Neuro-Fuzzy Approach to Estimate the Torque in Three-Phase Induction Motors with Unbalanced Power

The induction motors are the key elements of electromechanical energy conversion at various productive sectors. Torque is one of the major variables involved in the mechanical behavior of the motor and can be used in monitoring the energy efficiency in the design of the machine by knowing the load coupled to the shaft and also at fault diagnosis. However, the unbalanced diet stator winding provides electromagnetic torque oscillations in the different operation regimes. This paper presents an estimator applied torque in induction motors by means of fuzzy systems operating at steady state voltages unbalanced, presenting as input only the current system effective. This methodology shows the use of the tool Neuro-Fuzzy Anfis Matlab for adjusting the membership functions and establishment of rules and has the advantage of low computational cost, and express nonlinear dynamics of the machine in case of non-ideal power and variation parametric, mainly due to change in temperature of the windings. Simulation results are presented to demonstrate the behavior of the developed system and the validation of the proposal.