基于RBF神经网络的PID整定在直线驱动无绳电梯中的应用

为克服无绳电梯中直线驱动系统的常规PID参数只能离线调节的不足,提出了基于RBF神经网络的PID参数在线整定方法。首先从PMLSM动态数学模型出发,阐述了RBF神经网络原理及PI自整定的实现过程,给出了基于RBF神经网络的PMLSM控制系统框图。最后,借助MATLAB/Simulink软件对PMLSM控制系统进行了建模与仿真。仿真结果表明提出的方法鲁棒性强、转矩响应快速及转速超调量小,在直线驱动无绳电梯领域具有良好的应用前景。

Tuning of PID Based on RBF Neural Network in the Application of Cordless Elevator Drived by Linear Motor

To overcome the disadvantage of conventional PID controller which tunes parameters only offline in linear driven system of cordless elevator, an online tuning method for PID parameters based on RBF neural network is proposed. Starting from the dynamic mathematical model of PMLSM, the principle of RBF neutral network and the realization process of self-tuning PID controller are elaborated, and then the control system diagram of PMLSM is provided. Finally, modeling and simulation for control system of PMLSM are performed by MATLAB/Simulink. The simulation results show that the proposed method can work well with good robustness, quick response of torque and small overshoot. Thus, it has a good prospect in cordless elevator drived by linear motors.