一种动态软测量建模方法及其工业应用  

Dynamic Soft-sensing Modeling Method and Its Application in Industrial Process

作　　者：傅永峰[1] 

机构地区：[1]浙江工业大学之江学院,浙江杭州310024

出　　处：《自动化仪表》2011年第9期67-70,共4页Process Automation Instrumentation

摘　　要：针对目前静态软测量建模方法无法反映工业过程动态信息,造成模型预测精度低、鲁棒性差等问题,提出了一种基于模糊曲线和高斯过程的动态软测量建模方法。该方法采用模糊曲线法对输入数据进行处理,并利用处理后的数据构建新的数据集,最后采用高斯过程建立软测量模型。将提出的动态软测量模型应用于PTA氧化过程中4-CBA含量的预测,结果表明,所建模型运算速度快、预测精度高。At present,the static soft-sensing modeling method is unable to reflect dynamic information of industrial process,this makes low prediction accuracy and poor robustness of the model,thus the dynamic soft-sensing modeling method based on fuzzy curve and Gaussian process is proposed.With this method,input data are managed by using fuzzy curve;then new data set is setup by adopting these processed data;finally soft-sensing model is established by using Gaussian process.This dynamic soft-sensing model proposed has been used in prediction of 4-CBA content in PTA oxidation process;the result indicates that the model features fast operation speed and high accuracy.

关 键 词：软测量 静态建模 动态建模 模糊曲线 高斯过程 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]