Algorithms for automatic torus motor parameters identification: Comparative study

Purpose - The parameters of axial-field machines are very small compared with the parameters of conventional machines. Different measuring methods are normally used in order to obtain good estimates of the machine parameters. These methods are difficult to perform, costly and time consuming. This paper proposes the use of genetic algorithms to predict the self and mutual inductances of a specific type of axial-field machine, the Torus motor. Design/methodology/approach - The parameter extraction is reformulated as a search and optimization problem in which the only requirement is a set of values of current versus time and an approximate estimate of the parameters. Findings - The predicted machine self and mutual inductances are verified by comparing with several measuring methods and excellent agreement is obtained. Originality/value - Demonstrates that genetic algorithms can predict the self and mutual inductances of the Torus machine automatically with high accuracy.