ANSYS Solutions Fall 2006

39
ANSYS CFX software improves the design of a hydro-generator for use in remote areas. The latest addition to the ANSYS family, FLUENT 6.3 offers new CFD solver, modeling and other options. Industry Spotlight Researchers used ANSYS simulation technology to develop a housing for satellite electronic circuits that is 30 percent lighter than comparable structures.

description

analysis

Transcript of ANSYS Solutions Fall 2006

Page 1: ANSYS Solutions Fall 2006

ANSYS CFX software improvesthe design of a hydro-generator for use in remote areas.

The latest addition to the ANSYSfamily, FLUENT 6.3 offers newCFD solver, modeling and otheroptions.

Industry Spotlight

Researchers used ANSYS simulation technology to developa housing for satellite electroniccircuits that is 30 percent lighterthan comparable structures.

Page 2: ANSYS Solutions Fall 2006
Page 3: ANSYS Solutions Fall 2006

For ANSYS, Inc. sales information, call 1.866.267.9724, or visit www.ansys.com.To subscribe to ANSYS Solutions, go to www.ansys.com/subscribe.

ANSYS Solutions is published for ANSYS, Inc. customers, partners and others interested in the field of design and analysis applications.

Editorial DirectorJohn [email protected]

Managing EditorFran [email protected]

DesignersMiller Creative [email protected]

Art DirectorDan [email protected]

Ad Sales ManagerBeth [email protected]

Circulation ManagerElaine [email protected]

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Neither ANSYS, Inc. nor the editorial director nor Miller Creative Group guarantees or warrants accuracy or completeness of the material contained in this publication. ANSYS,ANSYS Workbench, CFX, AUTODYN, FLUENT and any and all ANSYS, Inc. product and service names are registered trademarks or trademarks of ANSYS, Inc. or its subsidiarieslocated in the United States or other countries. ICEM CFD is a trademark licensed by ANSYS, Inc. All other trademarks or registered trademarks are the property of their respectiveowners. POSTMASTER: Send change of address to ANSYS, Inc., Southpointe, 275 Technology Drive, Canonsburg, PA 15317 USA.

©2006 ANSYS, Inc. All rights reserved.

Editorial AdvisorKelly [email protected]

Editorial ContributorChris [email protected]

ANSYS, Inc. Welcomes Fluent Inc.ANSYS broadens its opportunities to provide leading-edge engineeringsolutions with the acquisition of Fluent Inc.

FLUENT 6.3: Major Advances in CFD SimulationThe latest addition to the ANSYS family features new solver, modelingand other options — resulting ingreater speed and flexibility.

ContentsIndustry Spotlight

Features

Factory and Plant EquipmentSimulation-driven design plays amajor role in developing industrialmachines for manufacturing andprocess facilities around the world.

6

11

14

Departments

Simulation at WorkCFD Simulation Recreates Aviation History. . . . . . . . . . . . . . . 24Weight-Optimized Design of a Commercial Truck Front Suspension Component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26Developing Construction Products with Better Fire Performance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Tech FileRunning Solutions from Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33

Technology UpdateBringing High-Performance Computing to the Mainstream . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30

Industry NewsAnnouncements and Upcoming Events . . . . . . . . . . . . . . . . . . . . . . . . . 3

Tips and TechniquesWorking with Coupled-Field Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35

EditorialTools for Product and Process Innovation . . . . . . . . . . . . . . . . . . . 2

Analyzing Composites forSatellite ComponentsResearchers used ANSYS softwareto study the behavior of a compositehousing for electronic circuits andquickly developed a design nearly30 percent lighter than a comparablealuminum structure.

16

Guest CommentaryIntegrating CAD and CAE to Enable Simulation-Driven Design. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

CFD Simulation Brings Electrical Power to Rural AreasANSYS CFX software improves thedesign and efficiency of a smallhydro-generator for use in remoteareas of developing countries.

20

Companies in the factory andplant equipment market useengineering simulation exten-sively, meeting the challengesof developing cost-effective,high-precision and efficientindustrial machines that must operate under harshconditions. Read more in this issue’s Industry Spotlightarticle beginning on page 6.

About the cover

Page 4: ANSYS Solutions Fall 2006

Editorial

Tools for Product and Process Innovation

Innovate or evaporate. That’sthe new business imperative.Until recently, getting a productto market faster, cheaper andbetter than the competitionusually was good enough. Not anymore. Now — in addi-tion to time-to-market, cost andquality — manufacturers mustfocus on innovation: designsthat take the market by stormand leading-edge developmentprocesses that transform conceptual ideas into saleable,reliable and cost-effective products.

Products must stand apart from others, breakingnew ground in performance, size capacity or otherattributes that compel consumers to pick a particularitem from among many, or that influence OEMs to dobusiness with one supplier over another. In manycases, companies improve existing products withimaginative functions and enhancements. Other timesthey create whole new classes of products that totallydominate a market segment as competitors scrambleto catch up. In a world economy of radical change andfast-moving trends, innovation has emerged as the bigmarket differentiator.

Innovation is clearly on the minds of top execu-tives, as reflected in a second annual survey on thetopic conducted by The Boston Consulting Group inconjunction with BusinessWeek magazine. Theyreceived input from 1,000 senior managers worldwide,making it their “deepest management survey to dateon this critical issue.” The report, “The World’s MostInnovative Companies,” discusses the importance ofdesign as a differentiator as well as how companiesare rewiring themselves to operate differently, with 72 percent of senior executives in the survey naminginnovation as one of their top priorities.

One of the most interesting parts of the report is alist of the top 25 innovative companies. Of the top 20,14 are industrial companies — and ANSYS users. Theothers include a coffeehouse chain, retail stores, air-

By John KrouseEditorial DirectorANSYS [email protected]

Simulation technology enables companies to stand out from the crowd with knock-out designs and leading-edge product development processes.

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

lines and Internet firms. Last year’s list reflected muchthe same.

The use of ANSYS by the world’s most innovativecompanies comes as no surprise, of course. Simulation-driven design is the basis for innovativedevelopment processes and product concepts at awide range of companies in nearly all manufacturingindustries. The ability to quickly perform what-if studies and readily evaluate alternative designs gives engineers valuable insight into product behavior,lets them make intelligent trade-off decisions and provides the freedom not only to imagine way-outideas but to easily test their feasibility. Design optimization and sensitivity studies augment engineering creativity and serve as guides to creativesolutions that are not always intuitively obvious.

Using these and other wide-ranging capabilities,virtual prototyping can simulate an entire system orsubsystem in its operating environments to study andrefine real-world product performance, thus enablingengineers to develop workable innovative designs forproducts that otherwise might turn out to be flops inthe market because of performance, warranty or reliability issues. Moreover, visualization of analysisresults that vividly depict product performance facilitates close collaboration and synergy betweenmembers of multidisciplinary teams in which peoplesynergistically create imaginative design concepts that might not have surfaced otherwise.

In this manner, simulation-driven design lever-ages the creativity of engineers and the intellectualcapital of the enterprise. This elevates the approach toa strategic role as an innovation enabler, allowing manufacturers who make smart use of the technologyto establish their brand value, strengthen their marketposition and boost top-line revenue growth by developing winning products.■

2

Page 5: ANSYS Solutions Fall 2006

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

3

Industry News

Recent Announcements and Upcoming Events

ANSYS Chosen as CAE Software for ChineseMechanical Design Engineer Qualification

The Chinese Mechanical Engineering Society (CMES)and the Examination Center of Ministry of Educationhave formally incorporated ANSYS software intoChina’s national Mechanical Design Engineer (MDE)qualification examination. MDE certification was begunin China in April 2006. It is designed to improve the skilllevel of professionals in the Chinese manufacturingsector, which also widely uses ANSYS software forcomputer-aided engineering (CAE).

CMES is a professional society engaged in promotingthe art and science of mechanical engineeringthroughout its host nation. In 2005, the MechanicalDesign Institution (MDI) of CMES approached ANSYSChina about using the software in developing a standard examination to certify mechanical engineers.

The first-ever MDE qualification examination was heldin test centers located in eight Chinese provinces, with619 students from more than 20 renowned universitiesparticipating. More than 400 passed the exam andreceived the MDE qualification certificate.

ANSYS Named to Honor Roll

ANSYS, Inc. has been named to the software industrySustained Success Honor RollTM for the third consecu-tive year. Culled from a list of more than 500 publicsoftware companies compiled annually by Cape HornStrategies, ANSYS is one of 20 that made the 2006honor roll.

Companies included in the list have an outstandingrecord of growing profitability for the past five consecutive years or more. With 10 consecutive yearsof profitable growth, ANSYS is the only engineeringsimulation software company that made the SustainedSuccess Honor Roll, as well as one of only eight software companies that reported at least 10 consecu-tive years of growing profitability. According to analysisfirm Cape Horn Strategies, honor roll members significantly outperformed industry averages.

ANSYS Software Offers 64-bit Support forMicrosoft Windows Compute Cluster Server 2003

The upcoming releases of ANSYS multiphysics simulation software — ANSYS 11.0 and FLUENT 6.3— will include support for Microsoft Windows Compute Cluster Server 2003. Enabling high-perform-ance computing (HPC) on the Microsoft Windows platform, the new solution helps customers deploycomputer-aided engineering at a higher level than inthe past, decreasing the time required for simulationsand increasing the accuracy of results.

ANSYS 11.0 and FLUENT 6.3 take advantage of theMicrosoft Message Passing Interface (MPI) softwarelayer in Windows Compute Cluster Server 2003 fordata communication between processors on the cluster. The new releases also use the Microsoft JobScheduler in Windows Compute Cluster Server 2003,providing an off-the-shelf solution for launching andcontrolling jobs on the cluster.

Fluent CAD Connection Software Facilitates Linkbetween Design and Simulation

The recently released Fluent Connection 1.1 softwarehelps streamline the process of creating simulationmodels based on design data from leading computer-aided design packages. Integrating core CAE technologies with the most popular independentdesign tools has been a key part of the ANSYS strategy for nearly a decade; this latest release bringsdirect integration to the Fluent products as well. Fluent Inc. recently was acquired by ANSYS, Inc.

The Fluent UGS-NXTM Connection, Fluent Pro/ENGINEER® Wildfire® Connection and Fluent Solidworks® Connection products operate within theCAD system user environments and provide tools forchecking and conditioning the 3-D geometry model inorder to ensure that it has been properly prepared forthe next step in the simulation process. Using FluentConnection, CAD users can eliminate or repair geometry issues that would otherwise impede thesimulation process. By providing a well-defined way to check the CAD model for possible simulation-related issues, Fluent Connection helps engineering organizations ensure a streamlined hand-off betweenCAD and simulation.

