in Block and Head, Case 1,

Baseline Fel-Pro Gasket

Based on the results of the CFD analysis of the baseline

Fel-Pro 1034 head gasket, along with previous coolant
flow tests [1] and experience gained from race engine
Figure 4, Fel-Pro 1034 Head Gasket teardowns, it was determined that the holes at the front
(cylinder #1) and rear (cylinder #7) ends of the gasket
were detrimental to good flow distribution. So, as a
result, three new cylinder head gasket designs were
conceived, analyzed, and then tested. As a starting
point with these new designs, to maintain comparable
flow through the head gasket, the total cross-sectional
area of the holes in all three new gasket designs #2, #3
and #4 was the same.

35
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

1034 Fel-Pro cylinder bores appears to be very stagnant. Without any

Coolant In: Side Only significant sized holes on the inboard side of the head
gasket, there is nothing to promote the coolant flow
Block Jacket around to the inboard side of the cylinder bores in the
block.
Improved Flow
Area

InB6
0.090” (3 places)

InB5

InB4

0.265” (8 places) 0.465” (3 places)

Figure 6a, Flow Patterns in Block and Head, Case 2, Figure 7, Gasket Design #2
Baseline Fel-Pro Gasket
Gasket Design #2
Coolant In: Side Only
1034 Fel-Pro Block Jacket
Coolant-In: Side Only
Very Stagnant
Cylinder Head Jacket Area
InB6

InB5

InB4

Figure 8a, Flow Patterns in Block and Head, Case 3,

Gasket Design #2
Figure 6b, Flow Patterns in Block and Head, Case 2,
Baseline Fel-Pro Gasket
Gasket Design #2
Coolant In: Side Only
Cylinder Head Jacket
CASE 3: GASKET DESIGN #2 (OUTBOARD BIASED)

The first prototype head gasket design (Design #2) is

shown in Figure 7. The intent of this gasket hole
arrangement was to direct the coolant flow up into the
cylinder head on the outboard side of the valve seats
and to promote coolant flow across the cylinder head
toward the center of the engine. As shown in the CFD
analysis (Figure 8 and Table 2), the coolant flow through
the cylinder head is well balanced and has good flow
around the valve seats.

However, as seen in Figures 8a and 8b and Table 3, the

gasket hole arrangement does not promote good coolant
flow around the inboard side of the cylinder bores in the
block. In fact, based on the very light density of flow Figure 8b, Flow Patterns in Block and Head, Case 3,
vectors, the coolant flow on the inboard side of the Gasket Design #2

36
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

CASE 4: GASKET DESIGN #3 (INBOARD BIASED) Gasket Design #3

Coolant In: Side Only
Cylinder Head Jacket
To promote better cross-block flow, the second new
head gasket design, Gasket Design #3, was developed.
As shown in Figure 9, it has one large hole located on
the inboard, or valley side, of the gasket. As shown by
the heavy density and even distribution of the flow
vectors in the CFD analysis of Figure 10, Gasket Design
#3 did promote improved coolant flow around to the
inboard, or valley side, of the cylinder bores.

As shown in Table 3, the coolant flow around the

inboard side of the cylinder bores with the Gasket
Design #3 was significantly greater than the coolant flow
around the inboard cylinder bores with the Gasket
Design #2. Although the coolant flow across or through
the cylinder head was anticipated to be potentially
worse, the CFD analysis showed that there appeared to Figure 10b, Flow Patterns in Block and Head, Case
be adequate flow, side to side and front to rear, across 4, Gasket Design #3
the head. However, the light density of the flow vectors
shown in Figures 10a and 10b and the lower flow rates CASE 5: GASKET DESIGN #4 (EVEN INBOARD AND
listed in Table 2 show that the coolant flow rate around OUTBOARD)
the valve seats in the cylinder head is slightly less than
that with Gasket Design #2. Gasket Design #4, shown in Figure 11, was intended to
establish balanced coolant flow to the inboard and
0.750” (1 place)
outboard side of the cylinder bores while maintaining
0.090” (2 places) good coolant flow through the cylinder head. As shown
in Figures 12a and 12b, the coolant flow distribution in
the block appears to be even and very similar to the flow
with Gasket Design #3, by comparing the flow vector’s
densities. However, as shown in Table 3, the average
coolant flow around the inboard cylinder bores was 31%
less. The coolant flow distribution through the cylinder
head appears to be evenly distributed and similar to
0.250” (1 place) Gasket Design #2.
0.265” (8 places) 0.125” (2 places)

Figure 9, Gasket Design #3

Gasket Design #3 0.620” (1 place) 0.090” (2 places)

Coolant In: Side Only
Block Jacket

Very Good
Flow
Distribution

0.265” (8 places)
0.350” (3 places)
InB5
Figure 11, Gasket Design #4

CASE 6: GASKET DESIGN #4 (COOLANT IN FRONT

AND SIDE)

Figure 10a, Flow Patterns in Block and Head, Case 4, Finally, as a variation on Case 5, CFD was utilized to
Gasket Design #3 determine how inputting the water in the front and into
the side of the block affects the coolant flow with Gasket
Design #4. As Figures 13a and 13b show, when the

37
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

coolant was introduced into the block from the front and
the side, the coolant flow around the cylinder bores Gasket Design #4
becomes biased toward the front. The results in Table 3 Coolant In: Front and Side
show 42% more coolant flows around the inboard side of Block Jacket
cylinders #1 and #3 compared to cylinders #5 and #7.
Stagnant Area

Gasket Design #4
Coolant In: Side Only
Block Jacket InB5
InB1

InB5 Figure 13a, Flow Patterns in Block and Head, Case 6,

Gasket Design #4, Front and Side Input

Gasket Design #4
Coolant In: Front and Side
Cylinder Head Jacket

Figure 12a, Flow Patterns in Block and Head, Case 5,

Gasket Design #4

Gasket Design #4
Coolant In: Side Only
Cylinder Head Jacket

Figure 13b, Flow Patterns in Block and Head, Case

6, Gasket Design #4, Front and Side Input

Table 2, Cylinder Head Section’s Coolant Flow Rate

CFD CYLINDER HEAD COOLANT FLOW RATE (gpm)
GASKET WATER-IN CYL. HEAD WATER JACKET CROSS-SECTION LOCATION
Figure 12b, Flow Patterns in Block and Head, Case DESIGN LOCATION #1 Exhaust #3 Exhaust #5 Exhaust #7 Exhaust AVERAGE
Fel-Pro 1034 Front & Side 1.03 0.53 1.96 2.41 1.48
5, Gasket Design #4 Fel-Pro 1034 Side Only 1.98 1.46 0.79 1.43 1.42
Gasket #2 Side Only 3.51 1.02 1.15 3.73 2.35
Gasket #3 Side Only 2.91 1.51 1.29 2.85 2.14
Gasket #4 Side Only 3.44 1.75 1.66 3.42 2.57
Gasket #4 Front & Side 3.39 1.43 1.27 3.37 2.37

Table 3, Block Section’s Coolant Flow Rate

CFD IN-BOARD BLOCK COOLANT FLOW RATE (gpm)
GASKET WATER-IN BLOCK W ATER JACKET CROSS-SECTION LOCATION
DESIGN LOCATION Cylinder #1 Cylinder #3 Cylinder #5 Cylinder #7 AVERAGE
Fel-Pro 1034 Front & Side 3.17 2.64 0.54 0.03 1.60
Fel-Pro 1034 Side Only 1.31 0.8 1.23 1.73 1.27
Gasket #2 Side Only 0.11 0.05 0.02 0.04 0.06
Gasket #3 Side Only 5.15 5.02 5.81 5.94 5.48
Gasket #4 Side Only 3.6 3.51 3.93 4.03 3.77
Gasket #4 Front & Side 3.58 5.48 2.6 2.7 3.59

38
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

DYNAMOMETER ENGINE TESTING

GASKET DESIGN #3 VS #4 (INBOARD BIASED VS
Dynamometer tests were conducted to evaluate the EVEN INBOARD AND OUTBOARD)
performance of each of the three new gasket designs.
The temperatures of the cylinder bore walls and valve Figure 15 shows that, on average, the cylinder bore wall
seats were recorded. The coolant system was temperatures were 5 degF hotter with Gasket Design #4
pressurized and maintained at approximately 25-28 psi than with Gasket Design #3. Unexpectedly, the valve
in the block, cylinder head, and valley plate. The seat temperatures with Gasket Design #3 were on
thermocouples used to monitor the temperature of the average 7.4 degF cooler than with Gasket Design #4.
exterior surface of the cylinder bore walls were installed The results of the valve seat temperature differential
on the outboard side of the end cylinders (#1, #7 and #2, contradict the expectations from CFD results that
#8) and the inboard side of the middle cylinders (#3, #5 predicted an average of 17% greater coolant flow
and #4, #6). The thermocouples used to monitor the around the outboard sides of the exhaust valve seats.
temperature of each of the exhaust valve seats were
installed 0.250” back from the machined valve seat Valve Seat Temperature Differential
insert counter bores in the cylinder head. Gasket #3 vs. #4

250.0

GASKET DESIGN #2 VS #3 (OUTBOARD VS 241

244
241.5 #3
238.1
INBOARD BIASED) 240.0
234.3 234.7 234.0 #4
230.7
230.0 226

Temperature (degF)
As shown in Figure 14, the improved flow and 220.0
219.7

distribution likely drove average valve seat temperatures

with Gasket Design #2 down by 8.5 degF cooler than 210.0

with Gasket Design #3. The inboard cylinder bore wall 200.0

temperatures of cylinders #3 and #5 were on average

28.5 degF cooler with Gasket Design #3 than with 190.0

Gasket Design #2. On average, all four of the cylinder 180.0

bore walls were 20.3 degF cooler with Gasket Design Cyl. #2 Exh Seat Cyl. #4 Exh Seat Cyl. #6 Exh Seat
Thermocouple Location
Cyl. #8 Exh Seat AVERAGE

#3. This correlates well with the expectations from the

CFD flow analysis. Cylinder Bore Wall Temperature Differential
Gasket #3 vs. #4

240.0
Valve Seat Temperature Differential
Gasket #2 vs. #3
230
230.0 #3
250.0
222.7
243.5
242 #2 #4 220
Temperature (degF)

238.5 220.0
240.0
232.3 233.3 232.3 #3
207.5 210.0
230.0 210.0
225.5
Temperature (degF)

205.0
217.7
220.0 200.0

210.0 191.0 190.5

190.0

200.0
180.0
Cyl. #2 Outboard Cyl. #4 Inboard Cyl. #6 Inboard Cyl. #8 Outboard
190.0
Thermocouple Location

180.0
Cyl. #1 Exh Seat Cyl. #3 Exh Seat Cyl. #5 Exh Seat Cyl. #7 Exh Seat
Thermocouple Location Figure 15, Temperature Test Data (Gasket #3 vs #4)
Cylinder Bore Wall Temperature Differential
Gasket #2 vs. #3 GASKET DESIGN #3 – COOLANT-IN: SIDE ONLY VS.
FRONT AND SIDE
250.0
246.7
243.7
#2
240.0 This next phase of the testing involved introducing the
#3 coolant into the front of the block through 0.500 inch
230.0
diameter restrictors in the coolant pump legs in addition
Temperature (deg)

220.0 216.5
217
to feeding the coolant to the side of the block. While the
210.0
212.0
207.0 207 average bore wall temperatures were the same, feeding
the coolant into the front and side reduced the inboard
200.0
temperatures of cylinders #3 and #5 by 8 degF and 4
190.0
187.5 degF, respectively, while increasing the bore wall
temperature of cylinder #7 by 9 degF. In addition, the
180.0
Cyl. #1 Outboard Cyl. #3 Inboard Cyl. #5 Inboard Cyl. #7 Outboard
siamesed bore walls between cylinders #1 and #3 and
Thermocouple Location cylinders #5 and #7 were about 10 degF cooler when
Figure 14, Temperature Test Data (Gasket #2 vs #3) the coolant was input from the side only. The average

39
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

valve seat temperature was 7.5 degF cooler as a result Valve Seat Temperature Differential
of introducing the water into the block on the side only, Gasket #4
WATER IN - Side Only vs. Front and Side
as illustrated in Figure 16. 260
255
253 252.5
250 Side Only 245.1
Valve Seat Temperature Differential 244
241.5
241
Gasket #3 240 Frt&Side 238.1

Temperature (degF)
WATER IN - Side Only vs. Front and Side
260
252 Side Only 230 226
250
250 220
245 243.5
242
Frt&Side 220
238.5
240
232.5 210
Temperature (degF)

230 225.5
200
220

190
210
Cyl. #2 Exh Seat Cyl. #4 Exh Seat Cyl. #6 Exh Seat Cyl. #8 Exh Seat AVERAGE
Thermocouple Location
200

190 Bore Wall Temperature Differential

Gasket #4
180 WATER IN - Side Only vs. Front and Side
Cyl. #1 Exh Seat Cyl. #3 Exh Seat Cyl. #5 Exh Seat Cyl. #7 Exh Seat
250
Thermocouple Location

Cylinder Bore Wall Temperature Differential

240
Side Only
230
Gasket #3 230 Frt&Side 226.5

Temperature (degF)
WATER INLET - Side Only vs. Front and Side
250
220
220

240 212
Side Only 210
207.5

230
Frt&Side 200
Temperature (degF)

200

220 216.5 217 216 190.5 191.5

190
213
208.5
210
207
180
Cyl. #2 Outboard Cyl. #4 Inboard Cyl. #6 Inboard Cyl. #8 Outboard
200 Thermocouple Location
191
187.5
190

Figure 17, Temperature Test Data (Gasket #4 with

different coolant inputs)
180
Cyl. #1 Outboard Cyl. #3 Inboard Cyl. #5 Inboard Cyl. #7 Outboard
Thermocouple Location

Figure 16, Temperature Test Data (Gasket #3 with SYMMETRICAL VS. BIASED VALLEY PLATE GASKET
different coolant inputs)
In previous SB2 coolant flow tests [1] using the Fel-Pro
1034 Head Gasket and feeding the block from the front
GASKET DESIGN #4 – COOLANT-IN: SIDE ONLY VS. coolant pump and a “size –10” hose line, it was
FRONT AND SIDE determined that in order to balance the valve seat
temperatures from the front to the rear of the engine, the
Illustrated in the graphs of Figure 17, as in the previous hole sizes in the valley plate gasket had to be biased
test, introducing the coolant into the front and side from front to rear. It was found that balanced valve seat
resulted in reduced inboard bore wall temperatures for temperatures from the front to the back cylinders could
cylinders #4 and #6 by 7.5 degF and 18 degF, be achieved with a valley plate gasket that had a front
respectively, but increased the bore wall temperature of hole 44% smaller than the rear hole.
cylinder #8 by 6.5 degF. This trend was expected from
the CFD analysis, which showed increased flow around However, when feeding the coolant through the water
the inboard side of cylinder #4 and reduced flow around pump and a “size –10” hose side line with either the
cylinder #8. However, the very significant reduction in Gasket Design #3 or #4, installing a biased valley did not
the temperature on the inboard bore wall of cylinder #6 have an impact on the temperature distribution of the
cannot be explained at this point in the testing. By valve seats from the front to the rear of the head. In fact,
feeding the coolant into the front and side of the block, using a biased valley plate gasket had no impact on the
the average valve seat temperatures increased by 7 overall temperature distribution in the block or the
degF. cylinder head.

CONCLUSIONS

Based on the results of this work, the following

conclusions were determined.

40
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

1. In general, CFD is an effective tool to predict engine performance and durability and possible structural
cooling system flow distribution performance. damage.
2. Cylinder head gaskets are a very effective engine
component that can be used for optimizing cooling ACKNOWLED GMENTS
flow distribution in both the block and head.
3. Specifically in this case, Gasket Design #2 provided The authors greatly appreciate the help of Andy Petree
better cooling of the exhaust seats in the cylinder Racing and Richard Childress Racing for their support,
head than Gasket Designs #3 or #4, but, Gasket particularly in the testing phases of this study.
Design #3 provides better cooling of the cylinder
bore walls than Gasket Designs #2 or #4. REFERENCES
4. Furthermore, introducing the coolant entirely into the
side of the block results in more even distribution 1. Covey, J.A.., Kiziah, C.B., Wilson, S.G., “The
than introducing the coolant into the side and front Optimization of the Coolant System Flow in a
together. NASCAR Winston Cup Engine”, SAE Paper 983024,
5. Eliminating the coolant holes from the ends of the 1998.
gaskets in front of cylinder #1 and rearward of
cylinder #7, in addition to reducing the number of 2. Covey, J.A., “2002 SB2 Coolant Flow Evaluation
potential leak paths, promotes improved coolant flow Report”, GM report, June, 2002.
around the cylinder bore walls.
6. The use of the biased valley plate gasket to improve 3. Managed Programs, LLC, Consultant Series Report
coolant flow distribution in the cylinder head is not to GM Racing, 2001 ~ 2003
required with any of these new head gasket designs.
7. Improper design of the flow paths in the coolant
system of a race engine can result in significant
overheating of engine parts with likely limitations on

41
Downloaded from SAE International by Purdue University, Monday, October 29, 2018
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Improvement of a High-Performance CNG-Engine 2011-24-0140

Published
Based on an innovative Virtual Development Process 09/11/2011

Marco Chiodi, Alessandro Ferrari, Oliver Mack, and Michael Bargende

FKFS

Donatus Wichelhaus
VW Motorsport

Copyright © 2011 SAE International

doi:10.4271/2011-24-0140

homogenization of the air-fuel mixture. This process is more

ABSTRACT critical because even high gas velocities at the injector nozzle
Methane as an alternative fuel in motorsports? Actually this cause a very low fuel penetration. Therefore, mixture
solution is well known for the reduction of CO2-emissions homogenization or stratification depends much more on
but apparently it does not really awake race feelings. At the charge motion as usual for liquid fuels. For this reason the
2009 edition of the 24-hour endurance race on the design of the intake system and the combustion chamber is a
Nürburgring the Volkswagen Motorsport GmbH, in addition crucial step for the optimization of a CNG engine.
to vehicles powered by gasoline engines, introduced two
vehicles powered by innovative turbo-charged CNG engines In this paper the engine development process has been
for the first time. The aim was to prove, that also an performed mainly in a virtual context. The implementation of
“environment-friendly” concept is able to provide the an innovative 3D-CFD-tool (QuickSim) that has permitted
required efficiency, dynamic and reliability for a successful full-engine simulations of this turbocharged CNG race-engine
participation in motorsports. After the success in the 2009 has allowed, within short time (few months), to remarkably
edition the engagement has been continued also in 2010, this increase the engine performance. The virtual engine
time exclusively with CNG-vehicles. development process has started with the 3D-CFD-analysis of
the fluid motion of the basic engine. Based on this analysis
Focusing on the CO2-emission, reclusively the higher many engine-design modifications have then been virtually
tested, so that at the end only a few promising solutions have
hydrogen content of methane which represents the main
been “concretely” realized and tested at the test bench. Since
component of NG leads to a CO2-reduction during the
the results at the test bench have finally confirmed the
combustion of about 20% compared to gasoline. Thanks to
expectations from the simulation results, following this
the laminar burning speed of methane which is approximately
procedure it has been definitively possible to speed up the
maximal for a stoichiometric mixture, CNG engines do not
engine development process even by limiting the budget.
require a mixture enrichment at WOT operating conditions,
so that the fuel consumption decreases. In addition the very
high knock-resistance of natural gas allows a further INTRODUCTION
efficiency increasing by using a higher compression ratio. Since the internal combustion engine was invented to satisfy
Conclusively the CO2-reduction of the CNG-version ranges the need of mobility and flexibility of the people, a constant
from ca. 30% using natural gas up to 80% for bio-gas. On the development of performance and efficiency was necessary
other hand gas injection in the intake manifold causes a loss and required by the market [1,2,3]. In this development
of charge due to both the low mass density of natural gas and process a very high contribution was given from motorsports,
the absence of heat of vaporization. The latter also produces a which adopted, since the beginning, the internal combustion
temperature level of the exhaust gas at the turbine which is engine as the best suitable power unit for racing. The human
more critical. Another drawback of CNG-engines is the passion for challenges broke through very fast in Europe and

43
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

America and car competitions became very popular. and will probably never exist. But there is another more
Motorsports is ever since then one of the most important realistic way represented by a selective coupling of different
boost-factors for engine and vehicle development through simulation tools depending on the different targeted tasks of
ambitious technical innovations that can be tested at their the engine development. E.g. focusing on the engine
limits. Nowadays Motorsports meet a new challenge. operating cycle, by means of the real working-process
Because of its role as a commercial promoter for every car analysis (thermodynamic approach), a full 1D-flow-model
manufacturer, and because the internal combustion engine is and 3D-CFD-simulations (full analysis of the fluid motion) it
still the best power unit for Motorsports, a trend is faced is nowadays possible to gain a deep insight of the
which aims to reduce the technical differences between race thermodynamic and chemical processes that govern engine
and commercial car engines. The use of efficient and performance and emissions. The combination of these
powerful Diesel engines in long distance races [1,2,4,5,6,7] simulation tools acts not only as a “virtual eye” within an
or low CO2 emitting fuels like CNG are some examples. existing engine but also this approach has turned into a
Downsizing, turbo charging or direct injection are keywords valuable way towards virtual development, i.e. as approach
that car buyers and motorsports followers take care of when beyond the test bench.
they recognize that these technologies can win racing
competitions. This means that typical technologies as well as A NEW APPROACH: THE 3D-CFD-
the development methods between commercial engine
manufacturing and Motorsports are getting more and more
TOOL “QUICKSIM”
similar, so that simulation procedures and a more scientific The Motivations
approach for performance development, which typically had
The 3D-CFD-simulation represents undeniably the most
a smaller role in Motorsports, are gaining importance. Two
detailed approach for the investigation of the engine
main conditions are required: the scientific approach has to fit
operating cycle. From its basic concept this approach should
the development velocity and the budget that are mandatory
allow an unlimited predictability of engine processes by
in Motorsports.
unrestricted varying of any parameter of both the engine
setting and the operating condition. Depending on the chosen
TOWARDS VIRTUAL degree of mesh discretization it is possible to fully record the
DEVELOPMENT OF INTERNAL fluid motion and any chemical and thermodynamic
phenomena acting in any part of the 3D-CFD-domain up to a
COMBUSTION ENGINES length scale which is not far away from molecular
In the engine development process, during the years, dimensions. But this is the theory, the practice looks quite
simulation and predictive programs have continuously gained different. The computational resources in terms of the
in reliance but mainly in the last two decades the rapid required hardware and the related CPU-time are very often
increase of computer performance has boosted their prohibitive. In addition the setting of initial and boundary
utilizability. At the present time the application of simulation conditions can easily introduce sources of inaccuracy that
programs towards a reliable “virtual engine development” irremediably compromises the overall quality of the results
represents one of the greatest challenges [1]. The role of the and drastically reduces the benefits of such resource
simulation in the development process of internal combustion investments. Another critical point is represented by few
engines and the process of virtual engine development are three-dimensional engine process models that due to the lack
still not well defined. Thus engine manufacturers have quite of phenomena understanding at the fundamental physical
different expectations regarding the considered aspect. For level, inaccurate mathematical formulations, numerical
developers the emphasis is most certainly towards the gain of dependencies on the mesh structure, ambiguous validation
knowledge. Designers aim to exploit the possibilities by processes, etc., are not able to ensure a high level of
means of calculations, but also the limits of their design ideas reliability in reproducing and predicting the requested engine
even during the phase of concept finding. Test engineers processes.
desire to find explanations for measured phenomena and to
be inspired to seek improvements. The management, on the Brief Tool Description
other hand, appreciates rather the reduction of the technical Since 1998 the 3D-CFD-tool QuickSim is being developed at
risk involved with a new development, a possible reduction the FKFS in cooperation with the University of Stuttgart.
of development time and investment (“engine test stands are This tool aims to take advantage of the potentiality of the
more costly than workstations”) and last but not least traditional 3D-CFD-approach combined with a remarkable
reductions in both the development budget, as well as reduction of the above mentioned drawbacks so that a major
subsequently the manufacturing costs during production. contribution to the engine development process can be
ensured. QuickSim aims to have a higher integration level to
An all-embracing tool for a reliable, detailed and predictive the existent engine development processes. This step requires
simulation of an internal combustion engine does not exist the possibility to analyze the same systems simulated in the

44
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Figure 1. The results of a 3D-CFD-simulation with QuickSim as a virtual test-bench.

1D-CFD or measured at the test bench: from a single-cylinder maps, which are for sure very impressive and allow a detailed
analysis e.g. for combustion process development up to the visualization of the engine processes, but very often do not
full engine for global investigation and performance permit an objective evaluation of the engine behavior or, e.g.,
development as required in motorsports. These are the main a quantification of the potentiality that can be exploited from
peculiarities of this tool (more details in [1,2,4]): a particular engine solution. For this reason the evaluation
tool and the integrated real working process analysis deliver
Optimized 3D-CFD-Tool: The tool is dedicated and parts of the results (diagrams and protocols) that are
optimized for the simulation of internal combustion engines equivalent to the output of a test bench with a modern
(gasoline, diesel, CNG and other alternative fuels). indicating system. By means of this procedure (the “virtual
test-bench”) the results are familiar to any engineer involved
Fast response: Thanks to improved or newly developed 3D- in the engine design process. In this context the term “virtual
CFD-models, an efficient and reliable calculation is ensured, test bench” does not mean a stand-alone engine development
also using coarse 3D-CFD-meshes (ca. 30,000 cells for the without experimental support. It means more the capability of
combustion chamber). Thus the CPU-time can be reduced up the tool to act as a “virtual eye” inside the engine, permitting
to a factor 100 in comparison to traditional 3D-CFD- to visualize, measure and record also processes that are
simulations. hardly to be analyzed - if at all - at the experimental test
bench (e.g. mixture formation, fluid motion, turbulence,
No limitations: Different engine designs, fuel injection or residual gas distribution, etc.).
valve motion strategies, etc. can be simulated with accuracy.
THE RACE: 24H OF THE
Integrated and automatic evaluation tool: This guarantees a
clear and comprehensive analysis of the engineering- NÜRBURGRING
scientifically relevant engine parameters with diagrams, At the 2009 edition of the 24-hour endurance race on the
images or animations. Nuerburgring the Volkswagen Motorsport GmbH, in addition
to three “Scirocco GT24” vehicles powered by gasoline
Flexibility: The appropriate flow-field domain extension can engines, used two “Scirocco” vehicles powered by innovative
be chosen in order to meet the technical, temporal and CNG engines for the first time. This wanted to proof, that
financial requests of the project. also an “environment-friendly” concept is able to deliver the
required efficiency, dynamic and reliability for a successful
In addition the “internal coupling” with the real working- participation in Motorsports. After this successful
process analysis (WP) allows both a supported analysis of the participation (double win in the class for cars with alternative
engine processes (e.g. self-calibration of the thermodynamic power) VW continued its engagement also in the 2010
cycle) and a better comparison and control with test bench edition, this time exclusively with five Scirocco CNG
results and other simulation tools. As shown in Figure 1 the vehicles.
results are delivered not only in form of colored flow field

45
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Table 1. Vehicle specifications: VW Scirocco GT 24 (GDI-engine) vs. GT 24-CNG.

The long racetrack of the Nuerburgring circuit (the efficiency) [5,7,9,10,12]. Moreover the laminar flame speed
Nordschleife) with a distance of 25.378 km with 33 left- is maximal approximately at the stoichiometric value, so that
handers, 40 right-handers and 300 meters of difference in no fuel enrichment is necessary [13]. Conclusively the CO2-
altitude is a very hard track. Particularly hard for the engine reduction of the CNG-version ranges from ca. 30% using
are the fast sections of the track: Kesselchen (1.7 km) and natural gas up to 80% for bio-gas. On the other hand this kind
Döttinger Höhe (2.7 km) which require the longest engine of fuel introduces limitations that are not crucial in case of
full-load operating conditions (special requisites are needed liquid fuels: the need, for racing purpose, of high fuel volume
for the design of the cooling system). Because of the track flow through the injectors is only possible with manifold
characteristics engine efficiency and reliability are crucial injection (because, among other things, the injection time is
factors. GT vehicles of different categories are taking part too long for direct injection), which has the consequence of
and are partitioned in different performance groups regulated drastically reducing the engine volumetric efficiency because
by the use of an air restrictor. The VW-Sciroccos used for the of the gas expansion process into the airbox. Moreover, the
race were equipped by regulations with a 38 mm restriction absence of heat of vaporization and the low lubricating
diameter and a boost pressure regulator at 2400 mbar. capability of gaseous fuel increase the thermal stresses of
mechanical parts and the temperature of the exhaust gases
THE BASE ENGINE FROM MASS (e.g. turbine, valve seats).
PRODUCTION: 2.0L TURBO-
ENGINE MODIFICATION: FROM
CHARGED DIRECT-INJECTED
GASOLINE TO CNG
GASOLINE ENGINE
As introduced before the CNG engine is able to run with a
The turbo-charged passenger car production engine VW-TSI higher compression ratio and consequently higher maximal
2.0 1 (4 cylinders - 16 valves) equipped with gasoline direct pressure in the cylinder, that easily reaches the maximal
injection has been used by Volkswagen Motorsport as base structural pressure of the cylinder head (140 bar - see Table
for different race applications. The first application of this 1). Here the piston shape has been modified with a top part
engine was the development of a gasoline version (GT24) that allows increasing the compression ratio (from ε=10 of
and since 2009 a monovalent CNG with the aim to participate the gasoline version to ε=12). Some mechanical parts have
in different race categories and championships. Among them, been reinforced or modified in order to better stand the higher
the 24 h of the Nuerburgring 2009 was the first relevant thermal and structural stresses (in particular the crank-shaft,
competition where one engine manufacturer brought two the connecting rod and the valve seats).
kinds of fuel motorization (gasoline and CNG) on the same
car (VW-Scirocco). Interesting is the fact, that the overall The most remarkable modification concerns the injection
performances were comparable and that the CNG engine system. Here the CNG engine has been converted from
performed lap times that did not disfigure against the sister gasoline direct injection to gas manifold injection (the hole
gasoline car. In Table 1 the characteristics of both engines are left by the injector in the combustion chamber has been
listed: closed). Due to the high demand of fuel even two gas
injectors for each cylinder (see Figure 4) are required for
The key-factors for exploiting the CNG at its working limits ensuring the fuel injection within a reasonable period of time.
for engine performance development is the understanding of
the thermo-dynamical characteristics during the engine Regarding the turbocharger the main differences are related
working cycle. With a RON of approximately 130, the to the exhaust gas temperature of the CNG engine, that is
natural gas is much less knock sensitive than gasoline, thus, if higher than in case of the gasoline engine. This is in
allowed, a higher compression ratio, boost pressure and particular due to the missing cooling effect of the gas and the
maximum cylinder pressure are possible (higher indicated

46
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

need to contain the maximal cylinder pressure by postponing Investigations at the Test Bench and Problem
the combustion (increasing of exhaust enthalpy flow). Since Recognition
the CNG engine has a very high level of exhaust enthalpy
Investigations at a test bench equipped with a full indicating
flow, even at the same mass of intake air, it requires a smaller
system for high and low pressure analysis in each cylinder
turbine.
have shown extremely high cycle-to-cycle variations (high
σ_imep) of the cylinders 2 and 3 (engine development-step
THE SIMULATION PROCEDURE 0). Since the maximal allowed pressure in cylinder, due to
The race engines have been developed with a combination of structural design of the cylinder head, is limited to 140 bar, it
experimental investigations at the test bench and simulation becomes evident that the cylinders 2 and 3 are not able to
analyses based on 1D-CFD and 3D-CFD simulations. The fully exploit their thermodynamic potential.
1D-CFD-simulations have been performed using the GT-
Power code coupled with the FKFS-User-Cylinder, which, The measurements, of course, are able to suggest mixture
among other things, allows a more accurate description of the formation problems as the causes of the behavior of the
working fluid and a predictive combustion analysis taking, central cylinders, but an accurate analysis of the engine
among other things, residual gas concentration and turbulence processes towards improvements cannot be performed.
into account [8,11]. The analyses based on the 1D-CFD- Therefore, like in this case, only the 3D-CFD-simulation is
simulations have been reported in another SAE-article edited able to find explanations and answers to this problem, so that
in 2011. The focus in this paper is exclusively on the 3D- successive engine development steps can meet
CFD-analyses using the before mentioned tool QuickSim. improvements. In this paper a selection of these steps are
Since this tool is able to perform reliable 3D-CFD- reported.
simulations within a short CPU-time (e.g. the simulation of
one cylinder of a gasoline engine over a whole operating Setting of the 3D-CFD-Simulation
cycle takes about two hours on a single processor-computer) Different extensions of the discretized 3D-CFD-domain have
this advantage can be used for the simulation of several been tested. These range from the 3D-CFD-simulation from
successive operating cycles (reduction up to completely one cylinder alone (Figure 2) up the simulation of the full
elimination of the influence of the initial conditions) and the engine (Figure 4). As shown in Table 2 both the CPU-time
extension of the simulated 3D-CFD-domain up to the full and the computer requirements (RAM and hard-disk) rises
engine (increasing of predictability and reduction of the with increasing the 3D-CFD-domain extension but by using
influence of boundary conditions). Using this simulation QuickSim still remain in an acceptable range. As introduced
procedure both the general reliability and the virtual before the simulation (with combustion) of several successive
development capability of the simulation can be remarkably operating cycles is required for the convergence of the results
increased meeting the tight timetable in Motorsports. through the cycles [1].

VIRTUAL DEVELOPMENT OF A
HIGH-PERFORMANCE CNG-ENGINE
ANALYSIS OF THE ENGINE
BEHAVIOR
As introduced before natural gas as fuel represents a very
innovative and promising solution towards decreasing of
CO2-emissions. This can be achieved also with remarkable
engine performances (e.g. race-engines). On the other hand,
in comparison to a gasoline engine, the homogenization of
the air-fuel mixture is more critical because even high gas
velocities at the injector nozzle cause a very low fuel
penetration (low mass density of gaseous fuels). Therefore, Figure 2. 3D-CFD-domain of the cylinder excluding the
mixture homogenization or stratification depends much more fuel injectors.
on charge motion as usual for liquid fuels. For this reason the
design of the intake system is a crucial step for the
Simulations with the 3D-CFD-domain limited to the cylinder
optimization of a CNG engine that must be accurately
alone are not able to reproduce the interactions between the
investigated, otherwise high cylinder-to-cylinder and cycle-
airbox and the cylinder. Moreover especially the location of
to-cycle variations compromise the engine output.
the intake ID-boundary conditions remarkably influences the

47
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Table 2. Vehicle specifications: VW Scirocco GT 24 (GDI-engine) vs. GT 24-CNG.

mixture formation process and the flow motion within the Results (Engine Development-Step 0)
cylinder [1]. The goal of a homogeneous fuel-air mixture in CNG-engines
is a very complex task. The fuel injectors are able of a very
In comparison to the simulation of the cylinder alone the low radial penetration and even an injection velocity of
extension of the 3D-CFD-domain to the airbox (Figure 3) approx. 400 m/s generates a poor axial penetration. Usually
allows a remarkable increasing of result accuracy, but after an axial penetration of ca. 25 mm the fuel does not have
indisputably the most comprehensive and accurate results a relative velocity to the background fluid anymore. From
have been performed using the full-engine variant (Figure 4). this distance the fuel can be assumed to be transported by the
This approach reduces the influence of the 1D-CFD- motion of the charge [12]. On the next pages some results of
boundary-conditions and makes evident the interaction the full-engine simulation at 5500 rpm WOT are shown. The
among injector-position, airbox design and all the cylinders. diagrams are provided automatically by the evaluation tool of
In these simulations the fluid motion, mixture formation and QuickSim. Due to secrecy often the variable profiles are
combustion can be clearly analyzed. normalized to the maximum value of cylinder 1. Starting
from the analysis of the fluid motion, e.g. the mass flows
through the intake and exhaust valves (see Figures 5 and 6), it
becomes evident how the airbox design influences the intake
phase of each cylinder. The external cylinders 1 and 4 have a
relatively sharp peak while the central cylinders 2 and 3 show
a more flat maximum. This behavior is mainly due to the
small volume of the airbox and the central position of the
throttle manifold (see Figure 4), that having a relatively
constant mass flow reduces the pressure waves in the facing
cylinders, i.e. the central ones.

Figure 3. 3D-CFD-domain of one cylinder with the

airbox.

Figure 5. (Step 0) - Mass flow through the intake-valves -

5500 rpm WOT - 5th cycle.

Figure 4. 3D-CFD-domain of the full engine.

48
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

flows are evaluated. Therefore, due to the low density of the

gas and the wide distance of the gas injectors from the valves,
a small asymmetry in the intake process generates a
considerable deflection of the fuel jets that increases the fuel
concentration in the intake channel of valve 1 in case of
cylinder 3 (see Figure 8) and the opposite for cylinder 2.

Figure 6. (Step 0) - Mass flow through the exhaust-

valves - 5500 rpm WOT - 5th cycle.

Figure 8. (Step 0) - Fuel mass fraction in the intake

channels of cyl. 3 (section 5mm from valve seats) - 60
deg after TDC (cylinder relative) - 5500 rpm - WOT (full
engine mesh - 5th cycle).

Figure 9. (Step 0) - Mass flow through each intake-valve

of cyl. 3 - 5500 rpm WOT - 5th cycle.

The result of the mixture formation process at the ignition

point (IP) is shown in Figure 10 (lambda-maps). Here it
becomes evident that the central cylinders have an inadequate
homogenization. In particular cycle to cycle variations of
lambda at the spark plug (from lean to a rich mixture and
back) are responsible for remarkable variations of the
combustion duration and consequently remarkable variations
Figure 7. (Step 0) - Fuel mass fraction and flow field of the indicated mean effective pressure (high σ_imep).
during the intake phase of each cylinder - 90 deg after
TDC (cylinder relative) - 5500 rpm - WOT (full engine
mesh - 5th cycle).

