Dynamics 34

8/10/2019 Dynamics 34

1/41

SSUE 34 SIMULATING SYSTEMS

TOUR DE FRANCE

Winning with Cicli Pinarello

ROTOR HUB DESIGN

WithSikorsky

FEATURES

AEROELASTICITY

Analysis with STAR-CCM+

THE HOLE GANG

F1 Slots Simulation

8/10/2019 Dynamics 34

2/41

8/10/2019 Dynamics 34

3/41

RECYCLED PAPER. VEGETABLE INKS.

Follow us online:

Global Offices CD-adapco

Contents

INTRODUCTION

03 Simulating Systems

Introduction by David L. Vaughn

05 Breaking News

STAR-CCM+ v8 High Performance Computing Steve

Multi-year Agreement with AREVA Calendar Competition Winner

AEROSPACE

09 Sikorsky Aircraft Corporation

Drag Prediction of Production Rotor Hub Geometries using CFD

13 Computational Aeroelasticity

A Key Enabling Technology for the Design of Next-Generation Aircraft

SPORTS

19 Cicli Pinarello

Winning the Tour de France 2012

MOTORSPORT

23 The Hole Gang

Simulating F1 Slots

27 Onda Solare

Taking on the World Solar Challenge

BUILDING SERVICES

30 M/E Engineering

Ventilating Cornell Universitys Biotechnology Laboratory

MANUFACTURING

33 Tetra Pak Cheese and Powder Systems

A Virtual Plant with STAR-CCM+

GROUND TRANSPORTATION

35 Hyundai Motor Company

Combining Automation with Scalability

37 Embry-Riddle University

Plugging into the Chevy Malibus Eco Future

Australia

CD-adapco Australia

[email protected]

New Zealand

Matrix Applied Computing Ltd.

[email protected]

Turkey

A-Ztech Ltd.

[email protected]

Israel

ADCOM Consulting Services

(Shmulik Keidar Ltd.)

[email protected]

South Africa

Aerotherm Computational Dynami

[email protected]

Russia

SAROV Engineering Center

[email protected]

EDITORIAL

Dynamics welcomes editorial from all users of CD-adapco software or services.

To submit an article, email: [email protected]

Telephone: +44 (0)20 7471 6200

Editor Deborah Eppel - [email protected]

Associate Editors Prashanth Shankara - [email protected]

Sabine Goodwin - [email protected]

Titus Sgro - [email protected]

Design & Art Direction Phillip Couzens, Stevie Miles Brewu & Ian Young

Press Contact Lauren Gautier - [email protected]

Advertising Sales Geri Jackman - [email protected]

US Events Bryant Aliaga - [email protected]

European Events Sandra Maureder - [email protected]

SUBSCRIPTIONS & DIGITAL EDITIONS

Dynamics is published approximately twice a year, and distributed internationally.

All recent editions of Dynamics, Special Reports & Digital Reports are now available online:

www.cd-adapco.com/downloads/dynamics

We also produce our monthly e-dynamics newsletter which is available on subscription.

To subscribe or unsubscribe to Dynamics and e-dynamics, please email [email protected]

05 09

13 19

27

30 33

35 37

23

RESELLERS

Corporate Headquarters

CD-adapco

60 Broadhollow Road

Melville, NY 11747

USA

Tel:+1 631 549 2300

Americas

Austin Cincinnati Detroit Houston Tulsa

New Hampshire Los Angeles Orlando

Sao Paulo Seattle State College

Europe

Glasgow London Lyon Nuremberg Paris

Rome Toulouse Turin Vienna

Asia-Pacific

Bangalore Osaka Seoul Shanghai

Singapore Yokohama

8/10/2019 Dynamics 34

4/41dynamics I S S UE 3 43

Simulating SystemsIntroduction by David L. Vaughn

CD-adapco has been in the CAE business for 33 years now, and its interest-

ing from that perspective to watch CAE fashions come and go. If you read or

listen to software marketing these days, then it would seem that multiphys-

ics simulation is all the rage. Similarly the word multidisciplinary is mak-

ing a big comeback as a buzzword in the CAE business.

In fact, the coupling of various physics models has been at the core of

CD-adapcos success for a third of a century. We were among the pioneers

of co-simulation, inventing methods to couple numerically difcult analyses

with each other. Whether it was uid ow, heat transfer, particle ow, sprays,

chemical reactions, or even structural stress and deections, CD-adapco

provided the solution with a full menu of physics models and a range of co-

simulation options with our CAE partners.

It is likely that the single most important factor in our success over the

years is the way that we actually partner with our customers. We work side

by side with them to solve their most complicated engineering problems.

And we listen carefully to what they tell us in order to improve our product

offerings and help them meet the challenges they face.

We talk with our customers a lot about their grand challenges.

What is it that keeps them up at night thinking; if I could just then that

would be a real game changer in my business. Regardless of which industry

they come from, their answers are very often quite similar. The overwhelm-

ing majority of these discussions relate to evaluating the performance of an

entire system.

Here is an example: A jet engine manufacturer might talk about being

able to evaluate the performance of an engine or the performance of cer-tain components during engine operation. Of course this could be done with

ight testing, but that would require manufacturing a prototype engine in

addition to everything else that goes along with testing, e.g. instrumentation,

test planning, etc. The cost and time scales of ight testing do not permit it

to affect engine design; rather, it is usually the last phase of engine devel-

opment that merely validates the safety and performance of the near-nal

engine design.

But now imagine that those engineers have access to a full set of ight

test-like data for every design candidate, and that this data could be had

for a small fraction of the time and cost of a ight test program. This would

transform the way business is done in that industry and have a signicant

impact on the bottom line.

The obvious solution is simulation. Certainly, CAE analysis is very preva-

lent in traditional engine design, but the grand challenge is putting every-

thing together, simulating the whole system with all its moving parts and

complicated physics.

Indeed this would require CAE software that incorporates multiple phys-

ics models and provides data to multiple disciplines, but it is a matter of

perspective. You see, the engineering management doesnt say, we need

to simulate multiple physics, nor do they say, we need to analyze several

disciplines. The goal is rather to have a virtual ight test program. And in

terms of simulation, this equates to a simulation of the full engine system.

You might think of it this way: its seeing the big picture using simula-

tion. This has been the direction of CD-adapco for a while now. There are

many examples in our current product offerings and in our development

plans. A few capabilities worth mentioning are: electro-chemistry (for bat-

tery simulation), electro-magnication, aero-vibro-acoustics, icing, casting,

chemistry, combustion, discrete element modeling and many more. Each

of these capabilities and features are implemented with the vision of simu-

lating full systems rather than merely coupling physics models to analyze

individual components.

An important point in this concept is using the appropriate del-

ity for each model within the simulation system. Its like the old saying,

when you go through life as a hammer, everything looks like a nail. In terms

of this discussion, if you approach it from an FEA perspective, for exam-

ple, you might tend toward modeling the entire system using FEA methods

(I could have just as easily said CFD, or 1-D, etc.). But the efcient simulation

of a complex system must incorporate models at different levels of delity inaddition to multiple physics/disciplines.

Words can be tricky; their meaning often depends entirely on ones per-

spective. But whether you call it multidisciplinary, multiphysics, simulation

of systems, tightly coupled, loosely coupled, or co-simulation, CD-adapco will

continue to focus on what we do best: technology innovation together with

customer partnerships to solve real world engineering problems through

simulation.

David L. Vaughn

VP of Marketing, CD-adapco

..::INTRODUCTIONDavid L. Vaughn

Whether you call it multidisciplinary,multiphysics, simulation of systems, tightly

coupled, loosely coupled, or co-simulation,CD-adapco will continue to focus on what wedo best: technology innovation together with

customer partnerships to solve real worldengineering problems through simulation.

8/10/2019 Dynamics 34

5/41

..::FEATURE ARTICLERotorcraft

8/10/2019 Dynamics 34

6/41

The STAR-CCM+ v8 releases are specifically

aimed at increasing engineering productivity,

making the software even easier to use and

interact with, while significantly reducing the time

required to get a high-quality solution, said SeniorVP Product Management Jean-Claude Ercolanelli.

STAR-CCM+ v8.02 is the first of our v8 releases

to benefit from our investment in a dedicated

User Experience Team, whose task is to dissect and improve every

aspect of the software to give engineers more effective and more

productive simulation tools.

New multidisciplinary enhancements allow users to tackle a wider range

of industrially relevant challenges, continued Ercolanelli. Among these

new features and enhancements is a new STAR-Cast add-on, developed

in collaboration with our partner ACCESS, recognized experts in casting

and metallurgy, which provides a comprehensive and intuitive process for

performing multiphase casting simulations and brings automation and

ease-of-use to casting and foundry processes.

Usability-related enhancements include:

- Parts-based meshing allows users to associate mesh denitions with

geometric entities, resulting in greater control, better automation and

reduced turnaround time.

- New surface preparation functionality greatly reduces the amount of time

required to clean-up imported CAD geometries, particularly those

which include large assemblies of components.

- JTOpen integration cuts import times from hours to minutes for large

complex CAD assemblies.

- A number of Graphic User Interface (GUI) enhancements signicantly

improve workow automation.

Performance-related enhancements include:

- Lagrangian and DEM dynamic load balancing improves runtimes for

applications such as SCR devices, IC engines and chemical sprays by

at least a factor 2.5.

- Improvements to the AMG algorithm dramatically decreases simulation time

on high-processor-count clusters for large scale unsteady simulations such

as underhood, aerodynamic and aeroacoustic analyses.

