Heat transport in magnetohydrodynamic duct flow regimes with conducting and insulating walls

Andreu Queralt, Dmitry Krasnov, Yuri Kolesnikov, Jörg Schumacher · 2026

Read Heat transport in magnetohydrodynamic duct flow regimes with conducting and insulating walls by Andreu Queralt, Dmitry Krasnov, Yuri Kolesnikov, Jörg Schumacher (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-12485.txt.

Introduction

Andreu Queralt McBride and three colleagues simulate liquid metal flow in a rectangular duct under a magnetic field. Side walls are heated evenly, and a transverse field runs across the flow. The team varies wall conductivity, from highly conducting to fully insulating. Buoyancy acts in horizontal and vertical ducts, sometimes with the flow and sometimes against it. Vortex promoters at the inlet stir the flow. Four distinct flow types emerge across the parameter range. Heat transfer for each type is scored with the Nusselt number, a heat transfer measure.

Catalog record

CREATORAndreu Queralt, Dmitry Krasnov, Yuri Kolesnikov, Jörg Schumacher
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE17105
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.12485

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Heat transport in magnetohydrodynamic duct flow regimes with conducting and insulating walls Andreu Queralt, Dmitry Krasnov, Yuri Kolesnikov, Jörg Schumacher Abstract ABSTRACT The flow of a liquid metal (LM) in a rectangular duct segment, subject to a uniform transverse magnetic field and uniform heating at the side walls is explored in an ample parameter space using Direct Numerical Simulation (DNS). We modify electrical wall conductivity, (either highly conducting or perfectly insulating) and investigate the effects of the buoyancy force, both in horizontally and vertically orientated ducts. In the latter case, it may be directed either with the flow or against the flow, creating backflow regions. In this parameter space and with the presence of vortex promoters at the inlet of the duct we identify 4 types of flow. We calculate the Nusselt number Nu(t) for each of them and study the statistical properties to compare their heat transfer capabilities in future fusion reactor blankets. INTRODUCTION One of the main candidates for coolant in future fusion reactors are liquid metals which flow inside the blankets. They satisfy operational requirements and, due to the flow of neutrons from the plasma core, tritium breeding can occur to provide fuel for future reactions. Despite these advantages, several engineering questions regarding their viability are still open. In the TOKAMAK context, the flow is subject to strong magnetic fields. The toroidal field component in particular is present in the blanket. Depending on the orientation of the duct, this component will be either in the streamwise direction or along two of the side walls. A second, smaller component is present in the poloidal direction of the TOKAMAK. For simplicity we will consider only the toroidal field, which will be perpendicular to the streamwise direction. In the case of conducting duct walls, the Lorenz force increases, dampens the velocity and increases the pressure drop \Delta P needed to pump the metal with sufficient volume flux ( 7 ) . Additionally, an M-profile is formed at the Shercliff layers. The strong velocity gradients at the walls are detrimental, both to the production of tritium and the lifespan of the duct, with the steep velocity gradients increasing corrosion at the walls ( 9 ) . To overcome these problems, ceramic Flow Channel Inserts (FCI) are mounted at the duct walls. These come with their own set of problems, such as different heat dilatation from the metal, development of cracks and disintegration under the neutron flux, the latter of which increases maintenance requirements (see, e.g., 10 for an overview). DUCT GEOMETRY Figure 1 shows the simulation domain, a rectangular duct of aspect ratio L_{y}/h=3.5 is considered. To maintain cohesion with previous studies, such as in 6 , the length of the duct L_{x} is 50 times its height h and at the inlet we also prescribe a velocity profile with two flat jets near the top and bottom walls. This profile mimics the velocity field past a round cylinder, which was used in experiments ( 6 ) to generate inlet instabilities by covering \sim 50\% of the inlet area (see figure 1 ). When heating is applied it is done only at the Shercliff walls, while the Hartmann walls remain adiabatic. Both Shercliff walls are heated to simplify the problem by imposing symmetry. The heating area begins at x=20 , again to maintain cohesion with previous studies. When the conductivity of the wall is present, it is set equal at all four walls. Figure 1: Simulated setup schematic for the duct. Left: side view. Right: downstream view. MATHEMATICAL MODEL In order to solve the flow of the liquid metal, the governing equations are used for an incompressible flow at low magnetic Reynolds numbers Rm , known as quasi-static approximation (see, e.g., 4 ): \displaystyle\nabla\cdot\bm{u} \displaystyle= \displaystyle 0 (1) \displaystyle\frac{\partial\bm{u}}{\partial t}\!