---
title: Incompressible Navier–Stokes–Maxwell System
url: https://www.emergentmind.com/topics/incompressible-navier-stokes-maxwell-system
type: topic
---

# Incompressible Navier–Stokes–Maxwell System

Searching arXiv for recent and foundational papers on the incompressible Navier–Stokes–Maxwell system and closely related limits.
arXiv Search Tool unavailable in this environment. Proceeding with the supplied arXiv records and citing them directly.
The incompressible Navier–Stokes–Maxwell system denotes a class of coupled PDEs in which an incompressible viscous charged fluid is linked to Maxwell’s equations through Ohm’s law. In the classical formulation, the unknowns are the fluid velocity \(u\), pressure \(p\), electric field \(E\), magnetic field \(B\), and current \(j\), and the model combines a parabolic incompressible Navier–Stokes equation with a hyperbolic–damped Maxwell subsystem. Across the literature, the term also borders on closely related systems: inhomogeneous variable-density versions, two-fluid formulations, thermally extended Navier–Stokes–Fourier–Maxwell models, and singular limits toward magnetohydrodynamics or toward solenoidal-Ohm-law reductions [1811.01364][1207.6187].

## 1. Canonical equations and constitutive structure

A classical incompressible Navier–Stokes–Maxwell system in space dimension \(d=2\) or \(3\) is written as
\[
\left\{
\begin{aligned}
&\partial_t u + u\cdot\nabla u - \mu\Delta u = -\nabla p + j\times B,\qquad \operatorname{div}u=0,\\
&\frac1c\partial_t E - \nabla\times B = -j,\qquad j=\sigma(cE+u\times B),\\
&\frac1c\partial_t B + \nabla\times E =0,\qquad \operatorname{div}B=0.
\end{aligned}
\right.
\]
Here \(u\) is the fluid velocity, \(E\) the electric field, \(B\) the magnetic field, \(p\) the pressure, \(j\) the current, \(c>0\) the speed of light, \(\mu>0\) the viscosity, and \(\sigma>0\) the conductivity. In this formulation the current is not an independent unknown: it is determined by Ohm’s law. In two dimensions, the fields are embedded into \(\mathbb R^3\) with
\[
u_3=E_3=B_1=B_2=0.
\]
“Incompressible” means \(\operatorname{div}u=0\), and the magnetic field is divergence-free as well, \(\operatorname{div}B=0\) [1811.01364].

A normalized version, used in wellposedness theory, sets the viscosity and conductivity equal to \(1\). The system then takes the form
\[
\left\{
\begin{aligned}
\partial_t v + v\cdot\nabla v - \Delta v + \nabla p &= j\times B,\\
\partial_t E - \operatorname{curl} B + E &= -j,\\
\partial_t B + \operatorname{curl} E &= 0,\\
\operatorname{div} v &= 0,\qquad \operatorname{div} B = 0,\qquad j = E+v\times B.
\end{aligned}
\right.
\]
This representation makes explicit that the Lorentz forcing in the fluid equation is \(j\times B=(E+v\times B)\times B\), while the damping term \(+E\) belongs to the Maxwell operator itself [1207.6187].

A distinct limiting constitutive law appears in the solenoidal-Ohm-law reduction. In that regime the current is given by
\[
j=\sigma(-\nabla \tilde p+cE+u\times B),
\]
with \(\nabla\cdot j=0\), so that the scalar \(\tilde p\) enforces the solenoidal constraint on the current rather than on the velocity field alone [2507.11881].

