摘要
提出一种分两步进行物料平衡的实时数据校正方法。充分利用测量数据的时间冗余性,剔除不可靠的数据。建立多元回归模型,根据历史数据对含有显著误差的测量数据预测,使其预测值参与协调。仿真研究证明了此法的有效性,并大大提高了协调精度,节省了协调时时间。
A two-step on-line data rectification method for material balance is presented. The method made the best use of temporal redundancy information, and deleted the undependable measurements. MLR is used to predict the value according to the historical data in the data reconciliation. And simulation results demonstrate the efficiency of the proposed method. The reconciliation time can be saved and the reconciliation precision can be ensured.
出处
《鞍山科技大学学报》
2007年第4期399-402,共4页
Journal of Anshan University of Science and Technology
关键词
实时
数据校正
多元回归
real-time
data reconciliation
MLR
