Estimación de la velocidad del motor de inducción en el rango de baja velocidad utilizando inteligencia artificial para el control vectorial Sensorless.

Abstract

In the present work the study, design and implementation of a speed estimation technique, using an Artificial Neural Network (ANN), for a three-phase induction motor type Squirrel Cage is approached. The main objective of this technique is to estimate the speed of the engine from its own conditions at low speed, in order to apply vector control without the use of a traditional speed sensor that feeds back to the control loop the speed of rotation the motor. When replacing the traditional sensor with an RNA, a "Sensorless" type control loop is obtained. As a first step, it starts with obtaining the mathematical model of the induction motor to be used. Once this model is obtained, the simulation of the control system subjected to various working conditions is carried out. In a second instance, the technique of vectorial control is applied, obtaining the set of pertinent equations for the control of the speed of the motor. The third step is to modify the control loop to replace the speed sensor with the RNA, previously trained. The training of the RNA to use, is done collecting data in different ranges of engine speed, both empty and loaded. The simulation of the system is carried out on the Simulink platform of the Matlab software package. As a next step to the simulation of the system, the control strategy is implemented using the Raspberry Pi 3B card.