The flow through the throttle manifold causes also a more

asymmetric mixture formation in the intake runners of the
central cylinders due to a major deflection of the fuel jets (see
Figures 7, 8 and 9). Here it is interesting to notice (see Figure
9) that the total mass flows through the intake valves (e.g. of
the cylinder 3) is minimally higher for IV1, but the
differences are remarkably amplified if only the fuel mass

49
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Figure 12. (Step 0) - (TKE) Turbulence at the ignition

Figure 10. (Step 0) - Lambda distribution at the ignition point of each cylinder - IP=25 deg before FTDC
point of each cylinder - IP=25 deg before FTDC (cylinder relative) - 5500 rpm - WOT (full engine mesh -
(cylinder relative) - 5500 rpm - WOT (full engine mesh - 5th cycle).
5th cycle).

The results of the 3D-CFD-simulations (see Figure 11) show

also a reduction of the tumble ratio in the central cylinders.
Since the breakdown of the tumble directly influences the
turbulence level within the combustion chamber (overall and
local at the spark plug - see Figures 12, 13, 14) it follows a
less beneficial turbulence profile for the combustion process.
The turbulence maps at the ignition point of each cylinder
(IP), in comparison to the fuel distribution, show a good
symmetry of the turbulence in the central cylinders between Figure 13. (Step 0) - (TKE) - Cylinder turbulence - 5500
intake valve 1 and intake valve 2. rpm WOT - 5th cycle.

Figure 11. (Step 0) - Tumble ratio - 5500 rpm WOT -5th Figure 14. (Step 0) - (TKE) - Spark plug turbulence -
cycle. 5500 rpm WOT - 5th cycle.

50
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Focusing now on the combustion process, the temperature

distributions in Figure 15 show the flame positions at CA=10
deg after FTDC (burned mass fraction ca. 35%). The flames
in the external cylinders remain in the central position of the
combustion chamber and still have a circular shape. In
contrast the centers of the burned zones in the internal
cylinders have moved near the exhaust valves and have a less
circular shape (indication of a more sensitive combustion).
The analysis of the combustion profile reported in Figure 16
shows at the end also a longer combustion duration of the
central cylinders. The combination of a lower turbulence
level and a worse mixture formation in the central cylinders Figure 16. (Step 0) - Burn rate profile - 5500 rpm WOT
cause definitively a loss of performance of the engine. -5th cycle.

Looking in general for improvement of a turbo-charged race-

engine many factors have to be taken into account.
Depending on the race track characteristic, race duration,
vehicle characteristics etc. not only the maximal power
output is a relevant parameter but also the torque profile
within the engine speed-range of interest and the engine
drivability (engine performance during transients) are very
important.

Within the given engine limitations fixed by regulations

(airbox pressure, max. air consumption) or structural reasons
(max. allowed cylinder pressure, exhaust gas temperature,
etc.) the engine must be developed finding the “best”
compromise among cylinder filling, turbulence (reduction of
combustion duration towards increasing of indicated
efficiency), charge exchange optimization (reduction of work
losses during the exchange phase) and mixture
homogenization within the engine speed-range of interest
(see Figure 17).

Figure 15. (Step 0) - Temperature distribution during

combustion of each cylinder - 10 deg after FTDC
(cylinder relative) - 5500 rpm - WOT (full engine mesh -
5th cycle).

Figure 17. Schematic of the evaluation of different

variants towards performance improvement of a race
engine.

51
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Improvements (Engine Development-Step 1)

Starting from the analysis of the engine “step 0” different
designs of the airbox have been simulated and evaluated
using QuickSim. Based on the results of this virtual
development process only the promising solutions have been
realized and then tested at the test bench. Thanks to this
virtual development it has been possible to speed up the
engine development process by limiting the budget.

Here a selection of these investigations is reported. Due to

secrecy the following results of full-engine simulations focus
mainly on the visualization of the mixture formation and
turbulence within the combustion chamber at the ignition
point.

Figure 19. (Step 1) - Fuel mass fraction in the intake

channels of cyl. 3 (section 5mm from valve seats) - 60
deg after TDC (cylinder relative) - 5500 rpm - WOT (full
engine mesh - 5th cycle).

Figure 20. (Step 1.0) - Mass flow through each intake-

valve of cyl. 3 - 5500 rpm WOT - 5th cycle.

Figure 18. Investigation on different airbox designs

(Step 1).

The starting idea here is to increase the volume of the airbox

by a factor of ca. 2.5. This aims to reduce the pressure waves
within the airbox towards a more “soft” intake process (see
Figure 18):
• Step 1.0: Greater airbox volume with throttle near the intake
runners.
• Step 1.1: Same airbox volume as “step 1.0” but higher
distance between the throttle and the intake runners.
Figure 21. (Step 1.1) - Mass flow through each intake-
valve of cyl. 3 - 5500 rpm WOT - 5th cycle.
Analyzing the concentration of the fuel within the intake
channels of cylinder 3 (see Figure 19) it becomes evident
how the position of the throttle is able to remarkably The analysis of the mass flows through the intake valves (see
influence the deviation of the fuel jets. In case of “step 1.0” Figures 20 and 21) shows in case of “step 1.0” higher
the asymmetry between the two channels increases also in differences in the peaks of the total mass flows through each
comparison to the “step 0” (smaller airbox volume). In case valve, i.e. in this case not only the fuel but also the repartition
of “step 1.1” the results show improvements, i.e. the of the intake air between the two valves becomes worse. This
repartition of the fuel between the two intake channels is now causes an increasing deflection of the fuel jets. Here the
better. shape of the curves of the total mass flow is more “waved”

52
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

and lightly retarded in the valve closing phase (no backflow comparison to “step 0”. Here, due to the particular design of
before IVC). In case of “step 1.1” (see Figure 21) the the airbox, the flow through the throttle manifold can first
asymmetries in general become lower and the curve profile of “calm down” in the airbox, so that the interactions with the
the total mass flow remains similar to the case of “step 0”. flow within the intake runners of each cylinder can be
reduced.

Improvements (Engine Development-Step 2)

In a second engine development step some investigations
have been focused on the influence of a filter plate within the
airbox and of different positions of the fuel injectors on the
charge motion (see Figure 24).

• Step 2.0: Similar airbox volume and design as “step 1.1”

with an internal filter plate.

• Step 2.1: Same airbox as step 2.0 but injector position

nearer to the cylinder head.

Figure 22. (Step 1) - Lambda distribution at the ignition

point of cylinder 3 - IP=25 deg before FTDC - 5500 rpm
- WOT (full engine mesh - 5th cycle).

Figure 24. Investigation on the same airbox design with

different injector positions (Step 2).

Figure 23. (Step 1) - (TKE) Turbulence at the ignition

point of cylinder 3 - IP=25 deg before FTDC - 5500 rpm
- WOT (full engine mesh - 5th cycle).

The results of the simulations (here briefly reported only for Figure 25. (Step 2.0) - Mass flow through each intake-
the cylinder 3 - see Figures 22 and 23) show for “step 1.0” valve of cyl. 3 - 5500 rpm WOT - 5th cycle.
also a worsening of both the mixture formation and the
turbulence level within the combustion chamber at the IP in
comparison to “step 0”. This means that only the increasing
of the airbox volume does not directly influence the charge
motion.

In case of “step 1.1” where the throttle is more distant to the

intake runners the mixture formation remarkably improves in

53
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

level in all combustion chambers shows a moderate reduction

of turbulence in comparison to the basis “step 0” or “step
1.0”, which sensibly increases the combustion duration
(negative influence on the indicated efficiency).

The configuration with fuel injectors nearer to the cylinder

head (“step 2.1”) shows a light worsening of the mixture
formation in comparison to “step 2.0” with the same level of
turbulence (as expected different fuel injector positions in the
manifolds do not influence the turbulence profile during the
whole operating cycle). Conclusively, solution 2.1, in
Figure 26. (Step 2.1) - Mass flow through each intake- comparison to “step 2.0”, thanks to the nearer position of the
valve of cyl. 3 - 5500 rpm WOT - 5th cycle. fuel injectors, does not have advantages in terms of indicated
mean effective pressure but only in terms of drivability due to
The analysis of the mass flows through the intake valves (see a faster response at any load changes.
Figures 25 and 26) shows very low differences in the profiles
of the total mass flows between the valves. Also the deviation
of the fuel jets becomes much lower than in the previous
cases. Due to the near position of the fuel injectors to the
intake valves in case of “step 2.1” the peak of fuel that flows
through the valves is brought forward in comparison to the
other cases (the fuel injection profile is the same in all the
cases).

Figure 28. (Step 2) - (TKE) Turbulence at the ignition

point of cylinder 3 - IP=25 deg before FTDC - 5500 rpm
- WOT (full engine mesh - 5th cycle).

CONCLUSIONS
Looking for performance enhancement of a race engine
(spark ignition) combined with high drivability there is a
Figure 27. (Step 2) - Lambda distribution at the ignition conflict among stable combustion (low σ_imep mainly due to
point of cylinder 3 - IP=25 deg before FTDC - 5500 rpm a homogeneous mixture formation), high imep (short
- WOT (full engine mesh - 5th cycle). combustion duration at full load mainly due to high
turbulence at the flame front) and short delay between fuel
The filter plate introduces a separation of the airbox into two injection in the intake runners and optimal mixture formation
chambers. Thanks to this approach the flow from the throttle in the combustion chambers during transient operating
can be better distributed over the whole filter-plate surface conditions. Considering the development steps reported in
with a resulting more uniform and lower velocity. It follows a this work, the most promising step for successive
reduction of the influence on the charge motion in the intake investigations has been represented by the step “1.1” (best
runners among the cylinders. compromise between quality of mixture formation and
turbulence level).
The airbox design with the standard injector position (“step
2.0”) shows a surprisingly good homogenization within the Based on these considerations, the development process of
combustion chamber (see Figure 27). As expected using an race engines requires an accurate analysis and evaluation of
airbox with integrated filter (see Figure 28) the turbulence the occurring engine processes towards a clear identification

54
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

of component design configurations, valve motion strategies, Working Fluids for Real Working-Process Simulations,”
injection strategies etc., which indisputably allow an SAE Technical Paper 2007-01-0936, 2007, doi:
enhancement of the engine performance. 10.4271/2007-01-0936.
12. Chiodi, M., Berner, H-J., Bargende, M., “Investigation
ACKNOWLEDGEMENT on Different Injection Strategies in a Direct-Injected
Turbocharged CNG-Engine,” SAE Technical Paper
The authors would like to acknowledge the team of the 2006-01-3000, 2006, doi:10.4271/2006-01-3000.
engine development at Volkswagen Motorsport GmbH in
Hannover for the wonderful cooperation and support with 13. Guelder, O. L., “Correlations of Laminar Combustion
experimental data and their feedbacks. Data for Alternative S.I. Engine Fuels,” SAE Technical Paper
841000, 1984, doi:10.4271/841000.
REFERENCES
1. CHIODI, M.: “An innovative 3D-CFD-Approach towards
DEFINITIONS/ABBREVIATIONS
Virtual Development of Internal Combustion Engines”.
CA
Dissertation, Universität Stuttgart, June 2010. Vieweg
Crank Angle
+Teubner Verlag 2011.
2. WICHELHAUS, D.: “Optimierung der dieselmotorischen
CFD
Verbrennung für hohe Leistung bei begrenztem
Computational Fluid Dynamics
Luftangebot”. Dissertation, Universität Stuttgart, Expert
Verlag 2009.
3. HEYWOOD, J.-B.: “Internal Combustion Engine CNG
Fundamentals”. McGraw-Hill International Editions, New Compressed Natural Gas
York, 1988.
4. Chiodi, M., Mack, O., Bargende, M., Paule, K., Brandt, V. EV
Fackh, J., Wichelhaus, D., “Improvement of a High- Exhaust Valve
Performance Diesel-Engine by means of Investigation on
different Injection Strategies,” SAE Technical Paper FTDC
2009-24-0008, 2009, doi:10.4271/2009-01-0008. Firing TDC
5. PAULE, K., WICHELHAUS, D., FERRARI, A.,
CHIODI, M., BARGENDE, M.: “CNG im Rennsport”. Imep
Gasfahrzeugekonferenz, October 2009. Indicated Mean Effective Pres.
6. CHIODI, M., FERRARI, A., MACK, O., BARGENDE,
M., WICHELHAUS, D.: “QuickSim in Motorsports”, 11. IP
Stuttgarter Symposium, Stuttgart 2011 Ignition Point
7. GEORGI, B., MEHNERT, J., DINGEL, O.:
“Möglichkeiten für CNG-Antriebe im Rennsporteinsatz”. 3.
IV
Tagung Gasfahrzeuge. Berlin 2008.
Intake Valve
8. GRILL, M.: “Objektorientierte Prozessrechnung von
Verbrennungsmotoren”. Dissertation, Universität Stuttgart.
IVC
9. CHIODI, M., BERNER, H.-J., BARGENDE, M.: Intake Valve Close
“Schnelle 3D-CFD-Untersuchungen des Ansaugsystems
eines aufgeladenen 11-Erdgasmotors mit dem
Entwicklungstool QuickSim”. Tagung: Ansaugsysteme STDC
aufgeladener Motoren, Haus der Technik, Ludwigsburg, May Scavenging TDC
2006.
10. BERNER, H.-J., CHIODI, M., BARGENDE, M.: TKE
“Berücksichtigung der Kalorik des Kraftstoffes Erdgas in der Turbulent Kinetic Energy
Prozessrechung”. 9. Tagung: Der Arbeitsprozess des
Verbrennungsmotors. Graz, September 2003.
11. Grill, M., Schmid, A., Chiodi, M., Berner, H-J.,
Bargende, M., “Calculating the Properties of User-Defined

55
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

WOT
Wide Open Throttle

The Engineering Meetings Board has approved this paper for publication. It has Positions and opinions advanced in this paper are those of the author(s) and not
successfully completed SAE's peer review process under the supervision of the session necessarily those of SAE. The author is solely responsible for the content of the paper.
organizer. This process requires a minimum of three (3) reviews by industry experts. SAE Customer Service:
Tel: 877-606-7323 (inside USA and Canada)
All rights reserved. No part of this publication may be reproduced, stored in a
Tel: 724-776-4970 (outside USA)
retrieval system, or transmitted, in any form or by any means, electronic, mechanical, Fax: 724-776-0790
photocopying, recording, or otherwise, without the prior written permission of SAE. Email: CustomerService@[Link]
ISSN 0148-7191 SAE Web Address: [Link]
Printed in USA

56
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

2008-01-2953

The Application of Energy-Based Fuel Formulae to

Increase the Efficiency Relevance and Reduce
the CO2 Emissions of Motor Sport
J. W. G. Turner and R. J. Pearson
Lotus Engineering

Copyright © 2008 SAE International

ABSTRACT Limiting the rate of energy usage by a vehicle at any

point in a race (either directly using fuel flow rate meters
Concerns over energy security and global warming are or indirectly using air flow restrictors) is unnecessarily
beginning to be a serious issue for society and are also restrictive and will constrain the evolution of energy-
starting to drive customer purchasing decisions across efficient technologies which are relevant to the
many areas. Against this background there is an mainstream automotive sector, such as hybridization.
increasing call for motor sport to improve its The paper proposes the adoption of a powertrain system
environmental image, despite the fact that the global formula with only a single major constraint: that the
energy consumption and CO2 emissions attributable to quantity of energy to tbe used for motive power which is
motor sport are a very low proportion of the total. The transferred to the vehicle before or during a the race be
real issue for motor sport in the face of the wider societal limited to a set maximum value. This value would be
concerns is that, if it is truly at the cutting edge of decreased over time to drive continual improvements in
relevant automotive engineering, it should be configured vehicle efficiency. It is argued that this approach would
and managed in such a way as to drive technology for incentivize the development of technology for fuel
the betterment of mankind. The status quo is, it is economy which is transferable to the automotive sector,
contended, increasingly seen to be blatantly energy- thereby enhancing both public awareness of such issues
profligate in the eyes of many people and this issue must and the environmental image of motor sport. Such an
be resolved if motor sport is to demonstrate the wider energy-efficiency approach will become increasingly
benefits of the technology developed by the huge important as society seeks to make the best use of any
financial investments committed to competing at the renewably-sourced energy available to it, i.e. to find
highest level. The paper proposes the adoption of much solutions for transport with the best well-to-wheels
broader rules based on limiting only the total energy (WTW) efficiency as a full vehicular system.
available to the vehicle to complete the race distance.
The paper also discusses some of the means of making
It is concluded that the application of this sole constraint renewable fuels and how, given a robust audit trail, the
would lead to the use of the substantial funding that the degree of renewability of the fuel could be used as a
sport attracts in the evolution of energy-efficient multiplier on the energy allowance in the interest of
technologies of direct relevance to the wider automotive encouraging race teams to move to more renewable
sector. fuels more quickly, to the benefit of the general image of
motor sport. The total energy-based approach covers a
INTRODUCTION wider range of energy storage media than conventional
fuels and allows battery-powered vehicles to be included
in the equivalence formula. This approach also
It is now widely accepted that anthropogenic greenhouse
encourages more aggressive uptake of the regenerative
gas (GHG) emissions are impacting the earth’s climate.
systems which are now beginning to appear in Formula 1
As the automotive industry is beginning to face targets
(F1) and in sports car racing.
for tailpipe CO2 emissions across the world the motor
racing industry will be subjected to increasing scrutiny,
despite its relatively small contribution to the problem. The authors accept that in absolute terms racing vehicle
Because of both its high profile and the considerable duty cycles are not the same as road vehicle ones (and
commitment of finances and resources made by the also that racing vehicles have no meaningful emissions
competing teams, top level motor sport formulae should restrictions placed upon them), but believe that adoption
be expected to take the lead in developing energy- of regulations in racing which promote efficiency is a
efficient technologies which have the potential to be worthwhile step since such a move at least has the
transferred to the automotive sector. possibility of promoting new technologies which could

57
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

feed back into the wider automotive sphere. This is likely because the overriding Lotus philosophy is one of
to become more possible as the degree of hybridization performance and efficiency through light weight, coupled
in a vehicle increases. with the intelligent application of technology, this being
taken from whatever field.
BACKGROUND AND HISTORICAL
PERSPECTIVE

The original thought process which led to the energy-

based fuel formula proposed in this paper was started by
the publication of an article in the January 2004 issue of
Automotive Engineering [1] concerning means of making
motor racing more relevant in terms of the efficiency with
which fuel energy can be converted into useful work.
Since that time, the proposals originally set out in that
article have, the authors believe, become more relevant
against the backdrop of global warming and energy
security. The authors have since conducted a great deal
of research into renewable fuels and energy supply [2]
and believe that these issues should be brought to bear
both in the motor racing arena and the wider automotive
world since energy (and the related issue of
anthropogenic CO2 release) is becoming increasingly
seen as the common denominator with regard to future
economic growth. The purpose of this paper is to update
the concept of an energy-based formula for powertrains
in motor racing and to show its ready transferability to the
issues facing society at large.

Although Lotus Engineering is largely known for work on

4-stroke reciprocating engines, it has pursued many Fig. 1: Compound-charged direct-injection racing engine
alternative concepts for power generation including the built to then-current Formula 1 regulations in the mid-
concept of a positive-displacement engine operating on 1980s
the Joule cycle (sometimes called a Brayton cycle
engine) utilising what would be widely recognized as
Wankel engine geometry [3], the automotive 2-stroke
(both conventional and free-piston) [4,5] and pressure
charged engines of unusual configuration [6,7], including
the development of one of the most ambitious charging
systems ever conceived for an F1 engine: compound
charging with both an axial compressor and variable
geometry turbochargers [8]. The simultaneous
incorporation into this last engine of a direct injection
gasoline fuel system controlled by piezoelectronics made
this an extremely advanced engine of any genre in the
mid-1980s. This V6 engine, built for the then-current 1.5
litre pressure-charged F1 regulations, is shown in Figure
1.

On the vehicle side, Lotus has raced four-wheel-drive

(4WD) cars powered by gas turbines in the form of the Fig. 2: The Type 56 gas-turbine-powered 4WD racing car
Type 56 (see Figure 2), has engineered road cars with was originally raced at Indianapolis in 1968 and later in
fuel cell engines and has even enjoyed victory with a this Type 56B specification in Formula 2
racecar with no engine at all at the Goodwood Festival of [Courtesy W. Taylor]
Speed (the Type 119 – see Figure 3).
The swept volume of engines has historically been the
Regardless of its image, therefore, as an organization it most commonly used means of establishing rules for a
does not feel beholden to any particular type of race series, and has also been used frequently for
powertrain, instead having a certain freedom of thought taxation purposes in road transport. Swept volume was
which others may not be permitted to exercise. This is always relatively easy to measure at least until the
especially the case when considering the vehicle, engine invention of the Wankel engine, which gave rise to a
and fuel as a complete system, and by extension, to the conceptual difficulty in equating it in an objective manner
efficiency of energy utilization from a WTW viewpoint to reciprocating engine operation. This problem rapidly

58
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

ensured it was banned from most forms of capacity- ENERGY, GREENHOUSE GAS EMISSIONS,
based motor sport. For related reasons, the same fate FUELS AND FUEL CONSUMPTION
befell the automotive gas turbine. However, with the
research and development of new technologies such as Concerns over energy security and the release of
series hybridization and 2-stroke engines for road greenhouse gas (GHG) emissions associated with the
vehicles, it is becoming clearer to society at large that combustion of fossil fuels, while not the same thing, are
there are better and more relevant ways to compare currently driving the simultaneous development of more
powertrains (and by extension, the vehicles to which they efficient vehicles and fuels which have a reduced fossil
are fitted) than the swept volume of a combustion CO2 footprint. This is happening around the globe. In
engine. Indeed, this orthodoxy may become moot if the the US, customer purchasing decisions are starting to
fuel-cell engine does become a viable mass-market move in the direction of reduced fuel consumption as a
proposition for passenger cars (a prospect highly result of increasing fuel costs. Congress has also
contended by some). enacted legislation requiring a improvement in Corporate
Average Fuel Economy (CAFÉ) from 28 to 35 US mpg
by 2017. In Europe, concerns over global warming as a
result of worldwide GHG emissions have prompted the
introduction of legislation designed to force
manufacturers to improve the fuel efficiency of vehicles.
Such CO2-based legislation is also being adopted in the
east.

The European situation has also been brought about by

a result of the failure of the manufacturers to meet the
CO2 reduction targets they originally agreed to with the
European Union (EU) under the ACEA ‘voluntary
agreement’ of 1998 (ACEA is the European Automobile
Manufacturers’ Association). Under this voluntary
agreement the OEMs undertook to reduce CO2 (as
measured at the tailpipe on the new European drive
cycle (NEDC)) to an average of 140 gCO2/km by 2008 in
exchange for no legislation being raised on the matter.
That this did not happen has led to the proposed
legislation which will require all manufacturers to meet an
Fig. 3: Type 119B ‘soap-box’ downhill racer racing at the average CO2 output for their products on the NEDC such
2003 Goodwood Festival of Speed that the average from the European fleet will be 130
1
gCO2/km from 2012 onwards . Failure to achieve their
Recently the CO2 emissions of a road vehicle on a given individual targets will result in swingeing fines per
drive cycle have been introduced as a more vehicle, the desired avoidance of which is expected to
environmentally-relevant basis for taxing vehicles. This drive significant investment in powertrain and vehicle
clearly takes into account not only the efficiency of the technologies by the OEMs to achieve their targeted
engine but also that of the vehicle to which it is mated – levels [9].
obviously aerodynamic drag, rolling resistance and mass
etc. all play a significant role in determining the amount As in the US, European buying patterns have been
of fuel a vehicle consumes (and hence its CO2 output), moving towards better fuel consumption cars both
as well as the fuel on which it operates, a subject which because of the increased cost of fuel (which has in any
will be returned to later. In terms of global warming, this case been significantly higher than in the US for some
new comparison of vehicles based on their CO2 output is time) and also rising environmental awareness. It is
very important indeed, and should perhaps be the interesting to note, however, that there is generally a low
ultimate decider of efficiency (as it is to the person level of understanding of the inherent link between fuel
paying the fuel bills, if they choose to take any notice). consumption and CO2 emissions in the mind of the
Discussing how this can be achieved with relevance general public. This link is illustrated in Figure 4, where
through motor sport is the purpose of this paper, but it is the CO2 released at the tailpipe is compared to fuel
important also to comprehend the link between energy, consumption for three different fuels: gasoline, diesel
greenhouse gas emissions, fuel types and fuel
consumption that will in any case become more
important in the future as some energy sources dwindle
and new technologies are developed to maximize the
potential of remaining resources whilst minimizing the
impact on the earth’s climate of exploiting them. 1
Note that while the legislation is being framed around 130 gCO2/km
on the NEDC, the EU actually intends to reduce the net fossil CO2
emission to 120 gCO2/km, with the additional 10 gCO2/km to come
from a renewable aspect of the fuel formulation. This will be returned
to later.

59
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

and E85 (a blend of 85% ethanol and 15% gasoline by measurements, knowing the properties of the fuel on
2
volume), using published results from the NEDC . which the vehicle is operating. In the future, knowing the
amount of energy used probably going to be much more
100
important, when both energy security and the renewable
90 nature of some forms of energy are likely to become
80 more and more important. Well-towheel (WTW) analysis
Fuel Consumption / [mpg]

70 of energy usage is an important approach to energy

60 utilization, but the authors would argue it is more relevant
50 Diesel to both the end user in general and society in particular
40 that this be resolved into well-to-tank (WTT) and tank-to-
30 Gasoline wheel (TTW) proportions. All of these efficiencies are
20
linked by the simple relationship:
10 E85

ηWTW = ηWTT x ηTTW

0
0 40 80 120 160 200 240 280 320 360 400 440 480 (Eqn 1)
CO2 Emissions / [g/km]

Thus the energy efficiency of the vehicle can be

Fig. 4: Theoretical fuel consumption versus CO2 for three considered separately from the efficiency of the energy
different fuels on the new European drive cycle (NEDC) source from which it draws. This is not the case with a
and test results for individual vehicle tests. [ ‘mpg’ refers simple volumetric fuel consumption or grams-of-CO2-
to miles per Imperial gallon] per-kilometer approach. Life cycle analysis (LCA) can
then be used to assess the energetic and environmental
That this link is immutable for any given hydrocarbon fuel impact of the fuel, while legislation could (and should)
is shown by the plotting on Figure 4 of individual test drive the energy efficiency of the vehicle instead.
results for European cars. Nearly all of the test results
fall on the theoretical lines (with the slight outliers
presumably being some function of test inconsistency). 9

8
The fundamental link between hydrocarbon fuel
Fuel Energy Input / [MJ/km]

7
consumption and CO2 emissions is a fact which all
6
customers should be made more aware of, since
120 gCO2/km =
increasing amounts of passenger-car taxation are being 5
1.64 MJ/km
related to the latter. However, from an energetic 4

standpoint, hydrocarbon fuels also produce a fixed 3

amount of CO2 per unit energy they contain. This is
X
2
shown in Table 1 for various viable transport fuels and
1
hydrogen, together with various other properties of
interest in automotive applications. 0
0 50 100 150 200 250 300 350 400 450 500 550 600
CO2 Emissions / [g/km]
Thus, from consideration of the data in Table 1 and the
fact that fuel consumption and the amount of CO2
emitted per km on any drive cycle are linked, it is Fig. 5: Fuel energy input on the European drive cycle
possible to adopt the more fundamental position of the versus tailpipe CO2 emissions for gasoline fuel
amount of energy used by the vehicle on the given drive
cycle rather than a notional volume of fuel. This
relationship is illustrated in Figure 5 for gasoline. This approach does, in fact, represent a far more
equitable split of the profits to be made from supplying
Here the relationship is a simple straight line and thus, the personal mobility sector. The vehicle OEMs would
for different fuels, it is possible to calculate the linear be called upon to make their products more energy
relationship between energy and CO2 performance. The efficient, while the fuel suppliers can separately be
two are clearly related by the grams of CO2 emitted per tasked with improving the renewability or energy security
MJ of energy that the fuel contains and the relationship is of their products, depending on how national
actually shown in Table 1. Expressed in this way the governments wish to prioritize those issues. At present
EU’s target of 130 gCO2/km is more fundamentally 1.774 only the vehicle is seen as the emitter, and is penalized
MJ/km if the fuel used is a typical gasoline with 73.3 as such, whereas the OEMs have little choice other than
gCO2/MJ. This is also readily discernible on Figure 5. to engineer their products to use fossil hydrocarbon
fuels, the raw material for which they do not own and
One can thus calculate vehicle energy efficiency from the sale of which they do not profit directly.
straightforwardly from the drive cycle CO2

2
The plot in Figure 4 shows fuel consumption in miles per Imperial
gallon, but it can be plotted in other units and will clearly yield defined
lines. There are 0.833 Imperial gallons to the US gallon.

60
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Table 1: Properties of several hydrocarbon and alcohol fuels and hydrogen

Property Gasoline Diesel Ethanol Methanol Natural Gas Hydrogen

(Typical) (Typical) (Typical)
Chemical Formula Various Various C2H5OH CH3OH CH4, C2H6… H2
Density at Atmospheric Pressure and 0.735 0.835 0.79 0.79 0.00083 0.00009
Temperature (kg/l)
Lower Heating Value (MJ/kg) 43.4 42.5 26.8 19.7 47 120
Lower Heating Value (MJ/l) (at STP) 31.9 35.5 21.2 13.6 0.0390 0.0108
CO2 per unit energy released (g/MJ) 73.3 74.2 71.4 69.8 55* 0
CO2 per litre of fuel (g/l) 2340 2633 1511 949 1.8* 0
* These values are for methane

The approach also allows the full energetic cost of plug- alternatively, if energy security is the main concern of the
in hybrids and pure electric vehicles to be assessed fairly government in question, fuels made from feedstocks
against vehicles which draw on a liquid (or possibly which originated from outside a country’s borders) can
gaseous) energy store. Many OEMs are expecting therefore easily be increased on a fossil-CO2-release
energy diversity to drive more varied solutions, and as a basis to encourage the use of fuels with a politically more
consequence are researching many new technologies favorable character. The concept of a factor such as this
for the future [10,11]. In fact governments are likely will be returned to in a later section related to the
going to have to move to such an energetically-based encouragement of renewable fuels in motor sport.
taxation regime for fuels as the market diversifies with
new fuels becoming available. In the UK, for example, Although this section has discussed relationships
electricity is taxed at approximately 10% of the energetic between energy, greenhouse gas emissions, fuels and
rate of gasoline, and natural gas is similarly cheap in fuel consumption from the position of the passenger
comparison. transportation, it can be seen how such parameters
could similarly be used in a motor sport environment to
Once this separation of fuel and vehicle efficiencies balance environmental and energy security concerns
occurs, the taxation system can be configured to favour with performance within a given formula. This is
fuels with either reduced WTT CO2 or energy security discussed in more detail below.
impact. Fuels therefore should be taxed on the amount
of energy supplied and not on a volumetric basis. This, PROMOTING FUEL EFFICIENCY VIA MOTOR
(the authors assert, is a historical anomaly caused by the SPORT
original ease with which volume flow could be measured
for the purposes of calculating the amount of fuel sold in As shown in Figure 6, progress was being made early on
early filling stations (necessarily using mechanical towards the ACEA voluntary agreement goal of 140
pumps), coupled with the fact that the two main gCO2/km, but from 2005 an interesting trend emerged:
hydrocarbon fuels used for transportation, gasoline and the average figure based upon units sold flattened off
diesel, fortuitously have similar volumetric energy markedly. The reason for this is very simple: for the
contents. In the modern world, however, fuel pumps at majority of people, fuel economy wasn’t very high on
filling stations are electronic, and thus it would be simple their list of priorities when purchasing a car until recently,
to calculate the volume of fuel supplied and hence the and was somewhat below the vehicle’s performance,
amount of energy; this would be the figure displayed on styling, image, etc.
the readout. Natural gas is already commonly metered
by the energetic unit, as is electricity. The move to an
The question is, how can one rationalize these issues in
energetically-based fuel taxation is therefore a
a manner promoting efficiency? There has been some
straightforward concept, and could be readily understood
significant movement recently towards more efficient
by the buying public by the fact that it is the energy a fuel
vehicles but this has primarily been as a result of the
contains, released in the engine, which moves the
recent marked increase in the price of fuel at the pumps.
vehicle, and not the volume it occupies.
But even accepting this pragmatism-driven trend, there
remains the question of how one can make fuel
Once this move to energy-based taxation is achieved, a efficiency attractive to both consumers and
further measure can be applied in the form of a multiplier manufacturers as a truly marketable quantity in its own
based on the audited amount of fossil energy used in a right. This image of efficiency should be one which is
fuels’ manufacture. Attaching absolute values to such attractive and enduring, rather than be something which
multipliers is beyond the scope of the current work is rapidly discarded should the price of oil come down.
(although some suggestions are made later on), but it One possible solution is through motor sport; indeed, it
can be seen that there is therefore a significant could be argued that motor sport has a responsibility to
inducement that governments could use to encourage reflect these issues of society as well.
the uptake of renewable fuels both for the fuel supplier
and the consumer. The cost of pure fossil fuels (or

61
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

200 efficiency of the engines except the above mentioned

190 engine speed limit (which manifestly does not have to
Tailpipe CO2 on NEDC / [gCO2/km]

180 have a beneficial effect); in fact, since the engine design

170 has effectively been frozen by the FIA (such that
160 manufacturers are not permitted to make changes to it
150
until the end of the current freeze period, expected to last
140
for five years) there is little that can be done to improve
130
the energy utilization of the engine in Formula 1 racing
120
cars.
110

100
In response to this status quo the FIA has permitted the
1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 introduction of kinetic energy recovery systems (KERS)
Year / [-] in an effort to improve the environmental image of the
Fleet Average 2008 Voluntary Agreement 2012 Proposed Legislation sport by promoting the use of energy regenerated during
braking [13]. However, the current regulations as framed
Fig. 6: Average CO2 emissions of passenger vehicles for 2009 limit the maximum amount of energy that can
sold in Europe as determined on the NEDC drive cycle, be captured and stored and the maximum amount of
with original 2008 target of 140 g/km and proposed 2012 power that can be delivered is 80.5 bhp (60 kW). It is
legislation point also shown presumed that this limitation is to control the amount of
energy stored aboard the car should a crash occur, in
Consider what is often championed by some as the the interests of safety.
pinnacle of four-wheeled motor sport, Formula 1 (F1).
We are in the middle of one of the longest periods of It must be noted that the requirements on a racing car –
stability in the F1 engine rules ever known. To achieve maximum power capability of the regenerating system –
that stability, however, the price has been high. The are somewhat different to the requirements of a road
engines could be perceived as being at the end of a long vehicle, which also suffers far greater challenges in the
and narrow development alley which leads nowhere form of durability and cost, and which also faces a
relative to the direction of mainstream technology in the significantly more difficult legislation set in the form of
automotive industry. Until the enforcement of an engine primary pollutant legislation. Nevertheless, the move to
design freeze, incremental gains came at the expense of fit KERS devices to racing cars is to be applauded as at
millions of dollars of development funding which least a step towards better energy utilization, with some
manufacturers were understandably loathe to see relevance to society in general. Whether this will be
wasted if a competitor had a good idea immediately developed in a logical manner, together with a long-
making what has gone before obsolete. Hence, as time overdue change in the fundamental engine regulations,
went on, internal pressure groups grew with the effect of however, is still to be seen; since the currently-proposed
making the rules more and more restrictive, which in turn regulations for KERS systems in 2011 effectively add up
had the effect of making incremental gains harder to to 268 bhp (200 kW) to the car, then KERS-equipped
achieve, meaning proportionately more money had to be cars in 2011 may have in excess of 1000 bhp unless
spent to move forward. This was clearly a vicious circle there is an engine rule change at that time [15], and they
which had the effect of removing the spectacle and will be able to accelerate significantly more rapidly than
making the series less and less relevant in engineering the previous year; this may be an issue with regards to
terms. For years, the only real avenue of development trackside safety before the issue of stored energy is
for the mandated natural-aspiration was ever-higher considered.
rates of engine revolution which had little equity in terms
of technology transfer to road cars, where pressure However, there is an argument that, correctly-framed,
charging and downspeeding the engine are clear trends even more open rules would increase the relevance of
[12]. Eventually this gave rise to the somewhat the engineering and therefore make the sport itself more
bewildering situation in current Formula 1 regulations interesting to many people. It would be a responsible act
where the engines are rev-limited to 19,000 rpm and now for the FIA to make the regulations force competitor
have to use a standardized engine control unit [13]. As a manufacturers to consider energy consumption, and
consequence of this scenario producing engines with hence CO2 generation, as a key driver on the type of
similar power output, coupled with unfortunate vehicle they field. Against this background the
aerodynamic and tyre regulations, overtaking introduction of KERS devices should be seen merely as
opportunities have been in decline on a year-by-year an excellent start. Removing limitations on both the
basis [14]. regenerative capability and energy storage capability of
the systems would encourage far more rapid
In fact, the engines are now built to such a tightly- development of concepts readily transferable to the
controlled formula that they are effectively stock units automotive arena since reduction of losses and removal
built by different manufacturers. Current rules are driving of energetic bottlenecks would allow much more of the
greater reliability into the engines used (with two race energy in the storage system to be converted into useful
meetings being the required life of an engine) but nothing work – in this case more rapid completion of the race
has been put in place to improve the fuel conversion distance.

