Important issues in validation of computational fluid dynamics codes against CFD grade experimental results

The continuing improvement in numerical methods and computer technology allows for more detailed and finer velocity and temperature predictions. The validation of such methods requires high resolution experimental results obtained by methods such as Particle Image Velocimetry (PIV) which are often referred to as “CFD-grade” experimental results. Such experiments provide some combination of fine spatial and/or temporal resolution which give rise to differing uncertainty contributions as compared to more conventional and coarse experiments. Some examples of these uncertainties relate to errors in the reported spatial position of the probe, the type and extent of spatial averaging in the measurement, and additional uncertainties that may arise in the application of these fine-resolution measurements. This paper uses a series of staged experiments and supporting CFD calculations as an example of the uncertainty quantification that is necessary when performing CFD grade experiments or validation. The papers examines these effects and shows that measured velocities can be impacted significantly and as a result need to be carefully examined by experimenters and considered for those performing numerical studies. It is noted that there are no general guidelines for examining uncertainty in such experiments as the uncertainties depend strongly on flow geometry, measurement technique and the region of interest. Based on our results, we propose a useful set of high-level guidelines which can be applied to CFD grade experiments to ensure that a robust set of uncertainties is quantified.