ansys advantage 1-2-07

COMPOSITESPAGE 7

HAIR CAREPAGE 10

INSOLE DESIGNPAGE S7

THE PROVEN WINNER IN CAEPAGE 3

A D V A N T A G EE X C E L L E N C E I N E N G I N E E R I N G S I M U L A T I O N

V O L U M E I I S S U E 2 2 0 0 7

www.ansys.comANSYS Advantage • Volume I, Issue 2, 20072

CAE technology, the competitive edge isn’t necessarilydetermined by which companies use simulation — mostmanufacturers now implement it in one way or another —but rather how they uniquely apply the technology in theirorganizations and integrate it into their product develop-ment processes.

To fully leverage a solution, many successful firms haveinitiatives for performing more upfront simulation to refinedesigns early instead of trying to hurriedly fix problemsnear the end of development. In most cases, this meansdeploying appropriate tools beyond the ranks of dedicatedanalysts to more rank-and-file engineers and designers forroutine use throughout development.

There is no cookie-cutter way to best implement suchan approach. Rather, companies have found that they mustcarefully evaluate their existing processes, skill sets, organizational structures, product strategies and businesspriorities to leverage simulation most effectively. Scheduling,funding and performance reviews generally are adjusted toallow for training; the approach also gives engineers and designers the time they need to perform analysis, what-if simulations and optimizations in the early stages ofdevelopment rather than the usual rush to finalize computer-aided design (CAD) models.

These and other necessary organizational changesrequire a significant investment in time and effort, ofcourse, but the level of commitment defines how companies uniquely leverage simulation; it also determineswhich firms will most likely lag behind while others reap the greatest business value from Simulation Driven Product Development.

John Krouse, Editorial Director

EDITOR’S NOTE

The Big Question

Companies are investing inengineering simulation at unprece-dented levels. According to thelatest statistics from marketresearch and technology assess-ment firm Daratech, sales revenuefor computer-aided engineering(CAE) software and services grewfrom $2.31 billion in 2005 to morethan $2.43 billion in 2006. Daratechforecasts a compound annualgrowth rate of 13 percent through

2010, when figures are expected to top $3.7 billion. Driving this expansion is the tremendous need for companies to shorten time to market, lower costs, improveperformance and develop steady streams of knock-out,innovative products. With survival on the line in many cases, manufacturers use simulation as a proven way ofaddressing these issues.

According to Daratech statistics, the number-one playerin CAE is ANSYS, Inc. So a substantial portion of the broadrange of simulation applications worldwide is based on thecompany’s suite of solutions. The breadth of applications isevidenced by the articles in this current issue of ANSYSAdvantage on simulation projects involving the design ofproducts ranging from trucks and turbojet engines to con-sumer goods and healthcare equipment. The content alsodemonstrates the vast range of company sizes, from theone-man design firm Stein Design in the story “No-HassleKitchen Appliance” to the $70 billion global consumer product giant Procter & Gamble Company in the article “TheDemocratization of Engineering Analysis.”

As these and other successful simulation users know,gaining market advantage now takes more than just utilizing analysis tools. Because of the ubiquitous use of

With engineering simulation becoming so widespread, the big questionnow for manufacturers is not if to use the technology but how. Andtheir answers will determine who gains the competitive edge.

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

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

Editorial DirectorJohn Krouse

Production ManagerChris Reeves

Art DirectorSusan Wheeler

EditorsMarty MundyFran HenslerErik FergusonRichard LaRocheThierry Marchal

Ad Sales ManagerBeth Mazurak

Editorial AdvisorKelly Wall

Circulation ManagersElaine TraversSharon Everts

DesignersMiller Creative Group

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, DesignModeler, ANSYS Mechanical, DesignSpace, ANSYS Structural, TGrid, GAMBIT and any and all ANSYS, Inc.brand, product, service, and feature names, logos and slogans are registered trademarks or trademarks of ANSYS, Inc. or its subsidiaries located in the United States or other countries. ICEM CFD is a trademark licensed by ANSYS, Inc. All other brand, product, service and feature names or trademarks are the property of their respective owners.

© 2007 ANSYS, Inc. All rights reserved.

About the CoverRacing yachts competed thissummer in the America’s Cupcompetition in which all the leading teams employed computersimulation to gain an edge. Seepage 3. Cover photograph ©ACM2007/Photo: Carlo Borlenghi.Simulation courtesy ChristosPashias, Team Shosholoza.

About the BiomedicalSpotlightThe biomedical industry is emerging as a strategic user of engineering simulation.One research team has foundthat improvements in cochlearimplants might be possibleusing shape memory alloys.See page s4.Email: [email protected]

ANSYS Advantage • Volume I, Issue 2, 2007

CONTENTS

www.ansys.com 1

3 SPORTS

The Simulation Race for America’s CupYacht designers used engineering simulation in a variety of applications to edge out the competition.

ContentsFeature

7

12

14

16

20

22

24

10

3

26

(Continued on next page)

7 MATERIALS/PARTNERSHIPS

Plying the Composite TradeESAComp software overcomes challenges in designing with composites.

10 CONSUMER PRODUCTS

Hair TodayProduct developers in the cosmetics industry can put simulation to use in performing hierarchical analyses of hair care product performance.

12 AUTOMOTIVE

Heavy-Duty LightweightAn innovative aluminum design gives a truck-body manufacturer the competitive edge in the worldwide construction industry.

14 POWER GENERATION

Gassing Up with CoalA two-fluid multiphase model allows for more accurate simulation of coal gasification.

16 PROCESS EQUIPMENT

Chopping Away at SolidsCFD simulation provides a pump company with a virtual test facility.

20 CONSUMER PRODUCTS

No-Hassle Kitchen ApplianceFinite element analysis helps redesign a countertop water filter.

22 AEROSPACE

Overcoming Big Challenges forSmall Turbojet EnginesEngineers used FEA to develop an impeller for a microjet turbine engine for unmanned drone aircraft.

24 POWER GENERATION

Keeping It CoolModeling fluid flow and heat transfer throughout a nuclear fuel assembly helps prevent reactor burnout.

26 PROCESS EQUIPMENT

The Greening of Gas Burner DesignSimulation assists in developing efficient and environmentally friendly recuperative burners used in heat-treating applications.

Applications


CONTENTS

s2 Making Life Longer and BetterThe biomedical industry is emerging as a strategic user of engineering simulation.

s4 Turning Up the VolumeThe use of shape memory alloys offers the promise of better functioning in cochlear implants.

s6 Hip to SimulationEvaluation of designs for a hip replacement prosthesis overcomes physical and scientific limitations.

s7 Walking Pain FreeNew insoles designed with the ANSYS mechanical suite relieve pain from foot disease.

s8 Engineering Solutions for Infection ControlSimulation assists in designing a hospital ward to reduce the airborne transmission of disease.

s10 Standing Up RightANSYS Multiphysics sheds light on the wonders of the human spine and how to fix it.

s12 Designing with HeartCFD-based design optimization for a pediatric implant can shave years off the development cycle.

s14 Going with the FlowFunctional biomedical imaging through CFD provides a new way of looking at pathological lungs.

s15 Battle of the BulgeRapid prototyping results in a new surgical tool to treat back pain.

19 TRENDS & PRACTICES

Managing Engineering KnowledgeWeb-based solution is aimed at hosting and integrating simulation data, processes and tools for more effective Simulation Driven Product Development.

28 THOUGHT LEADERS

The Democratization of Engineering AnalysisTo compete successfully in today’s business climate, Procter & Gamble makes analysis tools available to rank-and-file engineers as well as toanalysts and advanced simulation experts.

31 ANALYSIS TOOLS

Rotordynamic Capabilities in ANSYS MechanicalUseful features are available to study vibration behavior in rotating parts.

34 TIPS & TRICKS

Submodeling in ANSYS WorkbenchTo obtain accurate stress in a local region, submodeling separates local analysis from the global model.

Spotlight on Engineering Simulation in theBiomedical Industry

28

19

31

34

s2s4

s7

s8

s10

s14

s15

s12

s6

Departments

The America’s Cup is the most famous sailing regatta inthe world and also the oldest active trophy in internationalsport. The trophy, originally known as the Royal YachtSquadron Cup, was first awarded in 1851 when the NewYork Yacht Club schooner America defeated 15 Royal YachtSquadron challengers in a race around the Isle of Wight inEngland. In honor of America’s victory in the first competi-tion, and the subsequent dominance of American boats forover a century, the trophy officially became known as theAmerica’s Cup.

Despite its name, it is truly an international competition.In 2003, the Swiss challenger Alinghi defeated Team NewZealand to win sailing’s grand prize; Alinghi successfullydefended this summer at the 32nd America’s Cup in Valencia, Spain. The boat sizes and designs have variedthrough the years, ranging from the 130-foot J-class yachtsof the 1930s to a 60-foot catamaran in 1988. Since 1992though, the teams have sailed an International America’sCup Class (IACC) sloop, a monohull boat that has an average length of about 75 feet. To determine which

team would challenge Alinghi for the trophy in 2007, anambitious schedule of regattas was held, commencing in2004 and culminating with the Louis Vuitton Cup this pastspring. The America’s Cup match series was held in lateJune and early July, with Alinghi the winner in the closestCup in recent history.

The racing syndicates that compete for the cup are composed of the best sailors, designers, sailmakers,nautical engineers and boat builders in the world. The topteams expend more than 150,000 labor hours to optimizethe designs of their boats. All of the leading teams employcomputer simulation to determine the power generated bythe sails, the drag produced by the boat’s hull and the airresistance of the deck. Four of the top teams, includingBMW ORACLE Racing from the United States, SouthAfrica’s Team Shosholoza, Emirates Team New Zealand(ETNZ) and defending champion Alinghi from Switzerland,use computational fluid dynamics (CFD) software fromANSYS, Inc. to predict the effect of design alternatives onyacht performance down to the smallest details.

Emirates Team New Zealand used CFD to predict theeffect of design alternatives on yacht performance.

The Simulation Racefor America’s CupYacht designers used engineering simulation in a varietyof applications to edge out the competition.


SPORTS

www.ansys.com 3


The two most critical aspects of yacht performance arethe sail aerodynamics and the hydrodynamics of the hulland appendages. Picture this analogy: A racing yacht is likea plane floating on its side with one wing sticking up in theair and the other down into the water. The art of yachtdesign is to extract drive force because the two fluids (airand water) have different speeds and directions. The curva-ture of the sails generates lift in a manner similar to anairplane wing, while the keel of the boat generates lift in theopposite direction — like the opposite wing of the airplane— to prevent the boat from moving sideways. The keel canbe proportionately much smaller than the sails because it operates in a fluid 800 times denser than air. As in aircraft design, improving performance of a racing yacht isbasically a question of maximizing lift and minimizing drag.Small changes in geometry often make the differencebetween a competitive boat and an also-ran.

BMW ORACLE Racing: It’s In the DetailsIn the 2003 competition, BMW ORACLE Racing used a

public-domain CFD code to simulate the performance oftheir boat. However, they found that meshing and solutiontimes were so long that they were forced to simplify theirmodels to the extent that they could not distinguishbetween small design changes. For the 2007 race, the teamused ANSYS CFX software. BMW ORACLE Racing ranmodels with 10 to 15 million cells on large computer clusters that can resolve the performance impact of thesmallest design changes. The team’s designers simulatedthe performance of large numbers of different sail shapesand trims to understand performance under a variety ofconditions. They evaluated the aerodynamic effects of thedeck, such as the shape of edges and corners and the

position of the winches, and they also looked at the shapeof underwater components, such as the ballast bulb.

“Our new simulation methods make it possible to modelthe most complex problems down to the finest details in aday or two,” said Ian Burns, design team coordinator forBMW ORACLE Racing. “We now can determine the effectof the smallest changes, such as the shape of the deck orsmall hardware components on the mast. Some of thesechanges can have a significant impact on performance andare helping us make significant performance improvements.We have analyzed and improved nearly every detail of theboat with ANSYS CFX software.”

CFD simulates the wind flowing over the deck andcockpit of the Alinghi boat. Note the vortex thatformed in the bow where the wind wraps aroundon the deck.

An upwind aerodynamic simulation of the Team Shosholoza yacht clearly shows the tipvortices. Induced drag reduction is important for sails operating near their maximum lift.

SPORTS

ANSYS Advantage • Volume I, Issue 2, 2007www.ansys.com 5

SPORTS

Team Shosholoza: Big Things from Small PackagesTeam Shosholoza, South Africa’s first America’s Cup

entrant, was one of the smaller teams in this year’s compe-tition. Unlike some of the larger teams, Shosholoza has onlyone boat, so it can’t rely on running two boats against eachother to evaluate design changes. Therefore, CFD simula-tion is critical to the team, which has built a 42-node clusterthat places it near the top in terms of computing capabilitiesamong the smaller entrants. Shosholoza used computer-aided design (CAD) tools to develop a parametric model ofthe boat and then read the model into the ANSYS ICEMCFD Hexa meshing tool, which quickly generates a seriesof models by varying a key design variable over a definedrange. Shosholoza then solved the models with ANSYSCFX software, and designers used the results for force anddrag to predict the velocity.

“To date, the area where we have made the greatestimprovements is in the shape of the sails,” said ChristosPashias, fluid dynamicist for Shosholoza. “We are trying toget as much power out of the sails as possible because thewinds in Valencia are so light. We set up a parametric modelto automatically generate sail models. This enabled us tohave a quick turnaround and study more shapes. Being anew team, initially we made improvements of between 5 and 10 percent in driving force. A 1 percent improvementin driving force typically increases the speed of the boat by about 0.1 percent. We have tested boats with the new designs and discovered that they actually do providethe performance improvements that ANSYS CFX predicts.Since we made those initial big gains, we have made many other improvements that have provided smaller gains, typically in the area of 1 percent, which is what mostteams are after. Testing already has shown that these predictions are accurate, so we trust them to make more improvements.”

Shosholoza also used FLUENT CFD software to betterunderstand the flow of water around the yacht. The rankingof candidate hull shapes by FLUENT software agreed wellwith experimental results.

Emirates Team New Zealand: Location, Location, LocationETNZ has been focused on improving the ballast bulb

at the bottom of the boat. At about 21 tons, this torpedo-shaped lead component makes up nearly 80 percent of theboat’s mass and provides the craft with the stability to balance a very large sail area. Choosing a bulb shape with alower center of gravity increases the boat’s righting momentand enables the sail to provide a larger driving force. On theother hand, moving to a lower drag force wastes less of theavailable driving force and increases the speed of the yacht.In preparing for the 2003 race, the New Zealand designerswere able to lower the center of gravity substantially withoutany increase in drag. With these major improvements underits belt, the team’s goal for 2007 was to make more subtle

and site-specific changes, such as optimizing the bulbdesign for the expected conditions off Valencia.

“We developed a genetic algorithm that works by defining the geometry of the bulb with control points whosecoordinates and weighting are considered to be genes,”said Nick Holroyd, designer for ETNZ. “Then the populationwas seeded with a range of candidates, and mutationswere introduced into each generation to adequately spreadthe population across the design space. Each candidatewas simulated with ANSYS CFX software using the laminar-to-turbulent transition model to provide a drag value. This value is factored against the stability contribution of the shape to provide a fitness score for the design. Wedeveloped a family of new bulb shapes with a better

Simulations were conducted under a wide variety of conditions to determine performance. Velocity magnitude contours around the hull and sails of the BMW ORACLE Racing boat are shown (windward above and leeward below)with plane cuts that are perpendicular to the boat track.


SPORTS

drag/stability trade-off for the racing conditions expected atValencia. This approach made it possible to evaluate thedesign space with much less time than would have beenrequired manually.”

Alinghi: Defending Its HonorWinner and defending champion Alinghi used CFD to

evaluate every portion of the boat, including the sails, theunderwater portion of the hull and deck details. Alinghidesigners spent more than a year evaluating CFD resultscompared to wind tunnel testing and scientific papers. “We gained confidence in the ANSYS CFX software and

calibrated its results,” said Jim Bungener, CFD engineer forAlinghi. “The main areas where we have made performanceimprovements have been in the winglets on the ballastbulbs and the downwind sails or spinnakers. We also havemade smaller gains in areas such as winch placements andpillar shapes. These improvements have significantlyincreased the speed of the boat. When considered as awhole, the results that we have achieved with CFD aided usconsiderably in defending the America’s Cup.” Bungeneralso used ANSYS Structural software to identify the com-posite laminar structure that withstands the loads on thehull while minimizing weight.

