ADVANTAGE

TM

EXCELLENCE IN ENGINEERING SIMULATION

VOLUME III ISSUE 2 2009

SPOTLIGHT ON
THE BUSINESS VALUE
OF ENGINEERING
SIMULATION
LEVERAGING ADDING DRIVING PRODUCT
SIMULATION VALUE DEVELOPMENT
PAGE 6 PAGE 12 PAGE 14
 EDITORIAL

Who Will Survive?

Engineering simulation may help determine which companies
make it through the current economy and strengthen their
competitiveness when markets rebound.

Engineering simulation is an indispensable tool in 11,000 hours of usage on a single product from ANSYS in a
efficient product development processes at a growing typical year. Tier-one mechatronic system supplier Brose
number of companies. Engineers use the technology early Group had one simulation engineer in the 1990s and now
in the cycle to evaluate concepts, compare alternatives, has 45, with CAE usage growing by 50 percent annually.
identify problems and optimize designs. Upfront analysis That organization applies engineering simulation in
avoids the slowdowns and expenses of late-stage problem- developing higher-quality, lower-cost automotive
solving with frantic design fixes and trial-and-error testing. door and closure systems using tools such as ANSYS
These are among the powerful capabilities that enable multiphysics technology.
companies to reduce costs, shorten time to market, Gas turbine component supplier Power Systems
improve quality and create innovative designs. Manufacturing helps power-generation companies avoid
The ramifications of such benefits can be tremendous in expensive downtime using ANSYS Workbench based
terms of top-line revenue growth and bottom-line savings, technology to optimize the design of compressor blades.
and the increased profitability that companies reap is the Nozzle manufacturer Spraying Systems Co. strengthens its
staggering business value of engineering simulation, which relationships with customers and provides an additional
can provide the impetus for executive-level decisions to source of revenue by using software from ANSYS to study
invest in the technology. This is the theme of this issue’s and suggest improvements in the designs of customers’
Spotlight section. gas conditioning solutions that utilize nozzles in complex
Coverage of this topic is especially timely, given the pollution control systems.
importance of this business value to companies around Clearly, these companies recognize that they’re not
the world contending with continuing economic distress, just analyzing parts but are building customer relation-
financial uncertainty and volatile markets. Indeed, the ships, creating value-added services, growing revenue
competitive advantage provided by smart use of engineering streams and boosting their competitiveness. That’s the
simulation can be a deciding factor in determining which true business value of engineering simulation, which
companies survive the current economic chaos. As the NAFEMS Chief Executive Tim Morris aptly describes as a
articles in this section show, there is no cookie-cutter strategic weapon. In his article “Championing Simulation,”
approach to engineering simulation. Because of the wide he notes that best-in-class companies are often those that
range of corporate priorities and product portfolios, the make the greatest use of the technology. Investments in
ways in which technology is implemented and business simulation make companies more competitive, allowing
values are obtained are unique for each company. businesses to emerge from the current recession stronger
For example, mobile electronics and transportation — and probably with fewer competitors. ■
system supplier Delphi Electronics and Safety Systems
develops more-robust, reliable products and greatly
reduces validation failures in prototype testing through
a comprehensive program to train engineers at distributed
sites around the world in upfront simulation, logging in over John Krouse, Senior Editor and Industry Analyst

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 1

Table of Contents
SPOTLIGHT ON THE BUSINESS VALUE OF ENGINEERING SIMULATION
5 OVERVIEW
Profiting from the Investment in Smart Engineering Simulation
6 AUTOMOTIVE/ELECTRONICS
Leveraging Upfront Simulation in a Global Enterprise
Corporate initiative at Delphi focuses on the benefits of simulation as an integral
part of early product design at sites around the world.
6
8 TURBOMACHINERY
Designing for Quality
Simulation paired with optimization helps eliminate compressor blade failure.

10 ADVOCACY
Championing Simulation
NAFEMS champions CAE awareness, delivers education and sets simulation standards.

12 ENVIRONMENTAL
Extending the Bounds of Customer Service
Spray nozzle manufacturer expands value-added services by using simulation
to develop and validate gas conditioning solutions in complex pollution
control systems. 12

14 AUTOMOTIVE
Opening New Doors
Brose uses simulation to drive product quality, reduce testing and minimize costs.
14

SIMULATION@WORK
16 BUILT ENVIRONMENT
ANSYS Sets the Stage
Simulation was used to design the floating stage set used in the latest
Bond movie.

19 ENERGY
Harnessing the Power of Ocean Waves
Engineers use structural and hydrodynamic analysis to ensure that wave-
powered electrical generation machines produce maximum energy output
and operate effectively for decades.

22 MARINE 16
Designing for Strength, Speed and Luxury
Simulation software from ANSYS helps a yacht designer deliver the optimal
combination of luxury and performance.

24 ELECTRONICS
Seeing the Future of Channel Design
NVIDIA uses VerifEye and QuickEye as an extension to traditional SPICE-level
simulation approach to design high-performance graphics solutions.
22

2 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 CONTENTS

28 ELECTRONICS
Hot Topics: High-Capacity Hard Disks
Samsung uses simulation to improve thermo-fluidic performance of
hard disk drives.

30 AUTOMOTIVE 28
Fatigued by Stress Limitations
The combination of fe-safe and ANSYS software helps Cummins improve
life prediction accuracy.

32 ELECTRICAL
Managing Heat with Multiphysics
Multiphysics simulation helps a global company design better electrical products.

34 AEROSPACE
Up, Up and Away
Simulation-driven innovation delivers a new ejection seat design
for a military aircraft in less than 14 months. 32

36 MARINE
Propelling a More Efficient Fleet
Rolls-Royce uses simulation for propeller design to reduce
marine fuel consumption. 36

DEPARTMENTS
38 ANALYSIS TOOLS
Staying Cool with ANSYS Icepak
Thermal management solution predicts air flow and heat transfer in electronic 38
designs so engineers can protect heat-sensitive components.

40 Analyzing Vibration with Acoustic–Structural Coupling

FSI techniques using acoustic elements efficiently compute natural frequencies,
harmonic response and other vibration effects in structures containing fluids.

44 PARTNERS
Integrating CAE Tools: a Package Deal
Moldex3D and ANSYS Mechanical team up to simulate microchip encapsulation.

47 ACADEMIC 44
Sailing Past a Billion
Racing yacht design researchers push flow simulation past a meshing milestone.

50 TIPS AND TRICKS

Optimizing Options
Technologies converge in ANSYS Workbench for parametric fluid structure interaction analysis.

52 Analyzing Nonlinear Contact

Convenient tools help analyze problems in which the contacting area between
touching parts changes during the load history.

47

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 3

CONTENTS

WEB EXCLUSIVES
These additional articles are available exclusively on www.ansys.com/exclusives/209. For ANSYS, Inc. sales information, call
1.866.267.9724, or visit www.ansys.com.
ANALYSIS TOOLS For address changes, contact
The Immersed Boundary Approach to Fluid Flow Simulation AdvantageAddressChange@ansys.com.
The add-on immersed boundary module, jointly developed by ANSYS and Cascade Technologies, To subscribe to ANSYS Advantage,
is a preliminary design analysis tool that dramatically reduces the amount of time needed for go to www.ansys.com/subscribe.
fluid flow simulations and provides fast results by directly addressing the challenges Email the editorial staff at
associated with this meshing step. ansys-advantage@ansys.com.

ENERGY Executive Editor

Tracking Down Vibrations Fast with FSI Fran Hensler
Offshore oil and gas operations can lose significant revenue from even a few days of down-
Managing Editor
time, so they must efficiently study and rectify equipment failures that could shut down any Chris Reeves
part of the platform. With high-speed iterations between mechanical and fluids software,
Denmark-based Lloyd’s Register ODS used fluid structure interaction to quickly pinpoint the Senior Editor and Industry Analyst
John Krouse
cause of damaging vibrations and to assess new designs.
AUTOMOTIVE Art Director
Dan Hart
Designing Giants for Tough Work
In the mining industry, the engineering challenge is to design lightweight trucks strong enough Editors
to withstand harsh operating conditions. Liebherr engineers rely on ANSYS structural simulation Erik Ferguson
Shane Moeykens
technology to develop giant diesel electric trucks designed to withstand hostile conditions while Mark Ravenstahl
providing maximum load capacity.
Ad Sales Manager
ENVIRONMENTAL Helen Renshaw
Keeping New Orleans Flooding at Bay
In flood-prone coastal Louisiana, structural analysis of massive platforms holding storm Editorial Advisor
Kelly Wall
drainage equipment helped to predict and address vibration problems. Mechanical Solutions,
Inc. performed the simulation, the results of which were actually validated during a subsequent Designer
Category 2 hurricane. Miller Creative Group

SPORTS/AUTOMOTIVE Circulation Manager

Sharon Everts
Overtaking Race Car Design
Italian design firm Fioravanti has proposed an innovative Formula 1 car that could bring more
excitement to the race. Using aerodynamic simulation, the company designed a car that would About the Spotlight
allow drivers to more frequently overtake the other racers — safely. Upfront engineering simulation may
help determine which companies make
MARINE it through the current economy and
strengthen their competitiveness when
Designing Safe and Reliable Ships markets rebound. The spotlight begins
Delta Marine Engineering uses upfront simulation to minimize risk and maximize performance on page 5.
well before launch day. The Turkish shipbuilder identifies and corrects troublesome vibration
problems to comply with international standards as well as to ensure the safety of the crew Neither ANSYS, Inc. nor the senior editor nor Miller Creative
and a long life for the ship. Group guarantees or warrants accuracy or completeness of
the material contained in this publication.
ANSYS, ANSYS Workbench, Ansoft Designer, CFX,
PARTNERS AUTODYN, FLUENT, GAMBIT, POLYFLOW, Airpak,
Enabling Detailed Chemistry DesignSpace, FIDAP, Flotran, Iceboard, Icechip, Icemax,
Icepak, FloWizard, FLOWLAB, G/Turbo, MixSim, Nexxim, Q3D
CHEMKIN-CFD is an advanced chemistry simulation technology that efficiently and robustly Extractor, Maxwell, Simplorer, Mechanical, Professional,
couples accurate chemical kinetics to flow simulations. Its developer, Reaction Design, Structural, DesignModeler, TGrid, AI*Environment, ASAS,
AQWA, AutoReaGas, Blademodeler, DesignXplorer, Drop
provides the software free to ANSYS FLUENT users who are looking to improve the accuracy
Test, ED, Engineering Knowledge Manager, Emag, Fatigue,
of their analyses. Icepro, Icewave, Mesh Morpher, ParaMesh, TAS, TASSTRESS,
TASFET, TurboGrid, Vista, VT Accelerator, CADOE, CoolSim,
SIwave, Turbo Package Analyzer, RMxprt, PExprt, HFSS,
Full-Wave SPICE, VerifEye, QuickEye, Optimetrics, TPA,
AnsoftLinks, ePhysics, Simulation Driven Product
Development, Smart Engineering Simulation and any and all
ANSYS, Inc. brand, product, service, and feature names, logos
and slogans are registered trademarks or trademarks of
ANSYS, Inc. or its subsidiaries located in the United States or
other countries. ICEM CFD is a trademark licensed by ANSYS,
Inc. All other brand, product, service and feature names or
trademarks are the property of their respective owners.

ANSYS, Inc.
Southpointe
275 Technology Drive
Canonsburg, PA 15317
U.S.A.

© 2009 ANSYS, Inc. All rights reserved.

4 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

Spotlight on the
Business Value of
Engineering Simulation
Profiting from the Investment in Smart Engineering Simulation
To survive and profit in the current demanding business Capturing and reusing engineering data generated from
environment, organizations must engineer high-quality and numerous design iterations and increasing types of sim-
innovative products, design and manufacture them for the ulation help to maintain product design efficiency and protect
lowest possible cost, and then beat their competitors to the intellectual property. Managing engineering knowledge is
global marketplace. To meet that challenge, many of the most vital for individual users, small teams and enterprise-wide
innovative companies in the world use engineering simulation implementations.
to develop products that dominate the marketplace. Engineering teams and individual designers can be even
There is no single solution for every organization, more productive when using simulation technology with an
although one basic principle is universal: Engineering sim- adaptive architecture, which gives them the capability to
ulation is performed up front in the product development iterate with their existing CAD files, incorporate other
process at companies large and small because Simulation software tools, explore what-if scenarios and develop
Driven Product Development provides quantifiable value. reusable models all within the same working environment.
Tools used to test many alternatives for product designs Not only does this speed up the product development
before prototyping provide engineering teams with the process, it allows team members to apply their experience
ability to get products to market quicker. It also allows these and skills in the most efficient manner and generates more
teams to rapidly iterate on designs so they can determine opportunities to develop the innovative products required
the best and most innovative alternatives. In addition, for continued profitability.
reducing the number of expensive prototypes provides The world’s most successful companies turn to sim-
direct cost-savings benefits. ulation solutions from ANSYS, whose products are used by
As simulation technology continues to expand in depth 16 of the top 20 most innovative companies in the world
and breadth, companies are able to simulate a greater range of today, according to a BusinessWeek report prepared by
their product requirements. Multiphysics simulation delivers the Boston Consulting Group and by 97 of the top 100
results that approach real-world conditions, supplying an even industrial companies on the FORTUNE Global 500 list.
more reliable basis for making product design decisions. Today, more than ever, these companies have invested in
However, these larger solutions may come at the price of Smart Engineering Simulation to increase the efficiency of
computational time. High-performance computing continues their processes, improve the accuracy of their virtual
to address this problem, making it possible to solve complex prototypes, and capture and reuse simulation processes
problems on a laptop, which, several years ago, would have and data. ANSYS delivers solutions to help organizations
been impossible using an entire room of processors. profit in today’s turbulent economic environment. ■

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 5

AUTOMOTIVE/ELECTRONICS

Leveraging Upfront
Simulation in a
Global Enterprise
Corporate initiative at Delphi focuses on the benefits of simulation
as an integral part of early product design at sites around the world.
By Fereydoon Dadkhah, Senior Engineer, Mechanical Analysis and Simulation,
Delphi Electronics & Safety Systems, Indiana, U.S.A.

Most high-technology com- transportation systems for the automotive and consumer
panies now realize the potential product industries.
benefits of simulating the perform- Beginning in the late 1990s, Delphi Electronics & Safety
ance of their products using tools embarked on a program to take full advantage of FEA in the
such as finite element analysis product development process. Along with other companies,
(FEA). They also clearly know that Delphi Electronics & Safety had been using FEA in a
performing analysis early in the more limited way as a troubleshooting tool often later
design cycle has the potential to in development. The new initiative intended to employ
identify and solve design problems finite element analysis as an integral part of the product
much more efficiently and cost development process — especially focusing on the use of
Fereydoon Dadkhah, Delphi effectively compared to handling simulation up front in the design cycle.
Electronics & Safety Systems them later. One of the leading To achieve this goal, Delphi put into place a compre-
companies in employing upfront hensive program to train design engineers in the use of
analysis throughout the product engineering organization is FEA in the early stages of the design process. This program
Delphi Electronics & Safety Systems — a major division of began by classifying engineers according to their skill levels
Delphi Corporation specializing in mobile electronics and in use of FEA and interpretation of analysis results. Gradually,

Steady-State Thermal Analysis

The most common use of the ANSYS Workbench tool at Delphi is by design
engineers engaged in product development. Analysis types include steady-state
thermal, free vibration and linear static stress analysis. More advanced types of
analysis, including those involving material or geometric nonlinearity, transient
loading, fluid flow and multiphysics, are performed by full-time analysts using
products from ANSYS or other commercially available analysis tools.
Delphi produces a number of products including those for use in the automotive
and consumer product sectors that must meet stringent thermal requirements. A
steady-state thermal analysis is, in many cases, the first step in ensuring that the
final product will meet the thermal requirements of the customer. Based on usage
data collected annually, the ANSYS DesignSpace tool is widely used to perform this
type of analysis. Shown in the accompanying figure is an example of the results of
Steady-state thermal analysis
a steady-state thermal analysis of an integrated circuit package used in a trans- of an IC package used in a
mission controller unit. Once the steady-state performance is established, transient transmission controller unit
and system-level analyses are performed to completely characterize the system.

