INSIGHTS - AbaqusINSIGHTS is published by Dassault Systèmes Simulia Corp. Rising Sun MiIls 166 Valley Street Providence, RI 02909-2499 Tel. +1 401 276 4400 Fax. +1 401 276 4408

INSIGHTS

Distributed Memory Parallel Enables Faster Turnaround

Glass Service Improve BV Optimizes Bottle Forming

adidas Builds Better Running Shoe

10 2

007

2

Dassault Systèmes Realistic Simulation Magazine

It’s All in theSolder Jointsat Delphi

INSIGHTS is published by Dassault Systèmes Simulia Corp.

Rising Sun MiIls166 Valley Street

Providence, RI 02909-2499Tel. +1 401 276 4400 Fax. +1 401 276 4408

[email protected]

www.simulia.com

Editor:Tim Webb

Associate Editor: Lisa Miller

Contributors:Harun Bayraktar, Vladimir Belsky,

Dermot O’Brien & Paul Jukes (JPK), David Cadge, Dr. Mahesh K. Chengalva

(Delphi), Massimo Donatellis (Pirelli Tyre SpA), Karen Donovan, Matt Dunbar, Brad Heers,

Dr. Jonah Lee (U. of Alaska), Colin Mercer, The Parker Group, Rohit Ramkumar, (Dana Corp.),

Tim Robinson (adidas), Jon Wiening

Graphic Designer:Todd Sabelli

The 3DS logo, SIMULIA, and Abaqus are trademarks or registered trademarks of Dassault Systèmes or its subsidiaries. Other company, product, and service names may be trademarks or service marks of their respective owners. Copyright Dassault Systèmes, 2007.

Solution UpdateDistributed Memory Parallel Enables Faster Turnaround Time

Customer Case Studyadidas Builds Better Running Shoe from Inside Out

Executive LetterColin Mercer, Vice President, Product Development

In The NewsIndustry Press Coverage

Team SIMULIA Tackles the Pan-Mass Challenge

KTM Sportmotorcycle’s RC8 Superbike Gets Ready for the Street

Automatic Meshing Improves Glass Forming Analysis

Realistic Simulation—A Winning Edge for Sports Equipment Designers

••

•

•

•

186

4 16

3

In Each Issue

INSIGHTS

Inside This Issue

Alliances UpdatesEnhanced Powertrain Noise and Vibration Analysis with Abaqus and AVL EXCITE

Intel Creates Program to Simplify Cluster Deployment

•

•

17

Contents

Academic UpdateUniversity of Alaska Researchers Model Tire-Snow Interaction

New SIMULIA Academic Website

•

•

Customer SpotlightJP Kenny Pioneers FEA Software Use in Gas Field Projects

EventsAbaqus Regional Users’ Meetings

2008 Abaqus Users’ Conference Call for Papers

Webinars

••

•

10

9 14

9 14

OC

T_IN

S_Y

07_V

OL

02

October/November 2007

10

8 Product UpdateNew Release of Abaqus for CATIA V5

13 Electronics StrategyRealistic Simulation Strategy for the High Tech Industry

Cover StoryIt’s All in the Solder Joints at Delphi

�INSIGHTS October/November 2007 www.simulia.com

Two years ago, when Abaqus Inc. was acquired by Dassault Systèmes (DS), the SIMULIA Product Development organization was tasked with bringing realistic simulation to the world of Product Lifecycle Management (PLM). The past 24 months have been both exciting and challenging, and I am very proud to be part of a development organization so eager and passionate about delivering high-quality, technically strong products with high value to the end user.

I say exciting because the acquisition presented many opportunities. We merged two complementary groups of developers in the simulation domain: the Abaqus developers, who have experience predominantly focused on products for advanced mechanics users, and the CATIA Analysis developers, who have experience in developing analysis products mainly aimed at design engineers. The newest version of Abaqus for CATIA V5 exemplifies the progress being made by these teams to integrate advanced FEA technology in the design environment of CATIA (see page 8).

Our new organization has also faced challenges as we have transitioned from our origins, with varied histories and cultures, to become a first-class brand within a much larger company. Since 1990, when I first started in the development of Abaqus software, I have been involved in many transformations. The expansion of our product line to include Abaqus/Explicit and Abaqus/CAE was in many ways the most difficult, since it required developers to move from an entrepreneurial mindset with a focus on one product to a more structured environment with an emphasis on process.

By going through this transformation, we have made significant progress in maturing our processes while continuing to develop innovative simulation technology, including major performance improvements in Abaqus Version 6.7 (see page 9). Our progress is best illustrated by the significant increase in the number of product releases in production. We are now managing an order of magnitude more releases per year, including the Abaqus general and extended functionality releases, various Beta test programs, Abaqus for CATIA V5, as well as products for CAD associativity and multiphysics, among others.

Another aspect of being part of a larger company is that our Product Development organization is now geographically located in three countries. Nearly all development is split between the US and France, with a relatively small group in India, which includes a team focused on Quality Assurance. While this geographic distribution has presented communication challenges, the diversity, creativity, and depth of experience within this talented group of people is key to our ongoing success.

Advanced Mechanics development continues to be centered at SIMULIA HQ in Providence, Rhode Island. We continue to grow our staff in this domain by hiring engineers with the strongest mechanics skills from throughout the world. With my roots in computational mechanics, I strongly believe that the majority of this group needs to be co-located so that the interaction between the team members can foster technical innovation.

The broad spectrum of mechanics capabilities in our products—coupled with advances taking place in computational mechanics research, enhancement requests from customers and our perennial efforts to improve usability, performance, and quality—means we must make careful decisions and keep everyone focused on developing and delivering new and enhanced capabilities. We are confident that we have created a strong development organization that can execute on maturing our advanced technology, while being nimble enough to evolve and create new and valuable simulation products.

As we announced at the 2007 Abaqus Users’ Conference, we will soon release our new Simulation Lifecycle Management (SLM) solution. This new product line demonstrates how we are leveraging the extensive resources, technology, and PLM experience within the Dassault Systemes R&D organization. I encourage you to attend your Regional User Meetings and the 2008 AUC to learn more about our new products and initiatives. I look forward to continuing to work with the SIMULIA Product Development team to deliver many more exciting products and enhanced capabilities over the coming years.

Colin Mercer

Executive Message

Ensuring Our Focus on Advanced Technology and Quality

Colin Mercer Vice President, Product Development, SIMULIA

4 INSIGHTS October/November 2007 www.simulia.com

In The News

R&D June 2007, pp. 20–21 Modeling Earthquakes with Realistic Simulation Software More accurate prediction of earthquakes and tsunamis using FEA is the goal of University of Alabama geophysicist Tim Masterlark in this feature article illustrated with 3-D models of tectonic plates and a deformed ocean floor, created using Abaqus software. The researcher uses data collected from the December 2004 Sumatra-Andaman quake and tsunami to simulate the complexity of deformation of the Earth’s surface.

Design News July 2007, p. 54 Dassault Takes Life Cycle Approach to Simulation Dassault Systèmes’ new Simulation Lifecycle Management (SLM) strategy mirrors its life cycle approach to product development, according to contributing software/hardware editor Beth Stackpole in this review of SLM capabilities. The new product portfolio melds the SIMULIA simulation environment with Dassault’s V5 SOA for PLM, helping companies integrate simulation earlier and more effectively into their core product development processes.