Steady Wins the RaceComputer simulation has played a crucial role in the

boat design process for many of the top racing syndicates.With all entrants now using CFD to optimize the performance of their boats, different design groups havearrived at generally the same conclusions and made substantial performance improvements. As a result, theboats are closer together in terms of performance, makingtiny improvements that much more important. The teams now are all creating finer and finer meshes using larger clusters of computers so they can evaluate the effects of smaller design changes on yacht performance. The America’s Cup is thus becoming a showcase, not only for the world’s fastest yachts but also for its most powerful simulation tools.

This article was written through contributions from Alinghi, BMWORACLE Racing, Emirates Team New Zealand and Team Shosholoza.

BMW ORACLE Racing has analyzed and improved nearly every detail of the boat,including the keel–ballast bulb juncture.

Team Shosholoza

Alinghi simulation of typical downwind sail geometry illustrates the way air flowsover the sails. A large vortex is created behind the spinnaker, a billowing sail usedwhen the wind is behind the boat.


MATERIALS/PARTNERSHIPS

Plying the Composite TradeCoupled with technology from ANSYS, Inc., ESAComp software overcomes challenges in designing with composites,enabling engineers to evaluate part designs and better use these versatile materials to their full advantage.

Carbon-fiber reinforced plasticsand other composite materials areused in a wide range of applicationsbecause of their high strength-to-weight ratios. High-performancecomposites made of continuous fibersbound with thermoset resins can beused in making extremely efficientstructures, and laminated compositesare well suited for lightweight partswith complex surface contours.

Composites present some com-plex challenges in utilizing thesematerials to their best advantage,however. Material properties areanisotropic — that is, they are direc-tionally dependent on the orientationof the reinforcing fibers. Differences inthermal expansion of the matrix andreinforcing materials cause residualstresses, and asymmetric structuresespecially can yield unexpectedresponses to temperature variations.Moreover, sandwich structures exhibitcomplex behavior because of large differences in strength and stiffnessbetween layers.

When designing with laminatedcomposites, engineers must take intoaccount these and many other con-siderations in establishing importantdesign variables, including selection of material types, layer orientation and thickness, number of layers, andstacking sequence. Compounding thedifficulty, complete material propertydata cannot always be found from thesupplier data sheets.

Composite Analysis and DesignHigh-performance composites are

used extensively in the aerospaceindustry, where engineers rely on in-house tools developed specificallyfor composite analysis. These pro-grams require considerable resourcesto develop and maintain, however.Engineers need extensive training tounderstand the specialized command-based interfaces and numericaloutputs. In addition, users often haveto transfer data manually between multiple programs for modeling andanalyzing components.

Concerned about the inefficiencyand lack of consistency between thewide range of in-house codes used inthe aerospace industry, the EuropeanSpace Agency (ESA) initiated a projectin the early 1990s to standardize theanalysis approach with a single soft-ware platform combining various toolsunder a unified user interface. ESA,with headquarters in France and consisting of 17 member states, is incharge of shaping the development of Europe’s space capability andensuring that investment in space con-tinues to deliver benefits to the citizensof Europe. By coordinating memberresources, the agency can undertakeprograms and activities far beyond thescope of any single European country.ESA also works closely with spaceorganizations outside Europe.

Development work for the com-posite project was conducted byHelsinki University of Technology inFinland, and the first version ofESAComp software was released in1998. Development responsibility later

By Harri KatajistoComponeering Inc.Helsinki, Finland

Dialog boxes in the ESAComp plyspecification tool enable users toreadily enter various input data.



ANSYS, Inc. software and ESAComp were instrumental in the designof the 75-foot Wally-class racing yacht.

Image courtesy Johannes Schlieben, University of Applied Sciences ofNorthwestern Switzerland.

The esaplot viewing utility gives quick insight into the overall safety performance of the yachtdesign. For each composite element, a safety factor determined by the most critical layer of thesection is given. Options provide detailed data on the laminate failure mode and the most criticallayer, with this information overlaid on each element. Here, for example, cs indicates core shearfailure, and w(n) denotes wrinkling of the face sheet. The second option characterizes the mostcritical layer: the stacking number and the orientation.

Layer charts indicate the effect of layer orientationson interlaminar shear stress distribution in a shortbeam test sample.

can study constitutive relations andhygrothermal behavior of laminates, forexample, and compare laminate lay-ups with respect to strength and otherdesign requirements. Input checkshelp guarantee that analyses are notperformed with inadequate data.

Users can run preliminary designchecks to ensure that columns do notbuckle, plates withstand applied loadswithout deflecting excessively, pres-sure vessels carry specified internalpressures, joint configurations are efficient for load transfer, holes inplates do not cause severe stress concentrations, and scatter in materialproperties does not cause unexpectedproblems. The initial solution obtainedgives a starting point and benchmarkwhen going to finite element analysis(FEA) of the full structure, after whichpost-processing of the results helpssurvey the numerous failure mech-anism possibilities.

The software includes a materialdatabase of fibers and matrix materials, adhesives, sandwich corematerials, and reinforced material systems from commercial suppliers.

A ply specification tool in ESA-Comp is particularly valuable in settingup the input data for various materialconfigurations, such as a cured fiber-matrix system or a honeycombcore material. Since ply behavior istypically between isotropic and fullyanisotropic, the ply specification toolutilizes material symmetry rules to helpin defining the data. Ply data also canbe derived from fiber and matrix datawith micromechanics analyses.

With the laminate lay-up tool, lami-nates can be created and editedefficiently. The user has a wide range ofoptions for performing analyses as wellas selecting and combining result data.For example, several laminates, laminate orientations or failure criteria

was transferred to the spin-off Finland-based company Componeering Inc.,which now distributes and supports the software. Although the softwareoriginated in the aerospace industry, ithas been developed as a general toolfor engineers in other applicationsdesigning with high-performance com-posites, including automotive, marine,construction, machinery, rail trans-portation, sports and wind energy.

The software has analysis anddesign capabilities for solid–sandwichlaminates and micromechanical analyses. It further includes analysistools for structural elements: plates,stiffened panels, beams and columns, and bonded and mechanical joints.ESAComp focuses on the conceptualand preliminary design of compositestructures as well as detailed productevaluation using ANSYS Mechanicaland other analysis software. Engineers



can be selected for different types of analyses. The results displayoptions include numeric tables, lineand bar charts, failure envelopes, andcontour plots.

Integration with Software from ANSYS, Inc.ESAComp is fully integrated with

ANSYS Mechanical software. ESA-Comp FE export supports ANSYSpre-processing. Also, the program canbe launched from the ANSYS interfaceto perform detailed stress analysis andpost-processing. ANSYS Mechanicalsoftware allows defining FE modelinput files in text format using specificcommands, which is, in many cases,the best way to set up models; theESAComp FE export capability fits inthis scheme. The ANSYS Workbenchplatform supports these text formatlaminate definitions as well. For eachpart in the model tree, the user cangive ANSYS commands throughANSYS Workbench command objects.Laminates can be defined withESAComp FE export, and the defini-tions override the default materialdefinitions.

Currently, the best way to simulatecomplex composites structures is toimport computer-aided design (CAD)geometry in ANSYS Workbench assurface bodies and use enhancedcontact features, automated meshingand environment commands. Then,open the simulation model in ANSYSMechanical software and read in all

laminate definitions from an ESACompFE export file. When the geometry is imported as surfaces, ANSYS Workbench automatically uses shell181 elements. After elements havebeen updated to correspond to thecorrect laminate definitions, the modelis solved and post-processed.

Integration of ESAComp post-processing with ANSYS has been realized with the versatile ANSYSParametric Design Language (APDL)and is used through two commands:esapost and esaplot. The most rele-vant data can be combined in a singleANSYS contour plot showing safetymargins for the most critical failuremode, including layer failure, inter-laminar shear, or sandwich core shearand wrinkling. Text labels on elementsprovide additional information on the failure modes or critical layers.Through this procedure, the userquickly identifies design-driving areas,since all relevant failure modes areconsidered automatically and clearlydisplayed.

ANSYS, Inc. technology and ESA-Comp are complementary tools used routinely in developing products madeof high-performance composites. Thetechnologies were instrumental in thedesign of the 75-foot Wally-class racing yacht, for example, which features a unique canting keel for balancing the moments produced bythe sails. In this application, the weightof the boat was a dominant design

driver; simulation tools were critical in optimizing the lay-ups and certifyingthe laminates of heavily loaded compo-nents, such as the chain plates and thejunction between the keel box floor andthe hull.

Another application involved adesign project for a truck with a liquid-hauling tank made of filament-woundcomposites and sandwich structures.Advanced contact features and automatic meshing capabilities in theANSYS Workbench environment wereused to transform the CAD geometry ofthe tank support structure to the FEmodel. ANSYS parametric modelingfeatures and interfacing capabilitieswith ESAComp were further used tooptimize the design. Finally, ANSYSMechanical software was used for validating the design against the cer-tification authority’s requirements.Processing indicated how interlaminarshear (ILS) strength of the laminatestructure is a dominant design factor inthe discontinuity location while thetruck is decelerating.

The ANSYS ESAComp post-processing utility indicates to designersthe weakest point of the structure, the weakest ply in that location and the most likely mechanism of ply failure. This information gives usersvaluable insight for making informeddecisions on refining the design of the structure.

A liquid-hauling tank and associated structures of the truck were analyzed. The truck’stank was made of filament-wound composites and sandwich structures.

Analysis determined criticality of the interlaminar shearstrength for the trailer tank structure during deceleration.Margin to safety is indicated with contours, and results ofthe failure analysis are overlaid on the elements.


CONSUMER PRODUCTS

In the consumer-driven world ofcosmetics, consumer experience andexpectations are anything but an exactscience. Qualitative performance testing, to gather information such as“Does this product increase hair’sshine?” or “Does this product spreadthrough the hair well?” usually isachieved through subjective testing.As an alternative to such testing, product developers and researcherscan use computational fluid dynamics(CFD) coupled with appropriate surface science and emulsion decom-position mechanisms for virtual testingof hair care products.

Hair Today

By Aniruddha Mukhopadhyay, ANSYS, Inc.

Product developers in the cosmeticsindustry can put simulation to use in performing hierarchical analyses of haircare product performance.

In order to mimic the subjectivetest procedure, a standard (baseline)hair with the standard (baseline) product can be simulated at the outsetto establish quantitative correlationsbetween subjective characteristics andchemical or fluid properties. Examplesinclude the measurement of tackiness

associated with surface tension,greasiness with viscosity and resultingglossiness with optical reflectance. Assimulation progresses and correlationsare developed, product developersalso need to understand how and whatto model on various scales.

Consistency in simulation is onlyas reliable as the details of physics and chemistry in the models. Within a predefined scope, simulation providescontrolled test conditions. For exam-ple, a simulation-driven test procedurecould be set up to begin with a knowntest subject, possibly developed withina “hair library” in the simulation soft-ware, of specified morphology, age,pore size, moisture absorption proper-ties, temperature, and grease in and onthe hair. The researcher then coulddefine the environment around the testmaterial (a sample hair assembly ortress) and apply the product makingvarious assumptions, such as thechoice to define application such that it yields approximately a uniform layeron the head. More detailed optionsinclude an applicator or a fingertip forstudying the spreading and coating.

Varying size and scope of the CFD model can provide insight forbehaviors that are best observed onvarious scales. Product applicationand spreading can be accurately modeled on a relatively large “head-scale,” while functions such as glosserbinding, which actually occur at thehair surfaces, are best modeled at amuch smaller “hair-scale.” An effective

Simulation of water flow rinsing process using head-scale modeling

Simulation of shampoo concentration contoursusing head-scale modeling

Simulation of water velocity contours using head-scale modeling

Illustration of subjective test results representation fortwo hair care systems


CONSUMER PRODUCTS

overall modeling approach involvescoupling external flow with micro-phenomena near the hair surface. Withthis method, based on the large-scaleflow conditions, the model is used toextract useful hydrodynamics datadown to microscopic fluid volumesnear a single hair and locally evaluateperformance of various agents. Thiswould enable gathering detailed information about the effectiveness offactors such as grease removal rate or product decomposition, which is relatively difficult, if not impossible, toconsistently observe through tress-based tests.

Hair care products usually arepackaged as emulsions, multi-liquiddispersions with suspended ingredi-ents that don’t segregate while stored.They are designed to dilute and breakdown when applied to the head witheither fingers or a stream of shower.Variations in properties such as density, rheology, surface energy,

Spread of a complex oil droplet over a pair of cross-hairs: on the left, initial state in which ingredients suspended in the product’s emulsion are represented as sub-droplets in thelarger drop; on the right, the state after spreading has occurred

Dynamic simulation of long hairs in a liquid stream

chemical potential, temperature andphase-equilibrium of different immis-cible and dissolving ingredients posethe design challenges to productdevelopers. To understand the productbreakdown process that occurs duringapplication, a hair-scale simulation is required.

To examine the emulsion decom-position process, a complex, multi-phase simulation is performed. A dropof the specified product is deposited ata location at which two hairs cross.The drop being modeled is about threetimes the hair diameter and includessuspended sub-droplets intended torepresent the elemental ingredients inthe emulsion. The simulation demon-strates the capillary effects of the cross-hair assembly and provides theproduct designer with information onthe state of decomposition and spreadof the product that will occur on suchcross-hair configurations. Due to avariety of governing physics and the

dissolution kinetics, pretreated hairs aswell as conditioner ingredients greatlyaffect surface forces on the productdrop that is being decomposed andspread.

The embedded constituents canbe further defined to have their ownspecific material properties. For exam-ple, they could be defined as wettable,which means they stick to the hair,thereby serving as active depositionsites for various ingredients. In a case involving ingredients that areresponsible for “hold” qualities, thesub-droplets could be defined as polymers that will undergo glass tran-sition, leading to a firmer film at roomtemperature. This film structure willprovide added elastic strength for thehair strand and evolve as a hold quality. One complexity for these filmsis that they will neither be exactlyhomogeneous in content nor haveisotropic properties for factors such aselasticity, smoothness or thickness.

It is possible to set up a range ofsimulations for different starting compositions, sizes, temperatures andenvironmental dilutions and then toobserve the final state for each distinctmodel. Although each simulation will predict a single resulting state, as though the product is in fact homogeneous and isotropic, a heuristiccompilation of multiple simulations can together provide a more realisticstatistical representation and charac-terization of the relative performancesof various formulations.


AUTOMOTIVE

In developing hauler trucks, everyextra pound of vehicle weight increasesmanufacturing costs, lowers fuel efficiency and reduces vehicle payloadcapacity. So Axis Developments Ltd.had an idea for making one of thelargest parts of the vehicle out of light-weight aluminum: the tipper body bedthat tilts back to unload soil, rock,debris or other contents.

A designer and manufacturer of trailers and truck bodies for the world-wide highway transportation andconstruction industries, South Africa–based Axis is known for its Alutip seriesof aluminum tipper beds, which weighconsiderably less than comparablesteel bodies. Developing these struc-tures is an engineering challenge,however, since body strength must bemaintained with aluminum material,which has different properties thansteel; the amount of material must beminimized as much as possible for further reduction in weight and cost;

and there is the additional desire to get new designs released quickly without numerous physical prototypetesting cycles.

In redesigning an existing tipperbody having a capacity of 15 cubicmeters, Axis addressed these issuesupfront in the design cycle with softwarefrom ANSYS, Inc. by readily evaluatingstress levels for different configurations.Geometry of the existing design wasimported from Autodesk Inventor, acomputer-aided design (CAD) package,into ANSYS DesignModeler software,which has functions for preparing designmodels specifically for simulation. Theengineering team used a mid-surfaceextraction tool in ANSYS DesignModelerto convert the solid model of the tipperbody’s 10-mm-thick plates to a simplersurface representation. This simplifi-cation enabled the software to model thestructure with a minimal number of shellelements for greater solution speedwhile still retaining information on plate

Heavy-Duty LightweightAn innovative aluminum design gives a truck-body manufacturer thecompetitive edge in the worldwide construction industry.

By Mauritz CoetzeeAxis Developments Ltd.Pretoria, South Africa

This Volvo hauler truck is equipped with a lightweight aluminum Alutip tipperbed that tilts back to unload vehicle contents. The unique semicircular bed isdesigned by Axis Developments Ltd.


AUTOMOTIVE

www.ansys.com 13

thickness throughout the simulation.Additionally, the model was parameter-ized so engineers could quickly modifythe geometry of the model by changing only a few key parameters,instead of having to rebuild the entiremodel from scratch.

Next, design geometry passed from ANSYS DesignModeler to ANSYS Professional software for structuralanalysis. Since the two modules bothoperate on the ANSYS Workbench platform, transfer of data occurred with a menu pick, allowing switchingbetween design and analysis withouthaving to open and close differentapplications. In this way, Axis Develop-ments’ engineers quickly developed amesh and performed stress analysis inANSYS Professional software; theythen were able to appropriately modifythe geometry in ANSYS DesignModelerand immediately perform another analysis to ensure that stress concen-trations were eliminated.