6 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 AUTOMOTIVE/ELECTRONICS

Germany (1)
India (130) Luxembourg (711)
Unknown (56) Mexico (2,773) Poland (3) Development Process (PDP) as a requirement. The PDP
begins with the concept stage and proceeds to the validation
Singapore
(380) stage when prototypes are built and tested, and finally the
program is handed off to manufacturing. This has led to
developing much more robust and reliable products as well as
greatly reducing or eliminating validation failures. This process
is enforced by a Design Failure Modes and Effects Analysis
(DFMEA) plan represented by a spreadsheet of possible
failure modes for a product and the required analyses to
show that the product is immune to the specific failures.
U.S. (19,806) U.K. (8,697)
A large number of engineers around the world in the over-
all Delphi organization use these tools, including the full suite
of software from ANSYS within the ANSYS Workbench inter-
face, to perform thermal, stress, vibration and other general
analysis in the course of product development. In 2007, the
number of ANSYS DesignSpace users exceeded 200, and
approximately 30 percent were Delphi Electronics & Safety
engineers. Delphi Electronics & Safety users globally logged
11,151 hours of usage on the software, or 34 percent of
Number of hours of ANSYS DesignSpace usage at various Delphi sites internationally the total for all sites internationally. The licenses for
many CAE tools — including ANSYS products such as
the program incorporated use of FEA into the Delphi ANSYS DesignSpace — are supported from servers in
Electronics & Safety product development plan. Safeguards Michigan, in the United States, allowing engineering
such as peer reviews, engineering fundamentals training management to stay up to date on the use of these tools.
and mentoring were implemented to ensure proper use By adopting a comprehensive approach for implementing
of FEA. Furthermore, Delphi Electronics & Safety has FEA across the worldwide organization, Delphi has
restricted use of this technology to engineers and scientists effectively incorporated an extremely powerful technology
with a minimum of a bachelor’s degree. Training in the use into the product development process. The initiative to focus
of the structural mechanics simulation software — in this on upfront analysis in particular has resulted in outstanding
case, ANSYS DesignSpace that uses the ANSYS business value for Delphi in terms of improved designs devel-
Workbench platform — is a prerequisite at Delphi Electronics oped very efficiently. The use of the ANSYS Workbench
& Safety. Occasional users such as product engineers utilize platform has certainly facilitated this process by providing the
the software to perform linear and static analyses. ability to perform a variety of analysis types of different com-
More advanced analyses involving nonlinearity or transient plexities in the same familiar environment. Perhaps the best
loading are referred to full-time analysts. indicator of the effectiveness of this software in a business
Today, the company has incorporated the use of context is management support for its widespread use by
structural mechanics simulation into the Delphi Product such large numbers of Delphi engineers around the world. ■

Finding Natural Vibration Modes

ANSYS DesignSpace software is often used
for determination of the natural modes of vibra-
tion of a system. Many of Delphi’s products are
used by the automotive industry, and the first
step in establishing that a product can be used
in a vehicle is to ensure that the first few modes
of vibration of the product are beyond the
minimum values that can be excited by the vehicle’s
operation. The accompanying figure shows the first mode of vibration
for a bracket used to support an airbag control unit.
Using the ANSYS DesignSpace tool to perform modal analysis, product
engineers are able to determine if any changes to the initial design are First mode of vibration for bracket that
needed to improve the vibration characteristics of the system. The design supports an airbag control unit
then proceeds to the next stage, in which harmonic and power spectral
density (PSD) analyses are performed and any required changes are made.

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 7

TURBOMACHINERY

CAD model of redesigned

compressor blade

Designing for Quality

Simulation paired with optimization helps eliminate
compressor blade failure.

Power Systems Manufacturing (PSM) is a global blade. In addition, PSM engineers determined that the
provider of aftermarket gas turbine components in the blade design had several vibratory modes that were excited
industrial power generation industry. The company’s product during engine operation that when coupled with the static
line includes stationary and rotating airfoil components, low- stresses could result in failure. PSM’s analyses predicted
emission combustion systems, and advanced components
for GE Frames 6, 7 and 9 and Siemens 501F-class
machines. In one specific redesign case, an original
equipment manufacturer (OEM) first-stage compressor
blade design for a popular large gas turbine used for
generating electricity was not reaching its expected life.
Field failure scenarios included blade breakage resulting
in considerable downstream damage. Companies running
these engines then had to shut them down for lengthy
periods while making repairs, losing the revenue from the
a
electricity that the engines would normally generate.
“Several of our customers were running into the same
problem with this compressor blade and asked us if we
could improve and fix the issue,” said Page Strohl, former
lead structures engineer for PSM. “Any new design that we
developed had to fit into exactly the same envelope and to
have the same aerodynamics as the original blades.”
PSM designers and engineers worked together to
model and simulate the original blade design, taking into b

account the aerodynamic and centrifugal loads. Using Closeup of the maximum principal stress on the pressure side of (a) the original
simulation, PSM located the highest static stresses in the equipment manufacturer’s blade and (b) the redesigned blade

8 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 TURBOMACHINERY

that the blade would fail in the exact locations at which optimal design, the aerodynamics team made minor
failures occurred in the field. design modifications in order to ensure that the design
Using the ANSYS DesignXplorer tool to study a range of performed aerodynamically as required.
design variations, Strohl defined the blade parameters he The resulting design reduced all peak steady
wanted to vary, assigned acceptable parameter ranges, and stresses. The vibrational issues were eliminated as well with
identified the variables he wanted to optimize, which in this these design modifications. The blades are performing as
case were blade natural frequencies and peak steady expected and have been in operation since May 2008.
stresses at several locations on the blade. Based on the Additional sets are now on order.
peak stress locations, design space definitions were Strohl concluded, “Key to these improvements was the
created. ANSYS DesignXplorer software redefined the CAD ability of the ANSYS Workbench platform to interface
model and generated the series of design variations needed with our CAD system, allowing us to quickly prepare
to carry out experiments for the entire range of parameters. new geometries for analysis and to control the CAD
Strohl then used the ANSYS Workbench environment to system to explore a design space. We were able to
mesh each of these designs, solve the models, capture the quickly iterate to a design that was optimal, all while
results and perform the statistical analyses needed to maintaining the same aerodynamic properties as the
identify the optimal design. Once Strohl identified an original design.” ■

Transitioning into the ANSYS Workbench Environment

ANSYS Advantage editors talked to J. Page Strohl, former Lead Structures Engineer for Power Systems Manufacturing in Florida, U.S.A.,
about his challenges in redesigning blades for turbomachinery.

You’d been a long-time user of the traditional ANSYS Mechanical Has using ANSYS Workbench changed how you approach some
interface. Did you have any hesitance transitioning to the ANSYS of your projects?
Workbench platform? I find that I look for projects that can take advantage of
I found it hard to put down something that I was already the ease-of-use benefits that ANSYS Workbench offers. PSM
comfortable and proficient with in order to start using some- recently started a major redesign of another compressor air-
thing completely new. It wasn’t until I attended an “Intro to foil, and we jumped right into ANSYS Workbench and
ANSYS Workbench” training that I attempted to really utilize it. ANSYS DesignXplorer for it. I created the baseline model
When this compressor blade analysis came up and we and showed the aerodynamicist how to duplicate it,
were faced with a time crunch, I knew I could save time by regenerate a modified Pro/ENGINEER model, then solve and
taking advantage of the connectivity to Pro/ENGINEER® check the results. I actually gave my work away.
and the fact that the actual component settings stay with
the model as it is passed from the CAD environment to How did this new process work?
ANSYS Workbench. By using this approach, the long lead It was very easy to perform the simulation — a great
tasks became the CAD modeling and the aerodynamic benefit of ANSYS Workbench. If the geometry had
analysis efforts, not the structural analysis-specific ones. problems reading in or didn’t pass all of the built-in checks,
I would slide my chair over to the aerodynamicist’s
workstation and take a minute or so and fix things. It was
really a benefit in that I was able to do other work while we
were in the iterative phase of the design. I, like many, am
workload-challenged these days, and ANSYS Workbench
helped to relieve me of that particular task.
From a larger project perspective, it gave the aero guy
a good look at all the items that needed to be reviewed
outside of his aero world. Several of my coworkers joke with
me that my job consists purely of clicking the mouse button
one or two times for an analysis job; they say that the
hard ones are when I have to click the mouse three times.
The real trick in this case was setting up the original baseline
model. Once I did that correctly, it was a piece of cake
a b to have someone else turn the analysis around with
ANSYS Workbench.
Geometry of original equipment design (a) and modified geometry (b), as created using
ANSYS DesignXplorer software

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 9

ADVOCACY

Championing Simulation
NAFEMS champions CAE awareness, delivers education
and sets simulation standards.

NAFEMS, founded more than a Is the manufacturing industry taking full advantage of engineering
quarter-century ago, is an impartial simulation technology?
best-practice champion of computer- The power and scope of simulation technology has
aided engineering (CAE) standards. increased dramatically in the past 10 years. Simulation
A nonprofit organization head- should now be at the heart of the design process, driving it,
quartered in the United Kingdom, not merely validating it in the latter stages. To do this
NAFEMS provides information to requires change in how simulation technology is deployed
secure the best returns on invest- in many organizations. Simulation engineers need to be
ment in CAE software, to develop better integrated as a fully involved part of the product
Tim Morris, Chief Executive and enhance simulation capabilities, development strategy from the very beginning. They also
of NAFEMS and to ensure the safest and most need to understand the commercial imperatives driving
effective use of the software. About development. Product development managers need to
940 companies around the world, from large multinational better understand what engineering simulation can offer, in
corporations to small engineering consultancy firms, are areas such as shaping external design appearance, instead
members of NAFEMS, and this number is growing. of leaving this to marketing designers. By embedding
Although the largest proportion of members is involved in engineering simulation in the product development strategy,
finite element analysis, the computational fluid dynamics technical products that meet all market needs can be
(CFD) group is expanding rapidly. Members belong to realized, and engineering simulation can deliver best value
almost every industry sector. by increasingly compressing development processes in
ANSYS Advantage staff interviewed Tim Morris, chief order to reduce time to market.
executive of NAFEMS, on his viewpoint about trends in Engineering simulation is a strategic weapon inside
CAE and the value of engineering simulation. companies today, especially for nimble organizations
that have a philosophy of core adoption and deployment

10 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 ADVOCACY

because it leads to competitive advantage. Financial and might mean lighter, cheaper or stronger. While companies
commercial pressures from an ever-more competitive might need to cut back on manufacturing in these times of
market have led to companies’ increasing reliance on recession, the smarter companies will not cut back too much
engineering simulation to cut costs and reduce design on design or research but, rather, will use the opportunity to
and development cycles. improve both products and design processes.
Continued investment in simulation will continue to bring
How important is a multiphysics approach to the development of rewards in terms of making companies more competitive
engineering simulation? and should allow these businesses to emerge from
By relying on engineering simulation as the primary tool the recession in a stronger state, and quite probably with
to develop new products or processes, engineers and fewer competitors. What we do know, from independent
designers often want to simulate as near to the real world as research that we have been involved with, is that the best-
possible. A multiphysics approach is inevitably going to be in-class companies are often those that make the greatest
more accurate at simulating the real world than one that use of simulation.
uses only CFD or FEA, for example. As high-performance
computing (HPC) continues to improve, we will see more- Where is engineering simulation heading in the next 10 to 20 years?
and more-realistic multiphysics simulations in engineering. The developments that HPC makes possible are very
As engineering simulation becomes more powerful and exciting and could transform the complexity of physics that
additional companies come to rely on simulation to develop can be simulated to the point that it may be possible to
products and processes, being able to employ multiphysics simulate right down to the molecular level. One day, ambient
will become more and more critical. intelligent environments, ultra-high-bandwidth networks,
pervasive wireless communications, knowledge-based
What are the current challenges facing NAFEMS? engineering, networked immersive virtual environments and
As an organization, our broad challenge is to continue powerful games engines will transform multiphysics CAE for
to sharpen our focus on the commercial application of product design, creation, validation and manufacturing.
engineering simulation. For example, we are actively seeking In the future, mesh-insensitive iso-geometric
ways to bring about the kind of mutual understanding pre-processing techniques will become more common. We
among simulation engineers, product development and will see the gaming industry and Hollywood-style post-
business teams that is necessary to put engineering simu- processing and visualization being pulled into CAE more and
lation at the very heart of product development strategies. more. More stochastic simulation as opposed to determin-
Another example is to address the lack of CAE standards. istic predictions will be performed because, as more
In college education, we need to establish a set of computing power becomes available, it will be possible to
learning outcomes for simulation engineers, and knowledge study a range of analyses rather than worst-case/best-case
capture from more-experienced engineers is essential for simulations that are the trend today.
best practices. Simulation data management is an up-and-coming
Although bigger companies usually have rigorous issue in our industry. Good standards are required with
engineering simulation processes, small and medium petabyte-sized files that may soon become common.
enterprise companies may not. NAFEMS would like engi- Security of data and information is crucial with enterprise-
neers to understand the reliability of their CAE analyses. wide projects and collaborations across the world.
We aspire to establish grades of competencies for good FEA, CFD and other related technologies are still very
simulation based on experience for these engineers. much in their infancy. Engineers in the future may look back
and be amused at how crude and unreliable the methods of
What is the significance of CAE in the current economic climate? today are when compared with the technology that is yet to
Simulation ultimately helps companies to save money. It come. The technology itself continues to be developed
is all about making the design process more effective and at an ever-increasing rate, but the complexity of the
efficient. Simulation empowers engineers and designers to applications that industry would like to tackle continues to
envision and develop better designs — in which better exceed the available capabilities. ■

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 11

ENVIRONMENTAL

Extending the Bounds

of Customer Service
Spray nozzle manufacturer expands value-added services
by using simulation to develop and validate gas conditioning
solutions in complex pollution control systems.
By Rudolf Schick, Vice President, Spray Analysis and Research Services, Spraying Systems Co., Illinois, U.S.A.

Founded in it enters specialized pollution control Moreover, they must determine

1937 in a small equipment. For example, spraying effective spray patterns for complex
converted garage, water into an exhaust stream cools the ductwork — especially when retrofitting
Spraying Systems gas from 1,430 degrees F to 620 older exhaust systems that have
Co. is now a degrees F — a temperature that existing flanges, recesses and twisting
worldwide leader ensures optimal performance from geometries. Because there is no
in spray techno- downstream pollution control equip- standard computational development
logy, producing a ment. If the spray does not enter the gas method available to solve such difficult
Rudolf Schick, wide range of stream at the correct angle, or if too problems, companies often spend
Spraying Systems Co.
spray nozzles, much water is injected, droplets will not months of time and perhaps millions
automated spray systems, specialized fully evaporate. The resulting acid-laden of dollars in prototype tests and
fabricated products and accessories. mist can impinge on duct walls, equip- late-stage troubleshooting. Worse yet,
Customers that use this technology are ment and other parts of the structure, some of these poorly designed
in hundreds of industries, including causing erosion and damage. systems are put into service. Plant
steel, paper, food, chemical, petro- Sizing and positioning nozzles in owners then may face costly penalties
chemical, pharmaceutical and metal these gas conditioning systems to for failure to comply with emission con-
fabrication. Spraying Systems Co., avoid impingement and other spray trol regulations, as well as expensive
always on the lookout for ways to problems are demanding tasks. downtime for system redesign.
improve product development and Engineers must calculate numerous A far better approach to gas
expand the value-added services it variables such as flow temperature, conditioning system design uses
provides to customers, began using gas velocity and exhaust pollutants. computational fluid dynamics (CFD).
ANSYS FLUENT fluid dynamics soft-
ware a few years ago in analyzing nozzle
behavior. The company quickly
discovered that the simulation software
is also ideally suited for studying the
design of its customers’ gas condi-
tioning solutions that utilize nozzles
in complex pollution control systems.
Installed in industrial furnaces,
burners, refineries, processing facilities,
power plants and other sites, gas
conditioning systems remove toxins
such as nitrous oxide (NOx) and sulfur
dioxide (SO 2) from exhaust gases
prior to release into the atmosphere.
Spraying Systems Co. manufactures
Using fluid dynamics simulation, engineers studied swirling flow and uneven velocity in a gas conditioning system.
systems that use spray technology to Design change iterations optimized the design for much more uniform flow velocity (left) and lowered the temperature
cool and otherwise treat the gas before variations in the stack to within a nominal range (right).

