Two-motor Synchronous Speed-regulating System Based On Neural Network Control

The paper focuses on the multi-variable decoupling control of synchronous system of the AC induction motors. With the application of neural network control method, the decoupling control between the speed and tension of the two-motor synchronous system composed of two AC induction motors and transducers is studied.On the basis of mathematic model analysis of two-motor synchronous system, self-adaptive PID controllers based on diagonal recurrent neural network are designed in the speed and tension loop respectively combined with nonlinear mapping, adaptive and learning capability of artificial neural networks, which realize PID parameters adaptive adjustment in each loop and make the system possess stronger adaptive capability and other better performances. In order to minimize coupling effect between the two control loops, a neuron decoupling compensator is designed according to intelligent decoupling technology of multi-variable system. It integrates the effects of the two control loops and realizes the decoupling control between speed and tension by training the weights of networks to compensate the coupling relation.S7-300 PLC is used as lower computer controller, and neural network control algorithm is achieved with STEP7 structural programming method. Siemens upper computer software WinCC is applied in establishing monitoring picture of multi-motor system, realizing visualization of control process, making operation simplification and process information palpability. In addition, control network of two motor synchronous control system is established, including PROFIBUS-DP fieldbus communication between PLC and transducers, and MPI data communication between WinCC and PLC. Finally the long-distance control of the two-motor synchronous system is realized.Lots of experiments are performed on the platform of two-motor synchronous system. Experimental results show that the method designed realizes the decoupling control between the speed and tension of the two-motor synchronous system. The system is great robust to the disturbance, has better dynamic and static characteristics and can trance any defined route. Therefore, the method presented in the paper meets the requirements of many industrial control environments, and has good application prospects.