Using this approach, engineersquickly arrived at an optimal design by performing three iterations, with a total solution time of only five minutes per iteration. By experimenting with different types of designs, Axis Developments determined that the

traditional support beam configurationcould be replaced with a more effec-tive semicircular design having areinforced rib structure and end platefor additional stiffness. As the onlymanufacturer employing this uniquedesign shape, the Axis bodies are easily recognizable on the road andquickly are becoming the company’strademark. Body weight was reduced25 percent yet provided the additional

Using aluminum materials saves weight but presents a new set of challenges. The first step in redesigning the tipperbody was importing existing geometry into ANSYS DesignModeler software. Axis Developments’ engineers modeledthe truck body with shell elements and parameterized so models could be readily modified by changing a few keyparameters, instead of rebuilding the entire model from scratch.

strength needed for higher payloadcapacities — a benefit for customersand, thus, a definite competitiveadvantage for Alutip in the hauler truckmarket. The material cost savings paidfor the company’s software investmentwithin only 10 truck bodies.

Axis Developments’ engineers had no previous experience with finiteelement analysis, yet they were pro-ductive after only two hours of training.The tipper body design was completedin less than two days, which wouldhave been unfeasible using conven-tional hand calculations. Moreover, the design was refined with fewerhardware prototypes. Reduction inprototype testing was a huge benefit,since these large structures areextremely time-consuming and expen-sive to build and test. With the successof this tipper body redesign, AxisDevelopments now uses a simulation-based product development approachin which all new design concepts areevaluated, “what-if” scenarios arestudied, problems are fixed anddesigns are refined before detailedCAD work is started.

The authors would like to acknowledge theefforts of SolidCad (www.solidcad.co.za), aSouth Africa reseller of ANSYS, Inc. productsthat provides software and training.

Stress distribution contours (from ANSYS Professional software) plotted on the Axis tipper body structure


POWER GENERATION

The technology of coal gasificationhas existed since the early 19th century.Prior to the discovery of natural gas,coal was used to produce so-called“town gas” for lighting and heat in citiesacross the United States and Europe.Specifically, the gasification process isused to convert any carbon-containingmaterial into a synthesis gas, or syngas.Syngas contains mostly carbonmonoxide (CO), carbon dioxide (CO2)and hydrogen (H2) and can be used as afuel to generate electricity or as a basicchemical building block for a large num-ber of applications in the petrochemicaland refining industries. Gasificationthus adds value to low-rank coal feedstocks by converting them intomarketable fuels and products. Due tomore recent technological advances,gasification offers one of the most effi-cient and cleanest ways to convert theenergy content of coal into electricity,hydrogen, methanol and other usableforms.

Based on the mode of conveyanceof the coal and the gasifying medium,gasifiers can be classified into fixed- or moving-bed, fluidized-bed, andentrained-flow reactors. Entrained-flowgasifiers are normally dilute-flow withsmall particle sizes and have been suc-cessfully modeled with computational

fluid dynamics (CFD) using theEuler–Lagrange, or discrete phase,model approach [1]. For fluidized-bedgasifiers however, Eulerian–Eulerian (E-E), or two-fluid multiphase, model isthe most appropriate approach. The E-E model treats the solid phase as a distinct interpenetrating granular“fluid” and is the most general-purpose multi-fluid model.

Transport gasifiers are based oncirculating fluidized–bed (CFB) reactortechnology and have the ability toachieve higher throughput, better mixing, and increased heat and mass transfer rates compared to other conventional technologies. CFB reactors have been an establishedtechnology in the chemical and power

Gassing Up with CoalA two-fluid multiphase model allows for more accurate simulation of coal gasification.By Christopher Guenther, U.S. Department ofEnergy, National Energy Technology LaboratoryWest Virginia, U.S.A., and Shaoping Shi andStefano Orsino, ANSYS, Inc.

generation industries for years. However,new reactor designs to improve per-formance, reliability and safety havebeen slow to emerge due primarily to thelack of understanding of the complexhydrodynamics of the gas and solidphases.

The idea of describing fluidized bedsand CFBs with two-fluid hydrodynamicmodels has existed since the early1960s [2]. Even with today’s powerfulcomputers, numerical solutions of large-scale CFBs are rarely found in theliterature, and even fewer that consider3-D solutions [3]. Fortunately, the E-Emodeling approach is one that can helpresearchers understand the complexinteractions between the gas and solidphases and aid engineers in the designof new reactors. This approach can pro-vide detailed 3-D transient informationinside the reactor that otherwise couldnot be obtained through experimentsdue to the large scale, high pressuresand high temperatures involved.

To gain more insight into the processphenomena, ANSYS teamed with theU.S. Department of Energy’s NationalEnergy Technology Laboratory (NETL) todevelop different CFD models for simu-lating coal gasification applications.

PSDF gasifier schematics (left) and an exploded view of the mixing zone (right) colored by contours of CO fraction

Visualizations of the flow in the mixing zone of the PSDF gasifier for a case with air-blown and steam-enhanced lignitefuel. Included are flow pathlines colored by CO fraction (left); velocity vectors on isosurfaces of solid fraction of 0.2 and0.3, in which the formation of particle clusters can be seen (center); and contours of carbon reaction rate (right).

Riser

Mixing zone

Coal

Coal

Air

SorbentAir

Steam

Air/O2

steam Air/steam/O2

Recycling solids

Standpipe


POWER GENERATION

www.ansys.com 15

Their objective was to illustrate how CFDcan be used for complex large-scalegeometry with detailed physics andchemistry. Using FLUENT software, theteam developed a 3-D transient model ofKBR, Inc.’s Power Systems DevelopmentFacility (PSDF) transport gasifier. KBR is aglobal engineering, construction andservices company that has partneredwith other companies to build a com-mercial transport gasification unit, basedon the technology developed from thePSDF, at a 285-MW power generationfacility in Florida that promises to be thecleanest coal-fueled plant in the world.

In the FLUENT simulation of thePSDF, 11 species were included in thegas phase while four species wereassumed to be in the solid phase. A total of 16 reactions, both homo-geneous (involving only gas phasespecies) and heterogeneous (involvingspecies in both gas and solid phases),were used to model the coal gasifi-cation chemistry. The gas combustionreactions were simulated with a finite-rate combustion model. The coalreactions, including moisture releasing,devolatilization, char combustion, chargasification, tar cracking and water–gas shift reactions, were modeled witha heterogeneous reaction scheme anda set of user-defined functions. Thegeometry was meshed with 70,000cells, and each simulation case wasrun in parallel on an eight-processormachine. Post-processing the datawas done once the solution reached apseudo-steady state, which requiredrunning the simulation until it gener-ated physical data representing about40 seconds of time.

Time-averaged temperature distribution along the PSDF center line as compared to experiment

Outlet gas composition for the PSDF transport gasifieras compared to experiment

outlet

inlet

Fluctuations of the mass flux (including both solid and gas) at the gasifier outlet. The negativevalue represents the outgoing flow at the outlet. The magnitude of these fluctuations can deviateby as much as 70 percent around the mean of –47.12 kg/s.

The basic design of the PSDF transport gasifier included a mixing zone,which kept the recycling solids presentlong enough for the carbon left in the particles to react with the incoming gas (O2, steam or CO2). Visualizations of the system interior showed that the flow wasrecirculating and mixing in the mixing zonebefore it moved up into the riser section,and also that local conditions were verychaotic and turbulent. At the bottom of themixing zone, combustion of the carbonpresent in the recycle material depletedthe available O2. Further combustionoccurred as the solids moved up higherinto the mixing zone. At the same time,other reactions such as CO and H2

combustion were competing for the O2.These exothermic reactions generated thenecessary heat for the endothermic reactions, including steam gasification andCO2 gasification of carbon.

The research team validated the over-all computational results against PSDFexperimental data for both bituminous and sub-bituminous coals under both air-blown and oxygen-blown conditions.The computational difference between themass flux at the inlet and average massflux at the outlet was only 0.1 percent,which meant that the mass was balancedwell from the simulation standpoint. Theteam drew the same conclusion for theheat balance. For the temperature profile,the difference between the simulation andmeasurement was due mainly to the location of the probes relative to the centerline. Despite the finding of very uneventemperature distributions at any givencross section, the overall trends of the temperature profiles were in goodagreement with the measured data.

References

[1] Shi, S.; Zitney, S.; Shahnam, M.; Syamlal, M.;Rogers, W., Modeling Coal Gasification with CFDand the Discrete Phase Method, 4th InternationalConference on Computational Heat and MassTransfer, May 2005, Paris.

[2] Davidson, J., Symposium on Fluidization —Discussion, Trans. Inst. Chem. Eng., 1961, 39,pp. 230-232.

[3] Guenther, C.; Syamlal, M.; Shadle, L.; Ludlow, C., A Numerical Investigation of anIndustrial Scale Gas–Solids CFB, CirculatingFluidized Bed Technology VII; Grace, J.; Zhu, J.;de Lasa, H., Eds.; CSCHE, Ottawa, 2002, pp.483-488.

20

18

16

14

12

10

8

6

4

2

0

-29

-34

-39

-44

-49

-54

-59

-64

-69

-74

-79

CO H2 CH4 CO2 H2O

1300 1350 1400 1450 1500 1550 1600

260 270 280 290 300

Model

Exp.

%

Heig

htM

ass

flux

at o

utle

t (kg

/s)

Temperature (F)

t (s)

Model

Exp.

Chopper pumps utilize achopping action betweenthe impeller and the suctionplate to break down solidsthat pass through the pumpinto smaller pieces. VaughanCompany, an establishedpump manufacturer inWashington, U.S.A., designsand manufactures a line ofcentrifugal chopper pumps.These pumps originally weredesigned in the 1960s for usein the local dairy industry totransport manure to and fromstorage tanks. Since then,

Vaughan chopper pumps have been refined continually andawarded a number of patents; the company has earned wideacceptance for many applications that require solids handling. Today, Vaughan chopper pumps are used in various phases of municipal and industrial sewage treatment,food processing, and pulp and paper industries, in which the

By Glenn Dorsch and Kent KeeranVaughan Company Inc., Washington, U.S.A.

pumped liquid contains solids that need to pass through thepump without clogging or plugging.

The benefit of a Vaughan chopper pump over a typicalnon-clog or slurry pump is that it reduces the solids size ofmaterial passing through the pump. The unique choppingrequirements and suction arrangement of these pumpsmake it difficult to apply standard impeller design practicesin order to evaluate hydraulic performance. As energy costscontinue to rise, developing more efficient pumps becomesincreasingly critical for all pump manufacturers. VaughanCompany found that simulation was an effective and efficient way to approach the optimization of pump design.

Vaughan Company’s simulation process begins byimporting computer-aided design (CAD) models fromPro/ENGINEER® into ANSYS DesignModeler software. Theimpeller domain and casing domain are meshed separatelyand assembled within the CFX pre-processor in whichboundary conditions are applied. The ANSYS CFX solverperforms the required calculations; then, results are viewed and pump performance is calculated in the compu-tational fluid dynamics (CFD) post-processor. The ANSYS Workbench platform facilitates the entire simulation process,from geometry import through visualization.

Chopping Away at SolidsCFD simulation provides a pump company with a virtual test facility.

In a Vaughan chopper pump, the mainimpeller vanes extend all the way to the center hub of the impeller, and the suctionplate includes two stationary fingers that protrude to the center of the suction opening. As the main vanes pass by the stationary fingers, a chopping action results, which macerates any solids entering the pump.

Performance curve for a recently redesigned 6-inch pump. The simulation slightly underpredicts TDH because the geometry for the impeller and casing had to bereverse-engineered, and there were likely some differences between the model and the actual parts.

Comparison between the simulated existing impeller and the simulated redesignedimpeller ensured that the redesigned impeller had TDH characteristics that were asgood as the original impeller. The new design achieved an approximately 8-pointincrease in efficiency over most of the flow range.

Geometry of a typical casing, impeller and cutter bar assembly

1100 1300 1500 1700 1900 2100 2300 1100 1300 1500 1700 1900 2100 2300

GPM GPM

130

120

110

100

90

80

70

60

50

130

120

110

100

90

80

70

60

50

ACTUAL 6U VERSUS SIMULATED 6U @1750 RPM; 11.8'' DIA IMPELLER SIMULATED 6U VERSUS SIMULATED 6U @1750 RPM; 11.8'' DIA IMPELLER

TDH

(FEE

T),P

OWER

(HP)

,EFF

ICIE

NCY

(%)

TDH

(FEE

T),P

OWER

(HP)

,EFF

ICIE

NCY

(%)


PROCESS EQUIPMENT

A D V A N T A G E

Spotlight on Engineering Simulation in the

Biomedical Industrys10 Standing Up Right

s12 Designing with Heart

s14 Going with the Flow

s15 Battle of the Bulge

s2 Making Life Longer and Better

s4 Turning Up the Volume

s6 Hip to Simulation

s7 Walking Pain Free

s8 Engineering Solutions for Infection Control

SWEET SOUNDS FROM SIMULATIONCOCHLEAR IMPLANTS

www.ansys.comANSYS Advantage • Volume I, Issue 2, 2007s2

BIOMEDICAL: OVERVIEW

Recent analyses show that leading biomedical com-panies around the world are continuously growing theirinvestment into research and development (R&D), with anincrease of 12.5 percent in 2006 that reached total R&Dexpenses exceeding $9 billion [1]. This is no surprise, giventhe need for advanced medical treatments and care due to alarge and growing population of aging individuals, the needto find minimally invasive treatments for conditions such asdiabetes and heart disease, and the increasing demand forartificial organs. As medical product innovation continues tobecome more complex, there is a strong emerging need forSimulation Driven Product Development, which has beenseen and is broadly accepted in the semiconductor, aerospace and automotive industries.

Simulation is becoming an integral part of the productdesign cycle in biomedical applications ranging from prosthetics and artificial organs to endovascular techniquesto surgical devices, medical equipment and diagnostic

Simulation DrivenProduct Development:Making Life Longer and Better

By Thierry Marchal and Kumar Dhanasekharan, ANSYS, Inc.

products. There are a number of reasons for such simulationto continue its entrenchment in biomedical product develop-ment. First, the advancement in technologies such ashigh-performance computing (HPC) is able to meet thedemands of biomedical product development, allowinghealthcare institutions, life science researchers and theindustry to conduct large-scale simulation studies. Theincreasing ability to import computed tomography (CT)scans and magnetic resonance imaging (MRI) into simulationsoftware — a process now becoming routine — makes itfeasible to address in vivo device design needs (such as withrespiratory drug delivery and endovascular devices), essen-tially enabling virtual prototyping. In addition, the integrationof simulation techniques across multiphysics, from structuralanalysis to flow modeling to thermal analysis, is enhancingthe virtual prototyping needs of the biomedical industry. Forexample, in studying aneurysms, ANSYS simulation toolshave been used to import CT scans into the simulation

The biomedical industry is emerging as astrategic user of engineering simulation.

Simulation Driven Product Development is being applied regularly in the biomedical industry. This aneurysm study was performed within an integrated environment to analyze coupled fluid flow and structural simulation. The steps are: 1) CT scan; 2) segmentation from scans to extract branches; 3) cuts are written in form of splines; 4) creation of solid geometry composed of arterial wall/thrombus and automatic creation of fluid volume from the solid geometry; 5) independent mesh for each simulationtechnique (flow modeling and structural modeling); and 6) coupled fluid and structural model with model setup, analysis and post-processing in a single environment.

1 2 3

Arterial wall Thrombus

456

ANSYS Advantage • Volume I, Issue 2, 2007www.ansys.com s3

BIOMEDICAL: OVERVIEW

environment, allowing researchers to study a structural analysis of the weakened arteries along with the flow patternsin a single virtual environment, truly creating a virtual proto-type model with multiphysics, all in an integrated manner.

Another growing area is drug delivery, particularly withmedicines that are released into the bloodstream or respi-ratory system. There is a need to better understand theprocess and how adjustments can be made to acceleratedrug delivery to the point of highest efficacy, which then willallow healthcare companies to design better devices thatadminister appropriate dosages.

Similarly, orthopedic departments are paying moreattention to the virtual prototyping approach brought bycomputed-aided engineering (CAE). Bones are criticalpieces of the body, having complex, specific geometries;they are made of different materials exhibiting strongly nonlinear behavior. Until now, scientists have lacked proper,robust models that can be used to bring together, into a single simulation, characteristics as complex as poro-elasticity, nonlinear viscoelasticity and linear elasticity, whichare needed for an accurate description of an intervertebraldisc (ID), for example. The improved robustness of existingmodels together with the availability of reliable material properties now provides evidence that these numericalresults can bring new, invaluable information to doctors. As aresult, healthcare institutions now are studying how a hipprosthesis will perform related to a comfortable walk over along period of time as well as investigating — prior to plan-ning spinal surgery or even designing an ID implant —whether the remodeling procedure leading to the unificationof the pedicle screw and the vertebra is likely to progress smoothly. [See Standing Up Right on page s10.]