## 2. Energy structure and analytical difficulties

The natural energy level for the classical system is
\[
u\in L^\infty(\mathbb R^+;L^2)\cap L^2(\mathbb R^+;\dot H^1),\qquad
(E,B)\in L^\infty(\mathbb R^+;L^2),\qquad
j\in L^2(\mathbb R^+;L^2),
\]
together with the energy inequality
\[
\frac12\big(\|u(t)\|_{L^2}^2+\|E(t)\|_{L^2}^2+\|B(t)\|_{L^2}^2\big)
+\int_0^t\Big(\mu\|\nabla u(\tau)\|_{L^2}^2+\frac1\sigma\|j(\tau)\|_{L^2}^2\Big)\,d\tau
\le \mathcal E_0,
\]
where
\[
\mathcal E_0=\frac12\big(\|u_0\|_{L^2}^2+\|E_0\|_{L^2}^2+\|B_0\|_{L^2}^2\big).
\]
This energy level is natural but not by itself sufficient for general weak compactness of the nonlinear Lorentz term \(j\times B\): a central obstacle is that the magnetic field lacks compactness, so finite-energy weak approximations do not permit straightforward passage to the limit in the coupling [1811.01364].

This hyperbolic–parabolic interaction dictates much of the modern analysis. One strategy is to propagate extra regularity of \((E,B)\) in \(\dot H^s\), which yields compactness for \(B\) and thereby permits passage to \(j\times B\). In the large-energy weak-solution theory, this is coupled to refined energy estimates for Maxwell’s system, a Grönwall-like bootstrap, and a maximal parabolic estimate for the heat equation in Besov spaces strong enough to avoid Chemin–Lerner spaces altogether [1811.01364].

The damped character of the Maxwell subsystem is equally important in semigroup-based wellposedness theory. In the normalized formulation, the Maxwell equations imply dissipation for the electric field, and the magnetic field can be rewritten as a damped wave equation,
\[
\partial_{tt}B-\Delta B + \partial_t B = \operatorname{curl}G,
\]
which yields additional decay estimates for \(B\). This structure complements the parabolic smoothing of the Navier–Stokes component and underlies mild-solution frameworks based on the coupled linear semigroup [1207.6187].

## 3. Well-posedness theory and solution classes

The well-posedness theory separates sharply by dimension, regularity class, and whether the density is constant or variable. For the homogeneous incompressible system, a mild-solution theory establishes local existence for arbitrarily large initial data in spaces similar to the scale-invariant spaces classically used for Navier–Stokes, and global existence for sufficiently small data. In dimension \(2\), the theorem is formulated for
\[
(v^0,E^0,B^0)\in L^2(\mathbb R^2)\times L\log L(\mathbb R^2)\times L\log L(\mathbb R^2),
\]
while in dimension \(3\) it is stated for
\[
(v^0,E^0,B^0)\in H^1(\mathbb R^3)^3\times H^1(\mathbb R^3)^3\times H^1(\mathbb R^3)^3.
\]
The proof uses semigroup bounds, Littlewood–Paley theory, Besov and logarithmically weighted spaces, product estimates based on Bony’s decomposition, and a contraction argument for the mild formulation
\[
T(t)=e^{tA}T^0+\int_0^t e^{(t-s)A}N(T(s))\,ds
\]
with \(T=(v,E,B)\) [1207.6187].

A second line of results works at the energy level. In three dimensions, global weak solutions exist under the assumption that the initial velocity and electromagnetic fields have finite energy and that the initial electromagnetic field is small in \(\dot H^s(\mathbb R^3)\) for \(s\in [\frac12,\frac32)\). More precisely, if
\[
(u_0,E_0,B_0)\in L^2(\mathbb R^3)\times (H^s(\mathbb R^3))^2,\qquad \operatorname{div}u_0=\operatorname{div}B_0=0,
\]
and
\[
\|(E_0,B_0)\|_{\dot H^s}\, C_*\,\mathcal E_0^{\,s-\frac12}e^{C_*\mathcal E_0}\le 1,
\]
then there exists a global weak solution satisfying the energy inequality and
\[
E,B\in L^\infty(\mathbb R^+;\dot H^s),\qquad E\in L^2(\mathbb R^+;\dot H^s),\qquad
u\in L^1(\mathbb R^+;\dot B^{3/2}_{2,1})+L^2(\mathbb R^+;\dot B^{3/2}_{2,1}).
\]
In two dimensions, the same analysis yields global weak solutions with
\[
E,B\in L^\infty_{\mathrm{loc}}(\mathbb R^+;\dot H^s),\qquad
u\in L^2_{\mathrm{loc}}(\mathbb R^+;L^\infty),
\]
for \(s\in(0,1)\), together with estimates uniform in the speed of light \(c\) [1811.01364].

