IDENTIFICATION OF NONCIRCULAR PLASMA EQUILIBRIA USING A NEURAL-NETWORK APPROACH

A technique using a single hidden layer backpropagation neural network is described to establish a nonlinear mapping between a set of magnet ic flux measurements and some shaping parameters of a non-circular pla sma. The technique has been applied for the identification of limiter and X point equilibria in the ASDEX Upgrade geometry; the dataset of e quilibria required for training and testing the neural network has bee n generated by means of an integrated use of a fixed and a free bounda ry MHD code. The average accuracy of the identification procedure is q uite good, with a further improvement if a linear connection between t he input and output layers is introduced. A procedure is also proposed for the selection of the optimum location of a limited number of sens ors. The relationship existing between the behaviour of the neural net work and some statistical parameters of the dataset is analysed and di scussed.