Quench simulation of a CICC model coil subjected to longitudinal and transverse field pulses
✍ Scribed by Soren Prestemon; Stacy Sayre; Cesar Luongo; John Miller
- Publisher
- Elsevier Science
- Year
- 2000
- Tongue
- English
- Weight
- 218 KB
- Volume
- 40
- Category
- Article
- ISSN
- 0011-2275
No coin nor oath required. For personal study only.
✦ Synopsis
This paper describes the work done under a collaboration between the National High Magnetic Field Laboratory (NHMFL), and Indian Institute for Plasma Research (IPR). The objective of our work was to simulate quench initiation and evolution as a tool to help IPR with data interpretation on a series of model coil experiments. The model coil consisted of a solenoid pancake-wound with a NbTi CICC (forced ¯ow SHe at 4.5 K, 10 kA rated current at 5 T). The intent of the experiment was to determine the conductor stability when subjected to longitudinal and transverse magnetic ®eld pulses similar to those present in the full-size SST-1 fusion reactor. For that purpose, the model coil was mounted in a rather complex setup. The NHMFL's role was to create a computer simulation model to assist IPR in experimental data interpretation. The computer code Gandalf (L. Bottura, Quench analysis of large superconducting magnets. Part I: Model description, Cryogenics 1992;32(7):659 was used to simulate the quench tests performed on the model coil. The simulations uncovered the need to modify the standard AC loss models to signi®cantly increase the contribution of the longitudinal (toroidal) pulsating ®eld to match the observed quench initiation. With these modi®cations (pointing to higher than expected AC losses), the computer simulations matched the experiment very closely in terms of quench initiation. Given the complexity of the geometry (leading to very non-uniform ®eld distributions), quench evolution was also complex, with multiple initial normal zones that recovered, grew, or coalesced depending on the location. The computer simulations were able to shed light on the quench evolution; however, agreement with the experimental data was not achieved, pointing to a possible problem with the calibration methods used in the experiment.