摘要
在工业过程中,有很多重要变量往往无法在线检测,通常通过软测量方法进行估计,主元回归是其中1种常用方法。相比于主元,因子更具广泛意义,更能反映数据的本质特征。基于此,提出1种基于因子回归模型的软测量方法,先对过程日常运行数据进行因子分析,建立因子生成模型,并提取因子信息,然后建立因子与关键变量间的因子回归模型,在线应用时先将可测变量代入生成模型得到因子变量,然后将因子代入到因子回归模型,软测量出关键变量。将该方法应用到化工吸附分离过程中,比较了因子回归模型与主元回归模型的软测量效果,结果表明前者优于后者。
During the industrial process,there are many important variables available online which usually are estimated by soft-sensor.As a valid soft-sensor method,principal component regression is widely used.However,compared with principal component,factor has more extensive sense and can＂mine＂more intrinsic feature in process variables.Due to this,the paper proposes a soft-sensor technique based on factor regression model.Firstly,the method sets up factor generation model and extracts the factors from the routine process data through factor analysis,and then builds the factor regression model among factors and offline key variables.During the application online,the factors can be gained by putting the available variables into factor generation model and the key variables are estimated by putting the factors into the factor regression model.At last,the method is introduced into chemical separation process,the comparison of performance between principal component regression model and factor regression model shows the latter method＇s superiority.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2010年第1期38-40,共3页
Computers and Applied Chemistry
基金
国家高技术研究发展计划(863)资助项目(2007AA04Z198)
教育部博士点新教师基金(200802951038)
关键词
因子分析
软测量
主元回归
建模
factor analysis, soft-sensor, principal component regression, modeling