For More Information Visit www.simulia.com/news

Industry Press Coverage

Product Design & Development July 2007, pp. 38–39 Foretelling Failure Delphi Electronics research engineer Dr. Mahesh Chengalva (see INSIGHTS cover story and on pp. 10–12) and his group use Abaqus to predict the reliability of solder joints in automobile circuit boards. Months of testing time can be saved by using FEA models to simulate thermal cycling. Delphi’s web-based expert system, D-Cube, puts Abaqus FEA power in the hands of design engineers all over the world, day or night.

Aerospace Engineering August 2007, p. 18 SIMULIA Shaping Future of Simulation Recent announcements about Abaqus Version 6.7, the new multiphysics platform, and the strategic initiative to deliver simulation lifecycle management (SLM) caught the attention of assistant editor Matt Monaghan, who wrote about all these developments in the “Computers in Engineering” section of the magazine. Greg Brown, Product Management, SIMULIA, and Marc Schrank, Director of Crashworthiness and Occupant Safety, SIMULIA, were both quoted in the in-depth article.

Team SIMULIA Tackles the Pan-Mass ChallengeOn August 5, 2007, a group of 19 bike riders from SIMULIA headquarters volunteered to ride 47 miles across parts of Massachusetts in the 28th annual Pan-Mass Challenge (PMC). This was the premiere year for members of Team SIMULIA, who rode to raise money for the Dana-Farber Cancer Institute.

Now one of the largest fundraisers in the United States, the PMC began in 1980 and has evolved to include a cast of thousands who

have raised more than $102 million for cancer research and treatment. Team SIMULIA members were John Ballem, Scott Berkey, Greg Brown, Janet Buckley, Pierre Burgers, David Cadge, Charles Chopin, John Doll, Jon Foster, Glenda Jeffrey, John Mangili, Mike Mathers, Paul Montour, Steven Morse, Peter Nannucci, Matt Rees, Ken Short, Lisa Stein, and Michael Wood.

At the time of this article, the team has surpassed its goal of $25,000 with a significant amount being coming from individual contributions by SIMULIA personnel. The team would like to recognize their corporate sponsors, UVEX, Fuel Belt, CADNEXUS, and Providence Bicycle, who donated directly and through in-kind gifts such as cycling gear, hydration systems, discounts, and moral support.

For More Information To read the full stories highlighted here, please visit www.simulia.com/news/media_coverage

To share your case study, send an e-mail with a brief description of your application to [email protected].

5INSIGHTS October/November 2007 www.simulia.com

In The News

According to an article in the September 2007 issue of Science Daily, the Sports Technology Research Group (STRG) from Loughborough University, UK, uses Abaqus Unified FEA to predict the complex real-world behavior of products on and off the playing field. Designers of sports equipment are already taking advantage of this information to enhance their designs and accelerate time-to-market.

Dr. Andy Harland, STRG leader, explains, “Our computer models can provide invaluable technical input to the sports equipment design process. For example, by enabling the real-world behavior of different design options to be simulated with extreme accuracy, they can reduce the need to manufacture expensive prototypes and cut the time required to get improved equipment from the drawing-board to the shops.”

His group takes advantage of easily customized Abaqus algorithms to complete complex analyses, such as how a soccer ball or running

KTM Sportmotorcycle’s RC8 Superbike Gets Ready for the StreetTo help meet their customers’ performance expectations for their new street bike, KTM Sportmotorcycle AG, Europe’s second largest motorcycle manufacturer, is using Abaqus Unified FEA to evaluate their product’s performance and reliability.

“For the launch of our new Superbike RC8, which will be available to the market in the spring of 2008, we are using Abaqus software to evaluate thermal behavior, noise and vibration, structural performance of assemblies, and stress-based fatigue,” stated D.I. Gunther Hager, R&D Teamleader Engine Structural Analysis, KTM Sportmotorcycle AG. “This integrated simulation process gives us a better understanding of the physical behavior of our products during the design phase.”

KTM is using Abaqus Unified FEA software for all applications in engine and chassis development to analyze the stiffness and flexibility of vehicle frames and engine systems, including gaskets, pistons, and headblocks. They also use AVL EXCITE, which leverages Abaqus technology to study crankshaft dynamics that influence noise and vibration.

Automatic Meshing Improves Glass Forming Analysis Glass Service Improve BV, a provider of advanced solutions for glass manufacturing, is utilizing SIMULIA professional services and Abaqus Unified FEA to accelerate and optimize the development process of non-round glass bottle forming.

In collaboration with Glass Service Improve, SIMULIA has developed an automated remeshing technique to simulate a multistage manufacturing process involving viscous materials that are subjected to high temperatures, large geometry deformation, and complex contact. The innovative process allows designers to accurately evaluate the optimum blow mold process in a virtual environment.

“Customer partnerships, innovative methods development, and advanced simulation technology are core values for SIMULIA,” stated Ken Short, Vice President, SIMULIA Strategy and Marketing. “During a services engagement, our engineering teams work to gain a thorough understanding of our clients’ requirements and processes in order to develop a solution that helps get new products to market reliably and efficiently.”

The new method and the underlying Abaqus FEA technology are central to Glass Service Improve’s Galileo-Plus product, which enables designers to reliably simulate the glass forming process during the design phase. While the methodology has been developed to fulfill the needs of simulating boutique glass forming, like perfume bottles, it may apply to other forming and industrial applications.

shoe will behave on different surfaces when subject to varying stressors. This technology not only saves manufacturers time and money, it also allows designers to focus on specific user needs, enhance safety and durability, and encourage increased participation in sports.

This specially designed kicking robot is used to test the computer models’ accuracy. Image courtesy of Engineering and Physical Sciences Research Council.

Realistic Simulation—A Winning Edge for Sports Equipment Designers

For More Information http://www.sciencedaily.com/releases/2007/09/070913081035.htm

� INSIGHTS October/November 2007 www.simulia.com

Pipeline Design Challenges In a recent gas project approximately 150 miles off the western coast of Australia, J P Kenny applied state-of-the-art Abaqus FEA technology to evaluate the design performance and route mapping for a large system of subsea pipelines. Many critical issues affect subsea pipeline design including the great length of the pipeline, the depth of the sea bed at up to 1350m, the high temperature of the gas at up to 130°, and relatively high pressure at 360barg.

According to Pipeline Business Leader Dermot O’Brien, the mix of high pressure and high temperature also make pipeline material selection and corrosion

management a key design issue. Pipeline wall thickness, corrosion inhibitors, and claddings all impact costs; and FEA is helping in the selection of alternative materials. “The deepwater location of part of the gas field presents clear challenges for the design and installation of large diameter pipeline,” said O’Brien. “The pipeline requirements include about 260km of large gas delivery line and some 520km of small diameter pipe connecting the well heads, manifolds, and other equipment.”

Other issues being considered in the design of the pipeline are the effects of the local marine environment (which features steep escarpments at the continental shelf), the annual cyclone season, large tidal movements, and strong currents, which impact the seabed and the pipeline itself. A number of design criteria and equations fall outside the current design codes, so specialized engineering assessments are being applied.

Evaluating Subsea Loading J P Kenny is pioneering the use of Abaqus FEA to evaluate conceptual designs, perform Front End Engineering Design (FEED) studies, and conduct detailed design studies of the pipeline. Abaqus nonlinear and contact capabilities are readily applied to a deep water environment, coupling the analysis of water and seabed movements and pressures with high temperature, high pressure gas products.

J P Kenny Pioneers Software Use in Gas Field Projects

Customer Spotlight

In response to increasing demand for oil and gas, energy exploration companies are tackling challenging, deep sea projects that have been commercially and technically impossible until now. Advances in FEA software and growing industry knowledge have been instrumental to the success of these new projects. Companies such as J P Kenny are leveraging the latest technology to evaluate their designs and achieve long-term project sustainability.

