摘要
研究了稳态数据校正技术在含有化学反应过程的流程中的应用情况。在计算中通过两步投影矩阵分类算法进行数据分类,消除了已有算法中出现奇异矩阵的情况;采用两步矩阵投影与广义似然比法相结合的算法进行数据协调和过失误差的侦破、识别。以某乙烯装置的生产工况为实例进行了稳态数据校正,计算结果表明采用校正后的数据更符合实际的物料平衡关系。
Focus of this paper was to highlight application of data reconciliation in a steady-state process including chemical reactions.Two-step matrix projection classification algorithm was used to avoid the appearance of a singular matrix during data reconciliation which might exist in previous algorithms.Generalized likelihood ratio method (GLR) was used to detect gross errors.An example concerning an ethylene plant was presented.Results showed that reconciled and estimated values obeyed conservation laws much better.
出处
《石油化工》
CAS
CSCD
北大核心
2003年第12期1058-1062,共5页
Petrochemical Technology
基金
国家863计划项目(2001AA413220)。
关键词
数据分类
两步矩阵投影
数据校正
过失误差侦破
裂解
石脑油
data classification
two-step matrix projection
data reconciliation
gross error detection
cracking
naphtha
