多变量隐式广义预测自校正控制算法研究（毕业学术论文设计）.docVIP

PAGE PAGE II PAGE I 多变量系统的隐式广义预测自校正控制算法研究 摘 要 预测控制作为一种有效的控制算法，已被大量的仿真和实际应用所证实。国内外学者们对单变量广义预测控制算法(GPC)作了大量的研究，但是实际工业过程通常为复杂的多输入多输出(MIMO)系统。广义预测控制作为一种优化控制算法，其优点之一就是可以直接处理多变量系统的控制问题。本文就是针对复杂系统的广义预测控制算法研究而进行的。 首先，对多变量广义预测控制算法进行研究，在原有理论的基础上引出一种新的广义预测控制算法，它特有的隐式解耦功能可以有效地克服传统分散控制、解耦控制的烦琐和缺陷。在控制算法中利用GPC并列预报器之间的特点，直接辨识预报其中最远输出预报式的参数，并利用GPC与DMC控制律的等价性，来推求最优控制率的参数，提出一样简单的隐式自校正算法。它保留了GPC鲁棒性强等特点，可适用于任何稳定的最小相位/非最小相位和已知/未知时延系统。 然后，在对单变量隐式广义预测控制算法仿真的基础上，通过将目标函数分散化和把输入/输出的交换影响前馈解耦，使多变量系统问题转化为单变量系统设计问题。利用MATLAB建立多变量隐式广义预测控制系统，对系统进行仿真得出仿真图。 最后，以仿真图为基准，进行分析。仿真结果表明该算法在跟踪性、鲁棒性和抗干扰能力方面都有良好的性能，并且可控制不稳定非最小相位系统。 关键词：预测控制；多变量；隐式；自校正；MATLAB Multivariable Systems Implicit Generalized Predictive Self-tuning Control Algorithm Abstract Predictive control as an effective control algorithm，has been a lot of simulation and practical application of the confirmed domestic and foreign scholars on the single-variable generalized predictive control algorithm (GPC) made a lot of research, but the actual industrial process is usually complex multi-input multi-output (MIMO) systems. Generalized predictive control algorithm as an optimal control, one advantage is that you can directly handle multi-variable control system. This article is for the complex system of generalized predictive control algorithm carried out. First, Multi-variable generalized predictive control algorithm research, based on the original theory leads to a new generalized predictive control algorithm, its unique feature of implicit decoupling can effectively overcome the traditional decentralized control, decoupling control of the cumbersome and shortcomings. GPC control algorithm in parallel using the characteristics between the predictor, which forecasts the direct identification of the parameters of the farthest output prediction type, and use of GPC and DMC control law of equivalence, to estimating the parameters of the optimal control rate presented as simple implicit self-tuning algorithm. It retains the characteristics of GPC