1km long model

1m element length

Seabed NodesPipe-seabed interactionmodeled using Abaqus‘Contact Pair’ option

Non-LinearEnd AxialSpring to model pipe continuity

Effective Force (lay tension,

pressure, temp)

Pipe Nodes

Abaqus PIPE31pipe elements

SubmergedWeight

Non-Linear EndAxial Spring tomodel pipe continuity

Effective Force (lay tension,

pressure, temp)

Scarp crossing super span model simplified from Abaqus integration of Fledermaus �D view.


It is estimated that the combined characteristics of high pressure/high temperature gas flow and natural seabed behavior could cause a 7m expansion of the main line. The forces associated with this expansion include lateral displacement cycles of the pipe on the seabed of up to 10m. Abaqus is helping engineers develop and test designs to withstand the pipeline dynamics and the forces operating at the continental shelf crossing.

For example, detailed lateral buckling analyses are being done to assess forces, moments, and strains across the full range of behaviors, including ratchetting due to start up and shut down cycles, the cumulative effect that pressure and temperature fluctuations have on the highly stressed apex of the buckle, and the investigation of the potential for pipeline creep or walking.

In addition to helping the pipe lay team understand the loads on the pipeline due to pressure and seabed contact, Abaqus is being used to confirm the location where the pipe line should span the sea floor escarpment at the continental shelf, which drops from 200m to 800m water depth. The pipe lay team successfully integrated the Fledermaus interactive 3-D visualization system from IVS 3D with the Abaqus FEA tool to accurately map the escarpment. At the scarp crossing there is a potential pipeline span of 200 to 300m, so it is crucial that the

pipeline has the structural integrity to cross the escarpment. In addition to the usual lay tension, pressure, temperature, submerged pressure, and axial spring tensions to be calculated, there were additional elements defining the local geo-hazards, including mudflows on the scarp face. J P Kenny found that crossing the escarpment at the optimum point will reduce the pipe lay for the project by up to 40km—a significant cost saving.

Finite element analysis is being done for a complete range of conditions, including empty pipelines, those with operating contents, and those containing flushing media. The sensitivity cases being considered are different pipe outer diameters, wall thicknesses, addition of concrete coating, and residual lay tension. Results from the span analysis reveal the bending moment distribution, longitudinal strain profile and spanning pipeline profile along the route, and modal shapes and frequencies.

“To make the subsea infrastructure more secure, we have used Abaqus FEA software to plan for major event scenarios, including the impact of cyclones on pipe line dynamics,” explained O’Brien. “Abaqus has reduced simulation times, and improved the efficiency and accuracy of pipeline design and route mapping since the team switched from our former FEA tool.”

Meeting Future DemandsJ P Kenny, involved in more than half of all subsea projects around the world, is addressing challenging engineering problems on several cutting-edge projects. Paul Jukes, Advanced Engineering Manager for J P Kenny, suggests that key success factors in developing a viable engineering solution for these projects are employing engineers with a high level of experience, encouraging creative problem solving, and leveraging advanced numerical tools. It is the combination of these factors that is enabling J P Kenny to gain a deeper understanding of subsea pipeline performance.

Reference: Innovative Pipeline and Subsea Engineering Experience by Dr. Paul Jukes, Ph.D., CEng FIMarEST, J P Kenny. Presented at the International Conference on Subsea Technologies, SubSeaTech’ 2007, June 2007, St. Petersburg, Russia.

Image of the preferred pipe lay path dropping from 200m water depth to 800m at about 1�0km off the coast of Western Australia.

Customer Spotlight

For More Information www.jpkenny.com or www.simulia.com/solutions/power.html


Product Update

To meet customer demands for tighter integration and more advanced analysis features, Abaqus for CATIA V5 Version 2.5 offers a new level of ease-of-use and integration with CATIA, further enabling design engineers and expert analysts to access Abaqus technology and collaborate by using the same FEA models and methods. With this product your CATIA V5 models and your Abaqus models become synchronized, making the software a scalable solution for evaluating design performance.

“Abaqus for CATIA V5 allows our design engineers to immediately evaluate how new concepts and design changes affect product performance,” states Michael Thienel, MANN+HUMMEL GMBH. “This approach not only accelerates our development time, but also improves knowledge of our design’s physical behavior, giving us greater confidence in our product quality during the development stage.”

Delivering Enhanced FeaturesThis latest release continues our strategy of providing robust analysis solutions to the extended enterprise. A wizard-based approach greatly simplifies the way users define and manage contact interactions. When the user hits ‘Run,’ all likely contact pairs in the model are automatically detected. There are also several options available to help verify, merge, edit, and delete these detected pairs. This approach makes setup easier and reduces the likelihood of overlooking parts that may be in contact.

Deploying Integrated WorkflowsDesign engineers skilled in using CATIA V5 have access to a wide array of Abaqus simulation capabilities, while Abaqus experts can readily access CATIA V5 models for their analysis work.

Abaqus for CATIA V5 Version 2.5 is a valuable tool for deploying customer-specific workflows across the engineering enterprise. It enables advanced simulation capabilities to be used routinely by design engineers to accelerate and improve product development.

Tight integration with CATIA V5 is one of the most important usability features of Abaqus for CATIA V5 Version 2.5. The new release takes full advantage of CATIA Knowledgeware, publications, and sensors, further enabling the development of analysis templates, performance of interactive design loops, and leveraging of knowledge capture within Abaqus for CATIA V5 models. Specifically, support for publications enables the easy replacement of assembly components while automatically updating the FEA model. Support for sensors allows users to obtain specific locations in the model and to use the results to drive parameters in Knowledgeware formulas.

The latest version also provides support for 64-bit and 32-bit architecture. To take advantage of the 64-bit architecture, you must use Abaqus for CATIA V5 Version 2.5 in conjunction with a 64-bit installation of CATIA V5.

New Release of Abaqus for CATIA Supports Design Analysis Collaboration

CATIA’s Knowledgware tools enable companies to capture design knowledge inside the model and to manage complex design relationships. In this steering knuckle model a design table manages the data for a number of design variables and mandatory load cases.

In models such as this airplane fuselage panel (left) and this personal electronic device (right), the Find Interactions Wizard automatically detects pairs of surfaces (indicated in red and green) that will likely come into contact during the analysis. Setup time is greatly reduced for models with a large number of contact pairs.

Abaqus for CATIA V5 helps us to minimize the number of software interfaces that we use during the development of our exhaust turbocharger,” states Markus Staedeli, ABB Turbo Systems Ltd. “This approach allows us to work seamlessly with the same geometry model throughout our virtual development process and gain significant efficiencies.”

“

For More Information Request the Abaqus for CATIA V5 2.5 webinar replay at www.simulia.com/webinars


For companies relying on finite element analysis to support research and product development, the availability of faster and cheaper hardware has created the opportunity to solve larger more detailed models. For nonlinear implicit finite element analysis, turnaround time is largely determined by the performance of the equation solver. Furthermore, the solver’s relative contribution to analysis time typically increases as the problem size increases, making enhancements to solver performance even more crucial.

At SIMULIA we have invested significant resources into developing a high-performance, scalable distributed memory parallel (DMP) direct sparse solver. Abaqus Version 6.6 marked the first release in which users could run Abaqus/Standard on compute clusters while enjoying the robustness and accuracy of a direct sparse solver. In Abaqus Version 6.7 the solver has been enhanced to include expanded feature coverage, reduced memory usage, and improved performance and scalability. The improved scalability is particularly noticeable in large models with over a million degrees of freedom.

