Augmented Prediction of Turbulent Flows via Sequential Estimators
Among the numerous research aspects in the analysis of complex flow configurations of industrial interest, the accurate prediction of turbulent flows is one of the ultimate open challenges. Investigation via classical tools, such as experiments and numerical simulation, is difficult because of fundamental drawbacks which can not be completely excluded. Experiments provide a local description of flow dynamics via measurements sampled by sensors. A complete reconstruction of the flow behavior in the whole physical domain is problematic because of the non-linear, strongly inertial behavior of turbulence. While reduced-order models, such as POD (Lumley, Stochastic tools in turbulence. Academic Press, New York, 1970, ), have been extensively used for this purpose, they usually provide an incomplete reconstruction of turbulent flows for the aforementioned reasons.