Controle Neurofuzzy de motor de indução com estimação de parâmetros e fluxo de estator

Abstract

This work has as main propose the development of a neurofuzzy controller based on the Neo Fuzzy Neuron as an alternative to the conventional PI controllers. This controller, the Online Neurofuzzy Controller (ONFC), has as principal difference compared to the most of the neural and neurofuzzy structures used in control systems, the fact that the error is not only used to correct the network parameters, but also as network input. Moreover, the ONFC presents structure with low computational cost, once has only one input and one output, associated by two fuzzy rules. Take account this structure, the equations that describe the proposed controller and the weight corrections, for single and cascade configurations, are presented. A performance analysis of this controller is presented based on digital simulations and practical implementations. The main application of the ONFC is a vector control of an induction motor. Moreover, the proposed controller development is preceded by two stages. The firs one is related to a procedure to automatically identify the induction motor parameters using the frequency inverter. The second stage presents the development of a stator flux estimator based on the induction motor voltage equations with dc offset compensation of the estimated flux and the measurements of the stator voltage and current. From these stages, all necessary parameters to implement the vector control strategy adopted, as well as the stator flux used for field orientation and flux control are obtained.