Fuzzy logic and neural net current control on a salient pole permanent magnet synchronous machine

In this paper a digital fuzzy logic controller (FLC) is developed to control the sinusoidal currents of a salient pole permanent magnet synchronous machine. This fuzzy controller presents a good performance even at higher speeds, where the e.f.c.m. normally introduces a considerable steady state error in other digital current controllers. After tuning the FLC is replaced by a neural net, trained by the Levenberg-Marquardt method, with considerable reduction in processing time. Simulated and experimental results are presented for the FLC, the neural and a PI controller for performance comparison.