To illustrate recent concrete progress in addressing real-life problems and pain relief via CAE, this biomedicalspotlight describes applications in which simulation tech-nology has made a major difference. Both fluid flows andsolid mechanics, or the combination of the two, appear in surprising applications. Some are critical to patient life orfunction, such as lung air flow and spine implant; others simply make life more comfortable through better earimplants and insole design.

For the future, imagine the impact of simulation to drivethe development of patient-specific medicine and medicalcare. For example, tomorrow’s surgeons may be able to takeCT scans of patient physiology and use simulation to conduct virtual surgery as well as study the procedure’seffectiveness as part of the overall process. This is enabledthrough automation of simulation along with rapid designcomparisons through automated parametric studies — andit is rapidly becoming reality. The era of simulation in the biomedical world is rising.

References

[1] The R&D Scoreboard 2006, Volume 2, Department of Trade andIndustry (DTI), U.K.

Proper design of a medical insole required to develop an accurate modeling of the foot at different stance phases during required ambulation: 1) the initial contact state; 2) the mid-stance state; and 3) the toe-off state. The resulting data was used to calculate thepressure and stress induced on the plantar surface as well as inside deep tissues.

1

2

3


BIOMEDICAL: COCHLEAR IMPLANTS

Cochlear implants (CIs) are elec-tronic hearing devices designed torestore partial hearing to those who aredeaf or severely hearing-impaired. Thedevices consist of three external andtwo internal components. The externaldevice comprises a microphone thatpicks up sounds from the environment,a speech processor and a transmitter.The internal components include twosurgically implanted devices: a receiverthat works with the transmitter to convert speech processor signals intoelectronic impulses and an electrodearray that uses those signals to stimu-late the auditory nerves within the ear.One of the traditional limitations of the electrode array is the inability toachieve optimal depth of insertion intothe cochlea, the auditory portion of theinner ear. A German team including

CADFEM GmbH, the Hannover University of Applied Sciences andArts, and the Leibniz University of Hannover has found that an improve-ment might be possible using shapememory alloys (SMA).

Shape memory materials displaydistinct thermo-mechanical behavior.In the case of shape memory effect(SME), a body that has undergoneplastic deformation will return to theoriginal shape or form that it had priorto deformation by heating it above acritical temperature. After being heatedand returning to its original form, ashape memory material will not changeback to its deformed shape if cooled.This phenomenon can be observed inmany shape memory alloys, specific-ally nickel-titanium (Nitinol), which hasa wide range of applications in theautomotive and aerospace industries.In addition, due to its high biocompat-ibility, high resistance to corrosion and,above all, the thermal-induced SME,Nitinol is very useful in the field of medical engineering.

In the case of the CI, the researchteam thought that by taking advantageof the thermally induced shape memory behavior of Nitinol, greaterimplantation depth for the electrode

array could be achieved. The conceptwas to design an SMA componentwhose shape matched that of thecochlea. Prior to the insertion process,the component would be deformedpseudo-plastically, and then, relying on heating from the body itself, itwould return to its original form duringimplantation. To pursue this idea,implant simulations that accounted forthe pseudo-plastic deformation andshape memory behavior were carriedout using ANSYS Multiphysics tools.

For these simulations, the team created a material model for SMA andimplemented it in ANSYS Multiphysicsvia user-interface USERMAT for three-dimensional finite elements. The phenomenological material model was developed using stress–strain– temperature data for SMA and wasbased on a linear kinematic hardeningmodel. The stress–strain behavior ofshape memory materials, which is highly nonlinear in nature and varieswith temperature, was incorporated into the simulation with the addition of a temperature-dependent scalar parameter: the middle stress σm.

The shape memory stress–straincurve differs from the standard linearkinematic model in that the shape of

Turning Up the VolumeThe use of shape memory alloys offers the promise ofbetter functioning in cochlear implants.

By Dieter Kardas, Institut für Baumechanik und Numerische Mechanik (IBNM), Leibniz Universität Hannover, GermanyWilhelm Rust, Fachhochschule Hannover, Germany Ansgar Polley, CADFEM GmbH, Burgdorf, GermanyTilman Fabian, Hannover Medical School, Germany

Cochlear implant diagram: implant components (left) and insertion in the cochlea (right)

Image from Hals-Nasen-Ohren-Heilkunde, Boenninghaus, Hans-Georg, Lenarz, Thomas, 2005, Kapitel 5 “Klinik des Innenohres,” p 116. Published by Springer Berlin Heidelberg, ISBN 3-540-21969. With kind permission of Springer Science and Business Media.

Demonstration of one-way shape memory effect, from leftto right: initial shape of a component, deformed shape,shape on warming, shape on cooling after warming

Deform Heat up Cool down


the stress–strain hysteresis — whichone gets by periodically changing forcedirection — is ripped in a manner that varies with temperature. Shapememory alloys exhibit pseudo-plasticity at a low temperature rangeand pseudo-elasticity at a high temper-ature range. These temperature rangesdepend on the percentage compositionof nickel and titanium; generally both are equiatomic, which means thatthe rip of the curves increases withincreasing temperature.

The degree to which the curve isripped is determined by the mentionedmiddle stress, σm. If σm is set to zero,then the hysteresis experiences no rip,and pseudo-plasticity can be repre-sented. If σm is set to a higher valuethan the so-called amplitude stress σy*

(half value of the distance from upperflow curve to lower flow curve), pseu-do-elasticity can be represented. Theactual value of the middle stress wasdetermined using experimental datataken at various temperatures. In orderto obtain a smooth, nondiscontinuousrepresentation of the flow curve, atanh-function was included in theequations that describe the offset/ripbehavior as a function of σm.

By incorporating this offset- function σoff (tensor-function of ordertwo) into the material model, the shapememory behavior was effectively captured with only two sets of materialconstants: one set for pseudo-plasticityand another for pseudo-elasticity.ANSYS Multiphysics software itselfinterpolates between these parametersets to provide the material constantsfor the actual temperature. With this technique, it was possible to reproduceany intermediary state between pseudo-plastic and pseudo-elasticstress–strain behavior.

By including this shape memorybehavior, the CI development team wasable to simulate implantation of a shapememory cochlear implant (SM-CI) intothe cochlea. The results of a 65-secondsimulation of the implantation processsupported the idea that the temperatureof the human body could have enoughof a thermal effect on the array that,when implanted, it could return to theoriginal shape: that of the cochlea.These findings support the possibility of a solution that can provide deeper implantation and, thus, betterfunctionality for the CI.

Time-spaced results of the implant simulation for a shape memory cochlear implant. The red colorindicates that body temperature has been reached by the implant.Cochlear geometry data courtesy Hannover Medical School, Dr. Omid Majdani.

Pseudo-plasticity σm = 0 (left) and pseudo-elasticity σm > σy* (right). The middle stress (σm) rips theshape memory alloy stress–strain hysteresis as temperature increases.

BIOMEDICAL: COCHLEAR IMPLANTS

Temperature

Superposed Cochlea-model(Scale: 20˚C to 37˚C)

Final state superimposed(Scale: 20˚C to 37˚C)

Tim

e:0

Seco

nds

Tim

e:17

Sec

onds

Tim

e:35

Sec

onds

Tim

e:53

Sec

onds

Tim

e:65

Sec

onds


BIOMEDICAL: ARTIFICIAL JOINTS

Hip to SimulationEvaluation of designs for a hip replacement prosthesis overcomes physical and scientific limitations.

By Joel Thakker, Integrated Design and Analysis Consultants, U.K.

Hip replacement surgery involvesreplacing the damaged or diseasedball-and-socket joint configuration with artificial parts. During surgery, acup or hip socket — a dome-shapedshell/liner — is implanted into theacetabulum portion of the pelvic girdleafter the bone has been hollowed outusing a grater. The thigh, or femoral,portion of the hip replacement pros-thesis is composed of aball, which acts like abearing where it fits intothe cup and is attachedto a stem that furtherattaches to the femur.The Duraloc® unce-mented acetabular hipsocket, a replacementcup developed byDePuy Orthopaedics,Inc., in the U.K., uses an interference fitto hold the socket in place in the hipbone. To assist DePuy in the design ofthe Duraloc product, Integrated Designand Analysis Consultants (IDAC) used ANSYS Mechanical software todevelop parametric models that areused to establish both the necessaryimplantation and disassembly forcesfor variations of the replacement joint.

IDAC performed a two-dimensionalanalysis on the cup assembly in order

X-ray of a hip showing a prosthesis, including the socket,ball and stem. Image courtesy DePuy Orthopaedics, Inc.

Contour plot of stresses induced by the inter-ference fit between the prosthesis and the bone;the areas colored in grey illustrate the region of the bone that could be expected to yield during the assembly process.

Three-dimension finite element model meshof bone and prosthesis

Illustration of stress distribution in the hipjoint assembly after the prosthesis has beenpressed into place

to model the force required to removethe socket axially. A three-dimensionalmodel was used to analyze rotationalremoval of the joint, since a two-dimensional case would not representthe behavior fully. The ANSYSMechanical simulation used nonlinearcontact elements in the prosthetic hipsocket and accounted for frictionbetween the cup and bone. In all

analyses, the implantcup was modeled in titanium while the bonewas treated as an aniso-tropic material.

For both analyses,IDAC created parametricmodels in order to evalu-ate different bone andimplant cup geometries,material properties and

boundary conditions. The assemblyconditions involved inserting the cupinto the bone to overcome inter-ference, allowing the frictional effects to hold the cup in place, and subse-quently removing, either axially orrotationally, the cup from the bone toestablish disassembly loads.

This form of modeling allowsDePuy to evaluate different configura-tions of implant design numericallyrather than by physical testing, which

is time-consuming and expensive incomparison. Physical testing is limitedas real bone materials are not highlyavailable. Some synthetic and naturallyoccurring materials can be used, buttheir material properties do not pre-cisely match that of human bonematerials. Numerical modeling allowsDePuy to view detailed stress anddeflection distribution plots and loadversus time history plots that cannotbe created easily from physical tests.Comparisons between the resultsobtained through simulation and thoseobtained from previous testing reveal a close correlation.

As a result of this study, DePuy hasused this type of design evaluation inother orthopedic implant products,including artificial knee joints.

The Duraloc® uncemented acetabularhip socket is made from titanium andhas a porous coated shell.


BIOMEDICAL: BIOMECHANICS

www.ansys.com s7

Walking Pain FreeNew insoles designed with the ANSYS mechanicalsuite relieve pain from foot disease.By Bum Seok Namgung, Dohyung Lim, Chang Soo Chon and Han Sung KimYonsei University, Seoul, Korea

During ambulation (top to bottom), the highestpressure progressively shifts from the plantarregion under the heel bone forward to themetatarsal head bone.

Von Mises stress distributions on the plantarsurface of the foot using the flat (top) andtotal contact insoles (bottom)

The human foot does more thansimply enable mobility. Feet are animportant part of the body because theybear weight, absorb shock and stabilizebody structure, but they usually get littleof our attention. When foot diseaseappears and pressure and stressexceed a given limit, pain occurs —making a person suddenly aware of justhow critical a function the feet provide.For people with diabetes, subject topoor circulation and neuropathy, evenordinary foot problems can get worseand lead to serious complications.

One research project designed to benefit such patients involves developing insoles that will prevent pres-sure sores on the deep tissues inside theplantar surface of the foot. A team at theInstitute of Medical Engineering at Yonsei University in Korea is finding newways to gather information on themechanical response of the foot to vari-ous insole designs. They are utilizingfinite element analysis (FEA) softwarefrom ANSYS, Inc. to design new patient-specific insoles that reduce bothpressure during ambulation and stresswithin the feet, ultimately relieving pain. The team selected the ANSYSmechanical suite because of its reliabilityand flexibility for handling complex andirregular geometries. Furthermore, itsnonlinear, hyper-elastic models andadvanced contact conditions provide arealistic alternative to experimentalapproaches for gait analysis.

Using the ANSYS technology, theresearchers first created a three-dimensional model using computerizedtomography (CT) images obtained fromthe right foot of a subject with hallus valgus, commonly called a bunion. Commercial software, CANTIBio™(CANTIBio, Inc., Korea) and meshingsoftware were used to fine tune the contours of the foot.

Two insoles, one flat (left) and one shaped to contact the entire sole of the foot (right), were compared in this analysisto understand the impact of the geometry on foot pain.

Three geometries representing threeprimary states (initial contact, mid-stanceand toe-off) during ambulation then werecreated. The simulation models incorpo-rated two insole designs: one flat and one contoured to contact the entire bottom of the foot. Each design was analyzed at various values of elastic modulus (0.3 MPa, 1.0 MPa and 1 GPa) inorder to represent a variation in insolefirmness and identify which more effec-tively redistributed von Mises stresses onthe plantar, or bottom, surface of the footduring standing.

During ambulation, ANSYS softwareshowed that high pressures first appearon the plantar surface region overlyingthe heel bone for the initial contact state,progresses through the middle of the footfor the mid-stance state, and finally, forthe final toe-off state, is concentrated inthe vicinity of the metatarsal head bone atthe front of the foot. These results are inagreement with those obtained from afoot scan system used in experimentalgait analysis.

The results found that stresses on theplantar surface are significantly lower withthe total contact insole compared withthose of the flat insole; stresses also aredependent on the insole elastic modulus.This confirms that customized design ofan insole for patients with foot diseasemay be necessary, and the solutionshould include biomechanical and clinicalpoints of view.

Hospital Nacional Dos de Mayo in Lima, Peru, was thesite of a TB ward ventilation system redesign.

Extract (low, wall)

Bed 2Bed 1

Supply(ceiling)

Extracts (high, wall)

Bed 2Bed 1

Supply(ceiling)

Supply(ceiling)

BIOMEDICAL: INFECTION CONTROL


Hospital-acquired infection poses amajor problem in healthcare facilitiesaround the world. Although many infections are transmitted through hand-to-hand contact, airborne transmissionalso may play an important role; this isthe primary mechanism for a number ofinfections, including tuberculosis (TB)and influenza. Airborne routes also havebeen implicated in the transmission ofhospital-acquired infections such asmethicillin-resistant Staphylococcusaureus, Acinetobacter spp and noro-virus. Successful control of infectioninvolves breaking the chain of trans-mission. To do so, it is necessary to understand both the mode of trans-mission as well as the nature of thepathogen and its behavior in the environment.

The role played by airborne transport of pathogens has been the driving force behind the researchcarried out by the Pathogen ControlEngineering Group at the University ofLeeds in the U.K. for the past 10 years.The multi-disciplinary team of engi-neers, mathematical modelers and

Engineering Solutions for Infection ControlSimulation assists in designing a hospital ward to reduce the airbornetransmission of diseases such as tuberculosis and influenza.

By Cath Noakes and Andrew Sleigh University of Leeds, U.K.

Original room layout and ventilation system (top) and proposed new layout (bottom) showing the location of thepartition between the two beds, the additional ventilationsupply diffuser and the modified extract locations

microbiologists is based in the Schoolof Civil Engineering, with strong links to clinicians at the Leeds Teaching Hospitals and to academics and scientists around the world. Originallyset up to investigate ultraviolet (UV) airdisinfection devices to combat TB, thegroup now focuses on understandingairborne transmission routes with astrong emphasis on the hospital environment. This knowledge is usedto aid the development of new infection control technologies and tooptimize engineering strategies toreduce the risk of disease.

The suitability of a ward ventilationsystem design was the subject of arecent study carried out using ANSYSCFX computational fluid dynamics(CFD) software [3]. The two-bed ward inHospital Nacional Dos de Mayo, located in Lima, Peru, is one of a number of similar rooms housingpatients with TB. Unusual to a hospitalin this part of the world, the wards aremechanically ventilated. Any airbornetransmission of TB within the hospitalwill be strongly influenced by theimposed ventilation flow. As part of awider project researching TB trans-mission, led by Dr. Rod Escombe ofImperial College in London, U.K., theCFD study was carried out to examinewhether changes to the ward layout andventilation system could reduce the riskof cross-transmission between patients,staff and visitors in the hospital.

