[Paper Review] Continuous data assimilation for the magnetohydrodynamic equations in 2D using one component of the velocity and magnetic fields

Animikh Biswas, Joshua Hudson+2 more|arXiv (Cornell University)|Apr 7, 2017

Advanced Mathematical Modeling in Engineering47 references3 citations

TL;DR

This paper proposes continuous data assimilation algorithms for the 2D magnetohydrodynamic (MHD) equations using feedback control based on partial observations—specifically, one component of the velocity or magnetic field. It proves exponential convergence of the assimilated solution to the true solution in both $L^2$ and $H^1$ norms under suitable conditions on the control parameter and data resolution, even when feedback is applied to only a single component or horizontal variables.

ABSTRACT

We propose several continuous data assimilation (downscaling) algorithms based on feedback control for the 2D magnetohydrodynamic (MHD) equations. We show that for sufficiently large choices of the control parameter and resolution and assuming that the observed data is error-free, the solution of the controlled system converges exponentially (in $L^2$ and $H^1$ norms) to the reference solution independently of the initial data chosen for the controlled system. Furthermore, we show that a similar result holds when controls are placed only on the horizontal (or vertical) variables, or on a single Elsässer variable, under more restrictive conditions on the control parameter and resolution. Finally, using the data assimilation system, we show the existence of abridged determining modes, nodes and volume elements.

Motivation & Objective

• To develop continuous data assimilation methods for the 2D MHD equations using partial observations.
• To establish exponential convergence of the assimilated solution to the true solution in $L^2$ and $H^1$ norms.
• To show that feedback control on a single component (e.g., horizontal velocity or one Elsässer variable) still ensures convergence, albeit under more restrictive conditions.
• To connect data assimilation with determining modes, nodes, and volume elements in MHD systems.
• To extend the feedback control framework from Navier-Stokes to MHD equations with rigorous convergence analysis.

Proposed method

• Uses feedback control at the PDE level to correct the solution of the MHD system using partial observations.
• Applies the continuous data assimilation framework pioneered by Azouani, Olson, and Titi to the 2D MHD equations.
• Employs Elsässer variables to simplify the analysis and decouple the system dynamics.
• Imposes feedback control only on one component of velocity or magnetic field, or on a single Elsässer variable.
• Uses energy estimates and logarithmic Sobolev inequalities to bound nonlinear terms in the error dynamics.
• Derives a priori estimates in $L^2$ and $H^1$ norms to establish exponential convergence rates.

Experimental results

Research questions

• RQ1 Can continuous data assimilation with feedback control on a single component of velocity or magnetic field ensure exponential convergence to the true solution of the 2D MHD equations?
• RQ2 What are the minimal observational requirements (in terms of resolution and control parameter) for exponential convergence when only partial data is available?
• RQ3 How does the feedback control framework for MHD compare to existing Kalman filtering or variational methods in terms of convergence guarantees?
• RQ4 Can data assimilation systems be used to rigorously establish the existence of determining modes, nodes, or volume elements in MHD?
• RQ5 What conditions on the control parameter and data spacing ensure $H^1$-norm convergence in the presence of partial observations?

Key findings

• Exponential convergence in the $L^2$-norm is established for the feedback-controlled system when all variables are observed, under suitable conditions on the control parameter and data resolution.
• When feedback is applied only to the horizontal components of velocity and magnetic fields, exponential $L^2$-norm convergence holds under more restrictive conditions on the control parameter and resolution.
• Exponential convergence in the $H^1$-norm is proven for the full system under standard assumptions, and also under weaker assumptions on the interpolation function when only horizontal components are observed.
• The paper establishes the existence of abridged determining modes, nodes, and volume elements using the data assimilation framework.
• The convergence rate is exponential, with the error decaying as $e^{- u t}$ for some $ u > 0$, depending on the control parameter and system parameters.
• The results are robust to initial data uncertainty, as convergence is independent of the initial condition of the controlled system.