Benchmark ModelsDuring software development, we work closely with our customers to benchmark real-life industrial applications. This article demonstrates the benefits of running Abaqus on a compute cluster using a powertrain model (courtesy of Dana Corp.) and a tire model (courtesy of Pirelli Tyre S.p.A).Both are nonlinear analyses, which include contact and nonlinear material behavior.

The scalability of Abaqus/Standard was measured by calculating the speed-up for jobs run on up to 16 compute nodes (64 cores) with respect to jobs run on 1 node (4 cores). All jobs were run on a Hewlett-Packard XC3000BL Linux cluster with two Intel 3.0GHz Core2 Duo processors on each node. The powertrain model was run using Abaqus Version 6.7-1 and the tire model was run using a prerelease of Abaqus Version 6.7-EF (required for DMP execution of the steady-state transport procedure).

The results in Figure 1 show that the scaling of the powertrain model is much better than the smaller tire model. The speed-up of almost 11 for the powertrain analysis on 64 cores means that this analysis, which takes 60 hours (2.5 days) on a single compute node using 4 cores, can be completed in less than 6 hours on 64 cores.

In addition to completing the analysis much faster, a simple analysis token utilization calculation shows software costs for the powertrain model are 67% lower for the run on 64 cores. For the tire model, running on 32 cores is 46% cheaper than running on 4 cores. Summary These benchmark models clearly demonstrate the cost-effectiveness of running large nonlinear simulations in Abaqus/Standard on compute clusters. The scalability of Abaqus/Standard is capable of reducing analysis time from days to hours for large models. For medium size models, significant reductions in turnaround

time can also be achieved which justify parallel execution, particularly when taking into account token hours cost. As we work to improve the scalability of Abaqus/Standard, we are also pushing the boundary on the size of the models that can be solved. Recently, we have run in-house models with over 25 million degrees of freedom, and we are looking at scalability on up to 512 cores using Abaqus/Standard.

For those customers with large and medium size models, we encourage you to leverage the documentation, training, and support resources available to you to take advantage of DMP on compute clusters. The solver performance improvements can increase productivity significantly.

SIMULIA Authors: Harun Bayraktar, Developer, Implicit Analysis/Solvers; Matt Dunbar, Manager, Platform Specialists; Vladimir Belsky, Director, Solver Technology.

Special thanks to: Rohit Ramkumar, Dana Corp., and Massimo Donatellis, Pirelli Tyre S.p.A.

For More Information Request a replay of the DMP Webinar at www.simulia.com/webinars

Accelerating Nonlinear Analysis Using Abaqus/Standard on Compute Clusters

Powertrain TireFootprint Steady-

state transport

Elements 1.25M �50K �50K

Nodes �.28M 448K 448K

Equations (dofs) 9.��M 79�K 79�K

FLOPS per iteration �.�2E+1� 2.�9E+12 4.�8E+12

Steps 10 � 2

Iterations �5 51 �8

Pertinent Model and Analysis Information

Figure 1 Speed-up results for the analyses in comparison to ideal (linear) speed-up. Speed-up is the ratio of the analysis wall clock times on 4 cores and N cores (N=4, 8, 1�, �2, �4).

1�

14

12

10

2

4

�

8

0�4�2 40 48 5�8 1� 240

Spee

d U

p

Total Number of Cores

PowertrainTire-TotalTire-FootprintTire-SSTIdeal

Solution Update

Powertrain model. Courtesy of Dana Corp.


generation of automotive electronics. As one of the largest automotive systems suppliers in the United States, Delphi runs 170 manufacturing facilities in more than 30 countries. On a typical day, the Kokomo facility alone produces over a million IC packages.

In electronics design and manufacture, package reliability is sacrosanct. The reliability of electronics systems in automotive applications hinges on the reliability of every solder joint on every circuit board. A major factor in solder joint reliability is thermal cycling. “The single largest driver of failure in electronics systems in the field is the repeated temperature cycling that the product

There can be as many as 2,000 solder joints in a high-volume powertrain controller board such as this.

Lately it’s hard not to notice that automobiles have morphed into circuit boards on wheels. Electronics now control every important system in the average passenger vehicle, from entertainment and navigation to chassis and steering, brakes, security, and powertrain. A dazzling menagerie of electronic components populates automotive printed circuit board assemblies.

By Karen Donovan

“In terms of cost, there is more electronic content in vehicles today than even steel,” says Dr. Mahesh Chengalva, Staff Research Engineer at Delphi. “The performance requirements are challenging, and these systems are expected to last ten years or more. It’s a severe vibration and thermal environment, especially for underhood electronics.”

Chengalva is part of the Hardware Analysis and Simulation group at Delphi Electronics & Safety, a division of Delphi Corporation that is headquartered in Kokomo, Indiana. His group provides analysis services to squadrons of Delphi design engineers at work all over the world on the next

Cross-sections of leaded IC packages before and after extensive thermal cycling, showing copper leads soldered to the contact pad on the circuit board. Delphi gathers data on the thermal cycling failure of solder joints via extensive physical testing, which provides raw data for comparing and validating simulation results.

View of a typical leaded IC package assembled onto a high-density on-engine automotive controller board. The electrical connections are made by leads projecting from the package body that are soldered onto the circuit board.

Cover Story

It’s All in the Solder Joints


undergoes,” says Chengalva. “Eventually, in this harsh thermal environment, a solder joint will develop a crack, and we want to be able to predict when that will happen.”

Considering that the daily production of solder joints at Delphi is about one billion, the task of predicting thermal cycling reliability for all the electronic packages under design calls for very special tools. Chengalva’s simulation group has tackled the challenge from two sides. For analysis power, they rely on the advanced nonlinear solution capabilities of Abaqus simulation software from SIMULIA, the Dassault Systèmes brand for realistic simulation. To speed analysis throughput, they offer Delphi design engineers direct access to the Abaqus solver by means of a turnkey web-based expert system called D-Cube.

Simulating Solder JointsTo study solder joint performance and ensure field reliability, Delphi subjects its electronics products to a battery of thermal cycling tests that replicate automotive operating environments. Some tests require as many as a thousand or more thermal cycles, where a package is repeatedly heated and cooled between two temperature extremes, until solder joint failure occurs. For underhood applications, these extremes can range from −40˚C to 125˚C. Thermal-cycling tests are time consuming (a 3,000-cycle test would take four months), so the ability to simulate reliability testing gives Delphi a significant advantage in faster time to market.

Creating a valid FEA model for predicting thermal cycling reliability starts with characterization of material properties. Delphi has amassed a substantial amount of laboratory creep test data on the material

properties of solder and has developed validated material models for use with Abaqus. “Unlike steel or copper or ordinary plastic, solder is extremely complicated,” says Chengalva. “Its behavior is time dependent, temperature dependent, and stress dependent. This behavior cannot be simulated accurately without very in-depth nonlinear analysis tools.”

Along with the material model, Chengalva’s group has developed a library of prebuilt thermal-cycling reliability models for many families of electronics devices. These range from single components such as capacitors and resistors to full packages such as ball grid arrays, flip chip assemblies, and a variety of surface-mount leaded ICs. To validate the FEA models for each electronic package, the team simulates a set of thermal cycling tests and compares the results of simulations with the results of tests. “All our FEA models have been carefully correlated with test data,” he says, “so our design engineers can be confident that simulations give accurate predictions of package reliability.”

