Self-constructing fuzzy neural network speed controller for permanent-magnet synchronous motor drive

A self-constructing fuzzy neural network (SCFNN) which is suitable for prac tical implementation is proposed in this study. The structure and the param eter learning phases are performed concurrently and online in the SCFNN. Th e structure learning is based on the partition of input space and the param eter learning is based on the supervised gradient decent method using a del ta adaptation law. Several simulation and experimental results are provided to demonstrate the effectiveness of the proposed SCFNN control stratagem w ith the implementation of a permanent-magnet synchronous motor (PMSM) speed drive. Moreover, the simulation results of time varying and nonlinear dist urbances are given to show the dynamic characteristics of the proposed cont roller over a broad range of operating conditions.