基于BP神经网络的多升降台同步控制研究

大型分布式升降台控制系统中，其同步控制是最关键的问题之一。针对传统神经网络PID控制器在多升降台同步控制的结构复杂及同步时间长等问题，提出一种基于相邻偏差耦合控制结构的BP神经网络PID同步控制策略，在确定同步误差定义的基础上，建立异步电机的矢量控制模型，改进了BP神经网络同步控制器。对多升降台同步控制系统的仿真实验表明：所研究的控制策略同步误差小，收敛速度快，实用性强。

Study of Synchronization Control for Distributed Stage Elevators Based on Back Propagation Neural Network

The synchronization is one key issue in the large distributed multiple stage elevators. According to the problems of complex structure and big synchronization deviation occurring in the traditional neural network for multiple stage elevators, a neural network control strategy based on adjacent deviation coupling was proved. Vector control model of motor was built on the basis of synchronization deviation definition and tracking deviation definition, and back propagation neural network control was improved. Simulation result of multiple stage elevators system convinced that the studied control strategy is of small synchronization deviation, high convergence speed and strong practicality.