Aerodynamic Modeling with Heterogeneous Data Assimilation and Uncertainty Quantification

Clear Science Corp. proposes to develop an aerodynamic modeling tool that assimilates data from different sources and facilitates uncertainty quantification. The technical merit and feasibility of the technology will be demonstrated in Phase I through a series of verification and validation tests that utilize both computational and wind tunnel data in constructing aerodynamic models for the Orion launch abort system (LAS). Aerodynamic models provide inputs to the guidance, navigation, and control system. The proposed software will enable performance predictions over a wide range of operational conditions through the fusion of data from multiple sources including high-dimensional computational simulations, wind tunnel tests, and flight tests. The software will also facilitate uncertainty analyses to determine the propagation of variability in inputs into output variability and sensitivity analyses to identify critical design and modeling parameters and operational variables. Complex systems like the LAS are designed with a mixture of heterogeneous data, and uncertainties in the data can be a critical factor in evaluating designs. The objective is to develop assimilation methods that reduce the number of expensive wind tunnel tests and CFD simulations required during system design while maintaining and improving the quality of aerodynamic models and systematically assessing uncertainties.