Experimental investigations of MHD flows in a sudden expansion

Abstract

Reliable theories and models concerning MHD flows in manifolds for fusion applications are important, since most of the pressure drop occurs in those parts of the blanket. Abrupt changes of flow direction in feeding and draining lines as well as expansions and contractions cause flow redistribution and additional pressure drop. Numerical and asymptotic models have been developed during the last years, which require validation on the basis of precise experimental data. Since experimental results available in literature cover only the range of moderate magnetic fields, i.e. Ha < 300, and experiments in electrically conducting sudden expansions do not exist, the experimental data base has to be considerably extended in order to cover the fusion relevant range of parameters. For that reason a forced convective MHD experiment in a sudden expansion of rectangular ducts has been designed, manufactured and inserted into the liquid metal NaK-loop of the MEKKA laboratory at the Forschungszentrum Karlsruhe in the frame of the EFDA Technology Work Program TW2-TTBA-006a Dl. Experiments have been performed in which the induced electric potential is measured on the surface of the expanding test section by more than 300 surface potential probes. The pressure variation along the duct has been measured at 16 different axial positions. For high interaction parameters the pressure measurements confirm the theoretical results obtained by an asymptotic theory valid for inertialess flows. Local quantities of the flow, e.g. potential gradients, are inferred with traversable probes at different locations in the ducts. A comparison of experimental and available theoretical results shows quite good agreement and confirms both the validity of the used theoretical tools and the good quality of the measurements. Moreover, due to its valuable contribution to the data base of fundamental MHD flows these experimental results may serve as benchmark basis for future developments of computational MHD tools.