Page 6: ANSYS Solutions Fall 2006

Upcoming Events

Industry News

4

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

EuroBLECH – 19th International Sheet MetalWorking Technology ExhibitionHanover, GermanyOctober 24 – 28, 2006www.euroblech.com

24th CADFEM Users’ MeetingStuttgart, GermanyOctober 25 – 27, 2006www.usersmeeting.com/index.21.0.html

ANSYS Users ConferenceSan Miguel de Allende, MéxicoOctober 26 – 27, 2006www.grupossc.com

2006 Korea ANSYS CFX User ConferencePusan, KoreaOctober 26 – 27, 2006www.anst.co.kr

Taiwan ANSYS User ConferenceTaipei, TaiwanOctober 30 – 31, 2006www.cadmen.com

ANSYS User ConferenceSingaporeNovember 2 – 3, 2006www.cadit.com.sg

Benelux ANSYS User ConferenceBreda, NetherlandsNovember 3, 2006www.infinite.nl

China ANSYS User ConferenceSanya, China November 6 – 8, 2006www.ansys.com.cn/conference/con_06

ANSYS User ConferenceRio de Janiero, BrazilNovember 7 – 8, 2006www.softec.com.br

MicroMachine ConferenceTokyo, JapanNovember 7 – 9, 2006www.cybernet.co.jp

Fluent France Forum 2006Paris, FranceNovember 9, 2006www.fluent.com/worldwide/france/support/ugm06/index.htm

Italian ANSYS User ConferenceStezzano, ItalyNovember 9 – 10, 2006http://meeting2006.enginsoft.it

ANSYS User ConferenceBangalore, IndiaNovember 9 – 10, 2006www.ansysindia.com/index.htm

ANSYS Latin American User ConferenceFlorianopolis, BrazilNovember 9 – 10. 2006www.esss.com.br/ansys2006

Fluent Germany Forum 2006Bad Nauheim, GermanyNovember 14, 2006 www.fluent.com/worldwide/germany/support/ugm/index.htm

Electronica 2006 – Components Systems ApplicationsMunich, GermanyNovember 14 – 17, 2006www.global-electronics.net/?id=20307

Japan ANSYS User ConferenceTokyo, JapanNovember 15 – 16, 2006www.cybernet.co.jp

Fluent Asia Pacific Users’ Group MeetingTokyo, Japan November 16 – 17, 2006 www.fluent.co.jp

Fluent Italy Forum 2006Milan, ItalyNovember 21, 2006 www.fluent.com/worldwide/italy/support/ugm06

ANSYS User ConferenceMelbourne, AustraliaNovember 21 – 22, 2006www.leapaust.com.au

Fluent Forum 2006Madrid, Spain November 24, 2006 www.fluent.com/worldwide/spain/events/forum06.htm

Euromold 2006 – World Fair for Moldmaking & Tooling, Design & Application Development Frankfurt, GermanyNovember 29 – December 2, 2006www.euromold.com/splash/splash_em.html

Page 7: ANSYS Solutions Fall 2006

5

Page 8: ANSYS Solutions Fall 2006

Industry Spotlight

6

To keep up with the increasing demand for

manufactured products, companies rely on factory

equipment including machine tools, injection

molding equipment, robots, material handling

equipment, stamping machines, welders and other

industrial machines.

In the competitive drive for factories and plants to produce more with less, the increased speed and efficiency of today’s technology-based equipment iscredited as a major element in industrial productivitygains. According to the National Association of Manufacturers, manufacturing productivity grew 4.8 percent last year. That’s a 78 percent jump compared to the economy as a whole and amounts toa 24 percent increase during the past four years. So to remain competitive, companies around the world invest heavily in the latest state-of-the-art production equipment.

Statistics from the Manufacturing PerformanceInstitute indicate that 20 percent of sales are re-invested in factory capital equipment in China, forexample, and that 45 percent of U.S. plants expect to

Factory andPlant Equipment

From small job shops to giant superfactories and processing plants, facilities throughout the supply chain use production equipment to turn raw materials into products in the automotive,aerospace, telecommunications, electronics, heavy equipment, consumer products, petrochemical, pharmaceutical and food processing sectors, and even in service industries such as data processing, finance and insurance.

Images courtesy Hatch Australia.

By Achuth RaoProduct ManagerANSYS, Inc.

Simulation-driven design plays a major role indeveloping industrial machines for manufacturingand process facilities around the world.

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 9: ANSYS Solutions Fall 2006

7

increase their spending on production equipment in2006. According to the World Machine Tool Outputand Consumption Survey, Japan ranks number one interms of world output of machine tools and second inusage of this equipment. Meanwhile, figures fromfinancial services firm JPMorgan specify that the rateof manufacturing expansion in the European sector isreaching a six-year high.

On the flip side, industrial equipment buyers havea long list of demanding requirements. Machines mustoperate for decades under harsh conditions and oftenfor multiple shifts, seven days a week. Downtime isunacceptable, since daily revenue losses can run intomillions of dollars when production is halted. Energyefficiency is mandatory in lowering operating costs inthe face of rising electric utility prices. Noise emissionsmust be low to meet strict regulatory standards. Vibrations must be minimized to avoid fatigue failuresand unwanted resonances in precision machines. In addition, equipment must be cost-sensitive; newmodels must be launched quickly to meet fierce global competition.

Companies in the factory and plant equipmentmarket regard engineering simulation as an indispen-sable tool in meeting these challenges. ANSYS technology in particular is used by many of these firmsin their product development cycle. A range of leading-edge solutions provides a breadth and depthof analysis capabilities including meshing of complexparts and assemblies, computational fluid dynamics(CFD), optimization, structural and thermal tools, and awide range of multiphysics solutions. The ANSYSWorkbench environment brings these technologiestogether in a unified suite of software.

Boosting Machine Speed and Capacity

Simulation technology plays a key role in effortsaround the world to make industrial equipment moreproductive in terms of machine capacity as well asoperational speed. Spain-based technologicalresearch center Fundacion ITMA, for example, performed a coupled thermo–structural analysis withANSYS Mechanical software in developing a newdesign for a steel-making ladle — resulting in a 15 percent greater capacity for handling liquid metal.

Likewise, metalworking equipment manufacturerGebr. Heller Maschinenfabrik GmbH in Germany usesANSYS Mechanical in static, dynamic and thermalsimulations of the metalworking equipment it develops, which includes transfer lines, machiningcenters, flexible manufacturing systems, and milling

An ANSYS simulation model is superimposed on a representationof a Heller milling machine for manufacturing heavy truck crankshafts and camshafts. The handling system automaticallytransports the crankshafts into and out of the milling machine.

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 10: ANSYS Solutions Fall 2006

and broaching machines. Typical simulations includedeformation of parts, modal analysis and frequencyresponse of the machines together with tools andworkpieces, structural temperature distribution andtopology optimization. In one recent application, Hellercredits ANSYS simulation in achieving a 20 percentincrease in operational productivity of a machiningcenter for the production of automotive parts such asengine blocks, cylinder heads, transmission housingsand chassis components.

Building Better Robots

According to the Robotic Industries Association, nearly1 million robots populate global manufacturing, withalmost half working in Japan. Robots perform a widerange of production tasks including material handling,workpiece and tool positioning, arc and spot welding,packaging, and process applications such as inspec-tion, testing, spraying and dispensing. Simulation iscritical in developing these versatile machines for optimal speed, precision, lifting capacity and cost.

German-based Robo-Technology recently usedANSYS Workbench tools to develop a six-axis roboticsystem for ultrasonic testing of helicopter parts up tosix meters in length. The company reports that theability to analyze the design throughout the develop-ment process enabled them to verify that the rigidityand vibration behavior of the system met customerdemands of rapid testing movements, high dynamicprecision and accurate synchronization among multiple robots working together.

Similarly, Motoman Inc. in the United States usedANSYS DesignSpace software in developing a roboticoverhead transport with a two-meter boom capable ofcarrying a 50-kg payload. Simulation technology issaid to play a major role in the company’s productdevelopment group, which used ANSYS DesignSpaceto create a boom with less mass so engineers couldincrease the reach and payload of the equipment.

Insight into Complex Equipment Behavior

Simulation is a powerful tool for better understandingthe behavior of complex industrial machines. With thisinsight, engineers can then more effectively optimizedesigns and develop innovative concepts. The Non-Ferrous Metals Technology Group of Hatch Australiauses advanced analysis tools such as CFD for designevaluation, optimization and problem-solving in whichheat transfer, fluid flow, combustion and mass transferare critical issues. Hatch is a leading engineering consulting firm specializing in scale-up of processtechnology from prototype pilot systems to large production systems.

In one project, researchers used ANSYS CFXsoftware in analyzing a multiphase grinding mill that vigorously stirs incoming material together withsolid grinding media using a series of high-speedrotating disks. The CFD analysis showed the complex multiphase swirling flow through the mill’s intricategeometry and the nature of media distribution, secondary flows and wear characteristics of parts. Inthis way, simulation has improved the understanding

Industry Spotlight

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

ANSYS Workbench was used at Robo-Technology in developing a system in which two precision robots work together in ultrasonictesting of helicopter parts up to six meters in length.

8

Page 11: ANSYS Solutions Fall 2006

of mill behavior for scale-up of the design and also hasgenerally enhanced mill operation.

Adding Revenue by Shortening the Development Cycle

Computer simulation reduced time needed to developa new aggregate drying burner designed for use inasphalt plants from the normal six to 12 months toonly 32 days. Manufactured by Astec Industries in theUnited States, the burner is intended to remove moisture from rock so it will bind properly to cement informing asphalt. Getting dryers to market as quickly as possible necessitated development of the burner in an extraordinarily short time, yet there was barely time to build a single prototype.

The Astec design team used FLUENT CFD tech-nology, recently added to the ANSYS suite of softwaresolutions, to readily evaluate numerous virtual proto-types and quickly iterate to an optimized design. Theprimary concern was determining the best way ofinjecting fuel to obtain an optimal gas mixture. CFDsaved considerable time by determining the flow andchemical concentrations early in design, providing farmore information than ever would have been possiblewith physical experiments. In only two weeks, a working prototype was built; within a month, thedesign was optimized to meet stringent emission regulations. In this way, simulation drastically reduced time-to-market, thus providing up to a year of additional revenues while substantially reducingengineering costs.

More Time for Better Quality and Greater Innovation

ANSYS structural analysis software is a core tool for the state-of-the-art development facility at theheadquarters of Husky Injection Molding Systems Ltd.in Canada. The company designs and manufacturesthe plastics industry’s most comprehensive range ofinjection molding equipment, including machines,molds, hot runners and robots.

In developing these large injection moldingmachines, engineers face demanding design challenges. Machine weight must be minimized tokeep manufacturing and transportation costs low.Operating speeds must be fast enough for requiredthroughput of manufactured plastic parts. Reliabilityand precision must be maintained to provide satisfactory service with minimal downtime. Efficiency

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

In developing an asphalt plant burner, CFD simulation showsvelocity contours and pathlines indicating flow distribution aroundthe fanwheel (top). Diffusive mixing of methane is indicated bypathlines around the gas injection pipe assembly (bottom).Images courtesy Astec Industries.

9

Page 12: ANSYS Solutions Fall 2006

Industry Spotlight

is a requirement to keep energy consumption low andthus minimize operating costs.

