Identification of induction motor using a genetic algorithm and a quasi-Newton algorithm

This paper describes an identification procedure by output error to estimate the parameters of an induction motor. This approach requires an optimisation algorithm. Two different strategies are considered: a genetic algorithm and a quasi-Newton algorithm. The last one requires the evaluation of the gradient and the Hessian in order to minimise a criterion in opposition to the first one. These two techniques have the advantage to be applied to linear and nonlinear models. They can take into account some constraints such as knowledge on electrical parameters. The experimental results prove the effectiveness of these two algorithms.