rina advanced marine materials 9-10 oct 2003slide 1 design, verification, and forensic correlation...

RINA Advanced Marine Materials 9-10 Oct 2003 Slide 1

Design, Verification, and Design, Verification, and Forensic Correlation of Forensic Correlation of Composite Yacht StructuresComposite Yacht Structures

Paul H. Miller, D. Eng. P.E.Paul H. Miller, D. Eng. P.E.

United States Naval AcademyUnited States Naval Academy

Annapolis, Maryland, USAAnnapolis, Maryland, USA


Presentation OverviewPresentation Overview

• Background/Project Intro and ScopeBackground/Project Intro and Scope

• Analysis TechniquesAnalysis Techniques

• ExperimentationExperimentation

• Forensic CorrelationForensic Correlation– Deck PanelDeck Panel– MastsMasts– HullsHulls

• SuggestionsSuggestions


BackgroundBackground• Team Dennis Team Dennis

Conner’s America’s Conner’s America’s Cup Campaign 2003Cup Campaign 2003– Relatively small Relatively small

programprogram

• Private lightweight Private lightweight cruiser/racercruiser/racer

• Both “high Both “high performance but performance but risk averse”risk averse”

• Materials:Materials:• TDCTDC

– Prepreg carbon/epoxy Prepreg carbon/epoxy uni (100uni (100oo or 135 or 135oo, 1 or 3 , 1 or 3 atm)atm)

– Aluminum honeycombAluminum honeycomb

• S/V CascadiaS/V Cascadia– Wetpreg carbon/epoxy, Wetpreg carbon/epoxy,

latent cure (~45latent cure (~45oo, 1 atm), 1 atm)– Cedar/balsaCedar/balsa


Design MethodologyDesign Methodology

1.1. Risk AnalysisRisk Analysis– All team membersAll team members

2.2. Uncertainty Uncertainty IdentificationIdentification

3.3. Targeted Targeted Performance and Performance and Structural Structural Evaluation (FOS=1 Evaluation (FOS=1 to 8+)to 8+)

Risk Analysis

Test Design/Experimental Verification/Tool Calibration

Mat’l Tests

Final Design


Primary Structural ToolPrimary Structural Tool• COSMOS/M finite element COSMOS/M finite element

analysis (SRAC)analysis (SRAC)• Linear (Mindlin and Linear (Mindlin and

DiScuiva) Laminated Shell DiScuiva) Laminated Shell Elements & user-definedElements & user-defined

• Nonlinear Material and Nonlinear Material and GeometryGeometry

• Tsai-Wu and user-defined Tsai-Wu and user-defined (Hashin) failure criteria(Hashin) failure criteria

• Global/Local rig and hullGlobal/Local rig and hull• Loads from FLOW (rig) or Loads from FLOW (rig) or

SPLASHSPLASH


Deck Panel TestDeck Panel TestCompressionCompression

• Duplicated critical partDuplicated critical part• Ply stacking Ply stacking

investigatedinvestigated• Non-traditional stack Non-traditional stack

judged bestjudged best• FEA 180%/112% of testFEA 180%/112% of test


Rig TestRig TestCompression and ImpactCompression and Impact

• Wall buckling Wall buckling limitedlimited

• ““Risky” and “Likely”Risky” and “Likely”• Weight criticalWeight critical• Resin content effect Resin content effect

evaluatedevaluated• Sidewall thickness Sidewall thickness

decreased from 6 to decreased from 6 to 3 mm3 mm

• 72% to 89% of FEA72% to 89% of FEA 185 Volvo Station Wagons!



• Buckling limitedBuckling limited• ““Risky” and “Likely”Risky” and “Likely”• Weight criticalWeight critical• Resin content Resin content

evaluatedevaluated• Sidewall thickness Sidewall thickness

decreased from 6 to decreased from 6 to 3 mm3 mm

• 72% to 89% of FEA72% to 89% of FEA



• ““Highest Highest Consequence” Consequence” Event Simulated, Event Simulated, Load & GeometryLoad & Geometry

• Adopted “damage Adopted “damage tolerant” design tolerant” design (resin, stacking (resin, stacking sequence)sequence)

• Both passed, 1 Both passed, 1 puncturedpunctured


Forensic CorrelationForensic CorrelationMast #1Mast #1

• Failure after 13 Failure after 13 hours of usehours of use

• Winds of 14-17 kts Winds of 14-17 kts at the time (23 the at the time (23 the day before but less day before but less tension)tension)

• 6 possible failure 6 possible failure scenarios, all scenarios, all evaluated by FEA, evaluated by FEA, one had 1.0 FOS.one had 1.0 FOS.

Three minutes before failure


Rig (Reactive) ModificationsRig (Reactive) Modifications

• Mast 2: extra Mast 2: extra reinforcementreinforcement

• Mast 3: improved Mast 3: improved QAQA

• Mast 4: same as Mast 4: same as Mast 1Mast 1

• Mast 5: reduced Mast 5: reduced structurestructure


IACC USA-77 HullIACC USA-77 Hull

• Sank after “rudder Sank after “rudder failure”, traced to poor failure”, traced to poor QA and bad QA and bad communicationcommunication

• FEA identified possible FEA identified possible damage spotsdamage spots

• All but one foundAll but one found• Assisted in repairsAssisted in repairs• Vessel regained Vessel regained

competitive statuscompetitive status


S/V Cascadia HullS/V Cascadia Hull



Outer Ply Factor of Safety

Correlation better than 95%


ConclusionsConclusions• Matching boundary conditions is Matching boundary conditions is

critical to FEA accuracycritical to FEA accuracy• Maintain global FEA models (w/ as-Maintain global FEA models (w/ as-

builts)builts)• Initial ultrasound mapping is beneficialInitial ultrasound mapping is beneficial• QA is critical, but often under-stressed QA is critical, but often under-stressed

in modern high-tech, low FOS in modern high-tech, low FOS applicationsapplications

• Hopefully no further research!Hopefully no further research!

rina advanced marine materials 9-10 oct 2003slide 1 design, verification, and forensic correlation...

Documents