Avoidance of m=2, n=1 tearing mode wall-locking by torque-balance control with magnetic feedback in DIII-D and RFX-mod

Recent experiments performed in DIII-D and RFX-mod operated as a tokamak demonstrate that magnetic feedback realized by active saddle coils is able to prevent wall-locking of tearing modes. Disruptions caused by wall-locking of the m=2, n=1 tearing mode (TM) are considered one of the most critical roadblocks to achieve reliable operations in tokamak devices, and these experiments demonstrate a possible new tool for avoiding such disruptions. The result is remarkable, given the many differences of the two experiments both in terms of layout and plasma condition. In DIII-D we performed divertor, D-shaped, high-? plasmas, controlled by coils placed inside the vacuum vessel. Instead, in the RFX-mod experiments circular crosssection, limiter, ohmic discharges are realized with low edge safety factor 2<q(a)<3, and feedback control by active coils placed outside the stabilizing shell. The main result of the experiments is that feedback pushes a TM - which would otherwise lock to the wall - into slow rotation when the applied gain is above a critical value. This threshold depends on the machine layout and feedback time delays. We have developed a simple electromagnetic torque balance model, which successfully describes the observations in both devices and seems to indicate that the phenomenon is robust. Avoidance of wall-locking is generally able to prevent the related disruption in DIII-D. On the contrary, in RFX-mod this technique is necessary, but not sufficient, since under certain parameter regimes the m=2, n=1 induced disruptions are not due to a complete wall-locking but only to a slowing down of the mode rotation. The different behavior in the two devices will be discussed along with parametric dependences of disruption onset.