Lithium or any lithium alloy like the lithium lead alloy Pb-17Li is an attractive breeder material used in blankets of fusion power reactors because it allows the breeding of tritium and, in the case of self-cooled blankets, the transfer of the heat generated within the liquid metal and the walls of the cooling ducts to an external heat exchanger. Nevertheless, this type of liquid-metal-cooled blanket, called a self-cooled blanket, requires specific design of the coolant ducts, because the interaction of the circulating fluid and the plasma-confining magnetic fields causes magnetohydrodynamic (MHD) effects, yielding completely different flow patterns compared to ordinary hydrodynamics (OHD) and pressure drops significantly higher than there. In contrast to OHD, MHD flows depend strongly on the electrical properties of the wall. Also, MHD flows reveal anisotropic turbulence behavior and are quite sensitive to obstacles exposed to the fluid flow.

A comprehensive study of the heat transfer characteristics of free and forced convective MHD flows at fusion-relevant conditions is conducted. The general ideas of the analytical and numerical models to describe MHD heat transfer phenomena in this parameter regime are discussed. The MHD laboratory being installed, the experimental program established, and the experiments on heat transfer of free and forced convective flow being conducted are described. The theoretical results are compared to the results of a series of experiments in forced and free convective MHD flows with different wall properties, such as electrically insulating as well as electric conducting ducts. Based on this knowledge, methods to improve the heat transfer by means of electromagnetic/mechanic turbulence promoters (TPs) or sophisticated, arranged electrically conducting walls are discussed, experimental results are shown, and a cost-benefit analysis related to these methods is performed. Nevertheless, a few experimental results obtained should be highlighted:

1. The heat flux removable in rectangular electrically conducting ducts at walls parallel to the magnetic field is by a factor of 2 higher than in the slug flow model previously used in design calculations. Conditions for which this heat transfer enhancement is attainable are presented. The measured dimensionless pressure gradient coincides with the theoretical one and is constant throughout the whole Reynolds number regime investigated (Re = 103 → 105), although the flow turns from laminar to turbulent. The use of electromagnetic TPs close to the heated wall leads to nonmeasurable increase of the heat transfer in the same Re regime as long as they do not lead to an interaction with the wall adjacent boundary layers.

2. Mechanical TPs used in an electrically insulated rectangular duct improved the heat transfer up to seven times compared to slug flow, but the pressure drop can increase also up to 300%. In a cost-benefit analysis, the advantageous parameter regime for applying this method is determined.

3. Experiments performed in a flat box both in a vertical and a horizontal arrangement within a horizontal magnetic field show the expected increase of damping of the fluid motion with increasing Hartmann number M. At high M, buoyant convection will be completely suppressed in the horizontal case. In the vertical setup, the fluid motion is reduced to one large vortex leading to a decreasing heat transfer between heated and cooled plate to pure heat conduction.

From an analysis of the experimental and theoretical results, general design criteria are derived for the orientation and shape of the first wall coolant ducts of self-cooled liquid metal blankets. Methods to generate additional turbulence within the flow, which can improve the heat transfer further are elaborated.