论文部分内容阅读
针对复杂工业过程对象可能存在结构和参数变化的特点，提出了一种基于神经网络控制器，利用改进的算法在线修正权值。该控制器不依赖于对象的数学模型，可适用于变阶、变结构参数、多变量等场合。仿真结果显示，该控制器具有良好的控制特性和鲁棒性，并且可对多变量系统进行有效的解耦和控制，且控制器结构简单，易于实现，具有明显的工程应用价值。
Aiming at the characteristics that complex industrial process objects may have structural and parameter changes, a neural network controller based on neural network controller is proposed to modify weights online with improved algorithm. The controller does not depend on the mathematical model of the object and can be applied to variable order, variable structure parameters and multivariable applications. The simulation results show that the controller has good control characteristics and robustness, and can effectively decouple and control the multivariable system. The controller is simple in structure and easy to implement, and has obvious engineering application value.