Engineers at Husky met these challenges withANSYS Workbench and ANSYS DesignSpace tools.The integrated solutions provided efficient contactrepresentation for complex nonlinear assembly analysis, additional pre-analysis in construction and the benefit of common simulation methods in the various types of analyses. Engineers report that analyses formerly taking a week now can be completed in just half a day. This level of increasedanalysis efficiency is said to enable developmentteams to achieve better machine quality and greaterinnovation in products such as the company’s newReflex platens.

New Rotodynamics Capability IncreasesAnalysis Productivity

Design analysis of parts and assemblies in the industrial machinery industry involves complex computer-aided design (CAD) assemblies. To accurately handle these geometries in the designprocess, ANSYS offers close connection with CAD to access geometry and material parameters; it alsoallows quick turn-around while preparing geometry foranalysis. Geometry creation and editing tools allowgeometry manipulation for physics-based meshingand analysis.

U.S.-based Trane, a business of American Standard, Inc. and a leading worldwide supplier ofHVAC (heating, ventilating and air conditioning), usesANSYS software to design rotating equipment inindustrial chillers and air conditioning equipment usingthe new rotordynamics capability. The ability to importfull 3-D CAD models into ANSYS Workbench allowsthe user to analyze accurate 3-D models instead ofcreating simplified 1-D representation of the geometry.Productivity tools such as automatic contact detectionallow for easy problem setup and more time spent onengineering design decisions.

To survive in the global economy of the third millennium, manufacturers need to be inventive interms of factory equipment and raw materials, as wellas with the processes they develop. Using simulation-driven design efforts can bring value and innovation toa wide range of product development applications. ■

The author wishes to thank development, technical support

and marketing personnel at ANSYS, Inc. for their efforts and

contributions to this article.

Trane uses ANSYS geometry (top), meshing (middle) and dynamics(bottom) solutions to design HVAC systems and comprehensive facilitysolutions for factories and other large commercial and industrials.Images courtesy Trane, a business of American Standard, Inc.

10

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 13: ANSYS Solutions Fall 2006

On May 1, 2006, ANSYS, Inc. announced the comple-tion of the acquisition of Fluent Inc., headquartered inLebanon, New Hampshire. Fluent is a global providerof computer-aided engineering (CAE) products thatutilize computational fluid dynamics (CFD) principlesand techniques to enable engineers and designers tosimulate fluid flow, heat and mass transfer, and relatedphenomena involving turbulent, reacting and multi-phase flow. This acquisition reaffirmed the ANSYScommitment to providing the open interface and flexible simulation solutions that customers require.

What follows is some background information to help you get better acquainted with the newestmember of the ANSYS family.

History of Fluent Inc.

In 1982, when CFD was primarily of interest to academic specialists, engineers at Creare, Inc., a NewHampshire consulting company, collaborated withresearchers at Sheffield University in Sheffield, UK, todevelop an interactive, easy-to-use CFD softwareproduct for engineers. Called FLUENT, the first versionof this software was launched in October 1983. Theproduct was so successful that, in 1990, the FLUENTgroup at Creare split from its parent company, movedto a new location and formed Fluent Inc.

Rapid expansion of Fluent’s software businessensued, and, in May 1996, Fluent acquired FluidDynamics International, the developer of the general-purpose CFD software FIDAP. In 1997, Fluentacquired Polyflow S.A., the developer of POLYFLOW,a specialty CFD software product for the analysis ofmaterials such as polymers, plastics, food and rubber.

Since its inception, Fluent has continued to innovate and grow by offering superior CFD softwareand services to companies around the world.

Industry-Leading Technology

The broad physical modeling capabilities of FLUENTtechnology have been applied to industrial applica-tions ranging from air flow over an aircraft wing tocombustion in a furnace, from bubble columns toglass production, from blood flow to semiconductormanufacturing, from clean room design to wastewatertreatment plants. The ability of the software to modelreacting flows, aeroacoustics, turbulence, movingmeshes and multiphase systems has served to broadenits reach. Today, thousands of companies throughoutthe world benefit from using FLUENT software.

The suite of Fluent CFD products includes FLUENT, FIDAP and POLYFLOW for CFD analysis;FloWizard, a rapid flow modeling tool that allows

ANSYS broadens its opportunities to provide leading-edgeengineering solutions with the acquisition of Fluent Inc.

By Chris ReidVice President, MarketingANSYS, Inc.

ANSYS, Inc. WelcomesFluent Inc.

Image courtesy Sheffield University.

11

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 14: ANSYS Solutions Fall 2006

design and process engineers to quickly and accurately validate their designs much earlier in theproduct development cycle; FLUENT for CATIA V5,which integrates Fluent’s rapid flow modeling technology into the CATIA V5 product lifecycle management (PLM) process; and FlowLab, a student-friendly tool that uses the power of flow visualizationthrough CFD to teach basic fluid mechanics principlesin the engineering classroom. Fluent’s products alsoinclude the preprocessors GAMBIT, TGrid andG/Turbo. Application-focused products include Icepakto optimize thermal management of electronicsdesigns; Airpak for modeling airflow, heat transfer,contaminant transport and thermal comfort for thebuilt environment; and MixSim for the simulation ofstirred tanks.

Fluent always has taken pride in understandingcustomers’ strategic goals and helping them come to fruition through both software and services. The complete array of services available addressesthe specific needs of organizations and supportsthose organizations in implementing advanced technology solutions. Services include consulting,training and technical support.

Extensive Simulation Community

With the acquisition of Fluent Inc., ANSYS is pleasedto welcome Fluent software users to the world’s

largest simulation community. As a supplier to 94 ofthe FORTUNE 100, ANSYS serves a wide range ofindustries. They all have benefited from using ANSYSsoftware products and services — which continue toexpand, both with the addition of new technologiesdeveloped via innovative research and developmentand by acquisition, such as in the case of Fluent.

ANSYS now has one of the broadest ranges ofCFD simulation technologies in the world. ANSYSbelieves that success relies on ensuring customers’satisfaction. As such, in addition to focusing on investment in product development, the company willcontinue to provide the best possible service and support through technical centers of excellencearound the world.

The addition of Fluent products to the ANSYSportfolio significantly enhances the combined company’sability to provide world-leading simulation capabilitiesto customers, consistent with the ANSYS vision and strategy.

Broad and Integrated Solutions

ANSYS continues to concentrate on providing customers with best-in-class CAE tools integrated in aflexible manner that will enable easy and rapid analysisand optimization of engineering designs.

Clearly, there is now the opportunity for tighterlinkages between ANSYS products, such as ANSYS

Engineers at Wenger Manufacturing gained better insight intothe operation of a vertical cascade dryer (used to manufacturepet food) by utilizing CFD. The consulting team at Fluent wasable to provide Wenger with an understanding of the existingdryer in order to improve their ability to market, specify, applyand service it. Improved dryers allow food processors to meetguaranteed levels of product components, decrease costs andavoid recycling of fine particles that may detract from productappearance and present a fire hazard. Detailed airflow andpressure distribution in various sections of the dryer would beimpossible to obtain through physical testing and measure-ment — but were possible to simulate using FLUENT software.

Image courtesy Wenger Mfg.

In this example, FLUENT software is used to optimize the cooling package for a line of tractors. The cooling fancharacteristics, along with the placement of underhoodmodules, are varied to achieve optimum performance.

Courtesy Case New Holland.

12

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 15: ANSYS Solutions Fall 2006

Mechanical or ANSYS Multiphysics, and Fluentproducts, as the company has done already withANSYS CFX. The benefits of this strategy to customers and the engineering simulation industryhave been real and measurable in terms of increasedinnovation, greater productivity and lower costs.ANSYS, Inc. fully expects to extend the same benefitto today’s Fluent user community.

With the addition of more than 700 new employees from Fluent and its subsidiaries, the combined team with many years of simulation experience, deep industry expertise and world-classengineering talent will deliver even more excitingadvances in integrated CAE. To support the ability to provide industry-leading advancements, ANSYS willcontinue its focus on innovation and target approximately20 percent of revenue to be spent on R&D.

Moving Forward

ANSYS and Fluent always have had much in common. Now, the goals each company had for the future are shared, and progress toward these goals can be accelerated and fulfilled to thebenefit of customers. As in the past, ANSYS will

maintain a strong commitment to employees, partnersand customers as well as to the advancement of technology through innovation. ■

CFD is becoming critical to the furnace design process, particularly as environmental regulations become more stringent. To help meetthese regulations, John Zink Company selected FLUENT software to assist in modeling flames for a vertical cylindrical furnace used inan oil refinery. By using FLUENT, engineers were able to improve fuel mixing in order to decrease NOx production and maintain flameheight within the desired parameters. CFD modeling provided a proposed solution (right) to reduce burner interactions that had causedincreased flame height (left).

Image courtesy John Zink Co.

Some of the world’s greatest soccer goalkeepers have beenbeaten by unusual swerving balls that move left then rightbefore hitting the back of the net, even though they have littleor no spin applied to them. A team of researchers led by Dr. Matt Carré at the Department of Mechanical Engineering,University of Sheffield used FLUENT to demonstrate that theshape, surface and asymmetry of the ball, as well as its initialorientation, have a profound effect on how the ball movesthrough the air after it is kicked. The image shows high-speedairflow pathlines colored by local velocity over the Adidas®

Teamgeist 2006 soccer ball.

Image courtesy Sheffield University.

13

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 16: ANSYS Solutions Fall 2006

14

FLUENT 6.3: Major Advancesin CFD SimulationThe latest addition to the ANSYS family features new solver, modeling and other options — resulting in greater speed and flexibility.

By Christine Wolfe, FLUENT 6 Product Manager, Fluent Inc.Nicole Diana, Product Planning Manager, Fluent Inc.

Polyhedral meshes are being introduced in FLUENT 6.3. These meshes allow the flexibility of anunstructured mesh to be applied to a complex geometry without the overhead associated with alarge tetrahedral mesh. The polyhedral meshes arecreated using automatic cell agglomeration to combine tetrahedral cells into polyhedral ones. Thiscan reduce the overall cell count by a factor of 3 to 5.The automatic nature of these mesh agglomerationtechniques saves the user time and, since the poly-hedral mesh contains as few as one-fifth the number of cells in the original tetrahedral mesh, convergenceis faster.

In support of FLUENT software’s ongoing commitment to parallel processing, numerousimprovements to parallel efficiency and flexibility have

The new sliding mesh capability in FLUENT 6.3allows for many sliding interfaces within a single simulation. In the case of the V22Osprey, the rotating propellers gradually tilt as the craft lands on a platform, changing the propulsion from forward to hover mode.

The current version of FLUENT software continues toevolve, allowing difficult engineering problems to besolved faster and with greater flexibility than ever before. The upcoming release of FLUENT 6.3offers innovative technology for addressing a broadrange of applications. In all, there are more than 100new features that enhance core numerics and physicalmodeling capabilities in areas such as moving mesh, multiphase flow, combustion, reacting flow and radiation.

