Estimating the error in simulation prediction over the design space
- Rachel
- François M.
- Scott W.
This study addresses the assessrnent of accuracy of simulation predictions. A procedure is developed to validate a simple non-linear model defined to capture the hardening behavior of a foam material subjected to a short-duration transient impact. Validation means that the predictive accuracy of the model must be established, not just in the vicinity of a single testing condition, but for all settings or configurations of the system. The notion of validation domain is introduced to designate the design region where the model's predictive accuracy is appropriate for the application of interest. Techniques brought to bear to assess the model's predictive accuracy include test-analysis coi-relation, calibration, bootstrapping and sampling for uncertainty propagation and metamodeling. The model's predictive accuracy is established by training a metalnodel of prediction error. The prediction error is not assumed to be systcmatic. Instead, it depends on which configuration of the system is analyzed. Finally, the prediction error's confidence bounds are estimated by propagating the uncertainty associated with specific modeling assumptions.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE
- OSTI ID:
- 976579
- Report Number(s):
- LA-UR-03-2152; TRN: US201017%%725
- Resource Relation:
- Conference: Submitted to: 44th AIAA/ASME/ASCE/AHS Structures, Structural Dynamics and Materials Conference, Norfolk, Virginia, April 7-10, 2003
- Country of Publication:
- United States
- Language:
- English
Similar Records
TU-AB-202-03: Prediction of PET Transfer Uncertainty by DIR Error Estimating Software, AUTODIRECT
The myth of science-based predictive modeling.