Magnetohydrodynamics, commonly abbreviated MHD, is the study of the motion of electrically conducting fluids and their interaction with a magnetic field. Its central feature is a two-way coupling: fluid motion changes the magnetic field through electromagnetic induction, while electric currents interacting with that field exert forces on the fluid. The conducting medium may be a liquid metal, an electrolyte, or a plasma. MHD combines fluid mechanics with electromagnetism, usually treating matter as a continuous medium rather than following individual charged particles.
The subject connects phenomena across an exceptional range of scales. It helps explain the generation of planetary magnetic fields, disturbances in the solar wind, the structure of stellar atmospheres, and the confinement of plasmas for nuclear fusion. It also underlies electromagnetic pumps, liquid-metal flow control, and several approaches to electricity generation. Its equations support magnetic waves, shocks, instabilities, turbulence, and self-sustaining magnetic-field generation.
MHD is not a single universally applicable approximation. Ideal MHD neglects electrical resistance and several microscopic effects; resistive MHD permits magnetic diffusion and changes in magnetic connectivity. More elaborate formulations include Hall currents, separate ion and electron dynamics, partial ionization, anisotropic transport, and relativistic motion. Choosing among these descriptions requires comparing the spatial and temporal scales of a problem with the microscopic scales of the conducting medium.
Historical development and physical foundations
The foundations of MHD emerged from nineteenth-century work on electromagnetic forces and induction. Michael Faraday established electromagnetic induction in 1831 and investigated whether the motion of conducting water through Earth’s magnetic field could generate measurable electrical effects. The underlying principle is the same as that used in an electromagnetic flowmeter: charges in a moving conductor experience a magnetic force, producing charge separation and an electric potential difference. Early river experiments were difficult because the induced signals were small and electrical conditions at the boundaries were poorly controlled.
The theoretical synthesis depended on Maxwell’s equations and on continuum descriptions of fluid motion. A moving conductor does not merely respond passively to an imposed field. Its induced currents generate additional magnetic fields, which can alter both the original field and the motion that produced them. This feedback distinguishes a fully coupled MHD problem from the simpler calculation of a prescribed magnetic force acting on a known flow. Nevertheless, many engineering applications operate in a regime where the induced magnetic field is weak enough to neglect.
A decisive development came in 1942, when Hannes Alfvén proposed transverse waves supported by magnetic tension in a conducting fluid. These became known as Alfvén waves and supplied a concrete mechanism by which magnetic stresses could transmit disturbances through astrophysical matter. Alfvén received the 1970 Nobel Prize in Physics, shared with Louis Néel, for fundamental work and discoveries in magnetohydrodynamics with applications in plasma physics. The importance of his contribution lay partly in establishing that magnetic fields could have a dynamic, rather than merely diagnostic, role in space and astrophysical systems.
During the mid-twentieth century, several research programs developed in parallel. Studies of planetary magnetism led to dynamo theory; investigations of solar activity emphasized magnetic equilibria and eruptions; controlled-fusion research focused on equilibrium and stability; and engineering work examined conducting liquids in strong applied fields. Numerical computation subsequently made it possible to study nonlinear systems beyond the reach of analytical solutions. These branches share common conservation laws, but differ substantially in geometry, boundary conditions, collisionality, and the relative importance of rotation, compressibility, and electrical resistance.
The continuum approximation and its limits
MHD replaces the microscopic motions of many particles with macroscopic fields such as mass density , fluid velocity , pressure , and magnetic field . This is a form of fluid dynamics, but electrical conductivity introduces additional degrees of freedom and stresses. In a liquid metal, collisions ordinarily establish local fluid behavior on scales far smaller than an apparatus. In a dilute space plasma, the justification is more subtle because charged particles may travel considerable distances between collisions while remaining strongly constrained by the magnetic field.
The simplest plasma version is a single-fluid model. Ions and electrons contribute to a common bulk velocity, while their small relative drift carries the electric current. Approximate charge neutrality means that positive and negative charge densities nearly balance on scales much larger than the Debye length. This does not imply that the electric field or current vanishes. A very small charge imbalance can produce a dynamically important electric field, and substantial currents can flow through an almost neutral plasma.