62
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

A PROPOSAL FOR ENERGY-BASED 2. Allow any combination of several fuels currently or

POWERTRAIN REGULATIONS possibly relevant for the real world – gasoline,
diesel, CNG, alcohols, hydrogen etc. – since this
At this point it is proposed that an outright starting energy approach will allow stoichiometric spark-ignition
limit be the basis of future regulations for the powertrain. engines to compete with diesels, gas turbines or
Some motor sport formulae depend on air restrictors to even (hopefully) fuel cells.
achieve a notional parity of power output, but if fuel
efficiency of the whole vehicle is the aim (and hence a 3. Address the hybridisation issue by explicitly allowing
ready transfer of CO2-reducing technology to the road it (the aim is to promote the efficiency of the whole
car arena) then air restrictors should be avoided at all vehicle, in the same way low CO2 vehicles are
costs. This is because with an air restrictor, the most championed) .
powerful engine will be the one that consumes air in a
slightly rich manner. Conversely, the most fuel-efficient 4. For hybrid vehicles permit any kind of energy storage
forms of powerplant are those which run lean – the media, be it electrical, mechanical, pneumatic or
diesel, gas turbine and (in theory) the fuel-cell engine – hydraulic, etc.
and whose specific air consumption is therefore worse.
Consequently we should look at ways of controlling how 5. Permit refuelling, by any means – perhaps by
much fuel energy is consumed in the race, and not how cassette tanks for hydrogen consuming vehicles or
much air (which is not taxed and does not presently battery exchange for electric vehicles – but only up
suffer any energy-security-type issues). to the total amount of energy permitted.

The authors contend that the alternative of the fuel flow Vehicle regulations are outside the scope of the current
rate restrictor (i.e. energy flow rate restrictor) is too work, but it may be advantageous to remove weight
narrow in scope and also has the effect of seriously limits – upper or lower – in the interests of overall vehicle
compromising one of the key fuel efficiency technologies: efficiency.
hybridisation. This is because some forms of hybrid
powertrain might be more fuel efficient when running the The first of these suggestions is likely to be contentious
prime mover at full load for the short amounts of time. and will undoubtedly draw dire predictions of dull racing.
Hence, to limit the rate at which these powertrains can Interestingly, a fuel energy limitation is de facto the case
consume their fuel might therefore impose an artificial in most forms of sprint motorcycle regulations such as
restriction on efficiency, the latter of which is the ultimate MotoGP and World Superbikes, where tank size is fixed
aim. The broader specification of a total energy limit is and refueling is impractical, i.e. there is a fixed amount of
much less prescriptive and opens a plethora of energy available for the race. Crucially, these formulae
possibilities for energy management strategies which are provide arguably the most exciting racing of all, implying
not possible under an energy flow rate formula. that that limiting the fuel energy does not make for dull
racing and that other factors (aerodynamic and tyre
Considering road vehicles, an energy storage device is limitations etc.) do. It is perhaps no coincidence that
necessary (i.e. fuel tank or electric battery). It is used to they can function with significantly broader engine
store energy in the form required by the powertrain. The regulations than is the case in F1, Champ Car or
fiscal cost to the traveller of any given journey and is NASCAR.
affected by the amount of energy purchased in order to
complete the journey. The traveler actually wants the The concept of energy equivalency is not without
most efficient conversion in terms of MJ/mile (or MJ/km). precedent. There has recently been a move in the
A not dissimilar argument applies to a race distance, direction of fixed energy allowance by IMSA for the
where the most efficient use of the resources available American Le Mans Series (ALMS) [16]. This has been
can provide the shortest journey time. Therefore, limiting approached by balancing the fuel allowance with the aim
the total energy a racing car can use during a race that ‘the energy carried on-board the various cars should
distance would have the effect of race engineers seeking be (approximately) equal.” While in this case this applies
to maximise the efficiency of energy conversion. Since solely to the tank size aboard ALMS racing cars, it does
this is exactly what is being sought in the wider show that considering energy as a equilibrating factor is
automotive arena, there would be clear potentials for possible within the regulations of a given racing formula.
technological transfer between it and motor sport if this It is important to note that three fuel types (gasoline,
route were followed. diesel and E85) are permitted under ALMS regulations,
all with significantly different densities and energy
So, considering all of the above, the authors would like to contents (see Table 1).
suggest the following approach to regulation framing in
upper-echelon motor sport with regard to the powertrain: Under the above proposed regulations, ‘charge-
depleting‘ hybridization strategies would be permitted,
1. Throw the rules wide open by allowing teams only a but only under the overriding condition that all energy on
fixed quantity of energy at the start of a race (see board the vehicle at the start of the race is accounted for
below). and is within the energy limit. This would permit the use
of a degree of plug-in hybridization or even pure electric

63
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

vehicles. However it is policed, hybridisation should be reduction or energy security. Accepting the potential of
encouraged, and so the total quantity of fuel energy fuels to achieve the aim of reducing fossil CO2 release at
allowed for a race should be somewhat less than is a far greater rate than vehicle technology can (note that
currently uses during a race – say 25% less. If a vehicle manufacturers have vehicles in production now
regenerative braking system and the associated which can operate on fuels with near-zero TTW fossil
powertrain control system are performing correctly, the carbon such as alcohols; the issue is that these fuels are
peak power produced by the car should be very similar to not currently made available by the fuel suppliers in the
what is currently the norm (compare with the comments necessary quantities), various possibilities exist to
above regarding the introduction into F1 of KERS incentivize the adoption of fuel with a low fossil-based
devices to vehicles with an engine specification freeze carbon content:
[15]). It is accepted that a degree of simplicity is required
with regards to policing the regulations, so easily-
measured energy storage media may have to be 1. A separate score could be allocated to a team using
stipulated, such as the state-of-charge of any hybrid the lowest amount of fossil carbon in the fuel. This
storage media being no less than at the start of the race is analogous to separate championships for driver
– particularly in sprint racing it is important that the and constructor in some formulae. The team
spectator is aware of who has won the race on the basis scoring most highly in the season could be given a
of who crosses the line first, i.e. that the energy used in financial reward for use in the following season.
the race does not have to be determined after the race
has finished. 2. The amount of energy permitted could be multiplied
by a factor based on an audited fossil-CO2-
The specification of the entire powertrain configuration reduction or energy security trail. This is discussed
should be entirely free. Four-wheel-drive may be a very below and is suggested here merely to stimulate
advantageous concept when attempting to capture debate on possible future directions in the interest of
regeneratively as much energy as possible during a increasing relevance.
braking manoeuvre; this must especially be so in a car
with a very low centre of gravity such as a racing car Importantly, the encouragement of renewable fuel
whose braking is evenly divided between the axles. In development would compliment the direction which many
the interest of pursuing vehicle efficiency 4WD, and fuel suppliers are following as a result of the US Federal
infinitely variable gearboxes, should therefore be Renewable Fuel Standard’s requirement to provide 36
allowed. This would also open the door to billion gallons of biofuel per year; among other oil majors,
manufacturers who might have a desire to field 4WD both Shell (through its partnership with Choren) [17] and
machines, such as Audi, Subaru or Mitsubishi. BP (with Verenium) [18] have announced second-
generation ethanol programmes. This is in addition to
With regulations similar to the above, the large financial GM’s purchasing an undisclosed stake in Coskata [19].
resources allocated to F1 would be spent on technology Since it is unusual to find a vehicle OEM investing
which was likely to be applicable at some point in the directly in a fuel supplier, the potential of the Coskata
automotive sector. While the requirements of racing and process is perhaps illustrated; if other OEMs start to take
road cars are not exactly the same, the above proposal stakes in second- and third generation synthetic alcohol
does have significantly greater potential to yield useful producers, there may yet be a paradigm shift in the
technology than any overly-restrictive powertrain formula. overall personal mobility sector. The evolution of this
process should be encouraged by the adoption of a
A PROPOSAL TO ENCOURAGE THE ADOPTION mechanism to promote the development of the fuel.
OF RENEWABLE FUEL ENERGY SOLUTIONS Other new and predominantly chemistry-based
processes for the manufacture of alcohols would
The involvement of fuel companies in top level motor similarly be encouraged, such as catalytic processes
sport might be threatened by any or all of these changes. which can effectively use any hydrocarbon feedstock
Perhaps this would have to be accepted as a necessary including animal and municipal waste [20].
price to pay in pursuit of a greater societal good. Using a
variety pump (or series-specific) racing fuels would not With regard to a suggestion for setting the basic amount
be a bad thing: road cars are not developed solely to use of energy permitted for any given racing formula,
a single supplier’s fuel, and indeed it is even illegal for a consider the case of F1, although the methodology would
manufacturer to specify a single make of lubricant, let apply to any form of motor sport. Assume that 150 litres
alone fuel. Adopting this “high-street-fuel types only” of fossil-carbon-based gasoline fuel with the
approach is entirely relevant to the way road car characteristics shown in Table 1 are currently used per
customers buy their fuel, and would make fossil-CO2- car during an F1 race (i.e. a gasoline with lower heating
parity the only consideration with regard to the amount of value of 43.4 MJ/kg and a density of 0.735 kg/l). If all of
energy permitted for a vehicle. this fuel is consumed, this implies a total energy usage
of:
However, it would be possible to use the formula to
3
accelerate the development of more efficient processes 150 x 0.735 x 43.4 = 4784.85 MJ (Eqn 2)
for fuel manufacture, be this from the viewpoint of CO2

64
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

If there was a multiplier on the amount of energy such ethanol-based fuel would have an EEF of 1.066
permitted based on reduction of the amount of fossil CO2 and incur a taxation rate 93.8% that of gasoline.
released in the manufacturing cycle or alternatively the
energy return on investment (rE) in its manufacture, there Of course, EEF and taxation rate do not have to be
would be a very valid means of engaging fuel linked at all but this scenario is presented merely to
companies, both existing and new, in the competition. illustrate how degree of renewability could be made to
have a relevance to racing and the general. It is also
For example, the Coskata process for manufacturing worth pointing out that in the case of the third generation
ethanol is a third-generation process using lignocellulosic ethanol fuels, an effective 8.4% increase in energy
breakdown techniques and has been audited by Argonne allowable for a race would be a significant advantage for
National Laboratories for rE and CO2 [10]. These audited any racing team, making the adoption of a renewable
figures are 7.7 and 84% respectively (i.e. since energy is fuel a serious option when traded against the increased
added photosynthetically there is 7.7 times as much tank size or frequency of refueling required. For the
energy yielded by the process as is put in and this energy remainder of this paper it is assumed that the taxation
reduces fossil CO2 emission by 84%). This is very rate would be multiplied by a greater factor, still utilizing
usefully better than first-generation ‘corn ethanol’ at the degree of renewability but adjusted by a factor of 10
typical values of rE of 1.29-1.65 and even other and calculated as follows:
lignocellulosic methods at 4.5-6.6, with corresponding
reductions in fossil CO2 emission [21]. A straight Fossil CO2 taxation factor (%) =
multiplier based on energy security might therefore 1/(1 + (% reduction of fossil CO2)/100) (Eqn 4)
permit a vehicle using this fuel to have 7.7 times as
much energy as one using a fossil fuel, or one based on implying cellulosic ethanol would have a fossil CO2
reduction of fossil CO2 release might permit 100/(100-84) taxation rate of 54.3% applied to the basic energy tax
= 6.25 times as much energy. Such a massive increase levied per MJ. Similarly, the factor for E85 using this
would very quickly drive the take-up and development of ethanol would be 60.2%.
more renewable fuels, rather than the expensive
development of synthetic gasolines to comply with Of other alternative fuels, by this model both renewable
restrictive regulations. electricity and hydrogen manufactured by fully renewable
means would therefore have an EEF of 1.1 (from
Here a conjoining of road fuel taxation and race energy Equation 3) and incur taxation at a rate of 50% (from
allocation can be imagined. Consider an advanced Equation 4). Methanol, which can be made fully
renewable fuel such as ethanol made by the Coskata renewably from feedstocks supplied by air extraction of
process which has a fossil CO2 reduction factor of 6.25 CO2 and the electrolysis of water, could theoretically
(i.e. equal to (100-84)/100). It could be argued that to have an EEF of greater than 1.1 and correspondingly a
permit a vehicle using this fuel to have 6.25 as much CO2 taxation factor of less than 50%, because some of
energy with which to complete a race is perhaps the product resulting from this process can be
excessive, so for the sake of argument and purely to ‘autosequestrated’ by its conversion into plastics and
promote discussion assume a ‘CO2 energy equivalence other solid hydrocarbon products. That air extraction of
factor’ (EEF) related to fossil CO2 release and configured CO2 is feasible and the concept of a fully renewable
as follows: liquid transport fuel manufactured from it is possible has
been investigated and/or demonstrated by several
CO2 EEF = 1 + (% reduction of fossil CO2)/1000 (Eqn 3) authors [22,23,24,25]. The use of chemical looping can
help in achieving minimum energetic loss in the
For ethanol manufactured using the Coskata process its methanol manufactured from the atmospherically-
EEF would be 1.084, meaning that a car using this fuel scrubbed CO2 [26].
would be permitted 1.084 times as much energy that
available to one operating on a straight fossil fuel. The So called gas-to-liquid (GTL) fuels manufactured from
reciprocal of this could be the factor applied to energy for natural gas using the Fischer-Tropsch could be used and
taxation purposes, i.e. Coskata ethanol would in itself these are considered to be slightly worse than fossil-
incur 92.2% of the tax of fossil gasoline. By the same derived fuels in terms of gCO2 per MJ when the whole
logic and using the data in Table 1 for density and LHV process losses are considered [27]. This being the case,
then in admixture with 15% gasoline (by volume) in E85, under this regime GTL would have an EEF of slightly
less than 1 and a taxation rate of just over 100%, but it
may be possible to make the fuel and powertrain system
more efficient when operating on such fuels than
3
However, also from Table 1, it is interesting to note that gasoline conventional fossil hydrocarbons. Coal-to-liquid (CTL)
releases 73.3 gCO2/MJ, which in turn suggests that each car will would have an EEF less than 1 and a taxation rate
release a total of 350.6 kg of fossil CO2 during the race. Since the significantly over 100% because coal produces a
current minimum permitted mass of an F1 car (including driver and significant amount of CO2 per MJ. Biomass-to-liquid is
apparel) is currently 605 kg, this represents about 60% of the unfuelled
mass of the car. Furthermore, this leads to the rather uncomfortable
also possible and this would have an EEF greater than 1
realization that 20 such F1 cars on a starting grid will, assuming they all and taxation rate of less than 100%.
reach the end of the race, emit over 7 tons of fossil CO2 during the
event.

65
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Table 2: Suggestions for linked factors for road-fuel taxation and race energy allowance for various fossil and renewable
fuels

Property Fossil Fuel Ethanol Ethanol Methanol Hydrogen / Electric

rd
(Corn) (3 Gen.) (Renewable) Renewable Grid
Fossil CO2 mitigation 0 29%* 84%* 105%** 100% 36%***
Fossil CO2 energy equivalency factor 1 1.029 1.084 1.105 1.1 1.036
Fossil CO2 taxation factor 1 0.76 0.54 0.49 0.5 0.74
Total F1 race energy permitted (MJ) 3600 3704 3902 3978 3960 3730
* Argonne National Laboratory assessment from [10] ** Assumes air extraction of CO2 and renewable hydrogen production, with an assumed 5% of
fuel being sequestrated as hydrocarbon products [21,22,23,24] *** Based on an assumed European electricity grid fossil CO2 emissions

Table 2 presents these suggested figures based on the terms of fossil CO2 release or rE, will be an important
foregoing discussion and the data presented in Table 1. factor in ensuring fairness between competing fuel and
The row entitled ‘total F1 race energy permitted’ is based vehicle systems and simultaneously encouraging
on vehicles operating on conventional fossil fuels being renewability into either a racing formula or day-to-day
permitted a total of 3600 MJ for the race distance, or travel. However, the CO2 auditing proposed above is in
approximately 75% of the assumed energy used at line with proposals being considered in the EU for
present from equation 2. (In electrical terms, this is 1000 biofuels and the prize for racing is scrupulous credentials
kWh, in itself illustrating why liquid fuels are very in the eyes of the public and the development of relevant
attractive in terms of their energy density.) fuel systems.

It is reiterated that the figures presented in Table 2 are CONCLUSION

for discussion purposes only and it may be that from the
outset of any such formula these would be modified While it is true that there is little commonality in duty
depending on the audited CO2 impact of the fuels. The cycle between racing and road vehicles, it is also true
important issue is that fossil fuels are penalized due to that current and proposed technical regulations for motor
their CO2 impact although they still present an sport do not incentivize the development of innovative
opportunity to the system designer due to their high technologies which improve powertrain and vehicle
energy density. efficiencies and therefore reduce energy consumption
and CO2 emissions. The use of devices such as air or
Full BTL gasoline / diesel / kerosenes would be fuel flow restrictors which limit the rate of energy usage
advantageous since they can be made fully renewably. by the vehicle at any time impose stifling constraints on
This suggests that a total energy audit may be required technology evolution to improve efficiency.
in the future to bring in a factor for the energy loss in
manufacture, distribution and storage. Against this Adopting a formula based on limiting the total external
background, while renewable hydrogen may seem energy transferred to a vehicle at the start of or during a
attractive from Table 2, the very high energy loss that race provides a more fundamental constraint on energy
molecular hydrogen incurs through the distribution and profligacy which will allow a plethora of technical
storage network would count heavily against it [28]. It is solutions to be investigated within the motor sport arena,
not intended to discuss WTT weighting factors further thereby enhancing the public awareness of these
here, only to point out that this should perhaps be technologies and issues and improving the
considered in a long-term rules development. environmental credentials of the sport. This proposal
also obviates the arcane technical discussion regarding
Table 2 brings together racing and the real world insofar issues such as air restrictor sizes allowed for vehicles
as, because the fossil CO2 taxation factor is the a using different fuels, the need for equivalency factors (as
function of the fossil CO2 energy equivalency factor, the was the case for Wankel and gas turbine engines) and
spectator sees race cars using more ecologically sound the need for any assessment of energy state at the end
fuels being permitted greater fuel energy and of the race: any energy unused at the end represents a
simultaneously can choose a road vehicle-fuel potential loss of performance (and hence efficiency) for
combination for which he pays less in terms of currency the vehicle over the race distance. Importantly, the
per mile. Hence the approach, while requiring detailed results of the race stand at the time of its completion, in
discussion, could drive relevant behaviour patterns in the same manner as is the case now.
both arenas. Other details requiring thought are whether
refueling rates should be equalized based on energy flow In addition to this basic criterion, future evolution of the
rate, although this would make it difficult to truly equalize rules could include incentives to use energy resources
hydrogen and electrical storage systems with liquid fuels with reduced fossil carbon content when suitable auditing
(if cassette tanks were used). procedures can be agreed. At this point, a robust and
rational link to transport energy taxation could also be
From the above it is plain to see that robust and created further strengthening the validity of motor sport
transparent auditing of the impact of the fuels, be it in in the eyes of the public.

66
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

ACKNOWLEDGMENTS 9. Communication from the Commission to the Council

and the European Parliament: Results of the review
The authors wish to thank the directors of Group Lotus of the Community Strategy to reduce CO2 emissions
for permission to publish this paper. from passenger cars and light-commercial vehicles
{SEC(2007) 60} {SEC(2007) 61}. Available at:
The photograph of the Type 56B image in Figure 2 is [Link]
taken from ‘The Lotus Book’ by William Taylor, published [Link]/LexUriServ/[Link]?uri=CELEX
in the UK by Coterie Press Limited (ISBN 1 902351 00 :52007DC0019:EN:NOT
2). The author’s permission for its use is gratefully 10. Stephens, T.G., “GM's Advanced Propulsion
acknowledged. Technology Strategy - Solutions for Reduced CO2-
Emissions and Energy Diversity”, Year 2008 Series
CONTACT 12 No./Vol. 672-1 No. 2, 29th Vienna Motor
Symposium, Vienna, Austria, 24th-25th April, 2008.
James Turner 11. Schmidt, G., “Energy Storage and the Path to
Chief Engineer, Powertrain Research Sustainable Transportation”, Year 2008 Series 12
Lotus Engineering No./Vol. 672-1 No. 11, 29th Vienna Motor
Norwich Symposium, Vienna, Austria, 24th-25th April, 2008.
Norfolk NR14 8EZ 12. Königstein, A., Larsson, P.-I., Grebe, U.D. and Wu,
United Kingdom
K.J., “Differentiated Analysis of Downsizing
Concepts”, Year 2008 Series 12 No./Vol. 672-1 No.
Email : jturner@[Link]
22, 29th Vienna Motor Symposium, Vienna, Austria,
24th-25th April, 2008.
REFERENCES 13. [Link]
9__2009_F1_TECHNICAL_REGULATIONS.pdf
1. Turner, J.W.G. and Pearson, R.J., “Driving 14. CAR Magazine, July 2008, pp. 32-33.
powertrain efficiency forward through healthier 15. Ellis, C., “Two steps forwards but one backwards”,
competition”, Automotive Engineer, January 2004, Race Tech Magazine, June 2008, pp. 23-27.
pp. 36-38. 16. International Motor Sports Association Bulletin
2. Pearson, R.J. and Turner, J.W.G., “Exploitation of Number ALMS 07-03, “Fuel Energy Balance”, 5
th
Energy Resources and Future Automotive Fuels”, January 2007. Available at:
SAE paper number 2007-01-0034, SAE Fuels and [Link]
Emissions Conference, Cape Town, South Africa, n%[Link]
January 2007. 17. [Link]
3. “An Engine with a Trochoidal Rotor”, U.K. Patent esel_sup_sup_production/
Application No. 9910293.1, May 1999. 18. [Link]
4. Blundell, D.W. and Sandford, M.H., “Two-Stroke 2968&contentId=7046627
Engines - The Lotus Approach”, SAE paper number 19. [Link]
920779, SAE 1992 World Congress, Detroit, 20. [Link]
Michigan, USA, February 1992. [Link]
5. Blundell, D., Turner, J., Duret, P., Lavy J., 21. Hammerschlag, R., “Ethanol's Energy Return on
Oscarsson, J., Emanuelsson G., Bengtsson, J., Investment: A Survey of the Literature 1990-
Hammarström, T., Perotti, M., Kenny, R. and Present”, Environmental Science & Technology Vol.
Cunningham, G., “Design and Evaluation of the 40, No.6, pp. 1744-1750, 2006.
ELEVATE Automotive Two-Stroke Engine”, SAE 22. Zeman, F.S., “An Investigation into the Feasibility of
paper number 2003-01-0403, SAE 2003 World Capturing Carbon Dioxide Directly From The
Congress, Detroit, Michigan, USA, February 2003. Atmosphere”, Paper Number 235, Second Annual
6. Turner, J.W.G., Pearson, R.J., Bassett, M.D. and Conference on Carbon Sequestration, Alexandria,
Oscarsson, J., “Performance and Fuel Economy Virginia, USA, 5th-8th May, 2003.
Enhancement of Pressure Charged SI Engines 23. Weimer, T., Schaber, K., Specht, M. and Bandi, A.,
through Turboexpansion – An Initial Study”, SAE “Methanol from Atmospheric Carbon Dioxide: A
paper number 2003-01-0401 and SAE 2003 Liquid Zero Emission Fuel for the Future”, Energy
Transactions, June 2004. Convers. Mgmt, Vol. 37, Nos 6-8, pp. 1351-1356,
7. Turner, J.W.G., Pearson, R.J. and Parrott, A., “The 1996.
Performance of a High Compression Ratio, High 24. Olah, G.A., Goeppert, A. and Prakash, G.K.S.,
Speed Supercharged Engine”, Global Powertrain “Beyond Oil and Gas: The Methanol Economy”,
Congress, Dearborn, Detroit, Michigan, September Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim,
2004. Germany, March 2006, ISBN 3-527-31275-7.
8. Doble, I.M., “Supercharging with an Axial 25. Zeman, F.S., “Experimental Results for Capturing
Compressor”, SAE paper number 870722, 1987. CO2 from the Atmosphere”, R&D Note, AIChE
Journal Vol. 54, No.5, pp. 1396-1399, May 2008.

67
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

26. Zeman, F. and Castaldi, M., “An Investigation of

Synthetic Fuel Production via Chemical Looping”,
Envronmental Science and Technology, Vol. 42, pp.
2723-2727, January 2008.
27. EUCAR, CONCAWE and the Joint Research Centre
of the European Commission Directorate-General,
“Well-to-wheels analysis of future automotive fuels
and powertrains in the European context”, Well-to-
Wheels Report, Version 2b, May 2006. Available at:
[Link]
28. Bossel, U., “Does a Hydrogen Economy Make
Sense?”, Invited Paper, Proc. IEEE, Vol. 94, No. 10,
pp. 1826-1837, October, 2006.

DEFINITIONS, ACRONYMS, ABBREVIATIONS

ACEA European Automobile Manufacturers’ Associa-

tion
ALMS American Le Mans Series
BTL Biomass-to-liquid (fuel)
CAFÉ Corporate Average Fuel Economy
CTL Coal-to-liquid (fuel)
EEF Energy equivalency factor
F1 Formula 1
FIA Fédération Internationale de l’Automobile
GHG Greenhouse gas
GTL Gas-to-liquid (fuel)
IMSA International Motor Sports Association
LCA Life cycle analysis
NEDC New European drive cycle
rE Energy return on investment
TTW Tank-to-wheel
WTW Well-to-wheel
WTT Well-to-tank

68
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018




High Power Density Motor for Racing Use

Tamotsu Kawamura, Hirofumi Atarashi, Takehiro Miyoshi

Honda R&D Co.,Ltd., Automobile R&D Center
4630 Shimotakanezawa, Haga-machi, Haga-gun,Tochigi, 321-3393 Japan
(E-mail: Tamotsu_Kawamura@[Link])

Copyright © 2011 SAE International andd Copyright © d2011

d d
Society d
of Automotive d of Japan,
Engineers d d
Inc.

ABSTRACT: This paper presents the development of a motor that was supposed to be adopted in the Kinetic Energy
Recovery System for the Formula One world championship 2009 series. KERS can recover and utilize braking energy in the
form of power assist, with maximum energy use per lap of 400 kJ and maximum power of 60 kW. This hybrid system is
expected to provide benefits such as lap time reduction and straightaway speed gain for overtaking. The motor is required to
achieve downsizing, light weight, high power density, wide power band, and high efficiency. This paper describes how the
on-board layout, size, and specifications were decided and how the newly developed technologies were applied and test
result to confirm enhanced performance.

KEY WORDS: Motor, Hybrid, KERS, Formula One, Race

KERS-equipped vehicle, and in 4.22 seconds 10 m (two car

1. Introduction
lengths) ahead. This difference is enough to enable overtaking,
As for the Formula One World Championship, the regulation and so the system was considered effective in terms of race
amendments of 2009 allowed usage of Kinetic Energy Recovery strategy (Fig. 2).
System (KERS), which recovers and utilizes braking energy in
the form of power assist. This paper describes the development of [Link] Layout
a motor for KERS.
KERS on-board layout took not only the assurance of collision
safety but also the aerodynamic design of the vehicle into
consideration in securing the fuel tank capacity, optimizing the
longitudinal weight distribution in the vehicle, and lowering the
center of gravity. As a result, the motor was placed on the
engine’s left front side, the power control unit (PCU) on the left
side of the monocoque, and the battery at the bottom of the keel
and the nose (Fig. 3).

Fig. 4 shows an overview of the KERS. The motor and PCU

are on an anti-vibration mount on the left side of the monocoque.
Fig. 1 Honda KERS machine for 2009 F1 series The motor is cooled by engine oil, the PCU by its own cooling
water, and the battery by draft air from the front. The motor drive
2. Development Concept shaft is connected to the crank shaft through the engine's five-gear
train (Fig. 5).
The 2009 F1 regulations limited KERS connection to the rear
wheels, and the maximum amount of energy used in one lap 3. Motor Design
around the course was 400 kJ, while power output was limited to
a maximum of 60 kW. The effects of KERS introduction were Of course, the motor was designed with light weight,
projected using vehicle dynamics simulation. When the power compactness, and high output in mind, but making the
assist is initiated at 180 km/h, the point at which the tire grip motor more highly efficient is a crucial factor in securing
exceeds the drive system torque so that the tires stop spinning, the energy required for overtaking. A brushless DC motor
then use of the 60 kW of power for 6.67 seconds (400 kJ energy was adopted in order to realize these various factors at a
equivalent) enables the vehicle speed to exceed what it would be high level. Enabling the on-board layout described above
without KERS installation by 15 km/h. In terms of distance, that required the motor to be designed with a diameter of 100
difference is 20 m. Even if the course layout were such that the mm or less and a total length of 200 mm or less.
power could not be used for 6.67 seconds, in 2.78 seconds it Calculation from the required output shows that the power
would still put the vehicle 5 m (one car length) ahead of a non- density would be 8 kW/kg or more. This was very
demanding and far exceeds what is required of

69


 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

*Monza straight acceleration

+15 km/h
400[ J] assist Non-KERS car
finish @60kw

Assist start
180km/h

+20 m

Non-KERS car

6.67sec
Scale up
+2 car lengths

+1 car length

Non-KERS car
2.78sec 4.22sec

Fig. 2 Assist time VS car speed and distance gain

Engine
GearBox

Front

Motor
PCU
Battery

Fig. 3 Layout of KERS units

Energy storage
s
Li-ion × 106 Cells Water cooling

Water cooler

J/Box
Water pump

S-ECU M/G ECU

PCU
Engine
Engine
oil cooler Oil tank
Oil pump
Oil cooling 5-gear train

Motor
Front 60 kW 45 N-m 21000 rpm

Fig. 4 KERS overview

70


 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

190

100
Fig. 8 KERS motor
Fig. 5 KERS Engine with motor
The basic design of the motor was created with a
mass-production motors. KERS simulations were carried
particular focus on these three points: Achieving a balance
out using data from past races, and the target performance
between higher magnetic flux density and low loss, higher
was defined as the ability to operate continuously for 6.67
speed, and greater cooling. The following will describe the
seconds (400 kJ energy equivalent) with assist and recover
main constituent technologies.
operation available over the wide range of 13 000-21 000
rpm, which represents 40% of the maximum speed, at a Table 1 KERS motor specification
maximum output of 60 kW. This would enable the Type Brushless DC
minimum recover amount to be secured on any circuit. Fig. Phase 3
6 shows the estimated lap time reductions on various Stator teeth 12
circuits. Fig. 7 indicates that lap energy balances on several Rotor Poles 4
circuits were estimated to suffer shortage. That means Stator Lap Winding
KERS efficiency is quite important to maximize KERS Rotor SPM
performance. Achieving the high target performance Diameter 100mm
Size
Length 190mm
involved developing new electromagnetic steel plates and
Speed 21,000rpm
magnetic material, reviewing the winding techniques, Output 60kW
adopting new cooling methods and rotor structures, and Efficiency(Motor) 98%(Peak)
other such measures affecting the entire motor design. As a Weight 6.9kg
result, the high power density of 8.7 kW/kg was achieved Power Density 8.7kW/kg
while maintaining high efficiency. Fig. 8 shows an exterior
view of the motor and Table 1 shows the specifications. 3.1. Magnetic Design

In the development of the stator, new stator core material and

process were developed in order to reduce the size and weight of
the stator core, the heaviest component part in this motor. An
iron-cobalt alloy (49Fe-49Co-2V) was adopted for the
electromagnetic steel plates. This yielded a 30% increase in the
saturated magnetic flux density and 15% in the torque compared
to when silicon steel (Fe-Si) plate 200 m thick was used, and this
allowed the size to be reduced by a commensurate amount (Fig. 9).
A 60% reduction in iron loss was also realized through a
combination of technologies to lower eddy current loss by
reducing the plate thickness to 100 m, reduce hysteresis loss by
Fig. 6 KERS lap time reduction estimation post-rolling heat treatment, and enhancing the space factor by
means of an oxidized insulation coating (Fig. 10).
A newly developed high-coercivity magnet (iHc=1.1 MA/m or
more at 160°C) was adopted for the rotor. The magnetization
angles in and out were also tuned so as to concentrate the
magnetization direction of the magnet toward the polar center,
which raised the torque 8% (Fig. 11). Putting the rotor shaft to
work as another magnetic path realized a reduction in the rotor
diameter (Fig. 12). The magnet was also divided into 28 pieces in
the axial direction and 16 pieces circumferentially, making a total
of 448 pieces, thus reducing temperature increases occurring due
to eddy current loss in the magnet (Fig. 13).
Fig. 7 KERS lap energy balance estimation

71


 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

3.2. Mechanical Design

Magnetic flux density (T)
In order to achieve high torque and reduce cupper loss, double-
layer lap winding was used for stator winding, resulting in greater
density. Highly scratch-resistant magnet wire was also used, and
the ends were press-molded for greater compactness (Fig. 14).

22 24

Magnetizing force (A/m)

Fig.9 Increasing saturated magnetic flux density of stator core

100 Oxidized insulation coating

Iron Loss

% Reducing the plate thickness

Post-rolling heat treatment
40
% 16 Fig.14 Stator assembly
4 40
% % %
In order to reduce iron loss by lowering compressive stress on
Developed the stator core when fixing the stator inside the case, the
Prototype
conventional shrink-fit method was augmented with adhesive.
Fig.10 Reduction of stator core iron loss This reduced interference to an extremely small size. That
allowed the case wall to be made thinner, also contributing to
reductions in diameter and weight.
in The rotor was made capable of withstanding the centrifugal
force at the top speed of 21,000 rpm by using organic, high-
strength fibers in a high tensile strength filament winding as a
out measure to prevent magnet burst. This winding retained the
circumference of the magnet (Fig. 15). Grease-lubricated ceramic
ball bearings were adopted in order to increase speed and reduce
Prototype Developed loss in the high-temperature environment.
Fig.11 Tuning of magnetization direction Fig.15 Rotor assembly

Stator

Filament winding
Rotor

3.2. Thermal Design

Thermal design is critical to enhancement of the motor's power

Shaft density. Reduction in the heat quantity and strengthening of
Fig.12 Magnetic flux through motor shaft cooling performance are both essential. Steps were taken to
reduce loss in both the stator and the rotor, as described above,
but the cooling method still remains as a major issue. In the
Magnets divided 28 axially water-cooled method, a water jacket that has development
experience is attached on the periphery of the motor. However,
the motor does not achieve a balance between the target
performance and target size with this approach. Therefore, a new
structure was developed for direct cooling of the heat generating
locations by placing a duct for cooling oil inside the motor.
Common use of the engine oil as a motor coolant allowed the
radiator and other accessory parts used for cooling to be shared
with the engine. In order to avoid stirring resistance caused by
contact of the cooling oil with the surface of the rotor magnet, the
Fig.13 Section view of rotor
space between the stator and rotor was sealed with a cylindrical
sleeve (Fig. 16). The adoption of a sleeve structure enlarged the

72


 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Fig.16 Section view of motor with cooling oil flow

Table 2: Motor component technologies and effects

Size Flux Speed Loss Cooling
High saturated magnetic flux density
Oxidized insulation coating
Stator
Reducing the plate thickness
Magnetic Post-rolling heat treatment
design High BH magnet
Magnetization direction tuning
Rotor
Divided magnets
Magnetic flux through shaft
Compact coil winding
Stator
Mechanical Fitting minimization with adhesive
design High tension filament winding
Rotor
Ceramic ball bearing
Thermal Stator Direct oil cooling
design Rotor Oil seal sleeve

magnetic air gap, and the effect of the resulting reduction of Interlagos circuit simulation Assist
performance in the magnetic circuit was held to a minimum by 60
thinning the sleeve. The thickness of the sleeve wall was then
reduced by optimizing the angle of orientation of the filaments in 50

in Boost
the filament winding while still retaining the strength to withstand 40
the hydraulic pressure.
30 90
Motor output (kW)

The technologies developed as described above, together with 93 95 96 97 98

20
their effects, are summarized in Table 2.
10
4. Test 0
-10
The Measurements were made during bench testing to
determine the motor output curve, the operating efficiency map, -20
in Brake

and the operation points in circuit mode. The motor was capable -30
of power assist and recovery operation at maximum power output,
96
confirming that the motor power band was configured properly. The -40
operating efficiency of the stand-alone motor was an average 95% 87 90 93 94 95
at an average speed of 20000 rpm during power assist, and an -50
average 93% at an average speed of 16500 rpm during recovery -60
operation. Thus, the motor achieved a balance between high 9000 13000 17000 21000
output and high operational efficiency (Fig. 17). Speed (rpm) Recover

Fig.17 Motor efficiency map with operation points

73


 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Actual vehicle testing consisted of shake-down on a straight

course in April 2008 and driving tests conducted ahead of the
other teams on the Silverstone circuit in May, where the
superiority of this system and its high level of safety were proven.
Driving tests were conducted at the Jerez circuit in September to
verify the effects of KERS in an actual vehicle with the regulation
upper limit load of 60 kW. The following effects were confirmed:

•Lap times were shortened by approximately 0.4 seconds per lap

with power assist of around 400 kJ.
•A speed increase of 7 km/h and a distance increase of 7.8 m (1.6
car lengths) were achieved on a straightaway with a continuous
324 kJ of power assist.

5. Conclusion

Honda KERS motor achieved both compactness and

thermal toughness by newly developed technologies.

1)Stator core with high magnetic flux density CoFe

contributed to 15% greater torque with low iron loss.
2)Rotor gained 8% torque by magnetization direction tuning,
downsized by magnetic flux through shaft design,
reduced eddy current loss by divided magnets.
3)Direct oil cooling by thin oil seal sleeve accomplished
high cooling and motor performance with compactness.

Finally, quicker lap time and overtaking boost effect were

definitely demonstrated by the world 1st F1 KERS machine.