A simplified geometry representedthe key features in the ward, including

the basic furniture, the ventilation supply and extract vents. Isothermalairflow was modeled on an unstruc-tured tetrahedral grid using a standardk–ε turbulence model. Supply airvelocities were defined to ensure aroom ventilation rate of 6 AC/h for allsimulations, and a pressure of –10 Pawas imposed on the extracts to

BIOMEDICAL: INFECTION CONTROL


Contaminant concentration contours, at an elevation of 1.4 mabove the floor originating from patient 1. The figure on the tophas no partition, while the figure on the bottom uses a partitionand ventilation systems local to each patient.

Streamlines originating from patients 1 (red) and 2 (blue) show how a partitioned room with modified ventilation system (bottom) moreefficiently extracts contaminated air than the original room (top) does.

simulate the negative pressure that ismaintained in the real facility. As the study focused on the risks ofcross-infection, it was important toinclude a model to represent therelease of infectious material from TBpatients. To relate the CFD study topublished outbreak data, a scalarinfectious particle production variablewas defined in terms of units of infec-tious dose, known as “quanta.”

To represent a patient’s productionof TB bacteria, a small inlet conditionwas located close to the head of each bed. Scalars, representing the infectious particles produced by eachpatient, were introduced into the roomat a constant rate of 14 quanta/hour in order to represent the typical pro-duction rate of a pulmonary TB patient.

The CFD study made it quick andeasy to compare the impact of a number of proposed modifications tothe ward. The original room layout with

its single ceiling-mounted supply diffuser and wall-mounted extractresulted in significant mixing of TBcontamination throughout the room,demonstrating the high risk of cross-infection between patients. The simpleaddition of a partition between the twobeds yielded an immediate benefit,providing a physical barrier that limitedthe transfer of infection between thetwo areas. As a low-cost intervention,this could prove beneficial in resource-poor countries, although it may not be suitable for naturally ventilated environments. Combining the partitionwith a new ventilation system layout,comprising ceiling supply diffusersabove the foot of each bed with wall-mounted extracts at the head of eachbed, yielded the best results. Despitethe ventilation rate remaining constant,the transfer of infectious materialbetween the two beds was reduced byover 75 percent, representing a

significantly reduced risk of cross-infection between patients. These findings were of immediate benefit tothe architects redesigning the ward,who based the new ventilation systemand ward layout directly on the studyresults.

www.efm.leeds.ac.uk/aerobiology

References

[1] Noakes, C.J.; Sleigh, P.A.; Fletcher, L.A.;Beggs, C.B., Use of CFD Modeling in Optimising the Design of Upper-Room UVGIDisinfection Systems for Ventilated Rooms.Indoor and Built Environment, 2006 15(1),pp. 347-356.

[2] Noakes, C.J.; Fletcher, L.A.; Beggs, C.B.;Sleigh, P.A.; Kerr, K.G., Development of aNumerical Model to Simulate the BiologicalInactivation of Airborne Microorganisms inthe Presence of UV Light. Journal of AerosolScience, 2004, Vol. 35(4), pp. 489-507.

[3] Noakes, C.J.; Sleigh, P.A.; Escombe, A.R.;Beggs, C.B., Use of CFD Analysis inModifying a TB Ward in Lima, Peru. Indoorand Built Environment, 2004, 15(1),pp. 41-47.


BIOMEDICAL: SPINAL DISORDERS

The human spine is a wonder ofengineering work, one that is heavilyused in daily activities. An importantpart of it, the intervertebral disc (IVD), isone of the most sophisticated suspen-sion and shock absorption systemsever found. When disorders arise, backpain quickly can become a nightmare.The National Technical University ofAthens (NTUA) in Greece conducted a study using ANSYS Multiphysics software that revealed some secrets ofhow this precious structure works, aswell as ways to fix it efficiently when itmalfunctions.

Simulating the Intervertebral DiscThe IVD is located between the ver-

tebrae in the spine. In performing dailyactivities, it acts as a cushion andtherefore is exposed to a combinationof compression, bending and torsionstresses. Each disc consists of thenucleus pulposus, a gel-like inner por-tion of the disc; the annulus fibrosus,the outer portion made of about 20lamellae of coarse collagen fibers; andthe two cartilaginous endplates, com-posed of hyaline cartilage, located oneither side of the nucleus and annulus.

The IVD simulation model comprisedfour distinct volumes corresponding tothe disc’s regions: The nucleus wasmodeled as a nonlinear viscoelasticmaterial in a kidney-like cross section;the two cartilaginous vertebral endplateswere considered linear elastic bodies;and the annulus surrounding the nucleuswas simulated as dual laminated shellelements whose outer surfaces were viscoelastic in nature. The study analyzed various scenarios in order todetermine the contribution of each section of the IVD to the viscous char-acter of the entire structure.

The numerical model revealed thatthe maximum stresses appeared in thefibers of the intermediate volumes of theannulus, in the vicinities of the endplates.The nucleus was almost stress-free, asexpected due to its gel-like nature. The NTUA study also investigated thebehavior of the IVD during daily activities;the results found that the reduction ofdisc height related to a person’s 24-hourdaily cycle was in very good agreementwith the respective experimental data byTyrell et al (L3–L4 discs) [1].

Standing Up RightANSYS Multiphysics sheds light on the wonders of thehuman spine and how to fix it.

By Stavros Kourkoulis, Satraki Margarita and Chatzistergos Panagiotis, National Technical University of Athens, Greece

a b c

The numerical model of the intervertebral disc: a) nucleus pulposus, b) annulus fibrosus and c) cartilaginous vertebral endplates

The spine’s intervertebral disc is exposed to a combination of compression, bending and torsionstresses.


BIOMEDICAL: SPINAL DISORDERS

The von Mises stress distribution through the center of the disc horizontally (left) and at the point of minimumvertical cross-sectional area (right)

The distribution of the Mises equivalent stress in a typicalvertebra for a pull-out displacement of 0.02 mm

The two phases of model construction: (left) the screw and surrounding bone implantedinto the vertebra and (right) the regions of the vertebra (yellow: cancellous bone; red:subcortical bone; blue: cortical shell)

Studying the Surgical RemedySpinal stabilization using pedicle

screws and rods (or plates) is one of themost common invasive treatments forspinal disorders and injuries. In this procedure, the surgical team implantsscrews posteriorly into a number of vertebrae and bolts them to a rod orplate. This assembly actively fixes thevertebra in place, with respect to eachother, and thus stabilizes that section ofthe spine. After such a procedure, someserious problems can still exist. Pain inthe IVD adjacent to the fixed vertebraecan occur due to failure of the spinalinstrumentation, from either a fracture instructural elements or a loosening ofthe screws. Experimental and clinicalstudies alone cannot provide a com-plete view of the mechanical behaviorof such complex structures. Numericalsimulations introduce a unique tool forthe thorough and parametric study ofsuch systems.

From the moment a pedicle screwis implanted into the vertebra, the bonebegins to regrow around the screw.This regrowth leads to the eventualcomplete unification of the bone andthe implant, which occurs about twoyears postoperatively. A fundamentalrequirement for the success of thisprocedure is the stability of the screw’sfit into the bone. NTUA used mech-anical simulation to investigate theinfluence of the vertebra structure andscrew specifications — such as depthof implantation, pitch and inclination of the thread — on the value of theforce required to loosen the screw fromthe spine.

The parametric study assumedthat the vertebra consisted of cortical,subcortical and cancellous bone assuggested by measurements of bonemineral density of typical human lumbar vertebrae. The simulations estimated the force required to pro-duce a pull-out displacement of 0.02mm, the stress distribution onto thebone, and the contact pressure on thebone–screw interface. The results indi-cated that the pull-out resistance couldbe amplified significantly by ensuringthat the screw was anchored into theregions of stronger materials locatednear the cortical shell. Furthermore, theparameter found to have the strongestinfluence on the pull-out force was thescrew pitch. For pitch values varyingfrom 2 to 5 mm, the pull-out forceincreased linearly by approximately 30 percent. The variation of the screwdepth and the thread inclination hadlimited impact on the pull-out force.

A comparison of the numerical resultswith the experimental results foundthem to be in very good agreement,within the tolerance of experimentalerror.

The main advantage of the numerical models lies in the accuratesimulation of both the structure and theshape of the various portions of thebiological disc or vertebra as well as of the constitutive behavior of the different materials. In order to further improve the accuracy of thesenumerical analyses, researchers mustdevelop studies using models ofincreasing sophistication adapted tospecific groups of people with mor-phology and properties varying withage, sex, type of activities, degenera-tions and other factors.

References

[1] Tyrell, A; Reilly, T; Troup, J., CircadianVariation in Stature and the Effects of SpinalLoading, Spine, 1985, 10(2), pp. 161-164.


BIOMEDICAL: ARTIFICIAL ORGANS

An important challenge facing thedesign of turbodynamic ventricularassist devices (VADs) intended forlong-term cardiac support is the opti-mization of the flow geometry tomaximize hydraulic efficiency whileminimizing the peak shear stress in theblood flow. High efficiency reduces therequired battery size while low shearreduces the number of red blood cellsthat are ruptured by the pump. A pedi-atric heart-assist pump is particularlychallenging. Due to its small size(about 28 mm diameter by 51 mmlength), the design laws for adult-sizedpumps do not apply, and they cannotbe scaled. Therefore, the design ofpediatric blood pumps must rely onmodern design approaches to opti-mize the flow path. Computational fluiddynamics (CFD) has been widely usedin the field of artificial heart pumps forthe analysis of internal flow because itoffers an inexpensive and rapid meansof acquiring detailed flow field informa-tion that is expensive and painstakingthrough in vitro testing. LaunchPointTechnologies, Inc., in the UnitedStates, which developed the first mag-netically levitated (maglev) heart pump(the Streamliner ventricular assistdevice that reached animal trials in1998), finds that CFD is a powerful toolin the performance assessment andoptimization of artificial heart pumps.

LaunchPoint has developed a CFD-based design optimization approach

Designingwith HeartCFD-based design optimization for a miniature ventricular assistimplant can shave years off themedical device development cycle.

By Jingchun Wu, LaunchPoint Technologies, Inc., California, U.S.A.and Harvey Borovetz, McGowan Institute for Regenerative MedicinePennsylvania, U.S.A.

that integrates internally developed 3-D inverse blade design methods,parameterized geometry models, automatic mesh generators and math-ematical models of blood damage withthe commercial ANSYS CFX solver.The system provides rapid optimiza-tion for various types of centrifugal,mixed-flow and axial-flow bloodpumps. The ANSYS CFX solver waschosen because of its robustness forcomputations with multiple frames ofreference (MFR) (the coupling betweenrotating and stationary components).

A new LaunchPoint VAD, Pedia-Flow™ is intended to deliver a flow rateof 0.3 to 1.5 l/min against 100 mmHgpressure rise to neonates and infantsweighing 3 to 15 kg. The PediaFlowwas designed with a magnetically sus-pended, mixed-flow style impeller witha single annular flow gap between therotor and housing to avoid unfavorableretrograde flow and separation. Theshear stress transport (SST) model, alow Reynolds number turbulencemodel, was selected for the turbulentflow simulation, which was justified by the representative Reynolds number of ~30,000 based on theimpeller outlet diameter and the pumptip speed. Although blood exhibits non-Newtonian behavior at very lowshear rates, many studies have shownthat blood can be modeled as a Newtonian flow at a shear rate largerthan the threshold of a 100 s -1. The

shear rate in the computational modelof the PediaFlow is much larger thanthis threshold, so Newtonian blood witha constant viscosity of 0.0035 Pa-s anda density of 1040 kg/s3 was assumedfor the simulations.

The CFD-predicted velocity vectorsat both the mid-span blade-to-bladeregion of the impeller and the vane-to-vane region of the stay-vanes show avery smooth distribution without anyvortices at the nominal flow conditionfor the optimized PediaFlow model. Asliterature is replete with anecdotal evi-dence that recirculating flows lead toattachment of platelets to biomaterialsurfaces — which in the clinical VADsetting can promote blood clot forma-tion — reverse flows and vortices areundesirable. The CFD results foundthat a smooth and gradual transition inthe secondary flow velocity was present at the curvature of one inflowand outflow cannula geometry. Thisgraduation helps to prevent separationand reversal flow for the primary flowvelocity. In addition, the predictedpathlines of representative particlesthrough the entire flow region did notexhibit any vortices.

The exposure of blood elements toshear stress above a certain thresholdas a function of exposure time cancause hemolysis, which actively breaksopen the red blood cells; activateplatelets, which can cause clottingproblems; and denature proteins, which

The PediaFlow ventricularassist device provideslong-term cardiac supportfor infants.


BIOMEDICAL: ARTIFICIAL ORGANS

alters the proteins so they can no longer carryout their cellular functions. Thus, it is desirableto minimize the shear stress that blood passing through the pump may experience.Using the results of the CFD simulation, a plotof shear stress versus exposure time for particles passing through the pump demon-strates relative uniformity within the annularflow gap region, but it is less uniform withinboth the impeller and stay-vane regions. The overall mean blood damage through the entire domain of the model is dividedaccording to the three main regions of theflow path: impeller, annular gap and the stay-vane. The analysis reveals that the hemolysislevel in the annular gap region is highest,accounting for more than 50 percent of thetotal, while the level of hemolysis in theimpeller region and stay-vane region is almostthe same, each causing approximately 20 to25 percent of the total blood damage.

CFD-based design optimization with theintegration of the ANSYS CFX solver can significantly reduce the design optimizationcycle from years, compared to the traditionaltrial-and-error methods, to just severalmonths. It provides detailed and useful flowfield information from which blood damagemay be computed, and it also predicts thehydrodynamic characteristics such as therelationship of developed pressure and efficiency to flow rate.

This research was supported in part by NIH ContractNo. HHSN268200448192C (N01-HV-48192).

PediaFlow is a trademark of WorldHeart, Inc.

Shear stress history from impeller inlet to stay-vane outlet Proportion of total blood damage at different pump components undernominal flow condition

Predicted smooth velocity vectors at mid-span blade-to-blade region of the impeller (left) andmid-span vane-to-vane region of stay-vanes (right)

Secondary flow streamlines at sections of inflow cannula (left) and sections of outflowcannula (right)

Pathlines of particles at inflow cannula and impeller side (left) and stay-vanes side andoutflow cannula (right)

0.0025

0.0020

0.0015

0.0010

0.0005

0.0000Impeller Annular Gap Stay Vane Total

Giersiepen

Heuser

0.002337 0.002341

21.20% 19.5%

52.5%57.7%

26.27%22.8%

dHb

300

250

200

150

100

50

00.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18

Shea

r Stre

ss (P

a)

Diseases such as asthma, chronicobstructive pulmonary disease (COPD)and cystic fibrosis can have a signifi-cant adverse impact on the structureand integrity of the lungs’ airways.While functional magnetic resonanceimaging (MRI) allows for measure-ment of air flow, computational fluid dynamics (CFD) provides highlydetailed information of local flow characteristics and resistances. Thefirst requirement of a patient-specific analysis is knowledge of the bounding walls of thepatient’s flow domain — their lunggeometry. This type of informationusually comes from computedtomography (CT), a scan that indi-cates detailed information aboutlung geometry because of thenatural contrast between air andthe lung walls. The main draw-back of CT is that the resultingscan is a static image. Couplingcomputational analyses of air flowwith the lung scan has the poten-tial to provide significant addedvalue to the clinical evaluation of lung function.

FluidDA, a spin-off of theAntwerp and Ghent universities inBelgium, has successfully devel-oped a workflow for predicting airflow in healthy and diseased lungs

simulate and examine the air flow. Flowpatterns, relative pressure drops anddrug delivery profiles are readilyextracted from the simulation results.The resistance distribution — definedas the total pressure drop over variouslung segments — also is available.

The pharmaceutical and medicaldevice sectors also can benefit frompatient-specific flow analysis as a wayto evaluate performance and efficacy ina virtual patient population. In clinical

studies, it is possible to analyzethe effect of bronchodilating medication, which widens lung airpassages and relaxes bronchialsmooth muscle to ease breathing,on airway volume and flow resist-ance. A researcher then can beginto establish correlations betweendrug deposition patterns and clini-cal outcomes, thereby providingan indication as to why the drugdoes or does not work. Functionalimaging also can be used toassess the placement of intra-bronchial devices such as stentsand valves.

Coupled with CFD, suchimaging can dramatically increaseinsight into medical assessmentand improve the accuracy of medical interventions.

www.ansys.comANSYS Advantage • Volume I, Issue 2, 2007s14 www.ansys.coms14

BIOMEDICAL: IMAGING

Going with the FlowFunctional biomedical imaging through CFD provides a new way of looking at pathological lungs.

By Jan De Backer and Wim VosFluidDA nv, Antwerp, Belgium

Reconstructed airway of a patient with cystic fibrosis:The red arrows indicate regions in which inflammationhas restricted the airways.

Contour plots show the effect that the use of a bronchodilator has on the local values for airway volume (left) andresistance (right); red indicates high values and blue indicates low values.