Several scale separations support ordinary MHD. Characteristic speeds should be much smaller than the speed of light, and characteristic frequencies should be low enough that electromagnetic radiation and the displacement current do not control the dynamics. Spatial scales should normally exceed relevant ion kinetic scales, including the ion inertial length and, for magnetized ions, the ion gyroradius. The ion gyroperiod should also be short compared with the macroscopic evolution time when a magnetized-fluid description is invoked. These conditions are ordering assumptions, not a universal checklist sufficient for every plasma.
A scalar pressure requires an additional closure assumption. Collisions can make particle velocity distributions nearly isotropic, but weakly collisional plasmas may have different pressures parallel and perpendicular to the magnetic field. Heat flow and viscosity can likewise be strongly directional. Consequently, ordinary isotropic MHD can describe large-scale magnetic organization even where its thermodynamic closure is incomplete.
More detailed descriptions retain separate electron and ion fluids or evolve a particle distribution function through kinetic theory. They become important for collisionless damping, particle acceleration, small reconnection layers, and microinstabilities. The practical boundary between MHD and kinetic physics therefore depends on the phenomenon being studied: a global structure may be well described by MHD while its heating or dissipation requires a microscopic treatment.
Governing equations and conservation laws
The standard nonrelativistic MHD system combines conservation of mass, momentum, and energy with an induction equation. In the International System of Units, for a medium whose magnetic permeability is approximately the vacuum value , mass conservation is
Here is the mass density and is the bulk velocity. The momentum equation can be written
where is current density, is gravitational acceleration, and is the viscous stress tensor. The magnetic contribution is the Lorentz force per unit volume.
At low frequencies, Ampère’s law and the absence of magnetic monopoles give
With scalar electrical conductivity , the simplest moving-conductor form of Ohm’s law is
Combining this relation with Faraday’s law of induction gives, for spatially constant magnetic diffusivity ,
The two terms represent field transport and deformation by the flow, and resistive diffusion.
A thermodynamic closure completes the system. For a simple fluid with specific internal energy ,
The material derivative is , is heat flux, and represents additional volumetric heating or cooling. The term is Joule heating. An equation of state, such as for a calorically ideal gas, relates the thermodynamic variables.
These equations explicitly exchange energy between motion, magnetic fields, and internal energy. In an isolated ideal system, the sum of kinetic, magnetic, and internal energy is conserved. Resistivity does not simply destroy total energy: it converts magnetic energy into heat. Likewise, viscosity converts ordered motion into internal energy. Gravity, radiation, external currents, and boundary fluxes must be included consistently when assessing the energy balance of a particular system.
Characteristic scales and dimensionless parameters
Dimensionless parameters identify which terms dominate the MHD equations and whether results from one system can be compared with another. The magnetic Reynolds number is
where and are characteristic velocity and length scales. It compares inductive transport with magnetic diffusion. Equivalently, it is the ratio of the diffusion time to the flow time . When , an imposed magnetic field is only weakly perturbed by the flow. When , induction can strongly deform and amplify magnetic structures, although diffusion may remain important in thin layers.
The ordinary Reynolds number, , compares inertia with viscous effects, where is kinematic viscosity. Their ratio defines the magnetic Prandtl number,
Liquid metals generally have small magnetic Prandtl numbers, whereas some dilute astrophysical plasmas have very large effective values. This difference changes the ordering of viscous and resistive scales and complicates attempts to reproduce astrophysical magnetic turbulence in the laboratory. Similar magnetic Reynolds numbers do not guarantee similar flows if the other controlling parameters differ.
The Alfvén speed,
measures the propagation speed of magnetic-tension disturbances along the field. The Alfvén Mach number compares flow inertia with magnetic effects. Plasma beta,
compares thermal pressure with magnetic pressure. A low-beta plasma is magnetically dominated in this pressure sense, but beta alone does not determine the importance of inertia, gravity, or field geometry.
