Identifying dynamic model parameters of a BLDC motor

An off-line identification method founded on the least-squares approximation technique and a closed-loop disturbance observer is applied for identifying the parameters of a BLDC motor model. No special configuration of the motor is required besides the availability of experimental data for back-EMF, phase currents, rotor position, and rotor speed. This method is used to identify the back-EMF harmonics and mechanical parameters, where the mechanical parameters refer to cogging-torque, Viscous friction coefficient, and Coulomb friction coefficient. The proposed identification method is theoretically investigated, and the method's effectiveness is proved by experimental results performed on a low-power BLDC motor. (C) 2008 Elsevier B.V. All rights reserved.