Identification of Brushless Motor Based on Parallel Hammerstein Model

For the problem that the linear model of brushless motor can't fully reflect its dynamical characteristic, a parallel Hammerstein nonlinear model is applied to the identification of brushless motor. Firstly, the static nonlinear modes of Hammerstein model are established by analogy with Volterra series models. Then, in order to estimate frequency response functions of linear dynamic subsystems, a method based on the minimum mean square error criterion is proposed. Finally, the model of a certain type brushless motor is identified by the proposed method. The experiment results show that the identified nonlinear model has higher accuracy than conventional linear model and needs less calculating quantity than Volterra series model.