Magnetohydrodynamic Heat Transfer for Fusion Energy

Abstract

This thesis uses measurements from a new experimental flow loop and new analysis of existing literature data to provide a heat transfer correlation for turbulent magnetohydrodynamic flows. Fluids used as breeder/coolant in fusion power plant blankets must provide adequate heat transfer of up to 10 MW/m2 within strong magnetic fields . Because the two most popular choices for this function (liquid metals and molten salts) are electrically conductive, magnetohydrodynamic effects can alter the fluid flow fields, and therefore heat transfer, in these conditions. The primary mechanism by which these magnetohydrodynamic effects change the flow field is turbulence damping, where sufficiently high magnetic fields can fully relaminarize a flow and decrease heat transfer.

A flow loop was designed, constructed, and operated to measure heat transfer in a circular pipe with electrically conducting walls through a transverse magnetic field to study the impact of B field on modestly conducting fluids. Aqueous potassium hydroxide is used as a high Prandtl number simulant fluid for molten salts and is flowed through a heated test section with 15 diameters of flow development within a uniform transverse magnetic field of up to 1.7 T (provided by a copper dipole) and Reynolds numbers up to 15,000. Changes in wall temperature are used to measure magnetohydrodynamic Nusselt numbers. Due to the moderate electrical conductivity of the working fluid (~100 S/m), it is only possible to achieve high enough Hartmann numbers for partial flow relaminarization in the experiment. The flow loop allows for qualitative characterization of behavior in regions of steep magnetic field gradients and transitional flows .

This thesis uses existing liquid metal literature data to develop a new form of heat transfer correlation relating the heat transfer coefficient to Reynolds and Hartmann numbers of the flow over the full range of turbulent MHD flows (i.e. from non-MHD to fully relaminarized) for low Prandtl number fluids . Using the newly developed heat transfer correlational form, the experimental high Prandtl number data is extrapolated to propose a new heat transfer correlation for molten salts at any Hartmann number and Reynolds numbers up to 15,000. While the achievable experimental conditions only produce relative drops in heat transfer coefficients of around 20%, the analysis quantitatively predicts Nusselt number drops of up to a peak of >90% at relaminarization. Given the constrained design window, including these potentially large magnetohydrodynamic effects will be necessary in engineering successful fusion blankets.