12 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 ENVIRONMENTAL

In one recent case, engineers per-

formed such work on a gas
conditioning system retrofit project
for a refinery exhaust tower. Due to
physical constraints of the retrofit, the
design called for the quench system to
be installed and controlled from a fixed
height and single side panel of the
square tower. The custom solution
called for the installation of three
FloMax® FM25A air atomizing nozzles
from Spraying Systems Co. Numerical
calculations accurately determined the
resulting nozzle pressure, liquid flow
rate, atomization air flow rate and drop
size for the nozzles.
To analyze the performance of the
proposed system, the 3-D geometry
of the exhaust duct system was
generated in CAD based on information
provided by the customer. Spraying Simulation showed engineers the velocity and spray pattern of liquid emerging from the nozzle and streaming through
Systems Co. then imported this the air before eventually evaporating. This is indicated in color transitions from red to blue. The extended spray stream of
geometry into the pre-processing tool an initial design (left) was shortened considerably in an optimized configuration (right) to reduce damaging impingement
of the spray on the stack walls.
from ANSYS to create a mesh using
elements small enough to provide a design that optimized gas flow and engineering simulation in enhancing
high level of detail for the study. To improved the evenness of the gas value-added services to customers.
understand the behavior of the spray, velocity. The resulting flow uniformity Recently, they deployed ANSYS
engineers analyzed the model helped ensure better temperature Mechanical software to integrate
to determine droplet velocities and distribution and droplet evaporation. more structural analysis into the
trajectories, exhaust pressure and The final duct design improved development cycle for determination
temperature distributions and overall evaporation by over 10 percent with of stress and deformation of nozzle
spray concentration throughout the impingement of droplets on walls mounting lances. Also, use of the
duct tower. and other parts of the structure virtually ANSYS DesignXplorer tool enables
The CFD simulation indicated eliminated. In addition, temperature engineers to study alternative designs
significant problems with the proposed profiles at the duct exit were controlled quickly and to help zero in on optimal
gas conditioning system, including to within 7.7 percent of nominal temp- solutions.
strong swirling flow, several regions of erature requirements — a considerable In this manner, simulation has
low pressure, uneven velocity and improvement over the 42 percent become a key component in the
temperature profiles, and near-zero flow variation found in the initial design. company’s design capabilities, and its
in some local zones. These conditions Spraying Systems Co. now routinely value has been proven many times over.
would likely result in incomplete droplet uses this simulation approach to In addition to providing customers
evaporation along with impingement of validate many of its customers’ and regulatory agencies with docu-
droplets on internal walls, the duct proposed retrofit designs and to mentation of system performance,
structure and downstream equipment. design new gas conditioning systems. simulation has become a powerful new
Using CFD, engineers performed The method provides a unique value- communication tool for the company.
iterative simulations to study and added service, strengthens the Spraying Systems Co. has a 3-D
modify the proposed nozzle insertion company’s relationships with customers projection studio that allows customers
depth, rotation angle and insertion and provides an additional source of to stand virtually inside the application
angle configuration. Various nozzle revenue. Simulation enables engineers to to see and experience the technical
configurations were evaluated based efficiently complete several additional details of the proposed solution. In this
on the baseline gas flow through the projects each month, ensuring successful respect, simulation is as much a
existing tower. performance of installed equipment. sales tool as a design tool, enabling
These iterations enabled the The company continues to the company to increase business
engineers to develop a more effective investigate opportunities to leverage significantly in this growing market. ■

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 13

AUTOMOTIVE

Opening
New Doors
Brose uses simulation to drive
product quality, reduce testing
and minimize costs.

The Brose Group, headquartered in Coburg, Germany,

is a tier-one supplier specializing in developing and manu-
facturing mechatronic systems and electric drives for body
and interior of automobiles: components such as door
systems, seat adjusters, closure systems and electric drives
in particular — to more than 40 automobile manufacturers
and suppliers worldwide. The company has 52 locations
worldwide and a global staff of more than 14,000 people.
ANSYS Advantage staff recently interviewed
Sandro Wartzack, a simulation and knowledge-based
engineering manager at Brose. Dr. Wartzack completed
one of the first Ph.D.s in Germany for the integration of
finite element analysis (FEA) methods in combination with
knowledge-based engineering into design processes.

Can you describe the simulation strategy at Brose?

At Brose, we consider how our computer-aided
engineering (CAE) usage impacts each area in product
development. We try to optimize our CAE approach to max-
imize product quality, reduce testing and minimize cost.
Brose had one simulation engineer using software from
ANSYS in the 1990s. Today Brose has around 45 simulation
engineers across the company. When integrating simulation
into our development efforts, we often find that savings or
improvements in our product designs trickle on to client
savings, which can be significant.
Our door system design effort is a great example:
By using simulation, we have been able to reduce one
specific door system’s mass by approximately 50 percent
compared with its design 10 years ago. Because the new
designs use less material, original equipment manufacturers
(OEMs) that install these components in their vehicles
ultimately produce a lighter and, therefore, more fuel-
efficient and cheaper automobile. These cost savings get
passed on to the car vendor and buyer too.

How does the ANSYS vision for multiphysics tools, virtual

protoyping and a single simulation environment like
ANSYS Workbench fit in at Brose?
We have a lot of CAE specialists around the world who
need to work closely together. The multiphysics products
A selection of Brose door system designs

14 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 AUTOMOTIVE

“A typical door assembly from the Brose Door Systems Business Unit consists
of at least 20 to 30 components. By automating portions of our FEA processes
within the ANSYS Workbench environment, our engineers have shortened
their simulation times, for both static and transient analyses of these complex
designs, by as much as a factor of five.”
— Sandro Wartzack
The Brose Group

and ANSYS Workbench environment allow us to do this.

Being able to perform structural and fluid simulations in
one environment is a great benefit, and the continuous
improvements that ANSYS makes to its solvers — really
the foundation level for these technologies — can only
improve things.
We typically take more than 1,000 hours for testing of
new components today, but we want to reduce prototype
testing, replacing it with more virtual testing. I believe that
development processes tightly integrated with simulation are
the future. For example, if we were devising a new plastic
part with a certain fiber orientation, ideally the engineers
would feed designs into an automated production-like
analysis environment for optimization and virtual testing.

How has the growth in the high-performance computing

sector affected your simulation approach? Typical finite element model for a Brose transient dynamic door analysis in ANSYS
We invariably target a typical nonlinear simulation cycle of Mechanical software, consisting of 35 components and 4 million degrees of freedom
12 hours to 20 hours per simulation so that we can run
structural simulation cases to get overnight results. Our sim- automotive door and seat system sector, I see custom-
ulation model sizes and complexities increase every week, ization and implementation of modular components as a
and the need for powerful hardware is a never-ending focus for the automotive industry of the future. For Brose,
process. By using this approach, our CAE engineer productivity this means moving to develop “plug and play” modular door
is much higher now than it was 10 years ago. We typically had concepts, with separate outer panels and colors to match
one or two processors available per engineer a decade ago, different day-to-day situations.
whereas we now have a supercomputing cluster.
We incentivize our CAE engineers financially to find Where do you feel engineering simulation will move over that time?
innovative ways to streamline their CAE design processes, We are literally growing our CAE usage by 50 percent
targeting changes that might, for example, reduce a typical every year now, because it is so central to our business. The
250-hour CAE design process down to 80 hours or less. increased usage of simulation at Brose today has resulted
This allows us to introduce capture and use of our best in the development of products with lower mass and,
practices and experience to our benefit. consequently, lower costs. In 10 years’ time, I could foresee
that we will do significantly less physical prototyping
Where do you see Brose technology going in the next 20 years? but, rather, focus on full and accurate single-environment
With regard to environmental issues, we are looking at multiphysics simulations and virtual prototyping.
more lightweight products to minimize carbon dioxide I also think the area of visualization is very important. We
emissions. I also see acceleration in continuing trends for are already seeing 3-D co-operative visualization work with
automotive safety feature improvements. More and more, virtual meeting rooms. The ability to use simulation results
we need to produce products that are automatically to convey design and development choices together with
“presafe” electronically. From my viewpoint in the worldwide distributed teams is going to be vital. ■

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 15

BUILT ENVIRONMENT A European opera stage, which served as a background
set in the James Bond film Quantum of Solace, was
engineered using software from ANSYS. The stage is
dominated by a 9-meter-(30-foot)-diameter eyeball.
Copyright Bregenz Festival/Benno Hagleitner.

ANSYS Sets the Stage

Simulation was used to design the floating stage set
used in the latest Bond movie.
By Gerhard Lener, ZT Lener, Feldkirch, Austria

An important scene in the latest elements and in various stages of of the opera to become a projection
James Bond movie Quantum of Solace construction. Linear analysis was used screen, an opening door, an execution
takes place in and around the Bregenz to check the structure for serviceability, platform and a ledge from which a
Festival’s stunning open-air floating then nonlinear analysis was performed stunt-fall into the lake is performed.
stage. This European stage, originally to ensure that the structure could with- Bregenz is the capital of Voralberg,
constructed for the opera Tosca, was stand even higher loads without failing the westernmost state of Austria,
built at a cost of nearly $8 million and catastrophically. ZT Lener used the located near the border with Germany
features a huge eye, 31 meters (101 broad set of analysis capabilities in and Switzerland. Every two years, the
feet) high by 48 meters (157 feet) wide ANSYS Mechanical software to Bregenz Festival constructs a new
with an independently moving 9-meter- analyze the Tosca stage because it floating stage on Lake Constance for
diameter eyeball. The structural design enabled evaluation of the structure presenting a single opera. The latest
of the stage was validated to ensure that from every possible standpoint — all floating stage was built in 2007; its
it could safely withstand environmental within a single simulation environment. amphitheater has about 7,000 seats,
loads, loads caused by moving various and, over two years, approximately
elements of the stage and loading Opera Stage also a Movie Set 320,000 people will have seen Tosca.
during assembly of the stage. A key Quantum of Solace These stage sets always represent
Finite element analysis predicted sequence shot at the Bregenz stage complex engineering constructions
the stresses and deformations in the occurs during a production of the opera that have to simultaneously fulfill
structure at various wind speeds, at Tosca. The eye portion of the stage artistic and strength requirements.
different positions of the moving changes throughout the performance Because no stage set is similar to a

16 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 BUILT ENVIRONMENT

guided the structural design. ANSYS

Mechanical software was the ideal
tool for simulating the floating stage
because it provides a very wide range
of tools — including linear and
nonlinear analysis — addressing
materials ranging from metal to
rubber, a wide range of solvers and
the ANSYS Workbench environment,
which provides bidirectional commu-
nications with most CAD systems. In
performing structural analysis for the
last six floating stages for the Bregenz
Festival, the design team has encoun-
tered a very wide range of structural
analysis problems, and technology
from ANSYS has been able to handle
every one.
ZT Lener employed the traditional
ANSYS Mechanical tools predicted the stresses and deformations for the frame structure on the floating stage. version of ANSYS Mechanical soft-
ware for the main structure because it
allowed use of a script to generate
previous one, each becomes a of the analysis, since connecting the input files that automated the process
new challenge for the entire team, steel and wood in a shear plane of analyzing the structure at inter-
which has to find solutions in many provides additional stiffness beyond mediate positions. Engineers modeled
engineering disciplines. the sum of the properties of the the eye’s steel supports with beam
two materials. elements and its wooden surface using
Stage Presents Structural Challenges shell elements. Powered by two
The biggest challenge in the Broad Range of Analysis Tools hydraulic cylinders, the eye rotates 90
Tosca stage design was providing the The stage design began as a 100- degrees between the vertical and the
strength needed to safely move the to-1 scale model, provided by the horizontal position. Many structural
eye while staying within the weight stage designer, that was used to create members receive their highest loading
limits of the foundation. The moving a 3-D Pro/ENGINEER ® computer- in intermediate positions of rotation,
parts of the stage weigh about 250 aided design (CAD) model that so it was necessary to analyze the
metric tons, while the entire stage and
foundation weighs only 463 metric
tons. Another important limitation is
that each component must be moved
to the stage by a crane that can handle
only about 12 tons. The small bridge
between the stage and the land is
limited to a mere 1 ton per square meter.
This means that any larger components
must be moved in smaller pieces and
assembled on the stage itself.
Further challenges resulted from
the components’ materials: The eye
The eyeball portion of the stage is far more than a
and eyeball are made of a composite static background: The iris and pupil were engineered
construction with a steel frame and a to rotate and fold out via hydraulics, creating a horizontal The eye structure is a composite, a steel frame with a
performance space. The iris also serves as a screen for wood outer surface. Using the nonlinear capabilities of
wood outer surface. The composite special visual effects and a door that opens to reveal yet software from ANSYS, ZT Lener was able to accurately
construction increased the complexity another scene. Copyright Bregenz Festival/Karl Forster. predict the physics involved in this complex analysis.

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 17

BUILT ENVIRONMENT

Engineers used simulation to predict the stresses and deformations in the structure The triangular brackets that connect the eye to the rotating shaft are critical parts
at various wind speeds. Under wind loading, the eye structure deforms as much as of the steel support structure. Based on the analysis results of the triangular
27 millimeters (5 inches). The engineering team also performed dynamic analysis support, engineers changed the wall thicknesses and positions of the stiffeners
on various parts of the structure to ensure that it would not resonate and cause on the brackets to reduce deformations and stresses to safe levels.
vibrations that might interfere with a performance or damage the structure.

structure at many different positions to shell is assembled, but, on the other were calculated to ensure that it would
be sure that no structural member is hand, it also experiences less wind not resonate when several people
overstressed. The script moved the pressure. Simulation verified that the moved on it at the same time. When the
mechanism through a range of positions stage could perform all movements at decision was made to film the James
and tracked the highest stresses and normal speed at a wind speed up to 50 Bond movie on the stage, 1,500 kg
deformations on each area of the model kilometers per hour (kmh). There is a (3,306 pounds) of lights had to be
throughout the entire range. Under wind range of wind speeds above 50 kmh at installed in the upper corner of the eye
loading, the eye structure deforms as which the stage can be moved — but structure. This required a separate
much as 127 millimeters (5 inches). at a slower speed. At wind speeds simulation, which indicated that the
above this level, the stage needs to be structure needed to be strengthened.
Analyzing Structural Details moved to a specific position, where it In constructing the floating stages
The team modeled the triangular is best able to resist wind loading, and for the Bregenz Festival, there is
brackets that connect the eye held there. obviously no opportunity for building
to the rotating shaft, critical parts of prototypes or making design changes
the steel support structure, in Evaluating Ultimate Limits of Structure along the way. Since the opening date
the ANSYS Workbench environment. The ultimate limits of the stage of the festival is set long in advance,
ANSYS Workbench makes it very easy structures that take plastic elastic unlike many building projects, the
to bring the CAD model into the analysis capacity into account were also completion date for the stage cannot
environment. Based on the analysis evaluated with the loads multiplied by be changed. The safety of the singers,
results of the triangular support, the safety factors ranging from 1.3 to 1.5. the stage crew and the audience
team changed the wall thicknesses The structure had to be designed to depends upon getting the design
and positions of the stiffeners on the withstand these design loads and with right the very first time. The use of
brackets to reduce deformations and elastic deformations under character- ANSYS Mechanical software, whose
stresses to safe levels. istic loads (safety factor 1.0). Dynamic accuracy has been proven on a very
Because of the transportation analysis on various parts of the wide range of analysis tasks, gave
limitations already mentioned, it was structure ensured that it would not the entire project team confidence in
critical to model the structure at resonate nor cause vibrations that the analysis results. ■
various stages of construction. For might interfere with a performance
CADFEM, an ANSYS channel partner in
example, the steel structure of the eye or damage the structure. The mode Germany, supported Lener in his use of
is much weaker before the wooden shapes and frequencies of the eyeball software from ANSYS.