The D-Cube DoorwayIn fact, getting accurate simulation results for thermal reliability has never been easier for Delphi design engineers. They have Abaqus FEA power at their fingertips any time of the day or night, anywhere in the world, thanks to the D-Cube web-based expert system. Up and running on the Web since May 2000, the company-wide site offers a collection of over 200 design optimization tools, materials resources, self-taught tutorials, and notes on engineering lessons learned.

The brainchild of Chengalva and Scott Baxter, a former Delphi colleague, D-Cube

is fundamental to the company’s strategy of incorporating simulation upfront in design. To illustrate, Chengalva describes a scenario where a Delphi design engineer in Singapore needs to select a suitable ball grid array package from a number of suppliers. “The designer knows what he wants for package reliability,” he says, “but he needs a way to evaluate each supplier’s product and pick the best one. All he has to do is open the thermal tools section in D-Cube, click on the ball grid array icon, fill out an online form with input parameters, and hit submit.”

Input parameters vary with the type of package. Detailed dimensions of the package assembly, and material properties such as the coefficient of thermal expansion for each part in the assembly, are commonly required input. Once the designer submits the data, the input parameters go to Chengalva’s group for scheduling. The next step, FEA modeling, has been automated to cut preprocessing time from several days for complicated assemblies to about an hour at the most. In the case of this ball grid aray, a prebuilt parametric modeler would automatically create a model for the package the designer has specified. Regions of interest in the package geometry are automatically identified for greater mesh refinement.

Finally, the input file goes to the Abaqus solver. Complex three-dimensional simulations might take three or four days of continuous run time. “Once the simulation is complete,” says Chengalva, “the design engineer receives a reliability value. Perhaps this particular package, from this particular supplier, for this particular thermal-cycling profile will last 3,400 cycles. That number is what the designer needs. We can also compute a value for relative life prediction, (Story continued on page 12)

Cover Story

View of a ball grid array (BGA). These versatile components often form the brains of an electronics package. For example, the CPU of a powertrain controller is often a BGA. The electrical connections with the board are made through solder balls. Predicting the thermal cycling reliability of each one of these solder balls is important because a single failure can cause vehicle malfunction.

Sample contour plot of FEA results in Abaqus, showing equivalent creep strain for a ball grid array under thermal load. Thermal expansion mismatch between the BGA package and the circuit board causes differential movement between the top and bottom surfaces of the solder balls with every temperature cycle, eventually leading to fatigue failures.


For More InformationVisit www.delphi.com or www.simulia.com/solutions

which helps the designer compare one package with another. Relative life predictions are particularly useful when input data are incomplete.”

It is interesting to note that D-Cube helps tie together the far-flung global centers of Delphi analysis into a single integrated unit. In the example mentioned above, although the request originated in Singapore, it would be processed in Kokomo (USA), with the actual FEA modeling and execution conducted in either Krakow (Poland) or Bangalore (India).

Screen views of Delphi’s D-Cube web-based expert system. At left is a menu of tools for performing thermal cycling reliability analysis with Abaqus. At right are parameter input fields for analyzing the thermal cycling reliability of a ball grid array. A prebuilt parametric modeler automatically utilizes the designer’s input to create an analysis model for the simulation.

The FEA process automation that D-Cube offers design engineers is an efficient way to assess electronic packages, especially when numerous competing design configurations are being considered. “In electronics, parts may look identical, but their reliability can vary significantly, based on the mechanical structure or on the materials used,” says Chengalva. “For a leaded IC that is used on a FR4 substrate, if you change the material of the lead from copper to Alloy 42, you could have almost a tenfold decrease in solder joint reliability. If not selected carefully, additions such as conformal coats, which are applied to some circuit board assemblies to prevent moisture damage, can interact mechanically with components to degrade reliability. For design engineers, information about reliability as related to these multiple parameters is powerful because it enables tradeoff decisions very early in the design cycle.”

Simulating thermal-cycling reliability using Abaqus FEA saves Delphi an enormous amount of time and cost and opens doors to design innovation. With D-Cube, design engineers can quickly use Abaqus to explore different configurations for automotive electronics packages and select the optimal

solution. You might remember that the next time one of those rolling circuit boards passes you on the highway safely at 80 mph.

Mahesh K. Chengalva is a veteran Abaqus user and owner of 11 patents and trade secrets. He has worked for Delphi for 12 years, and was

inducted into the Delphi Hall of Fame in May 2006. He is currently technical manager of their mechanical analysis laboratory in their Electronics and Safety division. At Delphi, Mahesh conducts reliability simulations for electronics systems, manages the company’s web-based engineering expert system, provides consulting services to internal customers, and interfaces with overseas technical centers. In his spare time, he is also a founding member of a robot racing team headed for the semi-finals of the DARPA Grand Challenge this October.

Close-up view of an audio system board showing chip components, some as small as 2 mm in length. Since these components lack leads or other compliant structures, the entire thermal mismatch between component and board is borne by the solder joints.

Cover Story

1�INSIGHTS October/November 2007 www.simulia.com

Electronics Strategy

The competitive high tech industry requires innovative new products to be brought to market extremely quickly. In addition to time pressures, manufacturers face the challenge of squeezing more capabilities into smaller, yet design-conscious packages. The industry is also faced with regulatory compliance issues such as the elimination of lead-based solder, which has been in use since the beginning of the electronics revolution.

SIMULIA has actively engaged with a number of high-tech customers to gain a deeper understanding of their processes and simulation requirements. As a result, we seek to deliver more simulation functionality within easier-to-deploy environments. For example, we are making functionality like fracture modeling and advanced time- and temperature-dependent material models more accessible to designers though automated model creation, enhanced material libraries, and integrated multiphysics capabilities.

Electronics-Focused Capabilities Our Unified FEA strategy leverages existing Abaqus strengths to allow a single model definition and user environment to be used for multiple simulation types such as package reliability, thermal fatigue, moisture sensitivity, shock, and drop. This approach is being applied by several cell phone manufacturers to leverage their Abaqus models to simulate bending, thermal load, drop, and acoustics.

The Delphi article on page 10 of this issue provides an excellent example of the types of custom interfaces that are able to be created with Abaqus/CAE. These interfaces enable efficient and automated modeling for electronics applications such as ball grid arrays. Similar approaches have been taken by

other electronics customers to capture expert-generated workflows of advanced simulation, which are then packaged, automated, and deployed to design engineers.

SIMULIA is known for the development and deployment of innovative simulation methods. The direct cyclic solution method is a unique capability in Abaqus to automatically calculate the stabilized, steady-state response of a structure to cyclic loading, thus increasing solution accuracy and reducing solution time compared to traditional methods. In a recent IMECE paper, engineers at Intel described how they applied the direct cyclic solution method for thermal fatigue analysis of electronic packages.

We continue to make advances in our state-of-the-art material modeling capabilities by including special models for analyzing lead-free solder. Integrated and coupled multiphysics capabilities support robust coupled-field analyses of thermal, electrical, mechanical (both static and dynamic), and moisture-sensitivity load regimes. We have industry-leading capabilities for progressive fracture-failure modeling and nonlinear contact, and our recent Abaqus releases have also delivered best-in-class solver performance, allowing larger and more realistic models to be solved in much less time.

