基于神经网络控制的三电机同步系统

Three-motor synchronous system based on neural network control

设计了基于神经网络的三电机同步控制系统.首先,给出了三电机变频调速系统的数学模型.其次,基于该模型设计的新系统包括:3个BP神经网络在线自整定参数的智能PID控制器,分别对系统中的速度和2个张力变量进行自适应控制;1个神经元解耦补偿器,通过训练网络权值,补偿各回路之间的耦合影响.最后,基于西门子S7-300PLC构建平台,进行了解耦特性、跟踪性能和抗负载扰动能力的试验.结果表明,与传统PID参数控制系统相比,该系统能够根据不同的运行工况自动获得较优的PID参数,且对系统速度和张力实现了较好的解耦控制,具备了较好的动静态性能及较强的抗干扰能力.

Based on neural network,a three-motor synchronous system was designed. Firstly,the mathematical model of three-motor frequency control system was presented. Secondly,the new system was designed including three intelligent PID controllers and a neuron decoupling compensator. The three con-trollers based on BP neural network arithmetic adjust the parameters on-line and adaptively control the velocity loop and two tension loop respectively. The compensator compensates the coupling effects between the velocity loop and two tension loops by training the weights of networks. Thirdly,the experiments of decoupling property,tracking performance and anti-disturbance performance were completed on the platform constructed on the base of SIMATIC S7-300 power PLC. The results show that compared with traditional PID parameters control,the control system can obtain some optimal parameters of the PID controllers according to different running state of system,realize the better decoupling control of speed and tension,and possess better dynamic and static characteristics and stronger anti-interference ability.