Investigations on the Dynamics of Seven Level Inverter Fed Induction Motor Drive with Neural Based Rotor Resistance Estimator

Inverter fed induction motor drive dynamics is greatly dependent on the input voltage pattern. Indirect field oriented control (IFOC) is of industrial standard which is highly sensitive to motor parameter variations. This paper attempts to address the above cited issues. To ensure the quality of the input voltage multilevel inverter (MLI) is employed to feed the motor. Amongst all the motor parameters rotor resistance is of paramount important as it varies with temperature, frequency and skew effects and also the exact value of rotor resistance is required for the unit vector generation and slip calculation in an IFOC scheme. Hence neural network learning based model reference adaptive system (NN-MRAS) for the on-line estimation of rotor resistance (R-r) is utilized. The Neural learning algorithm determines the estimation speed, stability, weight convergence, accuracy of estimation, speed of tracking and ease of implementation. This work proposes a new neural learning strategy for Rr estimation. The entire system is modeled and simulated in Matlab/Simulink. The MLI fed drive is compared with the conventional inverter fed drive. Also the MLI fed drive dynamics are presented with and without the proposed estimator.