| Speed sensors and other devices are used to measure the actual motor speed in order to maintain closed-loop control in the AC induction motor vector control frequency control system. Speed sensorless can overcome the shortcomings of speed sensors such as inconvenience of installation and the entire system affected by the event of failure. Therefore, speed sensorless is one of the hotspots of research by scholars from various countries in recent ten years.Accurate identification of speed is the core speed sensorless of vector control technologies and basis of field oriented vector control system. Model reference adaptive method(MRAS) is the most mature of existed speed sensorless speed estimation methods. Precise motor parameters of reference model and adjustable model are required by the method in order to accurately identify the exact speed of motor. However, the motor parameters such as stator and rotor resistances are changed by temperature changes in case of motor running. This reason leads to inaccurate speed estimation. Therefore, estimation of speed with online parameters identification of motor is to achieve purpose of high precision speed estimation. The main points of this paper are as follows:Firstly, mathematical model of AC induction motor and its vector control algorithms are studied, and asynchronous motor vector control system with speed sensor is built.Secondly, Model reference adaptive speed estimation methods are researched and the advantages and disadvantages of flux speed identification and Back-EMF are analyzed, the impacts with stator and rotor resistances variation on accuracy of speed identification are studied.Thirdly, the influence of the integral parts of traditional flux based on MRAS stator and rotor resistances and speed interactive identification method are not eliminated, and an improved back-EMF MRAS based on stator and rotor resistances and speed interactive identification method is proposed and verified by Simulink simulation. Whether the motor is running at high speed section or at low speed section, highly accurate identification of stator and rotor resistances are obtained; and the accuracy of speed estimation is proved and the electromagnetic torque ripple.is reduced by the simulation data indication.Finally, the above proposed algorithm is applied to the semi-physical simulation platform based on dSPACE hardware in the loop system and the motor experiment platform based on DSP2812, and is verified that it can well identify the motor speed and has a high accuracy of recognition whether motor is running at high speed section or at low speed section.Simulink offline simulation and dSPACE and DSP2812 experiment results show that: the algorithm can recognize the motor speed and stator and rotor resistances simultaneously, speed sensorless closed-loop vector control system can well meet the dynamic and static performance requirements by the estimated speed, and it verified the feasibility of system in practice. |
PDF Full Text Request