Abstract: The ethylene column is a kind of separation devices used in the petro-chemical process. It is easy for the column to enter abnormal condition because of improper operation in the distillation process. The outlet temperature will decrease with reflux increasing in normal condition and the model gain is negative. However, if reflux continues to increase, plate efficiency will decrease then outlet temperature begins to rise and model gain becomes positive then the distillation comes into abnormal condition. A kind of model gain online identification method based on the least square method for ethylene column was proposed in this paper. The symbol of model gain was identified online and then if the ethylene distillation column enters abnormal condition could be determined according to identification of model gain. Normal and abnormal control systems of the ethylene column were established and the control effect of the column controller could be determined according to the identification result of model gain, making reflux and outlet temperature controlled in the permitted range of normal condition. It could guarantee the ethylene column to work under normal condition and achieve the control of ethylene column under normal and abnormal condition.

Key words: process control, dynamic simulation, fault diagnosis, abnormal condition, online detection and control, ethylene column, plate efficiency

摘要： 精馏塔是化工过程经常用到的一种分离设备,操作条件不当,易使精馏塔进入异常工况.正常工况时,出料温度随回流量增大而降低,模型增益为负;回流量继续增大,导致板效率下降,出料温度开始上升,模型增益变为正,精馏塔进入异常工况.针对此情况提出一种基于最小二乘法的乙烯精馏塔模型增益在线辨识方法,在线实时辨识出模型增益符号,根据辨识结果,判断乙烯精馏塔是否进入异常工况.建立正常与异常工况的乙烯精馏塔控制系统,根据模型增益辨识结果,由负反馈原理确定正常与异常工况中控制器正反作用,使得回流量和出料温度均控制在正常工况允许范围内,保证乙烯精馏塔在正常工况工作,实现乙烯精馏塔正常与异常工况控制.

关键词: 过程控制, 动态仿真, 故障诊断, 异常工况, 在线侦测与控制, 乙烯精馏塔, 板效率