基于α阶逆的大时滞非线性动态矩阵控制

Dynamic Matrix Control of Nonlinear Large-delay Systems Based on αth-Order Inverseo

针对一类大时滞非线性系统,提出了基于α阶逆的动态矩阵控制新方法。该方法采用BP神经网络辨识逼近原非线性系统的α阶逆系统,并与原系统串联复合组成伪线性系统;采用基于线性系统的动态矩阵预测控制方法设计系统附加控制器。在系统存在建模误差、存在扰动和模型参数发生较大变化等情况下,采用该控制方法依然具有很好的动、静态性能和很强的鲁棒性。给出了详细的设计原理和步骤,并通过大量的仿真分析与已有的大时滞非线性系统内模控制研究结果进行了比较:内模控制依赖于系统模型,当模型出现严重失配的情况下,系统性能变坏,而采用提出的方法则不依赖系统精确的数学模型,计算量小,简化了非线性系统的设计;研究与仿真结果证明了所提控制方法的有效性。

A novel method of dynamic matrix control （DMC） based on αth-order inverse is proposed for a class of nonlinear system with large-delay. The method cascades the αth-order inverse model approximated by BP neural network with the original system to get the composite pseudo-linear system. Then the dynamic matrix control method for the linear system is introduced into the pseudo-liear system. Both of the dynamic and static performances of the system are excellent even there are some modelling errors, disturbance or large change of model parameters, The principles and steps are presented, and compared with the internal model control for nonlinear system with large-delay. The performance of internal model control will be bad when internal models and the actual model mismatch. The proposed method does not rely on the accurate mathematical model and has small computation, therefore it simplifies the design of nonlinear DMC controller. The simulation result shows the validity of the proposed method.