New Solver OptionsIn FLUENT 6.3, a pressure-based coupled solver joinsthe existing solver options. The new solver canimprove solution efficiency as well as convergence androbustness for many cases. With this solver scheme,the pressure and velocity equations are solved in a fullycoupled manner, while the other equations are solvedsequentially. It is particularly beneficial for “stiff” prob-lems and for solving problems on unusually skewedand stretched meshes.

In addition, existing FLUENT solvers have beenenhanced to offer improved robustness, accuracy and efficiency. For example, strong shocks can be captured more effectively with the density-basedsolver, and transient simulations can be run more efficiently with the pressure-based solver. Furthermore,a new diagnostic case check algorithm can be used toassess case settings and offer recommendations toensure that commonly accepted best practices arebeing used.

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

In FLUENT 6.3, a discrete phaseof particles, droplets or bubblescan be launched from a surfaceat normal angles, as is shownfor a Rushton impeller in astirred tank.

Page 17: ANSYS Solutions Fall 2006

15

been implemented along with speed improvements forreading and writing case and data files. High-perform-ance computing also benefits from a 64-bit Windowsversion of FLUENT 6.3.

New Modeling OptionsNew models and extensions to existing models add tothe technology’s capabilities in the areas of movingmesh, reacting flow, multiphase flow and radiation.

FLUENT software’s industry-leading dynamicmesh capability for modeling moving objects — suchas pistons and valves in IC engines, store, separationand impellers in baffled mixing tanks — has beenenhanced. In FLUENT 6.3, the dynamic mesh capability can be applied to a series of related steady-state simulations, making them easier for users to setup and perform. For example, a control valve can be simulated with a range of open positions by buildingonly one mesh and having FLUENT software rebuildthe mesh for each new position. Other improvementsmake problem setup and user-defined mesh motioneven more straightforward and efficient.

In some cases, the motion of objects can be captured by using regions of mesh that slide along a common interface. This technique is useful for modeling two trains passing in a tunnel, for example.FLUENT 6.3 now is able to model more complexobject motion by allowing for multiple sliding meshregions on one side of an interface to be paired withmultiple sliding regions on the opposite side.

Reacting flow simulations benefit from new slowchemistry and micromixing models, useful for liquidreactions and certain combustion applications. A larger number of chemical species and reactions canbe handled in the non-premixed and partially premixedcombustion models. Emissions modeling is morecomprehensive through the addition of SOx predictionand the selective noncatalytic reduction of NOxthrough urea injection. Expanded in-cylinder combus-tion capabilities include the ability to model ignitiondelay in stratified engines.

Multiphase modeling continues to be an area of focus for FLUENT 6 development, and majorimprovements can be found in the accuracy of transient multiphase solutions. For the Eulerian multi-phase model, enhancements extend the regimes forwhich this model can be applied. For example, bothcompressible gas and liquid phases can be present,and the mixing plane model can be used, simplifyingthe solution of multiphase flows in pumps. For freesurface flows simulated using the volume of fluid (VOF)model, a new interface tracking scheme is available

The ability to solve on polyhedral meshes isnew in FLUENT 6.3. Tetrahedral cells areagglomerated to form polyhedra in the solver,resulting in a reduced overall cell count thatrequires less CPU time to reach convergence.

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

that improves solution stability when the viscosity ratiobetween the phases is high. FLUENT 6.3 technologyalso allows a user-defined function (UDF) to be utilizedto specify the wall contact angle, allowing a dynamicvalue to be calculated from the local flow field. Thisfeature is of primary importance for capillary-drivenflows in which surface tension is important.

For simulations involving surface-to-surface radiation, extensions to the existing model make theproblem definition easier, and increase the solution efficiency and range of applicability. For example, thistechnique now can be used for 2-D axisymmetriccases, and the participating boundaries can be specified more easily in the graphical user interface.

Add-On Modules and Third-Party ToolsSeveral capabilities can be added to FLUENT softwarethrough add-on modules. A population balance module is new with FLUENT 6.3. This module makes it possible to model multiphase flows with a particle or droplet size distribution. Three approaches are available that account for breakup and agglomerationso that applications such as bubble columns and crystallizers can be modeled. The proton-exchangemembrane (PEM) and solid-oxide fuel cell (SOFC)modules have been enhanced in FLUENT 6.3. For PEM fuel cells, transient simulations can be performed and electrical conductivity can be obtainedfrom the FLUENT materials database. For the SOFCmodule, the range of conditions that can be simulatedhas increased.

Another improvement in FLUENT 6.3 is the abilityto work with third-party CAE packages. It is now easier to import and export files to and from otheranalysis tools (for fluid structure interaction, for example) and postprocessing tools (such as EnSightor Fieldview, for example).

Along with many other features, these highlightsmake FLUENT 6.3 software a major step forward incommercial CFD capability. ■

Nylon is injected into a mold in this simulation of the production of a plasticgear part. The plastic has non-Newtonian rheology, and the ratio of the nylon toair viscosities is very large. To address this complex free surface flow, a newinterface tracking scheme in FLUENT 6.3 is used.

Page 18: ANSYS Solutions Fall 2006

Scientific, observation and reconnaissance missionsoften are performed by low-orbiting micro-satellites.These systems are much smaller and more compactthan larger telecommunications satellites, so space is severely limited and heat is more difficult to dissipate from closely packed electronic components.Traditionally, satellite electronics housings are made ofaluminum. This material is lightweight, has adequate

heat dissipation and provides good protection againstambient spatial radiation.

In one recent study, the European Space Agency(ESA) investigated the feasibility of fabricating thesehousings of composites to determine if this type of material systems could provide the same heat dissipation as aluminum but with less mass. In thisstudy, Verhaert Design and Development in Belgium

Researchers used ANSYS technology to study the behavior of a composite housing for electronic circuits and quickly developed a design nearly 30 percent lighter than a comparable aluminum structure.

By Harri KatajistoR&D EngineerComponeering Inc.Helsinki, Finland

Analyzing Composites for Satellite Components

Figure 1. The Proba 2 micro-satellite has instruments tomake solar observations and space weather measure-ments. The electronics housing from this micro-satellitewas used as a reference application in the ESA study.

Image courtesy Verhaert Design and Development.

16

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 19: ANSYS Solutions Fall 2006

provided the reference application for the study: theAdvanced Data and Power Management System(ADPMS) aluminum housing for the Proba 2 micro-satellite in Figure 1. Analysis of the composite housingwas performed by Componeering Inc., which specializesin simulation and design of high-performance com-posite structures. The Laboratory of LightweightStructures at Helsinki University of Technology was responsible for the design, prototyping and manufacturing of the housing, shown in Figure 2.

The Challenge of Designing with Composites

Designing composite structures with sandwich-typeelements or layered solid laminates is very challengingdue to the anisotropic behavior of the material. Moreover, the design depends on multiple variablessuch as material selection, number of layers, layer orientations and stacking sequence.

In the structure under investigation, performancerequirements for the composite housing were derivedfrom the requirements of the aluminum housing. For example, the goal in radiation protection was to provide shielding against spatial radiation in lowearth orbit comparable to the aluminum design with a2mm wall thickness. This was achieved by embeddingtungsten foil inside the carbon fiber-reinforced plastic (CFRP) laminate structure of the housing external panels.

One important requirement was that mechanicalinterfaces of the composite housing had to be identical to the aluminum counterpart. Inside and outside contours of the housing had to accommodate

internal circuit boards and connectors as well as external components of the satellite. These constraintslimited the composite design, in which smooth shapesare preferable.

To provide thermal management comparable to the aluminum housing, heat dissipation for the composite structure was provided by layers of plastic reinforced with K1100 pitch-based carbonfibers. Combined with the plastic matrix as a ply configuration, these fibers yield about four times higher thermal conductivity in the direction of fibersthan typical aluminum alloys. Moreover, the naturalblack color of CFRP provides high emissivity, and heatis dissipated very efficiently from the surface. From astructural standpoint, however, the K1100 fibersexhibit very low failure strain and break easily whenbent on a small radius.

Compounding the design challenge, mismatchesin coefficients of thermal expansion (CTE) between thecomposite housing and aluminum wedge locks andsatellite support structures cause deformations whenthe structure is subjected to temperature change.

Because of these complexities, determining thermal balance, structural integrity and resonant frequencies of the housing using conventional analysismethods can be an extremely cumbersome task.Results would most likely have a high probability oferror due to simplifications that would not adequatelyaccount for all design variables.

Advanced Tools for Simulation and Design

To meet these challenges, we performed a wide rangeof analyses using ANSYS Mechanical softwarethroughout the project. During conceptual design,laminate through-the-thickness direction behavior wasstudied with a solid thermal model using SOLID70elements as shown in Figure 3. (See next page.)Analysis results demonstrated that single-layer elements are adequately representative for thin laminates in steady state, so only in-plane thermalconduction capability of the SHELL131 element wasused. The effective conductivities of the laminate wereverified easily with capabilities of SHELL131 and simple test models. Mechanical interfaces and adhesively bonded joints were modeled using thermal LINK33 elements, with contact resistancedepending on joint materials, surface roughness andcontact pressure. Input data definition for thermal links

Figure 2. The housing with its back panel removed showscomposite laminate structures together with aluminumwedge locks and mounting rails.

Photo courtesy LLS/HUT.

17

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 20: ANSYS Solutions Fall 2006

was based on a literature study. Both conduction and radiation heat transfer modes were considered.The radiation effect was applied using the SURF151element that was overlaid onto the SHELL131 element.

Thermal analysis of this type is readily performedin ANSYS software, which provides the capability touse nodal temperatures resolved from the thermalmodel as an input for the structural model. The use ofthis feature is very simple, since ANSYS internallyconverts the thermal results file *.rth to equivalentforce and moment vectors in the structural model. In this study, in-service temperature variations in laminate structures were quite small, so thermal bending was not an important factor. However, thermalbending was found critical in some manufacturing testtrials in which laminate structures were bonded at high temperature with aluminum wedge locks. For thesecases, SHELL181 has proved to give excellent results,providing ease of controlling the laminate cross-section input data.

The wedge locks do not fully cover the hat section in the depth direction, and ANSYS thermalanalysis revealed that part of the heat-dissipationcapacity was lost with K1100 layers oriented in thelongitudinal direction of the housing, as shown in Figure 4. It was found that when K1100 layers wereoriented in ±30 degrees, the temperature distributionwas more homogenized, the peak temperature wasthe same and, structurally, the lay-up was acceptable.

The stiffness requirements of the housing dictated the number of required structural M40J CFRPlayers. The orientation and stacking sequence of bothK1100 and M40J layers was based on the laminatestiffness and CTE. CFRP structures have slightly negative CTE in the fiber direction. Respectively, in the perpendicular direction to the fibers, CTE is in the magnitude of aluminum. The wedge locks wereadhesively bonded to the laminate structures and acted like stiffeners. Due to this bonding and orientation, the CTE of the laminate was an important design parameter.

The laminate design was performed usingESAComp software (www.esacomp.com), which is adedicated tool for preliminary and detailed analysis ofcomposite structures. ESAComp interfaces smoothlywith ANSYS software. Laminate lay-ups and materialdata can be exported to ANSYS for composite solidand shell elements. Moreover, FEA results fromANSYS can be imported to ESAComp for detailedpost-processing. This capability was used, for example,in studying the criticality of the laminate interlaminarshear (ILS) stresses, which became high close to theinserts that were used to attach different panels.