18 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 ENERGY

Harnessing the
Power of Ocean Waves
Engineers use structural and hydrodynamic analysis to ensure that
wave-powered electrical generation machines produce maximum
energy output and operate effectively for decades.
By George Smith, Managing Director, and Tamas Bodai, Analyst Engineer, Green Ocean Energy Ltd, Aberdeen, Scotland

With rising fuel costs and environ-

mental concerns, governments around
the world are focusing on clean, safe
and sustainable alternative energy
sources for power generation. One of
the most unique and promising of these
concepts is harnessing the energy of
the earth’s oceans by converting the
relentless force of waves into electricity.
The idea has captured the imagi-
nation for centuries, but until now the
business justification has not been
sufficient to move such projects
forward. Significant engineering
hurdles must be overcome to develop
efficient, reliable and economical
wave-powered electrical generation
systems that could be deployed on a
mass-production basis. Green Ocean
Energy is meeting these challenges
with the help of simulation technology
from ANSYS: ANSYS AQWA software
for hydrodynamic analysis of the
wave action and ANSYS DesignSpace
software for structural analysis.
The engineering team uses this
advanced technology in the develop-
ment of the Ocean Treader, a floating
device designed to be moored five
kilometers offshore in open ocean
water with relatively high wave
Machines developed by Green Ocean Energy produce 500 KW of electricity from on-board generators powered by activity. A second device called the
wave action that raises and lowers floating arms, which sit atop buoyant sponsons. The Ocean Treader (top) is moored
to an anchor while the Wave Treader (bottom) mounts on the base of offshore structures such as wind turbines or Wave Treader mounts on the base of
tidal turbines. offshore structures such as wind

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 19

ENERGY

ANSYS DesignSpace structural analysis software was used to determine stress distribution in the Ocean Treader arm (top)
and deformation of a spreader beam structure that lifts the machine safely into the water (bottom).

turbines or tidal turbines. Both Software from ANSYS played a key • Added mass of the structure as
machines share a similar design, with role in meeting these objectives. The a result of the surrounding water
two 20-meter steel arms floating on a design team used the ANSYS AQWA set in motion by the oscillating
set of sponsons (components that product to determine how the structure body
make the machine buoyant) made of would respond to a particular wave • Hydrostatic stiffness and
glass-reinforced composite plastic. action. First, the team created a hydro- buoyancy
As wave action moves the floating dynamic model of the submerged part
arms up and down, hydraulic cylin- of the structure based on the geometry Hydrodynamic parameters were
ders spin generators that produce of components together with their entered into a proprietary code devel-
electricity sent back to shore via density and inertia. Next, they entered oped by Green Ocean Energy for
underwater cables. Each machine is wave data profiles, including wave computing the kinematic response and
designed to produce 500 KW of elec- height and frequency, obtained resulting power output of the machine.
tricity — enough to power 125 homes from empirical measurements in the Customized plots of the power output
— so a farm of 30 such devices would particular body of water. for a range of sizes of the major com-
have a rating of 15 MW. From these inputs the ANSYS ponents, such as the length of the arms
One of the primary challenges the AQWA application generated a variety and shape of the sponsons, enabled
engineers faced was reaching a of hydrodynamic parameters including: engineers to determine optimal design
balance between structural strength • Diffraction force accounting for parameters for the major structural
and weight restrictions. With an the deformation of waves as members.
expected 25-year design life, the they impact the structure The structural analysis model to
machines must withstand rough waters • Froude–Krylov force derived compute stress distribution and defor-
of the North Atlantic, where waves can from the pressure field of waves mation of components was efficiently
reach over 9 meters in height in gale- against the structure achieved through tight integration with
force winds. Conversely, structural • Hydrodynamic damping due Autodesk® Inventor®, which enabled
members must be lightweight to keep to radiation of waves induced part geometry to be automatically trans-
production costs within budget and by structure motions and the ferred from CAD to ANSYS Workbench
to allow for sufficient floatation. associated energy dissipation using the Geometry Interface for

20 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 ENERGY

5.0 5.0

4.0 4.0
Newtons/m

3.0 3.0

Kg/s
2.0 2.0

1.0 1.0

0.0 0.0
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Y*10+5 Frequency (radians/sec) Y*10+5 Frequency (radians/sec)
Dim 2 = -90 deg - Diffraction force - heave(z) Dim 1 = -180 deg - Radiation damping heave(z) - heave(z)

1.10 4.8
1.00 4.4
4.0
0.90
3.6
0.80
Newtons/m

Kilograms
3.2
0.70
2.8
0.60
2.4
0.50
2.0
0.40 1.6
0.30 1.2
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Y*10+6 Frequency (radians/sec) Y*10+5 Frequency (radians/sec)
Dim 2 = -90 deg - Froude–Krylov force - heave(z) Dim 1 = -180 deg - Added mass heave(z) - heave(z)

ANSYS AQWA hydrodynamic software determines how the Wave Treader reacts to wave action by computing hydrodynamic parameters such as diffraction,
Froude–Krylov force, radiation damping and added mass.

Inventor/MDT. The analytical meshing ANSYS DesignSpace software was In this complex development
was greatly simplified through the use instrumental in minimizing the weight of process — in which so many variables
of surface-to-surface contact element the entire structure while helping to must be considered — standard hydro-
features that automatically detect con- ensure that each part could withstand dynamic calculations alone would be
tact points of touching parts. The use the range of expected wave forces over too slow and not detailed enough to
of multiple parts allowed different time. The technology from ANSYS was provide sufficient insight into the
material properties to be assigned — crucial in achieving the sensitive behavior of the machines subjected
including the anisotropic nature of the balance of mass, moment of inertia and to severe environmental conditions.
glass-reinforced composite plastic parts. center of gravity so that the floating arms Moreover, because prototypes cost
After an initial simulation cycle was would react optimally to wave action. over $3 million each and take months
completed using ANSYS DesignSpace Currently, engineers are using this to construct, numerous rounds of
software, direct associativity with the procedure to develop Ocean Treader hardware test-and-redesign cycles
CAD system enabled engineers to scale-model prototypes — which are are impractical. To meet the rigorous
readily change the design and quickly undergoing wave-tank trials — and technical requirements, product
perform subsequent simulations on the several organizations are expressing delivery deadlines and business
new part geometry without having to strong interest in both the Ocean objectives for both the Ocean
re-apply loads or boundary conditions. Treader and Wave Treader. To fill future Treader and Wave Treader machines,
Green Ocean Energy engineers orders once tests verify power output Green Ocean Energy finds the virtual
performed successive iterations to and structural integrity, full-scale pro- prototyping capabilities of the
eliminate stress concentrations by duction models of the machines will be advanced tools from ANSYS a
adding or trimming material where developed using these same tools from critical element in getting the
needed. ANSYS, with detailed full analysis of product to market in a cost-effective
final designs to be performed with and timely manner. ■
ANSYS Mechanical software.

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 21

MARINE

The megayacht Happy Days at sea. Photo by Neil Rabinowitz.

Designing for Strength,

Speed and Luxury
Simulation software from ANSYS helps a yacht designer deliver
the optimal combination of luxury and performance.
By Chad Caron, Naval Architect, Delta Marine Industries Inc., Washington, U.S.A.

Purchasers of 100-foot-plus considerations; more recently, advance- thumb, are not adequate to achieve an
megayachts have come to expect the ments in composites have provided optimized structural assessment for
ability to customize the interior design designers with far more flexibility. Today, these new types of interiors. This
to a level that matches their wildest for example, the location of pillars means that analysis typically needs to
dreams. Award-winning yacht builder can be more readily accommodated be performed on a global basis, which
Delta Marine has become one of the by structural engineers to suit the in turn requires very powerful software
world’s leading builders of megayachts interior designers’ vision. and hardware.
— in part through expertise in designing Traditional design methods, such The horizontal structural elements
carbon fiber structures that enable as handbook formulas and rules of in a megayacht are the decks. In
virtually any interior configuration while Delta’s latest megayacht, the three
providing high levels of strength, decks are made of two-inch-thick
durability and performance. However, composite sandwich construction. The
giving interior designers the freedom vertical structural elements consist of
to place walls or partitions wherever free-standing pillars that are used to
they wish creates structural design support beams and also vertical
challenges by increasing the complexity beams that are attached to the super-
of the load paths. structure plating (the mullions).
Graphite composites provide the Delta selected ANSYS
ideal material for megayacht design Mechanical technology as
because they are stiffer and stronger its structural design
than metals per unit of weight, making
it possible to build a lighter and
stronger boat. Composites enable
more flexible designs because their
physical properties can be tailored to a
very high degree. In the past, interior Bimini top and mast first vibration mode for the Happy Days, the largest composite yacht built in the
design was constrained by structural Americas, showing total displacement sum

22 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 MARINE

software seven years ago because the Delta designed its new mega-
yacht maker believed that the soft- yacht in two stages: first for strength
ware’s composite analysis capabilities and then to resist vibration. Today’s
were well ahead of competitors. At the yacht buyers are interested in a
time ANSYS Mechanical was the only luxurious interior and a high cruising
finite element software Delta could find speed, but it is critical to optimize the
with a composite shell element. As structural elements to deliver the
composites simulation technology has required strength while avoiding any
progressed, according to Delta, the extra weight that would reduce the
ANSYS Mechanical package has speed of the boat. The designer uses
maintained an advantage in composite ANSYS Mechanical technology to
design capabilities. evaluate global and local stresses on a
The Delta Marine team models layer-by-layer basis. Most other finite
Delta’s latest project has a racking frame, shown here,
the major shapes of the yacht in element analysis packages merely designed to resist a roll acceleration. This is a relative
Rhinoceros®. The Rhinoceros model is average the loads over the stack. plot of displacement for a given frequency.
exported to a neutral file format and ANSYS Mechanical tells the
imported into ANSYS Mechanical soft- engineers exactly where the load is modeled the racking frame using
ware to provide the geometry for going, down to the individual com- 0/90 and +45/-45 biaxial laminate
the model. The naval architect uses posite layer. This simplifies the design and unidirectional carbon fiber. The
composite shell elements to model the of the mullions, beams and pillars. The analysis results identified the stresses
laminate stack layer by layer and uses ability to distribute the loads among on the structure and helped determine
solid elements for foundation parts that the different layers also helps to tune which layer buildup should be used on
are cast in resin and in scantlings (frame the laminate stack. Delta uses ANSYS each particular part of the frame. Delta
and structural support dimensions) that reports to detail and defend structural was also concerned about the longi-
have a core that is structurally significant. decisions that the regulatory body rule tudinal second mode of vibration
The prediction of vessel vibration books cannot cover adequately. identified in the modal analysis. This
frequencies is dependent on the total Even after a structure has been mode generated a high bending
weight distribution for the yacht. The designed to support the design loads, moment near the middle of the ship,
interior design has the potential effect the yacht may still vibrate. The Delta which was addressed by strengthening
of increasing overall weight through designer performs modal analysis to the hull and decks to make them stiffer
the substantial use of hardwood and investigate its primary modes of in the area that experiences the highest
stone, especially common in today’s vibration using the same ANSYS bending moment.
yachts. Delta has developed para- Mechanical model. The technology ANSYS Mechanical simulation
metric approaches to estimate interior determines the natural frequencies of makes it possible to determine exactly
weights using targets on a per-square- the mass matrix. The analysis results in how loads distribute through this
foot basis for various materials. The a recent project showed that the first complex structure, so that engineers
outfit weight along with the other mode of vibration was a racking mode, can tailor the properties of structural
structural and mechanical weight which meant that the superstructure of elements to provide strength and
components coupled with the hydro- the boat vibrated horizontally, with the stiffness exactly where it is needed.
dynamic added mass of the water decks decoupling from each other like These capabilities free the designers
directly affect the vibration frequencies a deck of cards sliding back and forth. to put walls and partitions wherever
and mode shapes the yacht will The engineers addressed the they want and to keep the weight to a
exhibit. Accurately predicting these concern by adding a racking frame, a minimum level. As a result, the new
frequencies and mode shapes is structure that spans two decks boat delivers the optimal combination
critical to successful design. and resists horizontal motion. Delta of luxury and performance. ■

Mr. Terrible cruising Alaska. This 154-foot semi-displacement design is built for high performance,
Mr. Terrible’s hull first mode of vibration reaching maximum speeds of 24 knots. Photo by Neil Rabinowitz.

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 23

ELECTRONICS

Images © istockphoto.com/MorganLeFaye, istockphoto.com/macroworld,

istockphoto.com/LPETTET, istockphoto.com/FarukUlay.
Seeing the Future
of Channel Design
NVIDIA uses VerifEye and QuickEye as an extension to traditional
SPICE-level simulation approach to design high-performance
graphics solutions.
By Chris Herrick, Technical Lead, Ansoft LLC

It’s hard to imagine, but there was once a day when One of the hardest challenges when designing these
slick high-resolution graphics were not the norm. With high-performance graphics solutions is ensuring that the
increasing monitor sizes as well as ever-sharper digital communication link is clear between the pixel generation
media and spectacular gaming technology, computer and pixel display. That means the signal, representing a zero
graphics have progressed in an amazing way. NVIDIA® is a or one, originating at one part of the system needs to prop-
world leader in visual computing technologies and the agate undistorted to another area so it may be detected
inventor of the GPU, a high-performance processor that without errors.
generates breathtaking, interactive graphics on work- As data link speeds increase, so do the problems
stations, personal computers, game consoles and mobile that affect signal quality. Every part of the physical routing
devices. NVIDIA serves the entertainment and consumer channel has some influence on the propagating electro-
market with its GeForce® graphics products, the professional magnetic fields and, thus, on the detected waveform. The
design and visualization market with its Quadro® graphics channel could be assembled from many combinations of
products, and the high-performance computing market with elements including packages, transmission lines, cables,
its Tesla™ computing solutions products. connectors and vias. A discontinuity or impedance

24 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 ELECTRONICS

Full channel simulation using the Ansoft Designer tool

mismatch to the propagating signal could occur at any point resulting graph is called an eye diagram, and it can
along the transmission path. very clearly show whether the received data is able to be
A common tool for the signal integrity engineer is circuit detected error free.
simulation. By modeling the channel virtually, engineers are Transient analysis is the most accurate means of
able to predict waveforms not only at the receiver but for determining signal waveforms, but it is limited by the scope
each section of their modeled channel. This level of detail of the problem. To fully analyze all the variations in a channel
allows engineers to verify signal detection as well as with nonlinear devices can take days to weeks. Trying to
to determine the contribution of signal distortion for each achieve low bit error rates (BERs) poses an additional
section of the channel. To improve or optimize a system, the challenge. Using transient analysis, simulating enough bits
sections of channel that produce the greatest signal to satisfy a BER of 10-12 could take years. In order to
distortion can be identified, and intelligent changes can be satisfy these engineering challenges, Nexxim technology
made. These changes could include geometric variations, has incorporated two specialized solvers, namely QuickEye
elimination or addition of components, or material selection. and VerifEye.
In order to construct the virtual channel, NVIDIA chose According to Ting Ku, director of signal integrity at
the Ansoft Designer tool as its simulation environment. NVIDIA, “The obvious reason for statistical transition is
Ansoft Designer allows the engineer to assemble each related to simulation coverage. Given there is a finite
piece of the channel as a black box model. These models amount of time and machine resources, the statistical
may comprise measured data, simple circuits, SPICE approach gives engineers systematic coverage without
components or dynamic links into any
of Ansoft’s circuit extraction tools. These Curve Info
600.00 AEYEPROBE (probe_out)
individual models may be rearranged, QuickEye_noFFE

bypassed or parametrically varied, providing 400.00

AEYEPROBE (probe_out) [mV]

the engineer with the ability to test all

200.00
possible configurations. This high-level
schematic approach also allows the design 0.00

to be easily shared among different groups, -200.00

which then can quickly see what is being
modeled and provide input into the design. -400.00

Under the hood of Ansoft Designer software -600.00

lies the powerful circuit simulator product 0.00 50.00 100.00 150.00 200.00 250.00
Time [ps]
Nexxim. Nexxim technology is a high-capacity,
high-accuracy engine for linear network Curve Info
150.00 AEYEPROBE (probe_out)
analysis, transient analysis, harmonic balance, QuickEye
fast convolution and statistical methods. With 100.00

this array of different simulators, NVIDIA is

AEYEPROBE (probe_out) [mV]

50.00
able to look at channel performance from
many different perspectives, all from within the 0.00

same environment.
-50.00
The traditional signal integrity simulation
methodology is to perform a transient sim- -100.00

ulation. Using this approach, the driver is toggled

-150.00
for a length of time, and the voltage is monitored 0.00 50.00 100.00 150.00 200.00 250.00
Time [ps]
at receiver. A common way of viewing this
received data is to overlay the voltage versus The top image demonstrates the receiver eye on a channel where the data cannot be
time for each bit period, or unit interval. The recovered; the bottom is a clean eye diagram where the data is recoverable.

