Beside conventional technique, ther

Beside conventional technique, there are two different approaches for parameter
estimation which are online and offline technics. The former one is using Kalman Filter
[2] and least square technics [3]. The latter one is offline [4] curve generation for the
experimentally measured data. Recently, artificial neural networks and various
4
evolutionary algorithms are used with both online and offline methods. In the literature,
Wishart and Harley [5] experimented a method that uses artificial neural networks for
induction motor parameter estimation in current and speed control [6]. Linear technics
based on dynamic model and neuro-fuzzy methods are also proposed for estimation of
induction motor parameters [7,8]. In [9-11] estimation of the stator resistance, transient
inductance and rotor resistance online has been discussed. An interesting approach for
tuning the rotor resistance is proposed in [12] based on model reference adaptive system
schemes [13]. Since some of the above approaches require derivative of the function
which is not always available or may be difficult to calculate, deterministic approaches
often cannot find optimal solutions [1].
applications. In the conventional technics, estimation of induction motor parameters is
based on no - load and blocked - rotor tests [1].