For the inhomogeneous planar system, the unknown density \(\rho\) satisfies the transport equation
\[
\partial_t\rho+\operatorname{div}(\rho u)=0,
\]
and the momentum equation becomes
\[
\partial_t (\rho u)+\operatorname{div}(\rho u\otimes u)-\nu\Delta u+\nabla p=j\times B.
\]
Global energy solutions are obtained in two dimensions when the initial density is bounded pointwise, bounded away from \(0\), and close to the constant state \(1\); the velocity satisfies
\[
u_0 \in \dot{B}^{-1+\frac{2}{p}}_{p,1}(\mathbb R^2),\qquad p\in(1,2),
\]
and the electromagnetic field obeys
\[
(E_0,B_0)\in H^s(\mathbb R^2),\qquad s\ge 2-\frac{2}{p}\in(0,1).
\]
The resulting solution is global and uniformly bounded with respect to the speed of light \(c\in(0,\infty)\); if \(s>\frac12\), then
\[
u \in L^1_{\loc}(\mathbb R^+; \dot{W}^{1,\infty}(\mathbb R^2)),
\]
and the energy solution is unique in the class of all energy solutions [2403.18500].

## 4. Singular limits and reduced equations

Several asymptotic regimes reorganize the incompressible Navier–Stokes–Maxwell dynamics into reduced systems. One such limit starts from a two-fluid incompressible Navier–Stokes–Maxwell system for cation and anion velocities \(u^\pm\), and introduces the bulk velocity and current variables
\[
u=\frac{u^++u^-}{2},\qquad j=\frac{u^+-u^-}{2\varepsilon}.
\]
The reformulated system contains the equations
\[
\partial_tu+u\cdot \nabla u+\varepsilon^2j\cdot \nabla j-\mu\Delta u=-\nabla p+j\times B,
\]
\[
\varepsilon^2\partial_tj+\varepsilon^2(u\cdot \nabla j+j\cdot \nabla u)-\varepsilon^2\mu\Delta j+\frac{1}{\sigma}j =-\nabla \tilde{p}+cE+u\times B,
\]
coupled to Maxwell’s equations and the divergence constraints. As \(\varepsilon\to0\), the \(\varepsilon^2\)-terms vanish and one obtains the incompressible Navier–Stokes–Maxwell system with solenoidal Ohm’s law,
\[
\partial_tu+u\cdot \nabla u-\mu\Delta u=-\nabla p+j\times B,\qquad
j=\sigma(-\nabla \tilde{p}+cE+u\times B),
\]
together with
\[
\frac{1}{c}\partial_tE-\nabla \times B=-j,\qquad
\frac{1}{c}\partial_tB+\nabla \times E=0,\qquad
\nabla\cdot u=\nabla\cdot j=\nabla\cdot E=\nabla\cdot B=0.
\]
The notable feature is that the convergence is strong and occurs without loss of regularity, through a frequency-envelope method combined with a low–high frequency decomposition [2507.11881].

A second family of limits concerns the speed of light. In the two-dimensional homogeneous case, the global weak-solution theory yields estimates uniform in \(c\), and along subsequences the solutions converge as \(c\to\infty\) to the two-dimensional magnetohydrodynamic system
\[
\partial_t u + u\cdot\nabla u - \mu\Delta u = -\nabla p + (\nabla\times B)\times B,\qquad \operatorname{div}u=0,
\]
\[
\partial_t B - \frac1\sigma \Delta B = \nabla\times(u\times B),\qquad \operatorname{div}B=0.
\]
The proof uses compactness for both \(u^c\) and \(B^c\), with the magnetic compactness relying on the stronger uniform control obtained for the electromagnetic field [1811.01364].

