Neural network generalized inverse permanent magnetism synchronous machine decoupling controller structure method without bearing

Abstract

The invention discloses a constructing method of a nerval net generalized inversing bearingless permanent magnet synchronous motor decoupling controlling device, which takes two Clark inverse transformations, two electric current tracing type inverters, a bearingless permanent magnet and a load model as a whole to form a composite controlled object, a nerval net generalized inverse of the composite controlled object is formed through adopting static nerval net added with a plurality of linear links, then the nerval net generalized inverse is reversely arranged before the composite controlled object to form a generalized pseudolinear system, the generalized pseudolinear system is decoupled to three single input and output subsystems through linearization, finally a nerval net generalized inverse, the two Clark inverse transformations and the two electric current tracing type inverters are all formed to a nerval net generalized inversing bearingless permanent magnet synchronous motor controlling device, the controlling device can not only realize dynamic decoupling between a radial position system of a motor and torque moment system and between radial forces, but also be taken as a nonlinear open-cycle controlling device to use directly, and stable suspension and operation of a rotor of a motor can be ensured.

Description

technical field [0001] The invention belongs to the technical field of electric drive control equipment, and is a generalized inverse bearingless permanent magnet synchronous motor control system based on neural network, that is, a control method for a bearingless permanent magnet synchronous motor, suitable for bearingless permanent magnet synchronous motors high performance control. Background technique [0002] The bearingless permanent magnet synchronous motor is a multivariable, strongly coupled nonlinear system. To realize the stable suspension and operation of the motor rotor, it is necessary to dynamically control the electromagnetic torque and the radial suspension force, as well as the horizontal and vertical suspension forces. Decoupled control. [0003] The research on the decoupling control of bearingless permanent magnet synchronous motor has always been the focus and difficulty. At present, the method that has been adopted is the rotor field oriented vector c...

AI Technical Summary

Problems solved by technology

However, when the neural network inverse control method is used, although the original nonlinear system can be linearized and decoupled into a single-input and single-output integral subsystem (pseudo-linear composite system) with a linear transfer relationship, these integral subsystems are open The loop is unstable, so the neural network inverse system cannot be simply used as the only controller to "open-loop control" the controlled system, but an additional controller must be designed to form a closed-loop controller together with the neural network inverse system to control the system , so the control is more complex

[0004] In order to essentially solve the complex problem of the closed-loop control system of the bearingless permanent magnet synchronous motor, and at the same time ensure the various control performance indicators of the bearingless permanent magnet synchronous motor, such as dynamic response speed and steady-state tracking accuracy, it is necessary to adopt a new Control technology and new control methods

[0005] The patent applications published in the same technical field in China include: 1. The name is "Magnetic Levitation Switched Reluctance Motor Radial Neural Network Inverse Decoupling Controller and Construction Method", publication number: CN200510038099.5, this patent application is aimed at Magnetic Levitation Switched Reluctance Motor The radial neural network inverse decoupling controller designed for the coupling between radial suspension forces only uses the neural network to inversely decouple the radial force, and does not consider the nonlinear strong coupling between radial force and torque. And its research object is magnetic levitation switched reluctance motor

2. The name is "Neural Network-Based Inverse Five-DOF Bearingless Permanent Magnet Synchronous Motor Control System and Control Method", publication number: CN200510040065.X, the application is a control method designed for five-DOF bearingless permanent magnet synchronous motors, This method needs to design a closed-loop controller after multi-variable decoupling using neural network inversion. The closed-loop control system is complex, and the controller parameters are difficult to adapt to changes in motor parameters, and its performance is unstable.

3. The name is "Control Method of Neural Network Inverse Decoupling Controller for Bearingless AC Asynchronous Motor", publication number: CN200610038711.3, the patent application is designed for the control method of bearingless AC asynchronous motor, the motor is a bearingless asynchronous For the motor, after decoupling the four variables using neural network inversion, it is necessary to design a closed-loop controller. The closed-loop control system is complex, and the controller parameters are difficult to adapt to the changes of the motor parameters, and its performance is unstable.

Images