Uncertainty Quantification for Structural Dynamics and Model Validation Problems Ben H Thacker, David S Riha, Daniel P Nicolella, Steve J Hudak, Luc J Huyse, Loren Francis Southwest Research Institute 6220 Culebra Road San Antonio, TX 78238 USA Simeon H.K. Fitch Mustard Seed Software 1634 Brandywine Drive Charlottesville, VA 22901 USA Jason E. Pepin and Edward A. Rodriguez Los Alamos National Laboratory Engineering Sciences & Applications Div. P.O. Box 1663, MS P946 Los Alamos, NM 87545 USA ABSTRACT The application of uncertainty quantification methodologies for structural dynamics requires the use of efficient and accurate analysis tools that can predict the uncertainty in a response due to uncertainties in the model formulation and input parameters. Uncertainty quantification techniques such as probabilistic methods are also being used to help develop quantitative model validation strategies. Using four different application problems— automotive crashworthiness, blast containment, spinal injury and space shuttle flowliner—this paper provides a broad overview of some techniques of applying probabilistic methods to computationally expensive structural dynamics simulations. For each application problem, a discussion on the model validation strategy used is also given. KEYWORDS Probabilistic, Structural Dynamics, Uncertainty Quantification, Stochastic, Model Validation 1 INTRODUCTION AND BACKGROUND As performance requirements and testing costs for engineered systems continue to increase, computational simulation is being increasingly relied upon to serve as a predictive tool. To meet these requirements, analysts are developing higher fidelity models in an attempt to accurately represent the behavior of the physical system. It is not uncommon nowadays for these models to involve multiple physics, complex interfaces, and several million finite elements. Despite the recent extraordinary increase in computer power, analyses performed with these high fidelity models continue to take hours or even days to complete for a single deterministic analysis. Structural performance is directly affected by uncertainties associated with models or in physical parameters and loadings. The traditional design approach has been to adopt safety factors to ensure that the risk of failure is sufficiently small, albeit not quantified. However, probabilistic analysis permits a more rigorous quantification of the various uncertainties, and ultimately will facilitate a more efficient design process. Areas in which probabilistic methods are being successfully applied include engineered or naturally occurring systems with high consequences of failure. Some of these areas include aircraft propulsion systems, airframes, biomedical