论文部分内容阅读
为了解决控制系统中一个回路参数变化导致其他回路的运行参数改变,提出了一种基于DRNN的在线自整定解耦控制算法。以某被控对象温湿度控制为例构建了数学模型,分析了系统变量之间的耦合关系,设计了解耦网络。将存在耦合关系的多变量控制系统变换为独立的单变量控制系统,以消除相关控制通道之间的影响。基于所提出的对角递归神经网络解耦算法进行了系统仿真实验。系统仿真响应显示:经过解耦后的温湿控制2个通道相互之间影响很小,实现了耦合变量的解耦。仿真研究结果表明:提出的解耦控制算法是可行与合理的。
In order to solve the change of the parameters of one loop in the control system, the online self-tuning decoupling control algorithm based on DRNN is proposed. Taking the temperature and humidity control of a controlled object as an example, a mathematical model is constructed, the coupling relationship between system variables is analyzed, and the decoupling network is designed. Multivariable control systems with coupling relationship are transformed into independent univariate control systems to eliminate the influence of related control channels. Based on the proposed diagonal recurrent neural network decoupling algorithm for system simulation experiments. The system simulation response shows that after the decoupled temperature and humidity control, the two channels have little influence on each other, and the decoupling of the coupling variables is achieved. Simulation results show that the proposed decoupling control algorithm is feasible and reasonable.