An intelligent adaptive vector control technique using a multi-neural networks based hybrid structure
Kurosh Madani, Abdennasser Chebira, Jean-Charles Depecker, Gilles Mercier · 2003
One of the most important tasks for a machine control process is the system identification. To identify varying parameters which are dependent to speed, control voltages and currents, one must have an adaptive control system. In the case of the synchronous machines the vector control concept is used. It supposes an appropriated plant's model taking into account internal parameters. In the real applications, the machine's parameters vary in a nonlinear way and are not constant. We propose a neural controller with a multi-neural networks based plant identifier. Simulations and experimental results based on a real plant database are reported validating our multi-neural networks based approach.