References

(1) [Link], [Link], [Link],

Development of Hybrid System for Formula One, Honda R&D
Technical Review, F1 Special (The Third Era Activities), p.215-
227
(2) [Link]
Development of Hybrid System for Formula One, JSAE
Motorsports sympodium 2010
(3) [Link], [Link], [Link], [Link],
Development of low loss stator core for formula-1 hybrid system,
IEEJ national conference 2010
(4) [Link], [Link], [Link], [Link],
Magnetic properties of Fe-Co-V alloy with low iron loss for
motor core material, IEEJ national conference 2010
(5) Tamotsu Kawamura, Hirofumi Atarashi, Takehiro Miyoshi
Development of F1 KERS motor, EVS25 in Shenzhen, China
2010

74


 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

JSAE 20119513 / SAE 2011-32-0513

The Development of a Low Viscosity, Highly Efficient Lubricant for Sport

Motorcycle Applications
Gianluigi Zoli, Matthew Symonds, May Turner, Ieuan Adams, Nick Solomon, Mark Leonard, Cliff
Newman
Castrol Ltd

Copyright © 2011 SAE Japan and Copyright © 2011 SAE International

ABSTRACT temperatures typical of sport operations. At the same time,

oil must protect small and heavily loaded gears from wear,
pitting and scuffing, as well as to ensure the correct level of
The highly competitive sport motorcycle market demands
friction between clutch plates in order to allow their smooth
continuous increase in engine performance. At the same
engagements, prevent slip under high load and protect
time, legislation in the main industrialized countries
friction material from premature wear.
establishes stringent limits for gaseous emissions from two
wheelers. As already witnessed in the passenger car
industry, motorcycle manufacturers are giving increased In a typical sport motorcycle engine, combination of high
focus to the lubricant as a potential tool to improve engine temperature, speed and load reduces oil film thickness in
performance and reduce gaseous emissions. When compared critical areas such as main and cone rod bearings and
to a conventional, high viscosity formulation, an optimized between gear teeth, making the application of ultra low
low viscosity, low friction engine oil can bring significant viscosity engine oils challenging from a durability point of
increase in power release and fuel economy, as well as a view. Moreover, strong mechanical shear caused by gear
reduction in tailpipe emissions. However, the use of highly teeth further reduces oil viscosity during service, increasing
efficient lubricants in sport motorcycle applications needs the risk of high wear rates and component failures if low
careful evaluation since motorcycle engines with an viscosity grades are used over extended periods.
integrated clutch and gearbox require the lubricant to
achieve very precise viscometric and frictional limits. This Despite challenges listed above, the demand for increased
paper describes the development of a new concept for high engine responsiveness, reduced fuel consumption and
performance motorcycle engine oils, capable of delivering a emissions, makes the application of low viscosity engine
tangible increase in engine efficiency whilst maintaining a oils to motorcycles a desired step. In order to allow safe
high level of wet clutch operability along with good engine application of low viscosity engine oils to sport motorcycles,
and gearbox durability. The development process included very careful formulation approach and extensive testing
the formulation of a specific additive package with an using representative bench motorcycle test are therefore
optimized friction coefficient and the definition of a base oil required.
mix with a suitable viscometric profile. A number of
candidate formulations were then screened using Castrol’s FORMULATION APPROACH
proprietary motorcycle engine test methods for power
release and clutch compatibility. A final candidate was field Three different fully synthetic, low viscosity formulations
tested for 30,000 km using a high speed track riding cycle. were developed for this project and tested for clutch
The candidate formulation showed a statistically significant compatibility and power / acceleration response; most
increase in engine power release over a number of high promising candidate was selected for extended engine /
performance engine oils. Wet clutch compatibility and gearbox durability testing. All candidates were based on a
powertrain durability was also equal or better to that newly developed additive system specifically designed for
obtained with high performance benchmark products. advanced motorcycle applications. The additive system
comprises metallic detergents, ashless dispersant, zinc –
INTRODUCTION phosphorus anti-wear and a combination of ashless and
metallic antioxidant. The additive system is formulated with
A modern motorcycle engine oil must fulfill a range of reduced phosphorus level, with the objective of allowing
conflicting requirements and calls for dedicated formulation good
approach. Unlike passenger car where lubrication of engine
and gearbox is separated and clutch is dry type, a typical
sport motorcycle powertrain uses the same fluid to lubricate
engine and gearbox as well as to cool the clutch. A good
motorcycle lubricant must be able to protect engine from
wear and deposits under very high speed, load and

SETC2011
75
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Table 1 – Main Characteristics of candidate oils and benchmark products

Oil code E0888M/001B E0888M/002A E0404M/046A E0698M/038A E0698M/103A E0404M/041A E0404M/042A
Type Candidate Candidate Candidate Internal Internal Market Market
benchmark benchmark benchmark benchmark
Viscosity Grade 0W20 0W30 5W30 5W40 10W30 0W20 5W40
API Classification SL SL SL SL SJ Nd Nd
JASO T903 : 06 MA2 type MA2 MA2 MA2 MA2 Nd MA
classification performance
Additive type A* A A B B Unknown Unknown
Base oil type Full synth. Full synth. Full synth. Full synth. Part Synth. Full synth. Full synth.
Kinematic Viscosity @ 8.21 9.98 10.73 13.0 10.5 8.83 13.32
100 °C, cSt
Kinematic Viscosity @ 41.98 54.96 61.74 77 66 45.3 72.7
40°C, cSt
Viscosity Index 175 172 167 170 146 179 187
HTHS Viscosity. cP 2.6 2.9 3.15 3.5 3.3 2.6 3.7
Noack evaporation, % wt 6 6 7 9 13 6.5 4
Phosphorus, % wt 0.096 0.096 0.096 0.12 0.12 0.15 0.093
*HTHS viscosity below JASO T903 limits

catalyst compatibility. In order to compare performance of PERFORMANCE TESTING –

the candidate oils in the different areas, a range of internal
(existing Castrol motorcycle products) and external CLUTCH
benchmarks (top tier products supplied from market) were
evaluated. The main characteristics of candidate oils and Clutch performance was evaluated using SAE#2 rig test and
benchmark products are reported in table 1. Castrol in house clutch slip test.

TEST METHODOLOGY SAE#2 test.

In order to evaluate performance of candidate oils and This established industry test is used in JASO T903
benchmark products, a combination of test methods specification in order to classify oils into high (MA2),
including rigs, in house bench tests and on-track field testing medium (MA1) and low (MB) friction. In the SAE#2 rig, a
was adopted. A summary of the test suite used for this work motored set of friction plates is loaded against a set of steel
is given in table 2. This set of tests has been designed with discs immersed in oil, and the friction is measured under
static (fixed load and speed) and dynamic (variable load and
Table 2 – Summary of rig, bench and field tests

Performance Area Test type Rig type / Engine / Test Method Measured parameter
Bike type
Rig SAE#2 JASO T904 Clutch friction
Clutch
Bench Superbike, water Castrol in- Clutch slip
cooled house
Power and acceleration Bench 600cc supersport, Castrol in- Power released and acceleration times
water cooled house
On-track 600 cc supersport, Castrol in- Used oil analysis, engine/ driveline visual
Engine and driveline durability
field test water cooled house rating, piston deposit rating

the objective of evaluating motorcycle oil performance in speed) conditions. In the JASO T904 clutch friction test
the key areas such as engine and driveline durability, clutch method, the conditions alternate between dynamic and static
compatibility and engine efficiency (power and mode, for a total of 1000 cycles. Oil Temperature is kept at
acceleration), giving a comprehensive and realistic 100°C, simulating typical motorcycle oil sump temperature.
representation of engine oil performance in real life. For each of the candidate oil, a low and a high friction
reference is run and results are used to calculate Dynamic,
Static and Stop Time Friction Indexes. Because response of
SAE#2 is known to be influenced by chemistry rather than
viscosity, for this experiment only one candidate,
E0888M/002A was run along with selected benchmark
SETC2011
76
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Table 3 – SAE#2 Results for Candidate and Benchmark Products

MB MA 002A 038A 041A 042A

(candidate) (internal (market (market
MA1 MA2 benchmark) benchmark) benchmark)
Static Friction Index 0.50-1.15 1.15-1.70 1.70 -2.5 1.9 2.02 1.51 0.78

Dynamic Friction Index 0.50-1.45 1.45-1.80 1.80-2.5 1.94 2.03 1.95 1.16

Stop Time Index 0.50-1.55 1.55-1.90 1.90-2.5 1.92 1.99 1.86 1.14

Classification MA2 MA2 MA1 MB

products. Results are reported in table 3. Static, Dynamic clutch slip; in particular, 041A gave clutch slip over three
and Stop Time Indexes for the candidate formulation all fall times 002A, in line with results from SAE#2 where this oil
within JASO MA2 high friction category, and results has shown very low friction coefficients.
obtained can be read across to other candidates based on
same additive chemistry. PERFORMANCE TESTING – POWER
Castrol in house clutch slip test AND ACCELERATION

The Castrol Fired Clutch Test has been developed as The effect of candidate and reference oils on power release
predictive tool for lubricant motorcycle wet clutch behaviors and acceleration was evaluated using in-house bench test
in real life conditions. The test uses a 1000 cc superbike procedure. This test uses a 600 cc, water cooled, in-line
engine run under high speed and load conditions. Clutch slip four cylinder engine installed on a transient dynamometer.
is measured as the rpm difference between the observed The procedure consists of a series of accelerations
engine speed and the actual dynamometer speed. Each of the
candidate oils is run at three different engine speeds and Figure 2 - Supersport Pow er and Acceleration Test
three oil temperature and data is expressed as a percentage Results
increase or reduction in slip against an internal reference oil
6
(an high friction oil which is known to have excellent clutch
% difference vs 10W40 reference

5
performance in the field). At the end of test, the entire
dataset is statistically evaluated in order to assess 4
Power
differences between candidates. Results obtained (average 3 Acceleration
2
Figure 1 - Fired Clutch Slip Test Results 1
0
-1
% difference in clutch slip vs reference

e
k)

k)
40
)

nc
20

30

30

ar
ar

5W

re

170
W

m
m

fe
(0

ch

(0

ch
(5

-2
ng

re
en

en
1A

2A

6A

ci

w
(b

(b
Ra

150
00

00

04

lo

-3
1A

2A

50
1
er
04

04

W
w

20

130
Po

-4
110
90
70 performed in fixed gear from a rolling start, during which
the engine accelerates up to a set maximum speed; the
50
average energy released and the time to reach maximum
30 speed are recorded for each run. Single run results are then
10 averaged for each candidate test. A reference oil is run every
-10 041A (mk t 042A (mkt 038A (internal 002A (0W 30)
three candidates in order to measure test repeatability, and
benchmark) benc hmark) benchmark ) candidate results are expressed as % gain or loss versus
closest reference run. At the end of test, entire dataset is
statistically evaluated in order to assess difference between
increase of clutch slip for all temperatures and speeds) are candidates and reference. Results are reported in figure 2;
summarized in Figure 1. Candidate 002A gave very good candidate 001A and 002A along with benchmark 042A, all
clutch slip control, equivalent to that of reference oil and gave statistically significant increases in power and
close to that of internal benchmark. Both market acceleration versus 10W40 reference oil.
benchmark products gave significantly higher levels of
SETC2011
77
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

PERFORMANCE TESTING –ENGINE closed test track, following a predetermined high speed
riding cycle, for a total length of 30,000 Km. Bikes were run
AND DRIVELINE DURABILITY in parallel in order to accumulate mileage under the same
climatic conditions, allowing a more precise comparison of
After review of the power and acceleration data candidate end of test results. Oil Samples were taken at 1000 Km
002A was submitted to engine and driveline durability test. intervals and analyzed in order to evaluate chemical
Candidate 002A was chosen because it gave a significant physical parameters such as viscosity, fuel dilution, acidity,
increase in engine performance, whilst retaining HTHS alkalinity and wear metals. At end of the test the two
viscosity within JASO T903 limits, and offering good engines were stripped and wear of key components visually
credential for sufficient hardware durability. A supersport, evaluated, with rating assigned on a scale 0 (excellent) to 5
600 cc in-line 4 cylinder bike model was selected for the (unsatisfactory). Gearbox pitting (macro and micro) was
test, and two bikes were run by professional testers on assessed using % area method, and piston measured and
Table 4 - Engine Component Rating
OIL E0404M/002A (Candidate) 038A (Internal Benchmark)
PART COMMENT RATING COMMENT RATING

Crankshaft
Main journals All with wear lines, No 5 the worst 4 Trace discolouration, clear defined 3.5
running marks, light scoring
Big end All with wear lines and traces 4.5 Trace discolouration, very clear defined 4
journals discoloration run marks, light to medium scoring,
Main bearings Top- Bearing 5 with 70% heavy polish, 3.5 Large polished areas, through first layer 3.5
all others good. Bottom - polished areas
and some trace cavitation on 2, 3 & 5
Big End Top- 75% heavy polish, large patches 4 Large polished areas with large patches 5
Bearings down to next layer- cavitation. Bottom- of first layer gone, light to medium
Large polished areas, medium to large scoring
patches down to next layer
Small End Covered in Light brown lacquer, traces 2 Covered in Light brown lacquer, outside 2.5
Bearings of polished patches, outside of rod eye of rod eye area -black
area –black

Gudgeon Pins Fine surface visual scoring, traces 2.5 All look to have "fine" scuffing/scoring 3
discolouration covering entire surfaces of both piston
bearing area and rod eye
Average 3.5 3.6
Camshaft
Inlet lobes Polished wear only, no scoring or pitting 2 Polished wear only, no scoring or pitting 2
Exhaust lobes Polished wear only, no scoring or pitting 2.5 Polished wear only, no scoring or pitting 2.5
Cam journals Polished with some light scoring 2.5 Light scoring 3
Cap bearings Light running marks only 2 Light scoring mostly exhaust side 2.5
Head bearings Trace contact 1 Trace contact 1.5
Bucklets inlet Trace discolouration, trace polish 2 Trace discolouration, trace to light polish 2
Bucklet Discolouration, some heavy polish 3 Trace discolouration, trace wear 2.5
h chain
Cam Wear detectable 3 Wear detectable 3
Tensioners Light to medium wear 3 Light to medium wear 3
Average 2.3 2.4
Cylinder bore All with heavy scoring on T & AT sides 4 All with medium to heavy scoring on T 3,5
Clutch assy ih f li h b l & A id l
Basket Medium wear on drive pillars 3.5 Medium wear on drive pillars 3
Hub Light wear on drive splines 2.5 Light wear on drive splines 2.5
Friction Plates Light wear on drive ears, no distress 2.5 Light wear on drive ears, no distress 2.5
Steel Plates Trace wear on drive ears 2.5 Trace wear on drive ears 2.5
Average 2.8 2.6
SETC2011
78
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Table 5 – piston cleanliness rating

rated according to DIN 51-361-02-VB method
Piston # Candidate Reference
where 100 = completely clean piston and 0 = completely
dirt. Summary of engine wear, gear pitting and piston 1 63 64
cleanliness rating is reported in tables 4,and 5 respectively, 2 62 67
whilst figure 3 shows trend of key wear element 3 60 67
accumulation throughout test. Overall, the candidate oil
4 58 74
gave levels of wear similar to those observed on reference
bike. Engines showed a certain level of bearing wear, but Average 61 68
since wear is comparable for the two oils, it can be deposit
attributed to hardware rather than oil effect. The gearbox of
candidate bike showed a higher level of micro pitting, but no
macro pitting was observed and the overall conditions of
SUMMARY
driveline are to be considered good. for the number of miles
accumulated. Candidate bike also gave slightly lower piston This work demonstrated that optimized low viscosity
rating than reference, but both engines scored high in engine oils formulated with specific technology can bring a
cleanliness with low level of piston deposits and no ring statistically significative increase in engine power and
sticking acceleration in comparison with internal reference and high
performance, commercially available benchmark products.
In particular, one of candidates developed, showed power
and acceleration statistically equivalent to benchmark
Table 6 – gear pitting rating, % area product of lower viscosity grade (0W30 versus 0W20),
Candidate Reference showing at same time much higher friction coefficient and
consequently a better control of clutch slip under high load.
Gear # Micro Macro Micro Macro This would indicate that tailor made formulation approach
pitting pitting pitting pitting can allow to develop highly efficient motorcycle engine oils
6th main 45 0 60 0 capable of retaining a high level of wet clutch compatibility.
4th main 50 0 43,8 0 In order to assess impact on durability, candidate oil was run
3rd main 29,7 0 10,2 0 for 30,000 Km on a supersport motorcycle using a high
6th counter 44,6 0 40 0 speed on-track riding cycle; the bike run throughout the test
without problems and engine inspection revealed key
4th counter 18,2 0 5 0 components to be in good conditions with overall engine
3rd counter 29,7 0 0 0 rating equivalent to that obtained on reference bike run with
oil conform to OEM
recommendation. This data indicates that, on the specific
hardware used for this work, the candidate formulation can
guarantee a good level of protection from engine wear and
gear pitting and could potentially be suitable for use as
service fill oil. However, in order to understand impact on
engine and driveline durability of ultra-low viscosity engine

Fig. 43--Wear
Figure Elements
Wear Elements forfor Reference
Reference and Candidate
and Candidate Oil Oil
90 Aluminium, candidate
80 Iron, candidate
Concentration (ppM)

70 Copper, candidate
Wear Element

60 Aluminium, reference

50 Iron, reference
Copper, reference
40
30
20
10
0
0 5000 10000 15000 20000 25000 30000 35000
Test Lenght (Km)

SETC2011
79
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

oils in real life riding conditions and on a wider range of

commercially available motorcycle models, more durability
testing on selected representative hardware is be necessary.

REFERENCES
1. F. Nonaka, M Akagi, I. Okamoto, [Link], Y.
Yashiro, “Study on Engine Oil for Motorcycles:
Engine, Characteristics and New Specification
Oils”, SAE Paper No 961217 (1996)
2. Moritsugo Kasai, Katsuhiko Nakajima, “Fuel
Saving Four Stroke Engine Oil for Motorcycle”,
SAE 2006-32-0014 (2006)

CONTACT INFORMATION
[Link]@[Link]

������������������������������������������������������������������������������ ��������������������������������������������������������������������������������
������������������������������������������������������������������������������������� ����������������������������������������������������������������������������������������
������������������������������������������������������������������������������������ ���������������������
�����������������������������������������
�������������������������������������������������������������������������������
�������������������������������
�������������������������������������������������������������������������������������� �����������������
����������������������������������������������������������������������������������� ������������������������������
�������������� �����������������������������������
��������������

80
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

JSAE 20119571 / SAE 2011-32-0571

An Improvement of a Small Displacement Engine’s Efficiency

with a Super Charging System

Changjoo Ahn, Takashi Suzuki, Yasuhumi Oguri,

Hiroki Toshitani, Tatsuyoshi Nakahuku, Yusuke Nakano
Sophia University (Japan)

Copyright © 2011 SAE Japan and Copyright © 2011 SAE International

1. Abstract 2. Introduction
Many environmental problems, such as global Formula SAE®, started back in the U.S in 1981
warming, drain of fuel and so on, are apprehended as a “student educational program by real manufac-
in all over the world today, and down-sizing is one of turing experience”, is a competition which is com-
the wise ways to deal with these problems. It is sig- peted by “all aspects of car manufacturing” includ-
nificant that a decrease of the engine power must ing research, design, manufacturing, testing, devel-
not be produced by using a small displacement en- oping, marketing, management and finances. After
gine, so more efficient engine system should be de- the first competition, it has become large and inter-
signed to increase the torque with a little fuel. national with over 100 participating teams. Formula
This study achieves an improvement of efficien- SAE® regulation based competition has been taking
cy for mounting the super charging system on the place also in Japan as Formula SAE® Japan since
small displacement engine. As a result, comparing a 2003, and has become a big event with over 70
super charged engine and a naturally aspirated one teams participating.
to drive the same course and laps, fuel consump- The racing course is designed to be technical
tions are 2547 [cc] and 3880 [cc], respectively, and with many chicanes and corners and few straight
an improvement of fuel consumption is 52%. areas, for that the regulation limits the maximum
Designing points to mount super charging sys- vehicle speed to 100[km/h]*1. For this reason, mid-
tem is introduced below. low engine speed range (around 5,000㨪9,000[rpm])
is more used than high speed range (over10,
1. It can be forecasted that intake air blow-by gas 000[rpm]). Improvement of torque in mid-low revo-
at the combustion chamber is increased in low en- lution range which is mostly used in the competition
gine speed because engine for motor cycle is used. is aimed in this study.
Therefore, the valve timing (cam profile) is changed.
2. Stress analysis and fluid flow analysis were done
3. Engine and Super Charger Specifications
for the new intake chamber made by powder lami-
nated casting. The sectional area of the new intake Adoption of super charger was considered to
chamber is set to become gradually smaller, in order improve the mid-low speed range. Specifications of
to increase the flow rate. Engine and Super Charger used in this study are
indicated in Table 1 and Table 2 respectively.
3. Inter cooler is adapted, because intake air com-
Genuine engine has a feature that the valve
pressed with super charger becomes hot and it caus-
over lap is set to large angle to have high-power in
es a decrease of charging efficiency and an increase
high revolution range. Two leaves roots type, which
of the knocking of an engine.
does not need the lubricating oil cooler system, is
4. The fuel injection map and the ignition timing adopted considering the reliability of super charger.
map are optimized for changes above. The outline diagram of test apparatus is shown
in Fig.1 and configuration of engine and super
charger is shown in Fig.2. Lubricating oil and cool-
These changes above made it possible to make high- ing water were externally circulated to have enough
er torques for a wider range of engine speeds, and cooling amounts. Also, they were controlled to have
this leads to low fuel consumption. It is because that constant temperature of 353[K] during the experi-
coming down torque peak and average engine speed ment by measuring the temperature at both outlets.
reduce engine friction. In this way, the value of fuel Furthermore, torque and power were measured un-
consumption with super charger can be lower than der controlled engine speed by connecting the en-
the one with naturally aspirated. gine to eddy-current dynamometer.
SETC2011
81
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Table 1 Engine Specifications 4. Effect of Cam Profile

㪜㫅㪾㫀㫅㪼㩷㪤㫆㪻㪼㫃 㪰㪘㪤㪘㪟㪘㩷㪌㪪㪣㪈㪄㪘㪜㪈
㪚㫆㫅㪽㫀㪾㫌㫉㪸㫋㫀㫆㫅 㪠㫅㪄㫃㫀㫅㪼㪋㪺㫐㫃㪅㪃㩷㪋㪄㫊㫋㫉㫆㫂㪼㪃㩷㪣㫀㫈㫌㫀㪻㪄㪚㫆㫆㫃㪼㪻 Blow-by ratio during naturally-aspiration is in-
㪛㫀㫊㫇㫃㪸㪺㪼㫄㪼㫅㫋㩷㩿㪙㫆㫉㪼㬍㪪㫋㫉㫆㫂㪼㪀
㪚㫆㫄㫇㫉㪼㫊㫊㫀㫆㫅㩷㪩㪸㫋㫀㫆
㪌㪐㪐㪅㪏㩷㪲㪺㪺㪴㩷㩿㪍㪌㪅㪌㩷㫏㩷㪋㪋㪅㪌㩷㪲㫄㫄㪴㪀
㪈㪉㪅㪋
dicated in Fig.3. Blow-by ratio was calculated by
㪤㪸㫏㫀㫄㫌㫄㩷㪧㫆㫎㪼㫉 㪏㪏㪅㪊㩷㪲㫂㪮㪴㩷㩿㩷㪈㪉㪉㪲㪿㫇㪴㪀㩷㪗㩷㪈㪊㪇㪇㪇㫉㫇㫄 measuring HC in the exhaust gas using gas analyz-
㪤㪸㫏㫀㫄㫌㫄㩷㪫㫆㫉㫈㫌㪼 㪍㪏㪅㪌㪲㪥㫄㪴㩷㪗㩷㪈㪉㪇㪇㪇㫉㫇㫄 er. From this result, increasing blow-by rate was
㪭㪸㫃㫍㪼㩷㪫㫉㪸㫀㫅 㪈㪍㪭㪸㫃㫍㪼㪄㪛㪦㪟㪚
㪣㫌㪹㫉㫀㪺㪸㫋㫀㫆㫅㩷㪪㫐㫊㫋㪼㫄 㪮㪼㫋㪄㪪㫌㫄㫇 confirmed as engine speed becomes lower. For ge-
㪫㫉㪸㫅㫊㫄㫀㫊㫊㫀㫆㫅 㪍㩷㪪㫇㪼㪼㪻㩷㪪㪼㫈㫌㪼㫅㫋㫀㪸㫃 nuine engine, it is general to set the valve over lap
㪠㫅㫋㪸㫂㪼
㪦㫇㪼㫅 㪊㪐䉙㪙㪫㪛㪚
㪚㫃㫆㫊㪼 㪌㪐䉙㪘㪙㪛㪚 angle to high angle (73degrees) as shown in Table.1,
㪭㪸㫃㫍㪼㩷㪫㫀㫄㫀㫅㪾
㪜㫏㪿㪸㫌㫊㫋
㪦㫇㪼㫅 㪍㪍䉙㪙㪙㪛㪚 for the aim to gain high-torque at high engine speed
㪚㫃㫆㫊㪼 㪊㪋䉙㪘㪫㪛㪚 range.

Table 2 Super Charger Specifications

㪫㪯㪇㪋 㪎㩼
㪪㫌㫇㪼㫉㪺㪿㪸㫉㪾㪼㫉
㩿㫄㪸㪻㪼㩷㪹㫐㩷㪦㪾㫌㫉㪸㪄㪚㫃㫌㫋㪺㪿㩷㪚㫆㪅㪃㩷㪣㫋㪻㪅㪀
㪍㩼 㪤㪼㪸㫊㫌㫉㪼㪻
㪛㫀㫊㪺㪿㪸㫉㪾㪼㩷㪩㪸㫋㪼 㪋㪈㪇㩷㪲㪺㪺㪆㫉㪼㫍㪴

㪙㫃㫆㫎㪄㪹㫐㩷㪘㫀㫉㩷㪩㪸㫋㫀㫆
㪤㪸㫏㩷㪩㪧㪤㩷㩿㪪㪼㫈㫌㪼㫅㪺㫐㪀 㪈㪉㪃㪇㪇㪇㩷㪲㫉㫇㫄㪴 㪌㩼 㪘㫇㫇㫉㫆㫏㫀㫄㪸㫋㫀㫆㫅
㪤㪸㫏㩷㪩㪧㪤㩷㩿㪠㫅㫊㫋㪸㫅㫋㪀 㪈㪌㪃㪇㪇㪇㩷㪲㫉㫇㫄㪴
㪤㪸㫏㩷㪧㫉㪼㫊㫊㫌㫉㪼㩷㪩㪸㫋㫀㫆㩷㩿㪪㪼㫈㫌㪼㫅㪺㫐㪀 㪈㪅㪏 㪋㩼
㪤㪸㫏㩷㪧㫉㪼㫊㫊㫌㫉㪼㩷㪩㪸㫋㫀㫆㩷㩿㪠㫅㫊㫋㪸㫅㫋㪀 㪉
㪤㪸㫏㩷㪛㫀㫊㪺㪿㪸㫉㪾㪼㩷㪩㪸㫋㪼
㪊㩼
㪉㪊㪌㩷㪲㫄㪵㪊㪆㪿㪴
㪸㫋㩷㫋㪿㪼㩷㪤㪸㫏㩷㪩㪧㪤 㪉㩼
㪮㪼㫀㪾㪿㫋 㪌㪅㪊㩷㪲㫂㪾㪴
㪈㩼
㪇㩼
Fuel consumption amount was measured by 㪋㪇㪇㪇 㪍㪇㪇㪇 㪏㪇㪇㪇 㪈㪇㪇㪇㪇
fuel flow gauge mounted on fuel line. Air-fuel ratio 㪜㫅㪾㫀㫅㪼㩷㪪㫇㪼㪼㪻㩷㪲㫉㫇㫄㪴
was measured via Air-fuel ratio meter attached to
exhaust pipe. Supercharger drive force is gained Fig.3 Blow-by Air Ratio at NA Condition
from crankshaft axle, and pulley ratio which is en-
gine speed for super charger speed is 1 for 0.89.
Moreover, inlet air temperature of intercooler and For that reason, by adopting the supercharger,
temperature of each cylinder’s exhaust manifold torque in mid engine speed range has increased
were measured. while significant blow-by phenomenon was observed
in low engine speed range. Therefore, controlling the
blow-by phenomenon by diminishing the size of
Fuel Tank valve over lap of low cam profile was considered.
Namely, from comparing with the data of natural
Fuel Flow Meter Inlet Surge Tank Oil Inlet Temp. aspiration, change the valve over lap angle from 73
Water Inlet Injector Dynamometer
degrees to 42 degrees by slowing the inlet valve
opening by 19 degrees, and hastening the exhaust-
㪧
Temp.
㪧 㪧

ing valve closing by 12 degrees. Also, for the aim to

Intake Pressure
Oil Cooler
Water Cooler decrease the acceleration of the valve opening and
Knock closing, reduction of the maximum lift by 0.5mm for
Oil Outlet Temp.
Super Sensor intake side and 0.6mm for the exhaust side due to
Charger
Muffler
the reduction of valve over lap was conducted. From
Water Outlet Temp.
Throttle this result, angle of range during valve over lap de-
Laminar Flow Meter Duct creased from 319.06[mm࡮rad] to 140.46[mm࡮rad]
Fig.1 Outline Diagram of Test Apparatus (about 56 percent) as shown in Fig.4.
Next Fig.5 shows torques with genuine profile
cam and new valve profile. Flat line of torque, nor-
Intake Air Temp. Sensor
Air Temp. Sensor malized as one, is used in order to compare the tor-
(Intercooler IN&OUT)
ques. For this result, with the new profile cam, it
reduces the decreasing rate in the low engine speed,
and this means blow-by ratio is inhibitedMoreover,
Inter-Cooler
Engine Super Charger torque depression in high engine speed range has
Throttle
become lenient. This is because blow-back has de-
Blow-Off Valve
creased owing to hastening the time of intake valve
Exhaust [Link] closing.
Restrictor
Crank Shaft
Fig.2 Configuration of Engine and Super Charger

SETC2011
82
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

10 Intake std. 5. Modification of Intake Collector Tank

Exhaust std. 2009 year’s collector tank shaped cylinder used

8
in this research is shown in Fig.7. The experiment
6
Intake mod. was conducted by measuring the exhaust tempera-
ture under constant fuel supply condition.
Lift [mm]

Exhaust mod.
4 For this, air-fuel ratio in each cylinder is esti-
Intake mod. mated, and the variability of intake air at each cy-
2 [Link] linder is evaluated.

0
0 90 180 270 360 450 540 630 720
Angle [deg]
Fig.4 Inlet and Exhaust Valve Profile

㪈㪅㪉

Fig.7 Old Collector Tank

㪫㪆㪫㪘㪻㫍㪅

㪇㪅㪏
㪣㫆㫎㩷㪚㪸㫄
㪇㪅㪍 Exhaust gas temperature time change of each
㪪㫋㪸㫅㪻㪸㫉㪻
cylinder at WOT is indicated in Fig.8. The experi-
ment indicates the case when changing the engine
㪇㪅㪋 speed from 5000[rpm] to 12000[rpm] in approx-
㪉㪇㪇㪇 㪋㪇㪇㪇 㪍㪇㪇㪇 㪏㪇㪇㪇 㪈㪇㪇㪇㪇 imately 25 seconds. From the Figure, large tempera-
㪜㫅㪾㫀㫅㪼㩷㪪㫇㪼㪼㪻㩷㪲㫉㫇㫄㪴 ture decrease can be seen in the furthest cylinder
Fig.5 Comparing Torques at Low Cam (cylinder No.1) from the air intake entrance. Miss
and Normal Cam fire, occurred by the reduction of air intake of cy-
linder No.1 which led to rich mixture, can be consi-
dered as the reason.
Moreover, shown in Fig.4, the experiment delay-
ing the intake valve open 21 degrees from the crank
shaft angle, making valve overlap area from 140.46 㪐㪇㪇 㪍㪇
[mm࡮rad] to 35.17[mm࡮rad], was done. Fig.6 shows 㪏㪇㪇
the torque difference before and after changing the intake 㪌㪇
㪎㪇㪇
cam phase.
The cam with slower crank angle showed lower

㪜㫅㪾㫀㫅㪼㩷㪫㫆㫉㫈㫌㪼㪲㪥㫄
㪍㪇㪇 㪋㪇
㪜㫏㪿㪸㫌㫊㫋㩷㪫㪼㫄㫇㪅㪲㷄㪴

torque and power. This is because residual gas in- 㪌㪇㪇

side the cylinders increased since the valve overlap
㪊㪇
㪋㪇㪇
was too small to let the residual gas out. 㪊㪇㪇
㪜㫏㪿㪸㫌㫊㫋㩷㪫㪼㫄㫇㪅㩷㪺㫐㫃㪅㪈㩷
㪉㪇
㪜㫏㪿㪸㫌㫊㫋㩷㪫㪼㫄㫇㪅㪺㫐㫃㪅㪉
㪉㪇㪇 㪜㫏㪿㪸㫌㫊㫋㩷㪫㪼㫄㫇㪅㩷㪺㫃㫐㪅㪊㩷
㪜㫏㪿㪸㫌㫊㫋㩷㪫㪼㫄㫇㪅㪺㫐㫃㪅㪋 㪈㪇
㪏㪇 㪎㪇 㪈㪇㪇 㪜㫅㪾㫀㫅㪼䇭㪫㫆㫉㫈㫌㪼

㪇 㪇
㪎㪇 㪍㪇 㪇 㪌 㪈㪇 㪈㪌 㪉㪇 㪉㪌 㪊㪇
㪫㫀㫄㪼㪲㫊㪴
㪍㪇
㪌㪇 Fig.8 Exhaust Temperature
㪌㪇
㪜㫅㪾㫀㫅㪼 㪫㫆㫉㫈㫌㪼䋨㪥㫄䋩

㪜㫅㪾㫀㫅㪼 㪧㫆㫎㪼㫉䋨㫂㪮䋩

㪋㪇
㪋㪇

㪊㪇
㪚㪸㫄㪶㫇㪿㪸㫊㪼㪶㪺㪿㪸㫅㪾㪼㩷㪫㫆㫉㫈㫌㪼 㪊㪇 Next, gas flow analysis was conducted using
㪚㪸㫄㪶㫇㪿㪸㫊㪼㪶㫊㫋㪻㪅㪫㫆㫉㫈㫌㪼
㪉㪇 FLUENT in order to reduce the variability of air
intake amount among each cylinder. Simulation for
㪉㪇
㪚㪸㫄㪶㫇㪿㪸㫊㪼㪶㪺㪿㪸㫅㪾㪼㩷㪧㫆㫎㪼㫉
㪈㪇 㪚㪸㫄㪶㫇㪿㪸㫊㪼㪶㫊㫋㪻㪅㪧㫆㫎㪼㫉
㪈㪇 collector volume was done by setting 1.2 times larg-
㪇 㪇 er amount than that of last year’s, referring to re-
㪉㪇㪇㪇 㪋㪇㪇㪇 㪍㪇㪇㪇 㪏㪇㪇㪇 㪈㪇㪇㪇㪇 sult of air-fuel ratio estimated from exhaust tem-
㪜㫅㪾㫀㫅㪼 㪪㫇㪼㪼㪻䋨㫉㫇㫄䋩 perature. The simulation was evaluated by following
Fig.6 Effect of Cam Timing aspects: smoothness of flow line, non-existence of
with Slower Crank Angle stagnation points, and variation of intake air
amount of each cylinder. Model of this year’s intake
collector tank used in this study is shown in Fig.9.
SETC2011
83
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

In order to stop the flow rate from decreasing, sec- Simulation result of each cylinder’s mass flow
tion area of collector tank is decreased incremental- rate is shown in Fig.11. Simulation was done under
ly along downstream. the condition of applying 1.3[bar], which is the same
pressure as manifold pressure during supercharger
experiment, to this year’s and last year’s intake
pipe’s entrances. As a result, variation in intake air
mass has decreased from 5% to 3%. Moreover, aver-
age mass flow has increased 9%.

㪇㪅㪊㪉
㪇㪅㪊㪈

㫄㪸㫊㫊㩷㪽㫃㫆㫎㩷㫃㪸㫋㪼㩷㪲㫂㪾㪆㫊㪴
㪇㪅㪊
㪇㪅㪉㪐
㪇㪅㪉㪏
㪇㪅㪉㪎
Fig.9 New Collector Tank 㪇㪅㪉㪍
㪇㪅㪉㪌
㪈 㪉 㪊 㪋
And Initial condition and boundary condition of 㪺㫐㫃㪅㪥㫆㪅㪲㪄㪴
the model used for calculation are shown table 3. 㪦㫃㪻㩷㪺㫆㫃㫃㪼㪺㫋㫆㫉㩷㫋㪸㫅㫂 㪥㪼㫎㩷㪺㫆㫃㫃㪼㪺㫋㫆㫉㩷㫋㪸㫅㫂
Result of fluent simulation is indicated in Fig.11 Intake Mass Flow Rate
Fig.10. It can be considered that air turbulence in-
side the collector tank is moderate, from smooth
flow line around funnel in Fig.10. 6. Non-equal Length Exhaust Manifold
Non-equal length exhaust manifold was
adopted for the purpose to increase the exhaust
Table 3 Analysis Condition
pressure and decrease the blow out the flesh air. It
㪘㫅㪸㫃㫐㫋㫀㪺㪸㫃㩷㪤㫆㪻㪼㫃 㫂㪄ε 㫄㫆㪻㪼㫃 is shown in Fig.12. From right to left, cylinders are
㪙㫆㫌㫅㪻㪸㫉㫐㩷㪚㫆㫅㪻㫀㫋㫀㫆㫅 named as one to four. Exhaust manifold was visua-
㪠㫅㫃㪼㫋㩷㪲㪹㪸㫉㪴 㪈㪅㪊 lized the flow in unsteady analysis using 3 dimen-
㪦㫌㫋㫃㪼㫋㩷㪲㪹㪸㫉㪴 㪇㪅㪏㪌 sion thermo-fluid analysis software, FLUENT, be-
㪤㪼㫊㪿㩷㪥㫌㫄㪹㪼㫉 㪉㪇㪇㪃㪇㪇㪇 cause of the effect of exhaust pressure wave to next
exhaust cylinder. Fig.13 shows exhaust mass flow
rates in each cylinder at high engine speeds. The
axis of ordinate is the value divided exhaust mass
flow rate of own cylinder by the mean value of four
cylinders. The simulation was done in high engine
speed range especially; 7000[rpm], 10000[rpm], and
14000[rpm], because the variability of among each
cylinder should be mainly concerned about.
In the result, it has little difference at each en-
gine speed, and exhaust mass flow rates of each cy-
linder were confirmed in acceptable value; maxi-
mum 0.065%.