Expanded coverage includes:

- STAR-Cast add-on is a new streamlined casting simulation process that

places industrial strength simulation technology in the hands of foundrymen,

casting designers and tool makers.

- The uid lm model can now be used with the moving reference frame (MRF)

model to simulate lms on moving objects such as pumps and break-disks,

as well as with the coupled solver, which is a key requirement for the

aerospace industry. In addition, it can be applied to simulate icing and

de-icing effects using a multi-component melting and solidication model.

- The Eulerian multiphase capability is improved through the addition of

interphase and intraphase reaction models, which is ideal for tackling

problems in the chemical and process industries.

- A new co-simulation capability through coupling with AMESIM, a 1D multi-

domain simulation tool, is implemented, enabling simulation possibilities

for hydraulics, IC engines, electro-magnetic and fuel injection systems.

i STAR-CCM+ V8: INFINITE POSSIBILITIES

www.cd-adapco.com/news

Follow the latest breaking news online:

CD-adapco has announced therelease of STAR-CCM+ v8.02,a major new version of theirintegrated solution formultidisciplinary engineeringproblems.

5 dynamics I S S U E 3 4

..::INTRODUCTIONBreaking News

8/10/2019 Dynamics 34

7/41

CD-adapco has always been engaged with high performance

computing partnerships. Recently, the company produced an

informational video on simulations using Penguin Computing

and Intel, featuring NASCAR Company, Michael Waltrip Racing

and Cervelo Bicycles, manufacturers of the worlds fastest and lightest

racing bikes.

To complement this video, CD-adapco hosted a webinar titled, The Need for

Speed High Performance Sports such as NASCAR, Formula 1, and Cycling

Turn to Digital Simulation for Competitive Edge. This online event, for which

the recording is now available on CD-adapcos website, showcased examples

from Michael Waltrip Racing and Cervlo, two companies who understand

the benets of speed on and off the track. Several additional examples of

engineering simulation in the sports world were also featured in the webinar.

Guest speaker, Ivn Sidorovich from Cervlo, showcased aerodynamic

analysis of their race-winning bicycle designs, and the benets of using digital

simulation combined with wind tunnel testing to save time and money.

Other experts in the eld demonstrated how smaller engineering organizations

are quickly realizing the importance of automated meshing, advanced physics

simulation, intuitive workows, high performance computing, and cloud

computing to reduce the cost and time of product development.



CD-adapco DIGS DEEPER INTO

HIGH PERFORMANCE COMPUTING

0dynamics I S S UE 3 4

..::INTRODUCTION Breaking News

i WE DIG DEEPER INTO HIGH PERFORMANCE COMPUTING

CD-adapco continues to demonstrate its ongoing commitmentto high performance computing through various outlets.

8/10/2019 Dynamics 34

8/41

CD-adapco signed a multi-year agreement with the leading

global nuclear supplier, AREVA, to deploy STAR-CCM+.

Nicolas Goreaud, Head of the AREVA Reactor CFD simulation team

in Lyon said, CFD technologies play a fundamental role in the design and

development of nuclear reactors. The ability to quickly and accurately simulate

the behavior and performance of our designs is key for ensuring safety

and improving durability of reactor design. We are pleased to continue our

relationship with CD-adapco, an acknowledged leader in this eld, as we work

towards our next generation designs.

AREVAs success is built around safety, performance and reliability of designs;

and CD-adapco is proud to be a part of that team, said Bill Clark, CD-adapcos

VP of Operations. In the past years, our software has been recognized as the

leading tool in the nuclear industry and we are happy to continue demonstrating

this by collaborating with industry partners like AREVA.

The agreement has been set-up to enable all AREVA engineering sites

worldwide to get access to CD-adapcos technologies.

This new agreement combines the expertise and experience from both

companies to enable state-of-the-art technologies like STAR-CCM+ to spread

within AREVA.

AREVA TO USE CD-adapcos TECHNOLOGIES WORLDWIDE

Steve is specifically designed to bring customers closer to their

Dedicated Support Engineer.

CD-adapco announced the availability of its new customer

support portal called Steve. Stephen McIlwain, Director Engineering

Customer Support (Americas), stated, CD-adapco has a highly responsive

and technically competent team of engineers to help its customers meet

their engineering analysis needs. Steve was built to be an extension of

this best-in-class customer support and offers yet more opportunities for

customers to interact with CD-adapcos engineering teams.

Steve includes the establishment of a brand new set of frequently

asked questions (FAQs) created from the rich database of knowledge

built up from assisting engineering companies all around the globe in

using CD-adapcos agship products over many years. This database

also includes multimedia les that provide users with short step-by-

step guides to many of its softwares features.

i HELP IS AT HAND WITH OUR NEW SUPPORT PORTAL



CD-adapco Introduces New Customer Support Portal, Steve

CD-adapco, Worldwide Leader in

CAE Engineering Software and

Services, announces Multi-year

Agreement with AREVA

dynamics I S S U E 3 47

..::INTRODUCTIONBreaking News

CD-adapco is committed to its customers Engineering Success, therefore

each customer has a dedicated support engineer (DSE), who is tasked with

understanding both the customers workow and engineering applications in

detail, identifying potential problems before they even arise, and providing

an immediate resolution when they do.

McIlwain continued, Traditionally, customers have accessed their DSE

over the telephone, or by dropping them an email. With the arrival of Steve,

CD-adapco is adding new methods by which customers can interact with

their DSE and opening up the capability to extract knowledge from across

a whole community of engineering professionals.

CD-adapco customers can gain access to Steve by contactingtheir DSE.

8/10/2019 Dynamics 34

9/41

i DONALD RIEDEBERGER WINS CALENDAR COMPETITION

dynamics I S S UE 3 4 0

..::INTRODUCTION Breaking News

CD-adapco is pleased to announce Donald Riedeberger

of the University of Stuttgart as the winner of the 2013

CAE Post-Processing Contest for his image of a dolphin.

Riedeberger submitted the image which shows off laminar-

turbulent transition on a common dolphin at 1 m/s and 1%

turbulence intensity.

For the seventh year running, the competition has given

users of CD-adapcos software the chance to showcase their

post-processing skills by producing striking images of their

work. In winning the competition, and claiming a Samsung

Galaxy SIII and Nexus 7 as his prize, Riedeberger beat off

tough competition from companies such as Volvo, CNH, Behr,

Atkins and Gruner.

CD-adapcos VP of Marketing, David Vaughn said, Post-

processing CAE results is no longer just about getting

numbers from a simulation, but also about being able to

engage and inform colleagues from other disciplines about

just how effective an analysis was. The quality of this years

entries has once again impressed everyone at CD-adapco

and has shown just how skillful our users are in effectively

deploying simulation across a huge range of different

applications. With over 100 entries to choose from, all of an

exceedingly high standard, voting proved to be a very close

race and is a reection of both the high levels of skill shown

by our users and of the versatility of STAR-CCM+.

The best twelve images are showcased in the2013 CD-adapco Desktop Calendar.

CD-adapco ANNOUNCES THEUNIVERSITY OF STUTTGART ASITS 2013 CALENDAR COMPETITIONWINNER! Donald Riedebergers dolphin simulation won the 2013 competition

Above:Donald Riedeberger receiving his prizes

8/10/2019 Dynamics 34


INTRODUCTION

E

stimation of helicopter parasiticdrag is an important step in thedesign process that will dictate thep o w e r a n d p r o p u l s i v e f o r c e

requirement at high speeds. Te total drag on ahelicopter is the sum of the parasitic, frictionaland lift-induced drag. Parasitic drag is due to the

non-lifting parts, frictional drag is caused bythe frictional resistance of the blades and lift-induced drag, as the name implies, is a result ofthe lif t production. In sing le-rotor helicopters,nearly of the total vehicle drag can becaused by the parasitic drag from the hub. Mini-mum possible drag is a key requirement in anyhelicopter design and reducing the hub dragplays a major role in achieving this.

One way to reduce hub d rag on conventionalarticul ated rotors is to use a fairing, which whileminim izing drag, leads to increased mainte-nance and inspection workload. Due to this,alternate methods of reducing hub drag aredesirable and one approach is to design the

SIKORSKY AIRCRAFT PREDICTS

DRAG OF PRODUCTION ROTORHUB GEOMETRIES USING CFDALANEGOLFSikorsky Aircraft MIKEDOMBROSKICD-adapco

Above:Sikorsky S-92 Helicopter

components of the hub such that they gener-ate less drag as a whole when installed in thehub. raditionally, hub drag estimation involvedpredicting the d rag build-up of the componentsbased on empirica l drag data from componentsof similar or almost similar shapes and sum-ming up their individual contributions. Asidefrom being based on historical data, this method

also involves estimation of interference effectsand is less valuable in a production environ-ment where optimization of component shapesis important. Eventually, the rotor hub designedbased on this subjective process is tested in a

wind tunnel, leading to an expensive processif design c hanges and improvements are to beimplemented and tested again.

Sikorsky Aircraft set out to explore analternate method of predicting hub drag ofproduction geometries based on numerical sim-ulation. his method can provide a reasonableprediction of hub drag for different designs in a

short time period, allowing easier optimizationof component design in a production environ-ment. his article showcases the applicationof CD-adapcos unstructured Navier Stokessolver, SAR-CCM+, to the blind prediction ofhub drag on two production rotor hub geom-etries, the S-A hub and the UH-A hub.

