A novel evolutionary technique to estimate induction machine parameters from name plate data

Owing to the fact that the performance and control design of large scale induction machines depend on accurate knowledge of its equivalent electrical circuit parameters, precise identification of these parameters is essential. Current methods used to quantify induction machine parameters call for performing several experimental testing such as no-load, locked-rotor and DC tests which may not be available due to the lack of hardware, experience and time required to perform the tests. In this paper, two different evolutionary computational techniques namely; bacterial foraging and genetic algorithm, are employed to estimate these parameters from machine nameplate data without conducting any experimental measurements. The accuracy of the proposed techniques is assessed through their application on squirrel cage and wound rotor induction motors of different ratings. The motors performance computed using the proposed techniques is compared with that computed using classical practical measurements. The obtained results reveal the ability of evolutionary techniques to estimate the equivalent electrical circuit parameters of induction machines with a reasonable degree of accuracy. Results also show that bacterial foraging approach is more accurate than genetic algorithm in estimating induction machine parameters.