Fig.10 Flow Line in the Collector Tank

Intake manifold was manufactured using glass

beads containing nylon material, which enables to
reproduce precise 3 dimensional objects. Because
supercharger is attached to the engine, stress simu-
lation to ensure whether collector has adequate
strength against its inner pressure was conducted, Fig.12 Exhaust Manifold
and attached ribs to parts where more strength is
necessary to increase reliability.
SETC2011
84
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

㪈㪅㪇㪏 㪏㪇㪅㪇㩷
㪎㪇㪇㪇㫉㫇㫄 㪎㪇㪅㪇㩷 㪩㪼㫊㫀㪻㪼㫅㪺㪼㩷㪫㫀㫄㪼
㪈㪅㪇㪋
㪈㪇㪇㪇㪇㫉㫇㫄 㪍㪇㪅㪇㩷 㪭㪸㫇㫆㫉㫀㫑㪼㪻㩷㪫㫀㫄㪼
㪈 㪈㪋㪇㪇㪇㫉㫇㫄
㪌㪇㪅㪇㩷

㪫㫀㫄㪼㩷㪲㫄㫊㪴
㪇㪅㪐㪍
㪞㪆㪞㪸㪻

㪋㪇㪅㪇㩷
㪇㪅㪐㪉 㪊㪇㪅㪇㩷
㪇㪅㪏㪏 㪉㪇㪅㪇㩷
㪈㪇㪅㪇㩷
㪇㪅㪏㪋
㪺㫐㫃㫀㫅㪻㪼㫉㪈 㪺㫐㫃㫀㫅㪻㪼㫉㪉 㪺㫐㫃㫀㫅㪻㪼㫉㪊 㪺㫐㫃㫀㫅㪻㪼㫉㪋 㪇㪅㪇㩷
㪚㫐㫃㫀㫅㪻㪼㫉㩷㪥㫌㫄㪹㪼㫉 㪇㩷 㪉㪇㪇㪇㩷 㪋㪇㪇㪇㩷 㪍㪇㪇㪇㩷 㪏㪇㪇㪇㩷 㪈㪇㪇㪇㪇㩷 㪈㪉㪇㪇㪇㩷
㪜㫅㪾㫀㫅㪼㩷㪪㫇㪼㪼㪻㩷㪲㫉㫇㫄㪴
Fig.13 Exhaust Mass Flow Rate
Fig.14 Fuel’s Residence and Vaporized Times

7 . Cooling Effect of Inhaled Air

This figure shows vaporized fuel increase over
There is a high risk of knocking because this 9000[rpm]. It is because discharge of the super
engine used for experiment has a high compression charger reaches the intake amount when the engine
ratio and super charging system. Addition, a rising speed at over 9000[rpm], and it leads intake pres-
intake air temperature leads charging efficiency de- sure decrease. As a result, decrease of intake air
crease, and then torque is decreased as a result. temperature change to 10K and it makes vaporized
Therefore, two methods to cool the intake air are time increase over 10000[rpm]. However, at under
considered and explained below. One way is sub- 9000[rpm], it is found out that vaporized fuel enters
injection system which applies fuel latent heat of into the cylinder.
vaporization to inject fuel at upper stream of intake In Fig.15, torques and intake temperature are
manifold. The other one is mounting an air-cooled compared with or without sub-injection system.
intercooler system. Temperature of sub-injection system couldn’t be
First, for the sub-injection system, main fuel measured correctly, because the fuel adhered on the
amount were sorted for primary-injection and sub- sensor vaporized.
injection in the experiment. In the low engine speed
range a time for intaking the fuel into the cylinder
is long because it was injected at upper of collector
80 370
tank, so a response lag was observed at changing
load. For resolving this problem, the fuel amount 70 360
balance with primary and sub was changed in every

Intake Temperature [K]

Engine Torque [Nm]

60 350
engine speed. In Fig.14, residence time*2 when in-
jected fuel enters the cylinder and vaporized time of 50
340
injected fuel are compared. For this calculation, di- 40
ameter of injected fuel is defined as 400[μm]. 330
30 Torque (Sub Inj.)
320
xA 20 Torque (Non Sub Inj.)
R  (1) Intake Temp. (Sub Inj.) 310
V SC 10
Intake Temp. (Non Sub Inj.)
       
0 300

 R : residence time x : Secondary position 4000 5000 6000 7000 8000 9000
Engine Speed [rpm]
A : Intake area VSC : Intake volume Fig.15 Sub Injection Performance Assessment

d 2 f c f T A  TB d 2 f q
V  ln  (2) From the sub-injection effect, intake tempera-
12 a T A  TB ' 8 a T A  TB  ture has decreased more than 5% at mid-high speed

range. From this result, improvement in engine tor-
 V : Vaporized time d : Diameter of injected fuel que at engine speed over 8000[rpm] is prominent.
 f : Fuel density
 c f : Fuel specific heat On the other hand, at the transient driving range
under 6500[rpm] in which the mixture velocity in
a : Thermal conductivi ty of air the intake manifold is slow, engine response delayed
Ta : Upstream temperature of injector about 0.25[sec] at throttle control because the lag of
TB : Temperatur e of fuel  injected fuel for sub-injection system occurred.
Next, conducted a research using air-cooling in-
TB ': Fuel boil temperature ter cooler. Considering the effect of pressure loss in
q : Fuel latent heat of vaporizati on the intercooler inside, flow pass section area at the
SETC2011
85
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

core member is set 1.2 times larger than intake pipe.

The side tank and core member arrays were ar-
30-40
ranged in order to minimize the weight increase due 40
to mounting the intercooler, and it increased only 20-30

Ignition Timing
1020[g]. 30 10-20
20 0-10

10 11000
㪏㪇 㪎㪇
㪎㪇 0 5500
㪍㪇
Engine

135
㪍㪇 0

115
㪌㪇 Speed

80
㪜㫅㪾㫀㫅㪼 㪫㫆㫉㫈㫌㪼䋨㪥㫄䋩

㪜㫅㪾㫀㫅㪼 㪧㫆㫎㪼㫉䋨㫂㪮䋩

40
[rpm]

0
㪌㪇
㪋㪇 Intake Pressure [kPa]
㪋㪇 Fig.17 Ignition Timing Map
㪫㫆㫉㫈㫌㪼㩷㩿㪠㫅㫋㪼㫉㩷㪺㫆㫆㫃㪼㫉㪀 㪊㪇
㪊㪇
㪫㫆㫉㫈㫌㪼㩷㩿㪪㫌㪹㩷㪠㫅㫁㪅㪀 㪉㪇
㪉㪇
㪧㫆㫎㪼㫉㩷㩿㪠㫅㫋㪼㫉㩷㪺㫆㫆㫃㪼㫉㪀
㪈㪇 㪈㪇 㪏㪇 㪊㪌㪇
㪧㫆㫎㪼㫉㩷㩿㪪㫌㪹㩷㪠㫅㫁㪅㪀
㪇 㪇 㪎㪇 㪊㪇㪇
㪉㪇㪇㪇 㪊㪇㪇㪇 㪋㪇㪇㪇 㪌㪇㪇㪇 㪍㪇㪇㪇 㪎㪇㪇㪇 㪏㪇㪇㪇 㪐㪇㪇㪇 㪈㪇㪇㪇㪇㪈㪈㪇㪇㪇 㪍㪇 㪉㪌㪇
㪜㫅㪾㫀㫅㪼 㪪㫇㪼㪼㪻䋨㫉㫇㫄䋩

㪙㪪㪝㪚㩷㪲㪾㪆㫂㪮㪿㪴
㪜㫅㪾㫀㫅㪼㩷㪫㫆㫉㫈㫌㪼㩷㪲㪥㫄㪴
㪌㪇
㪉㪇㪇
Fig.16 Intercooler instead of Sub Injection 㪋㪇 㪫㫆㫉㫈㫌㪼㩿㪠㪾㫅㪅㩷㪫㫀㫄㫀㫅㪾㩷㫄㫆㪻㫀㪽㫐㪀
㪈㪌㪇
㪊㪇 㪫㫆㫉㫈㫌㪼㩿㪠㪾㫅㪅㩷㪫㫀㫄㫀㫅㪾㩷㫊㫋㪻㪀
㪉㪇 㪈㪇㪇
㪙㪪㪝㪚㩷㩿㪠㪾㫅㪅㩷㪫㫀㫄㫀㫅㪾㩷㫄㫆㪻㫀㪽㫐㪀
Comparing torques and powers with intercooler 㪈㪇 㪌㪇
㪙㪪㪝㪚㩷㩿㪠㪾㫅㪅㩷㪫㫀㫄㫀㫅㪾㩷㫊㫋㪻㪀
or sub-injection system is shown in Fig.16. Effect of
㪇 㪇
cooling intake temperature is about -20 [K] with the
㪈㪇㪇㪇 㪊㪇㪇㪇 㪌㪇㪇㪇 㪎㪇㪇㪇 㪐㪇㪇㪇 㪈㪈㪇㪇㪇 㪈㪊㪇㪇㪇
sub-injection system and about -30~-40 [K] with the 㪜㫅㪾㫀㫅㪼㩷㪪㫇㪼㪼㪻㩷㪲㫉㫇㫄㪴
intercooler. This decreases the danger of knocking
Fig.18 Torques and BSFC
and increases the charging efficiency. Due to this
effect, the maximum torque and mean torque in-
crease to 9.24[Nm] and 5.89[Nm] in 3500
9. Knocking Control System
~9000[rpm], respectively.
For knocking prevention, knocking frequency
simulation and ignition timing control was con-
8. Optimization of Fuel Injection Amount and Igni-
ducted by attaching knocking censor and knocking
tion Timing
control unit*3. Knocking signal was measured syn-
Optimization of fuel injection amount and igni- chronizing with each cylinder’s combustion period.
tion timing (advance angle, retarded angle) was This enables to measure knocking per cylinder using
considered from the aim to improve fuel efficiency. single knocking censor. For this reason, ignition
And, Optimization of ignition timing was done by timing was set individually per cylinder.
obtaining MBT by using dynamometer. When a Fig. 19 is a typical example of time elapse ver-
knocking has occurred before MBT, ignition timing sus engine speed and knocking strength of each cy-
was decided at the angle of light knocking point. linder.
Map of optimized ignition timing is shown in
Fig.17. Torque curve and net fuel consumption rate
before and after alternating the ignition timing is
shown in Fig.18.
As shown in Fig.18, maintaining same air-fuel
ratio, engine torque at revolution lower than
5500[rpm] has increased. This reason is that im-
provement of indicated thermal efficiency owing to
the optimization of ignition timing can be considered.
Moreover, from this, net fuel consumption rate has
improved at 3300[rpm] from 319.7[g/kWh] to
260.9[g/kWh] at maximum, and it gets better 18.4%.
Fig.19 Knocking Sensor Signal

With controlling knocking in the light knocking

range, ignition timing was decided at MBT in this
experiment with engine bench. As one can see, level
SETC2011
86
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

of knock censor signal increases as engine speed 11. Conclusion

gets higher. This is because vibration of engine
By adopting super charger, it enabled to widen
block becomes larger in proportion to engine speed.
the torque band to lower revolution side and gain
For this reason, 8[Hz] band pass filter which is the
higher torque. The results of this study are concisely
knocking frequency is actually used. Moreover, to
described as follows.
eliminate the 8[Hz] background noise, threshold
amount called knock window is set per revolution. 1. The valve over lap angle was changed from 73
The part in Fig.19 where the knocking censor signal degrees to 42 degrees by slowing the inlet valve
level is rising up in the shape of a pulse shows that opening by 19 degrees, and hastening the ex-
it is actually knocking. hausting valve closing by 12 degrees. And the
And, for considering uncertain condition of a maximum lift was reduced by 0.5mm for intake
vehicle, knocking control system which delays igni- side and 0.6mm for the exhaust side. With
tion timing when the censor detects knocking is con- these modification, angle of area during valve
structed in order to optimize ignition timing for the over lap decreased from 319.06[mm ࡮ rad] to
safe engine driving. The following is the calculation 140.46[mm࡮rad] (about 56 percent).
used to determine the ignition retarded angle.
2. In order to reduce the variability of air intake
  ( KI  KI aim )  a (3) amount among each cylinder, intake manifold
    
was calculated with 3 dimension thermo-fluid
 : Amount of retard  a  Threshold value analysis software and manufactured using
glass beads containing nylon material. As a re-
KI  Knocking strength
sult, variation in intake air mass flow rate de-
KI aim  Aim knocking strength creased from 5% to 3%.

10. Torque Curve 3. For the cooling system, intercooler is used in-
Comparison between this year’s and last year’s stead of using sub-injection system. This re-
torque curve is indicated in Fig. 20. Torque peak of sulted in increasing the difference of intake
supercharger equipped engine appeared in lower temperature; from -20 [K] to -30~-40 [K].
engine speed than that of NA. Moreover, comparing
this year and last year, 80% torque band in this year 4. Optimization of fuel injection amount and igni-
was wider than that in last year; from 3500[rpm] tion timing when engine speed set at 3300[rpm],
~10000[rpm] to 3500[rpm] ~11500[rpm]. Maximum lead net fuel consumption rate increase from
torque increased to 73.9[Nm] from 67.3[Nm] and 319.7[g/kWh] to 260.9[g/kWh] at maximum.
engine speed at maximum torque decreased to This is an improvement of 18.4㧑.
6000[rpm] from 7000[rpm].
5. As a result of 1~4, wider range of 80% torque
band of SR09 was verified as 3500[rpm]
㪏㪇 ~11500[rpm]: this is greater than that of SR08,
3500[rpm] ~10000[rpm]. And maximum torque
㪎㪇
increased to 73.9[Nm] from 67.3[Nm] and max-
㪜㫅㪾㫀㫅㪼 㪫㫆㫉㫈㫌㪼㩷㪲㪥㫄㪴

㪍㪇
imum torque engine speed decreased to 6000
㪌㪇 [rpm] from 7000[rpm].
㪋㪇
㪊㪇 㪫㪿㫀㫊㩷㪰㪼㪸㫉㩷㩿㪪㪚㪀
㪣㪸㫊㫋㩷㪰㪼㪸㫉㩷㩿㪪㪚㪀 For future subject, optimization of retard
㪉㪇
amount and retard recovering time for knocking
㪈㪇 㪫㫎㫆㩷㫐㪼㪸㫉㫊㩷㪸㪾㫆㩷㩿㪥㪘㪀
control remains to be solved.
㪇
㪇 㪌㪇㪇㪇 㪈㪇㪇㪇㪇 㪈㪌㪇㪇㪇 Reference
㪜㫅㪾㫀㫅㪼 㪪㫇㪼㪼㪻㩷㪲㫉㫇㫄㪴 1. 2011 Formula SAE® Rules
Fig.20 Torque Curves 2. Combustion engineering, Yukio Mizutani, Morikita
Publisher, p143, 1997
This success is mainly considered as a result of 3. SKM and OKM User Manual, MoTeC Pty Ltd, 2008
mounting new intake collector tank, optimization of
ignition timing, adopting the intercooler instead of
intake cooler system.

������������������������������������������������������������������������������ ��������������������������������������������������������������������������������
������������������������������������������������������������������������������������� ����������������������������������������������������������������������������������������
������������������������������������������������������������������������������������ 
�����������������������������������������
�������������������������������������������������������������������������������
�������������������������������
�������������������������������������������������������������������������������������� �����������������
����������������������������������������������������������������������������������� ������������������������������
��������������  �������������������
    
87
Downloaded from SAE International by Purdue University, Monday, October 29, 2018
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

2008-01-3007

The Effects of Intake Plenum Volume on the Performance

of a Small Normally Aspirated Restricted Engine
L. J. Hamilton and J. E. Lee
US Naval Academy

ABSTRACT mass of air transferred to the combustion chamber over

desired RPM ranges.
Intake tuning is a widely recognized method for
optimizing the performance of a naturally aspirated While the actual wave dynamics are complex for a multi-
engine for motorsports applications. Wave resonance cylinder engine, the intake system can be modeled as a
and Helmholtz theories are useful for predicting the simple Helmholtz resonator [1]. When the intake valve
impact of intake runner length on engine performance. opens and the piston descends, a rarefaction or
However, there is very little information in the literature expansion wave travels upstream in the runner toward
regarding the effects of intake plenum volume. The goal the plenum. Upon reaching the runner opening in the
of this study was to determine the effects of intake plenum, this wave is partially transmitted (still as an
plenum volume on engine performance for a restricted expansion wave) into the plenum. At the same time, part
naturally aspirated engine for Formula SAE (FSAE) of the original expansion wave is reflected as a
vehicle use. Testing was conducted on a four cylinder compression wave that heads back to the cylinder. If the
600 cc motorcycle engine fitted with a 20 mm restrictor in compression wave arrives at the intake valve before it
compliance with FSAE competition rules. Plenum sizes closes, the increased pressure will effectively force more
were varied from 2 to 10 times engine displacement (1.2 air into the cylinder [2]. This tuning effect or resonant
to 6.0 L) and engine speeds were varied from 3,000 to supercharging leads to increased volumetric efficiency
12,500 RPM. Performance metrics including volumetric and torque. Meanwhile, the expansion waves will travel
efficiency, torque and power were recorded at steady to the inactive branches (with closed valves) and be
state conditions. Experimental results showed that reflected as expansion waves back to the active runner.
engine performance increased modestly as plenum These expansion waves reduce intake runner pressure
volume was increased from 2 to 8 times engine and have the effect of reducing volumetric efficiency [1].
displacement. Increasing plenum volume beyond 8 Typically, intake tuners focus on runner geometry (length
times engine displacement resulted in significant and/or cross section area) and plenum size and shape.
improvement in performance parameters. Overall, peak An excellent overview of resonant supercharging is
HP was shown to increase from 54 kW to 63 kW over the provided by Benajes et al [3].
range of plenums tested. While it was clear that larger
intake plenums yielded better steady state performance, This method is particularly useful for vehicles competing
their effect on transient performance was not explored in the Formula SAE (FSAE) design series where the
and will be the subject of further research. extra weight and complexity of mechanical boost
systems could be detrimental to overall performance.
The FSAE series provides college students from around
the world to conceive, design, fabricate and compete a
INTRODUCTION
small scale formula style vehicle based on a
comprehensive set of design constraints. Several of the
Intake system tuning is a well known and cost effective
constraints pertain to engine and intake design. For
method of increasing the performance of internal
example, engine size is limited to 610cc and all intake air
combustion engines without the use of mechanical
must pass through a restrictor of circular cross section
turbochargers or superchargers for motorsports
(20 mm for gasoline or 19 mm for E85) [4].
applications. In general, intake systems consist of an
airbox or plenum chamber (fed by a throttle body) and
Some excellent studies have been published to
runners that lead from the plenum to each cylinder.
document the effects of intake tuning on a restricted
Engine designers strive to take advantage of pressure
engine suitable for use in the FSAE design series.
wave dynamics in the intake system to increase the
Jawad et al evaluated four different intake designs with a
Honda™ CBR600 F4i engine coupled to a water brake

89
1312 SAE Int. J. Engines | Volume 1 | Issue 1
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

dynamometer [5]. Their performance criteria were based The intake plenum was designed as a simple aluminum
on total area under the power curve for each box with a movable ceiling to provide 10 discrete
configuration. They focused primarily on novel runner volumes for testing. The possible volumes were based
geometry approaches and did not document changes in on engine displacement and ranged from 1 (0.6L) to 10
plenum volume. (6.0L). In all cases, the entrance to the plenum via the
restrictor was not changed; only the ceiling was raised or
McKee et al provided a superb contribution to lowered to change plenum volume. The plenum can be
optimization of this class of engines by experimentally seen in Figure 1.
evaluating the effects of diffuser length (after the
restrictor), runner length, camshaft modifications and
exhaust tuning [6]. However, they selected a plenum Table 1: 2002 Honda F4i Engine Specifications
volume of 3.5L and did not vary it during testing. This
decision was based on model predictions which did not Engine configuration Inline four, 4 stroke, DOHC
show much benefit by increasing plenum volume beyond Displacement 600 cc
this amount. Fuel system EFI
Cooling system Water
The effect of plenum volume on engine performance was Lubrication system Wet sump
studied and reported by Ceviz [7] as a basis for variable Compression ratio 12:1
intake plenum technology. He installed plenum additions Valves per cylinder 4
(90 and 180 cc) to the original intake manifold of a 1.6L Bore 67.0 mm
four cylinder gasoline engine over a modest RPM range Stroke 42.5 mm
of 1250 – 3000 RPM. While he found that increasing
Valve timing
plenum volume increased power at the mid-RPM range,
IVO 30° BTC
plenum size had to be reduced for best performance at
IVC 60° ABC
higher speeds. It should be noted that this engine was
EVO 50° BBC
equipped with a carburetor and the author cited difficulty
with proper AFR control when changing plenum size. EVC 10° ATC

The objective of the current study is to determine the

effects of plenum size on volumetric efficiency, torque
and power for a restricted intake 600cc motorcycle
engine for use in Formula SAE applications under steady
state operating conditions.

EXPERIMENTAL SETUP

EXPERIMENTAL APPARATUS

All testing was conducted using a 2002 Honda™

CBR600 F4i engine equipped with a Microsquirt™
engine control unit (ECU). The engine characteristics are
listed in Table 1. Engine performance parameters were
measured using a Land and Sea™ water brake
dynamometer coupled to the output shaft. The
dynamometer was equipped with a load servo to
maintain a specified RPM regardless of throttle position.
The intake apparatus included a throttle body, an intake
pipe with a circular 20 mm diameter restrictor, a custom
variable volume intake plenum and four 5 cm aluminum
intake runners as shown in Figure 1. A 0 – 2 bar
Figure 1: Intake system
Kistler™ piezoelectric pressure transducer was mounted
in the plenum adjacent to the runner leading to cylinder
Volumetric efficiency, a measure of an engine’s air
#1 with a data rate of 40 kHz. An ECM™ 1200 oxygen
breathing capacity, was one of the desired parameters to
sensor was mounted in the exhaust header
be evaluated in this study as it directly relates to torque
approximately 12 cm from exhaust valve to provide
produced by an engine [8]. It is determined by dividing
positive feedback to the ECU to maintain proper AFR
the actual mass of air consumed by the ideal air mass
control. The AFR target in the ECU was set to
that could be drawn in based on displacement volume as
stoichiometric for all test points. Injector sequencing was
shown in equation (1).
selected to minimize pressure fluctuations in the fuel rail.

90
SAE Int. J. Engines | Volume 1 | Issue 1 1313
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018
80
m air , actual
ηV = (1)
m air , ideal 70

spark timing (deg BTC)

60
where
50

PatmVdisp  RPM 
m air , ideal =   (2) 40

RTair  2 
30

Since the engine was not equipped with a mass airflow 20

sensor, fuel flow rate and AFR were used to determine
actual air flow through the engine using equation (3). 10
0 2000 4000 6000 8000 10000 12000 14000

m air , actual = AFRm fuel (3) RPM

Figure 3: Spark timing curve

A fuel injector calibration test was performed to
EXPERIMENTAL PROCEDURES
determine actual fuel flow for a given pulse width. For
this calibration, the fuel rail was pressurized to 4.0 bar
All testing was performed in a climate controlled test cell
and a single injector was activated for 1000 ‘shots’ into a
with the engine fully warmed up. Prior to each test
graduated cylinder to get an accurate volumetric fuel flow
period, test cell temperature and barometric pressure
rate for various pulse width values. The result of this
were recorded. Standard SAE correction factors were
calibration is shown in Figure 2. Notice that the curve
used for all performance results. Plenum volumes were
intercepts the x-axis at 1.1 ms which represents the
varied from 1.2 L to 6.0 L in 0.6L increments. Special
opening time for these injectors. Using this calibration
care was taken to ensure that all seams and access
relationship yields the following equation for fuel mass
holes were sealed prior to testing. For each volume, the
flow rate:
engine was stabilized at WOT from 3000 to 12,500 RPM
at 500 RPM intervals using the RPM-hold feature of the
 RPM 
m fuel = ρ fuel ( 0.0036 PW − 0.0039 )   (4)
dynamometer. Absorber torque, fuel pulse width and
 2  AFR were recorded for each test point. Premium
gasoline with a (R+M)/2 octane rating of 94 was used for
all test points.
where PW is the fuel pulse width in msec.

0.035
RESULTS AND DISCUSSION

0.030
cal data
linear fit
FULL RPM SWEEP

0.025 Figure 4 and Figure 5 show graphical representations of

fuel flow (cc/shot)

the effects of plenum size on torque at various engine

0.020
speeds. At low RPM, the smaller plenum appears to
produce higher torque than the large plenum although
0.015
the effect is modest. However, above 7,000 RPM, it is
0.010
clear that torque tends to increase significantly as
y = 0.0036 x − 0.0039 plenum size increases. This effect is more pronounced
0.005 R2 = 1 at higher RPM. In fact at the maximum RPM tested,
torque dropped from 48 Nm to 36 Nm when comparing
0.000 the largest and smallest plenums; a 25% loss in torque.
0 1 2 3 4 5 6 7 8 9 10

pulse width (ms) It is also evident that there are two distinct torque peaks;
Figure 2: Injector calibration at fuel pressure = 4.0 one at approximately 6500 RPM and the second near
bar 10,000 RPM. It is also interesting to note that the torque
peak has a very weak dependence on plenum size. For
For all testing, spark timing was set according to the the 6.0 L plenum, torque peaks at 54.5 Nm at 10,000
curve shown in Figure 3. This curve represents RPM while the 1.2 L plenum results in a peak of 48.5 Nm
experimentally derived maximum brake torque spark at 9500 RPM. This is because runner length is the
timing for the range of 1300 to 12,500 RPM. Engine primary factor in determining the frequency (driven by
knock was not encountered on any of the test runs. engine RPM) for resonant supercharging [3]. Later in
this paper, the effects of plenum size on engine
performance operating at 9500 RPM will be more closely
examined.

91
1314 SAE Int. J. Engines | Volume 1 | Issue 1
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

70

60
60

55

)
corrected power (kW
50
m)

50
corrected torque (N-

40
45
30
40
20
35

30 10

25 0
6

ple
20

n
5

um
6
ple

vol
nu

u
mv

me
4 12000
olu

2 10000

(
3

L)
8000
me

12000 1 6000
2 10000 4000
(L)

8000 RPM
6000
1 4000 Figure 6: Power - RPM and plenum volume sweep
RPM
70
Figure 4: Torque - RPM and plenum volume sweep
6.0L plenum
60

60
50
6.0L plenum
55 1.2L plenum
power (kW)

40
50

30
45
torque (Nm)

20
40 1.2L plenum

35 10

30 0
2000 4000 6000 8000 10000 12000 14000
25
RPM
20
2000 4000 6000 8000 10000 12000 14000 Figure 7: Comparison of power output between
RPM smallest and largest plenums tested

Figure 5: Comparison of torque between smallest

and largest plenums tested Another interesting metric is comparison of integrated
area under torque and power curves over the useful
Figure 6 shows a similar RPM and plenum sweep with RPM range. This metric is useful in evaluating the
power as the variable of interest. As before, plenum size overall advantage offered by a particular plenum volume.
has little effect until approximately 7,000 RPM as shown Recall that in Figure 5 the large plenum clearly held the
in Figure 7. At the higher engine speeds, the increase in advantage at high RPM but the smaller plenum showed
power with plenum size is dramatic. Peak power output a modest advantage in the low RPM range. Figure 8
increases from 54 kW for the 1.2L plenum to 63 kW for shows integrated area (normalized to the 1.2L plenum)
the 6.0L plenum – an increase of 17%. Of note, plenum for both torque and power for engine speed ranging from
size appears to have a greater effect on the engine 3,000 to 12,500 RPM. From this chart, it is clear that
speed for peak power than it did on peak torque. In fact, gains in torque and power are relatively modest for
for the smaller plenums, power peaked at approximately plenum volume up to 3.6L or 6 times engine
11,000 RPM while the larger plenums drove the RPM for displacement. However, increasing plenum volume up to
peak power out to 12,500 RPM. This is largely due to 8 or 10 times displacement produces a relatively large
the precipitous reduction in torque for the small plenums increase in performance based on integrated area.
at high RPM seen in Figure 4. When compared to the 1.2L plenum, the 6.0L plenum
produces a 5% and 8% increase in area under the
torque and power curves respectively.

4
92
SAE Int. J. Engines | Volume 1 | Issue 1 1315
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

1.10 The plenum pressure traces shown in Figure 10 are

power helpful in understanding this significant effect. In this
1.08 figure, normalized pressure is plotted for several
plenums over the range of crank angle degrees (CAD)
1.06 corresponding to the open intake valve. In this case,
normalized area

CAD = 0° corresponds to top center. The dotted line,

1.04 torque representing the smallest plenum, shows two deep
pressure troughs; one just after IVO and another one as
the piston approached bottom center. The next two
1.02
bigger volumes, 3.6L and 4.8L show a similar (but less
significant trough) just after IVO. These two cases show
1.00
much higher pressure even as the piston is descending
implying greater resonant charging. Finally, the largest
0.98 plenum shows very little pressure drop just after IVO and
0 1 2 3 4 5 6 7
a significant peak just before IVC.
plenum volume (L)
Figure 8: Comparison of normalized integrated area Qualitatively, it appears that the largest plenum provides
under torque and power curves for various plenum the greatest average pressure throughout the intake
sizes. valve open period. This advantageous pressure profile
is likely the result of beneficial wave dynamics that occur
PLENUM COMPARISON AT 9500 RPM in the larger plenum chamber known as resonance
charging [9]. Due to the confined space in the smaller
Next, the effect of plenum size on engine performance at plenums, strong reflection of expansion waves from
9500 RPM was closely evaluated. This engine speed neighboring cylinders (with closed intake valves) may
was chosen because it corresponded to most of the cause interference waves that reduce the strength of
torque peaks seen during the RPM and plenum size compression waves in the primary runner thereby
sweep described earlier. Figure 9 shows the effect of negating the beneficial effects of resonant charging to
plenum size on volumetric efficiency and power at 9500 some extent. It is likely that the reflections of these
RPM. Note that it is not necessary to also plot torque as expansion waves in the larger plenums are not as strong
that would show an identical trend as power since RPM as those in the smaller plenums and therefore do not
is fixed. It is interesting to see that volumetric efficiency degrade resonance charging.
shows only modest gains as plenum size is increased
until 4.8 L (8 engine displacements) is exceeded. 1.04
Beyond that size, volumetric efficiency jumps from 0.83 1.2L
to 0.95, an increase of 14.5%. 1.02 3.6L
4.8L
6.0L
A similar trend is noted for power. Once a plenum size 1.00

of 8 engine displacements is used, power increases by

0.98
nearly 5%. Overall, power increased from 47 to 53 kW
P/Patm

across the full range of plenum volumes tested at this

0.96
engine speed. This corresponds to a 12.7% increase in
power with no other modifications to the engine. 0.94

1.0 58
0.92
volumetric efficiency
56 0.90
0.8
-30 0 30 60 90 120 150 180 210 240
IVO TC BC IVC
volumetric efficiency

54 crank angle degrees

corrected power (kW)

0.6

corrected power 52 Figure 10: Normalized plenum pressure for various

0.4 plenum volumes during intake valve open event for
50 cylinder #1. (Zero CAD = top center)
0.2
48 To quantify these pressure trends, normalized pressure
was averaged across the entire intake valve open period
0.0 46 and compared for the various plenum sizes. Figure 11
0 1 2 3 4 5 6 7
provides a summary of these averages and shows that
plenum volume (L)
they increase significantly as plenum volume increases.
Also plotted in this figure is volumetric efficiency. It is
Figure 9: Plenum volume sweep at 9500 RPM evident that the trend in normalized pressure averaged
over the intake valve open period corresponds directly

93
1316 SAE Int. J. Engines | Volume 1 | Issue 1
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

with an increase in volumetric efficiency. Thus intake -Due to resonant charging at high RPM, average
runner pressure can be used as a measure of intake pressure in the plenum adjacent to the intake runner
system effectiveness. throughout the intake valve open period was seen to
0.978 0.96 increase as plenum volume was increased and was
0.976
directly related to a rise in volumetric efficiency. The
0.94
smaller plenums may have negated this charging effect
0.974 avg P/Patm
volumetric efficiency
0.92
due to stronger interference waves from neighboring
0.972
0.90 cylinders.

volumetric efficiency
average P/Patm

0.970
0.88
0.968 ACKNOWLEDGMENTS
0.86
0.966

0.964
0.84
The authors would like to thank Mr. John Hein and Mr.
0.962
0.82 Charlie Baesch of the US Naval Academy Propulsion
0.80
Lab for providing support and access to engine test
0.960
facilities.
0.958 0.78
0 1 2 3 4 5 6 7

plenum volume (L) REFERENCES

Figure 11: The effect of normalized intake pressure 1. Lumley, J.L. Engines An Introduction, Cambridge
on volumetric efficiency at RPM = 9,500 University Press, 1999.
2. Heywood, J.B., Internal Combustion Engine
While the benefit of large plenum volume has been Fundamentals, Mc-Graw Hill, 1988.
demonstrated for steady state conditions, there may be 3. Benajes, J., Reyes, E., Galindo, J. and Peidro, J.,
some transient response drawbacks. For example, “Predesign Model for Intake Manifolds in Internal
when the throttle is rapidly opened, the plenum must fill Combustion Engines”, SAE Paper 970055.
with air before the cylinders can be charged. For larger 4. [Link]
plenums, this response time will be extended. Since the es//
FSAE vehicle must perform well in dynamic autocross 5. Jawad, B., Dragoiu, A., Dyar, L., Zellner, K. and
style events, throttle response will be crucial. Future Riedel, C., “Intake Design for Maximum
work will include evaluation of plenum size on the Performance”, SAE Paper 2003-01-2277.
transient response for a restricted engine. When this is 6. McKee, R.H., McCullough, G., Cunningham, G.,
better characterized, an optimum plenum size can be Taylor, J.O., McDowell, N., Taylor, J.T. and
determined for this specific application. McCullough, R., “Experimental Optimisation of
Manifold and Camshaft Geometries for a Restricted
CONCLUSIONS 600cc Four-Cylinder Four-Stoke Eninge”, SAE Paper
2006-32-0070.
Several conclusions can be drawn regarding the effect of 7. Ceviz, M.A., “Intake Plenum Volume and its
varying intake plenum volume on a restricted 600 cc influence on the Engine Performance, Cyclic
motorcycle engine in steady state conditions: Variability and Emissions”, Energy Conversion and
Management, Vol. 48, pp. 961-966, 2007.
-Plenum size has little effect on torque and power until
8. Yagi, S., Ishizuya, A. and Fujii, I., “Research and
approximately 6500 engine RPM. At high RPM, gains
Development of High-Speed, High-Performance,
were considerable; up to 31% in power and torque at
Small Displacement Honda Engines,” SAE Paper
12,500 RPM.
700122.
-For the plenums fitted with 5 cm intake runners, the 9. Neuber, H.J., Endres, H. and Breuer, M., “New
engine had torque peaks at approximately 7,000 and Variable Intake and Mixture Formation System for
10,000 RPM. Plenum size had a modest effect on peak Multi-Valve SI Engines”, SAE Paper 940449.
torque and power location (RPM).

-The 6.0L plenum produced 17% higher peak power than CONTACT
the 1.2L plenum (63 kW at 12,500 RPM vs. 54 kW at
11,000 RPM). This was largely due to the significant ljhamilt@[Link]
reduction in torque at high RPM for the smaller plenum.

-When compared to the 1.2L plenum, the 6.0L plenum

produced a 5% and 8% increase in area under the
torque and power curves respectively.

94
SAE Int. J. Engines | Volume 1 | Issue 1 1317
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

ABBREVIATIONS

ABC: After Bottom Center

ATC: After Top Center

BBC: Before Bottom Center

BC: Bottom Center

BTC: Before Top Center

CAD: Crank Angle Degrees

ECU: Engine Control Unit

EFI: Electronic Fuel Injection

EVC: Exhaust Valve Closing

EVO: Exhaust Valve Opening

FSAE: Formula Society of Automotive Engineers

m
 : Mass Flow Rate (kg/s)

V : Volumetric Efficiency

Patm: Atmospheric Pressure

ρfuel : Fuel density

PW: Fuel Injector Pulse Width

TC: Top Center

Vdisp: Displacement Volume

WOT: Wide Open Throttle

95
1318 SAE Int. J. Engines | Volume 1 | Issue 1
Downloaded from SAE International by Purdue University, Monday, October 29, 2018
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

2008-01-1774

Design and Development of a Turbocharged E85 Engine for

Formula SAE Racing
Jason P. Moscetti, Bryan Gilroy Smith, and Volker Sick
The University of Michigan

Copyright © 2008 SAE International

ABSTRACT used as fuel, and 19mm if E85 is used. Intake air can be
boosted, but only if the system is designed by the team.
A summary of the design and development process for a
Formula SAE engine is described. The focus is on three The goals for the engine design were to maintain
fundamental elements on which the entire engine reliability, minimize response time, and increase usable
package is based. The first is engine layout and engine output torque, while maintaining a low vehicle
displacement, second is the fuel type, and third is the air weight and center-of-gravity. The engine is broken down
induction method. These decisions lead to a design into sub-systems designated by the different fluids that
around a 4-cylinder 600cc motorcycle engine, utilizing a need to be managed: the fuel, air induction, lubrication
turbocharger and ethanol E-85 fuel. Concerns and oil, and coolant systems. The structural and dynamic
constraints involved with vehicle integration are also interaction (via a transmission and drive-train) between
highlighted. The final design was then tested on an the engine and the chassis also needs to be considered.
engine dynamometer, and finally in the 2007 M-Racing
FSAE racecar. An analysis of driving data that were recorded during
past FSAE competitions showed that the dynamic event
INTRODUCTION performance would benefit from an increase in torque at
engine speeds below 10,000 RPM to better
Starting in 1984 the Formula SAE (FSAE) Collegiate accommodate the vehicle acceleration demands.
Design Competition has gained popularity around the Increasing the overall peak power was not the primary
world, challenging the engineering knowledge and objective. Therefore, implementation of a turbocharger
creativity of generations of students [1]. This creativity is as a means to boost torque seemed attractive, even with
limited by a comprehensive set of rules that specify the challenges introduced with an upstream throttle.
design criteria and primarily aim at safety. The
competition is broken down into dynamic (Acceleration, The use of E85 fuel was also an interesting opportunity
Skid-pad, Autocross, and Endurance/Fuel Economy) to gain performance by taking advantage of differences
and static events (Design, Cost Analysis, and Business in physical properties. Specifically, ethanol has a much
Plan). Teams receive a score for each event relative to higher heat of vaporization than gasoline. This will help
the competition, and the best total score wins the overall intake charge cooling, and increase the potential for
competition. The engine package is extremely vital to more torque output and better fuel economy.
the racecar and the success of the team, so each design
decision has to be made carefully. Versatility is the DESIGN GOALS AND DECISIONS
primary goal because the engine needs to be high-
output, simple, efficient, reliable, cheap, and marketable. PLATFORM
To be competitive in all events, the final package will
have to strike a balance instead of being over-designed In observing the recently successful FSAE teams
for one criterion. throughout the world, there are two factions of engine
platforms; 600cc 4-cylinder and 450cc single cylinder
It is commonplace in all motor-sport for the engine engines. Both have their advantages and disadvantages.
output to be limited. FSAE rules do not allow a The 600cc 4-cylinder engine has the potential to be the
displacement greater than 610cc on a 4-stroke cycle, more powerful package, but is also the heaviest. In an
and all intake air must pass through a circular restrictor. OEM setting, each motor represents a very similar
The restrictor diameter has to be 20mm if gasoline is power-to-weight ratio of 1.6 kW/kg on-average. Both

97
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

platforms can be readily modified for increased output and 94 Octane) and the third is E85, a blend of
and reduced weight. Ultimately, two factors were found nominally 85% Ethanol and 15% gasoline. Each fuel has
to back up the decision to design around a 600cc 4- advantages and disadvantages. In any case the engine
cylinder platform. First, the availability of 600cc engines needs to be specifically prepared and tuned for
in the USA is much higher as they are sold more operation with the selected fuel. The anti-knock (post-
commonly by themselves, because they are discarded ignition) index varies significantly with each fuel. E85 has
from motorcycles that were damaged from collision. The the highest value and therefore will theoretically allow
450cc single cylinder engines are common to off-road the highest compression ratio. The difference in heating
motorcycling, where engine damage is more frequent. value between gasoline and ethanol is also significant.
Therefore, the engines are not separated from their This, along with its stoichiometric fuel/air ratio results in
factory chassis as often. Second, the magnitude of a projected better gas mileage for ethanol.
pressure waves through the intake/exhaust is smaller in
four-cylinder engines. This allows a smoother coupling Table 1: Comparison of ethanol and gasoline
with a supercharging device. properties. Data from [2].

