Fig. 1: The layout of the saddle coils in the RFX-mod experiment.

In Tokamak advanced scenarios for ITER and for future fusion reactors new magneto-hydrodynamic (MHD) instabilities are predicted to grow, whose active stabilization will be mandatory to achieve the expected performance. This class of instabilities are called Resistive Wall Modes, since their growth rate depends on the electrical resistance of the conducting wall that surrounds the plasma column. A perfectly conducting wall would fully stabilize these modes, but that is unfeasible in practice. As a consequence, active stabilization by means of suitable coils and feedbackcontrolled amplifiers will be needed in future fusion devices.

In the modified RFX experiment (RFX-mod), operating at Consorzio RFX in Padova, such active stabilization coils were installed as part of a programme to obtain a more stable plasma in RFX, and to conduct Resistive Wall Mode stabilization experiments. The experiment implements active feedback control of the magnetic field to effectively produce and maintain an almost ideally conducting wall at the plasma boundary. Similar experiments have been performed at low current on the EXTRAP-T2R device and on the DIII-D tokamak with a partial set of feedback coils.

To accomplish this, RFX-mod includes a new active MHD control system consisting of 48 toroidal x 4 poloidal saddle coils (see fig. 1), which are independently powered. They have been installed and tested during 2003-2004. The system has then been used under various control scenarios including experiments on local radial field cancellation over the entire torussurface to mimic an ideal wall, which is also called a “Virtual Shell”.

Fig.3: Comparison between the toroidal mode spectrum without (blue) and with (red) active control (virtual shell scenario). The amplitudes of the radial field components measured at the sensors located on the vacuum vessel are reduced to values of around 0.1 mT. The strong reduction of the mode amplitudes has a clear benefit in reducing energy transport, with a significant impact on plasma performance (see fig.2)

Fig.2 (left): Plasma current and toroidal loop voltage waveforms. Comparison between a typical shot without active control (blue) and one of the recent shots with the virtual shell active control scenario (black). The benefit of active control is evident both from the longer duration of the current pulse and from the lower loop voltage, which means less power losses.

Both in Tokamaks and Reversed Field Pinches, the MHD instabilities can be seen as helical deformations (“modes”), which are easily detected by radial magnetic field measurements. If a suitable number of coils is available, the feedback action can effectively reduce the amplitude of the modes across the whole spectrum.

During the experiments, successful Virtual Shell operation has been achieved leading to:

• A threefold increase in pulse length, from 100 to 300 ms, and well-controlled pulses up to nearly 800 kA of plasma current (see fig.2)

• A one order of magnitude reduction of the radial field at the plasma edge (see fig.3)

• A twofold increase in global energy confinement time.

When operating without active control, resistive wall modes are observed to grow in agreement with linear MHD theory predictions, i.e. on the timescale for the diffusion of the radial field through the shell. Recent experiments have clearly demonstrated feedback stabilization of resistive wall modes; both single and multiple helical modes have been successfully stabilized.

Overall, the results obtained in the first year of operation are very encouraging, and form a solid base for an intense experimental program which will focus on the study and control of the MHD dynamics, and on improving the confinement by increasing the plasma current to the mega-ampere range.