P. Pugnat et al., Statistical diagnosis method of conductor motions in superconducting magnets to predict their quench performance, IEEE APPL S, 11(1), 2001, pp. 1705-1708
Premature training quenches are usually caused by the transient energy rele
ased within the magnet coil as it is energised, Two distinct varieties of d
isturbances exist, They are thought to be electrical and mechanical in orig
in. The first type of disturbance comes from non-uniform current distributi
on in superconducting cables whereas the second one usually originates from
conductor motions or micro-fractures of insulating materials under the act
ion of Lorentz forces. All of these mechanical events produce in general a
rapid variation of the voltages in the so-called quench antennas and across
the magnet coil, called spikes. A statistical method to treat the spatial
localisation and the time occurrence of spikes will be presented. It allows
identification of the mechanical weak points in the magnet without need to
increase the current to provoke a quench. The prediction of the quench lev
el from detailed analysis of the spike statistics can be expected.