Speaker
Description
On July 22, 2016, the NSTX-U project team suspended plasma operations due to the inoperability of the PF1A Upper (PF1A-U) coil. Preliminary indications evidenced that the PF1A-U coil experienced a coolant blockage. An external coolant leak developed from the PF1A-U coil pack after the blockage was attempted to be cleared. A post-mortem physics analysis indicated that an undetected gradual deterioration of coil inductance preceded the coolant blockage in the weeks leading up to the operational suspension.
The subsequent investigation, documentation, and analysis relative to the failure of the PF1A-U are the subject of this paper. Initial non-destructive testing and examination was followed by extensive non-destructive radiography. The radiographic study confirmed the locations of four braze joints and identified five anomalies. Destructive examination was initiated by segmenting the coil pack with cutting planes in feature-free benign regions identified by the radiography results. Cuts were cautiously made in 0.03-inch depth increments through the coil pack at the identified cutting planes.
Two of the three coil sections were removed from the mandrel at the conclusion of cutting. The third section, containing the five anomalies identified in the radiographic study, the coil leads, and two braze joints, was initially left on the mandrel in order to minimize compromising evidence. All three sections were subjected to visual borescope/videoscope inspection through the cooling path, vacuum testing, and electrical testing of every conductor segment. One conductor cooling path visually evidenced a void through the sidewall of the cooling path in a layer-to-layer region. The identified void in the conductor cooling path was not proximal to a braze or joggle. Electrical testing indicated that the voided conductor segment was a member of a group of 14 conductor segments that evidenced low-resistance connectivity. The four braze joints and two lead segments were subjected to 400 psi hydrostatic pressure testing followed by helium leak testing and evidenced that there were no detectable leaks in the tested cooling path segments.
The coil section containing the conductor void was removed from the mandrel and split along a layer plane to provide visual access to the conductor void area. A region proximal to the center of the coil pack evidenced electrical pitting and molten conductor debris. Samples of the epoxy resin insulation system were extracted and analyzed with Dynamic Mechanical Analysis (DMA) and Differential Scanning Calorimetry (DSC) techniques. Epoxy resin insulation samples were also subjected to immersion testing per ASTM D570. Metallurgical samples were extracted from the coil pack conductors and subjected to hardness testing and grain structure analysis.
The sum results and interpretation of these analyses will be presented in the paper.
*This work is supported by US DOE Contract No. DE-AC02-09CH11466
| Eligible for student paper award? | No |
|---|
