串级控制对过程预测控制的影响分析与对策

化工过程控制中普遍设置以流量控制为副回路的串级控制来实现对温度、液位和成分等被控变量的控制。预测控制的操作变量在很多情况下也是流量,其控制作用的实现要靠底层的流量控制回路。本文针对由于现场串级控制结构不允许改变,流量副回路只能接收温度等主控制器的输出作为其给定值,造成上层预测控制的操作变量无法直接下载到流量控制回路的问题,分别提出了一种将上层优化输出通过一阶惯性滤波作用于主回路控制器和一种将串级控制中流量对主被控变量的传递函数嵌入预测控制模型的实施方案,通过Shell标准重油分馏塔的控制问题进行仿真实验证明了两种方案的可行性,并对其控制性能进行了比较分析。两种方法理论上构思简单,实际中易于实现,具有普遍适用性。

Implementation of model predictive control on cascade control

In chemical process control,the flow rate is usually set as the minor loop of cascade control for the other more important controlled variables such as temperature,liquid level and material component. At the same time,the flow rate is often the manipulated variable of model predictive control(MPC). In view of the implementation dilemma of downloading the manipulated variable of MPC while the flow rate control is set as minor loop of cascade control,two implementation methods of MPC were proposed. Firstly,the output of MPC is downloaded to the set-point of major loop through an one order inertia object. Secondly,the transfer function of flow rate to controlled variable of major loop is embedded into the predictive model of MPC. The control of the Shell heavy oil fractionator's benchmark problem has proved the feasibility of the two methods. They are simple in theory and easy to implement for commercial processes.