The Research A Of Speed Estimation Based On EKF And Improved MRAS

During the recent years, AC speed adjustment technology has a great development. Compared to vector control, direct torque control (DTC) for induction machine developed greatly because of its simple system scheme, better dynamic and static performances. This technology is based on the following control idea: by using the method of vector analysis, first, the motor's parameters of the flux and torque in the stator reference coordinate is calculated, then the inverter switching mode according to the errors of the torque and the flux position in space is chosen directly, thus realizing the high performances control of the induction machine.The paper discussed the speed sensorless induction machine control methods with high performance, and classified them into two categories: methods based on ideal motor model and based on non-ideal characteristics (or based on motor). In the methods based on non-ideal characteristics (or based on motor), the signal of rotor slot harmonics is used in the method and the high frequency signal injection is explained. In the methods based on deal motor model direct estimate, model reference adaptive system(MRAS), full order adaptive observer(Luenberger observer, LO), nerve network and extended Kalman filter(EKF) are introduced.How to use EKF to achieve speed-sensorless technology in detail is studied in this paper and the improvement is given here. The validity is verified through simulation. Because of complexity in computation and hardness in implement for using full order EKF, the reduced order extended Kalman filter (ROEKF) for rotational speed estimation of induction motor is presented. Based on the three order state equation of motor, this algorithm just needs measuring the stator voltage and current to estimate the speed. More importantly, there is an easier computation in this algorithm.This paper presents a novel method of rotor speed estimator based on various research of speed sensorless. It finds a solution to deal with the altering parameters for not only the stator resistance in the low speed but also the stator transient inductance in the magnetic field weakening. Rectifying parameters is used to improve the accuracy of estimators. Finally, Simulations have been performed to show the stability of the proposed estimators.