Hostname: page-component-7b7567c8c-z4zkx Total loading time: 0 Render date: 2026-11-16T08:04:04.582Z Has data issue: false hasContentIssue false

Energy flux enhancement, intermittency and turbulence via Fourier triad phase dynamics in the 1-D Burgers equation

Published online by Cambridge University Press:  06 July 2018

Brendan P. Murray
Affiliation:
Institute for Discovery, Department of Mathematics and Statistics, University College Dublin, Belfield D4, Ireland
Miguel D. Bustamante*
Affiliation:
Institute for Discovery, Department of Mathematics and Statistics, University College Dublin, Belfield D4, Ireland
*
†Email address for correspondence: miguel.bustamante@ucd.ie

Abstract

We present a theoretical and numerical study of Fourier-space triad phase dynamics in the one-dimensional stochastically forced Burgers equation at Reynolds number $Re\approx 2.7\times 10^{4}$. We demonstrate that Fourier triad phases over the inertial range display a collective behaviour characterised by intermittent periods of synchronisation and alignment, reminiscent of the Kuramoto model (Chemical Oscillations, Waves, and Turbulence, Springer, 1984) and directly related to collisions of shocks in physical space. These periods of synchronisation favour efficient energy fluxes across the inertial range towards small scales, resulting in strong bursts of dissipation and enhanced coherence of the Fourier energy spectrum. The fast time scale of the onset of synchronisation relegates energy dynamics to a passive role: this is further examined using a reduced system with the Fourier amplitudes fixed in time – a phase-only model. We show that intermittent triad phase dynamics persists without amplitude evolution and we broadly recover many of the characteristics of the full Burgers system. In addition, for both full Burgers and phase-only systems the physical-space velocity statistics reveals that triad phase alignment is directly related to the non-Gaussian statistics typically associated with structure-function intermittency in turbulent systems.

Information

Type
JFM Papers
Copyright
© 2018 Cambridge University Press 
Figure 0

Figure 1.Subset of time evolution of global system quantities, for the simulation under large spatial scale stochastic forcing ($k_{f}\in [1,3]$). Highlighted is the time range $[13.3,14.0]$, where the time approach to a synchronisation event will be studied in figure 2. (a) Energy flux $\unicode[STIX]{x1D6F1}[k=750](t)$ (magenta) at the high-wavenumber end of the inertial range. Total energy of the system $E(t)$ (red) and total energy dissipation rate $\unicode[STIX]{x1D716}(t)$ (black). (b) Triad phase order parameter $R(t)$ (blue) and $\unicode[STIX]{x1D6F7}(t)$ (green) for triad phases in the inertial range.

Figure 1

Figure 2.Analysis of system dynamics on approach to a synchronisation event ($R\rightarrow 1$), for the simulation under large spatial scale stochastic forcing ($k_{f}\in [1,3]$). (a) Zoomed time series of highlighted region in figure 1 showing the sharp jump in dissipation and energy flux when triad phases become synchronised. (b) $k^{2}$-scaled energy spectra $E_{k}$ at six snapshots (shifted for visualisation purposes). (c) Snapshots of PDF over all inertial-range triad phases $\unicode[STIX]{x1D711}_{k_{1},\,k_{2}}^{k_{3}}$. (d) Real-space solution snapshots $u(x,t)$.

Figure 2

Figure 3.Statistics over full statistically stationary state, for the simulation under large spatial scale stochastic forcing ($k_{f}\in [1,3]$). (a) Conditional time PDF of total dissipation $\unicode[STIX]{x1D716}(t)$ based on value of phase order parameter $R$. (b) PDF of ${\dot{a}}_{k}$ for a range of conditional time periods. (c) Structure functions $S^{p}(r)$ as a function of $r$ and (inset) associated local slopes $\unicode[STIX]{x1D701}_{p}(r)$. (d) Conditional time PDFs of the normalised velocity increments $\unicode[STIX]{x1D6FF}_{r}v/\langle (\unicode[STIX]{x1D6FF}_{r}v)^{2}\rangle ^{1/2}$ in the inertial range at a scale $k=450$ or $r=0.013$.

Figure 3

Figure 4.Subset of time evolution of global system quantities, for the simulation under extended spatial scale stochastic forcing ($k_{f}\in [1,12]$). Highlighted is the time range $[11.6,11.85]$, where the time approach to a jump in synchronisation event will be studied in figure 5. (a) Energy flux $\unicode[STIX]{x1D6F1}[k=150](t)$ (red), $\unicode[STIX]{x1D6F1}[k=600](t)$ (cyan) at each end of the inertial range. Total energy of the system $E(t)$ (magenta) and total energy dissipation rate $\unicode[STIX]{x1D716}(t)$ (black). (b) Triad phase order parameter $R(t)$ (blue) for triad phases in the inertial range.

