Thermal-stress analysis and testing of DIII-D armor tiles

It is planned to install about 1500 new armor tiles in the DIII-D tokamak. The armor tiles currently installed in DIII-D are made by brazing Poco AXF-5Q graphite onto Inconel X-750 stock. A small percentage of these have failed by breakage of graphite. These failures were believed to be related to significant residual stress remaining in graphite after brazing. Hence, an effort was undertaken to improve the design with all-graphite tiles. Three criteria must be satisfied by the armor tiles and the hardware used to attach the tiles to the vessel walls: tiles should not structurally fail, peak tile temperature must be less than 2500 K, and peak vessel stresses must be below acceptable levels. A number of alternate design concepts were first analyzed with the two-dimensional finite element codes TOPAZ2D and NIKE2D. Promising designs were optimized for best parameters such as thicknesses, etc. The two best designs were further analyzed for thermal stresses with the three-dimensional codes P/THERMAL and P/STRESS. Prototype tiles of a number of materials were fabricated by GA and tested at the Plasma Materials Test Facility of the Sandia National Laboratory at Albuquerque. The tests simulated the heat flux and cooling conditions in DIII-D. This paper describes the 2-D and 3-D thermal stress analyses, the test results and logic which led to the selected design of the DIII-D armor tiles. 5 refs., 7 figs., 3 tabs