Other parameters are useful in particular settings. The Lundquist number compares magnetic diffusion with an Alfvén crossing time. The Hartmann number,
measures magnetic effects relative to viscosity in conducting-fluid flows. The interaction parameter compares Lorentz forces with inertia in low-magnetic-Reynolds-number applications. Rotation introduces the Rossby number and, in many dynamo studies, the Elsasser number.
Each parameter requires a stated convention for , , and the representative field strength. Different choices can change numerical values substantially in inhomogeneous systems. Scaling analysis is therefore most informative when accompanied by an explicit physical balance, rather than by an isolated assertion that a parameter is “large” or “small.”
Magnetic stresses, equilibria, and flux freezing
Magnetic forces have a useful mechanical interpretation. Using , the Lorentz force becomes
The first term resembles a pressure force associated with magnetic energy density. The second describes magnetic tension, including the tendency of curved field lines to straighten. These are complementary parts of one stress, not independent forces that may always be considered separately. Their balance explains why a magnetic field can resist transverse compression while transmitting a restoring force along a bent magnetic structure.
A static equilibrium satisfies
Without gravity, pressure is constant along field lines because the Lorentz force has no component parallel to . In a simple planar configuration with a tangential magnetic field, equilibrium may reduce to constant total pressure, . More complicated geometries include curved magnetic surfaces, electric currents, and pressure gradients whose mutual compatibility strongly constrains the possible equilibrium.
A force-free field satisfies , so current is parallel to the magnetic field wherever both are nonzero. Such a field can contain substantial magnetic energy and electric current despite exerting no local Lorentz force. Force-free models are useful in magnetically dominated regions, including some descriptions of the solar corona, but fail where pressure, inertia, or gravity provides an essential part of the force balance.
In ideal MHD, electrical conductivity is taken to be infinite, giving . For sufficiently smooth fields and flows, the magnetic flux through a material surface moving with the fluid is conserved. This is Alfvén’s flux-freezing theorem. It is often described by saying that magnetic field lines move with the fluid, although field lines are geometric constructions rather than material objects.
Flux freezing permits amplification through stretching and compression. Under uniform isotropic compression, for example, conservation of mass and magnetic flux gives . It also restricts changes in magnetic connectivity. These restrictions depend on the ideal assumptions and regularity of the solution. Finite resistance, kinetic effects, or singular small-scale structure can invalidate the simple frozen-in picture even when large-scale magnetic Reynolds numbers are enormous.
Waves, discontinuities, and shocks
A uniform ideal-MHD medium supports three principal propagating wave families: Alfvén, fast magnetosonic, and slow magnetosonic waves. The Alfvén wave is transverse and, in the simplest linear setting, incompressible. Its frequency obeys , where is the component of wavevector along the background magnetic field. The restoring force is magnetic tension, and the velocity and magnetic perturbations are perpendicular to both the background field and the propagation plane.
The fast and slow modes couple pressure forces to magnetic stresses and generally involve compression. If is the adiabatic sound speed and the angle between the wavevector and the background field, their squared phase speeds are
The plus sign denotes the fast mode and the minus sign the slow mode. Their character depends on both propagation direction and plasma beta. For perpendicular propagation in this homogeneous ideal limit, the fast speed is , while the slow-mode frequency tends to zero.
Nonlinear compressive waves can steepen into a shock wave. Across a thin shock, integral conservation laws relate upstream and downstream mass flux, momentum, energy, and magnetic field. The magnetic field component normal to the shock remains continuous because magnetic monopoles are absent. Fast shocks, slow shocks, contact discontinuities, and rotational discontinuities have different characteristic properties; their admissibility also depends on the wave characteristics and entropy production.
Earth’s bow shock provides a large-scale example of a magnetized shock formed when the solar wind encounters an obstacle. Many space shocks are collisionless: their thickness and irreversible heating arise from collective electromagnetic processes rather than ordinary molecular collisions. MHD can describe their large-scale jump conditions without resolving those processes. This distinction is crucial because agreement with macroscopic conservation laws does not establish how electrons and ions divide the dissipated energy.