COMPUTATIONAL GEOMETRYAside from time savings in the design process,the real value of numerical simulation lies i n theaccuracy of the prediction of hub drag, particu-larly in blind calculations with no knowledgeof experimental data. he two rotor hubs in

STAR-CCM+ IS WELL-POISED TO TAKE ON THECHALLENGE OF PREDICTINGTHE WAKE STRUCTUREDOWNSTREAM OF THE HUBWITH HIGH FIDELITY.

..::FEATURE ARTICLEAerospace

8/10/2019 Dynamics 34

11/41dynamics I S S UE 3 4 1

conditions. Te surface representation ofthe S-A and UH-A hub are shown inthe accompanying image, in addition tothe hub/pylon geometry and the com-

putational domain.

MESHTe hub geometry was discretized at thesurface level using the surface wrappermethod in SAR-CCM+ before remesh-ing the surface. he surface wrappershrink wraps a mesh onto the geom-etry a nd creates a water-tight surface,preserving the geometric fidelity of thesurface, including minor details like nutsand bolts. he computational domainis th en dis cretized u s in g trimmedhexahedral cells in the volume, with a

prismatic boundary layer mesh near thesurface to capture the boundary layer

this ana lysis, S-A and UH-A weretested at a / size scale in in theURC main wind tun nel as part of theS-A aircraft development process.

Even though data on the d rag build-upof individual components was availablefrom this test, the numerical simulationswere performed as blind calculationswithout knowledge of the experimentalresults. he simulations were carriedout including the wind tunnel wal ls andtest pylon/splitter plate assembly, with-out considering the support structurefor the assembly. he swash plates inthe experiments were non-functionaland hence the link between the platesand their servos was removed in boththe experimental tests and the simula-tions. Te hub was tilted forward by five

degrees, while the test pylon/splitterassembly was kept level as per the test

Above:Surface representation of 1/2 scale S-92A hub

SIKORSKYAIRCRAFT

CORPORATION

is a world leaderin the design,

manufacture andservice of military

and commercialhelicopters; fixed-

wing aircraft;spare parts and

maintenance,repair and

overhaul servicesfor helicoptersand fixed-wing

aircraft; andcivil helicopter

operations.

Out

Splitter Plate Walls

Hub

Inlet

flow. he body-fitted boundary layermesh had four prismatic cells, with tenlayers of cells used on the hub cover toaccurately resolve the thick boundarylayer on this surface. Focused volumet-ric refinement based on the solutionfrom a coarse grid was used behind thehub to capture the hub wake. A slidingmesh was used around the hub assem-bly which will be rotational. he finalvolumetric mesh for the S-A hubconsisted of .M trimmed hexahedralcells, with the prismatic boundary layermesh accounting for .M cells. Si mi-lar process for the UH-A hub yielded.M advanced hexahedral cells, with.M cells in the boundary layer. Detailsof the volume mesh are shown in theaccompanying images.

SOLUTION METHODOLOGY

he solution methodology was a blindcalculation following the best prac-tices fo r mo vin g bo dy s imu la tio nwithin SAR-CCM+. Initial runs wereperformed on a coarse grid to obtainan initial solution that was used forverification of the setup and to identif yzones for mesh refinement. he solu-tion process followed the wind tunneltests in reverse, with a full configurationfor the S-A hub initi ally, followed byremoving the beanie, pushrods, scis-sors & servos, swash plate and bifilar inconsecutive runs. Simi larly, the UH-Aruns were started with a full configura-

tion, followed by removal of bifilar andpitch-link rods in subsequent steps. In

total, there were and configurationseach for the S-A and UH-A hubsrespectively. Steady state simulationswere conducted on both hubs with aninlet velocity of knots, hub rota-tional rate of rpm and an advanceratio of ., similar to advance ratios ona full scale rotorcraft. he simulationswere run at the same Reynolds numberand Mach number as the experimentsbut at scale values compared to flight

Above: Surface representation of the 1/2 scaleUH-60A hub

Above:Wind tunnel model and solution domain


8/10/2019 Dynamics 34

12/41

8/10/2019 Dynamics 34



perpendicular to the flow with large frontalarea () while min imum drag occurs when theblade attachments are at angle to the flowwith minimum frontal area. Te time-averageddrag value differed by between the and. time steps, while the di fference in averageddrag between URANS and DES runs was ..Te DES method resolved the turbulence in thehub wake better but the spectral content of thisturbulence did not have much impact on over-all hub drag. Te final va lidation simulations forboth hubs with their different configurationswere performed as DES runs with a ti me step of.

Results from the DES runs for the S-Ahub showed that the addition of componentsincreased drag and correlated well with thewind tunnel results. he numerical resultsgenerally over-predicted the drag slightlyand the l argest error between simulationsand test was below . For the UH-A hub,the numerical results under-predicted the

drag while other trends were the same as theS-A hub. Te drag shown in the tables belowis the normalized computed drag, based onthe experimental base hub drag. Contours ofsurface pressure and velocity magnitude atmid-plane are shown in the accompanyingimages. Te effect of rotation can be clearly seenin the surface pressure, while the unsteadyvortices shed from the rotating hub ca n be seenin the blue-green regions behind the hub in thevelocity magnitude.

he Sikorsky S-A hub showed a tailexcitation during tests in the early stages offlight development that were attributed to thewake coming f rom the scissor and associated

fittings on the rotor hub, the only structuresthat could induce a per rotor revolution

Base

Base+Bifilar

Base+SwashPlate

Base+SwashPlate+

Scissors&Servos

Base+SwashPlate+Scissors&

Servos+Pushrods

Base+SwashPlate+Scissors

&Servos+Pushrods+Beanie

1.4

1.2

1.0

1.8

0.6

0.4

0.2

0

Prediction

Test

Above:Normalized drag of S-92A hub configuration Above:Normalized drag on UH-60A hub configuration

forcing function. his was resolved by raisingthe vertical position of the hub and by makingchanges to the pylon. he unsteady analysison the S-A hub compared the fast fouriertransformation (FF) of the unsteady rotorhub drag with and w ithout scissor and scissorfittings. he graph on the left shows that thesimulations with all components included leadto large p drag force, while configurationswithout the scissors still exhibit a small pcontent coming from small drag from thescissor fittings.

CONCLUSIONSikorsky Aircraft set out to study the feasibilityof using numerical simulations to accuratelypredict hub drag of new design hubs early i n thedesign phase. Te blind numerical si mulationsshowed that SAR-CCM+ can predict the hubdrag reasonably well, with the largest errorbeing compared to the experiments. For an

initial st udy without grid convergence analysis,these are acceptable and the predictionswill only improve with solution-based gridrefinement and time step studies. he timetaken to go from CAD to results was around man-hours and ~ CPU-hours for MRFstudies and ~ CPU-hours for DES studies.Tese numbers show that an experienced usercan conduct a hub drag analysi s as part of thedesign study very quickly and effectively withacceptable accuracy, thereby providing earlyinsight i nto the hub design. Te results from thisstudy indicate that in addition to generating gridsin a timely manner for complex hub drag studies,SAR-CCM+ is also well-poised to take on the

challenge of predicting the wake structuredownstream of the hub with high fidelity.

Normalized computed Drag Relative to Experimantal Base Normalized compute Drag Relative to Experimantal Base

1.4

1.2

1

0.8

0.6

0.4

0.2

0

Base Base +Pushrods

Base + Pushrods+ Bifilar

Above: Drag convergence history in unsteadymodes of operation

Drag

URANS, 5 degrees / time step

DES,5

degree

s

DES,

0.5

degrees

0 1.0 1.5

Above:Harmonic content of unsteady drag forselected S-92A configurations

FFT of Unsteady Hub Drag

0p 1p 2p 3p 4p 5p 6p 7p 8 p 9p 10p

Base + Swash Plate + Scissors & Servos + Pushrods

Base + SwashPlate

Bse Hub

8/10/2019 Dynamics 34

14/41

T

he recent push towards

more fuel-efficient andenvironmentally friendlyaircraft is causing a dra-

m a t i c s h i f t i n t h e d e s i g n o f t h enext-generation of aircraft. With surg-ing fuel prices and more and more peopletaking to the skies, aircraft companiesh a ve been fo rced to reth in k th eirapproach to future designs, keeping inmind the goal of reducing fuel consump-tion by up to by []. o achievethis, increasingly li ghter and more flex-ible composite structures are beingintroduced, and in novative, unconven-tio n a l des ig n s res u ltin g in h ig h er

lift-to-drag ratios are on the table. Withsuch radical changes to the structure,propulsion and the aerodynamic shapeof aircraft, the use of high-fidelity com-putational aeroelasticity (CAe) early inthe design process will be critical tomeet tomorrows design challenges.

THE CRITICAL ROLE OF HIGHFIDELITY COMPUTATIONALAEROELASTICITYAvoiding negative aeroelastic impactswhile at the same time exploiting the

benefits of aeroelasticity will be critical

dMeeting air-worthiness requirements: Te aerospace

industry has entered uncharted territory with theintroduction of light, flexible composites and withthe development of revolutionary concepts such asmorphing vehicles, joined-wing aircraft, blendedwing-body configurations and innovative unmannedaerial veh icles (e.g. HALE UAV). As a result, companiesare no longer able to rely on their historical knowledgeaccumulated from traditional designs and insteadneed to gain a solid understanding of a new designsreal-world behavior early on through simulations.One major challenge for the industry is to accuratelypredict aeroelastic phenomena critical to fli ght safety,especially i n the transonic flight regime (e.g. flutter,buffet and buzz). o take full advantage of lighter designs,it will be of utmost importance to avoid surprises late in

the development cycle, as they almost always result inaeroelastic weight penalties and huge safety margi ns tomeet aeroelastic stability air-worthiness requirements.

