Parameter identification of a lumped parameter thermal model for a permanent magnet synchronous machine

In the thermal modeling of electric machines by lumped parameters, an important step is the tuning of influential poorly known parameters to calibrate the model. The use of Inverse methods based on the minimization of residuals between measured and calculated temperatures is therefore of great help. In this paper, the Gauss-Newton (GN) method, the Levenberg-Marquardt (LM) method and the Genetic Algorithms (GA) method are used to solve this optimization problem in order to identify 10 parameters of a lumped parameter thermal model for a permanent magnet synchronous machine (PMSM).