Akademska digitalna zbirka SLovenije - logo
VSE knjižnice (vzajemna bibliografsko-kataložna baza podatkov COBIB.SI)
  • Visualization of Karman vortex street for experimental modeling with neural networks [Elektronski vir]
    Hočevar, Marko, strojnik, 1972- ; Širok, Brane ; Novak, Matej, 1968-
    In the present paper a novel method of turbulent field modeling is presented. It is based on computer aided visualization and radial basis neural networks. Karman vortex street was used in the ... experiment. Concentration of passive tracer in the vortex street was measured with visualization method. Using radial basis neural networks concentration was modeled at several locations in the vortex street behind the bluff body. From the measured and modeled concentration power spectra and spatial correlation functions were calculated. Both measured and modeled concentration spectra exhibit basic peak of fluctuation power at the same frequency as measured power spectra. Measured and modeled power spectra also correspond well to the -5/3 turbulence decay law. The quality of modeling of the concentration spectra decreases with the growing distance between input and output area of the neural network. Agreement between predicted and measured concentration correlation functions is very good.
    Vir: CD Rom proceedings [Elektronski vir] (Str. 67-82)
    Vrsta gradiva - prispevek na konferenci
    Leto - 2002
    Jezik - angleški
    COBISS.SI-ID - 4952347

Vnos na polico