VEHICLE INTEGRATION Ethanol Gasoline

Boiling Point °C 78 35-210
More importantly than peak torque or minimal fuel
Lower Heating Value (MJ/kg) 26.8 ~42.7
economy is the packaging of the entire engine system
into the chassis. Many compromises need to be Heat of Vaporization (MJ/kg) 0.93 ~0.18
analyzed to balance performance, manufacturability, and Stoichiometric Air/Fuel ratio 9.0 ~14.6
time commitment. As approximately 25% of the vehicle Specific Energy 3.00 ~2.92
weight comes for the engine package, the engine
Blending RON 112-120 90-100
platform dictates many vehicle constraints, most
importantly the center of gravity location and total sprung Blending MON 95-106 80-90
mass.

Another important aspect of vehicle integration is with The higher heat of vaporization for ethanol and its
the driver. With benchmarked engine torque curve data blends should lead to greater charge cooling in the
from previous year’s FSAE efforts, the effects of final intake port of the port-fuel injected engine, further
drive ratio could be analyzed on how the driver can best reducing the auto-ignition tendency. This is especially
use the transmission to generate the most forward true at high engine speeds (see figure 2). From the
acceleration. Therefore, a final ratio of 1:4 was chosen to above data, it can be deduced that for every [additional]
rd
isolate 3 gear as the primary (middle) driving gear 5% Ethanol blended into gasoline, intake temperatures
placing the engine operating speed range between will decrease between 3°C and 5°C. With such a large
6,000 and 10,000 RPM, given a typical FSAE course presence of ethanol in the fuel, the need for a charge
layout. This will equalize engine speed changes between intercooler was eliminated.
shifts, and allow the driver to maintain the engine speed
within its power-band. The vehicle acceleration is still 1.1
rd
limited by tire grip through most of 3 gear, necessitating 100 Octane

the implementation of launch control and traction control 1

E-85

systems.
0.9
T / T_peak

0.8

0.7

0.6

0.5
8000 8500 9000 9500 10000 10500 11000
RPM

Figure 2: Back-to-back dynamometer tests comparing E-

Figure 1: Preliminary vehicle CAD model illustrating 85 to 100-Octane.
engine/chassis integration.
AIR INDUCTION
FUEL
As part of the motivation for using a 4-cylinder platform,
FSAE competition gives the option to choose from three the intake air was boosted using a turbocharger. The
different fuels. Two are types of gasoline (100 Octane, main concern with implementation was a loss of

98
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

drivability. Problems with low vehicle speed WAVE cannot capture. The next step is to integrate the
maneuverability caused by turbo lag were a big concern. 3-D model with the boundary conditions calculated in
Consequently, the engine response was made a higher WAVE.
priority than peak power. Therefore, response dominated
the turbocharger matching process, as well as
compression ratio, and intake and exhaust designs. By Thermodynamics
lowering the compression ratio through removal of piston
crown material, potential thermal stress concentrations
were eliminated near the valve pockets, thereby CAD
1-D WAVE
reducing the risk of auto-ignition. The turbocharger was Packaging
selected by setting targets for power output at several
engine speeds and calculating the required mass flow 2D/3D FLUENT Flow-bench
rates and boost pressures with steady thermodynamic
relationships. These flow rates and pressure ratios were
used to screen compressor maps to determine what 1D/3D
Engine Dyno
turbochargers would fit specifications. For a more WAVE/FLUENT
accurate prediction, a transient Ricardo WAVE model
was developed to simulate a typical dynamometer test. Vehicle Testing
The turbocharger was selected based on matching the
areas of maximum adiabatic efficiency with desired flow Figure 4: General design flow schematic for
and pressure calculations. The decision was confirmed intake/exhaust.
with WAVE, and those results were compared to data
from dynamometer tests. The most difficult aspect of The design of the intake system performance focused on
correlating the simulation to test stand data was the tuning effects from runner geometry, and trends in
initial transient period when the compressor begins to turbocharger transient response. Camshaft profiles
build positive intake pressure, as shown below (figure 3). remained unmodified to minimize design degrees of
freedom. The total intake volume was minimized as
1.1 much as possible without creating restriction. The intake
runner length was tuned to improve volumetric efficiency
1 below 6000 RPM and above 9000 RPM.
WAVE simulation
Dyno Test
0.9 Given the packaging constraints, the plenum inlet is
parallel to the crankshaft axis, leading to concerns about
T / T_peak

0.8 cylinder to cylinder variability. WAVE simulations

predicted a ~10% variation in volumetric efficiency
0.7 amongst the four cylinders. In addition, the 2D and 3D
CFD models were used to compare steady-state
0.6 pressure loss between each runner, also giving ~10%
variation. From this information, the fuel map could be
0.5
calibrated to compensate for the predicted volumetric
5500 6500 7500 8500 9500 10500 11500 12500 efficiency variations.
RPM

Figure 3: Torque curve comparison between WAVE and

dynamometer data.

Using the WAVE results, modeling of the intake and

exhaust began in CAD to package the proposed
geometry. The purpose was to address concerns with
serviceability, weight, and chassis integration. For
example, the WAVE model suggested a 625mm runner
would be effective for increasing volumetric efficiency
above and below the engine speeds at which the
turbocharger was most effective. However, it caused too
many packaging problems with the rear of the chassis,
so that proposal needed to be sacrificed. Figure 4 shows
the general flow of design decisions made throughout
the development process.
Figure 5: CAD rendering of intake plenum and runners,
which was welded from aluminum.
In addition to a WAVE model, 2D and 3D CFD models in
FLUENT of the intake manifold were used to help
identify any spatial pressure or velocity distribution that

99
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

The biggest constraint on the exhaust design was intake and exhaust is modified, and fuel mapping can
packaging in the chassis. In addition, enthalpy losses begin. Fuel mapping consists of first establishing
between the exhaust port and turbine inlet needed to be compensations for air and fuel thermodynamic state,
minimized. Therefore, tuning effects were sacrificed with which can be scaled for any desired air/fuel ratio. The
unequal header lengths in order to minimize volume and entire engine is mapped fuel-rich first, and then spark is
surface area. A small exhaust volume will aid pressure tuned preliminarily. Then the mixture is leaned out in
ratio generation across the turbine at low engine speeds, operating areas where the driver does not demand peak
but will typically increase exhaust temperatures. power out of the engine (such as braking and corner
However, a small surface area will reduce heat transfer entry, where the engine is closed/part throttle).
from the exhaust to the ambient. Careful material
selection and manufacturing were important to trade-off The 2007 development program consisted of two tuning
turbine response and heat-transfer loss. The small 4-1 phases. Both phases utilized the same turbocharger,
collector was investment cast in stainless steel, allowing but the intake and exhaust were changed for the second
for superior geometry and reduced surface phase to aid intake boosting in lower engine speeds and
inconsistencies. The header manifold was also ceramic better smooth out torque output when the engine is
coated to reduce conduction through the tubes, and operated in corner exiting conditions (throttle-on to wide-
radiation to surrounding components. The results fulfilled open).
packaging goals, and helped to shift 5% of the vehicle
weight to the rear wheels. 1.4

2006 (Naturally Aspirated)

2007 (Turbo-charged)
1.2

T / T_peak_NA
0.8

0.6

0.4

0.2
5500 6500 7500 8500 9500 10500 11500 12500
RPM

Figure 7: 2007 torque curve compared to 2006. Data is

normalized about the peak torque from 2006.

Figure 7 shows that the goal was achieved of

maximizing engine torque output in the range of typical
Figure 6: As-cast 316l stainless steel collectors and 3D competition speeds from 6,000 – 10,000 RPM.
rendering with laser inspection data. [3] Additional design precautions were taken in oiling and
cooling system design to make sure the engine can
RESULTS operate reliably, like the implementation of a 2-stage oil
scavenging system (dry-sump). The component design
ENGINE TESTING and tuning methodology was simple and well
established, thereby reducing troubleshooting efforts.
No amount of calculation and simulation can be This engine package makes more power than the
validated without experiments. Ultimately, the engine engines of 95% of FSAE teams, making it extremely
package has to perform on the test stand, and then on marketable to a potential racing customer (recall the
track tests in the racecar. competition involves a marketing aspect). The platform
decision, turbocharger selection, material selection, and
Engine test stand experiments were carried out on a manufacturing processes were based on availability of
Superflow 901 water brake dynamometer. The capacity parts and design simplicity, which saved development
is set to handle the torque and power output at the expense.
sprocket output, so, all power measurements include
clutch and transmission losses. For every engine that is VEHICLE TESTING
built, it is first broken-in with an existing intake, exhaust,
cooling, oiling, and fuel system to make sure the Most of the above analysis was applied to the production
cylinders hold compression. The first tests are performed of the 2007 FSAE racecar. M-Racing competed in 2007
to make sure the oiling and cooling systems are up to at Formula SAE (May), Formula Student (July), and the
st
the task of handling consecutive transient load sweeps, 1 Goodyear FSAE Shootout (October). The potential of
and many steady-state operating points. Then, the the system was shown in the beginning at FSAE where

100
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

rd
the team earned 3 place overall in the design event. Additional work needs to be focused on lower RPM
During the season, the engine packaged continued to throttle response by improving the turbo-matching
improve. In the end, the team won the acceleration process. This can be achieved by more accurate
event at Goodyear. simulation and more precise experimental correlation.
Performance gains will follow from iterative experimental
Drivability was a big concern in the design phase, and testing, paralleled with improved computer calculations.
still needs to be improved after several months of vehicle Specifically, the 1D and 3D CFD simulations need to be
testing. Hairpin corners that forced the vehicle to slow integrated. And in general, the combustion, heat
down enough to bring the manifold pressure to vacuum, transfer, and dynamic engine/vehicle models will be
led to a noticeable amount of acceleration lag as the car improved.
exited the corner. The lack of throttle response was
exacerbated as the season progressed, as the engine ACKNOWLEDGMENTS
started to draw oil into the intake from the turbocharger.
Past and present M-Racing engine-team contributors
Some testing sessions were partially dedicated toward
improving the straight line acceleration of the car, as it is BorgWarner Turbo Systems
worth 100 points at competition. With data analysis of
longitudinal acceleration and speed, acceleration times IAV Automotive Inc
decreased from 4.1 seconds to 3.8 seconds over the
course of the season. Suspension settings, shift RPM, Robert Bosch Corp
and launch control RPM were all tuned throughout the
testing program. University of Michigan 3D Lab

The team made substantial fuel economy gains Z-Corporation

throughout the testing schedule. The beginning weeks
of vehicle testing focus on the reliability of the racecar. REFERENCES
As always, the car has to first finish, before it can finish
first. With data analysis of air-fuel ratio, intake and fuel
1. SAE. Formula SAE. 2007 [cited 2007 11/29/2007];
conditions, fuel economy improved by nearly 30%.
Homepage of Formula SAE]. Available from:
[Link]
CONCLUSION 2. Owen, K. and T. Coley, Automotive Fuels
Handbook. 1990, Warrendale, PA: SAE
The need to develop an engine package that is tailored International. 684.
towards the demands of a Formula SAE competition was 3. Lyons B., 3DP-Based Investment Casting; Summary
addressed by designing around a commercial four- of an Evaluation for
cylinder motorcycle engine. A turbocharger was added Additive Manufacturing, TCT 2007 Conference
to increase torque over a [traditionally] naturally Proceedings, Coventy UK, September 2007
aspirated air induction package in the engine speed 4. Heywood, J.B., 'Internal Combustion Engine
range where typical driving takes place. Along with Fundamentals', ISBN 0-07-028637X (1988)
adding the turbocharger, the engine was re-calibrated to
operate on E85 fuel.
CONTACT
The design was guided by simulations using WAVE and
FLUENT, coupled with CAD and continued evaluation Jason P. Moscetti jpmos@[Link]
from dynamometer and vehicle tests. Overall, the
original design criteria of higher torque output in the
range of 6,000 -10,000 RPM was met. Continued vehicle
tests and fine-tuning of engine calibration settings
allowed improvements including a substantial gain in fuel
economy.

101
Downloaded from SAE International by Purdue University, Monday, October 29, 2018
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

2007-01-3553

Optimizing the Design of the Air Flow Orifice or Restrictor for

Race Car Applications
Harry C. Watson, Andrew Gauci, Faeez Yousuff, and Alberto Boretti
University of Melbourne

Copyright © 2007 SAE International

ABSTRACT that requires all of the engine combustion air to pass

through a hole of prescribed diameter. As the approach
Several race car competitions seek to limit engine power and departure wall shapes to this hole, usually termed
through a rule that requires all of the engine combustion orifice or restrictor are not prescribed, there is
air passes through a hole of prescribed diameter. As the opportunity for innovation in these shapes to obtain
approach and departure wall shapes to this hole, usually maximum flow and therefore power.
termed orifice or restrictor are not prescribed, there is
opportunity for innovation in these shapes to obtain Formula SAE is one such competition in which the air
maximum flow and therefore power. flow is limited by requiring all the combustion air
consumed by the engine to pass through a 20 mm
This paper reports measurements made for a range of diameter hole. Initiated by the Society of Automotive
restrictor types including venturis with conical inlets and Engineers (SAE) in 1982, the Formula SAE competition
outlets of various angles and the application of slotted provides the opportunity for engineering student
throats of the ‘Dall tube’ type. members to conceive, design, fabricate, and compete
with small formula-style racing cars. Restrictions placed
Although normal venturis have been optimized as on the vehicles design and engine performance ensure
subsonic flow measuring devices with minimum that the knowledge, imagination, and creativity of the
pressure losses, at the limit the flow in the throat is sonic students are challenged. One such restriction on engine
and the down stream shocks associated with flow performance is outlined below in the extract from the
transition from sub-sonic to sonic are best handled with 2006 FSAE® Rules: [Link] Intake System Restrictor. [1]
sudden angular changes and the boundary layer
minimized by the corner slots between the convergent Recent papers on optimizing FSAE induction system
and divergent cones. design include the shape of the inlet trumpets [2], and
the inlet system as a whole [3]. The latter paper clearly
The measurements made here were on a quasi-steady demonstrates that the flow into each cylinder causes an
flow rig using a Roots type blower to simulate repeatable associated period during which the flow at the restrictor
flows and for a restricted range of operating to become sonic. Of course, the maximum flow is
temperatures over which small corrections to datum achieved with a design which arranges for the flow to be
conditions could easily be applied. sonic at all times. This can be achieved in several ways,
for example a large number of cylinders, a sufficiently
The conclusions made are that best results are obtained large plenum, or the use of a turbocharger compressor
for the conical inlet nozzle conical diffuser devices with acting as a vacuum [4,5, 6]
slots at the throat, quite consistent with Dall’s original
claims as found in older editions of the BS 1042 flow Most common amongst restrictor designs are venturis
standard. This design maximizes the mass flow integral based on the wall shape profile developed for flow
over a range of volumetric flow rates from Mach number measuring devices with minimum pressure loss,
M=0.6 to a wide range of M=1 conditions simulating sometimes described as maximum pressure recovery,
increasing downstream MAP with increasing engine as found in BS1042 [7]
speed. This suggests ability to outperform more
conventional devices by a few per cent, which could be A simplified version based on conical shapes is the Dall
important in delivering more flexible and powerful venture or ‘Dall tube’, (also found in pre 1989 versions of
engines with more conventional nozzles. BS1042) particularly suited to measuring large flows
such as found in hydro-electricity where machining
INTRODUCTION classical venturis of several metres in diameter can be a
significant challenge. The two cones, the nozzle and
Several international and national car racing diffuser are connected by a throat which is slotted for
competitions seek to limit engine power through a rule boundary layer control, in a similar way to the slots on

103
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

leading edge flaps used on many large aircraft to In the third case with supersonic flow downstream of the
increase lift at take-off. At the limit the flow in the throat throat the pressure loss is larger.
which forms the minimum cross section is sonic and the
shocks associated with flow transition from sonic to sub The objective in restrictor design is to maximize the
sonic are best handled with linear rather than curved mass flow through the throat assuming that the engine
surfaces. has enough volumetric efficiency to deliver the required
GP at the exit, i.e. usually the plenum chamber of the
OBJECTIVE inlet manifold. The volume of the plenum is an important
parameter as found in simulations performed for recent
The purpose of this research was to determine the engine design analysis, as when even applying
influence of cone angles and throat slot dimensions on turbocharging the suction from the compressor is not
the performance of the Dall venturi that maximises the always sufficient to ensure continuous sonic flow under
flow under near sonic and sonic conditions that all operating conditions.[ 4]
correspond to full throttle operation in a well designed
FSAE engine intake. A subsidiary objective is to Figure 2 illustrates the influence of plenum size on the
determine the influence of upstream losses from devices duration of sonic flow for a four cylinder engine at a
such as the air filter and throttle plate on performance. condition when the plenum is pressure is not maintained
at the critical pressure ratio. Plainly, the air mass flow is
THEORY less, in both cases compared to the case when sonic
flow is maintained continuously.
Three cases of flow through an intake venturi are shown
in Figure . The first is subsonic flow where the pressure M/Mmax
recovery is good (exit pressure is almost unity, GP is 1.2
Small Plenum
small). With sonic flow at the throat (M=1),in the second Larger Plenum
1
P/Po
1
0.8
0.8

GP 0.6
0.6
Subsonic
0.4 0.4
Just sonic at throat
0.2 Supersonic
Shock 0.2
0
0 1 2 3 4 5 6 7 8 9 10
0
0 200 400 600 800
M
2.5
deg CA
2

1.5 Figure 2 Proportion of sonic mass flow through the

restrictor for a given upstream condition illustrating
1
the effect of plenum chamber size.
0.5
MAXIMUM THEORETICAL MASS FLOW RATE
0
0 1 2 3 4 5 6 7 8 9 10
The maximum mass rate of flow in an ideal gas through
5 an isentropic nozzle is limited by the stagnation
A/A 5 conditions and the minimum cross-sectional area. No
matter how much the downstream pressure is reduced
or the dimensions of the duct either side of the throat are
changed, this maximum flow rate cannot be exceeded
[4].

From this theory, the maximum mass flow rate is defined

-50 as,
0 1 2 3 4 5 6 7 8 9 10

Figure 1 Pressure ratio P/Po, Mach No M and area P0 k M

ratio A/A5(throat) showing subsonic, near sonic m ( theoretical max imum ) Au u ( k 1) .
T0 R § k 1
and supersonic flow in the divergent section of an 2·
2 ( k 1 )

intake venturi ¨1  M ¸
© 2 ¹
…… (1)
case, the GP at exit is a little larger because of the
irreversibilities associated with the shock at the throat.

104
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

The Mach number, M at the throat is taken to be one, ‘Co’, the expansibility factor ‘İ’ and the velocity of
since the maximum mass flow rate that can be achieved approach factor ‘E”.
through a convergent-divergent nozzle occurs when the
flow is sonic at the throat region. Table 1 summarises
the values of the parameters that appear in Equation 1
to calculate the theoretical maximum.

Table 1: Parameters used in theoretical choked

mass flow rate

Mass Flow Rate (Mach 1 at Throat)

Area restrictor throat A 0.00031 m²
Ambient Pressure P0 101300 Pa
Ambient Temperature T0 293.2 K
Ratio of Specific Heats k 1.40
Gas Constant for Air R 287 J/kg.K

Mass Flow Rate m max 0.0751 kg/s Figure 3 Picture of Roots type blower, variable
speed drive and flow bench

FLOW BENCH The mass flow rate is calculated by

The flow bench used for measuring the orifice mass flow m ( orifice ) C0 EHA 2 u 'p u U upstream …….. (2)
is that used for cylinder head and port development. As
depicted in figure 3, a Roots type blower is used as a
suction pump to draw air through the sonic nozzle. In
where the density of the air at the upstream tapping of
the duct between the bench and the blower is mounted
the orifice plate ‘ȡ upstream’ is calculated using the perfect
an orifice plate with 1/2D and 1/4D upstream and
gas equation;
downstream tappings, respectively. The volume
between the blower and the orifice is sufficiently large to
eliminate flow pulsations at the orifice – a necessary Pupstream
condition since the volumetric air consumption of this
U upstream . …….. (3)
Rair Tupstream
type of two lobed rotor device is not constant, although a
lobe form with an helix along the axial direction
minimises the pulsating flow to some extent. and the pressure drop across the plate, ‘¨p’ is calculated
from
MASS FLOW RATE CALCULATION

In order to measure the mass flow rate through the 'p U gauge u g u 'h ……. (4)
system a flow measuring device such as an orifice plate
can be utilized. The orifice plate creates a calibrated where ‘ȡ gauge’ is the density of the fluid in the pressure
pressure disturbance to the flow, and pressure tappings,
measuring device of differential height 'h .
inserted upstream and downstream of the plate,
measure the differential. The pressure drop across the
plate is calculated based on the differential head The orifice discharge coefficient ‘Co’ is determined from
measurements taken from the pressure tapping either a look-up table in reference [7], based on the bore ratio
side of the plate, this pressure drop across the plate ‘ȕ’ which is defined by the ratio of the diameter of the
causes the fluid in the manometer to be displaced. The orifice to the diameter of the pipe. The expansibility
pressure drop across the plate can be calculated based factor ‘İ’ was determined in the same way refer to [7].
on the local gravitational constant and from the
displacement and density of the manometer fluid. The velocity of approach factor ‘E’ is determined using
the relationship with the bore ratio,
The pressure of the air in the pipe at the upstream
tapping of the plate is measured to determine the E (1  E 4 ) 1 / 2 …….. (5 )
density of the air at this point. Other factors that
determine the mass flow rate are dependent on the
geometry of the plate and the pipe. These factors are;
the area of the orifice ‘A’, orifice discharge coefficient

105
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

ORIFICE PLATE varies linearly with speed. The mass flow is also shown
in Figure 4.
The orifice plate was designed using the parameters Volumetric Outlet Flow
Bench Characteristic
Mass Flow Rate (kg/s)

outlined in the Australian Standard (AS 2360.1.1) for (m3/s)

Measurement of Fluid Flow in Closed Conduits [8]. 0.14 0.14

Table 2 shows the key dimensions and factors that were 0.12 0.12
used in the calculation of the mass flow rate. Volume flow rate = 0.0000400*rpm + 0.0043
2
R = 0.9994658
0.10 0.10

0.08 0.08

Table 2: Orifice plate dimensions and Mass Flow = 3.33E-05rpm + 0.0036

2
R = 0.9995

correction factors 0.06 0.06

0.04 0.04

Orifice Plate AS 2360 0.02 0.02

0.00 0.00
0 500 1000 1500 2000 2500 3000 3500 4000
Pipe Inner Diameter D 104 mm Compressor Speed (rpm)

Orifice Diameter d 50 mm Figure 4 Volumetric air delivery volume corrected to

Bore Ratio ȕ 0.481 ambient conditions and corresponding mass flow
Discharge Coefficient Co 0.603
Velocity Approach Factor E 1.028
Expansibility Factor İ 0.989 It follows for the purposes of presentation that the
Area orifice A 0.00196 M² discharge coefficient and other flow rate related
parameters may be expressed in terms of the
compressor speed, which is a surrogate for engine
speed of a test engine.
The location of the pressure tappings are also advised in
the Standard, with the upstream tapping one pipe
diameter upstream of the plate and the downstream DALL FLOW TUBE
tapping lying half a pipe diameter from the plate. Their
location is just visible in Figure 3, where the tapping is The main difference between the Dall Flow Tube
the point from which a thin brown pressure line is (sectioned in Figure 5 and shown in 3-D format in Figure
placed, with the flow direction being from left-to right. 6) and the conventional Venturi is that there is a slot
The orifice is of course arranged so that its sharp edge located at the throat. It is intended that vortices are
is upstream. formed in the slot which act to prevent the flow from
contracting past the throat. It is known that strong
vorticity generated in the slot creates a favorable
DISCHARGE COEFFICIENT pressure gradient so that flow attachment occurs and
thus larger angles of divergence can be used, compared
to those used in the Venturi for which it is specified that
Here, the discharge coefficient for the intake tube venturi
the divergent angle shall not be greater than 7.5o.[1]
is defined as the ratio of the measured mass flow rate to
the theoretical maximum mass flow rate with sonic flow
at the throat.

m measured
C ds ……… (6)
m (theoretical  max imum )

The purpose of calculating the discharge coefficient is to

remove the effect of ambient properties of the air flowed
in the experiment. Accounting for the ambient pressure,
temperature and humidity will control this variable and
provide a common basis for comparison between the Figure 5 Dall tube showing throat details including the
devices tested. slot. Flow direction is from right to left.

CALIBRATION
Horace Dall developed the ‘tube’ at Kent Instruments UK
As the Roots blower is a volumetric displacement research center as a cheap alternative flow meter to the
device, volumetric flow rate is relatively insensitive to standard venturi because the larger pressure difference
suction side conditions as seen in figure 4. With no- between the throat and an upstream tapping improved
device mounted on the flow bench the mass flow also

106
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

accuracy or where the size of a venturi was The results shown in Figure 6 are for some of the more
impracticable such as in hydro-electric turbine flows. Its successful throat slot dimensions for the 20mm diameter
application to sonic flow conditions has not previously throat. It is clear that for low flows, at low blower
been reported. speeds, there is relatively little difference in Cds (or
mass flow) between the slots. However, sonic flow is
achieved much sooner, at 2700 rpm by the 3mm deep 2
mm wide slot. This means that at WOT (wide open
throttle) engine conditions corresponding to this
volumetric air consumption, more than 5% extra air
mass can be supplied, potentially allowing for more than
5 % increase in power or torque. Of course, the same
torque may achieved by a slightly higher speed where
the air consumptions match but the opportunity for a
wider power/torque band is lost.

Dall Slot Comparison

4mm Width

Figure 6 Computer graphic image of a sectioned

Dall Tube
3mm Width

INFLUENCE OF SLOT DIMENSIONS Average Discharge

Coefficinent

Here we explore whether the slot dimensions are the 0.9700-0.9800

same for sonic flow as for subsonic. The geometry 0.9600-0.9700 2mm Width
0.9500-0.9600
summarized in table 3 was the basis for these
0.9400-0.9500
experiments. A matrix of slot depths from 0-4mm and 0.9300-0.9400
widths from 0-4 mm was investigated on the flow bench.

Table 3: Dimensions of the Dall Flow Tube tested No Slot

with (0-4mm)x(0-4mm) slot

0 1 2 3 4
Slot Depth (mm)

Geometry
Total Figure 8 Contour plot of discharge coefficient for the
Nozzle Diffuser Length matrix of (0-4mm)x(0-4mm) slots tested
Diameter Angle Angle Diameter
mm deg deg mm mm Figure 8 shows the maximum discharge coefficients for
35 26 16 40 126 the matrix of slot depth by slot width. It can be seen that
2mm depth by 2.5 mm wide is the optimum.
Slot Depths and Width Comparison
1.0
EXAMINATION OF SOME PRACTICAL NOZZLE
0.9
AND DIFFUSER DIMENSIONS

Table 4 gives the nozzle and diffuser dimensions for

0.8
devices A through D. Their overall length including
Discharge Coefficient

2 deep 3 wide Dall Slot

2 deep 4 wide Dall Slot parallel connectors is given.

Table 4: Dimensions of the Dall Flow Tubes tested
0.7 2 deep 6 wide Dall Slot

3 deep 2 wide Dall Slot

0.6 Geometry
Total
Device Nozzle Diffuser Length
0.5
Diameter Angle Angle Diameter
mm deg Deg mm mm
0.4
900 1200 1500 1800 2100 2400 2700 3000 3300 3600 3900
A 50 20 12 40 182
Compressor Speed (rpm) B 50 40 14 40 147
C 35 26 16 40 126
Figure 7 Discharge characteristics of the more D 35 36 14 42 126
successful throat slots

107
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

The flow characteristics of the four configurations all with the car. It is noteworthy that design A achieves a flow
2x2mm slots is presented in Figure 9. Configuration A rate within 2.2% of the maximum theoretically possible.
with the largest intake diameter and narrowest nozzle
and diffuser angles delivers the highest mass flow (Cds) FLOW VISUALIZATION
although the A, B and D have similar flows up to 2000
rpm. C is characterized by the largest diffuser angle for A model was built of the core flow region of the C
which there is some penalty up to transonic speeds, design. The section was 10 mm in depth and with aid of
which as before, when attached to an engine could a transition piece mounted on the flow rig.
represent a reduction in mass flow and hence torque at
a given equivalent engine speed at WOT,

Dall Flow Tube Comparison 3600 rpm

1.00

0.95
Discharge Coefficient

0.90 (A) 50-20-12-40

0.85 s (B) 50-40-14-40 SHOCK

( C)35-26-16-40
0.80

(D) 35-36-14-42
0.75

0.70
900 1200 1500 1800 2100 2400 2700 3000 3300 3600 3900
Compressor Speed (rpm)

Figure 9 Comparison of the four configurations

mass flow as sonic based discharge coefficient Cds
Average Choked
Discharge Coefficient Dall Tube Comparison a) Core flow

1.000

0.990 0.9793 2600 rpm

0.9753
0.980
0.9666
0.970 0.9644

0.960

0.950

0.940
SHOCK
0.930

0.920 ATTACHED
0.910 BOUNDARY
LAYER
0.900
(A) 50-20-12-40 (B) 50-40-14-40 (C) 35-26-16-40 (D) 35-36-14-42
DALL TUBE CONFIGURATION

Figure 10 Comparison of the mass flow as sonic

based discharge coefficients averaged over the
coked flow range of compressor speeds of figure 8.
The average of the Cds once the Dall tube restrictor
becomes choked is given in Figure 9. The maximum
difference across the four configurations is about 1.2%. b) Wall flow
The ‘best’ design i.e. A, achieves the increased Cds or
mass flow at the expense of 60mm extra overall length. Figure 11 Visualization of the sonic flow using a planar
This extra length may represent a packaging problem in section of the Dall tube.

108
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Figure 11 shows images of smoke streams introduced INFLUENCE OF UPSTREAM FLOW

into the flow. The upper figure shows the flow CONDITIONS
introduced into the core. The contraction of the
streamlines through the throat is evidenced by the Designers of inlet systems often have concern about the
extremity of the smoke flow and the location of the shock upstream of the restrictor items such as throttle plate
can be seen. The post shock turbulence from the and air filter on the mass flow through the restrictor.
anisotropy across the strong shock at the higher speed Indeed device C was configured by the 2001 Melbourne
is evident. FSAE-A team to have a slide throttle that formed the slot
in the throat of the restrictor at wide open throttle, so as
In the lower image of figure 11 with the smoke to avoid having a throttle plate ahead of the restrictor.
introduced at the upstream wall and at a lower but still
sonic flow rate, the shock is seen to be weaker and Figure 13 reports for device D the effects of introducing
located closer to the throat. Excellent attachment of the two types of air filter and the throttle plate on the
boundary layer to the diffuser wall and the post shock maximum mass flow as Cds. It is plain that the choice
mixing of the smoke stream are evident. od air filter surface area and shape can be important.
The ideal filter allows the typical ‘mushroom’ shape of
These visualized effects are in line with Figure 1 stream line flow into the entrance of the nozzle section.
indicating the down stream movement of the shock with Flow losses of about 3% are shown for the small throttle
lower down stream pressure. and cylindrical air filter in this design. The losses for the
larger intake of device A were less.
PRESSURE RECOVERY ACROSS THE DEVICES
EFFECT OF INTAKE CONFIGURATION
The previous graphs for the devices have used blower
speed as a surrogate for engine speed as many readers 1.00

who work with engines are likely to identify better with 0.95

the engine speed base. For completeness we plot the

0.90 Bare
down stream pressure as a function of mass flow. As
Discharge Coefficient

expected device A shows the least pressure drop at the 0.85

Cylindrical Air filter

down stream static pressure measurement point of the 0.80

Throttle Plate

rig, which could be regarded as the plenum of an intake Conical Air filter

manifold. The higher pressure of device A compared 0.75

Cylindical Air Filter &
with device C is strong evidence of why an engine map 0.70
Throttle Plate

fuel recalibration might be needed if the fuel map is

calibrated on the usual speed-density relationship as 0.65

there is as much as 10kPa difference in pressures at the 0.60

same mass flow. Figure
900
13 1400 Effect
1900
of2400 upstream
2900
of 3400nozzle3900
Pressure Drop vs Flow Rate Compressor Speed (rpm)
configuration on mass flow as Cds including air
100 filter and throttle plate and combinations.
96
CONCLUSIONS
Downstream Pressure (kPa)

92

Maximising the mass flow through a restrictor limited by

88
a race car rules formula is a goal for maximizing engine
84 power. An experimental investigation has been
(A) 50-20-12-40
undertaken to optimize the design of a ‘Dall tube’ based
80 D) 35-36-14-42
restrictor that is relatively simple to manufacture
(C) 35-26-16-40
76 compared with the profile of a standard venturi. As at
the maximum flow limit the velocity in the throat of the
72 restrictor is sonic an optimized design will have the
smallest boundary layer possible in the throat so that the
68
0.025 0.035 0.045 0.055 0.065 0.075 sonic velocity extends across as much of the throat
Mass Flow Rate (kg/s) space as possible.

Figure 12 Downstream static pressure versus mass The best design found in this systematic study in which
flow rate showing significant difference in pressure the upstream and downstream (nozzle and diffuser)
that would be apparent in the plenum of a manifold angles and diameters were varied over a useable range,
for each of the devices in the transition to sonic flow it was found that design A, styled as 50-20-12-40,
region. outperformed three others. An approximately 2mm x
2mm throat slot was also determined to be close to
optimum. With this design nearly 98% of the maximum
mass flow was obtained.

109
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Whilst this design has the maximum integral under the 6. W Attard, What Limits Small Turbocharged Engine
flow curve as well as the maximum flow, it has the Design, PhD Thesis, University of Melbourne, in
disadvantage of being the longest. The best of the Preparation, 2007
shortest designs (D) suffered a 1.2% reduced maximum 7. British Standards BS1042 Measurement of fluid flow
flow capability. in closed conduits, ISI 1989, Section 1.1, 1.2., 1989.
8. Australian Standards AS2360 Orifice plate selection,
The visualization of the flow demonstrated that the slot design and applications Australian Standards, 1993,
in the throat ensured excellent attachment of the Section 1.3, 1.4, 1.5
boundary layer downstream. The shock moving 9. S Massey, Mechanics of Fluids, van Nostrand, 1972,
downstream with increasing differential pressure was pg 84, 198, 401.
observed in line with theory.

The flow studies demonstrated that the largest

CONTACT
differences in flow between the devices occurred in the
Prof Harry C Watson,
just subsonic flow conditions. Nearly 10% reduction in
Dept. Mechanical and manufacturuing Engineering,
mass flow difference could occur at some conditions. In
application to an engine this would have the effect of University of Melbourne
narrowing the constant power speed range (reducing the Australia, 3010.
engine speed at which maximum power is first
developed). It would also likely call for recalibrating the harrycw@[Link]
engine fuel map at this condition.

The testing alos illustrated that the upstream pressure

losses that result from the air filter, throttle and other
potential pressure losses are also important in achieving
the highest possible mass flow rates.

ACKNOWLEDGEMENT

The student authors (nos 2 & 3) wish to acknowledge

the contributions made by their predecessors in
designing and developing some of the devices tests;
particularly William Attard who designed device D and
assisted the Melbourne FSAE team over many years.

The first author wishes to acknowledge the influence of

Horace Edward Dall, one time chief engineer of George
Kent Limited (UK) who patented the device baring his
name, and introduced as much lateral thinking into flow
measurement and flow system design as he did to the
development of optics - a genius engineer.

