Research On PMSM Sensorless Control Based On Flux Observer

Permanent magnet synchronous motor (PMSM) has been widely used in high performance drive applications for its advantages, such as compactness, high efficiency, reliability and suitability to environment. Sensorless control of PMSM has been a hot research topic of motor control technology, which has the advantages of low cost and high reliability. Sensorless control technology has a great significance in some special operation field, such as high speed motor, micro-motor, submarine robot and home appliances.This dissertation presents an extended state sliding mode observer of PMSM, and make deep research on flux observer, parameter error influence, high order sliding mode, position and speed calculation, observer digital implementation. And the corresponding simulations and experiments are carried on.The traditional sliding mode observer is based on the second order state equation, of which state variance is stator current, and is used to observe the back electromotive force. The equivalent control is composed of symbolic function. Much chattering exists in observed results. The low-pass filter will change the amplitude and phase of the observed results. The influence can't be ignored and can't be compensated accurately.The extended state sliding mode observer of SPMSM is based on the voltage equation and flux changing regularity. In the current part, the control input is composed of the symbolic function of the hyperplane. The general reachability is proved by Lyapunov theory. In the flux part, the product of sliding mode control and feedback matrix is adapted as the input control. When sliding mode occurs, the observed flux error will converge to zero. The merits are that the chatting can be filtered by flux equation. So the rotor position and speed can be calculated directly from observed flux without low pass filter. The feedback matrix is selected according to the rule, which is that the observed fluxes are decoupled when the speed error existed. Then the speed can be calculated correctly even if the error exists in the observer.In the IPMSM, the inductance is time-variant in voltage equation of the stationary coordinate. The observer can't be utilized directly. The state equation with time invariant parameter is derived through the transform of voltage equation in rotating coordinate. The observer can be utilized by the definition of equivalent flux. The equivalence and hypothesis are proved to be appropriate by the analysis in the sensorless application field.According to the measure method or state varying, the parameters error in observer exist. The observed observation flux law is analysis when the angle speed error isn't zero, and the relationship between the amplitude and phase and feedback parameter in observer is deduced. In the aspect of stator resistance and inductance error the relationship between observed and real flux are derived including amplitude and phase based on the current method in field oriented control. An improved PLL method is presented in order to calculate the rotor position and speed directly from observed flux, with the merits of fast response and low pass filter. The response of speed estimation between observer and PLL is analyzed. And startup strategy based on virtual position and current closed control and the method switch to speed closed loop are introduced.In order to improve the transient response, a second order NTSM observer is presented by the nonlinear sliding hyperplane which is the extension of the traditional linear hyperplane. The control of NTSM observer contains the integral of symbolic function, in which the chattering signal is much less than the one in traditional SMO. The stability of the NTSM observer is proved by Lyapunov theory. The simulations and experiments prove that chattering is suppressing effectively and the transient response is improved further.The PMSM sensorless digital control system is designed including hardware and software. The observer discretization and simulation implementation are proposed, and the motor control system simulation is build up based on MATLAB. The observer implementation method in fix-point MCU is discussed especially, including data format process, observer digital implementation and power function implementation. The simulation and experiments in each chapter prove that the observer implementation method and control system are effective and feasible.