+\!(\bm{u}\!\cdot\!\nabla)\bm{u} \displaystyle= \displaystyle\!-\nabla\!p+\!\frac{1}{\mbox{{Re}}}\!\bigg(\!\nabla^{2}\!\bm{u}+\mbox{{Ha}}^{2}\!\bm{j}\!\times\!\bm{e}_{B}\!\bigg)\!-\!\frac{\mbox{{Gr}}}{\mbox{{Re}}^{2}}T\!\bm{e}_{g} (2) \displaystyle\frac{\partial T}{\partial t}+\bm{u}\cdot\nabla T \displaystyle= \displaystyle\frac{1}{\mbox{{Re}}\mbox{{Pr}}}\nabla^{2}T (3) \displaystyle\bm{j} \displaystyle= \displaystyle-\nabla\phi+\bm{u}\times\bm{e}_{B} (4) \displaystyle\nabla^{2}\phi \displaystyle= \displaystyle\nabla\cdot(\bm{u}\times\bm{e}_{B}) (5) Here \bm{u} , p , T , \bm{j} and \phi are the velocity, pressure, temperature, current density and electric potential fields. For conducting cases, we use the thin wall approximation (see 5 and references therein). This model assumes that the currents close within the (thin) walls. Non-dimensional parameters Re , Ha , Pr and Gr are the Reynolds, Hartmann, Prandtl and Grashof numbers respectively, defined as \mbox{{Re}}=U_{0}L/\nu , \mbox{{Ha}}=BL(\sigma/\rho\nu)^{1/2} , \mbox{{Pr}}=\nu k^{-1} , \mbox{{Gr}}=g\beta\Delta TL^{3}\nu^{-2} , where \Delta T is the temperature deviation from a reference temperature. The half-height of the duct L=h/2 and the mean velocity U_{0} in the streamwise direction are used as the length and velocity scales. Finally, \bm{e}_{g} and \bm{e}_{B} are the unit vectors in the directions of acceleration due to gravity and the imposed magnetic field. The equations are solved numerically using our in-house finite-difference solver described in ( 5 ). THE NUSSELT NUMBER To compare the heat transfer efficiency of the different flow regimes, we will use the Nusselt number Nu , defined as ( 2 ) : Nu(x,t)=\frac{q^{\prime}L}{k(T_{w}(x,t)-T_{b}(x,t))} (6) Where q^{\prime} is the heat flux across the duct wall, L the duct half-height, k the thermal conductivity, T_{b} the bulk temperature and T_{w} the wall temperature. For our simulations, we computed Nu(x,t) using both the top and bottom walls and found that the results were similar, for the insulating duct runs the most notable difference was with a 180 phase shift between them. For the conducting duct, the phase difference was not constant. Figure 2 shows an example of Nu(x,t) where this is clearly visible. Figure 2: Nu(x) at the top and bottom walls for Re=4000 , Ha=325 c_{W}=0 and passive heat transport The average Nu(x,t) along the duct length is calculated with equation 7 . Where l_{0} is the starting position for heating and l_{f} is the duct length. Due to the imposed symmetry of the problem, this value is statistically the same when measured from both walls. Nu(t)=\int^{l_{f}}_{l_{0}}Nu(x,t)dx (7) For the remainder of the paper, we take Nu(t) to be the duct-averaged Nu(t) at the top wall ( z=+h/2 ). In addition, \overline{Nu} will be the time average of Nu(t) : \overline{Nu}=\int_{t_{0}}^{t_{f}}Nu(t)dt, (8) where t_{0} is taken once the flow is fully developed. For our case, we take t_{0}=200 convective time units, which is the longest initial transient we encountered (see QM flow in figure 9 ). The end point of our simulations is t_{f}=600 in convective units. SIMULATION PARAMETERS Table 1 shows the parameters used for the simulations performed for this paper. We also performed simulations with c_{W}=0.1 with upward and downward flows and did not see any noticeable effect. The simulations were also performed at Ha=500,750 , however, the results are not shown in this paper. To evaluate the efficiency of the vortex promoters, we also performed simulations with laminar flow at the inlet, but only for the duct in a horizontal position. All simulations were run on a 4096\times 320\times 92 grid with uniform meshing in the streamwise x -direction and strong clustering towards the walls in the y - and z -directions. Simulation Re Ha Pr Gr c_{W} Flow Direction Observed Flow 1 4000 325 0.02 10^{6} 0.0 Horizontal QH 2 4000 325 0.02 10^{6} 0.1 Horizontal UL 3 4000 325 0.02 10^{6} 0.0 Upwards QM 4 4000 325 0.02

Read the complete text