validation of material models in uintah
TRANSCRIPT
Validation of Material Modelsin Uintah
Biswajit BanerjeeAugust 16, 2005
Outline
● Validation Process● Copper
– One-dimensional tests– Taylor impact tests
● 4340 Steel– Model parameters– One-dimensional tests
● Future Work
Validation Process : Goals
● Determine Model Error● Determine Numerical Algorithm Error● Minimize Errors
Validation Process : Models 1
● Elastic Models– Equation of state (for pressure)– Shear modulus model– Melting temperature model– Thermal expansion model
● Plastic Models– Shear modulus model– Melting temperature model– Model for heating due to plastic work– Specific heat model– Flow stress model– Yield condition model
Validation Process : Models 2
● Damage Models– Scalar damage model– Porosity model
● Failure Models– Energy balance – modifications of J-Integral
model– TEPLA-F empirical model– Material stability model
Validation Process : Step 1
● Verification of material models– Are the models correctly implemented ?– The material stress update algorithm is examined
and errors are quantified using 1-D tests or exact solutions.
● Validation of material models– Do the material models reflect “reality”?– What is the percentage error?
● “Reality” in this step means:– Controlled tests, usually 1-D
Validation Process : Step 2
● Validation of material models under more complex loading conditions– Non-uniform strain rate– Unloading of the material
● Examples:– Moderate strain rate: Taylor Impact Tests– High strain rate: Flyer Plate Impact Tests– Expansion of Copper and Steel Rings.
● The problem geometry is simple – May be 1-D or 2-D.
Validation Process : Step 3
● Coupling constitutive models with damage and failure– Taylor Impact Tests with Failure– Flyer Plate Impact Tests with Spall– Implosion/Explosion of 2-D cylinder (uniform load)– Impact of ball with plate– Fracture of cylinder due to expansion
● The tests are well characterized and involve simple geometries
Validation Process : Step 4
● Interaction with burn models and fluid-structure coupling (no fire)– Detonation and copper sleeve– Explosive fragmentation – Other tests ?
● The tests are complex but still simple enough to be well characterized in certain aspects.
● The final step is the full Monty.
Validation Tests: Copper
● Why copper ?– Well characterized and relatively simple material– A lot of experimental data – Useful for all the validation steps
● Verification and Validation steps– 1-D verification of plastic models– Taylor impact tests– Flyer plate tests (expt. mostly 6061-T6 Al )– Exploding ring (expt. mostly Al and Steel)– Detonation (Jim's simulations)
Verification Tests: Copper (JC)
Verification Tests: Copper (SCG)
Verification Tests: Copper (ZA)
Verification Tests: Copper (MTS)
Verification Tests: Copper (PTW)
1-D Validation Tests: Copper (JC)
1-D Validation Tests: Copper (SCG)
1-D Validation Tests: Copper (ZA)
1-D Validation Tests: Copper (MTS)
1-D Validation Tests: Copper (PTW)
1-D Tests: Summary
● PTW models data best● Stage IV hardening not captured by any
model● What happens during an overdriven shock ?
2-D Taylor Tests: Cu Profile 1
2-D Taylor Tests: Cu Profile 2
2-D Taylor Tests: Cu Profile 3
2-D Taylor Tests: Cu Energy 1
2-D Taylor Tests: Cu Energy 2
2-D Taylor Tests: Cu Energy 3
2-D Taylor Tests: Metrics
2-D Taylor Tests: Metrics Cu 1
2-D Taylor Tests: Metrics Cu 2
2-D Taylor Tests: Metrics Cu 3
2-D Taylor Tests: Summary
● Low temperature tests predicted well by Johnson-Cook.
● Higher temperature tests predicted best by Preston-Tonks-Wallace.
● Validation metrics should include different moments of the geometry – centroid, moment of inertia and higher order moments.
4340 Steel Validation Tests: Cp
4340 Steel Validation Tests: Tm
4340 Steel Validation Tests: µ
4340 Steel Validation Tests: EOS
4340 Steel Validation: JC Rc 30
4340 Steel Validation: JC Rc 38
4340 Steel Validation: JC Rc 45
4340 Steel Validation: JC Rc 49
4340 Steel Validation: MTS Rc 30
4340 Steel Validation: MTS Rc 38
4340 Steel Validation: MTS Rc 45
4340 Steel Validation: MTS Rc 49
4340 Steel Validation: Summary
● MTS fits data better overall.● Easier to determine parameters for MTS.● All submodels perform reasonably well.
4340 Steel Predictions: Rc 38 1000/s
Future Work
● Fit parameters for 4340 steel for PTW, SCG, ZA.
● Get parameters for 6061-T6 Al. Needed for a number of validation tests which are not available for 4340 steel. Anup's thesis.
● Validate damage and failure models.● Develop extended FEM-like fracture model to
create cracks – current approach does not do too well.
● Lot's of work .. so little time !