Direct numerical simulations of decaying isotropic turbulence with a passive scalar at different Schmidt numbers show that scaling on Kolmogorov and Batchelor variables provides a reasonable collapse at large wavenumbers of the energy and scalar spectra as well as the energy and scalar transfer functions. There appears to be an approach to a k(-1) viscous-convective range at small R. and for Sc not much larger than 1. To gain some information at larger R-lambda (= 200) and Sc (>1), EDQNM ( eddy damped quasi-normal Markovian) simulations are used, after validation, at small R-lambda (= 200) and Sc (= 1, 3), against the DNS. The EDQNM model reproduces, at least qualitatively, the difference between the inertial-convective and viscous-convective ranges, although the magnitudes of the corresponding power-law exponents are -1.5 and -1.1, respectively, which differ slightly from the expected asymptotic values of -5/3 and -1. The probability density function of some of the terms in the transport equation for theta(2) provides an indication of how large radius of curvature scalar gradients may be formed.
Dependence of a passive scalar in decaying isotropic turbulence on the Reynolds and Schmidt numbers using the EDQNM model / Orlandi, Paolo; R. A., Antonia. - In: JOURNAL OF TURBULENCE. - ISSN 1468-5248. - 5:(2004), pp. 9-9. [10.1088/1468-5248/5/1/009]
Dependence of a passive scalar in decaying isotropic turbulence on the Reynolds and Schmidt numbers using the EDQNM model
ORLANDI, Paolo;
2004
Abstract
Direct numerical simulations of decaying isotropic turbulence with a passive scalar at different Schmidt numbers show that scaling on Kolmogorov and Batchelor variables provides a reasonable collapse at large wavenumbers of the energy and scalar spectra as well as the energy and scalar transfer functions. There appears to be an approach to a k(-1) viscous-convective range at small R. and for Sc not much larger than 1. To gain some information at larger R-lambda (= 200) and Sc (>1), EDQNM ( eddy damped quasi-normal Markovian) simulations are used, after validation, at small R-lambda (= 200) and Sc (= 1, 3), against the DNS. The EDQNM model reproduces, at least qualitatively, the difference between the inertial-convective and viscous-convective ranges, although the magnitudes of the corresponding power-law exponents are -1.5 and -1.1, respectively, which differ slightly from the expected asymptotic values of -5/3 and -1. The probability density function of some of the terms in the transport equation for theta(2) provides an indication of how large radius of curvature scalar gradients may be formed.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.