Creating simulation models was facilitated usingANSYS Parametric Design Language (APDL), whichcould be linked to ESAComp for optimizing thedesign. With APDL, an external software such asESAComp can be readily linked to the design cycle,thus allowing simulation to effectively guide the development process toward the best design.

After completion of the thermal analysis withANSYS software and laminate design with ESAComp,ANSYS Mechanical was used for structural analysis ofthe housing, including a modal analysis for natural

Figure 3. Representative 3-D model illustrates temperaturevariation in the laminate through-the-thickness direction (left)and high thermal flux in the two K1100 layers located onboth surfaces (right).

Figure 4. ANSYS thermal analysis of the composite hat sectionrevealed that heat-dissipation capacity varied according tothe orientation of the K1100 layers of the laminate.

18

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 21: ANSYS Solutions Fall 2006

frequencies up to 800 Hz. Modal analysis resultstogether with the random vibration input specificationdefined the acceleration levels for the subsequent failure analysis. The design was predicted to withstandload levels multiplied by factor of safety of two, which is a typical value for composite materials inspace applications.

In the ANSYS model for these structural studies,the quadratic composite SHELL99 element was used for meshing the laminate structures. SHELL99 isapplicable to thin laminates, but it contains transverseshear deformations capability as well. Bolted jointsand inserts were modeled using the BEAM4 elementin order to be able to extract laminate bearing stressesand insert pull-out forces.

Impact and Benefits of the Solution

Thermal tests performed in the vacuum chamber atthe European Space Research and Technology Centre(ESTEC) in the Netherlands corresponding to theworst hot temperature condition of the equipmentconfirmed that simulation accurately predicted the satisfactory heat-dissipation capacity of the composite housing.

Using an electromagnetic shaker, sine and random vibration tests of a breadboard model wereconducted at the Royal Military Academy in Brussels,Belgium. The composite housing was found stifferthan its aluminum counterpart, and overall behavior ofthe system was as predicted with simulations.

In this project, ANSYS software worked smoothlyin exchanging data with ESAComp. Time was saved ingenerating simulation models by importing requireddata directly from ESAComp into ANSYS, as well astaking advantage of automated features of ANSYScontact elements and power of APDL for parameteri-zation. Considerable time also was saved through theability to use the same ANSYS model for both structural and thermal analyses, as shown in Figure 5.In the course of the project, details of the compositestructure were studied with simulation before goingthrough the time and expense of building physical prototype breadboard models. In this way, the modified structure could be analyzed and feedbackprovided almost instantly to the design team.

The ability of ANSYS software to work well withESAComp, to provide a robust parametric model representing all the different components and to reliably perform both structural and thermal analyseswas key to the speed and accuracy in successfullycompleting the study. With the help of this level ofadvanced analysis, the behavior of the structure couldbe properly understood, the design of the compositehousing was optimized to provide a mass saving of 29 percent over a comparable aluminum housing andthe project was completed in only 18 months from thekick-off meeting to the final presentation of results. ■

Refer to the following papers for more information on

this project:

Katajisto, H. et al., “Structural and Thermal Analysis of Carbon Composite Electronics Housing for a Satellite,” Conference Proceedings of the 1st NAFEMS Nordic Seminar “Component and System Analysis Using NumericalSimulation Techniques — FEA, CFD, MBS,” Gothenburg,November 23-24, 2005.

Brander, T. et al., “CFRP Electronics Housing for a Satellite,”Proceedings of European Conference on Spacecraft Structures, Materials and Mechanical Testing, Noordwijk,May 10-12, 2005.

Jussila, J. et al., “Manufacture and Assembly of CFRP Electronics Housing,” Proceedings of European Conferenceon Spacecraft Structures, Materials and Mechanical Testing,Noordwijk, May 10-12, 2005.

Figure 5. The same ANSYS model was used for both structuraland thermal analysis in determining characteristics such asthe thermal balance of the laminate structures and modeshapes of the system.

19

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 22: ANSYS Solutions Fall 2006

ANSYS CFX software improves the design and efficiency of a smallhydro-generator for use in remote areas of developing countries.

By Robert Simpson, Ph.D.Nottingham Trent University, UK

CFD Simulation Brings ElectricalPower to Rural Areas

In developing countries, many remote rural communitiesdo not have access to electricity due to the largeexpenses associated with extension of the nationalgrid. Where these communities have access to a suitable site, small hydroelectric schemes are found tobe a cost-effective and sustainable means of providingelectricity. The Micro Hydro Research Centre at Nottingham Trent University has been carrying out research into small-scale standardized hydro-generator units that are directly affordable for villagersin developing countries.

Low-head hydro sites (2 to 10m) offer the potential for providing electricity to many communities,but progress has been hampered by the lack of anappropriate turbine design. A research project has

been undertaken in collaboration with Practical ActionPeru (formerly known as Intermediate TechnologyDevelopment Group) to develop a standard designprocedure for the cost-effective local manufacture of pico-propeller turbines with good performance and reliability. Pico-hydro is the smallest classificationof power output with a maximum output of 5kW. Fixed geometry propeller turbines are one of the most cost-effective turbine options for low-head pico-hydropower.

The first objective of the project was to design a simplified prototype turbine based on current knowledge and to have it manufactured and installedat a field site in Peru. Stage two involved analyzing the turbine using computational fluid dynamics (CFD)

Visualization of the turbine velocity vectorsat the blade mid-span location

20

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 23: ANSYS Solutions Fall 2006

to improve the design and efficiency. Using CFD provides the flexibility to make design changes, investigate scale effects and vary the hydrologicalconditions (head and flow rate) without the need andexpense of creating a prototype for each test case.

ANSYS CFX software was used for all aspects of the CFD study because of its good track record with turbomachinery analysis and because it is part of an integrated system of software, whichincludes CFX-BladeGen (now ANSYS BladeModeler) and ANSYS TurboGrid for design and analysis of bladed geometry.

Prototype Turbine in Peru

The prototype turbine for the project is located on asmall farm in the northern highlands of Peru. Currently,the turbine is producing electricity for the owner’sfarmhouse located several hundred meters away, andthe horizontal shaft layout was designed to facilitateeasy connection to existing mechanical equipmentused by the farmer to produce feed for several chickenfarms. Water for the turbine is diverted into a concretechannel from an existing irrigation channel that runsacross the plot of land.

The spiral casing has a simplified design with a tapered rectangular cross-section. The rotor blades were manufactured using flat sheet metal bent and twisted into the required shape. Preliminaryfield testing of the turbine revealed several problemsduring operation. Water was being emptied from the forebay tank, resulting in a much lower head thanthe available four meters, and the turbine was not producing sufficient power to get the generator up to operating voltage.

Simulation Helps Refine the Design

Simulations performed on the original rotor geometryshowed that the maximum turbine efficiency was predicted to be approximately 55 percent at 600 rpmwith a flow rate of 284 liters per second (l/s). However,available flow rate at the site was measured to be inthe range of 180–220 l/s, which was not enough flowfor the turbine to operate efficiently. From this preliminary study, it was concluded that the bladeangles for the original rotor were incorrect. A new rotorwith flatter blades and a higher solidity ratio wasdesigned using conventional theoretical methods. Thenew rotor geometry was analyzed using ANSYS CFXsoftware, and the results predicted a significant reduc-tion in the flow rate required to obtain the same poweroutput. In addition, the power curve demonstrated animproved performance over the speed range with apredicted best efficiency for the new rotor of 80percent at 800 rpm.

This project was undertaken to provide power to a farmhousein rural Peru that is used to grind feed for several chickenfarms owned by the farmer.

The civil works required for power generation consisted of aconcrete channel, silt basin, forebay tank and the powerhouse.

The turbine drives an induction motor as generator (IMAG)with controller. The radiator-type ballast loads can be seenglowing red in the picture (upper left).

21

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 24: ANSYS Solutions Fall 2006

Ongoing Research

ANSYS CFX software has been used successfully toanalyze and identify an operational problem with theoriginal prototype turbine for the project. Furthermore,the software has proved to be a valuable design tool inthe process of developing and analyzing possible newrotor geometries for the turbine. At the time of writing,the turbine was producing approximately 4kW of electricalpower on site at a turbine efficiency of 65 percent.Ongoing research aims to further improve the turbinedesign and complement the existing CFD and field-test results with detailed laboratory testing. ■

This project is funded by a research grant awarded by

the Leverhulme Trust to Arthur Williams, Ph.D. (principal

investigator) and Shirley Ashforth-Frost, Ph.D., Nottingham

Trent University, UK.

For further information, contact Robert Simpson

([email protected]) or Arthur Williams

([email protected]) of Nottingham Trent University.

The new rotor was fabricated by a local manu-facturer in Peru, and field test data was obtained. A revised and more accurate CFD model also was created, which took into account parameters andgeometry changes that were not included in the preliminary simulations.

The revised CFD simulations had reasonableagreement with the field test results for the power output in the low-speed range. However, in the high-speed range (above 1200 rpm), the results tendto overpredict the power output when compared to the field tests. The flow rate through the turbine also is underpredicted by the CFD simulations byapproximately 10 precent. Further investigation is currently under way to determine the effect of changesto the ANSYS CFX model, including roughness effects, leakage losses through the hydrodynamic seal and cavitation modeling.

Preliminary CFD simulations were used to compare theoriginal and redesigned rotor geometry. A total pressureboundary condition was used for the inlet to simulate aconstant head of four meters.

The revised CFD simulations for the redesigned rotor showreasonable agreement with the field tests for power output.

The new rotor was manufactured locally using flat sheetmetal bent and twisted into shape and welded to the hub.

The fluid volume for the ANSYS CFX simulation wasseparated into four domains: the spiral casing, guidevanes, rotor passage and draft tube.

22

Visualization of the velocity vectors at theblade mid-span location

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 25: ANSYS Solutions Fall 2006

The explosion of computationalcapacity, now at rock bottomprices, will turn engineering simulation into an everyday tool for all engineers anddesigners across the full product development process.In such a paradigm shift, CAEbegins up front with conceptualdesign, engineers perform much of their own simulationand analytical results drivedesign decisions. Simulation

experts determine the right process, analysis assump-tions and check-points to validate results — but donot directly execute analysis work.

What’s holding us back from integrating CAD and CAE to enable simulation-driven design? Our Design/Simulation Council addresses these challenges with a proposed standard framework to integrate and optimize the divergent specialist activities that fragment design and simulation. A coreproblem relates to lack of knowledge capture and re-use regarding simulation.

Some forward-thinking companies are makingsignificant headway in leveraging the knowledge ofexpert specialists in their simulation-driven designefforts. Several initiatives rely on an abstract modelrepresenting analysis work that can be performedacross similar parts, at all stages of product develop-ment and across all analysis domains. The modelincludes three main types of information:

■ Input for the early concept phase coveringperformance requirements and functionalspecifications

■ Company rules and practices, product identification, context of the analysis and engineering knowledge or constraints that apply

■ Analysis-related information as a function of part and product type that includes the part or system definition, geometry,assumptions, loads, boundary conditions and materials data

Lack of knowledge capture and re-use holdsmany companies back from gaining the fullbenefits of analysis.

