Fusion Science and Technology / Volume 77 / Number 7-8 / November 2021 / Pages 676-686
Technical Paper / dx.doi.org/10.1080/15361055.2021.1912568
Articles are hosted by Taylor and Francis Online.
The possibility of thermal strain–induced delamination was anticipated in the original design of the NSTX-Upgrade (NSTX-U). Current “bunching” at the toroidal field (TF) flag causes nonuniform Joule heating of the TF conductor during a shot. This produced through-thickness tensions beyond the measured capacity of the insulation bond. This however occurred where the torsional shear was at a minimum, and the Upgrade design progressed with the understanding that delamination at the core of the TF flag might occur. During the Recovery project design reviews, concern over the extent of delamination was elevated. In various early simulations, the tensile stresses reached 50 MPa. With more accurate through-thickness insulation modulus and thickness, and at end-of-flattop versus end-of-pulse temperature, the tensile stress goes down to 25 MPa, and possibly lower based on higher-fidelity modeling. Insulation delamination has been predicted analytically in coils in projects other than NSTX-U, and indications of delamination have been observed in some machines.
Out-of-plane loads on the NSTX TF coil produce local and global twisting of the tokamak. The inner leg supports part of the torsion as a torque shaft or tube. Peak torsion is at the outside radius of the TF central column, away from the regions that don’t carry current, experience less Joule heat, and develop tension. Testing of the shear and tensile fatigue properties of the CTD-425 system was repeated and was a part of this requalification effort.
This paper addresses simulations done in ANSYS based on a simple assumption that when the tension and shear stresses exceed an allowable, the elements are “killed” or considered much less capable of carrying tension and shear loads established by fatigue S-N tests. A competing and complementary method, the Virtual Crack Closure Technique (VCCT) was used to augment and validate the EKILL procedure. Determination of Paris constants to support the VCCT analysis is described Fig. 8.