Adaptive Neuro-fuzzy inference systems into squirrel cage induction motor drive: Modeling, control and estimation

This paper presents application of adaptive neuro-fuzzy inference system (ANFIS) into a squirrel cage induction machine towards modeling, control and estimation. This paper contributes (i) Development of a simple and more realistic model of the induction motor using ANFIS. Using ANFIS, the parameter sets of the motor model are estimated. The simplified model contains eleven estimated parameters. In this paper, a new estimation technique for modeling of induction motor is presented. The identified model can be utilized for electric drives. (ii) Speed, torque and flux control using direct torque control (DTC) algorithm with ANFIS (iii) Design of Estimator through ANFIS which estimates the stator resistance with reference to the temperature when the DTC algorithm is involved. Better estimation of stator resistance results in the improvements in induction motor performance using DTC thereby facilitating torque ripple minimization. The values of stator voltage (Vs), stator current (Is) and rotor angular velocity (omega_r) are taken from the free acceleration test data of 5 HP motor for simulation.