工业过程采集的数据的可靠性和准确度直接影响到过程控制、调度及优化等。讨论了数据校正的原理及应用,分析了显著误差检测的意义以及显著误差检测的基本原理。在对两种传统的基于统计量的显著误差检测法讨论的基础上,提出了一种基于F统计量的NT-MT显著误差检测方法。该方法将两种传统方法运用其中,仿真结果表明,基于F统计量的改进NT-MT方法给出了很好的检测效果,对显著误差的灵敏度很高。
The reliability and accuracy of industrial process data directly affect the process control, scheduling and optimization.Thus,the measurement data with high accuracy,reliability and consistency are quite important.This paper discusses the theory and application of data rectification,and analyzes the significance of gross error detection method.By investing two kinds of traditional statistics-based gross error detection methods,this paper proposes an F-statistic-based NT-MT detection method.Compared with two traditional methods,the improved NT-MT algorithm can attain better detection results and has higher sensitiveness to gross error.