Estimativa da velocidade de rotação em motores de indução trifásicos usando histograma de gradientes orientados

Abstract

Three-phase induction motors are widely used in the production sector, for the most diverse segments, particularly given their robustness and relatively low cost when compared to machines of the same power. Measuring the rotation of its axis, for example, can assist in identifying the behavior of the motor, especially when it is subject to load conditions, or torque, different from those for which it was specified.In this sense, the present work proposes modeling oriented to the estimation of the rotation speed of the axis, of a three-phase induction motor with squirrel-cage rotor, using the Oriented Gradients Histogram technique, widely used in the image processing area and to predict the rotation speed of a neural network with the Levenberg-Marquardt algorithm was trained in different operational conditions.The results obtained allowed us to conclude that the use of histograms can serve to develop a computational system capable of estimating the axis rotation of an induction motor in real-time.