摘要
针对软测量技术在线实施时的数据预处理问题,提出了基于聚类分析的过失误差侦破方法。该方法不需过程的先验知识和假设,直接面向数据,可十分方便地在线实现。将该方法与滑动平均滤波算法相结合,可以有效处理过程测量数据的过失误差和随机误差,从而提高软仪表估计的精度。在二元精馏塔底产品组分浓度软测量仪表在线进行的仿真中,应用该方法进行数据预处理,使进入软测量模型的过程数据更接近真实值,取得了很好的效果。
Clustering technique is used to detect gross error of data pre-processing in soft sensor technology,The advantage of this method is no need of priory knowledge and assumption of the process. The gross error detection approach is integrated with moving time window average filter algorithm,so the random error and gross error can be handled simultaneously.The application result in a simulated binary distillation column shows that the proposed approach is a good data pre-processing method for soft sensing technology.
出处
《控制工程》
CSCD
2006年第4期298-300,共3页
Control Engineering of China
关键词
过失误差侦破
软测量
聚类分析
gross error detection
soft sensing
clustering