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 25

 ELECTRONICS

1.0000e-003
Name X Y 8.8889e-004
m1 0.4997 0.0000 7.7778e-004
m2 0.4997 35.2263 6.6667e-004
m3 0.5003 -34.2387 5.5556e-004
m4 0.6834 0.3292 4.4445e-004
m5 0.3172 0.3292 3.3333e-004
20.00 2.2222e-004
Amplitude [mV]

1.1111e-004
0.00 1.0000e-009

-20.00
Name Delta(X) Delta(Y) Slope(Y) InvSlope (Y)
d(m1,m2) 0.0000 35.2263 N/A 0.0000
d(m1,m3) 0.0007 -34.2387 -51985.7339 -0.0000
d(m1,m4) 0.1838 0.3292 1.7916 0.5582
d(m1,m5) 0.1824 0.3292 -1.8046 -0.5542
0.20 0.30 0.40 0.50 0.60 0.70 0.80

Interval

Eye contour plot generated with the VerifEye tool

running an astronomical number of simulation corners. One While it is critical to fully characterize the entire passive
other good benefit of the statistical approach is in dealing channel, the scope of the analysis does not stop there.
with design corner definition by projecting what the final To compensate for frequency-dependent effects of the
production yield would be. Using the statistical method- channel, such as inter-symbol interference (ISI), NVIDIA
ology allows engineers to make judgment calls between uses silicon-based compensation. Additionally, there may
cost and production yield.” be other influences on the signal in the form of jitter that
While both QuickEye and VerifEye methods offer must be accounted for. This jitter may be seen at both the
significant speedup over transient analysis, each offers a driver and the receiver.
different solution to the problem at hand. QuickEye is a fast As part of its investigation into silicon-based channel
convolution-based method that allows the user to explicitly compensation, NVIDIA can use either QuickEye or VerifEye to
define a bit pattern that is sent and to view the resultant evaluate feed forward equalization (FFE) or decision feedback
waveforms. VerifEye is a purely statistical-based approach equalization. If the silicon has already been characterized, the
that characterizes BER of a channel down to 10-16. number of equalization taps and their respective weights can
Both of these methods begin analysis the same way by be added to either the driver or receiver on the channel.
first computing the transfer function of the channel. This During early stages of design, the Nexxim tool can be used
computed channel response is assumed to be linear–time to automatically calculate the ideal weights necessary to
invariant. For QuickEye, the channel response is then invert the effects of the channel on a bit stream.
convolved with a user-specified bit sequence to obtain a Jitter characterization, and its inclusion in simulation, is
time vs. voltage waveform. For VerifEye, a cumulative another area critical to NVIDIA. Without including all sources
distribution function is derived from the step response of noise, accurate BER simulations would be impossible.
based on the conditional probability of various bit Random jitter (RJ) and duty cycle distortion (DCD) can also
transitions. The main outputs from these analyses would be added to each driver. ANSYS staff learned from this part-
be an eye diagram from QuickEye and a bathtub or a nership with NVIDIA that the inclusion of periodic jitter (PJ)
statistical eye contour plot from VerifEye. and sinusoidal jitter (SJ) would be useful features, so ANSYS
has since enhanced the Nexxim tool to include these features.
1.00E+000 Curve Info
1.00E-001
AEYEPROBE (probe_out)
VerifEye
Deterministic jitter (DJ), based on ISI, will inherently be
1.00E-002 __Amplitude=“0”
AEYEPROBE (probe_out)_1 modeled by the channel’s transfer function. At the receiver,
1.00E-003 VerifEye_noFEE
__Amplitude=“-0.005042782088” the source jitter will accumulate with the DJ of the channel to
1.00E-004
1.00E-005 create a new jitter distribution. This jitter in combination with
1.00E-006 the jitter defined at the receiver, either RJ or a user-defined
1.00E-007
distribution, will account for the total jitter (TJ) of the channel.
1.00E-008
Y1

1.00E-009
Reducing TJ is the main objective when designing a channel
1.00E-010 for low BER.
1.00E-011
With ever-increasing bit rates and channel complexity,
1.00E-012
1.00E-013
the landscape of signal integrity analysis is changing
1.00E-014 drastically. Transient analysis can no longer be relied on as
1.00E-015 the sole means of channel simulation, especially when trying
1.00E-016
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 to achieve extremely low BER. This challenge has been met
Unit Interval

Bathtub curve of feed forward equalization performance generated with the

head on at NVIDIA by adopting QuickEye and VerifEye
VerifEye tool analysis into their design process. ■

26 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 27
ELECTRONICS

Read and Write Heads

Head Stack Assembly (HSA)

Schematic of hard
disk drive assembly

Actuator Arm(s)

Pivot

Bobbin

Hot Topics:
Voice Coil Motor (Actuator)

High-Capacity Hard Disks

Samsung uses simulation to improve
thermo-fluidic performance of hard disk drives.
By Haesung Kwon, Senior Staff Engineer, Samsung Information Systems America, California, U.S.A.

Recently, the capacity of hard disk drives (HDD) has Samsung made an engineering discovery that allowed
reached the phenomenal level of more than one terabyte them to improve thermo-fluidic performance of the HDD.
per single drive. Robust mechanical design played a key The finding also provided insight into the design of high-
role in this achievement, since drive development performance HDDs.
challenges today are not related to just a single physics but As always in simulation-driven product design, sim-
to multiphysics. In the past, most mechanical-originated ulation during the early stage of HDD development is an
failure modes were identified using only a good under- important contributor to a successful time to market. The
standing of HDD dynamics. But as the tracks on the disk range of simulation available for the HDD industry includes
become more tightly packed to achieve higher capacity for both basic and advanced features of ANSYS Mechanical
the drive, nanometer-level positioning of read and write and ANSYS CFX software. Samsung uses software from
elements is very important in response to the external and ANSYS because of its expandability to multiphysics
internal vibration of the HDD. These vibrations are often capabilities. For instance, flow-induced vibration has been
caused by air flow and heat transfer. used to understand and predict the HSA’s dynamic
To achieve nanometer-level precision, faster seek and
access time is needed. This requires higher current, which,
in turn, leads to temperature rise in the voice coil motor (or
actuator), usually simply called the coil. The coil moves the
actuator arm holding the read and write heads, and the
arms, heads and coil together are called the head stack
assembly (HSA). Temperature rise in the coil can cause
undesirable mechanical performance. This temperature rise
is strongly dependent on the location of the HSA, and
convective heat transfer can affect the temperature rise
when the HSA is in different positions. Using both ANSYS
CFX and ANSYS Mechanical software, engineers at

28 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 ELECTRONICS

performance for different configurations. The model

can be extended in the future to run thermally induced
vibration or shock in a systematic manner with rela-
tively little effort.
The simulation models for the HDD contain solid
a b
and fluid regions. Samsung engineers used the fluid
region to solve for flow and heat transfer. They used
the solid domain to determine heat transfer only
and, in this study, without considering radiative heat
transfer. Temperature rise in the solid regions,
particularly the coil and actuator arms, is of great
interest because these regions are strongly tied to Temperature field and associated velocity vector of air flow in both (a) inner-located arm and
(b) outer-located arm models
the reliability of the entire drive.
Using ANSYS CFX capabilities, the Samsung
team calculated temperature, velocity and pressure
with two different models: with the HSA positioned at 12.00
the inner radius of disk (ID) and at the outer radius of ID OD
disk (OD). Intuitively, the engineers knew that different 10.00
outcomes for flow-induced vibration would occur at
each location. However, the team had seldom
Average Temperature Rise [K]

8.00
explored simulations of temperature discrepancies
because of the large model required to encompass the
two different physics — flow and heat transfer. 6.00

The model contained five arms and four disks.

The fluids simulation determined that a high flow rate 4.00
passes through the coil when the HSA arm is posi-
tioned in the ID. This creates a high convection
2.00
coefficient on the surface of the voice coil motor,
resulting in a relatively low temperature rise. The arms
actually block air flow from passing to other areas of 0.00
Bobbin Pivot Arm0 Arm1 Arm2 Arm3 Arm4
the disk, and the air flow tends to move toward the to Air to Air to Air to Air to Air to Air to Air
voice coil. However, with the arm on the OD, a path Temperature rise in different locations in the head stack assembly of a hard disk drive with the
for air to flow in the circumferential direction is actuator arm in different positions (ID = inner diameter, OD = outer diameter)
opened up, so that less air flow travels through the
coil and a low convective heat transfer coefficient is
generated. This observation is the critical reason that
temperature rise is quite different in the two cases. 0.25
Temperature rise within different parts of the HSA ID OD
were also investigated using structural mechanics 0.20
simulation. The bobbin has the highest temperature
Normalized Heat Transfer

swing compared to other parts since it is located 0.15

closest to the coil heat source. The temperature rise
for all five arms is almost the same. Samsung engi-
0.10
neers performed an analysis to determine if there was
uniformity of convective heat transfer from each part
0.05
compared to the total convective heat transfer. They
determined that more heat is transferred through the
inner arms located between two disks. The results 0.00
Air
Air

ir
ir

ir

ir
ir
Air

Air

ir

reveal that the heat transfer path of conduction and

oA
oA

oA

oA
oA
oA
to
o

to

to
il t

0t
nt

1t

4t
3t
2t
ut

in

ot
Co

Arm
e-I

Arm

Arm
Arm
Arm
e-O

bb

the convection modes are highly dependent upon the

Piv
Glu

Bo
Glu

location of the HSA. This understanding is critical to Heat transfer in different locations in the head stack assembly of a hard disk drive with the
the thermal packaging design of the HDD. ■ actuator arm in different positions (ID = inner diameter, OD = outer diameter)

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 29

AUTOMOTIVE

Fatigued by 325-hp model EPA 07 ISB Cummins diesel

used with bus, RV, medium-duty truck, and
fire or emergency vehicle applications

Stress Limitations
The combination of fe-safe and ANSYS software
helps Cummins improve life prediction accuracy.
By Jeff Jones, Technical Advisor, Cummins Inc., Indiana, U.S.A.

In developing cutting-edge design solutions, diesel mechanical analysis solver from ANSYS because of the tech-
engine manufacturer Cummins Inc. uses a deterministic nology’s flexibility and performance. The relationship between
approach for predicting product life, one that considers the two companies has expanded since then: Cummins has
complex materials and loading. Its current solution incorpo- been an active member of the ANSYS Advisory Board for
rates technology from two proven leaders — but the path to more than a decade.
this approach was not a straight line. However, there was reluctance at Cummins to replace
A recognized technology leader in the global diesel its internally developed fatigue analysis software because of
engine market, Cummins faces increasingly stringent rigorous internal requirements for depth and range of
design requirements as it develops cutting-edge solutions. fatigue theories along with the need to handle proprietary
The company’s roots are planted in soil nourished by materials and loads. In 2002, the company turned to Safe
innovation. For example, the firm was among the first to see Technology Limited, which offered fe-safe™ for fatigue
the commercial potential of diesel engine technology. Even and durability analysis. The partnership that existed
before the advent of commercial software tools, Cummins’ between ANSYS and Safe Technology ensured efficient and
engineers developed internal software for thermal, effective interfacing between fe-safe and simulation tools
structural and design applications from ANSYS.
to ensure that its engine designs To verify that fe-safe offered accurate life prediction
were cost effective, reliable and capability, Cummins executed a sophisticated test plan to
durable. Today, Cummins is no compare fe-safe results to internal fatigue analysis software.
longer just an engine business The test plan included four finite element models:
but a global power leader with
• Simple 2-D plane stress uniaxial model
more than $11 billion (U.S.) in
• Moderate 3-D biaxial stress model
annual sales.
In the late • Fully featured engine block
1970s, Cummins • Fully featured engine head
continued its
pioneering efforts, Cummins engineers subjected each model to a number
becoming one of of different and appropriate loading scenarios. By using a
the first compa- range of models, it was possible to gain fundamental
nies to embrace insights into the technology and to compare predictions
commercial tools against field data.
for finite element The baseline internal fatigue software was based on an
350-hp model 6.7 L Cummins turbodiesel
analysis. It stan- advanced Goodman approach: one that is stress-based,
used in Dodge Ram heavy-duty trucks dardized on the in which damage prediction is based on stresses.

30 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 AUTOMOTIVE

Stress-based fatigue analyses are severely limited when it

comes to low-cycle fatigue problems. Low-cycle fatigue
typically considers approximately 105 duty cycles, while
high-cycle fatigue is appropriate for more than about 109
cycles. Cummins needed a unified approach for predicting
product life for gray iron components that would be viable for
both low-cycle and high-cycle fatigue. The company also
wanted to determine if the strain-based methods employed
by fe-safe, such as the Smith–Watson–Topper (SWT)
algorithm with Neuber correction, were more suitable for
meeting this low- and high-cycle requirement for gray cast-
iron fatigue prediction.
For load cases dominated by tensile stresses, the
Goodman-based internal software provided results that
were consistent with the SWT approach. However, in the
Goodman method–equivalent fully reversed stress results identify spurious damage
biaxial case dominated by compressive stress, the internal prediction in addition to high fatigue damage locations.
software predicted much more damage than fe-safe,
implying that a stress-based approach in this case may
result in an overly conservative design.
Complex real-world models of a cylinder block and
head, considering standard proprietary loading conditions,
produced more noticeable differences. Stress situations
for two complex load cases were compared at more than 20
locations for which considerable test experience existed.
In nearly all cases, fe-safe results were in line with expecta-
tions, while the Goodman-based approach predicted less
damage at several locations (thus over-predicting product
life). A closer review at three critical locations revealed
that for cases with high mean, low alternating stresses,
fe-safe provided results that agreed very well with test and
field experience.
Factor of strength results from fe-safe plotted in ANSYS Mechanical software showing
Cummins engineers made the following observations actual high fatigue damage locations
when considering the test results:
• While internal software over-predicted product life in failure) or factors of strength (such as design margin).
some cases and under-predicted it in others, fe-safe Another important factor was the ability to use compre-
results used in conjunction with the mechanical hensive and user-configurable libraries, facilitating use of
analysis solver from ANSYS correlated very well with internal proprietary materials data with minimal effort.
Cummins’ industry experience. With the development of the integrative ANSYS
• Even with modifications, older stress-based Workbench platform, all structural analysis and flow
approaches for predicting fatigue have limitations modeling tools at Cummins are being brought into one
in comparison with modern strain-based methods. environment, further enhancing productivity.
• Use of the mechanical solver from ANSYS in Today at Cummins, nearly every engine component is
combination with fe-safe offers opportunities to analyzed using the ANSYS Mechanical product. fe-safe
further increase reliability and reduce costs. is used to perform fatigue analysis for many components,
• A better understanding of fatigue facilitates such as cylinder blocks, cylinder heads, pistons, connecting
design innovation. rods and main bearing caps. In engine cylinder heads with
high assembly stresses, significant compressive stresses,
and peculiar behaviors of gray cast iron, fe-safe software
A noted contributing factor to the successful outcome
plays a vital role in helping to develop reliable, cost-effective
of the testing was the tight integration between fe-safe and
designs. Advanced fatigue analysis using fe-safe with
ANSYS Mechanical software. Using fe-safe, the ANSYS
ANSYS Mechanical software helps to get the design right
results (.rst) file is read, material properties and fatigue cycle
the first time, and it reduces development costs. ■
(combinations of load steps) are specified, and the fatigue
damage is calculated and written back to an ANSYS results
file for display in ANSYS Mechanical software. Fatigue Reference
results may be plotted as contours of log-life (log-cycles to www.safetechnology.com

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 31

ELECTRICAL

Managing Heat with

Multiphysics
Multiphysics simulation helps a global company design
better electrical products.
By Arunvel Thangamani and Sanjeeva Reddy, Schneider Electric, R&D, Global Technology Centre, Bangalore, India

Multiphysics simulation is used Workbench, ANSYS

in electrical industries to predict product Multiphysics and
performance and failure conditions and ANSYS FLUENT tech-
to perform optimization. Product testing nologies. Researchers
is very costly, and repeated trials are there used ANSYS
not part of the preferred product Multiphysics to perform thermo-
development process, wherein electric simulations on two Schneider
products are optimized early in the Electric product assemblies — a circuit
design process using simulation. In breaker terminal and an automatic
addition, simulation assists product transfer switch — to determine the
designers in these industries to meet temperature generated due to Joule
standards required by bodies such as heating as well as to define conductor
Underwriters Laboratories (UL) and and insulator specifications to effectively
the International Electrotechnical manage heat. The study extended into Test rig of circuit breaker showing approximate location
of terminals inside the casing
Commission. Thermoelectric simu- analyzing the effects of convective
lations play a vital role in product cooling by varying the convective film
development in product areas such as coefficient, and the results from the is an electromechanical switching
final distribution enclosures, industrial ANSYS Multiphysics simulations were device used widely in Schneider
plugs and sockets, protection and compared with test results. Electric low-voltage products. For
control of low-voltage power circuits, Miniature circuit breaker (MCB) both the terminal and switch, the
electrical network management, energy- terminals are subassemblies of the CAD model was developed in
efficiency devices, automation and Schneider compact circuit breaker Pro/ENGINEER® and imported directly
control devices, power electronics series. The terminal connects the into the ANSYS Multiphysics environ-
cooling, and millivolt switching devices. circuit breaking device with external ment. Engineers meshed the geometry
Schneider Electric is a global circuitry. An automatic transfer switch with direct coupled-field elements,
specialist in energy management, with
operations in more than 100 countries.
The company focuses on making
energy safe, reliable and efficient. The
organization’s Global Technology
Centre (GTC) in Bangalore, India, has
460 employees working on product
development and resource enhance-
ment, and its resulting innovative
products and technologies are
available in markets across the globe.
To reduce costs and gain time in
their product development process, Temperature simulation of the 63-amp circuit breaker terminal The current density in the circuit breaker terminal
original design (left) and new design (right). The design
the GTC uses ANSYS Icepak, ANSYS improvement caused a 9-degree Celsius temperature reduction.

