8/6/2019 2007-06-28_Aeromat_rev6

1/31

Aeromat 2007Baltimore, 28 June 2007

Development of an Accelerated

Insertion of Materials (AIM) System foran Aluminum Extrusion

David Forrest*, Daniel Backman, Matthew Hayden*, Julie Christodoulou

*Naval Surface Warfare Center, West Bethesda, MDWorcester Polytechnic Institute and independent consultantOffice of Naval Research, Arlington, VA

8/6/2019 2007-06-28_Aeromat_rev6

2/31

2

NSWCCD: George Detraz (CAD), Al Brandemarte(metallography), Mary Beth Marquardt (HPC support),Catherine Wong (physical metallurgy)

Jan Backlund, Bo Bengtsson, Tolga Egrilmezer, Sapa(Extrusion processing, physical metallurgy, samples)

Paul Mason, Lars Hoglund, Thermo-Calc (API with PrecipiCalc)

Herng-Jeng Jou, Questek (PrecipiCalc)

Mike Foster, Scientific Forming Technologies (DEFORM 3D)

Jason Pyles, Engenious Software (iSight)

Yucel Birol, TBTAK Marmara Research Center(macroetchant)

Helpful discussions: ystein Grong (Norwegian U. of Science andTech.), Tim Langan (Surface Treatment Technologies)

Acknowledgements

8/6/2019 2007-06-28_Aeromat_rev6

3/31

3

Computational Materials Science:A natural and important advance in the evolution ofmaterials technology

Accelerated Insertion of Materials

Materials Engineering for Affordable New Systems

Integrated Computational Materials Engineering

Through Process Modeling

Important to Navy: developing internal capability

Insight into development process Appropriate goals for system demonstration/application

Commitment to infrastructure (hardware, software, trainedpersonnel)

System architecture that makes sense

Knowledge baseprocessing and physical metallurgy

Overview

8/6/2019 2007-06-28_Aeromat_rev6

4/31

4

Technology Transition

Technology transition is ageneric problem, including

materials technologies

http://www7.nationalacademies.org/nmab/CICME_home_page.html

General need to accelerate thetransition from science into

engineering practice Computational systems are

natural approach

MSE community needs toembrace, develop coherentframeworkCICME

8/6/2019 2007-06-28_Aeromat_rev6

5/31

5

Importance to Navy

Reduced cost of testing and evaluation

Reduced risk

Rapid insertion: advantages over enemy's systems

Mobility

Survivability

Lethality

8/6/2019 2007-06-28_Aeromat_rev6

6/31

6

System Demonstration

AA6082 Sidewall Panels for LCS

(1% Si, 0.9% Mg, 0.7% Mn)

8/6/2019 2007-06-28_Aeromat_rev6

7/31

7

System Demonstration

Issues :

Tradeoff between enough difficulty to show something substantial,yet easy enough to accomplish in ~2 years

6000 series alloy opportunity to model both extrusion processand microstructural transformations

Existing data Prior or published understanding of physical metallurgy

Target properties kept simple: hardness and tensile/yield strength

8/6/2019 2007-06-28_Aeromat_rev6

8/31

8

Infrastructure:Minimal Software Configuration

Process Integration and DesignOptimization

ISIGHT

Strength model based onmicrostructure

Custom software

Deformation and heat treatment

Recrystallization modeling

DEFORM

NGC of precipitate phases and size

distributions

PrecipiCalc

Aluminum thermodynamic databaseThermoTec TTAL

Mobility database for kineticsMOB2

Thermodynamic phase equilibriaThermo-Calc

UseSoftware package

8/6/2019 2007-06-28_Aeromat_rev6

9/31

9

System Architecture

CYGWIN (X-Windows)

Thermo-Calc & databasesFloating network license

iSight-FDFloating network license

PrecipiCalcNode-lockedMultiple, simultaneous users

Mobility DatabasesAccessed only by PrecipiCalc

Thermodynamic DatabasesAccessed only by PrecipiCalc

FLEX-LM serverThermo-Calc

iSight-FD

DEFORM 3DNode-lockedSingle user for runs

Code 60-MaterialsWindows 2000server

SEATech-HalseySGI Origin 3400Unix, 24 CPUs

RDT&E NetworkRDT&E Network

SEATech-BanyanBeowulf Linux Cluster96 Dual Xeon CPUs

8/6/2019 2007-06-28_Aeromat_rev6

10/31

10

Mechanics of Operation

Thermo-Calc,

Thermotec databaseMobility database

8/6/2019 2007-06-28_Aeromat_rev6

11/31

11

Knowledge Base

Compositional variations

Processing parameters

Physics of deformation and heat treatment operation

Recrystallization and precipitation models

Relationships between microstructure and strength

Microstructural data (e.g., initial precipitate distribution) Microstructural and mechanical data to verify numerical results