Integrating CAD and CAE toEnable Simulation-Driven Design

By Don Brown, ChairmanCollaborative Product DevelopmentAssociates, U.S.A.

Guest Commentary

23

The abstract model is used as a basis for creating aspecific physical model for analysis. Mesh generationand formatting then translate the physical model into ananalysis execution model for a particular solver.

Broad implementation of such approachesinvolves a transformation for product development.Considerable data management is required to storeexpert knowledge and data models. CAE solutions must be architected to leverage this knowledge. Mostimportant, organizational changes must meet the needsfor cross-functional collaboration to integrate designand analysis.

Leading-edge companies, including Airbus, Visteon, Whirlpool and John Deere, successfully utilizesuch an approach. At Visteon, the abstract model drivesall automotive air-handling systems across both CAD and CAE. The abstract model is defined by 15 components, 12 classes and 40 attributes. Airbus is standardizing structural analysis across all sites anddisciplines by automatically generating data needed fora particular analysis code from a CAE data model thatcontains analysis abstract models, rules, relationshipsand other information. For example, the shape anddimension of each rib in a wing can be defined from aset of design rules and constraints, and models mayapply to all frames of all aircraft based on particularplane geometry, loading conditions and materials.

Today, as in the past, simulation technology toooften is used only as a forensic tool for post-mortemchecks, validation and troubleshooting after the fact,when the worst of the damage already has been done.Too many critical product development decisions aremade based on assumptions and guesses rather thanon engineering analysis, even though the tools are readily available. Simulation should be applied up frontto avoid problems in the first place. ■

Collaborative Product Development Associates(www.cpd-associates.com) provides in-depth informa-tion for assessing technology, business goals and implementation road maps for engineering and manufacturing. The firm hosts a variety of events andengages in programs for critical analysis of decisiontrade-offs regarding design creation and validation,design/simulation council, PLM integration/product definition and product value management.

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 26: ANSYS Solutions Fall 2006

A native of Brazil, Alberto Santos-Dumont was agenius obsessed with the concept of flight. A hundredyears ago, he designed and built a plane from pineand bamboo poles covered with Japanese silk in a complex biplane, boxkite-like configuration. For flight testing, he attached the aircraft to his latest dirigible, the Number 14, and the plane becameknown thereafter as the 14-Bis.

On October 23, 1906, at the Bagatelli Field inParis, Santos-Dumont flew his aircraft 200 feet andwon the coveted Deutsch-Archdeacon Prize, whichwas created to encourage the growth of aviation. Theflight was witnessed by officials from what wouldbecome the Federation Aeronautique Internationale,and Santos-Dumont was credited with making the firstheavier-than-air powered flight in Europe.

CFD Simulation Recreates Aviation HistoryResearchers used ANSYS CFX software to study the 1906 flight of Alberto Santos-Dumont’s wood and silk aircraft, credited as the first officially recognized heavier-than-air flight in Europe.

By Leonardo O. Bitencourt, Aeronautical Engineering, ESSS, Florianópolis, BrazilRamon Morais de Freitas, Instituto Nacional de Pesquisas EspaciaisGrégori Pogorzelski, Aeronautical Engineering, Instituto Tecnológico de AeronauticaJoão L. F. Azevedo, Senior Researcher, Instituto de Aeronáutica e Espaço

Aviation pioneer Alberto Santos-Dumont From the collection of Jean-Pierre Lauwers.

24

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Simulation at Work

From the collection ofJean-Pierre Lauwers.

Page 27: ANSYS Solutions Fall 2006

To commemorate this milestone in the history ofaviation, a group of students in Brazil, assisted byEngineering Simulation and Scientific Software (ESSS,the ANSYS distributor in South America), establishedthe CFD 14-Bis Project to celebrate the historic occasion through advanced computational fluiddynamics techniques. ANSYS CFX software was chosen to analyze and understand airflow around the plane’s surface.

A CAD model was developed using historic pictures, plans and discussion. Due to the model’scomplex geometry, the students meshed surfacesusing hexahedral elements and represented volumeswith a tetra/prism mesh. ANSYS ICEM CFD softwarewas used to create both high-quality surface and volumetric meshes.

Some conclusions can be drawn from theANSYS CFX results. In his first attempt, Santos-Dumont used 24 hp nominal power and failed. In thesecond and successful attempt, he increased thenominal power to 50 hp. By analyzing drag and lift aswell as engine thrust, the possible angle-of-attack andflight speed values could be estimated: Approximatelyfive degrees and the range of 12 to 14 m/s are the predicted flight conditions. These values are higher than the speed of 11 m/s normally quoted in describing the flight.

As this value is a ground-related speed, the possible discrepancy could be due to the presence ofwind or ground effects during the centennial flight.This last factor currently is being studied. Aspectsrelated to aircraft stability still are under investigation,

but preliminary results indicate that the airplane wasstable even though small variations in the center-of-gravity position could make the airplane dangerouslyapproach unstable behavior. The simulation gaveresearchers insight into the airflow and dynamics ofthe plane as well as the genius and daring of the aviation pioneer Alberto Santos-Dumont.

The authors wish to thank Professor Paulo Greco from Escola

de Engenharia de São Carlos, Universidade de São Paulo,

who provided the geometrical CAD model, and Marcus Reis

from ESSS, who provided support and licenses for all the

software used.

Geometry of the 14-Bis

Lift dependence with flight velocity

Meshing the 14-Bis using ANSYS ICEM CFD software

Flow visualization using streamlines

Qualitative analysis of velocity

25

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 28: ANSYS Solutions Fall 2006

Headquartered in Toledo, Ohio, Dana Corporation is a leading supplier of parts and assemblies to the automotive industry. The company designs and manufactures a wide range of products for everymajor vehicle producer in the world and has twicereceived the Malcolm Baldrige National Quality Award, which recognizes U.S. businesses thatdemonstrate outstanding quality and performance.Dana is focused on being an essential partner to automotive, commercial and off-highway vehicle companies, which collectively produce more than 60 million vehicles annually.

Dana’s Commercial Vehicle Systems Group specializes in development of front-steer, rear-drive,trailer and auxiliary axles; driveshafts; steering shafts;suspensions; and related systems, modules and services for the world’s commercial vehicle market.Dating back to the company’s beginnings in 1904,Dana products have helped drive history’s greatestvehicles, from the Model T and World War II-era armyvehicles to London taxicabs, 18-wheel rigs, giantearth-moving machines and cars on the NASCAR racing circuit. Building on this foundation of

Weight-Optimized Design of a Commercial Truck FrontSuspension ComponentEngineers at Dana Corporation use topology optimization featuresof ANSYS Mechanical software to reduce upper control arm weightby 25 percent while maintaining required stiffness and strength.

Designed by Dana Corporation forcommercial vehicles, the upper control arm (highlighted) is a critical component in vehicle front suspension systems. (Dana patents pending.)

Simulation at Work

26

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 29: ANSYS Solutions Fall 2006

experience, Dana continues its commitment to qualityand innovation in advancing the science of mobility forthe benefit of its broad global customer base.

The Drive for Lighter Suspensions

Design of suspension systems and other assembliesfor heavy trucks is a formidable task due to heavyloads, harsh environments and long-life requirements.Historically, components tended to be over-designedheavy structures to meet these reliability requirements.But in today’s economy, the weight of commercialtrucks and its impact on vehicle cost, ride and fueleconomy are of significant concern for both truckmanufacturers and end users.

Lighter, well-designed suspensions provide better ride quality, lower initial cost, increased fueleconomy and greater truck payloads. The challenge isto design these parts with minimal material yet still maintain adequate strength and stiffness — allwhile meeting tight budgets and product launchschedules that rule out building and testing numeroushardware prototypes.

Working with Topology Optimization

Dana Commercial Vehicle engineers use topologyoptimization features of ANSYS Mechanical softwareto optimize component weight as part of the product design process. The method begins by determining loads from multibody simulation. Then aninitial rough solid model is constructed to fill the maximum available space envelope allowed for the component. Next, a finite-element mesh is developed, and the ANSYS topology optimization routine automatically eliminates elements with stiffness belowa specified threshold.

The result is an analysis model consisting of onlythose elements needed to maintain the required stiffnessof the component with minimal material. This topology-optimized model then is overlaid on the solid model to

guide engineers in completing the detailed design ofthe weight-optimized part.

Key Part of Product Development

Because weight reduction is a critical issue, this optimization approach is used extensively with considerable success at Dana. In the development ofan upper control arm for the front suspension of acommercial truck, for example, engineers reducedpart weight by 25 percent while maintaining requiredstiffness and strength.

The process was completed in less than a day,compared to weeks otherwise needed for trial-and-error iterations on expensive physical prototypes.Moreover, the optimization guided the design in adirection that was not intuitively obvious and providedengineers with greater understanding of componentbehavior and stiffness transfer paths. The approachhas been standardized as a best practice in the groupand now is applied readily to optimize the weight ofmost of Dana’s commercial truck components.

“The topological optimization capabilities ofANSYS Mechanical software represent a key part of our work in developing well-designed, lightweightsuspensions and other assemblies that meet the stringent requirements of the commercial truck industry,” explains Caner Demirdogen, senior principalengineer at Dana Corporation Heavy Vehicle Technologies and Systems. He notes that the optimization approach implemented at Dana enablesengineers to take advantage of this technology inreducing component weight much more quickly andcost-effectively than trying to accomplish the samegoals with physical prototypes.

According to Demirdogen, “Such techniquessave considerable time and expense in developingrefined designs, give us tremendous insight into component behavior and are a business requirementfor effectively designing tomorrow’s innovative products in the competitive automotive industry.” ■

27

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Starting with an initial mesh (left), the ANSYS topology optimization routine automatically eliminates elements with stiffness below a specified threshold (right). The result is an analysis model consisting only of the elements needed to maintain the required stiffnesswith minimal material. (Patent-pending design)

Page 30: ANSYS Solutions Fall 2006

Developing ConstructionProducts with Better FirePerformanceANSYS tools provide models to accurately simulatethe complex physics of a building fire.

By Yulian Spasov, Ph.D., Corus Research Development and Technology, Rotherham, UKand Yehuda Sinai, Ph.D., ANSYS Europe Ltd., Abingdon, UK

Simulation at Work

28

Fire performance is an extremely important aspect of the innovative new building products being developed for the construction market at CorusResearch Development and Technology (CorusRD&T). Although standard furnace tests are conductedto test compliance with current fire regulations, Corusis taking this a step further by using simulation and virtual testing to study the behavior of new products under conditions that accurately representreal building fires.

ANSYS CFX software was chosen as the CFDtool in this virtual testing because of its reputation formodeling combustion, radiation and buoyant flows,and because Corus has a long tradition of using thesoftware. A first step toward developing reliablesimulation practices has been to validate the softwareand modeling capabilities for specific situationsagainst available experimental data.