The same nonrelativistic passage appears in the inhomogeneous planar system, where uniform-in-\(c\) bounds imply convergence toward the inhomogeneous MHD equations
\[
\partial_t\rho+\operatorname{div} (\rho u)=0,
\]
\[
\partial_t (\rho u)+\operatorname{div}(\rho u\otimes u) -\nu \Delta u +\nabla\left( p- \frac{|B|^2}{2}\right)=B\cdot \nabla B,\qquad \div u =0,
\]
\[
\partial_t B  + u\cdot \nabla B - \frac{1}{\sigma} \Delta B  = B\cdot \nabla u,\qquad \div B = 0.
\]
Relative compactness follows from the energy and regularity bounds together with Aubin–Lions compactness [2403.18500].

A different singular regime begins with the full compressible Navier–Stokes–Maxwell system in a bounded smooth domain \(\Omega\subset\mathbb R^3\), with small Mach number and dielectric constant taken equal,
\[
\varepsilon_1=\varepsilon_2=\varepsilon.
\]
Under well-prepared initial data and slip/electromagnetic perfect-conductor boundary conditions, uniform-in-\(\varepsilon\) strong-solution estimates imply
\[
(\varepsilon\sigma^\varepsilon,u^\varepsilon,\varepsilon\theta^\varepsilon,b^\varepsilon)\to (0,v,0,B)
\quad\text{strongly in }L^2(0,T;H^1(\Omega)),
\]
and
\[
E^\varepsilon \to \operatorname{rot}B - v\times B
\quad\text{strongly in }L^2(0,T;L^2(\Omega)).
\]
The limit \((v,B)\) solves the incompressible MHD system
\[
v_t+v\cdot\nabla v+\nabla\pi-\mu\Delta v=\operatorname{rot}B\times B,
\]
\[
B_t-\operatorname{rot}(v\times B)-\Delta B=0,
\]
\[
\operatorname{div}v=0,\qquad \operatorname{div}B=0.
\]
This shows that simultaneous removal of the acoustic scaling and the displacement current leads from compressible Navier–Stokes–Maxwell to incompressible MHD in a bounded domain [1504.01071].

## 5. Kinetic derivations and the Fourier–Maxwell extension

A substantial part of the recent theory derives incompressible fluid–Maxwell systems from kinetic equations. The macroscopic limit obtained in these works is usually the two-fluid incompressible Navier–Stokes–Fourier–Maxwell system with Ohm’s law rather than the bare incompressible Navier–Stokes–Maxwell system. Its core structure contains the incompressible momentum equation, Maxwell equations, and Ohm-type closure, together with a temperature equation and Boussinesq-type constraints [1905.04739][2007.02286].

For the two-species Vlasov–Maxwell–Boltzmann system under diffusive scaling, one rigorous limit yields
\[
\partial_t u + u\cdot\nabla_x u - \mu \Delta u + \nabla_x p = \frac12\bigl(nE + j\times B\bigr),\qquad \operatorname{div}_x u=0,
\]
\[
\partial_t E - \nabla_x\times B = -j,\qquad \partial_t B + \nabla_x\times E = 0,
\]
and an Ohm law of the form
\[
j - n u = \sigma\left(-\frac12\nabla_x n + E + u\times B\right),
\]
supplemented by
\[
\partial_t\theta + u\cdot\nabla_x\theta - \kappa\Delta_x\theta = 0,\qquad \rho+\theta=0.
\]
One classical-solution result establishes global-in-time solutions uniform in the Knudsen number \(\varepsilon\), proves convergence of the fluctuations, and shows that the microscopic part \(P^\perp G_\varepsilon\) vanishes strongly in \(L^2(\mathbb R_+;H^s_{x,v})\) [1905.04739]. A companion Hilbert-expansion result constructs
\[
F^\varepsilon = M\bigl(1+\varepsilon g_0+\varepsilon g_1+\varepsilon g_2+\varepsilon g_{R,\varepsilon}\bigr), \quad
E^\varepsilon = E_0+\varepsilon E_1+\varepsilon E_{R,\varepsilon}, \quad
B^\varepsilon = B_0+\varepsilon B_1+\varepsilon B_{R,\varepsilon},
\]
and derives a global-in-time remainder estimate uniform in \(\varepsilon\) [2007.02286].