SABINEA. GOODWIN, CD-adapco

Above:Computational aeroelastic analyses are critical to modern aircraft design due to the increased interdependency between structures and aerodynamics caused by moreflexible composite materials.

THE TREND OVERTHE NEXT 20 YEARS IS NOTMORE OF THE SAME. IT IS

LARGER AIRCRAFT, CLEANER

AIRCRAFT, MORE FUEL

EFFICIENT AIRCRAFT.

AERO E L A S T I C I T Y :A KEY ENABLING TECHNOLOGY

FOR THE DESIGN OF NEXT

GENERATION AIRCRAFT

- John Leahy, COO Airbus, discuss ing AirbusGlobal Market Forecast 2012-2031

COMPUTATIONAL

for improved performance and reducedcost of the future fleet. As the industrytakes advantage of increasingly morepowerful and lower cost computers,integrating validated high-fidelity CAemethods early in the design process willsoon become a necessity for companiesto remain competitive.

Above:Flexible wing deflection

dynamics I S S UE 3 43


8/10/2019 Dynamics 34

15/41

d Developing and deploying game-changing

technologies: As airframes become more andmore flexible, they also become increasinglysensitive to dynamic atmospheric disturbancessuch as turbulence. High-fidelity CAe willbecome a key enabler in the development ofnovel active control technologies to limit aflexible vehicles dynamic response to thesedisturbances and to minimize critical design

loads.Game-changing struct ures technologiesare also opening the door for exploiting thepotential benefits of aeroelasticity. Flexi blecomposite materials now allow the designerto introduce directional stiffness to the wing,and smart structures in conjunction withactive control can be applied to aeroelasticallyshape the wing for drag reduction, improvedstability, and load alleviation. High-fidelityCAe enables for these innovative technologiesto be evaluated up front, and the knowledgeacquired through the simulations can thenbe transferred into the design resulting i n afurther weight reduction.

CHALLENGES OF HIGHFIDELITY COMPUTATIONALAEROELASTICITYh e co u plin g o f two dis tin ct en g in eerin gdisciplines like fluid and structural dynamicsis not trivial; they are inherently dissimilar andthe computational methods for each have beendeveloped largely independently of each other.Combining computational fluid dynamics(CFD) and computational structural dynamics(CSD) solvers for simulating non-linear fluid-structure interaction problems demands a

carefully designed implementation to ensurerobustness, stability, accuracy and efficiency ofthe resulting CAe capabil ity [].

dMapping and data exchange must be robustand efficient: he CFD and CSD grids are oftennon-conformal, each requiring different griddensities and topologies. In addition, thewetted areas on each model do not alwaysgeometrically match, making it challengingto identify the proper interface regions forinterpolating data. ypically, the mapping andsynchronization for data exchange in CAeare performed with third-party inter-codecommunication software or via f ile transfer,

greatly increasing overhead and reducingefficiency. In addition, if the interpolation ofgeometrical and loads data is not implementedcarefully, an imbalance in t he transfer of energybetween the models can lead to inaccurate andunstable solutions especially when predicti nghighly non-linear aeroelastic phenomena suchas flutter.

d Coupling strategy affects accuracy, sta-

bility and flexibility: he approach to use forcoupling the fluid and structure high ly dependson the type application and the complexity ofthe cross-coupling between the disciplines.Coupling methods can be largely divided intotwo classes: explicit and implicit. A trade-offbetween flexibility and accuracy must be

Figure 1:Experimental set-up Figure 2:Resonance on plate

66

66

66

66

66

66

66

66

66

66

66

66

Fluid Domain

Vertical Plate

Outlet

Inlet Velocity10m/s

T

h i s s i m p l e c a s e s t u d yinvestigates the stro ngt w o - w a y a e r o e l a s t i ccoupling between an elastic

plate with two funda mental modes (a 4 Hz1st bending mode and a 20 Hz 1st twistingmode) and compressible air movingnormal to the plate at 10 m/s (Figures 1and 2). he two-way coupled SAR-

CCM+/Abaqus FEA co-simulation wasused to perform an aeroelastic analysisand to demonstrate the presence ofresonance resulting from damping of the1st bending mode and excitation of the 1sttwisting mode during this experiment.

o validate the u se of the SAR-CCM+SS k- turbulence model for thisapplication, rigid unsteady Reynolds-A v e r a g e d N a v i e r - S t o k e s ( R A N S )computations were i nitially performed.Both the computed time-averaged dragcoefficient and Strouhal nu mber (a non-dimensional number describing thefrequency of vortex shedding) matched

w e l l w i t h p r e v i o u s l y d o c u m e n t e d

e x p e r i m e n t a l r e s u l t s , c o n f i r m i n gthat the turbulence model accuratelypredicts the complex turbulent flowpatterns in the wake behind the flatplate.

he shedding frequency o f thevortices in the wake was predicted at19.5 Hz, nearly identical to the naturalfrequency of the twisting mode. hese

results suggest that resonance willlikely occur when aeroelastic effects areincluded in the simulation.

During the two-way aeroelasticcomputation with SAR-CCM+/AbaqusFEA, the elas tic deformation of the platewas initially as expected: bending inthe direction of the wind . As the solutionprogressed in time, the bending modeof the structure was da mped out by theair flow and the plate began to twist asthe periodic driving force of the vortexshedding excited the 1st twistingmode. he simulation was successfulin predicting the resonance expected to

occur during this experiment.

PLATE VORTEX

INDUCED VIBRATION

CASESTUDY

01



8/10/2019 Dynamics 34

16/41

FEATUREAEROSPACE

Above: Streamlines from a MQ-1 Predator. The aerospace industry has entered uncharted territory with the introduction of revolutionary concepts such as innovativeunmanned aerial vehicles. CAe is a key enabling technology to accurately predict the aeroelastic phenomena critical to flight safety of these unconventional designs.

FLUTTER OF THE AGARD WING 445.6

CASESTUDY

02

Figure 3:Surface mesh on Wing 445.6

Figure 4:Flutter boundary (Wing 445.6)

Mach Number



T

he AGARD Wing 445.6 is a stan-dard aeroelastic configurations p e c i f i c a l l y d e s i g n e d f o rdynamic aeroelastic response.

his geometry was extensively tested in the16-ft ransonic Dynamics unnel at the NASA

Langley Research Center and resulting datahas been widely used for verification of vari-ous CAe codes for the past 20 years [3, 4]. his

study was performed to validate the two-waycoupled SAR-CCM+/Abaqus FEA co-simula-

tion for flutter prediction.Using the polyhedral meshing capabilityin SAR-CCM+, a fine viscous mesh withprisms in the boundary layer was generated,sufficiently refined to ensure accuratecapture of the shock locations at transonicconditions. he surface mesh is depicted inFigure 3.

o compute the flutter boundary, the

motion of the wing was initiated by applyingan impulse load to the st ructure, and unsteady

RANS time-marching calculations using thetwo-way coupled SAR-CCM+/Abacus FEAco-simulation were performed.

he computed flutter characteristics,represented in terms of the flutter velocityindex (FVI), are depicted in Figure 4. Resultsdemonstrate that the method nicely captures

the experimental flutter boundary includingthe significant drop in the flutter speed attransonic conditions (also called the flutterdip). Similar results were observed when

analyzing the flutter frequencies. Whencomparing these results to published resultsof other codes [2], the located boundaryis inside the range of the published dataspread with an error of less than 15%. his isconsidered good from an engineering pointof view and validates the ability of the CAecapability to accurately predict flutter.

8/10/2019 Dynamics 34

17/41

considered when deciding on a strategy.

With explicit coupling, the effects ofmoving the fluid mesh lags the solutionby one time step. Tis coupling approachresults in greater flexibility for the userbut can be unstable and less accurate,especially in applications when lightand compliant structures interact withheavy fluids.

Fully implicit coupling is the mostrigorous and robust approach becauseit solves a full system of cross-coupledfluid and solids equ ations. However, thiscoupling requires a much more inti-mate integration of the solvers, and itbecomes increasingly difficult to imple-

ment as more complicated physics areintroduced into the system.

d Dynamic mesh evolution is cum-

bersome and costly: One of the keydifficulties in CAe is that the physicalmotion of the geometry calls for a capa-bility to move all nodes in the CFD gr id atevery time step, a costly task when run-ning un steady simulations on complexgeometries. In addition, it is impera-tive that the mesh movement does notintroduce skewed cells and grid densitychanges as this can have a negativeeffect on the convergence and accuracy

GLOBAL VIGILANCE, REACH, AND POWER AT HOME ANDABROAD REQUIRE VAST AMOUNTS OF ENERGY - WHETHER IT ISFUEL FOR OUR AIRCRAFT, GAS FOR OUR VEHICLES, OR ELECTRICITYFOR OUR SPACE AND CYBERSPACE EFFORTS. AS THE LARGEST

ENERGY USER IN THE FEDERAL GOVERNMENT, THE AIR FORCE MUSTFIND WAYS TO REDUCE OUR ENERGY CONSUMPTION, ESPECIALLYGIVEN THE CURRENT ECONOMIC ENVIRONMENT.- Secretary of the Air Force Michael Donley

of the solution. With so many obstacles,dynamic mesh evolution is one of themost difficult problems to handle, oftenleading to costly and time consumingroad blocks where the user needs tostep in to manually fi x or regenerate themesh.

d Ease-of-use and automation are

imperative: Up until now, high-fidelityCAe has been mostly a research exer-cise, and running the simulation hasrequired a great deal of specializedknowledge and training. Ti s has oftenbeen a s h o ws to pper wh en co mpa -nies are considering integrating thecapability into the engineering designprocess. In todays fast-paced produc-tion environment, companies demandease-of-use and automation, so theycan focus on results rather than on fig-uring out how to set up, prepare, and runthe simulations.

