A capability is presented for updating a non-linear dynamic numerical model using test data in automotive crash applications. A development has also been completed for providing a variability assessment for response quantities of interest based on uncertainties which exist in the structures and the loads. Principal component decomposition is utilized to form an expansion basis from nonlinear structural response data. Metamodels are developed for the principal components by employing the Kriging method. The metamodels of the principal components are then utilized for generating the fast running models. The latter are employed in the MUVE process and for evaluating the statistical correlation metric. The fast running models are used instead of an actual simulation analysis for computing the responses of interest. This allows expediting significantly the MUVE process and making it feasible for analyzing large models in practical applications. A statistical correlation metric has been developed based on identifying the test data as part of the family of simulation models. The fast running models are also used in this process since a large number of evaluations for the numerical results is required when constructing the statistical information for the family of simulation models. The developed capabilities are demonstrated in an automotive crash simulation. Adjustments in the numerical models can improve significantly the correlation to the test results. In addition, variability bounds are introduced in the nonlinear response of the vehicle structure. The use of the fast running models allows to readily use existing non-linear finite element models and solvers for probabilistic analysis.
Model update and variability assessment for automotive crash simulations
Modellaktualisierung und Variabilitätsbewertung für Automobil-Crashsimulationen
2007
9 Seiten, 4 Bilder, 1 Tabelle, 20 Quellen
Aufsatz (Konferenz)
Englisch
Model update and statistical correlation metrics for automotive crash simulations
Kraftfahrwesen | 2007
|Model Update and Statistical Correlation Metrics for Automotive Crash Simulations
SAE Technical Papers | 2007
|Model Update and Statistical Correlation Metrics for Automotive Crash Simulations
British Library Conference Proceedings | 2007
|2007-01-1744 Model Update and Statistical Correlation Metrics for Automotive Crash Simulations
British Library Conference Proceedings | 2007
|GPU-based interactive, stereoscopic visualization of automotive crash simulations
Tema Archiv | 2007
|