As a result of our focused high tech strategy, our products and services are gaining momentum and growing rapidly within this industry. To support the market, we plan to host a series of half-day electronics seminars in select locations to enable customers to interact directly with our technical staff and gain first-hand knowledge of our latest technology. In addition, we have developed a Technology Brief and Electronics Webinar; both are available for download from the SIMULIA website. If you have specific requirements for high tech and electronics packaging simulation, please contact your local office. We look forward to working with you to provide the most robust high tech industry solutions on the market.

David Cadge has worked with SIMULIA since 1995 in various capacities within

our customer service and marketing teams. David initially worked at our SIMULIA UK office and then transferred to our Providence, RI, headquarters in 2003, where he is now responsible for developing and promoting our strategy for simulation within the high tech industry. He has traveled extensively, visiting our high tech customers around the world to understand their workflows and requirements while promoting our current capabilities and best practices for this industry. Information gathered during these visits will help us provide enhancements for advanced technology, usability, and productivity so that simulation can become an integral part of electronics design practices.

Realistic Simulation Strategy for the High Tech Industry

For More InformationAbaqus Extensions PCB modeler www.simulia.com/extensions

Abaqus Users’ Conference 2007 Paper Temperature Cycling Analysis of Lead-Free Solder Joints in Electronic Packaging by Shan Li, Zhenyu Huang, Jianfeng Wang, Intel Technology Development Ltd., Shanghai, China and Shaowu Gao, SIMULIA China. www.simulia.com/solutions/high_tech

Technology Brief Creep Analysis of Lead-Free Solders Undergoing Thermal Loading www.simulia.com/tech_briefs

Webinar Validate Electronics Packaging Performance with Realistic Simulation www.simulia.com/webinars

Screen capture of a PCB modeler in Abaqus

David Cadge Senior Solutions Engineer, Electronics, SIMULIA


Check out a store display of running shoes and you may conclude that it’s all about style, color, and fashion. In fact, people in the sports equipment industry admit that over half of all athletic shoes are bought solely for their looks.

But those people who actually run in their shoes—recreationally or professionally—will be pleased to hear that the designers and engineers at adidas are paying attention to more than flashy exteriors. They know that shin splints, a twisted knee, or one bad turn of the ankle can ruin any runner’s day. Using materials science, biomechanics, and finite element analysis (FEA) software, they are building motion-responsive, stress-reducing technologies—similar to ones found in earthquake-resistant buildings and car stability systems—right into the heels of their running shoes.

The shoes still look good, of course. But the best parts of the sneaker are now definitely those you can’t see.

Simulation Software is Key Historically, new athletic shoe models have been based on a retooling of existing ones, necessitating time-consuming rounds of real-world prototype testing and redesign. But in creating their new ForMotion™

line of running shoes, adidas has achieved significant design cycle reduction by coupling their surface-based computer aided design (CAD) tool with FEA software (Abaqus/Explicit and Abaqus/CAE) from SIMULIA.

“If you were going to retool and test a whole new shoe it could take six to eight weeks,” says adidas Mechanical Engineer Tim Robinson. “But when you run simulations it only takes a few days. In terms of speeding up the entire development process, the FEA software has proved invaluable.”

Robinson is part of the adidas innovation team made up of designers, engineers, and biomechanists working together from two locations: adidas’ birthplace outside Nuremberg, Germany, and Portland, Oregon, U.S.A. Declaring 2007 to be the Year of Running, adidas is releasing some details (but not all—the sports shoe industry is as secretive about R&D as anyone else) about how they create their high-tech products.

Unique Sliding Plate TechnologyA key element in the creation of the ForMotion™ line of running shoes is a unit comprised of a pair of sliding polyamide plates with a spherical bearing (similar to those earthquake protection systems in

buildings) located in the heel of the shoe. Stiffness is provided to the system by way of a thin polymer seal, and sometimes also polymer springs and washers, depending on the model.

“Instead of just absorbing the shock in a vertical axis, the sliding plates adjust to shock in three dimensions—vertical, horizontal, and rotational,” Robinson says. “Car banking systems also use a similar technology. There’s a lot of cross-fertilization within engineering these days.”

The heel unit benefits the runner in a number of important ways. For one, decoupling the heel with sliding plates slows down “sole angle velocity,” the speed at which your toes slap the ground after your heel makes contact—when it happens too fast you can end up with the common runner’s complaint called shin splints.

The plates can also rotate to counteract pronation, the amount your foot rolls inwards with each step (most people pronate to some degree). Too much pronation is one of the biggest causes of knee injury in running. The spherical plates automatically correct for this, with the rotational motion “give” protecting the knee. Overstraining in the opposite direction is prevented by displacement limiters.

adidas achieved significant design cycle reduction for

their ForMotion™ line of running shoes by coupling surface-based

computer aided design (CAD) with Abaqus finite element analysis software. A

unique sliding plate technology in the heel of the shoe absorbs shock in three dimensions. Wear testing

of the heel unit, using a bespoke test device, reproduces the motion and impact of an actual runner using the shoes.

Customer Case Study

Quick! Build a Better Running Shoe from the Inside Out


And finally, the unit automatically adjusts to each individual foot strike (how turned out your feet are when you run). Women tend to land straighter than men, notes Robinson. “Men can land at up to 30˚ away from a straight forward line. We’ve been able to engineer the heel unit to perform the same no matter which angle you land on the shoe.”

Turning the Concept into RealityWith the concept in hand, the next step for the innovation team is ensuring that the heel unit works as intended in an actual running shoe and can also be functionally graded to different shoe sizes.

Surfaced-based CAD modeling is the starting point at adidas, due to the design freedom it allows when dealing with the very organic geometry of objects like shoes. From the CAD models, 3D meshes are created and imported into Abaqus/CAE. “We create the modeling interface and then use it to apply boundary conditions and material properties to the mesh, and tie everything together defining contact, friction, connectors, and other factors,” says Robinson. “Then we solve it using Abaqus/Explicit.”

Material characteristics come from custom libraries adidas creates based on their highly proprietary materials. Biomechanics data, the basis of the boundary conditions, are equally confidential; they are generated in-house using a Vicon motion-capture system of tiny markers placed at strategic points on a human runner, which are then tracked with infrared cameras as the person moves.

Seeing Inside the Virtual ShoeFor postprocessing, Abaqus/CAE is used again. The engineers look for areas of strain and/or friction by applying simulated horizontal and vertical displacement forces. “At this stage of development it’s definitely very useful to be able to ‘see’ inside the actual shoe, and even cut through the modeling for a closer look, especially when we’re dealing with a complex unit including plates, seal, springs and washers,” Robinson notes. “It greatly helps communication with our designers about any changes we recommend.”

Since human feet—and bodies—come in many different sizes, an important part of FEA modeling at this stage is the functional grading of the shoes. “From our biomechanics results we have data about the correlation between shoe size and average weight,” Robinson points out. “The radius above the ankle joint will change with a larger size. We use this relationship to functionally grade the shape of the plates, employing FEA to predict the optimum stiffness of the heel unit required for each shoe size.” The shoes also come in four variations for control, cushion, trail, and competition.

Testing, TestingWhen prototypes are ready for testing in the real world, a bespoke adidas test device is put to use. Located just a few meters from the workstations running FEA software, it is used to measure the “horizontal stiffness” of the ForMotion™ shoes. The shoe is held at an angle of 30˚—the “male” foot strike angle—and a quasistatic displacement is applied to the heel. This non-destructive test is also used later for quality control testing of shoes taken directly from the factory line.

At this point there are rarely any surprises, Robinson notes. “Since we’ve already used FEA to run simulations that are comparable to movements of our test devices, we are very confident that when we get the first sample shoes back they will be very close to what we predicted them to be.”