We acknowledge gratefully the review of this manuscript

by Elisa Toulson.

REFERENCES

1. Rules Society of Automotive Engineers, “2007

Formula SAE ® Rules”, Section [Link], 2006
2. H C Watson and E E Milkins, Cylinder Head Design,
SAE AUST. Paper No 7246/1, 1-34pp, 1972.
3. H C Watson Power Generation Systems (436-409),
Lecture notes, University of Melbourne, 2003..
4. W. Attard, H C Watson, S Konidaris, Development
of a constant power engine for FSAE competition,
SAE paper 2006-01-0745. 2006.
5. W. Attard, H C Watson, S Konidaris, Highly
turbocharging a restricted, odd fire, two cylinder
small engine – design, lubrication, tuning and
control, SAE paper 2006-01-3637, 2006

110
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

2006-01-3653

A Theoretical and Experimental Study of Resonance in a High

Performance Engine Intake System: Part I
S. Brennan, R. J. Kee, R. G. Kenny, R. Fleck, and J. A. Gaynor
Queen’s University Belfast

B. Fleck
Advanced Engine Technology Ltd.

Copyright © 2006 SAE International

ABSTRACT INTRODUCTION

The unsteady gas dynamic phenomena in engine intake This work focuses on resonance in an intake airbox,
systems of the type found in racecars have been typical of those found in Formula 1, where the airbox sits
examined. In particular, the resonant tuning effects, above the engine and has a forward facing inlet above
including cylinder-to-cylinder power variations, which the driver’s head. Air flows through the inlet to a diffuser
can occur as a result of the interaction between an which leads to the main body of the airbox. While steady
engine and its airbox have been considered. flow losses may be important in the diffusing section of
the airbox, unsteady gas dynamic effects are also
Frequency analysis of the output from a Virtual 4-Stroke significant. It is the unsteady gas dynamic behaviour of
1D engine simulation was used to characterise the the airbox which is examined here.
forcing function applied by an engine to an airbox. A
separate computational frequency sweeping technique, Airbox behaviour is of interest because it influences
which employed the CFD package FLUENT, was used overall engine performance. In order to optimize the
to determine the natural frequencies of virtual airboxes performance of the engine and airbox system, the
in isolation from an engine. Using this technique, an geometry must be carefully considered. For example,
airbox with a natural frequency at 75 Hz was designed the dimensions of the airbox will affect the propagation
for a Yamaha R6 4-cylinder motorcycle engine. The of pressure waves through the intake system and hence
existence of an airbox natural frequency at 75 Hz was the airflow to each cylinder.
subsequently confirmed by an experimental frequency
sweeping technique carried out on the engine test bed. Traditionally, steady flow computational fluid dynamics
(CFD) is used to analyze airflow within the airbox and
A coupled 1D/3D analysis which employed the engine one-dimensional simulation tools are used to model the
simulation package Virtual 4-Stroke and the CFD intake system, engine and exhaust. Ideally 3D CFD
package FLUENT, was used to model the combined would be used to model the entire system. However,
engine and airbox system. The coupled 1D/3D analysis computational times are prohibitive for most users. An
predicted a 75 Hz resonance of the airbox at an engine advantage of one-dimensional methods is that they are
speed of 9000 rpm. This frequency was the induction less computationally intensive than multi-dimensional
frequency for a single cylinder. CFD and results for a complete engine model can be
obtained in minutes. The 1D code used in this work,
An airbox was fabricated and tested on the engine. Virtual 4-Stroke [1] by OPTIMUM Power Technology
Static pressure was recorded at a grid of points in the (Pittsburgh, PA, USA), is based on the GPB method [2].
airbox as the engine was swept through a speed range In the Virtual 4-Stroke code, one-dimensional elements
of 3000 to 10000 rpm. The measured engine speed are used to model pressure wave effects through the
corresponding to resonance in the airbox agreed well ducting, and 0D models are generally adopted to solve
with the predicted values. There was also good the thermodynamics in volumes where the performance
correlation between the amplitude and phase of the can be adequately described by averaged quantities
pressure traces recorded within the airbox and the without taking into account the wave action. Full non-
1D/3D predictions.

111
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

isentropic boundary solutions are used at each The frequency of the waves was found to remain
geometric and thermodynamic discontinuity. relatively constant at engine speeds of 10000 and 15000
rpm, and it appeared that some form of resonance was
While 1D methods are useful for areas within the engine occurring in the airbox. The low frequency pressure
where the flow is predominantly one-dimensional, they wave resulted in power variations between cylinders; the
cannot account for three-dimensional (3D) effects, and variations were significant and in the order of 15%.
so there are limitations in modelling regions of complex These were attributed to cylinder-to-cylinder variations in
geometry. A combination of 1D and 3D methods were the average pressure in the vicinity of the bell mouth
used in the present work. The airbox was modelled intakes during the inlet valve open periods. However, the
using the CFD package FLUENT, and the engine was presence of the low frequency pressure wave was
modelled using Virtual 4-Stroke. The boundaries neither explained nor validated. The work raised
between the Virtual 4-Stroke and FLUENT models were questions regarding the fundamental gas dynamic
initialized and run simultaneously, and information was behaviour within an airbox, and the importance of the
passed between the codes on a timestep-by-timestep geometrical and hence resonant properties of the airbox.
basis. The solution timestep was set dynamically by the In the work presented here, the low frequency pressure
Virtual 4-Stroke code according to the Courant- wave predicted by Maynes et al. was explained and
Friedrichs-Lewy (CFL) stability criterion. While experimentally validated.
convergence of the CFD models was ensured by the
CFD code at each timestep, convergence of the coupled While this work followed on from that of Maynes et al, it
model was determined by Virtual 4-Stroke. This coupled was not feasible to test a F1 engine. Instead, a high
approach enabled an entire system to be modelled, performance 4-cylinder engine from a Yamaha R6
while maintaining the benefits associated with each motorcycle was used to validate the coupled Virtual 4-
method individually. Coupled simulation is a powerful Stroke and CFD simulations.
tool for design optimization of engine components, and
in particular for those, like the airbox, in which the 3D FORCING FUNCTION
geometry is a major factor. The advantages of a coupled
1D/3D approach have been shown by many authors, Pressure oscillations in an airbox are an example of a
e.g. [3, 4, 5 and 6]. forced vibration. Every valve opening and closing event,
for each of the cylinders, will cause finite amplitude
Maynes et al [3] at Queen’s University used a pressure waves to propagate into the airbox; these
dynamically coupled 1D/3D model to predict F1 engine waves cause pressure fluctuations within the airbox. The
and airbox performance. A generic 50 litre airbox and number of cylinders, the valve timing, the firing order,
V10 engine were considered. The coupled simulations and the engine speed, will all influence this wave
were found to highlight the sensitivity of the coupled formation to varying degrees.
engine and airbox simulation technique to changes in
airbox geometry, and also to emphasize the three- Resonance will occur in the airbox if the forcing function
dimensional and unsteady nature of the flow in an applied by the engine to the airbox includes frequency
airbox. Of particular interest in Maynes’ work was the components which are close to a resonant frequency of
predicted presence of large amplitude, low frequency, either all or part of the intake system. The occurrence of
pressure fluctuations in the 1D/3D airbox. The pressure resonance in the intake system can be desirable; it may
fluctuations were at a frequency lower than the induction be possible to “tune” the engine, so that the system
frequency and their amplitude was an order of responds in a desired manner at specified engine
magnitude greater than steady flow pressure losses speeds. However, it is not always beneficial and in the
predicted by the steady CFD model. case of an airbox may lead to large cylinder-to-cylinder
power variations as predicted by Maynes et al. To
Generally for a multi-cylinder engine attached to an understand resonance in the airbox, it is first necessary
airbox, the engine induction frequency is the dominant to determine the frequency components in the forcing
frequency of the pressure waves in the airbox; this is function which are exciting the resonance.
intuitive and has been demonstrated for example by
Harrison et al [7]. To illustrate this point, in an airbox FREQUENCY ANALYSIS OF FORCING FUNCTION
attached to a 4-cylinder engine over the course of one
engine cycle there would be four significant pressure A Virtual 4-Stroke model of Maynes’ F1 V10 engine was
variations in the airbox; one for each valve opening used to examine the forcing function applied by the
event. Since the frequency of the airbox pressure engine to the airbox. It is not sufficient to simply use the
variation equals the multi-cylinder induction frequency, pressure-time history at the intake valves as a forcing
all cylinder inlets will experience similar pressure versus function, since normally this will be a superposition of
crank angle characteristics, i.e. for all cylinders, the outward (denoted leftward) and returning (rightward)
pressure at e.g. inlet valve opening will be similar. pressure waves. In order to look at the pure forcing
function supplied by the engine, one of the intake pipes
In the work carried out by Maynes et al, the frequency of in the model was extended to a length of 10 m. For this
the pressure fluctuations predicted in the airbox was pipe length, it is possible for the virtual engine to run (at
considerably lower than the engine induction frequency. high engine speed) for several cycles before reflections

112
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

from the end of the open pipe reach the intake valve. Maynes et al, it was frequency components from the
The pressure at the intake valve, on the cylinder with the individual cylinders which were exciting the pressure
long pipe, represents only the leftward moving waves, or oscillations in the airbox. Maynes et al observed a
the forcing function, for the first few cycles. pressure wave at 250 Hz in the airbox at 15000 and
10000 rpm. At 15000 rpm, 250 Hz is the first harmonic of
The engine, complete with 10 m pipe was simulated for the single-cylinder induction frequency, and at 10000
a range of speeds. To ensure that only the leftward rpm, 250 Hz is the second harmonic of the single-
moving waves were considered, the Dynoscope tool in cylinder induction frequency.
Virtual 4-Stroke was used to see at which cycle the first
reflected rightward moving waves returned to the valve. AIRBOX RESONANT FREQUENCIES
Figure 1 shows the static pressure due to the leftward
moving pressure waves, at the intake valve, for a single In order to verify that the airbox in the work by Maynes et
cylinder over the course of several engine cycles, prior al had a natural frequency close to 250 Hz, and gain an
to the arrival of reflections. understanding of airbox resonance characteristics, it
was necessary to find a way to determine the natural
frequency of the airbox. As a starting point, traditional
methods were considered.
Static pressure (Pa)

A common way to predict the behaviour of an intake

plenum in an IC engine is to model it as a Helmholtz
resonator. A Helmholtz resonator consists of a cavity,
communicating through an open hole or neck to some
external excitation. The air in, and near, the neck can
vibrate back and forth, as the air in the cavity acts like a
spring. Equation 1 is the most basic form of the
Crank angle (degrees) Helmholtz equation.
c A
Figure 1 Simulated static pressure due to leftward f
moving pressure waves for the first 6 engine cycles of
2S VL (1)
the F1 V10 at 15000 rpm
Equation 1 gives a single frequency f which is a function
An FFT of the data presented in Figure 2 shows of the velocity of sound c, the resonator cavity volume V,
frequency components at the single-cylinder induction the neck length L and the cross sectional area of the
frequency and its harmonics. It should be noted that the neck A, but is independent of the cavity geometry.
convention adopted here is that the first harmonic is
considered to be twice the fundamental frequency; in Extensive experimental work has been carried out to
some of the related literature, the first harmonic denotes investigate the limits of the Helmholtz equation. Almost
the fundamental frequency. all work published in the field, for example Ingard [8] and
Chanaud [9], suggests that adjustments and corrections
90
are necessary to make the equation more accurate. The
80
125Hz (Single-cylinder induction frequency) shortcomings of the equation are: the original equation
proposed by Helmholtz gives the same frequency
70 regardless of the cavity’s geometry, and it does not
include a length correction, or allow for a change in neck
60
position, or geometry. Typically, airboxes have a
Power spectrum

50
complex geometry and salient dimensions which are not
insignificant when compared to the resonant
40 wavelength. For this reason, the traditional Helmholtz
250Hz
equation, directly applied, may not be accurate, so it is
30
important to consider its limitations. Some alternative
20
equations have been presented in the literature. For
example Panton et al [10] presented Equation 2 for
10 375Hz
500Hz
calculating multiple resonant frequencies Ȧ, in radians
per second. The resonant frequencies were calculated
0
0 100 200 300 400 500 600 from the local speed of sound a, the neck length ln, the
Normalized Frequency (Hz) volume length lc, and An and Ac, the neck and cavity
Figure 2 Frequency content of the leftward moving cross-sectional areas respectively.
pressure waves of the F1 V10 at 15000 rpm
Zln Zlc An
The analysis of the leftward moving pressure wave from
tan tan (2)
a a Ac
a single cylinder suggests that, in the work presented by

113
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Equation 2 gives a slightly lower value for the monitored throughout the airbox. Both approaches gave
fundamental frequency than the lumped parameter identical resonant frequencies, and the neck was chosen
Helmholtz equation, as the length-to-width ratio of the as the input boundary for all subsequent simulations, as
cavity becomes greater than 1. However, while illustrated in Figure 3. Wall boundary conditions were
Equation 2 can account for multiple resonant applied at the exits from the 10 outlet pipes.
frequencies, it cannot account for complex geometries.

Neither the Helmholtz equation nor the equation

presented by Panton et al [10] can accurately predict
resonant frequencies for complex airbox geometries.
Consequently, the resonant frequencies of a complex
shape are most often found experimentally, which can
be both time consuming and costly as each permutation
in geometry must be fabricated. It is possible to add
correction factors to the Helmholtz equation, and the
approach used by Panton et al allows resonant
frequencies to be calculated more accurately for a wider
Figure 3 Frequency sweep applied to airbox inlet
range of shapes. However, the gap between theory and
experiment widens as the volume deviates from a
The static pressure at the centre of the airbox in the
perfect sphere or cube. A multi-dimensional approach is
region of a resonant mode is shown in Figure 4. There is
required to improve the accuracy of the predictions.
a clear increase in amplitude in the region of 260 Hz.
The same behaviour was observed for the subsequent
DETERMINATION OF NATURAL FREQUENCIES modes. However, these are not shown as the plot
USING CFD becomes less clear. A frequency analysis of these
pressure data over the course of the simulation is shown
In experimental work, a frequency sweep is often used in Figure 5. The first four resonant modes can be
to find the resonant frequencies of a cavity. An applied identified.
frequency is swept through a frequency range, and a
two-microphone approach is used to measure the input
and the response of the resonant cavity.

It was decided to mimic the frequency sweep approach

using CFD. In order to determine the resonant
frequencies of a cavity, a model was created in CFD,
and a FLUENT user-defined function (UDF) was used to
supply a sinusoidally varying pressure boundary. The
UDF supplied both a total pressure and a total
temperature at the chosen boundary. A starting
frequency was specified, and one cycle was completed
at this frequency. The frequency of each subsequent
cycle was increased by an increment of 2 Hz. The
pressure at a number of points in the model and the Figure 4 Variation of static pressure at the centre of F1
mass-flow rate at the boundary were monitored over the airbox
course of the simulation, and an FFT was performed on 60
the pressure and mass-flow data. This identified the
resonance frequencies. 50
56 Hz
The method was checked on models of a simple open-
40
ended pipe and a Helmholtz resonator, and the results 260 Hz
Power spectrum

for the resonant frequencies were in good agreement

with those determined using traditional lumped- 30

parameter calculation methods.

20

RESONANT FREQUENCIES OF A 50 L AIRBOX

357 Hz
10
The frequency sweep method was used to find the 542 Hz
resonant frequencies of the F1 airbox used by Maynes 0
et al in their paper. The fixed amplitude sinusoidally 0 100 200 300 400 500 600 700 800 900 1000
Frequency (Hz)
varying pressure could be applied to the airbox either at
the inlet (neck) or at the outlets to the engine. In both
approaches, the pressure at a number of points in the Figure 5 FFT of static pressure data from centre of F1
airbox and the mass-flow rate at the boundary were airbox

114
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

From Figures 4 and 5, it can be seen that the F1 airbox the side walls and the inclusion of additional volumes on
modelled by Maynes et al has a natural frequency at the sides near the base of the airbox, as often seen in
approximately 260 Hz. It was this resonant mode which Formula 1. The full results of the parametric
was excited at 250 Hz by the 2nd harmonic of the investigations are not presented here.
individual cylinder induction frequency at 10000 rpm and
the 1st harmonic at 15000 rpm. CFD methodology

Therefore, by characterising the forcing function and For all airbox geometries, there is more than one
determining the airbox’s natural frequencies, it was resonant mode and it is computationally expensive to
possible to explain why the work by Maynes et al sweep through the entire frequency range in one
predicted a low frequency pressure wave in the airbox at simulation. Instead, for each of the geometries
a frequency lower than the multi-cylinder induction considered, two simulations were run. A larger time step
frequency. However, the work carried out by Maynes et was used for the lower frequency range, and a smaller
al was never experimentally validated. The following timestep was used for the higher frequency range. There
sections detail the validation carried out as part of this was an overlap between the two models to ensure that
work. no mode was missed.

DESIGN OF AN AIRBOX FOR THE YAMAHA R6 For each of the airbox models considered, the standard
k-İ model was used for turbulence. The fluid, air, was
It was not feasible to test a F1 engine. Instead, the modelled as an ideal gas and the walls of the airbox
engine which was used in this work to validate the were modelled as adiabatic. A sinusoidally varying
coupled Virtual 4-Stroke and CFD simulations was a boundary condition was applied at the airbox inlet, as
4-cylinder Yamaha R6 (2003 model). The R6 is a high- detailed earlier, and the timestep was calculated with the
performance motorcycle engine, and has a maximum CFL stability criteria in mind. The model was double
engine speed of 15500 rpm. The models simulated by precision and the PISO algorithm was used for solving
Maynes et al were at speeds of 10000 and 15000 rpm, the pressure velocity coupling. Hexahedral cells were
but a typical F1 engine may have a maximum speed of used throughout the model.
between 16000 and 20000 rpm. It was decided to
perform the experiments at lower engine speeds in order After some trial models and mesh independency studies,
to limit the possibility of damage to the R6 engine. a reasonably coarse mesh was used with cell numbers
between 7000 and 15000, depending on the model size.
The lower the engine speed, the lower the frequency A frequency sweep on a 7000 cell model took a
components applied to the airbox. Therefore if the airbox minimum of 16 hours for each of the two stages on a 3.4
is to be excited, it must also have a lower resonant GHz computer with 1 GB of RAM. An example of a
frequency. It can be shown using the traditional mesh used is shown in Figure 6.
Helmholtz equation that the lower the desired natural
frequency of a volume, the larger the volume must be.
However, the larger the volume of the airbox, the lower
the amplitude of the pressure waves in the airbox.

Thus there were conflicting requirements in creating a

high amplitude pressure wave at lower speeds. In order
to limit the size of the volume, it was decided to design
an airbox which resonated in the region of 8000 to 9000
rpm. However, an airbox is a complex shape, and in
order to design an airbox which resonated in this speed
range, it was necessary to first understand how the
geometry influenced the natural frequencies of the
airbox.

PARAMETRIC STUDIES
Figure 6 Sample mesh for 20 litre airbox
FLUENT and the frequency sweep method described
earlier was used to examine a wide range of geometric The understanding gained of airbox resonant
factors in order to understand how they affect the characteristics made it possible to design an airbox with
resonance frequencies of the airbox. A simplified a specific natural frequency which resonated at a given
geometry was chosen to approximate an airbox. The engine speed. As explained earlier, it was decided to
geometry consisted of a long cuboid representing the design an airbox which would resonate at the single-
body of the airbox and a smaller cuboid representing the cylinder induction frequency in the region of 9000 rpm,
neck. Among the factors considered were the i.e. 75 Hz. This would also mean that the airbox would
dimensions of the main box, the dimensions and position resonate at 75 Hz when excited by the 1st harmonic of
of the neck, the position of the intake pipes, the angle of the single-cylinder induction frequency at 4500 rpm.

115
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Similarly, it would be excited by the 2nd harmonic at 2250 rpm the single-cylinder induction frequency is 75 Hz and
rpm, and the 3rd harmonic (or multi-cylinder induction the engine induction frequency is therefore 300 Hz. If
frequency) at 1125 rpm. there is no resonance in the airbox at 9000 rpm, or there
is a resonant mode in the region of 300 Hz, then 300 Hz
Out of the many geometries considered, a 570 mm by will be the most dominant frequency.
235 mm by 200 mm model, with a 100 mm by 100mm by
100 mm neck had a resonant frequency of 73.4 Hz. This The static pressure was recorded at a grid of points
frequency equals the single-cylinder induction frequency throughout the box. An FFT of the static pressure at the
at 8812 rpm. centre of the airbox is shown in Figure 7.
45
COUPLED 1D/3D ANALYSIS
40 75Hz
Single-cylinder
The airbox described in the previous section was 35 induction frequency at
dynamically coupled to a 1D Virtual 4-Stroke model of 9000rpm
the R6 engine. Inlet pipes of diameter 40 mm and length 30

Power Spectrum
250 mm were included in the CFD model. These pipes
25
were positioned 140 mm centre-to-centre distance apart
on entering the airbox. 20
300 Hz: 4 times single cylinder
induction frequency = engine
15
It was deemed impractical to record data for the entire induction frequency
3D flow field in the airbox over the course of the 10
simulation, as the volume of data would excessive. 450 Hz: 6 times single cylinder
induction frequency
Instead, variables such as temperature and pressure 5
were recorded at discrete points in the model over the 0
course of the coupled simulation. 0 100 200 300 400 500 600 700 800 900 1000
Normalised Frequency (Hz)

For the coupled airbox model, the standard k-İ Figure 7 FFT of the predicted static pressure at the
turbulence model was used. Species transport was centre of the airbox, 9000 rpm
enabled within FLUENT and the gas was modelled as
ideal to allow for compressibility effects. The walls of the Peaks in the power spectrum are seen at 75 Hz, 300 Hz
airbox were modelled as adiabatic. A “pressure outlet” and 450 Hz. The peak at 75 Hz indicates a resonance
boundary condition was applied at the airbox neck condition which is due to excitation of the fundamental
where atmospheric conditions were applied. At each of natural frequency of the airbox by the single-cylinder
four boundaries connecting the CFD model to the Virtual induction frequency. A fluctuation in pressure at 300 Hz
4-Stroke model, “pressure inlet” boundaries were is due to the overall (4 cylinder) engine induction
applied. Each of the intake pipes in the Virtual 4-Stroke frequency. The peak at 450 Hz corresponds to the 5th
model was matched by region number to the harmonic of the single-cylinder induction frequency.
corresponding intake in the CFD model. The values Note that some harmonics are not evident in the FFT
supplied at these coupled boundaries were calculated due to cancellation; this phenomenon will be described
through customized user subroutines supplied with in a subsequent publication.
Virtual 4-Stroke. This enabled time-dependent
thermodynamic properties along with species and Figure 8 shows the predicted static pressure for 3
turbulence details to be passed between the codes, with locations in the airbox versus crank angle. At all 3
no user intervention required. The PISO algorithm was locations, there is a clear dominant wave once per
used for solving the pressure velocity coupling, and the engine cycle (720°), i.e. at 75 Hz, with superimposed
timestep was set dynamically by Virtual 4-Stroke components at 300 Hz and 450 Hz.
according to the CFL stability criteria. Hexahedral cells 1500
were used throughout the model. Typically in this work it
was found that for a four-cylinder engine, a coupled 1000

model took between 12 and 15 cycles to converge. 500

Static Pressure (Pa)

The coupled simulation was run at a series of discrete 0

engine speeds. At each speed, after convergence was

-500
reached, the data for the next four cycles in Virtual 4-
Stroke were matched to the data for the corresponding -1000
four cycles in FLUENT. As the timestep was not
constant, the data were then interpolated so that the -1500

variables were known for each 1° crank angle. An FFT -2000

Midpoint 100mm from back wall 100mm from front wall

was performed on the data to get a picture of the 0 360 720 1080 1440 1800 2160

dominant frequencies within the airbox. If no resonance Crank angle (degrees)

occurred in the airbox, then the dominant frequency will Figure 8 Predicted static pressure versus crank angle, 3
be the engine induction frequency. For example, at 9000 engine cycles at 9000 rpm

116
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Figure 9 shows a FFT of the corresponding data for a predicted difference in power between Cylinders 3 and 4
different airbox geometry which had a predicted in the coupled 1D/3D model can be explained by
resonance at 88 Hz. Here, the engine induction considering the pressure in the airbox for one engine
frequency (300 Hz) is the dominant component, and the cycle, at a position mid-way between Cylinders 3 and 4.
component at 75 Hz has relatively low amplitude. These The predicted pressure at this position for one engine
characteristics are in marked contrast to those cycle is shown in Figure 10. The valve open periods for
presented in Figure 7 for the original airbox. Cylinders 3 and 4 are indicated on the graph. Inlet valve
opening is shown as IO, and inlet valve closing as IC.
55
50
45 300 Hz
engine induction
40 frequency
Power Spectrum

35
30
25 75 Hz
Single cylinder
20 induction
frequency
15
10
5
0
0 100 200 300 400 500 600 700 800 900 1000
Normalised Frequency (Hz)
Figure 10 Predicted pressure in airbox close to inlets of
Figure 9 FFT of the predicted static pressure at the
Cylinders 3 and 4
centre of an alternative airbox, 9000 rpm
From Figure 10 it can be seen that Cylinders 3 and 4
PREDICTED INFLUENCE OF LOW FREQUENCY
experience very different pressure-time histories in the
PRESSURE WAVE ON ENGINE PERFORMANCE
vicinity of their inlets during their inlet valve open
periods. The pressure in the airbox is close to
A 1D/0D model of the engine and airbox was created in atmospheric as the inlet valves on Cylinder 3 close. In
Virtual 4-Stroke. The airbox cavity was represented by a contrast, the pressure is almost 1 kPa below
0D element. In a 0D, element the pressure is constant atmospheric pressure as the inlet valves on Cylinder 4
throughout the airbox at any point in time; this will not be close. Thus, the cylinder-to-cylinder power variations
the case in a coupled model (or the real system), where can be attributed to pressure variations due to the low
it varies in three dimensions at any given time. frequency pressure oscillation in the airbox.
Therefore, the low frequency wave in the airbox at 9000
rpm will not be predicted by the 1D/0D model. It is important to note that if the dominant frequency
component in the airbox pressure fluctuation is the multi-
The performance results from the 1D/0D model at close cylinder induction frequency, then the pressure-time
to 9000 rpm were compared to a results from a coupled history at each inlet during the inlet open period will be
1D/3D model at the same speed. The overall predicted similar, and the effects described above will not arise.
power output from each engine was similar; 30.77 kW
for the coupled model and 31.52 kW for the 1D/0D
model. The individual cylinder powers for both the 1D/0D
VALIDATION
model and the coupled model are given in Table 1.
To validate the simulations, an airbox with a cavity of
Table 1 Predicted individual cylinder powers dimensions 570 mm by 235 mm by 200 mm, and a neck
of dimensions 100 mm by 100 mm by 100 mm was
Cylinder Number 1D/0D Model 1D/3D Model
fabricated from 9 mm thick MDF. The engine used for
Power (kW) Power (kW)
the tests was the four-cylinder Yamaha R6. In order to
1 7.73 7.93
perform the tests on the R6 engine, new throttle bodies
2 8.02 7.77 were fabricated for the engine and the carburettors were
3 8.03 7.84 replaced with a multipoint fuel injection system. A DTA
4 7.74 7.23 (Salford, England) engine management system was
Maximum variation 3.7% 9.8% used to control the ignition and fuel injection.

The predicted difference in cylinder-to-cylinder power is The engine was tested on a 145 kW Schenk AC
more significant in the coupled model when compared to dynamometer. A strain gauge based Kyowa PGM–02KG
those of the 1D/0D model. The largest difference in transducer with a range of a range of 0-20 kPa was used
cylinder power was 9.8%. This was predicted between to measure the low amplitude pressure fluctuations
Cylinders 1 and 4. Cylinders 3 and 4 are adjacent, yet a within the airbox. A specially designed mount allowed
8.6% difference in power output is predicted. The

117
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

the transducer to record the pressure at a grid of points ANALYSIS PROCEDURE

within the airbox. The pressure in the intake pipe
between Cylinder 2 and the airbox was measured by a To compare the experimental and simulation data, it was
0-2 bar Kistler 4045A2 piezo-electric pressure necessary to determine the airbox natural frequency,
transducer. and isolate specific engine speeds in the ramp so that
measured and predicted pressure data could be
The transducer was placed at one of three different compared.
points in the airbox, depending on the tests. All points
were along the mid-plane of the box, at mid-height. If the airbox is tested over a range of engine speeds, the
Position 1 was 100 mm from the rear wall, Position 2 airbox will resonate at speeds where frequency
was the centre of the box and Position 3 was 100 mm components in the applied pressure waves correspond
from the front wall, i.e. the neck end, as shown in Figure to a resonant mode of the airbox. When the airbox is
11. Cycle-averaged speed was calculated based on a resonating, the amplitude of the pressure component in
combination of the output from the crank position sensor the airbox which is excited will be increased. However,
and a square wave output from the DTA engine the pressure in the airbox will contain components at the
management system. single-cylinder induction frequency and all its multiples.
In order to see which component was exciting the
The high-speed data acquisition system PTrAc by resonance, the airbox pressure was deconstructed into
Optimum Power Technology (Pittsburgh) was used to its individual frequency components. To explain why this
record the output from the two pressure transducers, is important, consider the following example. An airbox
and the outputs from the crank position sensor and DTA with a natural frequency of 100 Hz will be excited by the
engine management system. Very high sample rates pressure component at the single-cylinder induction
were required to accurately record the square wave frequency at 12000 rpm (100 Hz). However, the same
output from the engine management system. However, resonant mode of the airbox will also be excited at
in Matlab, the pressure data were resampled at 2 kHz to 6000 rpm by the first harmonic of the single-cylinder
nd
reduce the volume of data for analysis. The data were induction frequency (100 Hz) and at 4000 rpm by the 2
then filtered with a 1 kHz Butterworth filter. harmonic (100 Hz). Therefore, to understand the
resonant behaviour in the airbox, it was necessary to
examine the amplitude of each individual frequency
component of airbox pressure. Before this could be
done, some pre-processing of the data was required.

NORMALISING THE AIRBOX PRESSURE DATA

Resonance is an increase in the amplitude of an output,

1 2 3 relative to the amplitude of the input, or forcing function.
In the case of a firing engine operating at WOT over a
range of speeds, the amplitude of the input to the airbox
is unknown and will vary with speed. Ideally, the
amplitude of the pressure fluctuation in the airbox would
Figure 11 Transducer positions within the airbox be divided by the amplitude of the forcing function in the
inlet, to give a gain for the system. This gain would show
The airbox natural frequency dictates the engine speed peaks at the resonant frequencies. The forcing function
at which resonant effects occur, so it was necessary to was considered to be the outward moving pressure
perform the tests over a wide range of speeds to ensure waves that propagate from the engine to the airbox.
that no significant effects were missed. The range However, this amplitude could not be measured, as the
between 3000 and 10000 rpm was investigated. For all inlet pressure transducer gives the superposition
tests, the engine was run at wide open throttle. pressure for the outward moving and reflected pressure
waves. One approach would be to ignore variations in
Rather than perform tests at steady engine speeds over the amplitude of the forcing function, but this would not
a discrete series of steps, a series of sweeps were necessarily show the true resonance characteristics, as
performed. This enabled a large amount of data to be peaks may be due to increases in the amplitude of the
recorded in a relatively short time. In a sweep, the forcing function at particular engine speeds.
engine speed was increased at a fixed ramp rate, e.g.
100 rpm per second over a specified speed range. In The amplitude of the suction waves at each cycle is a
order to check that this was sufficient to record the data function of how much air is drawn into the engine, so it
accurately, steady state tests were recorded at several seemed likely that the delivery ratio could be used as a
engine speeds and compared to the ramp results at measure of the amplitude of the forcing function.
those speeds. The results for the sweep at a given Therefore an alternative approach based on normalising
engine speed were found to be the same as the steady the amplitude of the airbox pressure fluctuation
state results. according to delivery ratio was considered.

118
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Several coupled FLUENT/Virtual 4-Stroke models were

run over the complete range of engine speeds to check
Constant
if the delivery ratio was an appropriate means of scaling
engine speed
the forcing function. For each speed, the pressure in the
airbox and the delivery ratio were recorded. The
diameter of the intake valves was then adjusted, thereby
changing the delivery ratio and the coupled analyses
were repeated. Note that the delivery ratio was varied in
this way, rather than by changing the throttle setting, to
avoid the effects due to pressure wave reflections from
the throttle. The amplitude of the pressure fluctuation in
the airbox was compared at each engine speed. The
amplitudes were found to be approximately proportional
to the delivery ratio. Therefore, the pressure
measurements in the airbox were normalised by the
delivery ratio, prior to data analysis. Figure 12 Typical FFT plot for data window at a given
engine speed
ISOLATING HARMONICS
For each window of data, the values of A1 to A6 were
In the frequency sweep used to determine natural calculated for every engine speed between 3000 and
frequency, a pure sine wave was applied to airbox 10000 rpm. The amplitude of each frequency component
models; this was equivalent to applying a single was then plotted against engine speed to give the
frequency component. However, the forcing function response of the airbox for each individual frequency
applied by the engine to the airbox comprises a component in the forcing function.
fundamental frequency and a number of harmonics.
Each frequency component will excite a response in the SINGLE-CYLINDER TESTS
airbox, therefore it was necessary to isolate each
frequency component within the measured data and plot In the first series of tests, a single intake pipe was
its normalised amplitude over the 3000 to 10000 rpm connected between the engine and airbox. This was to
speed range. verify the airbox natural frequency predicted by the CFD
models. Although, it is not a realistic situation to have
A sliding window approach in Matlab was used to only one intake pipe connected to the airbox and the
analyze the frequency components in the airbox other three cylinders breathing from the atmosphere, this
pressure trace. A window of 200 data points was moved provided a simple means to identify the natural
through the data in steps of 10 data points so that there frequency of the airbox experimentally. When the airbox
was an overlap between the windows. This minimized was coupled to an engine, it was expected that
discontinuities in the analysis. The engine speed was resonance would occur at the engine speed where the
assumed to be constant over the course of each single-cylinder induction frequency corresponded to an
window; the value used was the average of the speed of airbox natural frequency. If only one pipe is connected,
st
the 1 point and the last point within the step. For each there will be no interference or cancellation from
window, the airbox pressure trace was scaled by the frequency components due to the other cylinders.
delivery ratio and then normalized to a zero mean. An Therefore, if resonance occurs, the amplitude of the
FFT was performed on the data, and the result for each pressure fluctuation in the airbox will be greater than for
data window was a plot similar to that shown in the 4-cylinder case, and the resonant frequency should
Figure 12. be more clearly defined.

The first peak in Figure 12 is at the single-cylinder Static pressure was recorded in the intake pipe to
induction frequency. The amplitude A1 of the peak is the Cylinder 2, and in the airbox at Position 1. The airbox
amplitude of the pressure component in the airbox at the was tested over the speed range 3000 to 10000 rpm.
single-cylinder induction frequency at that engine speed.
Similarly, the amplitude A2 of the second peak is the Frequency analysis of pressure at Position 1
amplitude of the pressure component in the airbox at the
first harmonic of the single-cylinder induction frequency Figure 13 shows the amplitude of the normalised
at that engine speed. The third peak is at the second pressure components in the airbox at the single-cylinder
harmonic of the single-cylinder induction frequency and induction frequency and the first harmonic of the single-
so on. The fourth peak at the third harmonic of the cylinder induction frequency versus engine speed.
single-cylinder induction frequency will be at a frequency
corresponding to the engine induction frequency for a 4-
cylinder engine.

119
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

1 1
frequency component at single-cylinder firing frequency Predicted
0.9
frequency component at first harmonic of single-cylinder firing frequency Measured

0.8 0.8

0.7

Pressure amplitude (kPa)

Pressure amplitude (kPa)

0.6 0.6

0.5
73 Hz
0.4
0.4

0.3

0.2
0.2
0.1

0
0 3000 4000 5000 6000 7000 8000 9000 10000
3000 4000 5000 6000 7000 8000 9000 10000
Speed (rpm)
Engine Speed (rpm)

Figure 13 Amplitude of components of airbox pressure Figure 14 Measured and predicted amplitude of the
at the single-cylinder induction frequency and its first pressure component in the airbox at the single-cylinder
harmonic, single pipe connected. induction frequency at Position 1

The amplitude of the airbox pressure component at the There is very good agreement between the shapes of
single-cylinder induction frequency is a maximum at a the two curves. In particular, the frequency for
speed in the region of 9000 rpm, which is equivalent to resonance is well predicted, and corresponds to the
75 Hz. There is a peak in the amplitude of the 1st value given by the sweep method used in the design of
harmonic of the single-cylinder induction frequency in the airbox. The engine speed at which resonance
the region of 4450 rpm; this corresponds to a occurred was approximately 8800 rpm, which
fundamental frequency of 74 Hz. Both of these values corresponds to a frequency of 73 Hz.
are approximate. The first natural frequency of this
airbox geometry was predicted by FLUENT to be 73.4 The characteristics of the airbox pressure amplitude at
Hz. Therefore there was a difference of less than 1% the 1st harmonic of the single-cylinder induction
between the measured and predicted natural frequency frequency are presented in Figure 15. It was expected
for the airbox. that the 1st harmonic would excite resonance in the
airbox at 73 Hz, i.e. when the engine speed was
VALIDATION OF 1D/3D MODEL approximately 4400 rpm. The coupled simulation results
show resonance close to this speed, although it should
Once the natural frequency of the airbox geometry had be noted the simulation data are sparse in this region, as
been verified using a single cylinder, the more realistic interest was mainly focused on the region for 8000 to
configuration with four intake pipes connected was 10000 rpm. A resonance at 4100 rpm is shown in the
tested. measured results. This corresponds to a frequency of
approximately 68 Hz. The amplitudes of this frequency
Coupled models of the 3D airbox were run for a range of component in the pressure fluctuation in the airbox at
speeds so that the predicted and measured frequency this resonance condition for both the measured and
response over the full speed range could be compared. predicted results are approximately half those for the
The speeds modelled were 3000, 3500, 4000, 4440, fundamental at 8800 rpm.
5000, 5500, 6000, 6500, 7000, 7500, 8000, 8200, 8400, 1
Predicted
8700, 8880, 9000, 9500, and 10000 rpm. A greater 0.9 Measured
number of speeds were considered in the 8000 to 9000 0.8
rpm region, as it was in this region that the single-
cylinder induction frequency was close to the predicted
Pressure amplitude (kPa)

0.7

airbox natural frequency and resonance was predicted 0.6

to occur in the airbox. 0.5

0.4
Frequency analysis of pressure at Position 1
0.3

The coupled simulations and the experimental data were 0.2

both analyzed in the same manner, i.e. before a 0.1

frequency analysis of the data was performed, the 0

pressure in the airbox was scaled by the delivery ratio at 3000 4000 5000 6000 7000 8000 9000 10000
Speed (rpm)
that speed. The experimental results were compared to
the simulation results for the single-cylinder induction Figure 15 Measured and predicted amplitude of the
frequency and its 1st to 3rd harmonics. Results for the pressure component in the airbox at the 1st harmonic of
single-cylinder induction frequency are shown in the single-cylinder induction frequency at Position 1
Figure 14.

