A Speed Estimation Method For Induction Motors Based On Extended Kalman Filter And Markov Chain

In the high performance AC drive system, the closed-loop speed control is indispensable.Generally, the speed is measured by speed sensors. However, a lot of negative influences to the system are brought owing to installation of speed sensors. For example, the cost is increased, and it is difficult to adapt to the harsh environment, such as high temperature and high humidity. The simplicity and reliability of the AC drive system are reduced, and its application is limited. Therefore, the speed sensorless vector control technology has been paid extensive attention by scholars in recent years, which has has become a research hotspot in the induction motor control. In this paper, the robustness and anti gross error performance of speed estimation method based on multiple-model extended Kalman filter with markov chain(MC-MM-EKF) for induction motors has been researched in-depth.Firstly, the mathematical model of induction motor and the stability of the motor itself are analyzed in this paper. Secondly, the principle of extended Kalman filter is elaborated, and its application in speed sensorless vector control is described in detail. The influence of disturbance and motor parameter variations for speed sensorless vector control is discussed,especially for the motor speed identification. Thirdly, the principle of multi model theory is expounded, and its solution for model uncertainty is analyzed. The multiple-model robust mathematical model based on extended Kalman filter for speed estimation is established, and the robust mechanism of it is researched. During the wide speed range, especially at low speed,the stability and the parameter sensitivity of robust system basen on multiple-model extended Kalman filter with markov chain are analyzed. Fourthly, the robust performance of speed sensorless vector control system for induction motor based on multiple-model extended Kalman filter with markov chain is verified by Matlab/Simulink. Finally, the experiment platform based on TMS320F28335 is established, and the experimental verification of speed estimation method is implemented based on multiple-model extended Kalman filter with markov chain.Both the simulation and experimental results indicate that the proposed method can effectively improve the model adaptability to the actual systems and the environmental variations, compared with extended Kalman filter. The speed estimation error with disturbance and motor parameter variations is obviously reduced, and the steady and transient performance of speed sensorless vector control for induction motor are improved by using the proposed adaptive speed estimation method.