Search

From the energy balance in wavenumber space expressed by the Lin equation, we derive a new form for the local Karman-Howarth equation for forced isotropic turbulence in real space. This equation is then cast into a dimensionless form, from which a combined analytical and numerical study leads us to deduce a new model for the scale-independent nondimensional dissipation rate $\Ceps$, which takes the form $\Ceps = \Cinf + C_L/R_L$, where the asymptotic value $\Cinf$ can be evaluated from the third-order structure function. This is found to fit the numerical data with $\Cinf = 0.47 \pm 0.01$ and $C_L= 18.5 \pm 1.3$. By considering $\Ceps - \Cinf$ on logarithmic scales, we show that $R_L^{-1}$ is indeed the correct Reynolds number behaviour. The model is compared to previous attempts in the literature, with encouraging agreement. The effects of the scale-dependence of the inertial and viscous terms due to finite forcing are then considered and shown to compensate one another, such that the model equation is applicable for systems subject to finite forcing. In addition, we also show that, contrary to the case of freely decaying turbulence, the characteristic decline in $\Ceps$ with increasing Reynolds number is due to the \emph{increase} in the surrogate expression $U^3/L$; the dissipation rate being maintained constant as a consequence of the fixed rate of forcing. A long-time non-turbulent stable state is found to exist for low Reynolds number numerical simulations which use negative damping as a means of energy injection.

In order to analyse results for structure functions, $S_n(r)$, we propose plotting the ratio $|S_n(r)/S_3(r)|$ against the separation $r$. This method differs from the extended self-similarity (ESS) technique, which plots $S_n(r)$ against $S_3(r)$, where $S_3(r) \sim r$. Using this method in conjunction with pseudospectral evaluation of structure functions, for the particular case of $S_2(r)$ we obtain the new result that the exponent $\zeta_2$ decreases as the Taylor-Reynolds number increases, with $\zeta_2 \to 0.67 \pm 0.02$ as $R_\lambda \to \infty$. This supports the idea of finite-viscosity corrections to the K41 prediction for $S_2$, and is the opposite of the result obtained by ESS.

A parallel pseudospectral code for the direct numerical simulation (DNS) of isotropic turbulence has been developed. The code has been extensively benchmarked using established results from literature. The code has been used to conduct a series of runs for freely-decaying turbulence. We explore the use of power-law decay of the total energy to determine an evolved time and compare with the use of dynamic quantities such as the peak dissipation rate, maximum transport power and velocity derivative skewness. Stationary turbulence has also been investigated, where we ensure that the energy input rate remains constant for all runs. We present results for Reynolds numbers up to R{\lambda} = 335 on a 1024^3 lattice. An exploitation of the pseudospectral technique is used to calculate second and third-order structure functions from the energy and transfer spectra, with a comparison presented to the real-space calculation. An alternative to ESS is discussed, with the second-order exponent found to approach 2/3. The dissipation anomaly is considered for forced and free-decay. The K\'arm\'an-Howarth equation (KHE) is studied and a derivation of a new work term presented. The balance of energy represented by the KHE is then investigated. Based on the KHE, we develop a model for the behaviour of the dimensionless dissipation coefficient that predicts C{\epsilon} = C{\epsilon}(\infty) + C_L/R_L, with C{\epsilon}(\infty) = 0.47 and C_L = 19.1 obtained from DNS data. Theoretical methods based on RG and statistical closures are still being developed to study turbulence. The dynamic RG procedure used by Forster, Nelson and Stephen (FNS) is considered in some detail and a disagreement in the literature is resolved here. The application of statistical closure and renormalized perturbation theory is discussed and a new two-time model probability density functional presented., Comment: 292 pages; Ph.D thesis, University of Edinburgh, UK