These kinetic derivations have been extended beyond hard-sphere Boltzmann kernels. For soft potentials, uniform-in-\(\epsilon\) global classical estimates are proved for both non-cutoff and cutoff two-species Vlasov–Maxwell–Boltzmann systems. In the non-cutoff case the analysis assumes
\[
\max\{-3,\,-\tfrac32-2s\}<\gamma<-2s,\qquad \frac12\le s<1,
\]
while the cutoff case considers
\[
-1\le \gamma<0.
\]
The hydrodynamic limit again produces an incompressible Navier–Stokes–Fourier–Maxwell system with Ohm’s law. The same source explicitly notes that, for a Navier–Stokes–Maxwell-only discussion, the key equations are the momentum equation, incompressibility, Maxwell equations, and Ohm’s law, whereas the temperature equation is an extra feature of the Fourier extension [2308.12607].

An analogous program has also been carried out for the two-species Vlasov–Maxwell–Landau system with Coulomb potential. There, a global-in-time uniform-in-Knudsen-number estimate shows that the microscopic part is dissipated at order \(\varepsilon^{-2}\), and the limit is once again the two-fluid incompressible Navier–Stokes–Fourier–Maxwell system with Ohm’s law, with transport coefficients \(\mu\), \(\kappa\), and \(\sigma\) determined by Landau-collision correctors [2309.01052].

## 6. Terminological boundaries and adjacent formulations

A recurrent source of confusion is the phrase “Navier–Stokes–Maxwell.” One important counterexample is the incompressible Navier–Stokes–Maxwell–Stefan system for multicomponent gaseous mixtures. That model consists of the species continuity equations,
\[
\partial_t \rho_i + \operatorname{div}(J_i+\rho_i u)=0,
\]
the Maxwell–Stefan diffusion laws,
\[
\nabla x_i = -\sum_{k=1}^{N+1} \frac{\rho_k J_i-\rho_i J_k}{c\,M_iM_k\,D_{ik}},
\]
the Navier–Stokes momentum equation,
\[
\partial_t u + (u\cdot\nabla)u - \Delta u + \nabla p = f,
\]
and the incompressibility constraint \(\operatorname{div}u=0\). The corresponding 2018 paper develops an artificial compressibility approximation for this coupled fluid–diffusion model and proves existence of weak solutions for the approximating system together with convergence back to the incompressible Maxwell–Stefan limit as \(\varepsilon\to0\). It is explicitly not about the electromagnetic Navier–Stokes–Maxwell system [1805.06815].

A second adjacent body of work comes from gauge/gravity duality. In Einstein–Maxwell and Gauss–Bonnet–Maxwell theories, a non-relativistic long-wavelength expansion around a charged AdS black brane and a timelike cutoff surface \(r=r_c\) yields the incompressible Navier–Stokes equation with external force density,
\[
\partial_t v_a + v^b\partial_b v_a + \partial_a P_r - \nu\,\partial^2 v_a = f_a,\qquad
f_a=F_{aj}J^j,
\]
while the Maxwell equations enforce
\[
\partial_a v^a=0.
\]
This construction produces a forced incompressible Navier–Stokes equation on the cutoff surface rather than the full electromagnetic Navier–Stokes–Maxwell system with dynamical \(E\) and \(B\) fields [1107.1430].

Within these boundaries, the incompressible Navier–Stokes–Maxwell system is best viewed as a family of incompressible fluid–electromagnetic models organized by three structural choices: the Ohmic closure for \(j\), the regularity class in which the Lorentz term \(j\times B\) can be controlled, and the asymptotic regime under consideration. The current literature therefore treats the system not as a single fixed PDE, but as a central electromagnetic fluid framework linked upward to two-fluid and kinetic descriptions and downward to MHD-type limits.

Source: https://www.emergentmind.com/topics/incompressible-navier-stokes-maxwell-system