中文摘要：

长流程生产过程不同回路之间存在耦合和时滞，是一个复杂的多输入多输出（MultipleInputMultipleOutput，MIMO）多重时滞系统，其控制器参数整定尤为困难。为此，提出了一种基于谱离散方法的多重时滞系统PID参数整定方法，该方法运用多重时滞对象模型的伴随矩阵构造解耦矩阵，并将解耦的多重时滞系统以状态空间形式表达，通过采用谱离散方法限定闭环特征根的分布区域进行PID参数整定。由于无需模型约简即可进行参数整定，提高了多重时滞系统PID参数的整定效率。运用该方法对Wood—Berry二元精馏塔模型进行PID参数整定并仿真，并与其他方法比较，表明该方法在响应速度和扰动抑制方面具有优越性。预计提出的方法在其他长流程生产过程PID控制器参数整定方面也将有广泛的应用前景。

英文摘要：

Long processing plant has couples and time delays in control loops, and it is often a multiple input and multiple output （ MI- MO） system with multiple time delays. Such system is very difficuh to control. For this case, a new PID tuning method of MIMO sys- tem with multiple time delays based on Spectral Discretization is proposed. At first, the decoupler based on the adjoint matrix of the sys- tem is introduced. And then, the decoupled system is expressed by state space. Finally, the PID parameters are tuned by taking Spec- tral Discretization Methods to limit the distribution of closed-loop characteristic roots. The method can improve the efficiency of PID tun- ing without model reduction. The method is used to tune the PID parameters of Wood-Berry binary distillation column plant and the sim- ulation results show that this method has superiority for response speed and load disturbance rejection performance comparing with other methods. Promisingly, the proposed method will have a widespread application prospect in the PID tuning of other long processes.