Wear testing is done using another bespoke device, Performance Evaluation Test Equipment (PETE). This uses the previously obtained Vicon biometric input to reproduce the motion and impact of an actual runner wearing the shoes. Testing runs up to 100,000 cycles, comparable to 1000 km of running.

All that running will now take less of a toll on legs, thanks to adidas’ ForMotion line of shoes. So with shins and knees now better protected, can hips be far behind? “That’s a question for our biomechanists,” says Robinson. “Engineering shoes to work with hips might be next. We are continually making improvements using Abaqus software to advance the science of shoe design.”

For More Informationwww.adidas.com or www.simulia.com/solutions

External side view strain contour plots generated with Abaqus/CAE show the results of a biomechanical loading of the heel unit. View A shows the undeformed state, and B is the deformed state. In the first round of virtual prototypes, this information warns the engineers about areas that could be prone to excessive strain and lets them modify the design accordingly. Forces from the FEA outcome compared to mechanical tests from bespoke equipment show a very high degree of correlation.

A B

Internal cut view strain contour plot generated with Abaqus/CAE showing biomechanical loading of the heel unit. View A shows the undeformed state, and B is the deformed state. This function in Abaqus/CAE software allows engineers to cut sections through virtual shoes. This is extremely useful when analyzing the movement of internal components and making any necessary design changes.

A B

Customer Case Study

adidas, the adidas logo, ADIPRENE, and the �-Stripe trade mark are registered trademarks of the adidas Group, used with permission. FORMOTION and geoFIT are trademarks of the adidas Group, used with permission.

1� INSIGHTS October/November 2007 www.simulia.com

University of Alaska Researchers Model Tire-Snow Interaction for Improved Safety

For More Information Visit Dr. Lee’s webpage at www.faculty.uaf.edu/ffjhl/.

bead

sidewall

tread beltcarcass

rebar

rebar

Finite element tire model

Contact shear stress (CSHEAR2) for FZ= �,240N, α=8, and ix =0.407

Academic Update

and snow sinkage depend significantly on the longitudinal slip, slip angle, and vertical load. It was also concluded that vehicle-snow interaction modeling methods and results are similar to those for vehicle-sand interfaces, allowing this research to be shared by those modeling off-road tire behavior under desert conditions.

Dr. Jonah H. Lee is Professor and Chair of Mechanical Engineering at the University of Alaska Fairbanks (UAF), Alaska, U.S.A. He received his Ph.D. in Engineering Mechanics from Iowa State University in 1983. His recent research interests are in the area of vehicle-snow interaction and structure-property relationships of random heterogeneous materials.

Studying vehicle and tire behavior in arctic conditions creates a challenging simulation environment. Researchers encounter highly nonlinear conditions in sinkable terrain with low friction surfaces and uneven or uncertain ground, which cause complex, varying pressure, friction, stress, and contact distribution. Tire-snow interaction has been studied for decades at the U.S. Army Cold Regions Research and Engineering Laboratory. Extensive modeling, simulation, and tests have been conducted for a variety of vehicles. However, the studies reviewed did not consider the effects of controlled longitudinal slip, lateral slip, or wheel load. In addition, the traction and motion resistance were not separated from the reaction force at the hub of the wheel. Consequently, the off-terrain wheel states were not identified. In addition, contact stresses were not studied as a function of wheel state.

Professors Jonah Lee and Qing Liu of the Vehicle Terrain Interaction Group (VTIG) at the University of Alaska Fairbanks are building on previous research by using Abaqus to evaluate three-dimensional finite element simulations of pneumatic tire-snow interaction. Such analyses enable them to gain insight into contact mechanisms and obtain associated interfacial stresses and forces, sinkage of snow, and deflection of the tire as a function of vertical load under combined slip conditions where the longitudinal and lateral slips are explicitly controlled.

The pneumatic tire was modeled using elastic, viscoelastic, and hyperelastic material models; the snow was modeled using the modified Drucker-Prager cap plasticity model. The traction, motion

resistance, drawbar pull, snow sinkage, tire deflection, contact pressure, and contact shear stresses were obtained under combined slip conditions as well as the wheel states.

Since the deflection of a tire on snow is smaller than on the rigid ground, the tire can be considered as rigid; however, the deflection can influence the actual value of longitudinal slip. Due to the fact that the full range of longitudinal slip was prescribed in the simulations, the off-terrain wheel states were identified, which served as an indicator to categorize the interfacial forces and contact stresses. Trends of the distribution of contact pressure as a function of wheel state and vertical load were established. Multiple zones of the distribution of shear stresses at the contact patch were identified and analyzed as a function of wheel state.

The simulations indicated that the length of contact patch increases slightly with vertical load and that the interfacial forces

Learn more about what researchers, students, and teachers are accomplishing with the Abaqus Unified FEA product suite. Our academic website features compelling state-of-the-art Abaqus research applications as well as examples of innovative student projects that span many industries such as consumer goods, crashworthiness, high tech, and life sciences. Purchase Abaqus Student Edition online or find out how you may use our Academic and Research Editions at your university.

NEW! SIMULIA Academic Website

For More Information Visit www.simulia.com/academics

Contact between the side surface of tire and snow due to rotation of tire (zone 3, negative stress)

Contact between the side surface of tire and snow due to downward motion of tire (zone 1, positive stress)

Contact between the side surface of tire and snow due to rotation and lateral motion of tire (zone 4, positive stress)

Contact between the bottom of tire and snow due to lateral motion of tire (zone 2, transition from positive to negative)


According to IDC, a global provider of market information for the information technology markets, compute clusters captured more than 50% of technical server revenue in 2006. However, IDC is projecting a slow down in growth that they attribute to the increased complexity in deploying and managing compute clusters.

The Intel® Cluster Ready program addresses this industry concern by establishing a reference architecture for Intel-based systems. Hardware vendors can use this architecture to certify their configurations, and ISVs, such as SIMULIA, can use it to test and register their applications. The chief goal of this program is to ensure fundamental hardware/software integration, so that end users can focus on getting their job done, knowing that Intel® Cluster Ready ISV applications will run on their cluster that has been certified as Intel® Cluster Ready.

The Intel® program technology provides a standardized and replicable way to build clusters and run “off-the-shelf” high-performance applications. It includes a specification that sets minimum standards for software and hardware components.

A central component of the platform certification process is the Intel® Cluster Checker. This tool checks the cluster hardware and software components to ensure that the cluster was built according to the specification and that the cluster is operational before an ISV code is loaded. It also provides a fault isolation tool for early detection of cluster problems which helps improve productivity and reduce support costs.

To register their software applications as Intel® Cluster Ready, ISVs must test their applications on certified platforms with realistic workloads. Of particular importance to end-users, the ISV must

For More Information www.intel.com/go/cluster/

Designing low noise engines is a challenge due to the contrary demands of using lightweight materials while increasing the power and performance of vehicles. To meet these challenges, AVL and SIMULIA are leveraging significant advances in substructure (“superelement”) capabilities of Abaqus with AVL EXCITE to provide compelling new functionality for powertrain noise and vibration analysis.

A typical engine simulation model consists of various parts (bodies) such as a crankshaft, connecting rods, and an engine cylinder block, coupled with nonlinear contact (joints) such as journal bearings. The mass/moment of inertia and elasticity of these bodies are included by using Abaqus. This method enables engineers to calculate the dynamic behavior of powertrains, including the influence of hydrodynamic contact under realistic running engine conditions. By conducting a transient engine run-up simulation, the dynamic behavior in the full range of engine speed and loading condition may be investigated.

