Improved MRAS speed observation method

Abstract

The invention relates to an improved MRAS speed observation method. According to the improved MRAS speed observation method, a reference model and an adjustable model are established by means of a voltage model and a current model of a stator flux linkage, a stator flux linkage observer is constructed by adopting a first-order low-pass filter cascaded method, a flux linkage amplitude discrete iterative processing method is proposed, a self-adaptive law is designed based on a Popov hyperstability theorem, a variable PI regulator based on output generalized deviation is designed for speed observation, the speed sensorless vector control of a doubly-fed motor is realized, and a reference and a basis are provided for industrial field application.

Description

technical field [0001] The invention relates to an MRAS speed observation method, in particular to a closed-loop speed observation method based on an improved model reference self-adaptation. Background technique [0002] In the doubly-fed motor speed control system, in order to achieve high dynamic and steady-state performance control of the control system, it is necessary to install a speed sensor to obtain accurate motor speed for closed-loop control. However, the high-precision speed sensor is expensive and difficult to install and maintain. At the same time, the speed signal is easily connected with interference noise during the transmission of the signal line, which reduces the speed and causes excessive impact of the starting current, which will directly damage the inverter and the motor. [0003] In recent years, the speed sensorless control of doubly-fed motors has gradually attracted people's attention. Compared with the sensorless control of cage induction motor ...

AI Technical Summary

Problems solved by technology

The closed-loop observation method based on the torque angle needs to perform an integral operation on the rotor voltage, which introduces an integral drift; the speed observation method based on the power angle avoids the integral operation of the rotor voltage, but its speed observation is affected by the control accuracy of the current inner loop

Based on the observation method of excitation current and stator voltage, the principle is simple and the control is clear, but the mutual coupling between the speed observation loop and the doubly-fed motor control loop makes the system design more complicated and it is difficult to guarantee its efficiency

The extended Kalman filter method and the method based on neural network are respectively proposed. The former is too computationally intensive, and the speed estimation accuracy is also affected by the change of motor parameters; the latter has good fault tolerance and anti-harmonic interference, but its repeated Learning and training takes a lot of time, and there is still a certain distance from industrial applications

Images