结合测量误差的检测多元协方差矩阵的EWMA控制图

An EWMA Control Chart Combined with Measurement Error for Monitoring Covariance Matrix

目前,带有测量误差的一元控制图显著降低了测量误差对于控制图性能的影响,并在监控实际生产过程时发挥着重要作用.然而由于多元控制图的复杂性,尤其是在协方差矩阵发生漂移情况下,对带有测量误差的多元控制图的研究还有所不足.因此文章结合多元测量误差模型,构建了对监控过程协方差矩阵漂移具有显著优越性的ME-KQE、KME-KQE控制图,并分析了测量误差对于检测效率的影响.最后,文章通过实例展示了其在实际应用中对存在测量误差的数据有较好的监测效果.

At present,the univariate control chart with measurement error has significantly reduced the impact of measurement error on the performance of the control chart,and plays an important role in the actual production process monitoring.However,due to the complexity of multivariate control charts,especially when the covariance matrix shifts,there is still insufficient research on multivariate control charts with measurement errors.Therefore,this paper combines the multivariate measurement error model to construct ME-KQE and KME-KQE control charts that have significant advantages over the covariance matrix drift in the monitoring process,and analyzes the effect of measurement error on detection efficiency.Finally,this article shows that it has a good monitoring effect on the data with measurement errors in practical applications.