Simulating Fire Tests

In 1993, the Fire Research Station of the BuildingResearch Establishment (BRE), in collaboration withBritish Steel Technical, Swinden Laboratories (nowCorus RD&T) conducted nine fully developed fire tests in a large compartment in the BRE Cardington Laboratory. The photographic material and experimental data presented here are extracted from referenced documents. The fire load consisted of 33 dry wooden cribs located on the floor of the compartment. The ventilation opening wasobstructed by a column, and all cribs were set onfire simultaneously. The experiment lasted threehours, and the fire load was 20 kg/m2 of dry timber,which was completely burned by the end of the experiment. Temperature, velocity, radiation intensitiesand major species concentration were measured atselected locations. The weight of several cribs alsowas monitored to estimate the actual crib burning rate.

By validating simulation results with an experimental fire, Corusis confident in using ANSYS CFX to assist in improving the fireperformance of their products.

Experimental setup of dry wooden cribs in the fire test compartment

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 31: ANSYS Solutions Fall 2006

29

For simulation, ANSYS DesignModeler, ANSYSCFX-Mesh, ANSYS ICEM CFD Hexa and ANSYS CFX10.0 products were used to build the geometry, fluidmesh, solid mesh and simulation model, respectively.The model takes advantage of eddy dissipation combustion, Monte-Carlo gray radiation and the SST turbulence models in ANSYS CFX software. Conjugate heat transfer through the solid walls wasmodeled using the generalized grid interface capability for “gluing” together the unstructured fluid and structured solid meshes. The computational domainwas extended to include a region outside the compartment in order to minimize uncertainties in theboundary conditions. The heat-release rate of eachcrib was estimated by interpolation of the availabledata for crib mass loss rate. An inflated mesh wasused in the near-wall region, and a finer mesh close tothe cribs. The fluid domain contained 150,000 nodes.

Predictions Agree with Experimental Data

Temperature readings were taken in four groups oflocations. Comparison of the experimental data withthe predicted temperature shows a good agreementfor all groups for the duration of the simulation. Thepresent results are obtained only using best practices

References

B.R. Kirby, D.E. Wainman, L.N. Tomlinson, T.R. Kay, B.N.Peacock, “Natural Fires in Large Scale Compartments,” aBritish Steel Technical, Fire Research Station collaborativeproject, 1994.

GME Cooke, “Tests to Determine the Behaviour of FullyDeveloped Natural Fires in a Large Compartment,” FireNote 4, Building Research Establishment Ltd., FireResearch Station, 1994.

for modeling of this type, and no adjustments havebeen made to model parameters in order to obtainbetter agreement with experimental data.

Building fires involve complex physics, and theirmodeling requires accurate combustion, radiation and turbulence models. The model within ANSYS CFX software provided results that are accurate bothspatially and temporally. The intuitive nature of theuser interface, the embedded parametric capabilitiesand the reliability of the available physical modelsshowed that ANSYS CFX software is an invaluabletool that Corus can use to enhance the fire performance of its products. Demonstrating the fireperformance of whole modular compartments builtwith Corus products would be almost impossible todo experimentally. Corus now is able to use the ANSYS CFX model with confidence in situations in which experimental data is not available or would betoo expensive and time-consuming to obtain bymeans of physical models. ■

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Comparison with experimental data. Groups A and B (top left and right) thermocouples arelocated near the ceiling at the back and middle of the compartment, while Group C (bottomleft and right) thermocouples are located inside the compartment, just behind the column.

Simulation results for temperatureafter 20 minutes (top) and 40 minutes (bottom)

Page 32: ANSYS Solutions Fall 2006

Bringing High-PerformanceComputing to the MainstreamMicrosoft® Windows® Compute Cluster Server lets engineers andanalysts easily deploy, operate and manage workstation networksthat, until now, only IT professionals could set up.

By Kyril FaenovDirector, High Performance Computing GroupMicrosoft Corporation, U.S.A.

Traditionally, high-performance computing (HPC) has been used primarily by those lucky enough to have massive endowments or grants, with access to time on the supercomputer strictly rationed. HPC generally meant a single, large, symmetric multiprocessing (SMP) or vector supercomputer and a budget to match.

The ANSYS Workbench environment allows for remotesolutions, letting a high-performance computer clusteraccelerate work transparently.

Solving cutting-edge problems in science, engineering and business has always demanded capabilities beyond those of even the fastest work-stations. Market pressures demand an acceleratedproduct development cycle and reduced time-to-insight. The use of commodity servers in computeclusters now brings the power of supercomputing tothe workgroup or even desktop level.

30

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Technology Update

Page 33: ANSYS Solutions Fall 2006

In 1991, the cost of a 10 Gflop Cray supercom-puter was $40 million; the computers were availableonly in government labs, very large corporations oracademic research institutions. Today, that same 10Gflop computing power is available in a four-nodecompute cluster that can be bought off-the-shelf for $4,000.

This proliferation of HPC has an enormous effecton research and industry, making it possible to solveproblems that simply couldn’t be attempted before.But it also presents challenges for both vendors andconsumers of HPC.

Challenges of Compute ClustersWhen HPC was a single, large integrated supercom-puter, the researcher didn’t have to worry about howto deploy or manage the system — that was the job of(usually several) dedicated information technology (IT)professionals, with accompanying salaries and costs.But as we move to compute clusters of commodityservers, especially at the workgroup level, the

Windows CCS 2003 has a four-step wizard that easily configures networking, deploys compute nodes without IT interventionand manages cluster users.

challenge is to provide an installation and manage-ment experience that doesn’t distract from the reasonfor the cluster in the first place — without requiringmajor IT resources that would change both the economics and the immediacy of the interaction thatare driving this proliferation.

Another area of concern is security. When HPCwas a single supercomputer with carefully controlledand allocated access, security was inherent in theprocess: The system was physically isolated and hadno direct connection to corporate or other networks.Each researcher’s job ran as a discrete, self-containedbatch job. Today’s HPC compute cluster often isdirectly connected to the corporate network, and it isshared across a diverse group of scientists, engineersor analysts. Individual jobs may use only a portion ofthe cluster at any one point, with other jobs runningsimultaneously on other nodes in the cluster. Securitymust be built in to the overall HPC environment to protect the integrity of the cluster and the individualjobs that run on it, as well as the corporate network onwhich it resides.

Networking

31

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 34: ANSYS Solutions Fall 2006

Accelerating Work TransparentlyThe biggest benefit of an HPC compute cluster is availability. Compute clusters consisting of commodityservers have driven down the price per Gflop to thelevel at which anyone can reasonably afford his own HPC, and results are available faster and moreinteractively. By taking advantage of applications such as Distributed ANSYS, users can perform theirsimulations quicker, get more people involved in thesimulation and reach product decisions faster.

ANSYS, Inc. has embraced the client–servercompute model in the ANSYS Workbench environ-ment working with Distributed ANSYS. This enablesusers to focus on performing simulations and allowsfor remote solutions in the process, thereby letting the advanced computing power of an HPC cluster accelerate the work transparently.

Role of CCS 2003Microsoft® Windows® Compute Cluster Server (CCS)2003 provides a complete HPC cluster solution that:

■ Easily installs and deploys on commodity servers

■ Is highly secure and robust■ Easily integrates into existing data

environments■ Leverages existing IT infrastructure■ Scales from personal clusters to TOP500

clusters transparently■ Brings HPC mainstream

By integrating directly with Microsoft ActiveDirectory®, CCS 2003 leverages the existing IT

The familiar Windows infrastructure is used by CCS 2003 in itsMicrosoft Management Console for managing compute clusters.

Windows

For More Information

■ Microsoft HPC public Web site: www.microsoft.com/hpc

■ CCS 2003 community site: www.windowshpc.net/default.aspx

■ Windows Server x64 information: www.microsoft.com/x64

■ Windows Server System information: www.microsoft.com/wss

■ ANSYS Workbench information: www.ansys.com/products/workbench.asp

■ Distributed ANSYS information:www.ansys.com/products/parallel.asp

infrastructure, including Group Policy, to provide asecure cluster environment. A simple four-step wizardconfigures the networking, installs and configures theremote installation service (RIS) to deploy computenodes from bare metal without IT intervention, and manages the cluster’s users. Node deployment issimple and painless, making it possible to create aprototype solution on a small cluster and then scale itup as needed.

CCS 2003 leverages the familiar Windows management infrastructure as well, using MicrosoftManagement Console (MMC) 3.0 for cluster management while supporting Microsoft OperationsManager (MOM) for monitoring and management, and supporting Microsoft Systems Management Server (SMS) for node updates.

CCS 2003 includes Microsoft Message PassingInterface (MS-MPI), a secure and fully compatibleinterface based on the Argonne National LaboratoryMPICH2 reference implementation to simplify migration of existing code and integration with existingHPC environments.

The new Job Scheduler supports both a full command-line job management and a graphical interface to simplify job submission and monitoring.Job scheduling is part of the end-to-end security of CCS 2003, which leverages Active Directory toallow jobs to run with domain user credentials while protecting those credentials from other jobs in the queue.

Finally, CCS 2003 leverages Microsoft® VisualStudio® 2005 to provide a familiar, integrated develop-ment environment that supports remote paralleldebugging on the cluster, allowing developers toquickly test and debug their code. The rich Visual Studio environment gives application developers thetools and language choices they need.

SummaryThe wide availability of compute clusters running on commodity hardware is driving the proliferation of high-performance computing. Microsoft WindowsCompute Cluster Server 2003, working with applications such as Distributed ANSYS, provides a high-performance computing platform that is simple to deploy, operate and integrate with existing infrastructure and tools. ■

Technology Update

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

32

Page 35: ANSYS Solutions Fall 2006

By John CrawfordConsulting Analyst

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

33

These handy do-it-yourself files canbe real time-savers, both now andfor future projects.

Running Solutions from Macros

It’s interesting how the way weuse ANSYS evolves as webecome more familiar with theprogram. Most people begin byusing the menus for modeling

building and entering all the information needed to runan analysis. As they gain experience, their use ofmenus gradually is supplemented by entering commandsvia the command line because it allows them to workmore quickly. With more familiarity, it’s an easy transi-tion to use our knowledge of commands to writemacros and other input files in ANSYS ParametricDesign Language (APDL). Doing this taps into the realpower of ANSYS technology, whether it’s a macro for asingle specific application or one that can be used overand over again.

One of the practices I’ve used over the years is torun almost all my solutions with a macro or input file.Macros and input files are exactly the same; they differonly in how they are executed. A macro has a .macextension and can be executed by typing the macroname in the command line. An input file contains thesame data but doesn’t have the .mac extension, so youneed to use the /INPUT command to tell ANSYS toinput the file and execute the commands that residewithin it.

