摘要
复杂的机理模型包含较多的过程变量,过程变量数据的采集对于复杂机理模型在线计算的准确性十分关键,而现场数据在实时采集过程存在很多干扰因素。在现有系统及通讯平台基础上,针对变量与模型的关联程度,利用DCS或其他实时数据库的上、下限约束和算法滤波方法,对过程数据进行了预处理,然后将相关变量分为主导变量与辅助变量,对变量权重系数进行在线修正等一系列处理,尽可能消除模型的随机误差和系统误差。以实际工业用气分装置中典型的多组分精馏塔(脱丙烷塔)和典型的二元精馏塔(丙烯精馏塔)为例,将现场塔顶、塔底温度和组分含量的计算值与测量值分别进行了对比,结果表明开发的模型具有较高的精确度、可靠性和稳定性。
Mechanism model includes many process variables, the real-time data collection of which is often disturbed by lots of factors, such as noises, sensor troubles and so on. However, these data are very important to the online calculation accuracy of complicated mechanism model. On the basis of existing DCS system and communication platform, process data were pretreated by using the function of upper-low limit constraint on DCS or other real-time databases connected to DCS, and algorithm filter was also used. Anymore, all variables were divided into principal and auxiliary variables with different weight coefficients which could be modified online to eliminate almost all the statistical errors and random errors. At last, a depropanizator in a gas fractionating plant, including the typical multi-component distillation column and a propylene distillation column, the typical two- component fractionation equipment, was selected as a case for industrial application of the model. After comparisons of model calculation values of the top and bottom temperatures and component molar fractions of these two columns with their measured values, it was shown that this developed model was of high accuracy, reliability and stability.
出处
《石油学报(石油加工)》
EI
CAS
CSCD
北大核心
2012年第5期871-875,共5页
Acta Petrolei Sinica(Petroleum Processing Section)
关键词
精馏塔
气体分馏
在线分析
机理模型
distillation column
gas fractionating
online analysis
mechanism model