8/6/2019 2007-06-28_Aeromat_rev6

12/31

12

Knowledge BaseProcess Parameters

Sensitive information

Profile drawing

Die drawings

Thermal conditions of extruder components

Billet dimensions and temperaturesRam speed

Quenching rates

Aging times and temperatures

Billet microstructure

Material for analysis and validation of models

8/6/2019 2007-06-28_Aeromat_rev6

13/31

13

Extrusion simulation

3D problem

Lagrangian formulation

> 50:1 reduction Work still in progress

8/6/2019 2007-06-28_Aeromat_rev6

14/31

14

Extrusion Simulation

8/6/2019 2007-06-28_Aeromat_rev6

15/31

15

Thermomechanical

Modeling

Analytic approximation

Small Biot numberconstant temperature through

thickness

Stagnant for first few seconds

Then hbased on maintaining

7C/sec

T(x) = (To T)*exp(h

Cpw

t)+ T

0

100

200

300

400

500

600

0 20 40 60 80 100 120 140

Time (s)

Temp(C)

Case 1

Case 2

8/6/2019 2007-06-28_Aeromat_rev6

16/31

16

Thermo-Calc

ThermoTec TTAL database

Understanding ofphases beforeprecipitation modeling

Rejected all but:

FCC_A1ALPHA

MG2SI

8/6/2019 2007-06-28_Aeromat_rev6

17/31

17

Thermodynamics

Aluminum matrix (FCC_A1)

98.9 mole % Al with traces of the other elements in solid solution

(ALPHA, aka "Dispersoids")

Al72Si11Mn10Fe4

High temperature phasepresent at all temps up to m.p.

Mn is primary dispersoid-former

, , (MG2SI)

Thermodynamically stable below 498C

8/6/2019 2007-06-28_Aeromat_rev6

18/31

18

Thermodynamics

Actual composition of phases

(Mg2SI)

(Mg5Si3)

Not available in ThermoTec database

Mg5Si6

Various measurements Mg0.8-1.2Si

Not available in ThermoTec database

Myhr and Grong, "A Multiscale Modelling Approach to the Optimisation of Welding Conditionsand Heat Treatment Schedules for Age Hardening Aluminium Alloys," Graz, 26 May 2005.

8/6/2019 2007-06-28_Aeromat_rev6

19/31

19

Precipitation

Bratland, et al., Acta Mat., v. 45, No. 1, pp.1-22, 1997.

8/6/2019 2007-06-28_Aeromat_rev6

20/31

20

Precipitation Modeling

Using PrecipiCalc to Model Precipitation

8/6/2019 2007-06-28_Aeromat_rev6

21/31

21

Precipitation Modeling

Mg2Si

Surface energy = 0.055 J/m2

Lattice parameter = 4.034x10-10 m

Molar volume = 3.9541x10-5 m3/mol

Nucleation dislocation density = 1025/m3

Homogeneous nucleation

Multi-component interactions viaThermo-Calc

8/6/2019 2007-06-28_Aeromat_rev6

22/31

22

Exemplar Result

Approximate Thermal Cycle

8/6/2019 2007-06-28_Aeromat_rev6

23/31

23

Initial Distribution

Specified

8/6/2019 2007-06-28_Aeromat_rev6

24/31

24

Quench

8/6/2019 2007-06-28_Aeromat_rev6

25/31

25

Precipitation

www.sintef.no/static/mt/images/precipitation.pdf

8/6/2019 2007-06-28_Aeromat_rev6

26/31

26

Strength Model

pssiy ++=

Myhr, et al., "Modelling of the Age Hardening Behaviour of Al-Mg-Si Alloys,"Acta Materialia, 49 (2001) 65-75.

( ) 2/32/1

2/12

2

32 F

fGb

rb

Mp =

Strong, non-shearable particles, ri> r

c

Shearable particles, ri < rc

=c

ii

r

rGbF

22

22 GbFi =

=i

i

i

ii

N

FN

F

M Taylor factor, 3.1rc critical radius, 5x10-9 m

b Burgers vector, 2.84x10-10 m

G shear modulus, 2.7x1010N/m2

constant, 0.36

f volume fractionF mean interaction force between

dislocations and particles

8/6/2019 2007-06-28_Aeromat_rev6

27/31

27

Metallography

Grain size differences indicate influence of strain history

8/6/2019 2007-06-28_Aeromat_rev6

28/31

28

Metallography

No correlation between grain size and hardness (HV)

1.5 2.5 3.0 4.0 5.0 10.0

Grain size map (ASTM numbers)

Hardness map (HV)

8/6/2019 2007-06-28_Aeromat_rev6

29/31

29

Remaining Work

Complete DEFORM development

Vary speed, quench conditions

Integrate DEFORM using iSight-FD

Complete strength model and verify against published

data (tune PrecipiCalc)

Reconcile hardness variation with model results

Fully-coupled runs, generate response surface

8/6/2019 2007-06-28_Aeromat_rev6

30/31

30

Computational Materials Science: important advance

Navy committed to developing capability

Application balances difficulty vs. substantial result

Built infrastructure with iSight, Thermo-Calc, PrecipiCalc,DEFORM 3D, Custom microstructure model, databases

Networked system based on available resources Thermal history, extrusion processprecipitate distribution

PrecipiCalc can predict realistic anddistributions

Precipitate size distribution influences strength Actual part shows strength variations to be reconciled with

models

Summary

8/6/2019 2007-06-28_Aeromat_rev6

31/31

31

Dr. Julie Christodoulou, ONR Code 332

NSWCCD Code 61 Technology Enterprise

NSWCCD Code 0020 Bid & Proposal

Sponsors