Turbulence

The focus of turbulence related research has been the study of calibration, validation and uncertainty quantification in the context of standard RANS turbulence models. In particular, effort has been dedicated towards investigating model structural uncertainty and the resulting contribution to the uncertainty in quantities of interest from turbulence models in flow problems.

The proper incorporation of model structural uncertainty is crucial to accurate uncertainty quantification. In the current approach, the uncertainty due to the physical model structure is modeled via a stochastic extension of the physical model. Future investigation will focus on extending the current approach to include multiple stochastic representations of the model structural uncertainty, which can be compared in a Bayesian fashion using the evidence metric.

• Develop and apply the PECOS Bayesian validation methodology  to simple turbulence models such as the Spalart-Allmaras, k-omega, and k-epsilon
• Investigate effect of model structural uncertainty on quantities of interest in context of turbulent flows
• Investigate using model structural uncertainty in incompressible setting by adding a zero-mean Gaussian random field to the deterministic velocity field that solves the RANS/turbulence model system
• Study using additional stochastic models that will include stochastic representations of the Reynolds stress
• Model will require intrusive modifications to the underlying RANS solver – develop finite element RANS solvers for both incompressible and compressible regimes