PSO-Based EKF Estimator Design for PMBLDC Motor


The estimation of motor state variables is an important criterion in the drive performance, especially for high accuracy required, for that reason it’s-necessary to estimate rotor-position-continuously not for sixty-electrical-degrees as in most existing methods. In this work the speed and position for the rotor of Permanent Magnet Brushless DC Motor (PMBLDCM) was estimated by using extended Kalman filter (EKF), this work is divided into two parts, the first one deals with design and simulation of PMBLDCM with EKF as an estimator, the results are introduced by manually selected EKF parameters (Q & R) matrices, The second one deals with investigation the performance of the use of PSO technique to optimize the performance and operation of EKF, the main use of PSO here is to optimize value for EKF parameters (Q and R), the results proved that by tuning the EKF parameters by PSO the estimated values for speed and position is very-close-to the actual value-(estimation-accuracy is increased). The resultant error clearly decreases when tuned by EKF parameters for example at full load case the speed RMS error is 0.24 for 10μs sampling time, although the RMS error is 9 for 10μs sampling time trial and error selected EKF parameters.