For patients with deformation of the spinal column (kyphoscoliosis),simulation can be used to determine the site of obstruction and/or respiratory function.

Obstruction site (and subsequent location) of an intrabronchial stent,which re-inflated the blocked lower right lung lobe. Pressure contoursare plotted in the airway.

An increase in the volume of the lower lobe is clear in time followinginsertion of a stent.

Stent location

Lower lobe

using CFD. The fluid and structuraldynamics company combines clinicalexperience and capabilities withnumerical simulations to offer a varietyof services to the healthcare industry.

The workflow process begins withthe conversion of CT scan data into a3-D computer model of the airway,performed with the Materialise productMimics. FluidDA then uses TGrid software to create surface and volumemeshes and FLUENT technology to

ANSYS Advantage • Volume I, Issue 2, 2007www.ansys.com s15www.ansys.com s15

BIOMEDICAL: SURGICAL TOOLS

The SpineJet repairs a herniated intervertebral disc by removing a portion of thenucleus. The tool uses the Venturi effect created by high-velocity saline jets tocut and then aspirate targeted tissue. Image courtesy T.G. Communications.

In the United States, back pain is one of the most com-mon reasons for healthcare visits and missed work. Fourout of five adults have at least one bout of back pain atsome point in their lives.

A common source of pain is from a bulging interverte-bral disc impinging on spinal nerves, which can cause backpain or sciatica (pain down the leg) — a condition known as herniated disc. The intervertebral disc is sandwichedbetween the vertebrae of the back and acts as a shockabsorber during spinal movement. The disc is made of twoparts: a tough outer wall called the annulus and a gelatinousinner core called the nucleus. Trauma or aging of the disccan cause the annulus to bulge.

Most occurrences of lower back pain resolve with restand medication. For many people, though, the pain can bedebilitating and last for several months to years. Suchpatients typically require surgery.

Minimally invasive surgical techniques offer many bene-fits, since traditional back surgery can cause further painand complications. HydroCision, which develops and man-ufactures fluidjet-based surgical tools in the United States,used computational fluid dynamics (CFD) to improve anovel minimally invasive surgical treatment calledHydroDiscectomy™.

The goal of HydroDiscectomy is to decompress theherniated disc. When performing the procedure, a physi-cian uses a tool called the SpineJet® to remove a portion ofnucleus, which debulks the disc and retracts the bulge.The device uses a high-pressure jet of sterile water directed into an evacuation tube. The jet is attuned to cutthe softer nucleus but protect harder surrounding tissuessuch as the vertebrae and the annulus. The water jet natu-rally provides cutting and a low-pressure Venturi to drawthe nucleus to the jet, cut it and aspirate it through anevacuation tube.

As physicians adopt new technologies, their productdemands increase. HydroCision saw CFD as a technologythat could reduce development time and improve productperformance. Manufacturing limitations with the existingSpineJet nozzle affected the flow divergence, directionalityand alignment with the evacuation tube. By redesigning theSpineJet nozzle for better flow characteristics and greaterease of manufacture, the surgical device could be made more consistent and cost-effective. HydroCision’s productdevelopment team used FLUENT software in analyzing theperformance of the existing nozzle geometry. CFD simula-tions allowed new geometries to be designed and analyzedfor performance in a matter of hours to days. Optimizationof the device was faster and less expensive than the tradi-tional method of making and testing prototypes.

The CFD model included flow simulations through thesupply tube, nozzle orifice and evacuation region. CFDresults helped the HydroCision team visualize critical flowcharacteristics such as the velocity profile, pressure distri-bution and flow divergence (cone angle).

The team modeled six alternate SpineJet designs thatincorporated significant changes to the nozzle and/or thesupply tube. Engineers selected velocity magnitude andgeneral jet shape as the primary means for comparing thedifferent designs, since these two parameters are con-sidered the most accurate predictors of overall SpineJetperformance.

By Joe Richard, HydroCision, Massachusetts, U.S.A.Brenda Melius, consulting firm, New Hampshire, U.S.A.

Battle of the BulgeRapid prototyping results in a new surgical tool to treat back pain.

Supply and evacuation tube of the original SpineJet Image courtesy T.G. Communications.


BIOMEDICAL: SURGICAL TOOLS

CFD results for the existing SpineJet showed the influ-ence of a sharp-edge orifice and its location on the flowcharacteristic. As expected, the orifice creates a flow sepa-ration at the corner, and a vena contracta is formed. Inaddition, the proximity of the orifice to the 90-degree-bend inthe supply tube and the additional supply tube length pastthe orifice create a non-uniform flow condition at the orifice entrance. As a result, the region of highest flow velocity is concentrated in the lower portion of the orifice;therefore, the flow is neither symmetrical nor well developed.

CFD results for the alternate SpineJet designs showedsubstantial improvement compared to the existing design.Three of the alternate configurations had 20 percent highermass flow rates than the existing design as well as a 40percent reduction in cone angle (flow divergence). Thesedesigns had general jet shapes that were symmetrical andwell developed. They also retained higher flow velocitiesover longer distances from the orifice exit.

Historically, HydroCision manufactured prototypes ofnew geometries for testing to examine the feasibility of producing a new and improved design. Although fairlyeffective, this method was costly (more than $15,000 foreach design tested) and time-consuming (taking approxi-mately six months). Furthermore, testing did not alwayslead to a full understanding of the fluid flow characteristicsthat occur.

Computer modeling utilizing FLUENT software pro-vides a different approach to the problem. The onlyexpenses are computing and software costs; creating aCFD model and running it takes just a few days. This allowsHydroCision to model and refine many designs in a fractionof the time it would take to manufacture and test a singleprototype. In addition, computer simulation can yield betterinsights into the interactions between the geometry and thefluid flow. Finally, the graphics generated by FLUENT soft-ware help stakeholders better understand the operation ofthe surgical tool.

Cross-sectional view of all fluid volumes for original SpineJet design(top) with close-up section indicated by the red box at orifice (bottom)

Cross-sectional view of SpineJet alternative design colored by velocity magnitude

s16

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

ANSYS Advantage is published for ANSYS, Inc. customers, partners and othersinterested in the field of design and analysis applications. 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, DesignModeler, ANSYSMechanical, DesignSpace, ANSYS Structural, TGrid, GAMBIT, and any and allANSYS, Inc. brand, product, service and feature names, logos and slogans areregistered trademarks or trademarks of ANSYS, Inc. or its subisdiaries located inthe United States or other countries. ICEM CFD is a trademark licensed byANSYS, Inc. All other brand, product, service and feature names or trademarksare the property of their respective owners.

© 2007 ANSYS, Inc. All rights reserved.

About the Industry Spotlight

Cover image: Simulation demonstrates shape memory for a cochlear implant.Photo courtesy Cochlear GmbH. Simulation courtesy Fachhochschule Hannover– University of Applied Sciences and Arts, CADFEM GmbH and Dr. OmidMajdani – Hannover Medical School.

Cross-sectional view of all fluid volumes

Supply tube volume

Supply tube 90°bend volume

Supply tubevolume

Orifice volume Evacuation tube volume


Managing EngineeringKnowledgeWeb-based solution is aimed at hosting and integratingsimulation data, processes and tools for more effectiveSimulation Driven Product Development.

Managing simulation processesand data is a specialized subset of thelarger product lifecycle management(PLM) vision. But it is often overlookedor poorly addressed, since managingsimulation processes and data is moredemanding than the file/document-centric approach of PLM and relatedproduct data management (PDM) systems. Simulation data is both richerand typically many orders of magni-tude larger than other types of productdata: It can be many gigabytes in sizeand can require sophisticated datareduction techniques. In addition, toextract the true value and knowledgerepresented by simulation data, a usermust capture both the content and thecontext associated with the productbeing simulated.

The complexity of the task not-withstanding, the need to managesimulation data and processes is nowmore important than ever. Robust datamanagement systems have the poten-tial to provide significant benefits tocompanies by enabling users toaccess and reuse historical designinformation and expertise for speedingcreation of new designs, providingways to capture and leverage existing engineering knowledge, andaddressing the problems of loss ofengineering expertise and protectionof intellectual property.

Process management in the context of product engineering essentially means optimizing the design workflow through more effectiveuse of computer-aided engineering(CAE) simulation tools. This can result in a wide range of improvements,including enterprise standards for work

procedures, consolidation and auto-mation of best practices, and increasedquality and reduction in errors.

Data or knowledge managementapplies an archiving system to allowfor searches based on relevant anddescriptive tags that help identify filesand their contents. Thus, what isinvolved is knowledge management —capturing both data content and context — rather than just file or datamanagement. This information canlater be mined for insight into the how and why of a design or simulation. A managed simulation environmentcan address this issue by automatingmuch of the uploading and data entry steps.

The ANSYS Engineering Knowl-edge Manager (EKM), scheduled forinitial release this year, is aimed at meeting these challenges with

TRENDS & PRACTICES

capabilities for backup and archival,traceability and audit trail, processautomation, collaboration, and captureof engineering expertise and IP pro-tection. It is a Web-based design andsimulation framework aimed at hostingall simulation data, processes and tools(whether in-house or commercial) whilemaintaining a tight connection betweenthem. It provides three services: accessmanagement to address deploymentand collaboration, process manage-ment to address integration andprocess automation, and knowledgemanagement to address the issuesassociated with simulation data. AddingANSYS EKM to the capabilities of theANSYS, Inc. family of simulation products empowers organizations tocreate enterprise systems and achievethe goal of Simulation Driven ProductDevelopment.

By Michael Engelman, ANSYS, Inc.

Web browser

Desktop application

Firewall

Content managementrepository

Application server

Compute cluster

Relationaldatabase

File serverRepository of all files

and applications

Stores meta data

Executes simulationsand extracts data

using a batch systemas RSM, LSF, PBS

ANSYS EKM


No-HassleKitchen ApplianceFinite element analysis helps redesign a countertop water filter that iseasier to maintain, can be injection-molded in half the time and costs athird less to manufacture than previous models.

By Matthew Stein, Stein Design, California, U.S.A.

Even with degrees from top technical schools and considerabledesign experience, engineers find complex parts — especially ones withmodern ergonomic curves — difficult to analyze with traditional handbook thermal and stress analysis. As a smallone-man design shop, Stein Designcompletes several such projects eachyear that benefit from the application offinite element analysis (FEA).

The firm has used the technology to develop a wide range of plastic andcast parts, including water filtration systems, drinking fountains, medicalbacteriological filters, emergency chem-ical drench systems and computer diskdrives. Clients include Hewlett-Packard,Seagate, Plantronics and Duraflame —companies that value Stein Design forproviding fast-turnaround designs thatmeet their unique engineering and busi-ness requirements. In the developmentof consumer products in particular, thefirm recognizes that product aestheticsand visual impact often are critical ele-ments in the commercial success of aproduct.

In one recent project, Water SafetyCorporation of America in the UnitedStates commissioned Stein Design tocomplete a major redesign of theirEssence™ countertop drinking waterfilter, an appliance intended to beattractive as well as effective in turningordinary tap water into better-tasting,healthier water. The goal was to cut

production time and cost while makingit easier for consumers to change thecarbon filter cartridge and flow meterbattery annually. The previous housinghad incorporated thick walls to accom-modate the hydrostatic pressure of150 psi required for certification by theNational Sanitation Foundation (NSF),

a mark recognized for its value in international trade and respected byregulatory agencies at the local, stateand federal levels. These thick wallsresulted in slow injection molding cycletimes, excessive material usage and an undesirably expensive housing.However, arbitrarily reducing material

A countertop drinking water filter was redesigned to cut costs while making it easy for consumers to changethe carbon filter cartridge and flow meter battery.

CONSUMER PRODUCTS


from the overall design could potentiallycause part failures leading to waterdamage of consumers’ homes and highwarranty costs. To account for theseissues, Stein Design used FEA in developing a lightweight, reliable designfor an appliance that would be easier forconsumers to maintain.

The redesign was started by performing an FE analysis of WaterSafety’s existing product. When thehousing was subjected to an internalhydrostatic pressure of 150 psi, analysis with software from ANSYS,Inc. showed that, in spite of its 0.27-inch-thick bottom wall and threeinternal ribs, stress levels of 5,360 psiwere unacceptably close to the yieldstrength of the ABS thermoplasticmaterial. In redesigning the housing,one of the primary concerns wasreducing this maximum stress to half the material yield strength — thusproviding a safety factor around 2.0 —while reducing wall thickness andinjection molding cycle time for the parts.

To arrive at an optimal design satis-fying these complex requirements,Stein Design performed an iterativeprocess of evaluating different wallthickness and rib combinations. Three-dimensional models were designed in SolidWorks® software and thenimported into ANSYS DesignSpace.Once the initial pressure loads andboundary conditions were set for thefirst model, the project geometry wasupdated with each new model itera-tion, making quick work of the analysis

of “what if” scenarios. Since this was ahighly cosmetic part, the maximum ribthickness was kept to a maximum of70 percent of the wall thickness toensure that the part would not displayexcessive marks where the ribs joinedthe outer cosmetic surface. Suchindentations occur when the plasticcools and shrinks, and they are con-sidered problematic on products thatmust be highly attractive in nature.

The iterative process of analyzingvarious rib and wall thickness combi-nations using FEA yielded a domedsurface having a wall thickness of0.175 inch and 12 radial ribs with a

thickness of 0.125 inch minus 1/2degree of rib draft. Rib height was 7/8inch at the outside wall and slopeddown to 1/2 inch at the inside of the ribhub. The maximum rib stress on thenew design was reduced to 2,240 psi,giving a safety factor of 2.3 andexceeding the 2.0 target. At the sametime, by reducing the nominal bottomhousing wall thickness from 0.27 inchto 0.17 inch, injection molding cycletime was cut by a factor of two and part cost was lowered by morethan a third.

ANSYS DesignSpace software isan integral part of many Stein Designprojects — and part of the reason thecompany has succeeded in the highlycompetitive engineering consultingbusiness. Small consulting firms withno full-time analysts on staff can’tafford to spend a lot of time and moneyon training to run a complicated FEAprogram. Engineers who use ANSYSDesignSpace need little training to behighly productive, and the tool interfaces seamlessly with SolidWorksmechanical design software. SteinDesign finds it very easy to make quickchanges to the part geometry and toregenerate the ANSYS DesignSpaceFEA solutions to investigate “what-if”scenarios early in the design process,when design changes have littleimpact on project schedules and tooling. Even though several monthsmay pass between FEA applications,the software is designed so users canget up to speed quickly in producingmeaningful results.

Right: After the iterative process of testing variouscombinations using ANSYS DesignSpace software, thefinal design included 12 radial ribs with a thickness of0.125 inch.

Left: In spite of the device’s 0.27 inch-thick bottom walland three internal ribs, stress-levels in the original designwere excessive.This showed up as red areas on the ribs,as displayed in this color-coded stress plot.

Three-dimensional models of the water filter weredesigned in SolidWorks and then imported into ANSYSDesignSpace software for analysis of various ribbingconfigurations and wall thicknesses.

CONSUMER PRODUCTS


AEROSPACE

Overcoming Big Challenges forSmall Turbojet EnginesIn developing an impeller for a microjet turbine engine for unmanned droneaircraft, engineers used FEA to reduce stresses by 20 percent, preventfatigue in high-speed rotating parts and study resonances in the assembly.By Bulent Acar, Tusas Engine Industries (TEI), Inc., Eskisehir, Turkey

The concept of the unmanned air vehicle (UAV) isthought to have been envisioned first by Leonardo Da Vinciin 1488. The idea was not put into action until World War I,however, when radio control and gyro-stabilization tech-nology were available to make such an aircraft feasible.UAVs became more advanced during the Second WorldWar, when they were used to train anti-aircraft gunners andfly attack missions. Most of these early machines wereremote-controlled, full-sized aircraft, but more recent tech-nology advancements have led to the development ofminiaturized UAVs, providing opportunities for cheaper,highly functional military aircraft that can be used withoutrisk to aircrews.

One of the most challenging aspects in the developmentof these small aircraft is designing compact, lightweightpropulsion systems for delivering the required performance.In one recent project, Tusas Engine Industries, Inc. (TEI),based in Turkey, used finite element analysis (FEA) in developing the high-speed, precision radial compressorimpeller for a microjet turbine engine to be used in UAVapplications such as target drones for testing the accuracyof surface-to-air and air-to-air weapon systems.

Recognized as a leader in developing and producing arange of high-quality aircraft engine parts for the worldwideaerospace industry, TEI was established in 1985 for aircraftengine assembly primarily in the Turkish region and laterexpanded into design, testing and manufacturing of com-ponents for gas turbine engines and other precisionsystems. The firm began advanced research and develop-ment activities in 1996; since then, it has participated inmajor international projects such as the Joint Strike Fighter(JSF) and the A400M Airbus military transport aircraft withthe advanced TP400 turboprop engine.

