数据与未建模动态驱动的非线性PID切换控制

Data-based un-modeled dynamics driven nonlinear PID

中文关键词  非线性系统，PID，未建模动态，数据，水箱

英文关键词  Nonlinear system  PID  Unmodeled dynamics  Data  Tank

中文摘要

      本文充分利用系统的数据信息和知识，把数据驱动控制、PID控制与一步超前最优控制策略相结合，提出了数据与未建模动态驱动的非线性PID切换控制方法。该方法首先利用被控对象往往运行在工作点附近的特点把被控对象表示成低阶控制器设计模型与高阶非线性项(未建模动态)和的形式。与以往方法的本质区别在于，所提的方法直接将未建模动态分解为前一拍数据与未知增量的和，并利用未建模动态可测数据补偿系统未知的非线性动态特性，设计非线性PID控制器，对未建模动态的未知增量采用自适应神经模糊推理系统(adaptive-network-based fuzzy inference system, ANFIS)进行估计，从而设计带有未建模动态增量估计的非线性PID控制器。两个控制器通过切换机制协调控制系统，既保证系统的稳定，同时提高系统的性能。给出了PID控制器参数选择的一般原则和方法，推导了保证闭环系统输入输出稳定性的条件；最后,通过数值仿真实验以及在水箱液位控制系统的物理对比实验，实验结果验证了所提算法的有效性和实用性.

英文摘要

      To solve the control problem for a class of complex industrial process which are difficult to establish their accurate mathematical model, data based unmodeled dynamics driven nonlinear PID control method which combined the advantages of data driven control, classic PID control and one-step ahead optimal control strategy is proposed in this paper. First, the model of the controlled plant is expressed as the linear model plus the higher order nonlinear term (unmodeled dynamics) , with the above development, the posterior unmodeled dynamics measurement data are made use to compensate the unknown nonlinearity of the system, and the unknown increment of the unmodeled dynamics is estimated by a adaptive-network-based fuzzy inference system (ANFIS), then, a nonlinear PID controller with compensation of the posterior unmodeled dynamics measurement data and the estimation of the increment of the unmodeled dynamics is designed. The two controllers are used to the control system coordinately through the switching mechanism, the stability and the performance of the system are all thus improved. Moreover, the general principle and basic design method for PID controller parameters are griven. The condition of analysis on stability and convergence of the proposed method are established. Finally, through the experiment based comparative experiment on a tank level control system, the effectiveness of the proposed method is justified.