Skip to Main Content
In this paper a hardware design of a behaviour predictor of a nuclear fusion plasma is presented. It is based on a fuzzy neural network modelization of the nonlinear relation existing among the six forcing magnetic fields and the geometric displacement of the high-temperature plasma. This approach fits quite well with the intrinsic nonlinearity of this phenomena, providing a good degree of approximation. Each design step is described and the final board is suitably tested in order to validate the proposed approach.