Integrated Analysis WorkflowA noise and vibration analysis begins with the generation of FE models for engine components such as the crankshaft, connecting rod, and power unit. The first analysis step is a natural frequency analysis of all parts using Abaqus as a model check and to investigate dynamic behavior. A combined static and dynamic condensation of the FE models in Abaqus provides reduced mass and stiffness matrices for the following nonlinear forced vibration calculation with AVL EXCITE.

AVL EXCITE performs an internal recovery of the results from Abaqus using a transformation matrix. This enables

powertrain engineers to obtain all motion quantities and structural vibrations for the entire FE model (Figure 1).

This process provides the necessary integration of simulation methodologies to automate data transfer, improve workflow efficiencies, and gain improvements in simulation fidelity. In addition, leveraging the latest improvements, Abaqus/AMS engineers are able to significantly shorten project turnaround times and reduce costs in reaching performance targets.

Intel Creates Program to Simplify Cluster Deployment

Improving Powertrain Noise & Vibration Analysis

For More Information www.avl.com or www.simulia.com/solutions

Alliances

describe any software dependencies beyond those required by the Intel® Cluster Ready Specification; this ensures the smooth installation and proper execution of registered ISV applications on any certified cluster. SIMULIA engineers have tested and registered both Version 6.7-1 and Version 6.7-2 of Abaqus/Standard and Abaqus/Explicit to run on clusters certified as Intel® Cluster Ready.

Total Surface Integral Level

Engine Selection Area

500

�000 Speed [rpm]

5dB

Leve

l [dB

]

Frequency [Hz]

Figure 1: Result evaluation and determination of critical speeds and frequencies based on structure borne noise results (inline 4-cylinder engine)


Events

2007 Abaqus Regional Users’ Meetings and Update SeminarsOur regional offices and representatives are in the midst of hosting the 2007 Regional Users’ Meetings (RUMs). The twenty-six RUMS held around the world provide a local forum to learn about innovative analysis methods and how our customers are applying Abaqus FEA technology within their industry or academic institution. More than 2,000 attendees will learn first-hand about the latest enhancements to Abaqus and the overall SIMULIA strategy. The meetings also offer our customers the opportunity to provide input on future software enhancements, network with other users, and share experiences and new ideas.

While most RUMs are held September through November, Abaqus users in China, Korea, and South Africa gathered in June and July. Our China RUM, held in Guilin, was the gathering place for more than 160 attendees. Of the 84 papers that were published in the proceedings, eight received awards, and four were presented at the conference.

The Korea RUM was held in Daejon, with more than 150 Abaqus users in atendance. Among the papers presented, LG Chemistry provided insight into creep behavior of plastics, Schaeffler Korea Corporation discussed contact stress analysis of seal mechanism of a bearing, and Kumho Tire reviewed the simulation of the driving performance for a vehicle tire.

As this issue goes to press, RUMs in Erie, Central, Scandinavia, and Deutschland will have just concluded. We trust that you were able to take advantage of the meeting in your area. For those of you in regions that still have upcoming meetings on the calendar, we encourage you to take part in these conveniently located and valuable events.

Abaqus experts from our regional offices and our headquarters will be on-hand to provide specific product and technology information and answer your questions. As an additional educational opportunity, several of the regions are including another seminar either before or after the user meeting.

To learn more about the Regional Users’ Meeting in your area, contact your local office or visit the events page on the SIMULIA website.

AmericasDate LocationOctober 1–2 Córdoba, Argentina

October 17 Houston, TX

October 22–24 Las Vegas, NV

November 5–� Toronto, Ontario

November 7–8 Plymouth, MI

November 14–15 Providence, RI

Asia PacificDate LocationOctober 24 Bangalore, India

October 29–�0 Malaysia

October �1 Tokyo, Japan

November 2 Osaka, Japan

Europe / South AfricaDate LocationOctober 9 Prague, Czech Republic

November 5–� Bilbao, Spain

November 8–9 Istanbul, Turkey

November 12–1� Salzburg, Austria

November 1�–15 Daventry, United Kingdom

November 15–1� Poznan, Poland

November 15–1� Leuven, Belgium

November 21–2� Turin, Italy

November 2� Herzelia, Israel

China RUM: Mingwei Chen, KC Jen, Scott Berky, Barry Bai

China RUM

For More Information www.simulia.com/events/rums

Korea RUM

Erie RUM


2008 Abaqus Users’ Conference May 19–22, 2008 Newport Marriott - Newport, Rhode Island

Webinars

SIMULIA provides free informational webinars to help you get the most from our products and solutions. Attend upcoming live webinars or request a replay of a recorded webinar and view it at your convenience.

Upcoming Live Webinars Introducing Simulation Lifecycle Management from SIMULIA

December 4: 4:00–5:00 p.m. EST

December 6: 9:00–10:00 a.m. EST

Webinars On-DemandRequest a replay of a recorded Webinar and view it at your convenience.

••

For More Information www.simulia.com/webinars

Events

Call for Papers November 8, 2007: Final date for Abstracts and Permission Forms January 17, 2008: Draft Manuscript Due February 21, 2008: Final Manuscript and Registration for Conference Due

The 2008 AUC will mark the 30th anniversary of the commercialization of Abaqus FEA software. Released for the first time in 1978, Abaqus FEA software has undergone significant development over the years, making it the world’s technology-leading simulation software.

Our customers’ tradition of sharing information about their industrial applications is one of the reasons that we have been able to grow and advance this important technology.

Submit Your Abstract Abstracts for conference papers are now being accepted. Prospective authors are invited to submit abstracts of 250 words or less. Suggested topics can be found online at the 2008 AUC website. Paper acceptance will be based on the review of the abstract and draft manuscript by SIMULIA management and technical professionals. Please take time to submit your abstract today.

For More Information www.simulia.com/auc2008

Benefits of PresentingEnhance your professional credentials by being published in the Conference Proceedings.

Receive $100 off your registration fee to the conference and $100 off an Advanced Seminar.

Gain recognition as an expert by the simulation community

Establish beneficial industry contacts

•

•

•

•

Realistic Simulation for Offshore Exploration Applications

Realistic Simulation with SIMULIA Multiphysics Solutions Abaqus Version 6.7 for Unified FEA

Improving Powertrain Noise & Vibration Analysis

••

•

Simulation for the Real World Aircraft manufacturers are using advanced composites to create the next generation of quieter, more comfortable airliners. Our customers use SIMULIA solutions to understand the behavior and improve the application of composite materials for aerostructures. We partner with our customers to invent new methods and deploy realistic simulation technology, which helps them drive innovation and make flights more relaxing. SIMULIA is the Dassault Systèmes Brand for Realistic Simulation. We provide the Abaqus product suite for Unified Finite Element Analysis, multiphysics solutions for insight into challenging engineering problems, and an open PLM platform for managing simulation data, processes, and intellectual property. Learn more at: www.simulia.com/ae07

The �DS logo, SIMULIA, and Abaqus are trademarks or registered trademarks of Dassault Systèmes or its subsidiaries. Other company, product, and service names may be trademarks or service marks of their respective owners. Copyright Dassault Systèmes, 2007.

www.

3ds.c

om

SIMULIA Helps Get Me There.

Documents

INSIGHTS - AbaqusINSIGHTS is published by Dassault Systèmes Simulia Corp. Rising Sun MiIls 166 Valley Street Providence, RI 02909-2499 Tel. +1 401 276 4400 Fax. +1 401 276 4408