Observed waves can also diagnose inaccessible plasma conditions. Oscillations of solar coronal structures, for example, constrain combinations of magnetic field strength, density, and geometry. Interpretation requires care because realistic structures are nonuniform, multiple modes may coexist, and damping can involve phase mixing, resonant transfer, thermal conduction, or kinetic effects absent from the simplest wave equations.
Magnetic reconnection and nonideal effects
Magnetic reconnection is the rearrangement of magnetic connectivity in a region where ideal flux freezing fails. It can convert magnetic energy into bulk motion, heat, and accelerated particles. Reconnection is central to solar flares, magnetospheric energy release, fusion-plasma relaxation, and many turbulent systems. A complete description must distinguish the small region where ideal behavior breaks down from the much larger region through which magnetic stresses drive inflow and outflow.
In resistive MHD, reconnection can occur in a narrow current sheet where the resistive term becomes important despite a large system-scale magnetic Reynolds number. The classical Sweet–Parker model, developed in the 1950s, balances inflow, outflow, advection, and diffusion in a long, thin layer. For a steady, approximately incompressible configuration with comparable upstream and downstream densities, its scaling is
where is the sheet thickness. At very large Lundquist numbers this predicts a reconnection rate too slow for many observed rapid events.
Several mechanisms change this result. Long current sheets can become unstable to the formation of multiple magnetic islands, or plasmoids, producing a time-dependent hierarchy of shorter sheets. Three-dimensional turbulence can alter field-line transport and the geometry of reconnection. In collisionless plasmas, ions and electrons decouple at small scales, so ordinary resistive MHD no longer supplies the appropriate electric-field balance.
A generalized Ohm’s law may include the Hall term , electron-pressure forces, and electron inertia, where is electron number density and is the magnitude of electron charge. The Hall term changes how magnetic flux is transported relative to the ion flow, but does not by itself provide ordinary Joule dissipation, since . Electron-scale physics can determine where the electron frozen-in condition breaks down.
Partially ionized media introduce another distinction. Collisions between charged particles and neutrals can permit ambipolar drift, in which the magnetized charged component moves relative to neutral material. This can concentrate currents and dissipate energy through friction. Reconnection rates therefore cannot be inferred from conductivity alone: geometry, guide fields, plasma composition, pressure anisotropy, boundary driving, and the hierarchy of kinetic scales can all be decisive.
Instabilities and magnetic confinement
An MHD equilibrium may satisfy force balance yet remain unstable to small disturbances. Stability concerns whether perturbations grow by drawing energy from currents, pressure gradients, shear flows, gravity, or differential rotation. For static ideal-MHD equilibria under appropriate boundary conditions, the energy principle expresses stability through the change in potential energy caused by a fluid displacement. A displacement that lowers the relevant energy signals an unstable direction.
Current-driven instabilities are especially important in magnetically confined columns. A sausage, or axisymmetric pinch, disturbance produces alternating constrictions and expansions. A kink disturbance bends the column or magnetic axis into a helical shape. Their behavior depends on the relative strength of axial and azimuthal magnetic fields, the current profile, the column length, and whether a conducting wall supplies stabilizing constraints. Field-line twist is therefore important, but no single twist threshold applies to every geometry and boundary condition.
Pressure-driven interchange modes exchange neighboring plasma elements in a way analogous to heavy fluid moving downward in a gravitational field. Magnetic curvature can act favorably or unfavorably in this process. Ballooning modes localize in regions of unfavorable curvature and are particularly important in toroidal confinement. Magnetic shear can reduce the coherence of some disturbances, while pressure gradients provide their free energy.
Nonideal effects introduce additional modes. A tearing instability uses finite resistivity or other nonideal physics to reconnect field lines and form magnetic islands. Its growth is typically slower than an Alfvén crossing rate in the classical resistive regime, but it can strongly change confinement by flattening pressure profiles across islands. In fusion plasmas, bootstrap-current perturbations can drive neoclassical tearing modes, illustrating how kinetic transport and large-scale MHD interact.
Flow supplies other energy sources. The Kelvin–Helmholtz instability arises from velocity shear, with magnetic tension suppressing some perturbations depending on field orientation and strength. The magnetorotational instability destabilizes certain differentially rotating magnetized flows, notably those whose angular velocity decreases outward. It can operate even when the corresponding unmagnetized flow satisfies the usual centrifugal stability condition.