One of the most critical parts of the Tusas TEI-TJ-1Xmicrojet engine, the impeller compresses air entering theengine inlet to a high pressure and delivers it to the com-bustion chamber. Rotational speeds in the order of 100,000rpm are necessary to achieve high compression, resulting indesign challenges related to vibration, resonance, transonicflow, shock waves in diffusers and high stress levels.

Studies performed for the TEI-TJ-1X using FEA included structural analysis to determine stresses anddeformation of the impeller, modal analysis of the impellerand rotor, and rotordynamics analysis of the entire assembly

Finite element analysis was used in developing TusasEngine Industries’ turbojet engine as well as theadvanced turboprop engine shown here. These smallturbine engines are designed to power unmanned airvehicles for applications such as military target drones.

to study the response of the components to rotationaleffects. TEI used ANSYS Mechanical software to minimizestress and deformation in their impeller designs. Variouscombinations of mechanical, fluid and thermal loads wereconsidered. By using this approach, stresses in the criticalregions of the impeller were reduced by 20 percent.

TEI engineers also used ANSYS Mechanical technologyto examine the centrifugal and aerodynamic loads that can affect vibration of the blade and potential deformationof its geometry. Such deformation is a major concern inmaintaining proper tip clearance — the spacing betweenthe outer edge of the impeller blade and the inlet housing —under the range of operating conditions. If not carefullyaccounted for, excessive deformation could create the riskof contact between the blades and their housing.

Following the initial structural analyses that minimizedstress and deformation, Tusas engineers performed modalanalyses to determine dynamic characteristics of theimpeller. Analyses indicated that none of the impeller frequencies coincide with any of the resonance frequenciesfor the engine in the operational range of impeller speeds of100,000 to 120,000 rpm. Since rotational speed is very high,rotating parts (such as impeller and turbine) can undergomillions of cycles in a relatively very short time. Vibrationcharacteristics of the impeller were investigated in detail toprevent high cycle fatigue (HCF) as well as contact betweenthe impeller blade tips and the stationary inlet as a result ofexcessive vibration.

TEI performed full rotordynamics modal analysis on thecomplete assembly, including the impeller, shaft and turbine, to determine the resonant frequencies of each individual component. The most challenging aspect of thefull modal analyses was defining realistic boundary condi-tions for the rotor’s bearings and bearing housings, whosestiffnesses substantially affect modal response. In order tocalculate the bearing housing stiffness values correctly andprecisely, the engineering team created a whole enginemodel. ANSYS contact elements were used to blend thedifferent mesh patterns of the impeller, shaft and turbine fordynamic analysis of the assembly.

As a result of the analyses, three critical frequencieswere determined. The first and second frequencies affectthe impeller and turbine respectively, while the last frequency has impact on the shaft. The impeller and turbinecritical frequencies are especially important since they mayexist in operational range and/or during startup or shut-down cycles of the engine. This led the TEI team to makedesign modifications, including incorporation of integratedblades. Subsequent tests validated that critical frequenciesfor the impeller and turbine were within approximately 10 percent of the FEA simulated values, which was accept-able. The shaft-related critical speed occurred 25 percentabove the maximum operation speed. Critical shaft speeds could not be validated due to the requirement thatrotational speeds were higher than operational speeds. TheTusas engineers noted that, while the test apparatus was

operated at its maximum speed, there were no indications of vibration-induced problems related to the shaft.

Simulation in the early stages of the development cycleprovided valuable insight for quickly identifying potentialproblems and evaluating alternative solutions. This preventedlarge numbers of costly and time-consuming late-stagedesign changes, and it enabled TEI engineers to verify thedesign with the minimum number of physical tests. Simula-tion was a critical tool in TEI’s successful development of theTEI-TJ-1X microjet engine, which has successfully under-gone initial performance tests and is being used as a basisfor the design of an advanced turboprop engine TEI-TP-1X,now under development.

Rotordynamic analysis of the complete assembly was performedto determine resonant frequencies of each individual component,including the impeller, shaft and turbine.

The radial compressor impeller is one of the most criticalparts of the engine. In designing the microjet turbineimpeller, TEI engineers used ANSYS Mechanical technologyfor structural analysis to determine stresses and deforma-tion (top) and for modal analysis in showing displacementat various harmonic frequencies (bottom).


AEROSPACE


POWER GENERATION

Keeping It CoolModeling fluid flow and heat transferthroughout a nuclear fuel assembly helps prevent reactor burnout.

By Fahri Aglar, Turkish Atomic Energy AuthorityAnkara, TurkeyMustafa Ozer Gelisli and Emre Ozturk ANOVA Ltd., Istanbul, Turkey

Adequate cooling of fuel in nuclear reactors hasalways been an important safety concern. The bulk of theradioactive inventory of a nuclear reactor is contained inthe fuel elements, and, normally, their integrity can bedestroyed only by excessive temperature. Insufficientcooling of the fuel leads to burnout that can cause struc-tural damage, and subsequent leaching of radioactivefission products. Therefore, the main goal of nuclear safetystrategy is to avoid an imbalance between the heat gener-ation and heat removal in all operational states. Suchimbalance could result from transients in which either theheat generation exceeds the nominal values or heatremoval falls below these values. Another cause of imbal-ance could be the loss of coolant from accidents thatresult in the partial or total depletion of coolant required forthe heat removal. In past investigations of the problemsencountered in cooling the fuel used in nuclear reactors,thermal hydraulic studies have been carried out bothexperimentally and theoretically [1, 2].

As part of its work studying reactor safety, the TurkishAtomic Energy Authority (TAEK) needed to evaluate theflow and heat transfer characteristics of a material testreactor (MTR)-type fuel assembly. As a provider ofadvanced engineering fluid mechanics solutions in Turkey,ANOVA Ltd. performed this study to assist TAEK in its evaluation.

The fuel assembly consisted of plate-type fuel elements, with light water serving as both the coolant and the moderator. Based on geometry and boundary conditions provided by TAEK, ANOVA generated a mesh of the assembly using GAMBIT software. By assuming symmetrical flow and geometry, only one-quarter of the fuelassembly needed to be modeled. When the cross section ofthe fuel assembly was examined, distinctive geometrieswith variable cross-sectional area — such as narrow coolingchannels, slender fuel elements, and sudden enlargementsand contractions — could be seen. Therefore, during theGAMBIT modeling, the fuel assembly was divided into threeregions: the diffuser, the fuel plates and cooling channelsbetween them, and the outlet region. A generally hexagonalmesh was developed, and the three regions were connected through non-conformal interfaces. Accurateevaluation of wall shear stress and local heat transfer coefficients at narrow cooling channels was required, whichnecessitated a boundary-layer meshing scheme. Underthese circumstances, and following a sensitivity analysis,ANOVA analysts created a grid containing 2 million cells.

In an evaluation of the safety of a nuclear reactor, ANOVA simulated the rotating andseparating flow through the cooling channel and also modeled the wall shear stressand local heat transfer coefficients. Geometry created in GAMBIT 2.3 software of amaterial test reactor-type fuel assembly shows the diffuser and fuel plates (left) andoutlet region (right).

Comparison of velocity profile in channel 1 [1]

Temperature distributionon the nuclear reactor’sfuel plates

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

1.2

1

0.8

0.6

0.4

0.2

0

U/Ua

vg

1.2

1

0.8

0.6

0.4

0.2

0

U/Ua

vg

y+

y+


Pathlines colored by velocity showing a vortex observed in the space between thediffuser and plate region (left) and corresponding velocity vectors (right)

Pressure drop along the fuel assembly

Computational meshes in GAMBIT software for the cooling channel (left) anddiffuser section (right)

∆p= 119 kPa ∆T= 15.26K

∆p= 167 kPa

The realizable k–ε turbulence model with standard wallfunctions was used throughout the computations to exploitits advantage for simulating flow possessing rotation andseparation. The ANOVA team performed pressure andvelocity coupling with FLUENT software using the semi-implicit method for pressure-linked equations, or SIMPLE.The convection and diffusion terms of the equations ofmotion were obtained by cell-based discretization. One ofthe main input variables for the FLUENT simulation was thevolumetric heat generation, which TAEK extracted from theneutronic calculations using the WIMS-D/4 and CITATIONcodes. The power peaking factors, which describe the localpower density at the hottest part of a fuel rod, also were estimated and used to correct the volumetric heatgeneration term.

One of the main concerns of the simulation was thecomparison of the velocities at peripheral and central cooling channels. Engineers observed that the magnitude ofthe velocities at the peripheral cooling channels was slightlylower than the channels located at the center of the assem-bly. The reason for this became apparent when the influenceof the vortex observed in a region between the diffuser andthe fuel plates was taken into account. The vortex and itsinfluence are extremely important from the reactor safetypoint of view, and estimations revealed that this velocityreduction seemed to be negligible. The outcomes related tovelocity reduction also matched those obtained from exper-iment [1]. Further channel-to-channel flow distributionanalysis showed that the relative flow rate, evaluated as aratio of the flow rate in the individual channel to that of theassembly, decreased from the central channel to the outer-most channel. The plate-to-plate temperatures showed theopposite behavior; that is, the temperature increasedtoward the outer channels.

A final point of interest was that the pressure differencebetween the inlet and the outlet of the fuel assembly was inthe acceptable range and did not cause flow instability andphase change during normal operation. The pressure dropalong the fuel region was 70 percent of the total pressuredrop, which was in accord with experimental data [1]. The FLUENT results thus have been instrumental in understanding the complex 3-D flow in an MTR-type fuel assembly. Such CFD simulations have contributed significantly to the design and licensing of nuclear power systems.

www.taek.gov.tr www.anova.com.tr

References[1] Ha, T.; Garland, W. J., Hydraulic Study of Turbulent Flow in

MTR-Type Nuclear Fuel Assembly, Nuclear Engineering and Design,2006, 236, pp. 975-984.

[2] Franzen, F. L., Nuclear Power Plant Operational Safety — SafetyStrategy and its Technical Realization, IAEA Interregional TrainingCourse, Karlsruhe Nuclear Research Center, 1981.

POWER GENERATION


PROCESS EQUIPMENT

The Greening of Gas Burner DesignSimulation assists in developing efficient andenvironmentally friendly recuperative burnersused in heat-treating applications.By Örjan Danielsson and Marcus AnderssonKanthal AB, Hallstahammar, Sweden

Companies that depend on gasburners for heat-treating materials arechallenged with adapting to toughernitrous oxides (NOx) and carbon dioxide (CO2) emissions rules and legislation, along with maintaining high efficiency. To address this, KanthalAB uses a simulation-driven designprocess to develop innovative productssuch as the ECOTHAL® single-endedrecuperative (SER) burner, the latestaddition to the Kanthal family of heatingsolutions.

The key to success in delivering a low-emission, high-efficiency gasburner lies in well-defined combustion.Delivering too much air reduces heatoutput and increases the amount ofharmful NOx produced, whereas too

little air results in incomplete combus-tion that causes unburned residue inthe form of carbon monoxide (CO) and hydrocarbons. In the ECOTHALSER burner, fresh air and fuel are com-busted in an inner tube within a burnerassembly while exiting exhaust gasesare recovered, passed back through anouter tube that surrounds the innertube and used to heat the incomingfresh air in a recuperator regionupstream of the combustion area.Kanthal used computational fluiddynamics (CFD) to model and optimizeflow behavior, gas mixture control andcombustion efficiency. CFD simula-tions using software from ANSYS, Inc.,together with physical testing, resultedin an SER burner that had an efficiencyof approximately 80 percent — 10 to20 percent higher than conventionalSER burners — while still keeping NOxlevels below 50 ppm (or 20 mg/MJ).

To help ensure that Kanthal provides accurate recommendationsregarding procedures for maintainingproper performance, the companyuses ANSYS Mechanical software tomodel creep. Kanthal’s burner systemsoften are mounted horizontally, andcreep, or deflection, of the tubes canaffect flow characteristics and opera-tion within the burner. The deflectionrate typically is measured throughphysical testing in which a sample tubeis placed in a furnace and the deflec-tion is measured at specified intervals.This is a very time-consuming test that can take up to 3,000 hours, or 125 days. On the other hand, whenmodeling creep, existing test data isused to provide the coefficients for the creep equation that is used in thesimulation inputs, and the testingprocess is simulated in less than a day.

Geometry of a single-ended recuperative burner, in whichfresh air and fuel are combusted in an inner tube within a burner assembly while exiting exhaust gases are recovered, passed back through an outer tube that surrounds the inner tube and used to heat the incomingfresh air in a recuperator region

Contours of velocity and streamlines predicted by ANSYSCFX, viewed looking into the inner tube from its inlet area

Streamlines indicate exhaust gases that are exiting theinner tube and recovered back through the outer tubeof Kanthal’s SER burner.

The SER burner from ECOTHAL isthe first in a family of five burners. The second burner in the series wasdesigned entirely using CFD in combination with traditional computer-aided design (CAD) software. Theability to go directly from a 3-D CADmodel to meshing and simulation within the ANSYS Workbench platform,and then to pass design changes backto the CAD program, greatly improvedthe speed of product development forKanthal. By using simulation in conjunction with CAD tools, expensiveand time-consuming laboratory testingwas kept at a minimum, and develop-ment time was reduce by severalmonths.

www.kanthal.com

Furnace wall

Gas inlet

Air inlet

Flue gas recirculation

Recuperator

Inner tube

Outer tube


THOUGHT LEADERS

The Democratizationof Engineering AnalysisTo compete successfully in today’s business climate, Procter & Gamblemakes analysis tools available to rank-and-file engineers as well as toanalysts and advanced simulation experts.By Fred Murrell and Tom Lange, Procter & Gamble Company, Ohio, U.S.A.

Early practitioners of techniques such as finite element analysis (FEA) andcomputational fluid dynamics (CFD) typically were confined to industries inwhich the risks to human life or mission success were such that the expensecould be justified. It is, therefore, no surprise that the first commercial FEA pack-ages came from and were used by industries that could afford access toexpensive computational resources — and for which a failed component couldhave catastrophic results.

As these techniques spread to other industries, computer-aided engineering(CAE) remained the bailiwick of the expert analyst, requiring advanced degreesand long apprenticeships to cope with the difficulties of the technique and toensure accurate results.

The rapid and unrelenting improvements in hardware, the personal com-puter and low-cost cluster computing — and technology such as the ANSYSWorkbench platform — has truly democratized CAE analysis. A common desktop PC has more than 10 times more computing horsepower than a high-end workstation from just 10 years ago costing 10 times that price. No longer is engineering analysis a luxury that costs many thousands of dollars requiring theservices of highly trained experts.

Fred Murrell Tom Lange


THOUGHT LEADERS

www.ansys.com 29

The Procter & Gamble Company (P&G) is best knownfor its brands. Three billion times a day, P&G brands touchthe lives of people around the world. The company has oneof the strongest portfolios of trusted, quality leadershipbrands, including Pampers, Tide, Ariel, Always, Whisper,Pantene, Mach, Bounty, Dawn, Pringles, Folgers, Charmin,Downy, Lenor, Iams, Crest, Oral-B, Actonel, Duracell, Olay,Head & Shoulders, Wella, Gillette and Braun. Chances arethat you have used one of these brands recently, if not today.

P&G also is well known for advertising these brands.According to Advertising Age magazine, P&G was thelargest advertiser in the United States in 2005, spendingmore than $4.6 billion. Advertising, packaging, display andname recognition are aimed at what P&G refers to as the first“moment of truth,” when a customer decides to purchase aproduct they have never used before. But, as any manu-facturer knows, you won’t have a customer for long if yourproduct doesn’t deliver as promised. If you fail the consumerthe first time, you will not be rewarded with repeat business.P&G calls this the second moment of truth, when a customeruses the product and judges whether you have delivered onyour advertised promise. This is where the science behindthe brands comes into play.

All manufacturers face similar tensions — rapid innova-tion, keeping down costs and improved time to market. P&Gis a leading proponent of CAE technologies in its drive forimproved innovation. In fact, in a 2003 conference call withWall Street analysts, P&G Chief Executive Officer A.G. Lafleystated, “We are significantly expanding capabilities in computational modeling and computer-aided engineering,so we can do an increasing percentage of product andprocess design through virtual simulation.”

In the consumer packaged goods business, this wouldnot have been realistic or feasible just 10 years ago. High-end CAE analysis was then the domain of experts, most

likely employed in the defense, aerospace or automotivebusiness. The expert also was armed with complicated,high-end analysis software and an expensive UNIX®

workstation. Today, the ubiquity of inexpensive, fast andpowerful desktop PC workstations has made the use of CAE analysis available to the rank-and-file engineer in waysunimaginable in the past.

