Q控制图初始性能的改进

Improvement of initial performance of Q control chart

针对单值Q控制图在过程初始阶段所面临的突出问题,即检出力不强的弱点,提出一种改进单值Q控制图初始性能的质量控制图.它是建立在贝叶斯预测理论的基础上,应用在过程方差是未知的稳态生产过程的条件下.对控制图处于稳态和过程均值发生偏移时的性能及基于贝叶斯模型先验估计值对控制图的影响进行了讨论,结果表明控制图的特性适用于监视稳定生产过程的初始阶段,而这种特性恰好弥补了Q控制图在初始阶段性能不佳的弱点.

Aiming at the remarkable problem confronting individual Q control chart during start-up phase-poor detection ability, a quality control chart improving initial performance of individual Q control chart is presented. This chart is established on the basis of Bayesian forecasting and dynamic model and is applied to a stable process during which process variance is unknown. Performance of this chart when the process was in a statistical stable state and the mean of the process had a drift and its effects when Bayesian dynamic models have different pre-estimated values are discussed. The results indicate that property of this chart can be applied to monitor the start-up phase during stable process well, which compensates the drawbacks of Q control chart.