A linear instability calculation identifies early growth rates and mode structures, not necessarily the final state. Nonlinear evolution may saturate into a persistent deformation, generate turbulence, trigger reconnection, or destroy the original equilibrium. Predicting that outcome requires the energy sources, dissipative mechanisms, and boundary response to be modeled together.
Magnetohydrodynamic turbulence
MHD turbulence involves irregular, nonlinear motion coupled to magnetic fluctuations over a range of scales. Unlike ordinary hydrodynamic turbulence, it contains both kinetic and magnetic energy reservoirs and supports wave propagation along a local magnetic field. Energy can pass between fluid motion and magnetic fields while also cascading toward smaller scales. The magnetic field often makes the resulting structures strongly anisotropic.
A useful formulation employs the Elsasser variables,
where is the magnetic fluctuation about a background field and density is taken constant. In ideal incompressible MHD, nonlinear interactions are expressed as interactions between oppositely propagating Elsasser fluctuations. A disturbance traveling in only one direction along a uniform guide field is therefore a special case with no ordinary counterpropagating nonlinear cascade.
Theories differ according to the strength of the fluctuations and the relative timescales of propagation and nonlinear distortion. Weak-turbulence approaches treat individual wave interactions as small perturbations. Strong-turbulence phenomenologies often invoke critical balance, in which the Alfvén propagation time along a structure is comparable to its nonlinear distortion time across the field. This leads to scale-dependent anisotropy rather than a cascade that is statistically identical in all directions.
Energy spectra alone do not uniquely identify the governing mechanism. Measurements and simulations often find approximate power laws, but slopes vary with imbalance between oppositely traveling fluctuations, compressibility, intermittency, alignment, and the extent of the resolved scale range. Distinguishing a spectrum from a spectrum over a limited interval is not sufficient to settle the underlying dynamics.
MHD turbulence also produces intermittent current sheets and concentrated vorticity. These structures can become sites of reconnection and enhanced dissipation. At scales approaching the ion gyroradius or ion inertial length, a fluid cascade can transition into regimes requiring Hall or kinetic physics. The ultimate division of heating between electrons and ions is therefore not determined by ideal MHD.
The solar wind provides unusually detailed in situ observations of these processes. Spacecraft measure magnetic fluctuations, particle distributions, and flow velocities along their trajectories. Translating a time series into a spatial spectrum often uses an assumption that structures pass the spacecraft faster than they evolve. That approximation can fail when wave propagation or intrinsic evolution is comparable to the sampling speed, especially close to the Sun.
Dynamo action and magnetic-field generation
A dynamo maintains or amplifies a magnetic field by converting mechanical energy into magnetic energy through electromagnetic induction. A conducting fluid need not contain a permanent magnet to sustain a field. Instead, suitable motion stretches, folds, and rearranges an existing weak seed field faster than electrical resistance removes it. Dynamo action powers the large-scale fields of Earth and many stars and contributes to magnetic fields in galaxies.
The kinematic dynamo problem prescribes the velocity field and asks whether a weak magnetic field grows. The induction equation is then linear in , although the flow can be complicated. Exponential magnetic growth indicates that the prescribed motion overcomes diffusion. In the nonlinear problem, the growing field exerts Lorentz forces that modify the flow, eventually leading to saturation, time dependence, or a transition to a different state.
A large magnetic Reynolds number is often necessary but is not sufficient. The geometry and temporal structure of the flow matter. Cowling’s anti-dynamo theorem, established in 1933, rules out the maintenance of an axisymmetric magnetic field by axisymmetric dynamo action under its assumptions. Other anti-dynamo results exclude classes of overly simple flows. These restrictions explain why rotation, three-dimensional motion, and broken symmetries play prominent roles in successful dynamo models.