When P&G creates new products, there are three goals — it has to fit, do what it is supposed to do and, mostimportant, make financial sense. The company wants tomake the first prototypes virtually and make the physicalitem only when confident it will work. In the consumer packaged goods business, companies make billions ofitems and sell them for a relatively small amount. Analysisallows P&G to optimize those products and processes tosave a penny or two here and there. The focus is in makinglots of high-quality products very quickly.

Just as the needs of individual projects vary, so doschemes for utilizing CAE. P&G has developed a three-tierapproach for CAE. Tier 1 consists of a small cadre ofexperts. They face new-to-the-world kinds of problems thatrequire a great deal of preparation and development. Here ahighly trained, advanced-degree individual will stretch thebounds of a high-end commercial code or require special-ized codes from national laboratories to solve the problem.

The second-tier analysts use very high-end analysistools, but the problems are such that the tools can be auto-mated to some extent. A common example at P&G is theanalysis of bottles. P&G sells billions of packages each year.Design optimization is critical to maintaining competitive-ness and profitability. The sheer numbers of projectsannually require a different level of expertise to achieve effec-tive results. P&G has chosen to automate a number of theseanalyses in a product called the Virtual Packaging System(VPS). VPS is a collection of common analysis tasks that

Simulation on the ANSYS Workbench platform was used to determine stresses (left) and thermal distribution (right) in these components of high-speed equipment used inProcter & Gamble Company production operations.


THOUGHT LEADERS

have been developed over the years and automated to a large degree by internally written code. This allows a journeyman analyst to feed various geometries to the systemand view the results in short order. The time to complete an analysis is reduced substantially. “This system frees analysts to focus on the physical parameters of the designproblem rather than on setting up analytical solutions,” saidDavid Henning, manager of packaging analysis at P&G.Typically, there are three times as many analysts in this category as there are experts.

A third tier is that of the rank-and-file engineer engagedin project work. Occasionally, this individual is faced withthe need for an analysis to determine the suitability of astructure for a particular load or other such question. In thepast, a call would go to the expert practitioner who may (or may not) have the time or resources to assist. For thesetypes of analyses, P&G uses the ANSYS DesignSpaceproduct as the software of choice. ANSYS DesignSpacewas selected after a careful investigation of solutions available in the marketplace.

In making the decision on which software to use, thecomparison requirements were ease of use, accuracy, fullassociativity with a number of 3-D computer-aided design(CAD) systems, and widespread training and support. In the end, ANSYS DesignSpace software was selected.

Proctor & Gamble rank-and-file engineers routinely use ANSYS DesignSpace software in product development projects. Sample plots here show loads on the slotted concentric shafts of a converting machine assembly, enabling engineers to quickly evaluate the design early in development.

The solution allows for escalation of the problem to ANSYSMechanical or ANSYS Multiphysics software if a particularanalysis requires nonlinear materials, large deformation oradvanced contact. The ANSYS Workbench platform alsocontains tools for convergence studies that serve to ensurean accurate solution.

The ANSYS Workbench environment is available tothousands of engineers and scientists within the P&Gorganization. Training is available to those who wish to utilize the tools. P&G also finds that more and more newhires are already trained in CAE tools. These software products allow the engineer to rapidly screen numerousdesigns before having to commit to a physical prototype.The overarching goal is to make sure the first physical prototype has the best chance for success that the engineercan provide.

This translates to fewer, more meaningful tests,decreased innovation cycle times, and, most important,reduced time to market. This is where analysis makesmoney: in improving the decisions that are made every dayand getting a better product to the market faster.

Pampers, Tide, Ariel, Always, Whisper, Pantene, Mach3, Bounty, Dawn, Pringles,Folgers, Charmin, Downy, Lenor, Iams, Crest, Oral-B, Actonel, Duracell, Olay, Head &Shoulders, Wella, Gillette and Braun are registered trademarks of the Procter & GambleCompany, all rights reserved.


ANALYSIS TOOLS

www.ansys.com 31

Rotordynamics is a collective term for the study ofvibration of rotating parts found in a wide range of equip-ment including turbines, power stations, machine tools,automobiles, home appliances, aircraft, marine propulsionsystems, medical equipment and more. In these applica-tions, resonant vibration — in which mechanical systemscan oscillate excessively when excited by harmonic loads at their natural frequencies — is of particular concern.These large-amplitude vibrations can bend and twist rotating shafts, leading to premature fatigue failure in thesecomponents as well as bearings and support structures.Also, deformation of shafts and other components cancause rotating systems to impact adjacent parts in whichclearances are tight, causing potentially catastrophic damage in high-speed equipment.

Analysis of rotating systems typically involves the studyof many different variables related to vibration including thecritical rotational speeds that set up natural-frequency reso-nances, the response of the entire system to unbalancedloads and instabilities, deflection of the shaft during vibra-tion, torsional vibration in which shafts also twist aroundtheir axes, and flow-induced oscillations produced by fluidsmoving through the system. Calculation of these and othervibration-related variables can be performed in ANSYSMechanical software using some of the most advancedrotordynamics simulation capabilities available in com-mercial finite element analysis (FEA) codes.

Rotordynamics usually is best studied in the rotatingframe of reference, in which Coriolis terms are used in theequations of motion to describe rotational velocities andaccelerations. Introducing these Coriolis terms for static,modal, harmonic and transient analysis provides a modifiedequation of motion:

[M], [C] and [K] are the structural mass, damping and stiffness matrices, respectively. [Kc] is the spin softeningmatrix, and [G] is a “damping” matrix contribution due to therotation of the structure or the Coriolis term.

This modified equation of motion is at the foundation of performing the most common types of rotordynamicsanalyses.

Modal analysis: When components are spinning, the Coriolis term adds nonsymmetric terms that introduceforces to the system, causing natural frequencies to splitand shift up and down. These natural frequencies must bedetermined, therefore, to avoid excitations at the criticalspeeds. Modal analysis predicts how speed affects frequency by running at speeds from zero rpm up to themaximum rotational velocity of the system.

Harmonic analysis: A harmonic analysis sweeps through a range of frequencies to determine how the systemresponds to various rotating speeds and excitation forces.Again, the Coriolis terms shift the frequencies, and dampingplays a greater role. If the excitation is different from therotating frequency, ANSYS Mechanical technology offersoptions to scale it up or down.

Static and transient analysis: Static and transient analysesdetermine loads exerted on structures, joints and bearingsof rotating structures. This can be done as a static analysis(by applying initial conditions to specify velocities) or transient dynamic simulation in which the Coriolis effectsare included.

Rotordynamic Capabilitiesin ANSYS MechanicalUseful features are available to study vibrationbehavior in rotating shafts, bearings, seals, out ofbalance systems, instability and condition monitoring. By Achuth Rao, ANSYS, Inc.


ANALYSIS TOOLS

Capabilities for Rotordynamics AnalysisANSYS software offers a complete set of capabilities for

studying the dynamics of rotating machinery.

Solids, shell and beam elements: For decades, rotor-dynamics has been performed with in-house andcommercial codes using beams and masses. For mostrotor assemblies, this still is the most efficient and the mostaccurate method. However, sometimes a system does notlend itself to this type of approximation. ANSYS Mechanicalsoftware provides a unique solution to address such issuesusing 2-D and 3-D solid and shell elements to accurately

Case in Point: Beam Model Analysis of a Multi-Spool RotorThe following is an example of a harmonic analysis of a two-spool rotor on symmetric bearings with unbalance

force. An unbalance is located on the second disk of the inner spool, and harmonic response is calculated. The example uses an excitation frequency that is synchronous with the rotational velocity of the structure. ANSYSMechanical software calculates the rotational velocity Ω of the structure from the excitation frequency and anunbalance excitation force (F = Ω2 * Unbalance) is applied on the nodes.

In a typical rotordynamics harmonic analysis, quantities of interest such as nodal amplitude as a function of frequency, orbit plots at a given frequency and displacements plots at a given frequency are often output as part ofthe analysis.

Beam model of a two-spool rotor with symmetric bearings (left) and displacement plot (right)

model rotating machinery starting from computer-aideddesign (CAD) geometry.

Bearings and damping: In real-world rotating systems, bearings are not infinitely stiff, and the friction and lubricant inthem introduce damping. Also, springs in these systems oftenhave stiffness that varies with speed and direction. The same goes for damping. ANSYS Mechanical offers spring-damper elements like COMBI14, or the newer COMBI214 for modeling bearings in rotor dynamics, allowing users to specify stiffness and damping ratios for their particular systems.

VALU

FREQ

1.OE.03

1.OE.04

1.OE.05

1.OE.06

1.OE.07

025

5075

100125

150175

200225

250

Amplitude versus frequency (left) and orbit plots (right) for harmonic analysis of a beam model


ANALYSIS TOOLS

www.ansys.com 33

Case in Point: Solid Model Analysis of a Hard Disk AssemblyIn a hard disk assembly, modal analysis is run to predict how speed affects frequency by running at zero rpm

and then several speeds up to the maximum rotational velocity the system is expected to see. The primary post-processing tool for modal analysis is the Campbell diagram.

Hard drive assembly modeling using 3-D solid, beam and spring elements

Stationary and rotating frames: ANSYS Mechanical soft-ware provides both rotating and stationary reference framesfor rotor-dynamics analysis. The primary application for astationary frame of reference is a case in which a rotatingstructure (rotor) is modeled along with a stationary supportstructure. The primary application for a rotating frame of ref-erence is in the field of flexible body dynamics in which,generally, the structure has no stationary parts and theentire structure is rotating.

Unbalance response: ANSYS Mechanical allows users tospecify whether the excitation frequency is synchronous orasynchronous with the rotational velocity of a structure.New capabilities in the software such as the SYNCHROcommand update the amplitude of the rotational velocityvector with the frequency of excitation at each frequencystep of the harmonic analysis.

Campbell diagram: The primary post-processing tool forrotordynamics work is the Campbell diagram showing howvibration modes split because of whirling. The Campbelldiagram assists users in finding the critical speed for arotating synchronous or asynchronous force as a functionof rotation speed.

Whirl orbit plot: When a structure is rotating about an axis and undergoes vibration motion, the trajectory of anode around the axis generally is an ellipse designated as awhirl orbit. ANSYS Mechanical software provides plottingtools of the whirl for beam/mass and solid rotordynamicmodels. The orbit (ANHARM macro) can be animated forfurther examination.

The author would like to thank the development and technical supportteam at ANSYS, Inc. and Eric Miller from Phoenix Analysis & DesignTechnologies (PADT) for their efforts and contribution to this article.

Campbell diagram (top) and mode shapes (bottom) from modal analysis withCoriolis effects

Freq

uenc

y (H

z)

Spin velocity(rd/s)

1040

832

624

416

208

00

376.992753.984

1130.9761507.968

1884.962261.952


TIPS & TRICKS

Submodeling in ANSYS WorkbenchBy Dave Looman, ANSYS, Inc.

Submodeling utilizes two separatemodels. A full or global model repre-senting the entire structure is used totransform global loads to local defor-mation. The submodel includes the localgeometric details with an appropriatemesh density. The submodeling algo-rithm then interpolates the deformationfrom the global model to the submodel“cut boundaries” and solves for the localstress state.

This method typically requires exten-sive planning and documentation of theworkflow, especially if many submodelsand numerous load cases are involved.In addition, setup of a submodel maytake considerable time. However, theANSYS Workbench tree and efficientcomputer-aided design (CAD) interactionmake the procedure easier. With smallANSYS Parametric Design Language(APDL) enhancements applied to themodel tree, the submodeling techniquemay be combined with the ANSYSWorkbench Geometry handling andprocess documentation. Thus, a work-flow can be presented that covers thewhole process from CAD to fatigueanalysis in five steps.

Step 1. Build/import the model fromCAD. To illustrate this process, a sampleanalysis is performed to determinestresses on a tubular welded assemblywith a regular pattern of joints. It repre-sents a small sample section of arepetitive structure that forms the trackof a roller coaster. The model shown inFigure 1 has been created using ANSYSDesignModeler software.

The loading on this structure iscaused by a trolley rolling along the twoupside tubes. This loading is transferredto the larger tube via the joint elementsand passed to the supporting structure.Loading is generated by gravity and centrifugal forces.

Figure 1. Full CAD model of a curved tubular assemblyApplication courtesy Klaus-Dieter Schoenborn, CADFEM.

Figure 2. Full global model


TIPS & TRICKS

Figure 3. Submodel geometry of a critical structural detail

To obtain accurate stress in a local region, submodeling separates localanalysis from the global model. This allows mesh refinement in a regionthat might not be possible on the full model without exceeding size limits.

Step 2. Mesh and solve the globalmodel. Each part was meshed independently and connected withsurface-to-surface contact. Contactpairs may be part of the interpolatedregion in the full model, as long as they are completely enclosed by the submodel cut boundary. Figure 2shows a sample mesh with surface-to-surface contact.

Step 3. Analyze the global modelto identify critical spots. The fullmodel gives a general impression ofthe structural deflection and is used tofind the location of critical spots ofinterest. The structural result is savedand maintained in database andresults files in the parent directory. The global solution is needed for inter-polation of the submodel boundaryconditions.

Step 4. Generate the submodelfrom CAD data and mesh. Once critical spots are known, submodelssuch as the one shown in Figure 3 may be created from the original CADgeometry. This is done by cutting themodel and suppressing the remainingsolids. This process automaticallyachieves geometrical consistency sincethe remaining solid does not change itslocation with respect to the global coordinate system. In order to inter-polate displacements from the globalmodel to the submodel, a NamedSelection should be created on the cut boundaries, which will be referenced in the macro described next.

Step 5. Interpolate boundary conditions from the global model tothe submodel and solve. The inter-polation is done by inserting a smallmacro into the environment branch of

!!! ANSYS 11.0 Procedure !!!

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Copy results to parent directory !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!/COPY,file,rst,,coarse,rst,..\..\/COPY,file,db,,coarse,db,..\..\

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Copy results from parent directory !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

/COPY,coarse,rst,..\..\,coarse,rst/COPY,coarse,db,..\..\,coarse,db!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Submodeling Commands !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

FINI/PREP7CMSEL,s,cut_boundaryNWRITEALLSEL

SAVE,fine,dbFINI

/POST1 RESUME,coarse,db ! OPEN COARSE MESH FILE,coarse,rst ! DEFINE RESULTS FILE NAME

SET,1CBDOF

FINISH /PREP7

RESUME,fine,db ! RESUME SUBMODEL FROM FILE /INPUT,,cbdo

FINI/SOLU

Figure 4. Simulation tree showingthe submodeling procedure


TIPS & TRICKS

the submodel tree as shown in Figure 4.This macro resumes the full modeldatabase and results file, and it performs the displacement interpola-tion (CBDOF command). After that, thesubmodel is restored, the interpolatedboundary conditions are read and thesubmodel is solved. Note that anyexternal loads present in the submodel(including gravity effects or temp-erature loading) also should be applied.ANSYS Workbench Simulation solvesthe model and performs post-process-ing just as on any regular model. Fromthe interpretation of the resulting localstress state in the submodel, the criticallocations now may be reviewed withgreater fidelity.

For this example, the initial sub-model mesh is found to be still toocoarse to accurately predict fatigue lifefrom the resulting stress, so a locallyrefined mesh is needed. The “sphere of

influence” method of the ANSYSWorkbench platform is ideally suited toobtain the type of mesh needed. Figure 5 shows the refined mesh onthe submodel.

This step is then repeated simplyby creating the “sphere of influence”tab on the mesh branch and solving.The interpolation now is performed onthe new FE mesh since the macrooverwrites any files that were createdon a previous run. The interpolation isdone from the original results file,which still resides in the parent direc-tory. Equivalent stress in the submodelthen can be solved.

In the submodeling process,model consistency is maintained byusing ANSYS DesignModeler softwareto create both full models and sub-models. Capturing the process in thetree clearly archives the analysis andallows a user who later opens the

Figure 5. Submodel mesh detail using the sphere of influence feature

database to understand immediatelywhat was done. Note that an arbitrarynumber of submodels may be createdand solved by interpolation from a single run of the global structure. All of these submodels may be included in the model tree, and variants may bestudied without having to repeat thewhole process. Users should remem-ber to review and compare stresses atthe cut boundaries between the globalmodel and submodel to verify that thecut boundary is far enough from theregion of interest. Submodels may bealtered using the bidirectional CADinterface to ANSYS DesignModelersoftware or other CAD programs.

For more information, refer to chapter 9 on submodeling in the ANSYS Advanced AnalysisTechniques Guide.

ansys advantage 1-2-07

Documents