Research On Some Issues Of Speed Sensorless Vector Controlled System

The fast development of power electronics,micro-electronics and modern control theory brought huge impulsively effects to the development of the advanced control strategies for asynchronous motor. For the practice engineering applications, this paper takes the sensorless vector controlled system of asynchronous motor as the main research object. Around the pivotal technologies such as rotor flux estimation,speed sensorless and stability et al, This paper carried deeply researches from the theory analysis,computer simulation,soft and hardware design of control system,prototype examination et al.The intrinsic physical meaning of three different coordinate transformation methods based on equivalent motor and non-equivalent motor principle respectively is studied, and the relationship among them is analyzed. The mathematics model in different coordinates and the equation of state space are established. The states observable and controllable are verified. Those established theory foundation for the magnetic field direction vector control and speed sensorless adaptive design.The decoupling characteristics of flux and electromagnetism torque in different magnetic field directional modes are expatiated. In rotor magnetic field directional mode, the rotor flux and electromagnetism torque are decoupled thoroughly. The flux observer is the core technology in speed sensorless system. In voltage mode flux observer, the pure integral can cause the saturation and excursion of output flux. Based on analyzing the cause of the rotor flux observation error, several typical improved integral methods are analyzed and compared in theory,simulation and experiment. It's pointed that these methods are only improved the limitation of integral initial value, but the input dc offset isn't settled availability. A new scheme was presented which can eliminate the dc offset of sampling stator current and improve the pure integral loop. The experiment results show that this scheme can obtain good sinusoidal of observed rotor flux wave and eliminate the influence of input dc offset to integral loop.Based on the Popov hyper-stability theorem, the rotor speed is estimated in the model reference adaptive system (MRAS), the dynamic stability of estimated subsystem is verified by the small signal stability analysis. The asynchronous motor is separated into electromagnetism subsystem and mechanism subsystem based on passive analysis. The stability of electromagnetism subsystem is analyzed by the Lyapunov stability theorem. The analysis results show that the consistent stability of the asynchronous motor with constant motor parameters and the sufficiency condition of the asynchronous motor with variable motor parameters is derived. Those afford stability theory base for the application of high performance control strategy to asynchronous motor. After analysis the stability of asynchronous motor and speed sensorless system, this paper proposed that the stability range of close loop system is judged and the influence to system stability of timing references of asynchronous motor and controller references is analysis by establishing the dynamic small signal model of the close loop system, obtaining the root locus and step response by MATLAB simulation. According to the close loop root locus simulation, the influence of timing variable motor parameters and controller parameters to the stability of close loop system is analyzed. The experiment results show that this analysis method can instruct the tuning of controller parameters during experiment.A prototype of the asynchronous motor drive system that based on the speed sensorless vector control is designed. To higher computer precision, the software of control system is designed based on the normalized mathematical model of the asynchronous motor. To sure the veracity of computer, a digital scaling mode that includes the sign bit is adopted to realize floating-point operation to the fix point DSP. An instantaneous average voltage dead-time compensate method is brought forward based on analysis the causing of dead-time effect in SVPWM which realized by computer the average error voltage in one switch period according to the sector in which stator current vector stay. To overcome the low sampling precision of stator voltage, a stator voltage structure method is adopted which used the dc link voltage and the switch state of inverter. The delay caused by filter is avoided for omitting the voltage filter loop. The experiment results show that the speed sensorless vector controlled system that was designed in this paper has good static and dynamic performances.