Large-scale and small-scale dynamos describe different aspects of magnetic organization. A small-scale dynamo amplifies fluctuations whose characteristic scales are comparable to or smaller than the driving motions. A large-scale dynamo generates organized fields on scales exceeding the principal turbulent eddies. Mean-field theory separates average and fluctuating quantities, introducing a turbulent electromotive force . Under suitable conditions this can produce effects associated with helical turbulence, while differential rotation converts poloidal field into toroidal field.
Magnetic helicity, which measures aspects of field linkage and twist, constrains dynamo evolution. In a closed volume with , it can be written as , where . Open boundaries require a gauge-consistent relative-helicity formulation. Approximate helicity conservation at high conductivity can restrict the growth of large-scale fields unless helicity is transported between scales or through boundaries. Dynamo saturation consequently depends on more than a local balance between magnetic and kinetic energy.
Planetary interiors and magnetospheres
Earth’s main magnetic field is generated by dynamo action in the electrically conducting liquid outer core, composed predominantly of iron alloy. The solid inner core extends to a radius of about kilometres, while the core–mantle boundary lies at a radius of roughly kilometres. The outer core between them is therefore a conducting fluid shell more than kilometres thick. Its motion is strongly influenced by rotation, buoyancy, and magnetic stresses.
Cooling and inner-core growth contribute to the energy and buoyancy available to drive core motion. As the inner core freezes, the partitioning of lighter elements can make adjacent liquid buoyant. The resulting flow interacts with the magnetic field through induction. Coriolis forces organize motion in the rapidly rotating shell, while Lorentz forces modify that organization. Geodynamo models often investigate balances involving Coriolis, pressure, buoyancy, and magnetic forces, although the relative importance of individual terms varies with scale.
Geomagnetic observations show secular variation on timescales accessible to historical and modern measurements, while rocks preserve evidence of polarity reversals over geological time. Reversals are not periodic clockwork events. Their occurrence demonstrates that a dynamo can sustain a field while its global dipole changes strength and orientation. Numerical simulations reproduce reversals under some parameter choices, but computational limits prevent direct reproduction of all terrestrial core parameters.
Other planets illustrate the range of conducting fluids and dynamo geometries. Jupiter’s field is associated with its deep conducting interior, including metallic hydrogen. Uranus and Neptune have strongly non-dipolar fields, interpreted through models involving conducting fluid regions unlike Earth’s iron core. Planetary magnetic fields therefore provide indirect evidence about otherwise inaccessible interiors, but the inference depends on thermodynamic, compositional, and dynamical models.
Outside a magnetized planet, the magnetosphere forms through interaction with the solar wind. Global MHD describes the bow shock, magnetopause, magnetotail, and large-scale plasma circulation. Reconnection at the magnetopause transfers magnetic flux and energy into the system; reconnection in the tail contributes to rapid energy release. These processes couple to the ionosphere through electric currents and electromagnetic energy flux. Kinetic physics is essential in thin boundary layers and particle-acceleration regions, so operational and research models often combine global MHD with more specialized descriptions of those regions.
Solar, stellar, and relativistic astrophysics
The Sun contains several distinct MHD regimes. In its convection zone, buoyant motion and rotation participate in magnetic-field generation. Near the visible surface, convection concentrates magnetic flux into structures ranging from small magnetic elements to sunspots. In the corona, low density and high temperature allow magnetic forces to dominate much of the large-scale organization. The same equations must therefore be used with very different pressure balances, transport coefficients, and radiative treatments in adjacent regions.
Solar magnetic activity includes an approximately 11-year sunspot cycle, with a full magnetic-polarity cycle lasting about 22 years. Differential rotation, convective motions, and flux emergence are central ingredients of solar dynamo models. The detailed organization of the dynamo and the transport of magnetic flux through the interior remain active research topics. Observations constrain the surface field and some internal motions, but do not uniquely determine the complete three-dimensional subsurface field.
In the corona, photospheric motions can stress magnetic structures and inject energy. Waves and reconnection are major components of proposed heating mechanisms. Coronal mass ejections involve large expulsions of magnetized plasma, while solar flares convert stored magnetic energy into radiation, heat, bulk motion, and energetic particles. MHD models can follow large-scale eruptions and their propagation through the heliosphere, but particle acceleration and the smallest dissipation regions require additional physics.