32 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 ELECTRICAL

“The Global Technology Centre uses multiphysics

technologies from ANSYS to reduce costs and gain
time in their product development process.”
which automatically account for and also corresponded well with
the bi-directional coupling between experimental results.
electric Joule heating and temperature Similarly, using simulation, design
in either steady-state or transient improvements in insulation material
analyses. These elements accept both were applied to the transfer switch
thermal and electric boundary condi- terminal model that resulted in a
tions and excitation. Electric boundary temperature reduction of 6 degrees C
conditions were used to prescribe in the assembly. Once again, the
the net DC current passing through simulation results compared well with Transfer switch test setup
individual solid conductors. Thermal the experimental results, with a plus or
boundary conditions consisted of minus 4-degree C deviation.
thermal convection coefficients The R&D experts then studied the
defined on surfaces exposed to current flow path and the variation
25-degree Celsius (C) air. of the convection coefficient with
The team modeled two MCB the maximum temperatures obtained.
terminals: a 63-amp terminal and a These simulation results gave
25-amp terminal. The design for the researchers confidence that the same
63-amp terminal was modified in modeling approach could be applied
thickness to reduce the temperature to all products in these families.
by 9 degrees C. Correlations between Electrical conductor thickness and
simulation and experimental lab results insulation material selection can be
were impressive, with only a plus or optimized using simulation to reduce
minus 2-degree C deviation. For the the need for prototyping at the beginning
25-amp terminal, temperature results of the design cycle and to save valuable
confirmed the safety of the product development time and costs. ■
ANSYS Multiphysics simulation results for the transfer
switch terminal: original design (top) and new design
(bottom) showing temperature reduction
Effect of Convection Coefficient on Joule Heating
Temperature

Convection Coefficient

Effect of convection coefficient on Joule heating for the transfer switch terminal Current density in the transfer switch terminal

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 33

AEROSPACE

Up, Up
and Away
Simulation-driven innovation
delivers a new ejection seat
design for a military aircraft
in less than 14 months.
An ejection seat, used in emergency situations in military aircraft. An explosive charge or rocket
By Park O. Cover, Jr., Senior Mechanical Engineer, Concurrent motor thrusts the seat out of the aircraft, carrying the pilot with it. Once airborne, a parachute is
Technologies Corporation, Pennsylvania, U.S.A. deployed. This photo shows the ACES II ejection seat. U.S. Air Force photo by Staff Sgt. Bennie J. Davis III.

The military’s advanced-concept ejection seat, ACES II®, is Analysis of the seat was split into three phases. The first
one of the most successful aircrew escape systems in U.S. Air analysis phase was conceptual design development. During
Force history and is credited with saving more than 450 lives this time, engineers designed the seat structure to meet
since it was introduced in 1976. With more than 8,000 seats functional requirements, while simulation was used to verify
delivered to date, the ACES II is currently used on F-15, that the structure was sound and weight was optimized.
F-16, B-1B, B-2, A-10, F-117 and F - 2 2 a i r c r a f t . U s i n g Functional, structural and safety requirements were derived
t h e strengths of the ACES II as a foundation, Goodrich from the performance-based specification supplied by
Aircraft Interiors and Concurrent Technologies Corporation aircraft manufacturer Lockheed Martin for the JSF ejection
(CTC), both in the United States, developed the next- seat. To reduce maintenance downtime, a modular seat
generation ACES 5 seat for the F-35 Joint Strike Fighter structure was developed to allow the seat to be easily
(JSF). The new seat was optimized to enhance safety for removed from the aircraft. The modular seat consists of the
aircrew, to reduce maintenance downtime, to reduce seat back, seat bucket, parachute, survival kit and aircraft
weight and to integrate with the F-35 cockpit. However, interface module. Assembly costs and part count were
the biggest challenge was developing and delivering a reduced by designing the new seat to use a few machined
brand new seat structure in less than 14 months. components instead of many sheet metal components.
The parametric link between the ANSYS Workbench Engineers evaluated designs for tough load require-
platform and Pro/ENGINEER® Wildfire® software was a ments, such as ejection from an aircraft traveling at 750
critical factor in successfully developing a design that met mph, parachute load and crash loads.
all the requirements while maintaining the aggressive The first simulation phase evaluated individual
schedule. Engineers at CTC were able to quickly update components of the preliminary seat design. Equivalent
simulations for multiple design iterations. This concurrent stress plots of various stages of the bucket design evolution
design and analysis approach enabled the team to optimize demonstrated how, during ejection, the occupant’s legs are
the seat for both function and weight from the earliest forced apart by the windblast. The structure had to be
developmental stage. optimized to contain this splitting force, or else the

Pressure due
to windblast
at 750 mph

Area that required

a submodel

Acceleration due to
ejection catapult

Closeup of high stress region

Loads imparted on the seat when ejected from an aircraft traveling at 750 mph Stress loads that result from static analysis of ejection at aircraft speed of 750 mph

34 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 AEROSPACE

Iteration 1 Iteration 2 Iteration 6 Iteration 20

2.1 pounds 3.0 pounds 2.7 pounds 2.8 pounds

Equivalent stress plots from various iterations of the ACES 5 ejection seat bucket design.
Weight of the bucket, one of the design considerations, is shown for each.

occupant would sustain critical injuries. The engineering experience these loads only one time during deploy-
team analyzed the structure for ejection and crash loads ment. Because the system model used a static simulation
within the ANSYS Workbench framework. Once the sim- approach without nonlinear material properties, the simu-
ulation was set up, design iterations were quickly lation revealed small areas of stress concentration that
evaluated for all the applicable load cases simply by exceeded the allowable ultimate strength of the material.
updating the geometry from the CAD system. Engineers scrutinized these high-stress zones using sub-
During the second analysis phase, the CTC team built a models that allowed material yielding during the third
system model of the seat structure. Analyzing the seat phase of the analysis.
structure as a whole gave the most representative view of To produce the submodel, the team first cut out the area
how the actual seat structure would behave and eliminated of interest using the ANSYS DesignModeler tool. The CTC
compromises associated with analyzing individual seat team developed a submodeling subroutine using the
subsystems or modules. To prepare the system model commands object in the mechanical simulation area
for analysis, the team imported the CAD geometry of ANSYS Workbench. The subroutine interpolated the
into the ANSYS DesignModeler tool system model displacements
where defeaturing operations, such onto the submodels’ cut bound-
as elimination of rivet holes, were aries. The submodel results
performed. In addition, a few components typically showed that some
were converted to mid-plane surface permanent deformation occurred,
models using the software’s automatic but the ultimate strength of the
mid-plane feature. material was not exceeded.
The CTC team assigned material Furthermore, the submodel provided
properties, defined boundary conditions more accurate stress results due
and applied loads to the system model. to the finer mesh. Roughly 30
Contact regions were characterized for high-stress areas were evaluated
each riveted face on the seat. This allowed using this technique to ensure
contact reaction forces to be used to that the structure would not
determine the number of rivets required at fail when loaded in extreme
each joint. Point masses were used to Submodel of high stress regions in ANSYS DesignModeler
conditions. These results proved
represent nonstructural seat subsystems, software. Cut boundaries are shown in red. that the ultimate load requirements
such as the parachute and survival kits. were met.
The model was meshed using a hex- After 10 months of develop-
dominant mesh control and a 0.125-inch ment, five prototype seats were
global element size. A single linear static built for test purposes, and the
structural analysis of the seat model was first ejection test of the ACES 5
solved in less than 30 minutes using the F-35 JSF seat occurred after 14
direct solver within the mechanical months. The seat performed
software available through the ANSYS flawlessly the first time out. This
Workbench platform. The quick analysis extraordinary outcome is the result
turnaround time allowed the engineering of a great deal of teamwork
team to quickly evaluate various what-if between Goodrich and CTC
design scenarios. and would have been unattain-
Actual loads on the seat are very Submodel results provide more-accurate stress results
able without using engineering
dynamic in nature, and the seat will than the global static model. simulation software. ■

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 35

MARINE

Propelling a

Image © istockphoto.com/dan_pratt.
More Efficient Fleet
Rolls-Royce uses simulation for propeller design to reduce
marine fuel consumption.
By Johan Lundberg, CFD Engineer, and Per Aren, Project Manager in Hydrodynamic Design, Rolls-Royce Marine, Kristinehamn, Sweden

Rolls-Royce is a name associated analysis, such as flow velocities and

worldwide with quality — and not pressure at every point in the problem
limited just to automobiles. Through domain, as well as the inclusion of
subsidiary Rolls-Royce Marine, its viscous effects. The challenge at Rolls-
equipment is installed on 20,000 com- Royce Marine was to incorporate the
mercial and naval vessels around the full complexity of flow around the
world, and its comprehensive range of propeller into the CFD model in order to
products includes gas turbine and accurately match physical experiments.
diesel engines, nuclear propulsion Engineers developed the CFD
systems, steering gears, stabilizers, models using a hexcore volume mesh
thrusters, water jets, winches, cranes, generated using TGrid pre-processing
rudders and main propeller systems. Closeup view of the unstructured surface mesh used for software from ANSYS. The hexcore
CFD simulation of the Kamewa CP-A propeller
The company is a key part of the Rolls- mesh, which maintains cell surfaces
Royce Group, with 7,000 employees in physical reality but involve the perpendicular to the main flow in the
serving 2,000 customers. The Rolls- expense and time of building and core fluid region, kept the number of
Royce Hydrodynamic Research Center testing a prototype. In addition, poten- cells to a reasonable level. Rolls-Royce
in Sweden is Rolls-Royce Marine’s tial flow analysis is restricted in that it Marine used the full multi-grid initial-
center of excellence in hydrodynamic does not account for the full geometry ization method together with the
propeller and waterjet design research. of the propeller. CFD, on the other pressure-based coupled solver in
The center has combined the best of hand, can incorporate the full geomet- ANSYS FLUENT software, which has
computational fluid dynamics (CFD) rical complexity of a propeller’s proven to be both robust and fast for
simulation and physical experiments to operation. CFD can also provide many applications. The engineering
help the company to develop the much more detailed results than either team simulated turbulence using the
Kamewa CP-A, its latest controllable- physical experiments or potential flow renormalization group (RNG) k-ε model,
pitch propeller (CPP).
A CPP is a special type of propeller XF5 STD 2 1 1 2
with blades that can be rotated around
their long axis to change their pitch.
Changing the pitch makes it possible
to provide high levels of efficiency and
maneuverability for any speed and
load condition. Stopping distance Type A 21 1 2
can be cut in half compared with a
conventional fixed-pitch propeller.
Traditionally, propeller development
has been driven by a combination of
Open-water simulation of previous XF5 propeller (top) and new Kamewa CP-A (bottom). Contours of velocity
physical experiments and potential magnitude are shown at planes in two locations (1 and 2) for each design and indicate the low momentum,
flow analyses. Physical experiments or viscous losses, close to the hub. The arrows indicate the thicknesses of the boundary layers.
have the advantage of being grounded

36 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 MARINE

Contours of pressure coefficient for the XF5 (left) and the new Kamewa CP-A (right). Insets: Photographs of the blade indicating the locations of the simulation where cavitation is present
(noticeable as pitting). ANSYS FLUENT results helped reduce pressure at the blade root in the CP-A design, indicated by the lack of cavitation erosion present in the CP-A photo.

since it was considered to be stable designed the propeller as the last step in waves that cause localized stress on
and, thus, a conservative approach. designing the ship, so there was no and damage to nearby components.
The engineering team began by opportunity to improve the hull design ANSYS FLUENT results identified low-
analyzing propeller calculations for to optimize the propulsion system. pressure areas in which cavitation
open-water operating conditions. These The ability to simulate the interaction could occur on the Kamewa CP-A hub.
calculations considered the operation of of the propeller and hull has now made it Changing the geometry increased the
the propeller in a uniform flow field with- possible to address such a concern. pressure above the critical level and
out looking at the influence of the ship’s CFD simulations allowed Rolls- eliminated cavitation, which made it
hull. They then used the rotating Royce Marine to evaluate a wide range possible to increase the load at the
reference frame method to simulate the of alternative hub geometries. The sim- blade root to further improve efficiency.
rotating propeller. With this method, the ulations also helped the engineering According to a 2003 study from the
team solved the flow equations in the team reach a higher level of knowledge University of Delaware, international
rotating frame of the propeller blade. by providing far more information than commercial and military shipping fleets
Integration of the calculated pressures physical tests could. For example, consume approximately 289 million
and shear stresses on the CFD made it feasible to easily metric tons of petroleum per year,
propeller blades yielded determine the boundary which is more than twice the
thrust and torque, and layer in any prospective consumption of the entire population
the propeller’s efficiency design. Generally, as of Germany[1]. The ANSYS FLUENT
was then calculated the boundary layer simulations run on the modified
using these values. gets thinner, the design propeller geometry predicted that
Design develop- becomes more efficient. the efficiency would increase by 1
ment then moved into By performing simu- percent to 1.5 percent, and physical
a detailed study of the lations of a number of experiments confirmed that this was,
interaction between different designs quickly, in fact, the case. This seemingly
the propeller and ship the team concluded that small improvement, however, has the
The new Kamewa CP-A propeller
appendages. The Rolls- from Rolls-Royce Marine they could reduce the potential to reduce fuel costs by
Royce Marine team boundary layer and improve several billion dollars if applied across
simulated the complete ship hull efficiency by modifying the hub contour. the board to the world’s commercial
in order to calculate the effect of the Rolls-Royce Marine engineers shipping fleets. It also has the
wake field on the propeller design. were next concerned about the possi- opportunity to significantly reduce
Engineers used a sliding mesh model bility of cavitation on the propeller hub energy consumption and emissions of
to simulate the operation of the caused by the boat’s wake. Cavitation greenhouse gases. ■
propeller in the flow field under the is the formation of vapor cavities in a
influence of the ship hull. The sliding liquid due to a localized reduction in
mesh model is a transient approach fluid pressure below certain critical Reference
that calculates the flow field as one grid values. The vapor cavities collapse [1] Corbett, J.J.; Koehler, H.W. Updated
Emissions from Ocean Shipping, Journal
region rotates (or translates) relative to violently as they move to regions of of Geophysical Research – Atmospheres,
another. Historically, Rolls-Royce Marine higher pressure and generate pressure 108(D20), 2003; pp. 4650–4666.

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 37

ANALYSIS TOOLS

Staying Cool with

ANSYS Icepak
Thermal management solution predicts air flow and heat transfer in
electronic designs so engineers can protect heat-sensitive components.
By Stephen Scampoli, Lead Product Manager, ANSYS, Inc.

ANSYS Icepak technology is aimed at one

of the most significant challenges facing
engineers designing electronic assemblies:
dissipating thermal energy from electronic com-
ponents to prevent premature component failure
due to overheating. This fully interactive software
is used to evaluate the thermal management
of electronic systems in a wide range of
applications, including simulation of air flow
in enclosures, analysis of temperature
distributions in chip and board-level packages,
and detailed thermal modeling of complex
systems such as telecommunications equipment
and consumer electronics. By predicting air flow
and heat transfer at the component, board or
system level, the software improves design
performance, reduces the need for physical
prototypes and shortens time to market in the
highly competitive electronics industry.
Based on powerful computational fluid
dynamics (CFD) simulation, ANSYS Icepak
technology has a specialized user interface that ANSYS Icepak software predicts the temperature profile in a computer graphics card.
speaks the language of electronics design
engineers. Models are created by simply
dragging and dropping icons of familiar
predefined elements including cabinets, fans,
circuit boards, racks, vents, openings, plates,
walls, ducts, heat sources, resistances and
heat sinks. These “smart objects” capture geo-
metric information, material properties and
boundary conditions — all of which can be fully
parametric so a user can easily enter values to precisely
match application requirements or to study what-if scenarios. The
software also includes extensive libraries for standard materials,
packages and electronic components such as fans — including fan
geometry and operating curves.
As a further modeling aid, the software can import both
electronic CAD (ECAD) and mechanical CAD (MCAD) data from a
variety of sources. Geometry from ECAD and MCAD data sources
can be combined with smart objects to quickly and efficiently Simulation results of a fan-cooled processor heat sink attached to a
create models of electronic assemblies. For instance, a system printed circuit board

38 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 ANALYSIS TOOLS

model of a computer enclosure could easily be

generated by combining MCAD data for the
enclosure, ECAD data for the printed circuit boards
(PCBs) and electronic packages, and smart objects for
other components. In addition, the ANSYS Icepak solution
includes many macros to automate the creation of geometry,
including different types of packages, heat sinks, thermo-
electric coolers and industry-standard test configurations.
Another productivity feature is the ability to automatically
generate highly accurate body-conformal meshes that
Thermal simulation of a PCB using the power map data imported from SIwave
represent the true shape of components rather than a
rough stair-step approximation. Meshing algorithms can
generate both multi-block and unstructured hex- all available. The software also offers customized reports
dominant meshes. Algorithms also distribute the mesh that allow users to identify trends in the simulation along
appropriately to resolve the fluid boundary layer. While the with the ability to report fan and blower operating points.
meshing process is fully automated, users can customize Reports including images can be created in HTML format
the meshing parameters to refine the mesh and optimize the for distributing the results data.
trade-off between computational cost and solution accuracy. ANSYS Icepak tools can interface with other products in
By grouping objects into assemblies, the mesh count can be the software portfolio from ANSYS to allow comprehensive
further optimized by meshing each assembly separately and multiphysics simulation of electronic components. One
automatically combining them before running the solution. option is the ability to import a power distribution map from
This meshing flexibility results in the fastest solution times SIwave; this simulation software from Ansoft extracts
possible without compromising accuracy. frequency-dependent electronic circuit models of signal
ANSYS Icepak uses the state-of-the-art ANSYS and power distribution networks from device layout data-
FLUENT computational fluid dynamics solver for the bases for modeling integrated circuit packages and printed
thermal and fluid flow calculations. The CFD solver solves circuit boards. Based on the results from an SIwave simu-
the fluid flow and includes all modes of heat transfer — lation, users can import the DC power distribution profile of
conduction, convection and radiation — for both steady- printed circuit board layers into ANSYS Icepak software for
state and transient thermal-flow simulations. The solver also a thermal analysis of the board. The coupling between
provides complete mesh flexibility, and this allows the user the two packages allows users to predict both internal
to solve even the most complex electronic assemblies using temperatures and accurate component junction temp-
unstructured meshes, providing robust and extremely fast eratures for printed circuit boards and packages.
solution times. ANSYS Icepak software can export temperature data
Once the solution is complete, ANSYS Icepak software from a thermal simulation to a structural mechanics model
provides a number of different methods for visualizing and to calculate thermal stresses of electronic components.
interpreting results. Visualization of velocity vectors, With the demands of today’s high-performance electronic
temperature contours, fluid particle traces, iso-surfaces, devices, electronic components are becoming more
cut-planes and two-dimensional XY plots of results data are complex and using more exotic materials. These newer
materials have widely varying thermal and mechanical
properties and are being subjected to higher temperatures
during both manufacturing and usage. These varying
material properties and temperatures can result in signifi-
cant thermal stresses, which can bring about fatigue-based
failure of the components. ANSYS Icepak software,
together with ANSYS Mechanical technology, allows users
to evaluate both the thermal and mechanical aspects of
the design.
ANSYS Icepak technology in conjunction with SIwave
and ANSYS Mechanical products provides a full portfolio of
software to meet the simulation requirements of the
electronics design engineer. ANSYS continues to be a
leader in providing solutions to the electronics industry —
solutions that provide the high-fidelity electrical, thermal
Mesh around a pin fin heat sink follows the geometry of the part without and structural simulations required to meet the challenges
any approximation. of today’s product development demands. ■

