Zur JKU Startseite
Many industrial (and environmental) processes are governed by the underlying flow phenomena. Conventional CFD simulations of such flows are typically extremely expensive. In case the flow under consideration exhibits recurring flow features, we propose to switch towards a more efficient statistical description of flow. With recurrence CFD we can accelerate flow simulations by four orders of magnitude. In many cases this speed-up enables real-time simulations of flows at high resolution.
In the picture two snap-shots of species concentration (in this case we consider hydrogen concentration in steel) are given. While a conventional CFD simulation (top) took 26 hours of computational wall-clock time, a rCFD simulation (bottom) took only 1.6 seconds although the results are obtained at the same (high spatial and temporal) resolution. Actually, in this case rCFD runs more than an order of magnitude faster than real-time.
Hydrogen removal in a RH plant for vacuum degassing of steel
Prediction of heat transfer in fluidized beds via recurrence CFD
particulate and gas phase takes much longer. Employing recurrence CFD, one can easily extrapolate their motion using results from conventional CFD-DEM simulations at low numerical costs.
In systems governed by reappearing structures, one may generate sequences of generic flow patterns of arbitrary length using recurrence statistics. Passive or weakly coupled processes such as heat transfer may then be studied on the extrapolated fields, which represents a strong decoupling of fast from slow degrees of freedom. Speed-up factors of about 300 are achieved at minor accuracy impairments. It is shown in Figs. 1 and 2 that while the particle temperature distribution turns out to be a bit too sharp in comparison to CFD-DEM calculations, the average values agree very well.
Most surprisingly, even after 60 s, heat transfer was pictured correctly with fields from the first 2.5s. With how little information can we get away to describe complex flows?
References:
Patil, A. V. et al. Chem. Eng. J. 277 (2015): 388-401
Lichtenegger, T. and Pirker, S. in: 12th International Conference on CFD in the Oil & Gas, Metallurgical and Process Industries (2017): 47-51
Lichtenegger, T. et al. Chem. Eng. Sci. 172 (2017): 310-322
