| Define “Validation Metric ” by What You Want to Do with It (2007) | |||||||||||||
Abstract | |||||||||||||
| • See [O&B 2006] for CFD perspective. • Quantitative approach to code calibration, • Determine where model & experiment agree well & where they don’t, – Know where to focus future effort. • Test hypotheses, – E.g., does one model match experiment better than another? • Quantitatively characterize uncertainty in model – experiment comparison, – Is resulting confidence interval acceptable? – Estimate uncertainty in predictive simulations. Factors Desired in Validation Metrics • Experimental measurement uncertainties, – Including those introduced in post-processing. • Uncertainties in code inputs, • Code errors. – E.g., inadequate spatial resolution. • Number of experiments, • “Primacy ” of variables used for comparison. Example Validation Metrics: [O&B 2006], [O&T 2002] | |||||||||||||
Details der Publikation | |||||||||||||
| |||||||||||||