Research On The System Of The Multi-motor Synchronous Operation Based On The Single Neuron Adaptive Control

The paper focuses on the multi-variable decoupling control of synchronization system of the AC induction motors—decoupling the synchronically modulating velocity system composed of two AC induction motors and transducers into velocity and tension subsystems by the application of the single neuron adaptive control method.Firstly, it is found that the control method of the single neuron adaptive PSD can hardly realize the control of the system.Many experiment results demonstrate that when integral component of the tension controller works,it takes great abrasion on the magnetic poweder braker.While it does not work, the tension can not follow up the given value. A single neuron adaptive PSD based on intergral saparation control algorithm is presented to realize the control of the system. Through simulation and experiments, the control effect is justified. However, the coefficient of the integral component in the tension controller is constant, so it could not realize the adaptive adjust according to the system demands with incomplete decoupling of tension and speed, as well as great overshoot of the tension exist. For better decoupling control,the coefficient of the integral component must be adjusted adaptively.Thus, combining the single neuron adaptive PSD control algorithm and the variable speed integrating PID, the variable speed integrating single neuron adaptive PSD control algorithm is presented and applicated in the tension-controller, which gets fine control effect.The presented multi-variable single neuron adaptive controller controls the object directly without modeling the control object. Many experiments demonstrate that the designed controller successfully decouples the velocity and tension, and solves the problem that tension overshoots greatly when the system starts. The system is great robust to the disturbance and has a good character in both dynamic and static status, which can validly trace any given track. While transducers operate in different operation mode(constant V/Hz principle control mode or Vector control mode), it does not need change the controller's parameters. The multi-variable single neuron adaptive controller presented in this paper is suitable to many industrial applications, and has good application prospects.