Kinematic dynamo in random flows

Publication date 25 June 1992

diffusion eigenvalue structure fast dynamo invariant measure kinematic equations large deviations theory Lyapunov exponent magneto-hydrodynamics stochastic excitation

Mathematics Subject Classification ID

Stochastic partial differential equations (aspects of stochastic analysis) (60H15) Stability and instability of magnetohydrodynamic and electrohydrodynamic flows (76E25)

Review text

The kinematic equations of magneto-hydrodynamics with external stochastic excitation may be written in the form \[ \partial y^ \varepsilon/\partial t=(\varepsilon^ 2/2)\Delta_ x y^ \varepsilon- v(x,\xi(t))\cdot\nabla_ xy^ \varepsilon+(\partial v/\partial x)(x,\xi(t))y^ \varepsilon-\hbox{div} v(x,\xi(t))y^ \varepsilon, \] with \(x\in\mathbb{R}^ n\), \(y^ \varepsilon\mid_{t=0}=y_ 0(x)\) and \(\hbox{div} y_ 0=0\), where \(\xi(t)\) is a diffusion process on a compact manifold. Without diffusion term, i.e. for \(\varepsilon=0\), the solution can be written as \[ \begin{gathered} y^ 0(\gamma(t)x,t)=\Theta(\gamma(t)x,t)y_ 0(x), \text{ with } (d/dt)\gamma(t)x=v(\gamma(t)x,\xi(t))\\ \text{and } (\partial/\partial t)\Theta=[(\partial v/\partial x)(\gamma(t)x)-\hbox{div} v(\gamma(t)x)]\Theta,\;\Theta\mid_{t=0}=\text{Id}.\tag{*}\end{gathered} \] The existence of a fast dynamo for small magnetic diffusivity is shown in the following way: One first proves that (*) has a positive Lyapunov exponent. This is done via moment exponents and large deviations theory for the a.s. exponent. For linear velocity fields \(v\) the Arnold/Kliemann theory is sufficient to guarantee positivity under suitable assumptions on the interactions between \(\xi\) and \(v\). If \(v\) is nonlinear, the author employs Lyapunov function techniques to assure that the \((\gamma,\xi,\theta)\)-process has a unique invariant measure (here \(\theta\) is the projection of \(\Theta\) onto the projective space). Using the Donsker/Varadhan large deviations theory, he then shows positivity under assumptions on the eigenvalue structure of the operator associated with \((\gamma,\xi.\theta)\). The second step consists of proving that positivity of the Lyapunov exponent is preserved for small \(\varepsilon>0\). The author shows this for the norm \(|\| u^ \varepsilon(\cdot,t)\||^ 2\), with \(|\|\cdot\||^ 2=E\|\cdot\|_ 2^ 2\), where \(\|\cdot\|_ 2\) is the \(L^ 2\)-norm on \(\mathbb{R}^ n\), and \(E\) is the expectation with respect to the Wiener measure. For linear velocities he obtains \(\lim_{\varepsilon\downarrow 0}\lim_{t\to\infty}(1/t)\log|\| y^ \varepsilon\||^ 2>0\), which establishes a fast dynamo effect in the sense of Molchanov et al. For nonlinear velocity fields a slightly weaker version is proved. These results are remarkable, their usefulness, however, is somewhat diminished by the fact that various assumptions are not given directly in terms of the data \(v\) and the vector fields defining the stochastic excitation \(\xi(t)\).

Recommendations

Cited in

(27)

This page was built for publication: Kinematic dynamo in random flows

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3978798)