Estimation of two-axis synchronous machine parameters using non-deterministic tools

Optimization is employed in this work to obtain the two-axis parameters of a typical solid-rotor synchronous machine. SSFR data is used to set-up the identification problem, leading to a fitness function that must be minimized. However, the function is plagued with local minima, making the parameter identification impossible with deterministic methods. Genetic Algorithms have become an important tool to deal with local minima and software codes are readily available. Determination of the global/true minimum is not easy though since big computational resources are required even for relatively small problems. Moreover, the final minimum may not be found in spite of proper tuning of the Genetic Algorithm. Some tuning aspects are addressed in this work.