STARCCM+ PROVIDESTHE SOLUTIONSSAR-CCM+, CD-adapcos flagshipsoftware, offers practical solutions formany of the chal lenges encounteredwhen tackling highly non-linear fluid-structure i nteraction problems such asaerodynamic flutter and buffet.

d Built-in mapping and co-simula-

tion coupling are robust and efficient :SAR-CCM+ has a direct link to Abaqusfinite element analysis (FEA) through aco-simulation application programminginterface (API) developed by SIMULIA,delivering a fully coupled, two-way,fluid-structure interaction. Direct co-simulation coupling provides efficiency

and reduced overhead associated withdata transfer through file exchangesand use of external middleware soft-ware. As data is passed back andforth via sockets, the API manages allexchange synchronization (how oftendata is passed back and forth) whileboth codes are runnin g in memory. Inaddition, the SAR-CCM+/Abaqus co-simulation gives the user the flexibilityto choose between explicit or implicitcoupling, depending on the application.

he built-in mapping implementedin SAR-CCM+ is robust and accurateand efficiently ha ndles non-conformal

meshes with no need for writing scriptsand input files. Mapping is done ina distributed manner (local on eachpro ces s o r) en s u rin g th a t th ere isenough memory available to reliablyhandle the most complex geometries.

dJava automation results in flexibility

and customization: In addition to directBelow:State-of-the-art meshing inSAR-CCM+



8/10/2019 Dynamics 34

18/41

Figure 8:Phase from prescribed motion

0.0 0.2 0.4 0.6 0.8 1.0

/

Phase

180

135

90

45

0

-45

-90

-135

-180

7U

Exp

STAR-CCM+

STATIC & DYNAMIC COMPUTATIONS

ON THE HIRENASD WING

comparisons with expe rimental data [6].

he wing aeroelastic equilibrium shape andpressure coefficients (C

p) were computed

for various angles of attack () using thetwo-way coupled SAR-CCM+/Abaqus FEAco-simulation. Figure 6 depicts chord-wiseC

pdistributions on an outboard section of the

wing (M=0.8, =2), dramatically showing theeffect the deformation of the structure has onthe loads and indicating a good comparison

with experiments. Simila r results were obtainedwhen looking at t he span-wise displacementsand lift/drag distributions vs. . [7]

he accuracy of SAR-CCM+ was furthervalidated using t he one-way coupled prescribed

motion technique. In this approach, the second

eigenmode was extracted from Abaqus andused to prescribe a harmonically-varying gridmotion about the wing aeroelastic equilibrium.he geometry was moved using mesh morphi ng

in SAR-CCM+. Several periods of prescribedvibration were computed, aimed at verifyingS A R - C C M + ' s a c c u r a c y c o m p a r e d t oexperiments and other simulation codes.Figures 7 and 8 depict the magnit ude and phaseof the Fourier transforms of C

p on the upper

surface on an outboard station of the wing(M=0.8, =-1.34), showing good comparison

to experi mental values. Simi lar resultswere obtained on the lower surface and onspan stations further inboard on the wi ng.

Although not shown, these results alsocompare well to resu lts of other CAe codes [8].

Free and forced wind-on vibrations using2-way coupling with SAR-CCM+/AbaqusFEA co-simulation were also performedand results compared well to the publishedexperimental results [9].

his investigation confirms that theS A R - C C M + / A b a q u s F E A c a p a b i l i t yaccurately predicts both static and dynamicaeroelastic effects at transonic conditionsand realistic flight Reynolds numbers.

GOAL FOR SUSTAINING OUR FUTURE: TO DEVELOP ANDOPERATE AN AVIATION SYSTEM THAT REDUCES AVIATIONSENVIRONMENTAL AND ENERGY IMPACTS TO A LEVEL THATDOES NOT CONSTRAIN GROWTH AND IS A MODEL FORSUSTAINABILITY. - FAA Strategic Plan - Destin ation 2025

Figure 5:HIRENASD surface mesh

Figure 6: Static aeroelastic solution

-1

-0.8-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

10.0 0.2 0.4 0.6 0.8 1.0

/

Experiments

STAR-CCM+ rigid

STAR-CCM+ staticaeroelastic

CASESTUDY

03

Figure 7:Magnitude from prescribed motion

F

or this validation study, staticand dynamic aeroelastic com-putations were performed onthe High Reynolds Number

Aero s tru ctu ra l D yn a mics (HI RE N ASD )wing (Figure 5). he wing was originallytested in the European ransonic Windunnel [5] and offers both static and dynam-

ic measurements at transonic conditionswith realistic flight Reynolds numbers. hiswork was part of the first Aeroelastic

Prediction Workshop.o ensure CFD solution accuracy, a griddensity study was performed on the rigidwing at transonic conditions. As the meshwas refined, lift and drag were comparedwith previo u s ly pu blis h ed rig id bo dycomputational data and a drag convergedmesh was identified for this validation study

[6]. o v a l i d a t e t h e F E A m o d e l , a n

eigenfrequency extraction analysis wasperformed in Abaqus resulting in good



8/10/2019 Dynamics 34

19/41

c o - s i m u l a t i o n c o u p l i n g w i t hAbaqus, SAR-CCM+ enables CAecomputations through import-in g /expo rtin g CSD mes h es inother native formats (e.g. Nastran,Ansys) as it leverages the powerof Java to give users the ability to

customize every step of the simu-lation workflow. Although thisfeature requires work up front toset up and manage the simulation,the pay-off is greater flexibility touse legacy codes and simplifiedCSD models.

F o r e x a m p l e , w i t h t h i sapproach, it is possible to performan aeroelastic analysis on therotating blades of a helicopter byusing high-fidelity CFD in SAR-CCM+ in conjunction with a simplebeam-rod approach for modelingstructural deformation and pitch

of the blades.

d State-of-the-art meshing and

morphing reduces turnaround

time : he viability of deploying aCAe simulation in a productionenvironment strongly dependson the ability to quickly generateh i g h - q u a l i t y c o m p u t a t i o n a lmeshes. With unrivaled polyhedrala n d t r i m m e d c e l l m e s h i n g ,S A R - C C M + c u t s g e o m e t r ypreparation and meshing timeo n very co mplex g eo metriesdown from months to hours. In

a dditio n , th e mu lti- qu a dra ticmorphing capability robustly andsmoothly moves these meshes(of any topology) based on thedeformation it receives from theCSD solver. he resulting CFDmesh conforms to the shape of

the deflected structure, and theredistribution of the mesh verticesnicely preserves the quality of theoriginal mesh.

In addition to a superior mor-phing technology, SAR-CCM+also has an overset mesh capabil-ity. When using overset meshes,in the case of flow around bodiesat various relative positions, oneneeds to generate individual gridsonly once and then compute theflow for many combinations orrelative positions and orie ntationsby simply moving grids, with no

need to re-mesh or change bound-ary conditions. his capability isa key enabler for applications inaero-servo-elasticity such as gustload alleviation, where it is vital tomodel the control surfaces deflec-tions as part of the simulation.

d Intuitive user simulation envi-

ronment with high-fidelity physics

delivers engineering solutions:

SAR-CCM+ drives innovation asit seamlessly fits into any exist-ing engineering process and givesthe designer the power to handle

the most complex multi-physicsproblems with ease. In additionto performing high-fidelity CAesimulations, with the same code,the user can easily include tem-perature effects, aero-acoustics, degrees-of-freedom (DOF)

and other high-fidelity physicsin a fully coupled manner. he netresult is more time analyzing dataand less time preparing and set-ting up simulations.

CONCLUSIONIn todays competitive climate,driven by climbing fuel costs andincreasing demand for air travel,high-fidelity CAe is a key enablingtechnology for aerospace compa-nies to develop innovative mini-mum structural weight designs

while meeting the tight scheduleand cost constraints of a typi-cal production environment. It isimperative that the CAe capabil-ity is accurate, robust and efficientand that it easily fits into the cur-rent engineering design process-es, producing high-quality resultswith minimum user efforts. Withits unrivaled meshing technology,high-fidelity physics, intuitiveuser environment and direct linkto Abaqus FEA for co-simulation,S A R - C C M + s e a m l e s s l y i n t e -grates CAe into the design process,driving i nnovation and resulting inengineering success.

REFERENCES

01. NASA Facts NF-2010-07-500-HQ

02. Schuster, D., Liu, D. and Huttsell, L .,

Computational Aeroelasticity: Success,

Progress and Cha llenge, AIAA -2003-1725

03.Yates, E.C., AGARD Standard

Aeroelastic Configurations for Dynamic

Response. Candidate Configuration I.

