An alternative approach to estimate load torque in industrial environment using neural networks

Abstract

The induction motors are largely used in several industry sectors. The dimensioning of an induction motor has still been inaccurate because in most of the cases the load behavior in its shaft is completely unknown. The proposal of this paper is to use artificial neural networks as a tool for estimating the load torque applied to the induction motor shaft rather than conventional methods, which use classical identification techniques and mechanical load modeling. Simulation results are also presented to validate the proposed approach.