Performance analysis of a neural flux observer for a bearingless induction machine with divided windings

This study presents a performance analysis of a neural observer applied to speed and radial position controls of an induction-type bearingless motor with divided windings. The speed and radial control actions are carried out in conjunction with current control in the windings. The neural flux observer aims at compensating possible parametric variations of the machine caused by agents such as temperature or nucleus saturation. Control and estimation of flux are executed by a program developed in ANSI C language. The following resources were used in the DSP: the analog/digital converter, the PWM outputs and the RS-232 parallel and serial interfaces, which are responsible, respectively, for programming the DSP and capturing data using the supervisory system. Compared with a fixed parameter, the observer validates the proposed method.