Wing 445.6, NASA M 100492, Aug. 1987

04. Yates, E.C., Land, N.S., and Foughner, J..,

Measured and Calculated Subsonic and

ransonic Flutter Characteristics of a 45

Degree Sweptback Wing Platform in Air

and Freaon-12 in the Langley ransonic

Dynamics unnel, NASA N D-1616, March

1963

05.Ballman n, J. et al., Aero-Structural

Wind unnel Experiments with Elastic

Wing Models at High Reynolds Numbers

(HIRENASD-ASDMAD), AIAA -2011-0882,

January 2011

06.Florance, J. Chwalowski, P. and

Wieseman, C., Aeroelasticity Benchmark

Assessment, Aeroelasticity Branch, NASA

Langley Research Center, Subsonic Fixed

Wing Program, Interim Report, March 2010

07.Heeg, J., Florance, J., C hwalowski, P.,

Perry, B. and Wieseman, C., InformationPackage: Workshop on Aeroelastic

Prediction, Aeroelasticity Branch, NASA

Hampton, Virginia, October 2010

08.Schuster, D., Chwalowski, P., Heeg,

J., and Wieseman, C. Su mmary of Data

and Findings f rom the First Aeroelastic

Prediction Workshop, ICCFD7-3101

09. Ballmann, J. et al., Experimental

Analysis of High Rey nolds Number Aero-

Structural Dyna mics in EW, AIAA 2008-

841, Presented at the 46th AIAA Aerospace

Sciences Meeting and Exh ibit, Reno, NV,

January 7-10, 2008

Above: Integration of computational aeroelas ticity into the design process of unmanned aerial vehicles drives innovation, avoidssurprises late in the development cycle and results in engineering success (University of Washington).



8/10/2019 Dynamics 34


C

icli Pinarello has earned a solidreputation for consistently pro-ducing winning bicycle designs,both on the road and in the velo-

drome. Recently, the Italian bike manufacturerjoi ned t he worl dwi de com mun ity of CFD, u singSAR-CCM+ to assess aerodynamic perfor-mance of their bike frames with the goal ofidentifying the trends towards increasinglymore aerodynamically efficient designs. hiswork played a critical role in the design and opti-mizati on of the our De France Dogma winning bike and will be invaluable in thedevelopment of the next generation of CicliPinarello race bikes.

A BIT OF HISTORYTe first aero bi ke (Te Espada) was producedby Cicli Pinarello in , and was specifi-cally designed for the Spanish cyclist MiguelIndurain. With this innovative bike, Indurainbeat the hour record on S eptember , atthe Bordeaux Velodrome, reaching a speed of. km/h, which was . km/h faster thanthe previous record held by the Scot GraemeObree. he year following this i mmense suc-cess, the road version of he Espada wascreated and used by Miguel Indurain in the seasons time trial s, includi ng those ofthe our de France, which he won for the fif thconsecutive year.

During the next few years, bike designers,

including Pinarello, started to introduce increas-

ingly more extreme aerodynamic concepts,

heavily involving not only the frame of the bike,

but also changing the positions of the cyclists,

now riding in a very elongated position with

their arms stretched out in front. Tis resultedin he Atlanta, a frame that was developed

and was used by Andrea Collinell i to win a

gold medal at the Atlanta Olympics in the indi-

vidual pursuit. Subsequently, a road version of

Te Atlanta was also produced and helped lead

Jan Ullrich to victory in the our de France .

he turni ng point for innovative aero bikedesign came in when, at the end of thatyear, the International Federation took thestance that the bike was becoming too impor-tant, to the extent that the merits of the athleterisked being shunted into the background. olevel the playing field, the decision was madethat bike frames (including time-trial mod-

els) should be of a traditional shape based on aregular triangle, wit h a minimum cross sectionthickness of mm and a max imum cross sec-tion length of mm. As a result of these newrules, the interest in bike aerodynamics waned.

In recent years, this interest has bouncedback with a vengeance because advance-ments in carbon technology have enabledthe concepts of aerodynamics to be extendedto bikes used in non-time-trial races (alsodefined as massed start races in the sport-ing regulations). his has opened the doorsfor the industry to apply novel CFD methodsto improve the aerodynamic efficiency of t hemassed-start bicycles.

TOUR DE

2012FRANCE

CICLIPINARELLO (TREVISO, ITALY)

PHOTOBETTINI

9

..::FEATURE ARTICLESports

8/10/2019 Dynamics 34

21/41 2

CICLI PINARELLO AND CFD

Cicli Pi narello joined the worldwide communityof CFD when aerodynamics was beginning tocome out of the realm of time trials to ventureinto the area of frames used in massed startraces, including the stages of the major tours(our de France, Giro dItalia, Vuelta de Espana,etc.) and the classics (Paris-Roubaix, Milan-San Remo, etc.).

During this time, Cicli Pinarello startedusing SAR-CCM+ to compile a databaserecording frame performance, with the goalto identify trends towards increasingly more

efficient aerodynamic designs (designs withreduced wind resistance or drag). he intro-duction of CFD into an environment that wasfairly new to these numerical simulation tech-niques required a lot of consideration whenpost-processing and visualizi ng data becausethe company needed to make sure that thenumerical results could be easily understoodby everyone and would be of valuable use tothe technicians. In addition, making the resultsfrom this R&D activity accessible and compre-hensible to everyone, even to non-experts, wasalso extremely important from a commercialand marketing point of view because clientsand fans needed to be kept informed (throughcatalogues, the Internet, e tc.) in a way that wasboth clear and intuitive.

APPROACHFor the CFD results in the database to beconsistent from frame to frame and from year

to year, baseline run conditions were defined ata speed of km /h (. m/sec) with standarda tmo s ph eric pres s u re a n d tempera tu re(,. mbar and C). Keeping theseconditions constant for all CFD simulationsresults in meaningfu l data when comparingthe aerodynamic efficiency and identifyingthe beneficial aerodynamic trends of variousframes. A typical resistance value for a cycli striding under these conditions is approximately- Newton (fra me + whee ls).In addition to defining fixed run conditions, the

The road version ofthe Espada used bythe Spaniard MiguelIndurain in the 1995Tour de France

6 767

6

7

6

7

Cross

Section

1 cm min

2.5 cm min

8cmmax

8cmmax

Above:Sporting Regulations, introduced in 1997,indicating the main rules to be followed whendesigning a racing frame

6

7

16cm max

4

2

1

7

3

6

5

configuration for each simulation was alsoclearly identified to make sure that the CFDcalculations on different f rames remainedcomparable. he configuration was fixed toinclude a typical shape of a cyclist, the wheelsand the brakes which, although not strictlypart of the frame, are important due to theinterference they generate. Each of thesecomponents (cyclist, wheels and brakes) arepresent as if they are part of the frame andmust be kept the same for all simulations in

the database, resulting in consistent relativevalues of aerodynamic efficiency. o build amore extensive and univocal data bank, thedecision was also made to maintain the sameconfiguration for both the time trial and roadrace frames. Although this may not be veryrealistic, at least in an initial phase, it wasconsidered to be more important to accu mulate alarge amount of comparable data within a si nglehomogeneous family, rather than to generatetwo subsets, one with a road configuration andthe other with a time-trial configuration.

AN ACCURATERECORD SHOWING PASTPERFORMANCE OF FRAMESWILL BE INVALUABLE INPREDICTINGPERFORMANCEOF THE NEXT GENERATIONOF CICLI PINARELLO RACEBIKES.



8/10/2019 Dynamics 34

22/41



Race bikes operate at a low Rey nolds num-ber (of the order of ) and CFD results in that

regime can be very sensitive to variations in thenumerical grid. o maintain grid consistencywhen evaluating the performance of t he frames,great care was taken in defining the method forgenerating a numerical mesh to guarantee thesame resolution and accuracy for all computa-tions in the database. A trimmed grid with wallprism layers and a k-turbulence model withwall fu nctions is used for the CFD solutions. Tethickness of the pri sm layers is kept constantthroughout the grid and consistent betweenframes. his does not always allow for an opti-mal y+ to be maintained as at times the smallerparts on the frame have dimensions of thesame order as the layers themselves. With this

approach, one cannot automatically assume ahighly accurate prediction of the absolute dragvalues but it can be successfully used to iden-tify the trends towards geometries that are toa greater or lesser extent aerodynamically effi-cient, which is the purpose of thi s work.

When presenting the CFD results, it is impor-tant to make it easy for the non-expert audienceto quickly assess the aerodynamic efficiency ofthe product. For this reason, a key was devel-oped to help interpret the data produced by CFDboth qualitative ly as well as quantitatively. oaccomplish this, reference values for aerody-namic quality were defi ned ranging from lowquality to medium quality to top quality

and geometric reference profiles were associ-ated with each of these. Wall sheer stress in thelongitudinal direction was selected as a goodindicator of aerodynamic quality and the ref-erence profiles of varying aerodynamic qualitywere selected to be a cylinder (low qualityaerodynamics), an ellipse (average quality aero-dynamics) and a NACA profile (top qualityaerodynamics). he latter must be consideredonly as the reference object and cannot be usedin practice as the sporting regulations stipulatea minimum thickness of . For visualiza-tion of the CFD results, the scale of wall shearstress is fixed so that the image of a new framedesign immed iately gives a good indication ofits aerodynamic efficiency when compared tothe image of the t hree baseline profiles (usingthe same scale). o further improve the visual-ization of the CFD data, the shear stress graphsare sometimes displayed together with the walllimiting streamlines, making the most of thefact that streamlines can be easily interpreted

by anyone (whether an expert or not) becausethey are remini scent of what you would see if oilwere spread over the surface.

For reporting performance results, theuse of numerical values is kept to a mi nimumand when numbers are used, they are usual lyreported in the form of increments or ratherthan in terms of absolute values. From thispoint of view, the accumulated drag force func-tion, which shows the development of drag ina longitudinal di rection, is an excellent way toget easy and immediate comprehension whencomparing technically precise CFD data.