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 39

ANALYSIS TOOLS

Analyzing Vibration
with Acoustic–
Structural Coupling
FSI techniques using acoustic elements efficiently compute
natural frequencies, harmonic response and other vibration
effects in structures containing fluids.
By Marold Moosrainer, Head of Consulting, CADFEM GmbH, Munich, Germany

When designing equipment such as system when a water valve is abruptly segments of multiphysics simulation.
vessels, tanks, agitators, hydraulic shut off. FSI simulations are usually performed
piping systems, hydraulic turbines, To fully study structural vibration in using the ANSYS multi-field solver,
transformers and sensors, engineers these types of applications, engineers which employs implicit sequential
often must take into account a must model the coupling mechanisms coupling to calculate interactions
contained fluid. Presence of such fluids for fluid structure interaction (FSI). For between fluid and structural solutions.
may add mass, stiffness and damping, these detailed studies, software from As an alternative to these types of
which change the structural mech- ANSYS has an outstanding breadth FSI analyses of fluid-filled structures,
anics of the system. Also, the fluid may and depth of capabilities for structural engineers may want to consider an
act as an excitation mechanism such as and fluid analysis. Models are getting approach based on the use of ANSYS
occurs in water hammer, which is the more and more realistic, and FSI FLUID30 elements available in the
shock wave that occurs in a piping continues to be one of the largest ANSYS Mechanical and ANSYS

How Fluids Influence Structural Vibration

The presence of a fluid can significantly change rule of thumb in analytically determining the required
the vibration characteristics of the containing added mass because results depend so much on
structure. To determine the extent of this effect, frequency and mode shape. A fully coupled FSI
engineers must model all relevant dynamics, simulation is required to answer this question.
especially the fluid–structure coupling that represents
the interaction between these two domains. Models
based on the ANSYS FLUID30 element must, there-
fore, account for factors, such as mass, stiffness and
damping, that the fluid adds to the overall system.

Mass. The additional mass of a heavy, rather incom-

pressible fluid such as water in a vessel usually must
be included in the analysis model, such as the thin-
walled box in the example. Note that the first vibration The influence of the additional mass of water on vibration modes of a rectangular
mode for the partially filled container is about a third tank is shown in models of a partially filled tank (left), an uncoupled structural dry
mode of the empty tank at 33 Hz (middle), and a coupled wet mode at 10 Hz taking
lower than that of the dry container. There is no easy into account FSI using ANSYS SHELL181/FLUID30 coupling (right).

40 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 ANALYSIS TOOLS

Multiphysics products. These elements

have their origin in acoustic appli-
cations; typically, they are used for
simulating sound radiation. Their
elasto-acoustic and hydro-elastic
capabilities, however, are very helpful
in solving FSI vibration problems
by providing straightforward fluid–
structural coupling in a given range of
vibration analyses in which:
• The fluid is quiescent or at least
ANSYS model of a hydraulic turbine (top) and
moderately slow. FLUID30-based pressure field (bottom)
Images courtesy Voith Hydro.
• Vibration amplitudes are small
(linear theory).
• The influence of fluid viscosity
or shear layers is negligible,
meaning an ideal gas
assumption.

ANSYS acoustic elements accom- engineers can use ANSYS FLUID30 frequency domain. Consequently,
plish the required fluid–structural elements to attach the piezoelectric for the latter, simulation of the
coupling because they have four part of the multiphysics problem via desired stationary peak response
degrees of freedom (DOF): one for matrix coupling. In this way, three within one single frequency step is
the sound pressure and three strongly coupled physical domains performed very efficiently. This
optional displacement DOFs. Thus, can be solved simultaneously: piezo- is conveniently done without
a consistent matrix coupling is set up electric, structural and fluid. having to account for lengthy initial
between structural and fluid elements Coupling structural elements to transients (particularly for weakly
in which strongly coupled physics acoustic elements in this manner damped structures) required in
cause no convergence or performance allows for transient analysis and, most time-domain solutions coupling
problems. Additionally, in analyzing even more important, for modal structural and fluid domains by a
sensor applications, for example, and harmonic analysis in the load vector.

Stiffness. Lightweight, rather compressible fluids such as gases do

not add appreciable mass, but they can add stiffness to a closed air-
filled structure. To imagine the “stiffness of air,” think of an air spring
(gas shock absorber) or a bicycle air pump that you close with your
thumb. In the analysis of a rectangular piston, for example, the added
stiffness of air included in the model doubled the natural frequency,
from 40 Hz to 80 Hz. Further, instead of one mode, analysis of the
air-filled cylinder indicated many resonances from acoustic cavity.

Damping. In an unbounded fluid domain (ANSYS FLUID30

combined with FLUID130 for the external absorbent boundary layer),
structural vibration may lead to pressure waves that propagate
through the entire fluid system. In these cases, the energy spent on
these compressive longitudinal acoustic waves is dissipated in an Gases add stiffness to the containing structure, as in the piston shown
effect known as “radiation damping.” Particularly large plate-like here supported by a spring and an attached air enclosure (top) with an
uncoupled dry mode at 40 Hz and a coupled wet mode at 80 Hz.
structures in heavy fluids may encounter considerable added Pressure mode of the cavity (bottom) was found using ANSYS FLUID30
damping that must not be neglected, but no fluid viscosity is required elements to analyze the system.
for analyzing such cases.

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 41

ANALYSIS TOOLS

Another advantage of using acoustic

elements is their ability to quickly solve for Why Shallow Containers Slosh
fluid–pressure fluctuations. Since fluid and To avoid severe load instabilities, engineers often face
structural vibration systems have different strict design requirements to control sloshing of liquid in moving
temporal and spatial scales, properly containers, such as tanker trucks or rockets. In these applications,
discretizing wavelengths in time and space designers usually insert interior baffle plates or similar structures to
for these structural and fluid domains often impede the flow of liquid. Other applications include harbor design or
requires extensive amounts of CPU time and study of long-wavelength tsunami waves. In all these cases, simu-
resources. In contrast, acoustic elements lation plays a key role in predicting sloshing and evaluating ways to
account for pressure fluctuations and fluid solve the problem.
properties much more efficiently when struc- An example of sloshing is the carrying of a dog bowl full of water,
tural behavior such as resonant frequencies, in which the liquid has the tendency to slop from side to side and
mode shapes and peak vibration amplitudes often spill. Simulation reveals that this behavior occurs because the
must be calculated. Recent ANSYS solver first sloshing mode of the bowl is roughly at 2 Hz — the typical human
step frequency that excites this undesired resonance. Repeating the
improvements have significantly sped up this
analysis for a glass of water reveals a first sloshing mode of 4 Hz,
task, solving the resulting unsymmetric
which shows why water glasses are much less prone to spilling than
coupled system of equations in a fraction of
bowls. In these simulations, the structural walls have been assumed
the time previously required.
to be rigid. Using the hydro-elastic coupling capabilities of software
The hydraulic turbine study is an
from ANSYS, however, engineers also can study sloshing in elastic
excellent example of a contained fluid
vessels such as reactor containment structures. Note that sloshing
application that can be analyzed quickly and analysis with ANSYS FLUID30 coupling is restricted to small
easily by a structural engineer using ANSYS amplitudes, and that full-fledged finite element and FSI analysis must
acoustic FLUID30 elements. When coupled be applied for simulating very large vibration amplitudes or
to the housing and the rotor, the elements fluid–surface motion.
include two important features. First, they
account for the vibrating mass of water that
strongly affects the vibration frequencies
and vibration deformation of the structure.
Secondly, the elements allow the engineer to
easily model the excitation mechanism: the
fluctuating pressure field of the water induced
by the rotating rotor. Analyzing this coupled
system by modal and harmonic response
analysis with acoustic elements is
considerably easier than with conventional
finite element and fluid dynamics FSI
methods, which require extensive effort
to perform with transient analysis in
both domains.
Today, users can complete modal
analysis of FLUID30-based coupled systems
within hours, even for large assemblies with Golden retriever Alex assists in a demonstration
millions of DOFs that otherwise would take of sloshing. Analysis indicates that the first
sloshing mode is 1.6 Hz, the frequency at which
days to perform using conventional FSI the bowl is prone to resonance and spills its
methods. Solution speed for huge models contents.
can be further improved by using special
component mode synthesis (CMS) methods,
Krylov subspace-based order reduction
methods and symmetric formulations of the
originally unsymmetric FSI system matrix. ■

42 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 43
PARTNERS

Integrating CAE Tools:

a Package Deal
Moldex3D and ANSYS Mechanical team up
to simulate microchip encapsulation.
By Anthony Yang, Director, Technical Research Division, CoreTech System Co., Ltd., HsinChu, Taiwan

Integrated circuits (ICs), also x10 (sec) 0

called microchips, are at the core 3.201 Advancing

Molten Resin
of the modern electronic device. 3.004
(EMC)
2.790
Packaging — the final stage of IC fab-
2.575
rication — is the essential technology 2.361
that distributes electrical signals from 2.146
Die
a silicon chip onto the printed 1.932
1.717
circuit board and provides protection
1.503
against environmental stresses. 1.288 Leadframe
(Paddle)
Customer demand for highly sophisti- 1.074

cated and ever-smaller electronic 0.859

0.645
products has made IC packaging a
0.430
challenging procedure and, thus, 0.216 Metal Leads
critical to system performance. 0.001
Gold Wires
Moreover, this trend is driving the
technology toward higher packaging
Moldex3D simulation results showing the melt front advancements, colored by elapsed time,
densities with thinner and smaller pro-
at a level of 45 percent filled. The die (pink, center), leadframe (salmon, extending outward)
files, which makes the encapsulation and gold wires of the package are also displayed.
process much more complicated and
unpredictable.
Despite recent trends in copper internal heat spreaders, of wire deformation, or sweep, during encapsulation is
a significant fraction of ICs are encapsulated by plastic, a affected by the resin viscosity, wire mechanical properties,
process that is usually accomplished by transfer molding. package dimension and process conditions. If wire defor-
A typical transfer molding begins with loading a leadframe mation exceeds the limit, it can cause wire breakage or
into the mold, placing and preheating the preforms, and adjacent wires to touch, and the package will subsequently
then closing the mold. The molten epoxy molding fail. The paddle, on the other hand, is the flat base of the
compound (EMC) is then forced into the mold system by the leadframe and is used to support the chip. Usually, the
transfer ram. The packages are ejected once the packing paddle is connected to the leadframe by long and thin metal
and curing processes have been completed. Common leads, which give the paddle system a quite flexible
defects in the transfer molding process arise from the structure during mold filling. Uneven resin flow in the cavity,
improper selection of processing conditions, molding however, can apply unbalanced loading on the paddle
material, leadframe layout or mold design. These defects system and, hence, deform or shift the paddle. When the
include short shot, air trap, wire sweep and paddle shift. amount of paddle shift is too large, it will significantly reduce
Among these defects, wire sweep and paddle shift in the thickness of the cavity, further resulting in excessive wire
particular are caused by fluid structure interaction prob- sweep or exposing the die.
lems. Gold wires are common IC package components Applying the conventional trial-and-error method to
used for transferring the electronic signals and power resolve these problems is difficult and costly because of the
between the die and the leadframe contacts. During the complex interactions among fluid flow, heat transfer, structural
encapsulation process, the melted resin flow exerts drag deformation and polymerization of the EMC. However, through
force on wires, which causes them to deform. The amount the coupling of Moldex3D® from Coretech System Co. and

44 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 PARTNERS

force distribution can be calculated along

x100 (%)
6.762
each wire. With a single click in Moldex3D,
6.312
the drag force, boundary conditions and
5.861
relevant mesh data are then exported to
5.410
the ANSYS Mechanical solver, which
4.959
performs the deformation calculation
4.508
transparently in the background for
4.057
each wire at the instant when the melt
3.607
touches it. After the analysis in ANSYS
3.156
Mechanical is completed, the wire
2.705
deformation result can be checked
2.254
either in Moldex3D or in ANSYS
1.803
post-processors. As to the paddle
1.352
shift problem, Moldex3D outputs
0.902
the spatially and time-varying
0.451
pressure loads on the leadframe
0.000
directly to ANSYS Mechanical, which
then performs steady-state simulations
Wire sweep simulation result showing the original and deformed wires colored by wire sweep index
(WSI). The WSI distribution is the maximum deformation divided by the projected length of each wire. based on the data for each instant of
A high WSI value indicates the potential failure of adjacent wires touching or wire breakage. time. For the wire formation calculation,
both Moldex3D and ANSYS Mechanical
ANSYS Mechanical simulation software, engineers can can display the paddle shift results.
analyze the complicated physical phenomena inherent in the The combination of Moldex3D and ANSYS Mechanical
encapsulation process and further optimize the package software provides a promising simulation solution for the
design. The nonlinear capability of ANSYS Mechanical tech- microchip encapsulation process. By using the integrated
nology is critical for wire sweep and paddle shift simulations, analysis, molding defects can be easily detected and
since the deformation of wire or paddle could be quite large. moldability problems can be improved efficiently to reduce
The IC Package module of Moldex3D is a fully manufacturing cost and design cycle time. These reliable
integrated analysis environment connecting pre-processing, and powerful simulation tools can help IC package designers
post-processing, mold filling and structural analyses for to meet the difficult demands from legacy devices to
microchip encapsulation simulation. One of the major tomorrow’s innovative packaging solutions. ■
challenges of three-dimensional modeling
of microchip encapsulation is generating
a suitable mesh for analysis. In the
Moldex3D pre-processor, an efficient NODAL SOLUTION
SUB = 1
volume mesh generator with high-quality TIME = 7.037
and arbitrary grid type (for example, tet, USUM (AVG)
RSYS= 0
hex, wedge, pyramid or boundary layer DMX = .975E-04
SMX = .975E-04
elements) allows meshing of the
package geometry with minimal model
simplification. Furthermore, a parametric
capability can assist in creating wires,
thereby greatly reducing meshing effort. In
the mold-filling stage, Moldex3D can
calculate the resin flow considering
nonlinearities such as viscosity change
and the curing reaction of the EMC.
Engineers can thus predict how the mold
0 .217E-04 .433E-04 .650E-04 .867E-04
will fill, identify potential air traps and weld .108E-04 .325E-04 .542E-04 .758E-04 .975E-04
Moldex3D IC Package Paddle Shift Model
lines positions, and evaluate the runner
and gate design.
An example of a total displacement distribution, measured in centimeters as shown in the
Having obtained the local flow field ANSYS Mechanical post-processor. This paddle shift is caused by the unbalanced pressure
from the mold-filling analysis, the drag loading during mold filling.

