This would be computationally demanding!
Nevertheless, currently we tend to determine Material Property Constitutive relations by fitting simple uniaxial isothermal experimental test results (to measurements such as stress and/or strain and endurance or life).
More complex experimental test results such as thermo mechanical fatigue ( TMF ), complex geometries such as notched bars, crack growth tests and real-plant measurements/experience tend to be used for Validation and Verification of those Material Property Constitutive relations and Damage Mechanics models etc...
How many 10s or 100s of years will it be before we could input all simple and complex, experimental tests and all real-plant experience into a set of Finite Element Models with all known inputs and all known measurements and actually solve the Material Property Constitutive relations and Damage Mechanics models so that "the right result (with realistic error distribution) is predicted in all cases"?
I would suggest that so long as a selection of (more than 3) different Material Property Constitutive relations are trialled then Success can be claimed for the best model tried.
- Badges
- Science topic