120
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

The predicted results given in Figure 15 also show a 1

peak at approximately 7500 rpm, i.e. 125 Hz. At this Predicted

Measured
speed, there is only a small indication of an increase in 0.8
amplitude seen in the measured data.

Pressure amplitude (kPa)

The measured amplitudes of the components of airbox 0.6

pressure at the single-cylinder induction frequency and

4-cylinder engine induction frequency (3rd harmonic) are
0.4
shown in Figure 16. As also shown in Figure 14,
resonance at the fundamental frequency of the airbox
(73 Hz) has been excited by a component at the single- 0.2

cylinder induction frequency. In the region of this

resonance, the amplitude of the single-cylinder induction
0
frequency component in the airbox is greater than that of 3000 4000 5000 6000 7000 8000 9000 10000
the component at engine induction frequency. This Speed (rpm)

means that over the course of one complete (720˚)

engine cycle, the low frequency pressure fluctuation is Figure 17 Measured and predicted amplitude of the
more dominant than four individual pressure fluctuations components of airbox pressure at engine induction
due to each induction. As noted earlier, this low frequency (3rd harmonic) at Position 1
frequency pressure fluctuation in the airbox may result in
cylinder-to-cylinder power variations. Comparison of pressure traces at 8880 rpm

0.6
In this section, the measured pressure traces recorded
120 Hz
frequency component at single-cylinder firing frequency in the intake pipe to Cylinder 2 are compared to the
frequency component at engine induction frequency (3rd harmonic
of single-cylinder firing frequency)
pressure traces predicted by the coupled models at a
73 Hz single engine speed. Also, the measured and predicted
airbox pressure traces at Positions 1, 2 and 3 are
Pressure amplitude (kPa)

0.4
compared. The chosen speed was 8800 rpm, at which
the single-cylinder induction frequency matched the
predicted natural frequency of the airbox, and a low
frequency pressure wave was predicted in the airbox.
0.2

The measured and predicted static pressures in the 2nd

intake pipe at 8880 rpm are shown in Figure 18 for two
engine cycles. Good agreement is evident.
0
3000 4000 5000 6000 7000 8000 9000 10000
Speed (rpm) 1.4
Experimental
Figure 16 Measured amplitudes of the components of Simulated
1.3
airbox pressure at the single-cylinder induction
frequency and engine induction frequency (3rd harmonic) 1.2
at Position 1
Pressure (Bar)

1.1

Around 3600 rpm, there is an apparent resonance at

120 Hz, i.e. the engine induction frequency. This was not 1

expected, as no mode at this frequency was predicted

0.9
by the CFD frequency sweep of the airbox. However, the
comparison of these measured results with those 0.8
predicted by the full 1D/3D coupled model of the engine
0 =TDC Cyl 1
and airbox, as given in Figure 17, shows that this peak 0.7
was well predicted. 0 180 360 540 720 900 1080 1260 1440
Crank Angle (Degrees)

It is possible that the application of the sinusoidal Figure 18 Measured and predicted static pressure in the
pressure fluctuation at the neck of the airbox during the 2nd intake pipe at 8880 rpm for two engine cycles
sweep of frequencies may not have excited the mode at
120Hz. Therefore, further work is required to investigate The predicted and measured pressures at Positions 1, 2
if the sweep method will show a resonance at 120 Hz and 3 in the airbox are compared in Figures 19, 20 and
when the forcing function is applied at the inlet pipes. 21.

121
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

1.5 pressure-time histories in the vicinity of the inlets was

Experimental
different for each of the 4 cylinders during their
Simulated
1 respective inlet valve open periods.

0.5 At Position 3, there is poor agreement between the

Static Pressure (kPa)

measured and predicted pressure traces. Position 3 was

0
close to the neck boundary and the modelling of the
neck boundary may be a factor in the discrepancies
-0.5
between measured and simulated results. In the coupled
models, an atmospheric pressure boundary was
imposed at the plane of the neck opening. Thus,
-1
pressure waves propagated from the airbox volume
0 =TDC Cyl 1 toward the neck will be reflected from this plane. In
-1.5
0 180 360 540 720 900 1080 1260 1440 reality, the pressure waves propagate a small distance
Crank angle (degrees) beyond the neck, before reflection. To take account of
Figure 19 Position 1 - Measured and predicted pressure this end effect, the CFD model should be extended to
in the airbox at 8880 rpm for two engine cycles include a region outside the neck.

1.50 CONCLUSIONS
Experimental
Simulated
1.00 1. A CFD frequency sweeping technique has been
developed and successfully used to determine the
0.50 natural frequencies of complex volumes.
Static pressure (kPa)

0.00 2. This technique was used to design an airbox with a

specific natural frequency. The natural frequencies
-0.50
predicted by the CFD model were within 1% of
experimentally determined values.
-1.00
3. A 3D CFD model of the airbox was coupled to a
0 =TDC Cyl 1 one-dimensional engine model of a 4-cylinder
-1.50
0 180 360 540 720 900 1080 1260 1440 Yamaha R6 engine. The model showed that
Crank angle (degrees) resonance of the airbox could be excited by the
Figure 20 Position 2 - Measured and predicted pressure single-cylinder induction frequency, or its harmonics.
in the airbox at 8880 rpm for two engine cycles If the resonance was excited by the single-cylinder
induction frequency, or one of the lower harmonics,
1.5 then the pressure in the airbox oscillated at a
Experimental
Simulated
frequency lower than the overall multi-cylinder
1.0 induction frequency. Thus, the coupled simulations
explained the presence of the low frequency wave
0.5
predicted by Maynes et al [3].
Static Pressure (kPa)

0.0
4. The simulations showed that pressure oscillations at
-0.5 a frequency lower that the overall multi-cylinder
induction frequency resulted in significant cylinder-
-1.0 to-cylinder power variations.

-1.5
5. The predicted amplitude and phase of the pressure
0 =TDC Cyl 1 traces in the airbox correlated extremely well with
-2.0
0 180 360 540 720 900 1080 1260 1440 the experimental results for the Yamaha R6 engine.
Crank angle (degrees)

Figure 21 Position 3 - Measured and predicted pressure 6. There was good agreement between the measured
in the airbox at 8880 rpm for two engine cycles and predicted amplitude of the individual frequency
components of the pressure waves in the airbox.
For Positions 1 and 2, the shapes of the measured and
predicted airbox pressure traces are similar, and there is
excellent agreement for the frequency and phasing of
the low frequency pressure fluctuation. Both the
measured and predicted pressures for Positions 1 and 2
clearly show that the dominant pressure fluctuation has
a period of 720°, i.e. 2 crankshaft revolutions in which 4
induction events take place. This confirms that the

122
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

ACKNOWLEDGMENTS 6. SINCLAIR, R., SCHINDLER, P. and STRAUSS, T.,

2002. Validated 1D/3D Coupling Method to Solve
Funding to support this research work was provided by Transient Flow in Internal Combustion Engines.
the European Social Fund. OPTIMUM Power ASME -Publications- Pvp, 448 (Part 1), pp. 197-204.
Technology and FLUENT are gratefully acknowledged 7. HARRISON, M.F. and DUNKLEY, A., 2004. The
for provision of the software. The expert advice and acoustics of racing engine intake systems. Journal
assistance of Dr Brendan Maynes is greatly appreciated. of Sound and Vibration, 271(3-4), pp. 959-984.
The Virtual Engineering Centre (VEC) at Queen’s is 8. INGARD, U., 1953. On the theory and design of
gratefully acknowledged for provision of engine test and acoustic resonators. The journal of the acoustical
computing facilities. The assistance of final year student society of America, 25(6), pp. 1037-1061.
Russell McKee during engine testing is much 9. CHANAUD, R.C., 1994. Effects of geometry on the
appreciated. resonance frequency of Helmholtz resonators.
Journal of Sound and Vibration, 178(3), pp. 337-348.
REFERENCES 10. PANTON, R.L. and MILLER, J.M., 1975. Resonant
Frequencies of Cylindrical Helmholtz Resonators.
1. OPTIMUM POWER TECHNOLOGY, Virtual 4- 57(6 pt 2), pp. 1533-1535.
Stroke engine simulation software user's guide.
Bridgeville, PA, USA.
2. BLAIR, G. P., 1999, Design and Simulation of Four- ACRONYMS
Stroke Engines, SAE, Warrendale, PA, USA.
3. MAYNES, B.D.J., KEE, R.J., KENNY, R.G., FLECK, CFD Computational Fluid Dynamics
R., MACKEY, D.O. and FOLEY, L., 2002. Prediction CFL Courant Friedrich Lewy number
of Formula 1 Engine and Airbox Performance Using F1 Formula 1
Coupled Virtual 4-Stroke and CFD Simulations. SAE FFT Fast Fourier Transform
Transactions, 111(Part 3), pp. 2779-2795. IC Intake Valve Closing angle
4. DE VITA, A., DI ANGELO, L. and ANDREASSI, L., IO Intake Valve Opening angle
2003. ICES2003-0680 CFD Analysis of Engines: An MDF Medium density fibreboard
Advanced Approach Based on Codes Dynamically PISO Pressure Implicit with Splitting of Operators
Coupled, N.X. BLYTHE, ed. In: ASME; Internal UDF User Defined Function
Combustion Engine Division; Design, application, WOT Wide Open Throttle
performance and emissions of modern internal 0D Zero Dimensional
combustion engine systems and components, May 1D One Dimensional
2003, ASME pp585-594. 3D Three Dimensional
5. RIEGLER, U.G. and BARGENDE, M., 2002. 2002-
01-0901 Direct Coupled 1D/3D-CFD-Computation
(GT-Power/Star-CD) of the Flow in the Switch-Over CONTACTS
Intake System of an 8-Cylinder SI Engine with
External Exhaust Gas Recirculation, Modeling of SI Síle Brennan: [Link]@[Link]
engines and multi-dimensional engine modeling, Robert Kee: [Link]@[Link]
Mar 2002, Society of Automotive Engineers pp 205-
216.

123
Downloaded from SAE International by Purdue University, Monday, October 29, 2018
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

2004-01-3511

Horsepower Retention by ISF (Isotropic Superfinishing) of

Automotive Racing Components
William P. Nebiolo
REM Chemicals, Inc.

Copyright © 2004 SAE International

ABSTRACT

Automotive power and drivetrain in frictional heat generation and a loss of

components typically transfer the energy engine horsepower.
from fuel combustions through a series of
metal-to-metal mechanical components. This paper will examine the use of the ISF
Typical examples of metal-to-metal Process to improve the micro surface of
components include bearings, camshafts automotive metal-to-metal contact surfaces.
rotation and gear teeth in a transmission or Evidence will be presented demonstrating
differential. how the improved micro surface of ISF
processed parts retains horsepower,
The normal manufacturing steps for increases component durability and reduces
bearings and camshafts typically include operating temperature.
various machining operations that normally
end with a final grinding step. For INTRODUCTION
economical reasons, automotive
transmission and differential gears are Traditionally, automotive engine parts such
usually lapped to generate the final surface as camshafts, rocker arms and valve springs,
finish. A careful examination of the as drivetrain parts such as transmission gears,
manufactured surface of these components shifting linkages and splines, differential
reveals a polishing pattern associated with parts such as rings and pinions, and
the final grinding or lapping step used to universally mounted components such as
finish the component. Typically the pattern bearings operate by rolling, sliding or
is visible as parallel rows of micro-asperity pushing against their complementary
peaks and valleys. partners.

In operation these mechanical components Several OEM manufacturing steps are

work by engaging their complementary required to produce these components.
partners. As such, bearings roll in races, These include, but are not limited to various
gear teeth roll and slide against other gear combinations of forging, blanking, cold
teeth and camshafts engage cam followers. heading, and machining, hobbing and
The presence of the micro-asperity peaks hardening operations such as thru-
and valleys on the surface of these hardening, induction hardening, case
components reduces the actual available hardening, nitriding, etc.7, 8
area of metal-to-metal contact. This results

125
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

OEM FINISHING been ground to a final surface finish of Ra = 35

µin./0.9 µm. Note the parallel rows of surface
asperities corresponding to the directional pattern of
OEM manufacturers will typically employ a the final grinding or machining operation.7, 8, 9, 12, and 13
final grinding operation to remove heat
treatment warping, generate the final THE ISF SURFACE
dimensions/geometry and achieve the
ultimate surface finish. Automotive gears It has been reported1, 2, 3, 7, 8, 9, 12, 13, 14 that metal-
are typically lapped for economy reasons. to-metal contact areas, if refined by
See Photo 1.

Photo 3; Shows a 3-dimensional, magnified view of

an isotropic surface finish Ra = 1.5 Pin./0.04 µm. The
surface has been produced by chemically accelerated
Photo 1; Shows a beveled spur gear from an OEM vibratory finishing. Note the elimination of the
that is ready for installation. Note, grind lines are unidirectional surface asperities, grinding/machining
clearly visible on the flanks of the gear teeth.7, 8 lines, when compared to Photo #2.7, 8, 9, 12, 13

An examination of the surface of any of Chemically accelerated vibratory finishing1,

these parts will reveal a unidirectional, 5, 6, 7, 8, 10, 12, 13, 14
have a resultant Isotropic
microscopic, surface pattern consisting of Superfinish. An isotropic surface is a
parallel rows of asperity peaks and valleys. surface that has a random, non-directional
The directional pattern of the parallel rows surface finish.7, 8, 13 See Photo 3.
corresponds to the direction of the final
grinding or machining operation used to METAL-TO-METAL CONTACT
finish the component.7, 8, 12, 13 See Photo 2. TRADITIONAL SURFACES VS. ISF
SURFACES

When placed into operation for the first

time, camshaft lobes, gear teeth, bearings,
rear end rings and pinions, etc. generate a
contact, or wear pattern on their
complementary partners.2, 8, 9, 10, 12, 13, 14

If the components are traditionally finished,

the percentage of actual metal-to-metal
contact between the two complementary
partners is minimal because initially contact
is only peak-to-peak.8, 14 This contact
Photo 2; Shows a 3-dimensional, magnified view of a
section of the surface of a gear tooth. The tooth had pattern concentrates all the surface contact

126 2
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

stress associated with the two parts

functioning against one another into a few
isolated locations.8, 14

On the other hand, parts that have been

finished with the ISF process have an
improved metal-to-metal contact pattern.8, 14
Since the asperity peaks have been removed
from the complementary components the
resultant surface is smoother. Contact stress
in any one location is reduced, 8, 14 because
the smoother finish spreads the stress across
a wide surface area. Photo 5; Shows a traditional vibratory tub. Ideal for
processing long parts such as splines, camshafts,
crankshafts, drive shafts and pinions.
GENERATING THE ISF FINISH
During the initial refinement step, a
According to the techniques described
chemistry is added to the vibratory
elsewhere, 1, 5, 6, 7, 8, 12, 13 asperity refinement
machine.1, 5, 7, 8, 12 Independent testing has
occurs in a chemically accelerated vibratory
shown that this chemistry will not generate
finishing process. The camshafts, valve
hydrogen embrittlement within the steel
springs, gears, bearings, rings and pinions,
parts.4 The chemistry reacts with the surface
etc. to be refined are placed into a vibratory
of the part to produce a soft, conversion
machine containing a high-density, non-
coating.1, 5, 7, 8, 12 As the part rolls in the
abrasive, ceramic media.6 See Photos 4 and
vibratory machine, the conversion coating is
5. The chemically accelerated vibratory
wiped from the peaks by the weight of the
finishing process consists of two steps that
non-abrasive, high density ceramic media
are conducted sequentially within the same
exposing unreacted, underlying metal.
vibratory machine.
The peaks, being elevated, are wiped
preferentially.7, 8, 12, 13 Since the wiped
coating was formed by reacting with the
steel’s surface, coating removal results in a
lowering of the peak.
The recessed valleys between the peaks are
untouched by the media, leaving the coating
intact and the valleys untouched.7, 8, 12, 13 The
coating reforms on lowered peak to
propagate final peak removal.7, 8, 12, 13 In
rapid order, the peaks are wiped thereby
generating the improved micro-finish.7, 8, 12, 13
Photo 4; Shows a traditional vibratory bowl. Ideal
for processing short stubby parts such as gears,
rocker arms, valve springs and bearings.

3 127
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

Round Bowl
Finishing
Machines (6)
Rinse/Dry Unit

Central Electrical enclosure

provided with system
automation

Air operated directional chute Inclined media

return conveyor

Rinse/Dry Unit Drip Pan

Photo 6; Shows a close-up of the teeth of the beveled

spur gear shown in Photo 1. The gear was processed
to achieve an isotropic finish by using the chemically
accelerated vibratory finishing technique. The final Schematic #1; Shows a 6 bowl, high volume ISF
surface finish was Ra = 1.5 Pin./0.04µm. Note when vibratory operation for an OEM manufacturer of
compared to Photo 1, that all the grind line asperity planetary pinions. The operation is capable of
peaks have been removed from the tooth flanks. generating 3,000 ISF finished pinions per hour or in
JIT terms 1 pinion every 1.2 seconds.
Once the asperities are removed and the
improved micro-finish has been achieved, a HORSEPOWER LOSS AS
mildly alkaline, soap-like burnish chemistry DEMONSTRATED BY FRICTIONAL
is added to the vibratory machine.7, 8, 12, 13 HEAT GENERATION
The burnish neutralizes and removes any
residual coating remaining on the parts from As noted earlier, engineered metal-to-metal
the refinement step and yields a mirror-like, contact surfaces such as camshafts, valve
final finish.7, 8, 12, 13 See Photo 6. springs, transmission or quick change gears,
bearings and rear end rings and pinions
In high volume, production operations, cost transfer energy by mechanical contact. This
savings are realized through the ability to contact can be a rolling, sliding or pushing
process parts in mass, thereby, achieving the force against a complementary component.
superior ISF finish at per piece costs lower The asperity peaks on these surfaces
than conventional grinding, honing and introduce friction (i.e. inefficiency) into the
lapping operations. See Schematic #1. mechanical transfer of engine horsepower
resulting in horsepower loss most noticeable
as heat.3, 7, 8, 11, 12, 14

In a recent evaluation, a major bearing

manufacturer9 monitored the performance
features of two functional metal-to-metal
contact surfaces by means of an ASTM
D2714 and D2792 Block-on Ring test. See
Schematic 2.

128 4
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

the break-in cycle on the ground test

specimens.3, 7, 8, 11, 12, 14 Once the metal-to-
metal contact surfaces of the ground
specimens were broken-in; i.e. asperities
removed, the monitored temperature
dropped to a steady state operational
temperature of 145°F/62°C. 3, 7, 8, 11, 12, 14 See
Graph 1.

Schematic 2; a simplified schematic of the Block-on-

Ring apparatus.

In the experiment, two steel blocks and rings

were evaluated sequentially using the same
test rig. During each evaluation, the sump Graph 1; Superimposes the temperature curve for
temperature of the lubricant system was normally ground, control standard specimens with the
monitored as the ring was rotated at a temperature curve for the ISF specimens. Note the
constant 800 rpm. 9 The block was held on temperature spike for the ground specimens
corresponding to asperity peak break-in and the
the rotating ring at a constant 8 lb (3.64 Kg)
absence of a spike for the ISF specimens3, 7, 8, 11, 12, 14
force. also, note the lower, final operating temperature of
the ISF specimens.
In the first evaluation (the control standard
test) the block and ring specimens were When the testing was repeated using the ISF
finished using a grinding process to generate finished specimens, it was noted
a 25µin/0.64µm surface finish. In the immediately that there was a complete
second evaluation the block and ring were absence of the temperature spike, (i.e. no
ISF finished to remove surface asperities asperity peaks present therefore no break-in
and generate the improved ISF finish. (See period required.) The ISF finished
Table #1) specimens showed a gradual rise to a steady-
state temperature, 132°F/54°C. 3, 7, 8, 11, 12, 14
Sample I.D. Test Surface Ra
Normal Ground Block 25 µin / 0.64 µm When compared to the normally ground
Normal Ground Ring 29 µin / 0.75 µm specimens this represents a 13°F/8°C
ISF Finished Block 2.3 µin / 0.06 µm difference in final operating temperature. It
ISF Finished Ring 5.1 µin / 0.13 µm can be inferred from this experiment that the
Table #1; Block-on-Ring Test, specimen surface reduction in operational temperature at the
conditions. metal-to-metal contact interface indicates a
reduction in metal-to-metal friction thereby
Concomitant to the mechanical abrading of allowing this contact interface to retain
the asperities on the ground specimens was horsepower that would ordinarily be lost to
a higher, monitored bearing temperature.3, 7, 8, frictional heat generation.
11, 12, 14
Additionally, there was a definitive
temperature spike 165°F/74°C, associated
with the completion of the loading phase of
5 129
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

LOADING EFFICIENCY
IMPROVEMENT OF BEARING
SYSTEMS BY THE ISF SUPERFINISH

A major manufacturer of industrial roller

bearings recently performed an evaluation of
its bearing sets under various loading
levels.11 The manufacturers’ bearing sets
consist of roller bearings and an inside and
outside bearing race. The evaluation was
performed to determine the extent of the
possible benefit achieved by ISF finishing
when applied to complete or partial bearing
sets.
Table #2; Shows the combined results for all three
The company chose as its test specimens, a evaluations. 11
common bearing set from its traditional line
of roller bearings. The testing was An examination of each loading level shows
performed using three sets of roller bearings that the standard ground, roller/races,
that were sequentially mounted in a pillow bearing set consistently had the highest
block. operational temperature.

Test engineers varied the applied load on the Bearing sets in which only the rollers were
bearing sets in the pillow block. Beginning ISF finished demonstrated cooler operating
at 455 kg applied load, loading was temperatures at every loading level. At the
increased to a maximum of 4,550 kg applied lower load levels of 455 – 1,365 kg, the
load. The applied load was adjusted upward temperature differential was modest.
in 455 kg increments. However as the loading was increased from
the 1,820 kg level and up, the temperature
In each evaluation the bearing sets were differential between Evaluation #1 samples
turned at a constant 2,400 rpm and only the and Evaluation #2 was significant.
applied load was varied. The pillow block Additionally, that differential was
was equipped with a thermocouple that maintained throughout the balance of the
allowed test engineers to monitor frictional loading variations.
heat generation at the roller/race interface.
At all load levels and most especially at the
Evaluation #1 served as a control standard higher end loading level of 4,550 kg.,
and consisted of the company’s normally Evaluation #3 bearing sets demonstrated the
ground roller bearings coupled with most significant temperature benefit.
normally ground inner and outer races.
The results of this experiment clearly
In Evaluation #2 the rollers were ISF suggest that the ISF finish is of significant
finished prior to being coupled with the benefit even if applied to select components
traditionally ground inner and outer races. in a metal-to-metal, friction inducing
Finally, in Evaluation #3, both the roller application. Additionally, the ISF finish can
bearings and the races were ISF finished
prior to testing. See Table #2.

130 6
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

be of maximum benefit when applied to and the third dynamometer test was
both, complementary metal-to-metal contact conducted.12 See graph #1.
surfaces.

DYNAMOMETER TESTING
DEMONSTRATES HORSEPOWER
RETENTION

As an actual performance evaluation to

determine if the experimental benefits of the
ISF finish were legitimate, a series of
dynamometer tests12 were conducted using
three Nextel Cup, T101 transmissions.

The testing was conducted sequentially on

the same dynamometer stand. The Graph #2; Superimposes the resultant horsepower
transmissions were run using the same curves of 3rd gear, at specified rpm levels, for the
engine12 and in each evaluation a tested T101 transmissions.12 Note an approximate
horsepower curve for 3rd gear was plotted. 1% horsepower retention with the ISF transmission.
During each test, the engine with the test
transmission attached was preheated by Examination of the graph between the 6,400
idling until the transmission oil (Unocal and 6,600 rpm levels shows output levels of
90W) reached a uniform 82°C.12 Drive shaft 383, 385 and 388 hp respectively for the as
angles were monitored to ensure a consistent OEM supplied, coated and ISF gears.
angle was maintained for all evaluations.12
Load was then applied and horsepower The 5 hp difference between the as OEM
output was monitored. supplied and ISF gears represents a 1.3%
retention of horsepower that would have
Dynamometer evaluation #1 served as the been lost had the gears not been ISF
baseline control standard. A T101 finished. Likewise, when compared to the
transmission was run as received from the coated gears there was nearly a 1%
OEM. The transmission was not horsepower retention benefit due to ISF
disassembled, nor was it improved upon. finishing. Since ISF is an actual refinement
to the base metal there is no risk of coating
In evaluation #2, a second T101 loss via wear through, flaking or peeling.
transmission was disassembled and the
individual components were removed. The Component Durability Increase; Virtual
components were then coated with a Elimination of Contact Fatigue
commercially available product valued as a
friction reducing, performance enhancer. It is a reasonable expectation that ISF will
The transmission was then reassembled, retain horsepower by reducing friction and
attached to the engine and the second will also increase a component’s long term
dynamometer test was conducted.12 durability by eliminating contact fatigue,
asperity stress risers and metallic debris in
In Evaluation #3, the transmission was the lube system.
completely dissembled. All components
were then ISF finished. The transmission As an example consider the operational
was then reassembled, attached to the engine mode of automotive transmission and rear
end gearing. Gears by their nature undergo

7 131
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

two primary modes of motion, rolling and

sliding. Rolling motion is, as one would
expect, the rolling of one gear against its
complementary partner.

Sliding motion occurs when the flank of a

gear tooth slides against the opposing flank
of another gear tooth. Sliding occurs both as
gear teeth engage and then again, as they
disengage. Since this sliding action is a
metal-to-metal contact phenomenon, it is
here that ISF will benefit the gear by
eliminating contact fatigue problems.

Recently, the Gear Research Institute (GRI) Photo #7; Shows a 1” diameter specimen pin on the
lower left and a 5” loading roller on the upper right
used a Rolling/Sliding Contact Fatigue side of the photograph.
(R/SCF) Test Rig12, 13 to evaluate this very
hypothesis. See Schematic #3. The loading roller and the specimen pin
have different diameters and the rpm level
of the motor shaft on the R/SCF Test Rig
can be adjusted by the rig’s operator. These
differences result in a phenomenon where
the specimen pin becomes the sacrificial,
friction-affected, sliding wear component.

During the GRI tests, the R/SCF test

operator was able to produce a 43% negative
sliding ratio on the specimen,12, 13 whereas, a
mid-20% range is typical of normally
paired gears. Adjusting the sliding ratio to
an artificially high level facilitated having a
failure mode on the 1” specimen pin in a
minimal period of test time and at a minimal
Schematic #3; Schematic of the R/SCF test rig used
by GRI to conduct its evaluations.12, 13
test rig run cost.12, 13

The GRI, R/SCF test rig utilizes a crowned, Once the rpm level of the R/SCF test rig was
5” diameter loading roller and a 1” diameter optimized, it was maintained as a constant
specimen pin.12, 13 See Photo #7. throughout the evaluations. To facilitate
specimen pin fatigue, the R/SCF test rig
During R/SCF operation, the 1” specimen operator varied the stress load with which
pin is secured to a rotating motor shaft. The the 5” diameter contact roller was rolled
shaft is in turn connected by two gears to the against the 1” specimen pin.
5” diameter loading roller. The two
components roll against one another during Normal stress loading for the R/SCF test rig
rig operation. is 400 ksi. However to facilitate accelerated
testing the stress loading can be varied

132 8
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

upward. In this evaluation, the stress loads Sample Stress in Millions Failure
were increased upward in 25 ksi increments I.D. ksi of cycles Mode
to 425 and 450 ksi respectively.
Gd/hd #1 400 3.6 Pitted
Gd/hd #2 400 4.2 Pitted
Two types of sample sets were evaluated
during the R/SCF tests. The first Gd/hd #3 400 3.5 Pitted
evaluations were performed on 3 sets of ISF #1 400 20.0 None
rollers and pins. These rollers and pins were ISF #1 425 20.0 None
the baseline, control standard and were ISF #1 450 22.4 None
prepared to have a traditional ground/honed #1 Sum 400-450 62.4 None
surface finish typical of the standard surface ISF #2 400 5.0* None
condition of gear tooth flanks as received ISF#2 425 5.0* None
from an original OEM manufacturer. See ISF #2 450 20.0 None
Table #3. #2 Sum 400-450 30.0 None
Table #2; Tabulated results of R/SCF Testing 12, 13
Where Gd/hd = ground/honed
Sample I.D. Test Surface Ra
Ground/honed set 16 µin./0.4 µm * Note, after the extensive time it took to complete
ISF Finished Set 1.5 µin./0.04 µm the full run outs at all three contact stress levels of
Table #3; Surface conditions of the R/SCF test ISF specimen set #1, it was decided that when testing
specimens. ISF specimen set #2, the cycle count would be
stopped after 5 million cycles at the lower contact
stress levels of 400 and 425 ksi respectively to
The second set of evaluations was
advance to the next highest stress loading level and to
performed on 2 sets of rollers and pins. save time and reduce cost. The maximum 450 ksi
These rollers and pins were ISF processed to stress loading level however would be run to full run
generate the improved isotropic finish using out.12, 13
the chemically accelerated vibratory
finishing technique described earlier. See In the first set of evaluations, the GRI
Table #3. R/SCF test rig operator was able to generate
the typical failure mode of pitting, caused by
Prior to beginning the evaluation GRI the significant, applied sliding ratio on the
defined the test’s limitations and specimen control standard specimens. All three
performance to determine test completion. baseline standard samples failed in less than
Success was determined to be cycle run out 5 million cycles at the lowest contact stress
where cycle run out was defined as 20 loading level of 400 ksi. In each case, the
million specimen pin rotations without a failure mode was by pitting.12, 13
failure. Likewise, failure mode was defined
as specimen pin pitting. Testing results of the ISF specimens were
exemplary. In fact the ISF specimens could
To facilitate testing the R/SCF test rig was not be made to fail, even with the artificially
equipped with a cycle counter to monitor the high sliding ratio and the 450 ksi contact
rotations. Additionally the R/SCF was stress load.
equipped with a vibration monitor that
would record the onset of pitting by the It should be noted that the initial ISF roller
inception of specimen pin vibration and and pin set ran for a total accumulated cycle
automatically shutdown the test rig to count of 62.4 million cycles. This included
preserve the cycle count. 20 million cycles at 400 ksi, 20 million
cycles at 425 ksi and 22.4 million cycles at
450 ksi before the test rig was shut down.12, 13

9 133
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

being finished they leave a consequential

unidirectional pattern on the surface of the
part. This pattern can easily be seen as
parallel rows of asperity peaks and valleys.

Ground Pinion

Photo #8; R/SCF test, pin specimen results

Top – Ground/honed specimen #3 with pitting failure
mode after 3.5 million cycles at 400 ksi load level.
Bottom – ISF Processed specimen #1 after 62.4
million cycles. No failure, no pitting.
Photo #9; Shows pinion with traditionally ground
Testing was repeated with ISF specimen set gear teeth. Note, rows of asperity peaks.
#2 to determine if the favorable results were
repeatable. The test was abbreviated to 5 The presence of asperity peaks results in:
million cycles at the 400 and 425 ksi stress x poor metal-to-metal contact 7, 8, 12, 13
loading levels so as to reach the maximum x increased friction, promoted initially
loading pressure of 450 ksi quicker.12, 13 Five by peak-to-peak contact 8, 14
million cycles were chosen for the x increased surface contact stresses
abbreviated, lower ksi loading, run out limit isolated in the peak-to-peak contact
because it was itself, a longer cycle life than locations 8, 14
exhibited by any of the 3 sets of failed x Heat generation due to friction 11
control standard, ground/honed specimens. x Lose of engine horsepower transfer
through these inefficiencies 12
CONCLUSION x Shortening of component life due to
contact fatigue wear and pitting 12, 13
Grinding, honing and lapping are traditional
x Metal debris in the lubricant
surface finishing methods applied to
recirculation system 12, 13
automotive components such as camshafts,
bearings, transmission gears and rear end
Chemically accelerated vibratory finishing
rings and pinions.
offers an efficient way to mass produce an
ISF finish on metal-to-metal contact
Since these finishing techniques are
automotive components.1, 5, 6, 7, 8, 12, 13 The ISF
mechanically applied to the component

134 10
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

finish is non-directional and asperity free. 7, 8, x Increase in component durability via

12, 13
a reduction on contact fatigue failure
12, 13

x Elimination of metallic debris in the

ISF finished Pinion lubricant recirculation system 12, 13

REFERENCES

1. J. Arvin, Arrow Gear, Manesh, A., Power

Transfer System Manufacturing, Michaud,
M., Sroka G., Winkelmann, L., REM
Chemicals Inc, Proc. AGMA Technical
Conference, 2002 “The Effect of Chemically
Accelerated Vibratory Finishing on Gear
Metrology”

2. R.D. Britton, Elcoate C.D., Alanou M.P.,

Evans H.P. & Snidle R.W., Proc.
STLE/ASME Tribology Conference;
Orlando, FL (1999) “Effect of
Surface Finish on Gear Tooth Friction”

3. F.F.H. Hashimoto, Zhou R.S. & Howlett

Photo #10; Pinion as seen in photo #9 after being K.; U.S. Patent 5,503,481 (1996) “Bearing
chemically accelerated vibratory finished generating Surfaces with Isotropic Finish”
the ISF finish
4. R.D. Kane & Maldonado J., InterCorr
The ISF finished part contains no contact International, Inc., Report L99825GK
fatigue initiation sites because the asperity (1999)
peaks have been removed by the wiping
action of the media during the chemically 5. M. Michaud, Tirendi F., Zobbi R.G.;
accelerated finishing process.8, 14
U.S. Patent 4,491,500 (1985) “Method for
Additionally, metallic debris is not present Refinement of Metal Surfaces”
in the lube system since the asperities were
removed prior to component installation. 8, 14 6. M. Michaud, U.S. Patent 4,818,333
(1989) “Metal Surface Refinement Using
An isotropic surface finish offers several Dense Alumina-Based Media”
advantages to the operation of the metal-to-
metal contact surfaces of automotive parts: 7. W.P. Nebiolo, Proc. of the 37th AESF
Aerospace/Airline Plating & Metal
x Efficient metal-to-metal contact 7, 8, 12, Finishing Forum; Portland, OR (2001)
13
“Isotropic Finishing of Helicopter and
x Diffusion of contact stresses by the Turboprop Gearbox Components.”
elimination of asperity peaks thereby
spreading the stresses over a larger 8. W.P. Nebiolo, Proc. of the SAE/PRI
surface area 8, 14 Conference, Indianapolis, IN; 2002 “The
x Elimination of parasitic, frictional, Reduction of Parasitic Friction in
heat generation 11 Automotive Gearbox and Drive Train
x Retention of engine horsepower 8 Components by the Isotropic Superfinish”

11 135
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

9. H.P. Nixon, Cogdell J.D., Proc. of the

SAE International; 1998 Earthmoving
Industry Conference and Exposition Peoria,
IL (1998) “Performance Evaluation of a
Potential New Engineered Surface for
Enhanced Concentrated Tribological
Contacts”

10. R.W. Snidle, Evans H.P., Alanou M.P.,

“The Effect of Superfinishing on Gear Tooth
Profile” Report 2284; contract N68171-96-
C-9043 (1997)

11. Technical Bulletin, “Heat Generation

Test Results for REM Processed Bearings
versus Standard Spherical Roller Bearings,”
REM Chemicals, Inc., Southington, CT

12. L. Winkelmann, Michaud M., Sroka,

G., Swiglo, A.A., Mahan D.; American Gear
Manufacturers Association Meeting, Detroit,
MI, (2001) “Chemically Accelerated
Vibratory Finishing for the Virtual
Elimination of Wear and Pitting for Steel
Gears.”

13. L. Winkelmann, Michaud M., Sroka,

G., Swiglo A.A., SAE International Off-
Highway Congress, Las Vegas, NV (2002)
“Impact of Isotropic Superfinishing on
Contact and Bending Fatigue of Carburized
Steel”

14. Zhou R.S.; Hashimoto F.; Proc.

ATLE/ASME Tribology Conference; Maui,
HI (1994) “A New Rolling Contract Surface
and “No-Run-In”

136 12
 Downloaded from SAE International by Purdue University, Monday, October 29, 2018

About the Editor

Doug Fehan is Corvette Racing program manager, responsible for
coordinating engineering and marketing for Chevrolet’s championship-
winning team in the American Le Mans Series. Under his leadership,
Corvette Racing has become the world’s premier sports car team.

Fehan was introduced to racing as a youngster in Buffalo, NY. After

attending Eastern Michigan University, he raced his own stock car and
managed a major motorsports sponsorship program before joining
GM, where he was instrumental in securing SCCA Trans-Am, IMSA World
Sports Car, ALMS, and Le Mans championships.

137
Downloaded from SAE International by Purdue University, Monday, October 29, 2018
Downloaded from SAE International by Purdue University, Monday, October 29, 2018