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 45

46 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com
 ACADEMIC

Sailing Past a Billion

Racing yacht design researchers push flow
simulation past a meshing milestone.
By Ignazio Maria Viola, Yacht Research Unit, University of Auckland, New Zealand,
Raffaele Ponzini, High-Performance Computing Group, CILEA Consortium, Milan, Italy and
Giuseppe Passoni, Maritime Hydrodynamics Department, Politecnico di Milano, Italy

Over the last few decades, the repeatability of the measurements, and
development of techniques in compu- the ability to simulate nonstandard
tational fluid dynamics (CFD) together sailing condition scenarios.
with the increasing performance of In the 2003 America’s Cup, in New
hardware and software have helped Zealand, only a few racing syndicates
engineers understand the role of had adopted fluid flow simulation as an
geometrical and mechanical factors on effective design tool, though by the
external aerodynamics in ways that 2007 Cup, in Spain, almost all of the Oil-flow pathlines just above the yacht model
were nearly intractable in the past. In 12 competing teams had recognized surface. The observed tracks, colored by velocity,
recent years, several leading America’s the value of investing resources in both are painted by the wind and simulate a classic wind
tunnel experiment. Converging lines show separating
Cup sailing teams have become top- experimental tests and computational or re-attachment regions.
shelf users of flow simulation software research. Nevertheless, for several
by pushing the envelope of existing technological reasons, there is still a be simulated — from the largest, which
meshing and solver technology. Just a reliability gap between experimental- draw energy from the mean flow, to
decade ago, experiments on physical and simulation-based results. One of the smallest, which are associated
models — using wind tunnels and these is the extremely complex flow with the viscous dissipation that
towing tanks — were the main tools around a racing yacht, particularly in extracts energy as heat. It is possible
for the top teams in their external downwind conditions. to estimate the overall number of cells
aerodynamic and hydrodynamic To design the sail plan for an required to simulate all of the turbulent
investigations. The option of simulating International America’s Cup Class scales. This theoretical cell count is
a number of boat designs in a virtual yacht, a model-scale boat is commonly directly related to the Reynolds
environment has been shown to have tested in a wind tunnel. To perform the number, which is the ratio of inertial
several advantages, including full same test in a virtual environment, all of forces to viscous forces, and it is of the
control of all the parameters involved, the turbulent scales of the wind need to order of 10 billion. If such a number

GFlops in top500.org Ranking Number of Cells in Sail-Plan Computations

100,000,000 1,000,000,000 Viola & Ponzini

1st CFD Grid
10,000,000 500th 100,000,000
1,000,000 10,000,000
Viola
100,000 1,000,000
32nd AC
Number of Cells

10,000 100,000
GFlops

1,000 10,000 31st AC

100 1,000 Miyata

10 100 Hedges
1 10
0 1
1993 1996 1999 2002 2005 2008 1993 1996 1999 2002 2005 2008
Time [Years] Time [Years]

Increasing computational capabilities of the last 15 years, expressed in The increasing trend of the number of cells adopted in downwind CFD
Gflops (billions of floating point operations per second) and published by simulations. Very similar behavior is shown compared to increasing
the official worldwide ranking top500.org computational capabilities, with the exception of the groundbreaking
billion-cell computation.

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 47

ACADEMIC

of cells were achievable, a direct years. By calculating the

numerical simulation (DNS) could be pressure at each cell of the
achieved, which is widely considered to computational mesh, the team
be as accurate as a full-scale measure- could determine the aerodynamic
ment. Unfortunately, generating such coefficients and compare them
a high number of cells would require a with experimental tests
huge amount of memory to create performed in the Politecnico
the mesh and to then perform the di Milano’s twisted flow
fluid dynamics computation. wind tunnel.
Prior to the 2007 Cup races, Running an ANSYS
a state-of-the-art mesh had just Vortex separating from the asymmetric spinnaker. The
FLUENT simulation with
10 million cells, meaning that it was yellow region is evidence for separated flow, and the grey a billion cells — the first
necessary to use turbulence models to regions show the low-speed zones. commercial simulation of its kind
account for the effects of the smallest focused on a single computational
turbulent scales on the mean flow. imported into ANSYS FLUENT flow model — shows the possibility of
In 2008, a researcher involved simulation software and partitioned performing very accurate CFD
in design for a leading contender in into 512 parallel processes so that it modeling in the aerodynamics of
the 2007 Cup [1] collaborated with could be run on CILEA’s powerful downwind sails using leading-edge
members of the CILEA inter-university supercomputer, known as Lagrange. hardware and software resources.
consortium and the Maritime Hydro- Using a hanging-node algorithm, each Though this simulation had 100 times
dynamics Department from Politecnico tetrahedral cell was subdivided into more mesh density than other recent
di Milano to achieve the most realistic eight smaller cells, which grew the CFD computations performed in
simulation of a racing yacht thus mesh to 128 million cells. By repeating America’s Cup boat design, the mesh
attempted [2]. The resulting 1-billion- this procedure a second time, the team was still 10 times coarser than what
cell CFD model was therefore two arrived at a final mesh of just over would be needed to resolve all
orders of magnitude greater than the 1 billion cells. turbulence scales using DNS and,
previous state-of-the-art mesh size in Running on 512 CPUs, the job hence, numerical models, with their
wind engineering. To achieve such an occupied 2 terabytes (TB) of RAM for inherent assumptions, were still
enormous cell count, the researchers just over a week (170 hours) to used to simulate the flow. As high-
reconstructed the sail shapes along complete the calculation of flow performance computing becomes
with a simplified model of the hull velocities and pressures. Performing cheaper and more accessible, however,
and rig using GAMBIT and TGrid such a large calculation in parallel was the DNS goal is in sight. Another
pre-processors from ANSYS, which imperative, as the time necessary for a important goal in the near future of
resulted in an initial grid of 16 million serial process to complete the same racing yacht modeling is to couple the
tetrahedral cells. This grid was then computation would be more than 10 CFD computations with a fully 3-D
shape optimization procedure. This
would overcome the sail designer’s
requirement to perform a physical wind
tunnel test to determine a finite number
of trims for different sails beforehand.
Until that time arrives, though, the
benefits of complementing the global
accuracy of wind tunnel testing with
the local insight of the flow simulation
continue to be affirmed by top
America’s Cup design teams from
around the world. ■
References
[1] Viola, I.M. Downwind Sail Aerodynamics: A
CFD Investigation with High Grid Resolution.
Journal of Ocean Engineering, to be
published in 2009.
[2] Viola, I.M.; Ponzini, R.; Passoni, G.
Downwind Sail Aerodynamics: Large Scale
Computing vs. Large Scale Wind Tunnel Test.
Two leeward (downwind) views, from behind (left) and front (right) showing the air flow velocity over the Journal of Wind Engineering and Industrial
yacht. The low-velocity regions in blue show vortices from the asymmetric spinnaker and the hull. Aerodynamics, submitted in 2009.

48 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 49
TIPS AND TRICKS

Optimizing Options
Technologies converge in ANSYS Workbench
for parametric fluid structure interaction analysis.
By Aashish Watave, Technology Specialist, ANSYS, Inc.

The ANSYS Workbench environment The ANSYS Workbench project

provides a convenient way to conduct schematic shows the two analysis
parametric studies for geometry shape systems used in this project and how
and size variations as well as boundary they are connected to share data with
conditions, which can be extended to cover each other: a fluid flow analysis system
multiphysics simulations within a single using ANSYS FLUENT and an ANSYS
working environment. In this example, the static structural system. The fluid and
effect of a butterfly valve position on the structural systems share the same
fluid flow rate and pressure drop across the geometry. Results from the fluid flow
valve is calculated using ANSYS FLUENT analysis are used to assign boundary
software, and the effect of fluid pressure on conditions for the structural analysis,
Fluid flow simulation of a butterfly valve
the valve deformation is determined using and the blue lines in the project with disk angle of 15 degrees (top) and
ANSYS Structural capabilities. schematic indicate information sharing 50 degrees (bottom)
between different analysis systems.

ANSYS Workbench project schematic Deformation of a butterfly valve (magnified)

as a result of fluid pressure with disk angle
of 15 degrees (top) and 50 degrees (bottom)

1 The geometry can be defined in two ways: It can

be imported from a CAD system and, if necessary,
simplified, repaired or prepared for simulation
using the ANSYS DesignModeler tool within
ANSYS Workbench; alternately, the geometry can
be created in ANSYS DesignModeler. Certain
geometric inputs can be set as parameters in the
tree outline view. For the butterfly valve, the disk
angle is defined as an input parameter, and a
15-degree initial value was assigned.

106 DPI

50 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 TIPS AND TRICKS

2 All the design point parameters are defined in this parameters, and those pointing toward the parameter set
manner using the design point table, which is created bar indicate output parameters from the given system.
as soon as the first parameter is
defined. This table is represented
by a bar called Parameter Set
on the project page. It may be
viewed by double-clicking on
the parameter set bar. Arrows in
the ANSYS Workbench project
schematic pointing toward the
analysis system indicate input

3 For the fluid flow simulation, the geometry is first Meshing and an input parameter in ANSYS FLUENT, and is
meshed using the ANSYS Meshing tool. Zones that shown in the design table on the project schematic as P8-ip1.
are unnecessary for fluid flow simulation, such as the
valve body thickness, are suppressed when meshing
the fluid zones. However, the valve body and valve
disk thickness are maintained as part of the actual
geometry since they are required for the ANSYS
structural analysis. In ANSYS Meshing, the boundary
zones (such as inlet and outlet) required for fluid flow
analysis are identified using named selections. These
named selections are persistent throughout the
project and appear in other tools, such as ANSYS
FLUENT. Setup of the problem in ANSYS FLUENT
proceeds as normal with input parameters defined as
part of the problem setup. For our example, inlet
pressure is defined as a named selection in ANSYS

4 The results obtained from the ANSYS FLUENT surface pressure (on a given surface; for example,
solution can be post-processed using the CFD-Post the valve disk). The output parameters defined in
tool by clicking on the Results cell in the project CFD-Post are available in the design point table as
schematic. Expressions can be used to define output output parameters. Alternatively, output parameters
parameters such as pressure drop (ΔP) or average can be defined in ANSYS FLUENT.

5 The static structural analysis shares the same

geometry used by the fluid flow system. Geometry
parts that are not required for structural simulations,
such as fluid flow volumes, are suppressed in the
ANSYS Meshing tool. For each design point,
the pressure distribution on the valve body and valve
disk obtained from the ANSYS FLUENT results
is used as a boundary condition for the ANSYS
Structural simulation, with the data for the
corresponding zones being mapped and interpolated
between the two systems accordingly. Information
such as maximum deformation and maximum/ updates the geometry and mesh
minimum stress levels from the static structural and then performs both simu-
analysis are defined as output parameters. After lations, including automatically
completing the analysis for a base point (Design Point updating the results in the design
No. 0), additional design points can be directly point table. The project instances
defined in the design point table. Subsequent simu- for each design point can be saved
lation results are obtained by updating all or selected separately using the Export check-
design points from the table. ANSYS Workbench box in the design table. ■

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 51

TIPS AND TRICKS

deselected, and plotting/listing of only the contact or target

Analyzing side is possible from this worksheet. Multiple Contact

Tool branches may also be inserted for reviewing contact
regions in different groups.
The Initial Information branch is included by

Nonlinear default, although users may insert contour results of initial

Status, initial Penetration or initial Gap as well. If you
right-click on Contact Tool and select Generate Initial
Contact Results, the initial contact information will be

Contact calculated and presented in tabular form, as shown in

Figure 2. The rows conveniently summarize the type of
contact and highlight possible problems in different colors:
orange (possibly large penetration or gap), yellow (friction-
Convenient tools help analyze less or frictional contact pair having an initially open state) or
problems in which the contacting red (bonded or no-separation contact initially having an
open state). This allows models with large numbers of
area between touching parts contact regions to be easily examined.
changes during the load history.
By Sheldon Imaoka, Technical Support Engineer, ANSYS, Inc.

In a wide range of structural applications, bonded

contact element methods are sufficient to calculate stresses
between parts in assemblies in which multiple components
are bolted, welded, glued or otherwise joined. In cases such
as gears, cams, levers and other assemblies with moving
parts, however, the contact area between components
Figure 2. Types of contact are summarized, with potential problems highlighted in
changes during the load history. For these types of nonlinear various colors.
analyses, mechanical solutions from ANSYS provide robust
contact technology along with diagnostic tools that can help Contact Result Tracker
obtain converged, accurate solutions to problems that Nonlinear solutions of large models may consume
otherwise would be quite challenging to handle. considerable CPU time, after which users may be
disappointed to find that incorrect model setup or
Initial Contact Information unanticipated contacting areas led to an invalid solution.
Rigid-body motion in which parts are not initially in The ability to track results can help alleviate such prob-
contact is often a common convergence problem. Defining lems. Prior to solving, you can request certain results for
and verifying contact between parts, therefore, is an specific contact regions and monitor these results during
important first step in the analysis. Initial contact status the course of the analysis. Then, if the contact solution
is easily checked in mechanical solutions from ANSYS, starts to deviate from the expected behavior, the analysis
including whether or not parts that are thought to be in initial can be stopped without having to wait until the end of the
contact are truly touching. run to find out that the analysis setup may not be correct.
Using mechanical simulation within the ANSYS To track contact results, drag-and-drop a contact region
Workbench environment, you may insert a Contact Tool branch from the Connections branch to the Solution
underneath the Connections branch, as shown in Information branch. In the Details view of the Result
Figure 1. Specific contact regions can be selected or Tracker that appears, the user may select a number of items
for a given contact region, including but not limited to the num-
ber of contacting elements and the maximum contact pressure.
Add as many Result Tracker items as necessary.
As an example, Figure 3 shows the number of
contacting elements for seven contact regions while the
nonlinear solution is progressing. Note that the contact
region Frictional-seal3 is in near-field (open) contact
throughout the solution. On the other hand, the contact
region Frictional-opening was open until a time of 0.4, when
a large number of elements came into contact. This helps a
Figure 1. The contact tool can be used to check initial contact status.

52 ANSYS Advantage • Volume III, Issue 2, 2009 www.ansys.com

 TIPS AND TRICKS

Number Contacting on Frictional-seal1 Number Contacting on Frictional-seal2 branch. In the Details view, a value of “4” can be entered for
Number Contacting on Frictional-opening Number Contacting on Frictional-upper
Number Contacting on Frictional-lower
Number Contacting on Frictional-seal3
Number Contacting on Frictional-faste the Newton–Raphson residuals. In cases of an incomplete
1.06e+3 solution, contours of Newton–Raphson residuals for the
last four iterations will be available under the Solution
Information branch, and contour plots can be generated
750 as shown in Figure 5. In this example, a solid cylinder pushes
Number Contacting

down on two hollow cylinders; half of the model is displayed.

500 The highest residuals are between the two concentric
hollow cylinders, indicating that the contact stiffness defined
250 for that region may be too high and, consequently, should
be lowered.
-1.
5.e-2 0.1 0.2 0.3 0.4 0.5 0.6 0.692
Time (s)

Figure 3. The number of contacting elements is shown as the nonlinear

solution is progressing.

user understand if each contact region is increasing or

decreasing in the contacting area. If the behavior is
unexpected, the solution may be stopped to examine the
intermediate results.

Nonlinear Diagnostics
The contact stiffness kn is the most important contact
parameter for the penalty-based approach, influencing both
convergence behavior and accuracy. During equilibrium
iterations, if the force residuals plateau (as shown in the Figure 5. Contour plot shows highest residuals at the point of contact between two
example in Figure 4), chances are high that contact stiffness concentric hollow cylinders.
is preventing force convergence. While contact stiffness
may be a cause for the high residuals, you may not be Contact Post-Processing
certain simply by looking at the force convergence behavior. Post-processing is the most important step of any
analysis, and contact problems are no exception. Always
Force Convergence Force Criterion Bisection Occurred
review contour plots of contact status, pressure and pene-
105
26.1 tration to verify that the mesh adequately captures the contact
6.48 behavior and that results are correct. Contact penetration is in
1.61
0.399 units of length, so deformation can be compared in the same
Force (N)

9.89e-2 direction as contact. If the penetration is a small fraction of

2.45e-2
6.08e-2
the deformation, you can safely assume that any variation in
1.51e-3 penetration would not affect results significantly.
3.74e-4
9.27e-5
1. 25. 50. 75. 102

0.1
Time (S)

3.75e-2
0.
1. 25. 50. 75. 102

Figure 4. Contact stiffness might be preventing force convergence if force

residuals plateau.

During the Newton–Raphson iteration, convergence is

achieved when force equilibrium is satisfied. When using Figure 6. Maximum penetration can be compared to deformation to verify that penetration
mechanical simulation in ANSYS Workbench, the user can is negligible.
request Newton–Raphson residual output, so regions of
high out-of-balance forces can be reviewed. This helps in In Figure 6, maximum penetration is 4.256x10-3 mm.
determining where force imbalance is high and, if the area is This value can be compared to the deformation on the same
associated with a contact region, which contact regions contact surface to verify that the penetration is negligible.
may have defined a contact stiffness that is too high. Checking the contact status may indicate that contact
In mechanical simulation within ANSYS Workbench, prior detection is occurring at a very localized region and may
to initiating a solution, select the Solution Information warrant a finer mesh. ■

www.ansys.com ANSYS Advantage • Volume III, Issue 2, 2009 53

ANSYS, Inc.
Southpointe
275 Technology Drive
Canonsburg, PA U.S.A. 15317

Send address corrections to AdvantageAddressChange@ansys.com.