@article{G.SRINIVAS_2021, title={ARTIFICAL INTELLIGENCE BASED METHODS FOR SENSORLESS VECTOR CONTROL OF INDUCTION MOTOR}, volume={9}, url={https://innovareacademics.in/journals/index.php/ijet/article/view/42147}, abstractNote={<p>Three phase induction motors are the most sought after machines in industry as they are relatively less costly and rugged due to the absence of commutator. They are the driving mechanism for majority of operations in industries, agriculture, commercial complexes etc. but in the separately excited DC machines, because of the presence of commutator, the flux axis and the armature axis are always in quadrature. Hence, there is always inherent decoupling between main flux and the armature flux called vector or decoupled control which leads to flexible operation and hence accurate control. Since, induction motor is singly fed, stator current has to meet with both torque and flux requirements. Hence, it is not possible to control both components independently and this is the main cause for the sluggish behavior of induction motors. Induction motor performance can be made similar to that of DC machine by resolving the stator current into flux producing component and torque producing component of current. The difficulty here is the determination of flux axis so that the flux component of stator current can be along that axis. In order to carry out this, the information regarding the exact position of rotor position as well as its magnitude is required. Depending on how the information is collected, vector control is divided in to two classes, namely, direct and indirect field oriented control schemes. In the direct method the use of hall probes or search coils used for flux measurement, destroy the ruggedness of the motor. Instead, rotor position can be estimated using machine models where indirect method or sensor less control aims at using mathematical expressions.</p> <p>The basic foundation needed for vector control of induction motor</p> <p>(IM) is decoupling of stator currents into flux and torque components along the rotor flux axis. For this information the instantaneous rotor position is necessary. Depending on the methods employed for finding rotor position vector control are two types. Direct vector control (DVC) and indirect vector control (IVC). In direct vector control the rotor position is sensed by Hall Effect sensors introduced in the stator. The basic drawback is it introduces harmonics in the output voltage and results in addition of cost and size. In Sensorless vector control (SVC) the rotor position is estimated by using mathematical analysis and machine dynamic model which eliminates speed sensors, encoder and motor shaft extension and hence reduces cost and ruggedness.</p> <p>The basic methods employed for detection of rotor position by sensor less control involve: Conventional methods like Kalman filter method, Conventional PI Controller, Artificial intelligence methods like Fuzzy and ANN and Evolutionary methods like Genetic algorithms and Particle Swarm Optimization.</p> <p>In conventional methods like PI Control method and Extended Kalman filter method of estimation is done by using motor equations to directly compute speed and are prone to numerical and steady state errors. Hence a new method is suggested which employs latest simulating and computing techniques like Artificial intelligence methods like Fuzzy controller and ANN and Evolutionary control methods like GA and PSO are used for multivariable state feedback linearization method whose load torque is estimated by the above suggested artificial and evolutionary methods, hence the errors in the above conventional methods can be minimized.</p> <p>This research aims at designing a controller using conventional methods like PI controller ,Extended Kalman filter and artificial intelligent methods like fuzzy ,neural networks and evolutionary methods like genetic algorithms and particle swam optimization and to find speed and torque responses using parameters like peak overshoot ,peak time rise time etc. and to reduce steady state errors in conventional methods and to implement hardware using FPGA Module for conventional methods like Kalman filter and Conventional PI Controller and compare with Proposed methods like GA and PSO.</p>}, number={1}, journal={Innovare Journal of Engineering and Technology}, author={G.SRINIVAS}, year={2021}, month={Jun.}, pages={1–17} }