精馏塔软测量建模中数据校正的计算机实现

Program Design for Data Rectification in Soft-sensing Modeling of Distillation Columns

针对精馏塔软测量模型建立过程中数据校正技术进行了研究。根据精馏塔反应过程的特征,对从现场采集的数据中的几种不同误差,运用相应的方法进行查找剔除,并通过多项式拟合法和预测校正法等方法对剔除的数据进行补偿,使得数据更加准确可靠,从而建立更为精确的软测量模型。同时设计了计算机程序可以实现的数据校正算法,运用VisualBasic语言进行编程对数据校正基本实现了自动化处理,在实际应用中取得了良好的效果。

The technique of data rectification in soft-sensing modeling of distillation columns is researched in the article. Several kinds of errors in data collected from the distillation columns are found and deleted by corresponding methods. And those deleted data are compensated by polynomial-fitting and prediction-amending so that the data and the soft-sensing model become more precise and reliable. At the same time, a kind of data rectification method is designed which can be programmed in the computer. Date is processed automatically by program written with Visual Basic. Example analysis proves that these methods are effective.