The axisymmetric, incompressible, viscoresistive magnetohydrodynamic equations for purely poloidal currents I(r,z) are transformed into three coupled scalar partial differential equations. These show that finite dl/dz acts as a volumetric source of fluid "vorticity" which drives strong poloidal flows (jets). PACS numbers: 52.80.Mg, 52.55.Ez, 52.75.Di, 98.60.Qs Axisymmetric, purely poloidal current configurations are found in a wide variety of conducting fluids and plasmas. Terrestrial examples include electric furnace [l] and other industrial arcs [2], magnetoplasma-dynamic (MPD) thrusters [3], z pinches [4], Marshall guns [5], thyratrons [6], ignitrons, lightning, liquid metals [7], and electrolytes. Although astrophysical plasmas are typically assumed to have both toroidal and poloidal currents, there presumably exist situations where the current is primarily poloidal (Refs. [8,9] are possible examples). The most interesting property of these configurations, a strong axial acceleration of fluid or plasma away from regions of constricted current, was first discussed by Maecker [10]. Axially directed fluid velocities of the order of the Alfven velocity can result from this acceleration, and in a high current arc most [11] of the electrical energy input can go into translational kinetic energy, rather than into thermal energy or radiation. The thrust associated with this axial flow is the basis of the MPD rocket engine, while in electric arc furnaces, viscous dissipation of the arc kinetic energy heats the metal being processed. Arc axial flow velocities have been measured by Bowman [12] and by Irie and Barrault [13]. Reed [14] graphically demonstrated the flow in liquid mercury. Numerical magnetohydrodynamic (MHD) models for electric arc furnaces have been discussed by McKelliget and Szekely [15] and for MPD thrusters by La Pointe [16].We present here a new interpretation of this phenomenon which shows that axial current inhomogeneity acts as a volumetric source of fluid "vorticity," Our model assumes an incompressible, constant density fluid; this is reasonable for liquids but is an oversimplification for gases (compressibility considerations will alter the response to the vorticity generation mechanism discussed here, but should not aff'ect the mechanism itself).
The relevant equations are the MHD equation of motion p[9U/d/-fUVU]=JxB-V/>+pvV2u,where p and v are, respectively, the mass density and kinematic viscosity; the MHD Ohm's law and Ampere's and Faraday's laws VxB=A^oJ, VxE=-9B/a/.(2)The electric field is E= -VO -9A/9r and we choose the Coulomb gauge V-A=0, i.e., A=V/xV^, where / is a scalar function and we use cylindrical coordinates (/•,0,z) sothat V^ = ^/r. Since the 0 component of Eq.(1),
birUe) dt -^-V-'VirUe)=pv ri--f(rUe) or r or
+ -^(rUe) dz'(4)
contains no driving force term, finite Ue is possible only in the exceptional situations where (i) it has been imposed as an initial condition at the initial time ^o (after whichUe transiently decays), or (ii) it is imposed as a boundary condition on some ...