Figure 4

Figure 5.Analysis of system dynamics on approach to a partial synchronisation event ($R\rightarrow 0.85$), for the simulation under extended spatial scale stochastic forcing ($k_{f}\in [1,12]$). (a) Zoomed time series of highlighted region in figure 4 showing the sharp jump in dissipation and energy flux when the triad phase synchronisation parameter $R$ increases. (b) $k^{2}$-scaled energy spectra $E_{k}$ at four snapshots (shifted for visualisation purposes). (c) Snapshots of PDF over all inertial-range triad phases $\unicode[STIX]{x1D711}_{k_{1},k_{2}}^{k_{3}}$. (d) Real-space solution snapshots $u(x,t)$.

Figure 5

Figure 6.Joint PDF, over all times $t$ of our statistically stationary simulation with stochastic forcing over wavenumbers $k_{f}\in [1,12]$, between the normalised time derivative of $R$ $(|(\text{d}/\text{d}t)(R)/R|)$ at time $t$ and the normalised time derivative of total system dissipation $|(\text{d}/\text{d}t)(\unicode[STIX]{x1D716})/E|$ at the slightly later time $t+0.03\unicode[STIX]{x1D70F}_{\unicode[STIX]{x1D702}}$, where $\unicode[STIX]{x1D70F}_{\unicode[STIX]{x1D702}}$ is the Kolmogorov time microscale.

Figure 6

Figure 7.Phase-only model data and statistics. (a) Time evolution of triad phase order parameter $R(t)$ (blue) and $\unicode[STIX]{x1D6F7}(t)$ (green) for inertial-range triads. (b) Change in the PDF of triad phase as underlying spectrum slope is modified. (c) Structure functions $S^{p}(r)$ as a function of $r$ and (inset) associated local slopes $\unicode[STIX]{x1D701}_{p}(r)$. (d) Conditional time PDFs of normalised velocity increments $\unicode[STIX]{x1D6FF}_{r}v/\langle (\unicode[STIX]{x1D6FF}_{r}v)^{2}\rangle ^{1/2}$ at inertial-range scale $k=450$.

Figure 7

Figure 8.Stability study of fixed point $\unicode[STIX]{x1D719}_{k}=\unicode[STIX]{x03C0}/2,k=1,\ldots ,N$. For resolutions $N=128,256,512$ and $1024$ (red circles, blue triangles, orange rhombs and magenta squares respectively), plots of the proportion of unstable eigenvalues as a function of the spectrum exponent $\unicode[STIX]{x1D6FC}$ in $a_{k}\propto |k|^{-\unicode[STIX]{x1D6FC}/2}$. A sharp transition is observed at a value of $\unicode[STIX]{x1D6FC}$ that depends on the resolution $N$.

Figure 8

Figure 9.Stability study of fixed point $\unicode[STIX]{x1D719}_{k}=\unicode[STIX]{x03C0}/2,k=1,\ldots ,N$. Same parameters and colours as in figure 8. This is a more detailed study of the transition values of exponent $\unicode[STIX]{x1D6FC}$ at each resolution. Solid lines represent the results of nonlinear fits done for each resolution separately, using all points that are shown in the figure. The fits are of the form $\text{ratio}=C(N)(\unicode[STIX]{x1D6FC}_{c}(N)-\unicode[STIX]{x1D6FC})^{z(N)},\unicode[STIX]{x1D6FC}\leqslant \unicode[STIX]{x1D6FC}_{c}(N)$, where $C(N),\unicode[STIX]{x1D6FC}_{c}(N)$ and $z(N)$ are fit parameters depending on the resolution $N$.

Figure 9

Figure 10.Plots, for several values of spectrum exponent $\unicode[STIX]{x1D6FC}$ near the transition value $\unicode[STIX]{x1D6FC}_{c}$ (red circles, $\unicode[STIX]{x1D6FC}=1.566$; blue triangles, $\unicode[STIX]{x1D6FC}=1.569$; orange rhombs, $\unicode[STIX]{x1D6FC}=1.572$; magenta squares, $\unicode[STIX]{x1D6FC}=1.575$; green inverted triangles, $\unicode[STIX]{x1D6FC}=1.578$), of the real part of the non-zero eigenvalues of the Jacobian matrix with the largest real part, centred about the eigenvalue zero for the purpose of better presentation. There are 1023 eigenvalues and only the largest ones are shown. The lowest eigenvalues (not shown) take values as low as $-14\,000$.

Supplementary material: File

Murray and Bustamante supplementary material

Murray and Bustamante supplementary material 1

Download Murray and Bustamante supplementary material(File)
File 1.5 MB