VALIDATIONDimensionless values such as the drag coef-ficient (C

d), while not often used in the data

dissemination, are indispensable for the vali-dation of a CFD simulation methodology. ovalidate the numerica l approach used for popu-lating the database, th ree cases representingtypical frame cross sections were defined: anellipse with an aspect ratio of : (cylinder), anellipse with an aspect ratio of : and an ellipsewith an aspect ratio of :. A grid densit y studywas performed in each case by keeping theprism mesh constant ( layers with a st retch-ing factor of . and a thick ness of . mm) andrefining the tri mmed portion of the grid. Tree

levels of grid refinement were generated result-ing in Gr id A (coarse), Grid B (medium) and Grid C(fine) for each elliptical cross section and resultsobtained with SAR-CCM+ were compared todata available in the literature (Hoerner, FluidDynamic Drag).

At a speed of km/h, the Rey nolds numberfor each of the validation cases is in the criti-cal region where the transition from laminarto turbulent flow occurs (Reynolds numbersfor the cylinder, ellipse : and ellipse : are.x, .xand .xrespectively). Inthis transition region, the physical drag coeffi-cients significantly vary depend ing on whether

the flow remains laminar (subcritical flow) orbecomes turbulent (supercritica l flow) and as aresult, the predicted CFD drag coefficie nts wereexpected to be very sensitive to variations inthe grid densities. Drag coeff icients (based onfrontal area) computed with SAR-CCM+ on thethree grid-refined meshes ranged from: C

d_(cylinder)=.-.

Cd_(ellipse:)=.-., and

Cd_(ellipse:)=.-..

Te range reported in the literature i s: Cd_(cylinder)= .-. C

d_(ellipse:)= .-., and

Cd_(ellipse:)=.-.

confirming that al l three meshes resulted in

Above:The three reference profiles for varyingaerodynamic quality using shear stress in thelongitudinal direction as the quality indicator

LOWQUALITY

MEDIUMQUALITY

HIGHQUALITY

Above:The CFD grid refinement used for the validation cases

4GRID B 4GRID C

Above: Grid density study of validation cases comparedto results from Hoerner Fluid Dynamic Drag

From Hoemer FluidDynamic Drag

Cord/Thickness

GRID A

GRID B

GRID C

0.0

0.2

0.4

0.6

0.8

1.0

0 2 4 6 8 10

Cd(basedonfro

ntalarea)

Above:Typical mesh (consistent with Grid B) forframes in the database

4GRID A

1

8/10/2019 Dynamics 34

23/41



realistic values for Cd

. he finest mesh (Grid C)showed in the best trend but was consideredcomputationally too expensive. he coarsestmesh (Grid A) was not considered for the com-

putations because the overall trend of Cdwasnot well captured. Grid B was chosen as a rea-sonable compromise between resources andaccuracy. Because this study looks at incre-mental changes in drag between frames ratherthan absolute values, as long as the run condi-tions, meshing approach and configurationsare consistent between simulations, Grid B isexpected to predict the correct trends.

FRAME PERFORMANCERESULTShe database of cases analysed to date con-sists of four frames whic h are referred to in

this article as grey, red, blue and green.Te mesh for each of the frames was generatedusing the guidelines from Grid B obtaineddurin g validat ion studies. Effic iency of eachframe is visualized by plotting the develop-ment of drag (accumulated force f unction) from

the front wheel ax le (x .) to the rear wheelaxle (x .). Te grey frame (the first to bestudied) represents the baseline and the resultsfor all other frames are all presented as a per-

centage of this baseline data. he red frameshows a small improvement in the area of thefork head (x .) with respect to the greyframe, whereas both the green and bluesecond-generation frames show more sub-stantial improvements mainly in the area ofthe fork and the rear body. he blue frame isthe most aerodynamic of the four but becauseof rigidity reasons, this frame is not consideredas a practical design. Modifications were madeto the blue frame to make it a feasible design,resulting in the green frame.

In addition to looking at the accumulatedC

dfunction, the wall shear force proves to be

very useful to visually show the areas in which

the gains of the new frame designs are concen-trated. For example, by comparing the front ofthe grey frame with the most recent bluedesign, those who are not experts in CFD andaerodynamics can visualize major improve-ments in efficiency: blue is good, red is bad.

CONCLUSIONCicli Pinarello is using SAR-CCM+ to gener-ate a performance database of their frames toidentify and visual ize trends towards increas-ingly more efficient designs. Te configuration,meshing approach and CFD run conditions forpopulating the database were clearly defined toensure a consistent and accurate comparison ofaerodynamic performance from frame to frameand from year to year. Validation stud ies of thesimulation methodology were performed a nd aprocess was developed for quick and intuitivevisualization of the aerodynamic efficiency of

the frames. o date, the database has been pop-ulated with four generations of frames showingan increasingly i mproved performance witheach new frame design. his accurate recordshowing past performance of f rames played acrucial role in the design of the our De France winn ing bike and wil l be invaluable in pre-dicting performance of t he next generation ofCicli Pinarello race bikes. D

Above:Aerodynamics (shear stress in thelongitudinal direction) on the front fork from the(grey) base frame to the latest (blue) version.

BLUE

GREY

Above:Accumulated drag function on the frames currently existing in the database

Cicli Pinarello, founded by Giovanni Pinarello in

Treviso Italy in 1952, is one of the worlds leading bicycle

manufacturers and uses cutting edge technology and

materials to build award-winning bikes that are ridden to

race victories around the world.

PHOTOBETTINI

2

8/10/2019 Dynamics 34


W

hen Nico Rosberg won the

Chinese Grand Prix in

Shanghai, illustrations were

immediately circulating of a

strange slot on the underside of the front

wing from which a gush of ai r would blow out

jus t at the ri ght momen t to canc el o ut (or at

least reduce) the load generated by the wing

itself. This was a revised and improved

version of the McLaren F-duct from the year

before, connected to the drag reductionsystem (DRS) so as to unload both wings

(front and rear) at the same ti me and reduce

the drag generated when the DRS is open. In

addition, these slots kept the si ngle-seater

well balanced because the two wing surfaces

were unloaded simultaneously.

It seemed such an ingenious invention

that the other teams immediately asked

the FIA, in vain at first, to deem the device

illegal. In Monte Carlo, both Mercedes

c a r s o n c e a g a i n p e r f o r m e d v e r y w e l l

during qualify ing, although when Michael

Schumacher was interviewed, he did not

seem that enthusiastic about the system:

First of all, we can only use it in qualifyingand for overtaking in a race, in the designated

zones. Generally speaking, it offers a small

advantage, otherwise we wouldnt use it,

but the extent depends on the nature of

the circuit. For example, in Monte Carlo, its

practically useless. Ross Brawn, for his part,

went so far as to say that the advantage could

be evaluated to around one-tenth of a second

per lap, which, in actual fact, seems quite a lot

to us. In any case, the Mercedes experience

gives us the opportunity to address the

debate surrounding the use of holes, slots and

the like in F.

DISCOVERING SLOTSAerodynamics experts have a lways found holesand slots intriguing, and those used in racing areno exception. However, openings are generallyused to increase, rather than reduce, the loadthat a wing profile is capable of generating. Teywork because they allow air to be transferredfrom high-pressure areas to low-pressureareas, helping the tired particles, which havelost speed through friction, to follow their path.

Wing profiles were the first to demonstrate thisconcept and as a result their typical bananashapes have become more pronounced to gainfull advantage of the increase in load that theyproduce. But nothing is free in this world and

a high load value comes at the price of a highresistance. Tis is where the stroke of genius ofthe DRS comes in. Tis system, i n fact, gives thewing a variable geometry and the gap createdwhen the driver activates the device is so largethat it is no longer a slot but two separate wi ngswhich produce less vertical load and thereforeless drag. he obvious benefits of this systemcan be seen on straight-line sections, wherethe load is of no use and where a low resistance

boosts the speed to help with overtaking.But the DRS only acts on the rear wing and

so Mercedes came up with the clever idea toplay with the slots, inventing a reverse one onthe front wing (the double-DRS or DDRS) to

To win races, you not only need the geniusof the designers but also their imagination,

as in the case of using go-faster holes.

MARCOGIACHI, ASSOMOTORACINGTranslated and reprinted from original article in Paddock magaz ine

..::FEATURE ARTICLEMotorsport

8/10/2019 Dynamics 34


reduce the load rather than increase it, thussynchronising both wings: they are loaded andunloaded together, so that the car is alwaysbalanced. he most important requirementfor the DDRS to succeed is to feed the frontslot at the same time as the DRS is opened. In, the driver was still allowed to open theDRS anywhere on the track during qualifying ,which enabled him to take full advantage ofthe devices benefits; in competition, however,its usage was restricted to the designatedovertaking zones, where the interferencecaused by the wake from the front vehiclelimited its effectiveness. he benefits werethen only measured in terms of resistance wit hrespect to a car whose front wing had no spec ialdevices.

Te Ross Brawn/Aldo Costa duo believed inthe solution from the onset and both carriedthe development of the idea right through tothe end. As far as we could see, the systemconsisted of a long tube, to cm in diameter,which ran along the entire length of the vehicle,connecting an opening on the rear-wing

system, we at Paddock were fascinated bythis success and wanted to try out our ownslot (an original, not the same as Mercedes, ofwhich there is insufficient detail), with the help

of numerical simulation and Lucia Sclafani, aC D - a d a p c o e x p e r t i n t h e s e c a l c u l a t i o ntechniques. O ur tool of choice was SAR-CCM+,CD-adapco's flagship CAE software, which iswidely used for aerodynamic design acrossthe motorsport industry. In a way, we werefollowing the same procedure as the designengineer from an F team who has seen the ideaon a rival car and wants to assess what effect itwill have on his own car before constructing aprototype. His approach would be to performsimulations with CFD to check out and developthe design, just as we did ourselves.

