On Line Neurofuzzy Modeling for Permanent Magnet Synchronous Machines
Edmary Altamiranda, Essam S Hamdi · Chalmers Publication Library (Chalmers University of Technology) · 2006
-This paper presents a neurofuzzy structure for on line modeling of Permanent Magnet Synchronous Machines (PMSM). The model structure is based on recurrent fuzzy neurons (RFN) presented in [1,2], which are used to synthesize a single layer RFN network for nonlinear discrete state space representation of the PMSM dynamics in the d,q rotating reference frame. The proposed scheme allows obtaining on line, a time varying nonlinear model with an appropriate structure for linearizing control laws design. The efficiency of the neurofuzzy structure for PMSM modeling is illustrated by computer simulations. Key-Words: Fuzzy Systems, Neural Dynamics, Nonlinear Models, Permanent Magnet Synchronous Machines.