Characterization of induction machines with a genetic algorithm

In this paper, the characterization of an induction machine is formulated as a nonlinear estimation problem. The solution of the estimation problem is achieved with a genetic algorithm that aims at minimizing a quadratic cost function. The data sources used in the proposed formulation correspond to the values of the current and power absorbed by the machine in standard laboratory tests. The characterization obtained with the proposed technique is compared with that obtained with the classical no-load and locked rotor tests either by simulation or experimentally