Te Paddock virtual Formula car comparedwell with the actual Mercedes design. At

first glance, the air flows from the numericalsimulations appeared complex enough togive us a headache, but clearly indicated anadvantage: with the jet of air coming out ofthe slot, the resistance of the front wing waseffectively reduced by . and the verticalload by as much as . When looking closerat the details of the analysis, however, ourenthusiasm suddenly began to wane.

In aerodynamics, modifications almostalways result in multiple conflicting effectson the performance of the system. heseeffects are usually geometry-dependent andas a consequence, the innovative conceptso b s e r v e d i n t h e p i t l a n e a r e n o t e a s i l y

transferable from one vehicle to the next. Forexample, on the Paddock Formula car, thevariation in load on the front wing was just oneof the effects produced by the air flowing outof the slot. he front wheel was also affectedby the modification, with its resistance beingincreased by a s much as .(!), whereas therear wheel seemed . less resistant. Te effectof the slot on the remaining par ts (body and rearwing) of the car was insignificant, resultingin an overall drag reduction of ., which isa considerable improvement. If we investedsome more time and managed to eliminate thenegative effect on the wheels, our car would

record a reduction of around one tenth of asecond in the time it took to travel one kilometrein a straight line. Furthermore, supplying thedriver with a balancing effect could give him

more confidence to keep his foot down on theaccelerator for a longer time, which would resultin additional benefits.

All of this is possible with our in-housesystem. Mercedes ha s no doubt created a moreenhanced design and positioned the slot at a dif-ferent point on the surface of the wing in orderto achieve optimum results. But in our case, is itreally worth the effort? Tis is where the differ-ence lies between the well-organized, winning

Above: When the DRS is closed (top), a standardslot is created, but when it is open (bot tom), the slot

disappears completely: the main plane and the flap

become two separate wings and the load generated

(and as a result the drag) drop.

endplate to the front wing. Te hole on the rearendplate was blocked by a flap which was onlyopened when the driver activated the DRS. Atthat point, air entered the hole and almostinstantaneously flew out through the slot onthe underside of the front wing.

In addition to its benefits in terms of lowerresistance and increased balance, the DDRScould enable the use of a smaller, more flexiblefront wing, which could be placed closer to

the ground at low speed and still generatethe maximum load at high speed. his wouldavoid the risk of overstressing the front wingsextended flaps, which have not been designedto withstand the higher load generated atmaximum speed. T is is a hypothesis proposedby Gary Anderson (the renowned designer ofthe s); somewhat complicated, but in themagical world of F, nothing is impossible.

SIMULATING SLOTSAlthough the F IA technical regulations for outlawed the Mercedes-style double-DRS

1The FIA ruled that as of the 2013 season, the DRS system should only be used in designated overtaking zones during practice, qualifying as well as competition.It also banned the use of Mercedes-style double-DRS, unless it is passive, such as the Lotus device.


Above: The wing on Michele Alboretos Ferrari 126 C4 in1984, showing that Mauro Forghieri had very clear ideas onhow to best use slots for aerodynamic purposes. After this,

slots immediately started appearing on wing prof iles and

were a complete success on all levels. The load instantlydoubled without adding any penalty, e.g. increase in

weight or cost.

Above:A slot that produces a je t of air perpendicular to the surface of the wing is a bad slot because it tends to push the f luidthreads away from the profile and causes the wing to stall; exactly the opposite of what is achieved by all the conventional(good) slots, which, instead, drag the air particles to keep them attached to the sur face of the profile.

8/10/2019 Dynamics 34

26/415



Above, left & right:In

2008, Ferrari opened

up a hole on the nose

to connect the bottom

to the top. With the

nose raised, the car

offers its underside

to the air. This exerts

pressure and tendsto lift the nose, partly

thwarting the action of

the wing. By contrast,

Brawn GP, the 2009

winning car driven by

Jenson Button had an

opening on the surface

of the rear diffuser in

order to increase the

local airflow section.

Above:In a very curved wing profile, the air flowing

over the lower surface receives a real thrust from

being blown out through the slot.

maybe joined together by a piece of hydrau-lic hose, fixed to the outside of the chassiswith something as ordinary as adhesive tape.hen you would go to the track to try it out.But times have cha nged: nowadays wings are

no longer made of aluminium, no private testson the track are al lowed, and even the F teamshave their own bureaucratic procedures to fol-low. Tis explains why certa in ideas, even oncethey have appeared and have been shown to bevalid, are not always implemented by everyone.Whoever thinks of it first, and works on it allwinter long, will always have the advantage.

HOW IT ALL STARTEDhe idea conceived by the Ross Brawn/AldoCosta duo is not the first occurrence of a holeor magic slot, although up until , the "holegang" of racing car desig ners had not producedmuch more than slots on profiles. Ten, in ,

teams and those with fewer resources, whomay not have the time or money to explorethe possibilities. At least eight to ten wings areneeded (including spa res) for two single-seatercars and new moulds must be made for both the

front and back wings, on which a channellingmust be created to transfer the air. his is nomean feat in the midd le of a championship, par-ticularly if resources for the calculations andplanning need to be reallocated to new projects.Furthermore, it is always good practice to vali-date the CFD solutions in a wind tunnel, whichwould involve building a scale model especiallyfor this purpose. Finally, the balancing effectthat involves the drivers feeling for the carmust be evaluated in the simulator. Anothertwo or three days work!

Everything used to be much easier. Youwent into the office with a hand-drawn sketchand, with a bit of metal plate and a bodywork

expert, the new wings would be ready in a day,

DALESSIO

DALESSIO

8/10/2019 Dynamics 34

27/41 2



Above:Maps showing the p ressure under the nose and the front wing in STAR-CCM+. The blue indicates lowpressure, the green a neutral area and the red shows high pressure. With the slot closed (left), the area of lowpressure is significantly wider. With the slot open (right), even the side skirt has a different behaviour. With thisinformation available, if we had been in charge of a team, we might have considered trying out the idea.

Above: Our slot under the front wing also has adisruptive effect on the aerodynamics of the wheels,which are completely immersed in the wake of thewing.

Ferrari opened up a hole on the nose of theircar to connect the bottom to the top. With the

nose raised, in fact, the car offers its undersideto the air upfront and the air inevitably exertspressure in that area, which tends to liftthe nose, partly thwarting the action of thewing. his idea was not new as it had beenused previously by Giacomo Caliri, at Ferrariin the 's/'s, on their PB prototype.

In , the regulations changed and one ofthe most brilliant designers of the time (RossBrawn, again) focussed his attention on the reardiffusor as a possible site for integrating holes.But there was little physics behind his famousdouble diffusor; it consisted of an opening toexpand the airflow section on the rear diffusora fter th e n ew tech n ica l reg u la tio n s h a d

introduced very strict li mits to the geometry ofthis part of the chassis. Tus the Brawn GP carwas born, driven by Jenson Button, who wasWorld Champion that year. Its rear is not thatvisible on television, but we can a ssure you thatunder those diffusors was a gap big enough topass a cat through! Up until this point, as far asthe wonderful world of racing cars is concerned,the engineers had not come up with anythi ngthat was not already known and slots and holeshad served a fairly conventional purpose inthat they helped the a ir to do its work: generatemore load. Te first McLaren F-duct was used on the rearwing i n and the slot was operated directly

by the driver who, through his movementsinside the cockpit, could open and close thehole feeding the slot on the rear wing. Withthis approach, the slots started being usedto actually remove the load on the wing. hisinitial version of the F-duct soon becameredundant when, for safety reasons, thedriver was prohibited from blocking holes andslots of any nature with his own movements.he subsequent introduction of the DRS putthe final nail in the coffin for McLarens idea.

But by now, the idea of slots being used toreduce the load generated by a wing had beenborn and there was no turning back. Havingalready gained control over the rear wing with

the DRS, it seemed only natural to transfer the

concept to the front wing, drawing the air offthe rear wing and conveying it under the frontwing to reduce its capacity to generate load(and drag). And this bri ngs us to Mercedes idea,which enabled Nico Rosberg to join the select

group of drivers who have won at least oneGrand Prix and helped Michael Schumacher tosavour once again the joys of the pole positionwhen perhaps he was no longer expecting it. D

Above:Paddocks virtual Formula 1 car comparedwith Rosbergs Mercedes. We are more or less there.

This article was originally published in the

Ital ian ma gaz ine Pad do ck. Sin ce it was first

created in 1992, Paddocks main focus has been

on motorsport and more particularly on the

technical aspects of Formula 1. In 2010, Paddock

star ted to use STAR -CCM+ to exp lain to thei r

readers what the engineers do to design their cars

to go faster and faster.

% variation in resistance

Above: It is always very difficult, almost impossible,

for a modification to have an impact on only one areaof the bodywork. Above all, the modifications to the

front part of the vehicle affect the whole of the chassis

and often the effects do not go in the same direction,

as in this case. When all the positive and negative

effects are taken into consideration, there remains

a drag reduction with the slot open of 1.4%, which is

not insignificant. If there were no collateral effects, the

front wing alone would register a drag reduction of

2.4%. Only the rear wing seems to be unaffected by the

existence of the slot.

frontwing

rearwing

body frontwheel

rearwheel

8/10/2019 Dynamics 34

28/41

his is what the team OndaSolare, based in Castel SanPietro Terme, Italy, did whenthey decided to take part in theWorld Solar Challenge. The

challenge, which is staged every two years,involves solar-powered vehicles racing for days across the , km separating Darw in

from Adelaide on the Australia n continent.After first participating in the Challenge in (Figure ), Onda

Documents

Dynamics 34