A neural network approach for the detection of the locking position in RFX

Recently in RFX, where wall locked modes were always present, a new technique has demonstrated the possibility to induce a continuous rotation of the modes with respect to the wall. In this technique the non-linear coupling of the m=0 and m=1 modes has been used to decouple the modes themselves. In the present experiments the mode rotation is induced with a preprogrammed waveform of a toroidal magnetic field rotating ripple. A feedback system able to create a continuous rotation with variable and increasing speed is now under implementation. A Neural Network (NN) has been developed to identify the locked mode position. In the paper different NNs are presented, discussed and compared.