基于非线性类PID算法的控制器优化过程仿真

Design of Nonlinear PID-Like Controller and Its Simulation

研究控制器的优化控制问题。在较为复杂的工业控制环境下,被控对象具有不确定性、非线性及时变性。传统PID算法不能满足控制复杂过程的要求,会造成传统的非线性类PID神经元网络控制方法计算量相对较大,且整定困难,造成控制不稳定。为了解决上述问题,利用本地混合递归神经网络设计了一种低复杂度类PID神经元网络控制器。构造一个适用于强耦合大时滞MIMO系统的多变量控制器。实验中,运用Lyapunov稳定性理论分析电力线载波通信流量控制、塑料注射成型机通道温度控制过程,仿真结果表明,提出的类PID控制器模型闭环稳定且非线性控制性能优异,对强耦合时滞系统具有良好的跟踪与解耦能力。

A mix locally recurrent neural network was used to design a low complexity PID-Like Neural Network Controller aiming at reduce calculation and tuning difficult. And then, a multiple-input muhiple-output multivariable controller was constructed for strong coupling and large delay system. The result of analysis using Lyapunov stability theory and flow control of PLC and plastic injecting-moulding machines ＇ pipe temperature control show that the con- troller is stability and has a outstanding non-linear control ability, the multivariable controller constructed has a good tracking and decoupling capability for strong coupling systems with time delay.