On-line rotor time constant estimation for induction motor using two new methods, RLS and SDBP algorithms

This paper presents two new methods of online estimation for the rotor time constant of the induction motor for indirect vector control drives. These methods are presented using artificial neural networks with steepest descent back propagation training algorithm and recursive least square algorithm. These methods use measurements of the stator voltages, stator currents and the rotor speed. The problem is formulated as finding the rotor flux error, which occurs as a result of rotor time constant variation, and then the actual value of the rotor resistance is estimated. The effectiveness of these algorithms is demonstrated with simulation.