09 Feb Quantifying Numerical Uncertainty in Background – Oriented Schlieren
In Background – Oriented Schlieren (BOS), distortions of a pattern are used to characterize flow fields containing refractive gradients. In the process, they can lead to some uncertainty. We proposed a method to quantify one aspect of uncertainty – that introduced when the density gradient field obtained from the displacement of the background pattern is integrated to obtain the density field. We define numerical uncertainty using the Richardson extrapolation, which has been used for grid convergence studies in computational fluid mechanics. The method was tested on synthetic noisy density fields and achieved up to 91% accuracy when the numerical uncertainty was combined with the random uncertainty. We also included a demonstration of our methodology by applying it to experimental BOS images.
No Comments