Permanent magnet synchronous motor (PMSM ) state estimation method based on sliding mode reference self adaptation

Abstract

The invention discloses a permanent magnet synchronous motor (PMSM) state estimation method based on sliding mode reference self adaptation. According to stator voltage and stator current detected inreal time, a sign function is adopted, a model based on a sliding mode current observer is constructed to estimate the state of the PMSM, an error function based on the sliding mode current observer is constructed, a rotor position is used as a to-be-estimated parameter through dynamic tracking on output of the sliding mode current observer, the flux difference between a reference model and an adjustable model is used as the error function, small-range optimization around the rotor position in the last sampling period is carried out, and the optimal rotor position which enables the error function to be the minimum is found out quickly. Through continuously adjusting the parameters of the adjustable model, the purposes of enabling the adjustable model to follow output of the reference modeland effectively reducing tracking errors are finally realized, the motor can realize quick and accurate rotation speed estimation in a wide rotation speed range, and the accuracy of PMSM speed estimation is improved.

Description

technical field [0001] The invention belongs to the field of position sensorless state estimation of permanent magnet synchronous motors, in particular to a method for state estimation of permanent magnet synchronous motors based on sliding mode reference self-adaptation. Background technique [0002] Permanent magnet synchronous motor (PMSM) has many advantages such as small size, small moment of inertia, high efficiency, large torque-to-inertia ratio, strong overload capacity, and wide speed range. This makes the PMSM speed control system has become a research hotspot in the field of AC speed control transmission. The main idea of ​​the Model Reference Adaptive (MRAS) algorithm is to build two models simultaneously, one is an adjustable model, which includes equations with estimated parameters, and the other is a reference model, which mainly consists of parameter equations without estimates . These two models exist and work at the same time, and the output error of the ...

AI Technical Summary

Problems solved by technology

[0003] The commonly used MRAS method generally uses the voltage model to estimate the flux linkage, but the DC component of the pure integral link of the voltage model has always troubled the control performance of the MRAS method at low speeds. For this reason, many scholars have proposed various improvement methods, including Using high-pass filter to filter DC, using low-pass filter with variable cut-off frequency to replace pure integral link, and using coordinate transformation to filter under synchronous rotating coordinates, although the problem of amplitude compensation is better solved, it is still not possible. Completely eliminate DC component errors

Moreover, although the traditional model reference adaptive method has good accuracy, it is highly dependent on the model parameters. When the external disturbance is large, it will greatly affect the robustness and accuracy of the system.

Images