Development, installation, and initial operation of DIII-D graphite armor tiles

An upgrade of the DIII-D vacuum vessel protection system has been completed. The ceiling, floor, and inner wall have been armored to enable operation of CIT-relevant double-null diverted plasmas and to enable the use of the inner wall as a limiting surface. The all-graphite tiles replace the earlier partial coverage armor configuration which consisted of a combination of Inconel tiles and graphite brazed to Inconel tiles. A new all-graphite design concept was chosen for cost and reliability reasons.The ten minute duration between plasma discharge required the tiles to be cooled by conduction to the water-cooled vessel wall. Using two and three-dimensional analyses, the tile design was optimized to minimize thermal stresses with uniform thermal loading on the plasma-facing surface. Minimizing the stress around the tile hold-down feature and eliminating stress concentrators were emphasized in the design. The design of the tile fastener system resulted in sufficient hold-down forces for good thermal conductance to the vessel and for securing the tile against eddy current forces.The tiles are made of graphite, and a program to select a suitable grade of graphite was undertaken. Initially, graphites were compared based on published technical data. Graphite samples were then tested for thermal shock capacity in an electron beam test facility at the Sandia National Laboratory (SNLA) in Albuquerque, New Mexico, USA.The expected operational heat load to the tiles is 500 W/cm2 for 10 s over the entire tile during limiter operation and 1000 W/cm2 over a 2 cm wide band for 10 s for divertor operation. Comparative stress analysis was performed on the basis of 1000 W/cm2 uniformly distributed on the front surface for 5 s with 10 min between pulses. Prototype tiles survived beam heating to the limit of the beam system which was 1000 W/cm2 for 4 s or 1750 W/cm2 for 1.75 s. The electron beam facility at the Sandia National Laboratory was used to deposite higher heat flux in a 2 cm wide band in order to model divertor heat flux conditions. These tests were successful with heat fluxes up to 2500 W/cm2 for 5 s. Both the ion beam and the electron beam tests produced surface temperatures up to 3100 °C which resulted in minor surface erosion but no structural cracking.Installation of approximately 1600 armor tiles was accomplished in November/December 1987. Initial operation experience with the new armor system is summarized.