Three-level Inverter-fed Induction Motor Drive Performance Improvement with Neuro-fuzzy Space Vector Modulation

Abstract

—The space vector modulation technique is an optimal pulse-width modulation technique used for inverter control. This article presents a neuro-fuzzy-based space vector modulation technique for a three-level inverter fed induction motor. It uses a hybrid learning algorithm (combination of back-propagation and least-squares methods) for training the input–output data pattern. The training data for neuro-fuzzy-based space vector modulation are generated from the conventional simplified space vector modulation method. The proposed scheme uses the space vector rotation angle and change of rotation angle information as input and generated duty ratios as output. The dynamic and steady-state performance of a neuro-fuzzy-controlled induction motor drive is compared with the conventional space vector modulation and neural network-based space vector modulation methods. The simulation results obtained are verified experimentally using a dSPACE kit (DS1104). The performance measure in terms of the total harmonic distortion of inverter line–line voltage of the neuro-fuzzy-based system is compared with the neural network-based space vector modulation and conventional space vector modulation methods.