Auto-correlated charts based on dynamic Bayesian model
-
摘要: 传统控制图应用的假设前提是从过程得到的观测值彼此独立.但许多过程存在自相关现象,并且这种自相关现象对控制图性能有严重影响.从工程实际出发,对AR(1) 过程模型提出了基于贝叶斯动态模型的自相关控制图,阐明了此自相关控制图的原理和方法;并将该控制图与基于过程模型的自相关控制图的性能进行比较.此控制图借助贝叶斯先验估计,建模迅速,不以时间序列模型为基础,对强自相关模型具有很好的性能.对强自相关过程,并要求迅速建立过程控制时,可以使用此自相关控制图,进行过程监视.Abstract: In statistical process control it is usually assumed that the observations taken from the process are independent. However, for many processes the observations are autocorrelated, and this autocorrelation can have a significant effect on the performance of traditional control chart. The problem of monitoring the mean of process in which the observations can be modeled as an AR(1) process was considered. A one-step forecasting autocorrelated control charts based on dynamic Bayesian model was presented for engineering practices. The principle and methodology to build autocorrelated control charts was described in detail. This control chart′s performance was compared to the performance of control chart based on process model, and the effect of process parameter estimation on the control chart was also investigated. When the process model is strong autocorrelation AR(1), the autocorrelated control chart has good performances with few samples. In this case, autocorrelated control chart was suggested instead of traditional control charts.
-
Key words:
- statistical process control /
- control charts /
- time series /
- Bayes methods
-
[1] Montgomery D C. Introduction to statistical quality control[M]. 3rd ed. New York:John Wiley & Sons,1996 [2] Awin R W, Lee S J. The effects of autocorrelation and outliers on two-sided tolerance limits[J]. Technometrics,1998,40:24-38 [3] Atienza O O, Tang L C, Ang B W. A SPC procedure for detecting level shifts of autorrelated process[J]. Journal of Quality Technology,1998,30:340-351 [4] Dyer J N, Adams B M, Conerly M D. The reverse moving average control chart for monitoring auto-correlated process[J]. Journal of Quality Technology,2003,35:139-152 [5] Wang F K. A simple data transformation of auto-correlated data for SPC[J]. International Journal of Production Research,2005, 43(3):981-989 [6] Vasilopoulos A V, Stamboulis A P. Modification of control chart limits in the resence of data correlation[J]. Journal of Quality Technology,1978,10(1):20-30 [7] Wardell D G, Moskowitz H, Plantz R D. Run length distribution of special-cause control charts for correlated observations[J]. Technometrics,1994,36:3-17 [8] Jiang Wei, Tsui K L. A new SPC monitoring method:the ARMA chart [J]. Technometrics,2000,42(4):399-410 [9] 张孝令,刘福升,张承进,等. 贝叶斯动态模型及其预测[M]. 济南:山东科学技术出版社,1992 Zhang Xiaoling, Liu Fusheng, Zhang Chengjin, et al. Bayesian forecasting and dynamic models[M].Jinan:Shandong Science and Technology Press, 1992(in Chinese) [10] Cox D R. Prediction by exponentially weighted moving average and related methods[J]. Journal of the Royal Statistical Society,1961,23:414-412 [11] 孙静.自相关过程的自相关控制图[J]. 清华大学学报,2002,42(6):735-738 Sunjing. Residual charts for autocorrelated processes[J]. Journal of Tsinghua University, 2002,42(6):735-738(in Chinese) -
点击查看大图
计量
- 文章访问数: 2550
- HTML全文浏览量: 41
- PDF下载量: 1061
- 被引次数: 0
下载:
