中文摘要：

讨论了基于连续时间模型的广义预测控制软件的开发和应用,避免了传统离散时间方法的采样时间选择问题。本文中的软件采用OPC数据采集技术,以基于随机数搜索的连续时间模型辨识方法为基础,实现了连续时间控制律的模块化仿真与在线控制。通过该软件平台可以计算具有多种组合形式的连续时间控制律,并能实时在线的获取和观察各个模块的输入、输出以及状态,设计整定控制器参数。基于VC平台开发的控制软件,既有与Matlab类似的仿真功能,又可以用于工业现场实时控制。通过对济南炼油厂一个实际模型的控制效果,证明了该软件的有效性,推广了连续时间控制律的应用。

英文摘要：

In this paper, we discuss the development and application of continuous-time generalized predictive control （CGPC） software, in order to avoid the arbitrary choice of sampling time necessary in traditional discrete-time theory. This software is based on a New Luus and Jaackola （NLJ） continuous-time module identifieation method, incorporates Object Linking and Embedding for Process Control （OPC） data sampling technology, and allows module continuous-time simulation and online control. Using this software, various kirids of continuous-time control laws can be obtained, the real-time input, output and status can be displayed and observed, and the controller parameters can be set in real-time mode. Developed with a virtual channel （VC）, this software is characterized not only by its simulation function, but also by its real-time control ability in different fields. Through its successful application on a module in a Jinan refinery, the software is shown to be effective and extends the application of continuous-time control laws.