Well Worth the Effort

Regardless of whether you use a macro or an input fileto run a solution, the benefits of working in this mannermake it worth the effort to write them. When doing ananalysis, I usually build the model and get everythingset up and ready to go, with the exception of most ofthe loads and solution settings. I may include some ofthe boundary conditions that I don’t expect to changein the model, and I write a macro that contains otherloads and solution settings for each load step that willbe run. I usually define loads as parameters at the topof the macro so they can be changed easily by myself

or anyone else who might use the macro to rerun theanalysis at a later date.

The solution macro usually begins by resumingthe startup model; then it assigns values to parametersthat I want to use in the solution. When doing a modelanalysis, I might define the frequency range and numberof frequencies to solve for as parameters and then usethese to control how the solution is performed. Or Imight use a parameter to tell ANSYS whether theexcitation for a PSD analysis is in the X, Y or Z direction.The macro then uses these parameters as it progresses through the application of loads andsolution settings and runs the solution for each loadstep or substep. I always include comments in themacro that explain what is being done and why.

Saving Time and Trouble

Writing solution macros may sound like a lot ofunnecessary extra work, but, for all but the simplestsolutions, it’s a tremendous productivity enhancer.Here are just a few reasons why I use solution macrosin almost all my work:

■ It gives me a history of the loads and solutionsteps to which I can refer when reviewingresults. Sometimes it’s difficult to rememberwhat loads or options I applied for a specificload step or substep. Having them listed in amacro removes all uncertainty regarding howthe solution was obtained.

■ Rather than starting ANSYS and resuming amodel, I can refer to the macro when someoneasks about the loads, the number of loadsteps and other solution controls.

■ I can revisit an analysis later and rerun it in anefficient and consistent manner. This is usefulwhen I need to run different loads and want toduplicate the earlier solution as a startingpoint.

Tech File

Page 36: ANSYS Solutions Fall 2006

34

Tech File

element stiffness, number and duration of time steps,and many other things that I needed to adjust toensure that the results were accurate and useful.

I’ve used solution macros for quite a while andnow have a substantial library of them. The next time I want to do a power spectral density analysis, I have amacro that I can refer to that will help me do it again.Single point response spectrum? No problem. Tran-sient dynamic analysis? I have macros that will helpme with that. Modal cyclic symmetry? I have a coupleof those macros. An axisymmetric transient heat transfersolution that has the results applied to a 3-D structuralanalysis? That’s in my library as well. Superelementgeneration, use and expansion passes used in severaldifferent types of analyses? I have all of these.

Whether I last ran a specific type of analysis lastweek or five years ago doesn’t matter. I can review the solution macro to retrieve the insight and understanding I had at that time and quickly performthe analysis that I want to run today. BecauseANSYS emphasizes upward compatibility in new releases, the commands that worked in my earlieranalysis will work in my new analysis, too.

Leveraging Knowledge of ANSYS Tools

Solution macros not only benefit me, but they alsohelp my customers. When I finish an analysis, I turnover the model and the macros I used to run the solution and perform the post-processing. This allowsthe customer to duplicate the results at a later date orto make changes to the loads and run a new solution.

By putting the loads at the top of the macro asparameters and taking the time to thoroughly makecomments on what they are and how they are used,the customer is able to change the parameters andrun a new solution without any special understandingof APDL or ANSYS. This allows the customer toleverage my knowledge of ANSYS to his benefit, so he can run additional solutions without my beingdirectly involved.

I didn’t use solution macros on a regular basisuntil I had been an ANSYS software user for quite a while. But when I began writing and using them, I discovered that it was a worthwhile investment intime and effort that has paid off over and over again.Maybe it will pay off for you as well. ■

■ I can easily vary solution settings (contactelement stiffness, convergence criteria, substep size, etc.) to see what effect they have on the analysis. In this manner, everysolution begins from the same starting point.

■ The macro can automatically calculate settingsthat are unique to a specific solution. Forexample, if I want to run a series of harmonicanalyses that vary from one frequency toanother, I can have the macro calculate alphaand/or beta damping for each frequency.

■ It greatly reduces the possibility for errorsbecause I have a document that confirms what I did to get the latest solution.

■ The macro can be listed in the appendix of a report so readers know exactly what was done.

■ The macro is easily shared with others whowish to do similar solutions.

■ I can refer to the macro the next time I do a similar type of solution and use it as a foundation for future work.

Benefits of a Macro Library

While all the above reasons make macros well worththe time needed to write them, I think the last reasonbenefits me the most. I have a library of solutionmacros that cover a wide range of analyses done inthe past, and I frequently refer to these when tacklingsimilar projects. Here is an example.

Several years ago I did a transient analysis of apressure wave striking a panel to determine how thestress wave propagates through the panel. This was afairly complicated analysis with very small time incre-ments and several thousand load steps. A few yearslater, I was asked to model a sphere impacting amachined part to see how quickly the stress wavesmoved through the part, what the alternating stresseswere, and how much reflection and damping weretaking place. Rather than re-invent the wheel, I pulledup the macro from my earlier analysis to refresh mymemory of how I had previously solved this type ofproblem, and I used it as the basis for the morecomplicated analysis I was now beginning to work on.This enabled me to work more quickly and avoid someof the pitfalls I had worked through previously. Once Ihad the new macro running, I could vary contact

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 37: ANSYS Solutions Fall 2006

Tips and Techniques

When multiple physics problems are coupled via material response, the physics need to be solved simultaneously. This is a typical application of the use of 22x coupled-field elements, which have multiple degrees-of-freedom (DOF) and can support piezoelectric, piezoresistive, thermoelastic and thermoelectric materials.

PiezoelectricityIn piezoelectric materials, structural and electric fieldsare coupled so that an applied voltage generates astrain (and vice versa). Consequently, piezoelectricceramics are used as transducers to convert electricalenergy to a mechanical response or as sensors to con-vert mechanical energy to an electrical signal.

Mechanical stress {T} and strain {S} are related toelectric displacement {D} and electric field {E} via thefollowing constitutive equations:

Here, [sE] is the compliance matrix evaluated atconstant electric field, [εT ] is the permittivity matrixevaluated at constant stress and [d] is the piezoelectricmatrix relating strain to electric field.

The above relationship provides a basis for theFEA piezoelectric matrix equations:

Description of terms:■ [KZ ] contains the piezoelectric effect, and the

piezoelectric constants can be input as either[d] form (strain/electric field) or [e] form(stress/electric field).

■ [Ku ] contains (anisotropic) stiffness coefficients, and these are either compliance[sE ] or stiffness [cE ] coefficients.

Working with Coupled-Field Elements

■ [KV ] is the (anisotropic) permittivity matrix withpermittivity values evaluated at constant strain[εS] or constant stress [εT ].

■ [Cu ] is the structural damping matrix, whereas[CV ] represents dielectric losses.

Issues to consider: Since coupling is via the coefficient matrix, a single iteration is required for calculating coupled effects. Elements support nonlinearstatic, modal, harmonic response and transient analyses. The dielectric loss tangent tanδ can be inputvia MP,LSST, which is part of [CV ]. Manufacturers’ datatypically have mechanical vectors in the form {x, y, z, yz,xz, xy}, whereas ANSYS requires the input to be {x, y, z,xy, yz, xz} — so the 4/5/6 terms need to be rearranged.

PiezoresistivityFor piezoresistive materials, an applied mechanicalstress or strain causes a change in the material’s resistivity for use as sensors, for example, in which amechanical load affects the electrical signal.

The electrical resistivity [ρ] is related to stress as follows:

in which [ρo] is the input (nominal) resistivity and [π] is the piezoresistive stress matrix. Alternatively, thepiezoresistive strain matrix [m] may be input instead,relating relative change in resistivity to strains.

The resulting matrix equations are:

Description of terms:■ [KV ] is the (orthotropic) electrical conductivity

matrix, which includes piezoresistive effects as described above.

■ [Ku ] contains the (anisotropic) stiffness coefficients.

PLANE223, SOLID226 and SOLID227 elements readily handle multiphysics problems coupled via material response.

By Sheldon Imaoka / Technical Support Engineer / ANSYS, Inc.

35

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 38: ANSYS Solutions Fall 2006

Tips and Techniques

Issues to consider: Coupling is in the coefficientmatrix, but resistivity is based on calculated stress orstrain, so multiple iterations are required for calculatingcoupled effects. Elements support nonlinear static andfull transient analyses. The change in resistivities canbe measured by a Wheatstone bridge configuration,since this is how piezoresistive materials usually arearranged. Input of permittivity allows for definition of[CV ] to account for dielectric effects.

ThermoelectricityIn thermal–electric applications, two types of couplingare present. Joule heating is an irreversible processoccurring when current flows through material withelectrical resistance, proportional to the currentsquared and independent of the current direction:

Thermoelectricity consists of the reversible Seebeck, Peltier and Thomson effects.

The Seebeck effect, defined by the coefficient α,relates a temperature gradient with a potential difference. An example application is MEMS powergeneration converting heat to electrical power:

The Peltier effect, noted by π, is the reverse condition in which a current causes a heat differential,and the direction of the current determines whetherheat is removed or input. Typical examples are thermo-electric coolers:

The Thomson effect, described by the coefficientµ, illustrates what occurs when a current flows througha material with a temperature gradient:

Seebeck, Peltier and Thomson coefficients are related using absolute temperature To, so only Seebeck coefficients need to be defined:

Thermal–electric equations are incorporated intothe FE matrices as follows:

Description of terms:■ [KVT] is the Seebeck coefficient coupling matrix.■ Joule heating {Qj} and the Peltier effect {Qp} are

included in the load vector.

The Thomson effect is not explicitly includedabove, since it is accounted for when temperature-dependent Seebeck coefficients exist.

Issues to consider: Since Joule heating and Peltier effect are accounted for by load-vector coupling, thermal–electric analyses are iterative innature. Material behavior supports nonlinear static and transient analyses. Absolute temperature needs to be defined via TOFFST.

ThermoelasticityThermal–stress analyses are commonplace, in whichthe temperature field is calculated and imposed as aload vector on the structural model. However, thepiezocaloric effect (thermoelastic damping) also can bemodeled with 22x coupled-field elements for dynamicapplications. The constitutive relations are:

in which {α} is the coefficient of thermal expansion, S isentropy density and To is the absolute temperature. The combined system of equations is expressed inmatrix form:

Description of terms:■ [KuT] is the thermoelastic stiffness matrix

(thermal expansion term) while [CTu] is the thermoelastic damping matrix.

Issues to consider: The system of equations isunsymmetric, although by having it matrix-coupled, theeffects are considered in a single iteration. Nonlinearstatic, full transient or harmonic response analyses areavailable. The piezocaloric term (also known as thermoelastic damping because of the coupling of theenergy equation) is present only for dynamic (harmonicor transient) analyses and does not act like structural orviscous damping by always lowering the resonant frequencies. Coefficient of thermal expansion providesthe coupling response for both [KuT ] and [CuT ] terms. A temperature offset via TOFFST is required, with the reference temperature designating strain-free temperature. ■

Contact the author at [email protected] for theentire paper from which this article was excerpted.

36

www.ansys.com ANSYS Solutions | Volume 7, Issue 4 2006

Page 39: ANSYS Solutions Fall 2006