In accretion disks around stars and compact objects, magnetic stresses transport angular momentum. The magnetorotational instability provides one route to turbulence in sufficiently ionized differentially rotating disks. In cold, weakly ionized regions of disks around young stars, Ohmic diffusion, Hall effects, and ambipolar drift can alter or suppress this behavior. Disk winds can also remove angular momentum, so turbulent transport is not the only magnetic mechanism available.
Jets and flows near a black hole may require relativistic MHD. Special-relativistic formulations account for flows approaching light speed, while general-relativistic MHD incorporates curved spacetime. Magnetic fields can channel outflows and mediate extraction of rotational energy from magnetized accreting systems. In very strongly magnetized regions with negligible material inertia, force-free electrodynamics supplies a related limiting description.
Radiation can introduce another essential coupling. In luminous accretion flows and some stellar environments, radiation transports energy and exerts dynamically important forces. Radiation-MHD therefore evolves or approximates radiation transport alongside the fluid and magnetic field, rather than treating all emitted energy as a simple local cooling term.
Fusion plasmas and laboratory experiments
MHD supplies the principal large-scale description of equilibrium and stability in magnetic-confinement fusion. In a tokamak, externally produced toroidal magnetic field combines with a poloidal field associated largely with plasma current to form helical field lines. In a stellarator, external coils produce much of the required three-dimensional magnetic geometry. Both approaches aim to confine hot plasma long enough for fusion reactions to contribute substantial energy.
Axisymmetric tokamak equilibria are commonly described by the Grad–Shafranov equation, which expresses force balance in terms of magnetic flux surfaces and pressure and current profiles. The safety factor characterizes the winding of field lines around the torus. Its radial profile affects magnetic shear, resonant surfaces, and the accessibility of particular instabilities. A computed equilibrium is only a starting point: pressure, current, and shape must also remain compatible with stability and transport constraints.
Ideal-MHD calculations identify limits associated with kink and ballooning modes. Resistive and extended-MHD models address tearing modes, resistive-wall interactions, and events involving reconnection. Tokamak disruptions can involve rapid loss of thermal energy and plasma current, with substantial electromagnetic and thermal loads on surrounding structures. Their prediction and mitigation require models that extend beyond ideal MHD to include radiation, impurities, wall coupling, and, in some cases, runaway electrons.
MHD also describes laboratory pinch devices, in which a large current produces a magnetic field that compresses plasma. Rapidly evolving pinches can generate extreme pressures and intense radiation, but current-driven instabilities strongly affect their behavior. These experiments test magnetic compression, instability growth, and radiation-coupled dynamics under conditions quite different from the approximately steady equilibria sought in many confinement systems.
Liquid-metal experiments investigate induction, dynamos, and magnetized flow without the thermodynamic complexity of a hot plasma. The Riga and Karlsruhe experiments reported self-excited dynamo action in 1999 using liquid sodium. Later experiments explored more turbulent flows, field reversals, and the effects of boundary materials. Such devices offer controlled measurements of velocity, electric potentials, and magnetic fields, although achieving large magnetic Reynolds numbers requires substantial size, speed, or electrical conductivity.
Laboratory similarity is necessarily selective. Matching magnetic Reynolds number may be feasible while matching magnetic Prandtl number, rotation, stratification, and boundary conductivity simultaneously is not. Experiments therefore test particular mechanisms and scaling laws rather than reproducing an entire planetary core or astrophysical plasma in miniature.
Engineering applications and low-induction flows
Many practical MHD systems use an externally imposed magnetic field and operate at low magnetic Reynolds number. In this quasi-static regime, the induced field is small relative to the applied field, but induced currents can still exert strong forces. The electric potential is obtained from charge conservation, , together with the moving-conductor Ohm’s law. Electrical boundary conditions are essential because currents must close through the fluid, walls, electrodes, or an external circuit.
An electromagnetic pump drives a conducting liquid using , avoiding some moving mechanical components in contact with the fluid. A flowmeter uses the complementary effect: motion through a magnetic field produces an electric potential difference related to flow velocity. For a simple geometry with mutually perpendicular velocity, field, and electrode separation, the characteristic induced voltage scales as , where is the electrode spacing. Real instruments require calibration for geometry, conductivity distribution, and flow profile.
