Reducing uncertainties in a wind-tunnel experiment using Bayesian updating

Reducing uncertainties in a wind-tunnel experiment using Bayesian updating

Boon, D.J.
Dwight, R.P.
Sterenborg, J.J.H.M.
Bijl, H.

Aerospace Engineering

Aerodynamics, Wind Energy & Propulsion

2012-04-23

We perform a fully stochastic analysis of an experiment in aerodynamics. Given estimated uncertainties on the principle input parameters of the experiment, including uncertainties on the shape of the model, we apply uncertainty propagation methods to a suitable CFD model of the experimental setup. Thereby we predict the stochastic response of the measurements due to the experimental uncertainties. To reduce the variance of these uncertainties a Bayesian updating technique is employed in which the uncertain parameters are treated as calibration parameters, with priors taken as the original uncertainty estimates. Imprecise measurements of aerodynamic forces are used as observational data. Motivation and a concrete application come from a wind-tunnel experiment whose parameters and model geometry have substantial uncertainty. In this case the uncertainty was a consequence of a poorly constructed model in the pre-measurement phase. These methodological uncertainties lead to substantial uncertainties in the measurement of forces. Imprecise geometry measurements from multiple sources are used to create an improved stochastic model of the geometry. Calibration against lift and moment data then gives us estimates of the remaining parameters. The effectiveness of the procedure is demonstrated by prediction of drag with uncertainty.

http://resolver.tudelft.nl/uuid:505a42b5-a5ef-49f7-b6a5-b8c50e1f45d1

https://doi.org/10.2514/6.2012-1856

American Institute of Aeronautics and Astronautics (AIAA)

978-1-60086-937-2

53rd AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference, Honolulu, Hawaii, 23-26 April 2012; AIAA 2012-1856

Institutional Repository

conference paper

© 2012 Dwight, R.P.