Improvement of an induction motor drive-based direct torque control strategy using a neuro-fuzzy controller

This work describes how a Neuro-Fuzzy network (ANFIS, Adaptive Network-based Fuzzy Inference System) is used to control an Induction Motor Drive-based Direct Torque Control strategy to improve the dynamic behavior of the system and simplify the conventional direct torque control scheme. The Neuro-Fuzzy controller's structure guides the torque and stator flux error signals through the fuzzy inference to get an output that takes the form of a space voltage vector to be modulated by the inverter. Besides this intelligent controller, Space Vector Modulation is applied to avoid hysteresis bands and switching tables which are common in the conventional scheme. Under this frame it is possible to set a constant value for the switching frequency and, in general, improve the dynamic behavior of the system. Simulations are presented to show that the proposed Intelligent System is characterized not only by simple mathematical models, but also by very fast torque and flux responses. Furthermore, zero-steady-state error in torque and flux is gotten thanks to inverse learning and the tuning algorithm applied to the network. This control also provides a natural and well-defined human-machine interaction.