In metallurgy, magnetic fields can stir, brake, or shape molten metal. Electromagnetic stirring modifies mixing and heat transfer during casting and solidification. Magnetic braking can suppress unwanted jets or fluctuations, while carefully selected forcing can improve homogenization. The outcome is not always simple damping: spatially varying fields and current paths can generate vortices, redistribute momentum, and produce nonuniform thermal conditions.
MHD generators convert part of a conducting fluid’s mechanical or thermal energy directly into electrical energy as the fluid crosses a magnetic field. Their electromagnetic principle resembles that of a conventional generator, but the moving conductor is a fluid rather than a solid winding. Practical systems face demanding requirements involving conductivity, electrode durability, high-temperature materials, pressure losses, and overall thermodynamic efficiency. These constraints have prevented broad replacement of conventional electricity-generation technologies.
Strong-field liquid-metal flows are also relevant to proposed fusion blankets, where conducting coolants or breeder materials encounter intense magnetic fields. Lorentz forces can create substantial pressure drops and alter heat transport. Thin boundary layers become important: Hartmann layers form at walls normal to a simple imposed field, while different shear layers occur along field-parallel walls. For large Hartmann number in the classical planar setting, the Hartmann-layer thickness scales as .
Wall conductivity can change the entire flow because it alters current closure. Consequently, a model that accurately represents the liquid but treats insulating and conducting walls interchangeably can give qualitatively wrong pressure losses and velocity distributions.
Numerical methods, validation, and unresolved problems
Most realistic MHD problems require numerical solutions because nonlinear coupling, complex boundaries, and broad scale ranges prevent closed-form analysis. Compressible codes commonly use conservative finite-volume methods designed to capture shocks and discontinuities. Incompressible and weakly compressible systems may use finite-difference, finite-element, or spectral methods. Adaptive mesh refinement concentrates resolution around shocks, current sheets, and other localized structures, while parallel computing makes large three-dimensional calculations possible.
A central numerical requirement is preserving . Constrained-transport schemes evolve magnetic fluxes so that a discrete divergence remains zero to numerical precision. Divergence-cleaning schemes instead transport and damp numerical divergence errors. Vector-potential formulations enforce a related constraint through , but introduce gauge and discretization considerations. Uncontrolled magnetic divergence can generate spurious forces and corrupt the physical solution.
Resolution and dissipation require equal attention. A code solving nominally ideal MHD still introduces numerical diffusion through discretization and stabilization. Reconnection or heating observed in such a calculation may therefore depend on the grid unless a converged macroscopic regime has been established. Conversely, adding explicit resistivity does not guarantee that the resistive layer is actually resolved. Convergence tests, energy budgets, and comparisons across algorithms are necessary to determine what a simulation demonstrates.
Verification asks whether the equations are being solved correctly; validation asks whether those equations adequately describe the physical system. Standard verification problems include linear waves, shock tubes, advected magnetic structures, and instability growth. Validation compares predictions with experiments or observations, accounting for diagnostic limitations and measurement uncertainty. In astrophysics, uncertain boundary conditions and line-of-sight integration can be as important as numerical error.
Several unresolved problems concern the connection between fluid and microscopic scales. These include how turbulent energy becomes particle heat, how reconnection rates emerge in complex three-dimensional systems, and how weakly collisional plasmas regulate pressure anisotropy and transport. Dynamo predictions remain limited by the difficulty of reaching planetary and stellar parameter regimes. Fusion applications face the additional challenge of predicting rare, rapidly developing events from evolving profiles and imperfect measurements.
Hybrid approaches address these limitations by embedding kinetic regions within fluid calculations, coupling separate plasma components, or developing closures informed by particle simulations. Their validity depends on conserving exchanged quantities and avoiding inconsistent assumptions at interfaces. The resulting models remain tied to specific scale hierarchies: greater complexity does not remove the need to establish which physical processes control the phenomenon under study.