Abstract:

Chemical processes are usually multivariable systems, which non-square system with more outputs than inputs is known as thin system. Till now, the most common method of designing control configuration for thin systems is to form square systems by increasing or decreasing variables such that only fully decentralized PID control could be achieved with one input pairing to one output. Some critical process variables cannot be used as controlled variables or non-critical variables cannot be included in feedback loops, so the control system often does not meet requirements. With introduction of a non-square gain array relative to average frequency, a method for designing cascade control configuration was proposed through analyzing variable pairing in thin systems. This method fully utilized feedbacks of all output variables without addition of new input variables, so system feedback was more complete and a cascade control was developed for both critical and non-critical variables. Two case study showed that this method achieved not only proper variable pairing but also good control performance especially in speed and efficiency of interference reduction.

Key words: process control, process systems, multivariable system, thin system, interaction analysis, variable pairing, cascade control system

摘要：

化工过程一般为多变量系统，其中输出变量个数多于输入变量个数的非方多变量系统称为瘦系统，现有的非方系统控制结构设计一般采用方形化处理方法，只能形成一个输入变量与一个输出变量配对的单回路控制，要么作为工艺控制指标的重要变量不能成为被控变量，要么次要变量未纳入反馈控制，无法满足控制要求。本文介绍了非方系统的平均频域相对增益阵，对瘦系统进行了变量配对分析，提出了一种瘦系统串级控制系统结构设计方法。这种方法在不添加输入变量的同时有效利用了所有的输出变量反馈，使系统反馈信息更完备，构造了重要变量与次要变量结合的串级控制系统。最后通过实例分析说明了瘦系统的串级控制系统结构设计方法不仅能够得到合理的变量配对，而且系统控制性能良好，尤其在进行干扰抑制的过程中体现出了快速性和高效性。

关键词: 过程控制, 过程系统, 多变量系统, 瘦系统, 关联分析, 变量配对, 串级控制系统