《SPC连续-第二部分》.pptVIP

检验 7 可能由于需重新计算控制限或由于极差图变化范围很大所导致.极差图变化范围很大也就是说组内偏差大于组间偏差. 这可以通过检查极差图加以识别. 高度变化的极差图和变化较小的X-杠图常表明选择的编组不合理. 应重新进行编组以回答这类问题. 对特殊原因引起的偏差采取行动最容易进行. 按实际情况让最基层的人员去采取行动有好处. 这样减少了反应时间. 因为行动由技术人员采取, 故应记录在故障排除指南中从而能让操作员下次采取同样行动. 减少常见原因引起的偏差通常是一个大项目, 甚至是一个大的黑带项目. 我们期望黑带在改进阶段已经减少了偏差, 在控制阶段作控制图只是为了进一步改进或监控. 按前面描述过的由d2常数计算总体的标准差. OCAP 是指 Out of Control Action Plan(失控行动计划). 有许多方面不同控制图的标准程序. 所有程序都在帮助系统和 Minitab 手册中有记录. 下一节我们会介绍离散变量的控制图. 在前一节我们已介绍了选择适当控制图的流程图. 尽管所有八个检验都存在, 正如前面所论述的, 除非要寻找细微的特殊原因, 应先用检验 1 和 5作为开端. 在消去那些偏差后在进行其他检验. Minitab 虽提供了一个好的图以供分析和计算, 却不允许让操作员或实时计算机系统输入数据以随时更新控制图. 事后分析大大失去了控制图的效用, 即当失控信号一发现就可以对流程采取措施的效用. Minitab directly calculates values for each subgroup standard deviation, rather than estimate using range values, since ease of calculation is not an issue. These within-subgroup values for s must then be pooled to obtain the “average.” Recall that standard deviations are not additive, so that they cannot be averaged in the ordinary way. Pooling them squares each one to get within-subgroup variances, averages the variances, and takes the square root of the result. This pooled standard deviation is then used in place of Rbar in the chart calculations, as shown on the next slide. Comparing the values for control limits and center lines for the two methods, very small differences can be seen. These differences do not have a significant impact on the results from the charts. Since Minitab is using actual standard deviation values, rather than estimates from Rbar, its control limits and center line values are actually slightly more rigorous. In the previous example it was discovered that during the time period represented by subgroups 9-11 24-25 a vacation substitute was running the process at the wrong settings. So, since they are explained, they can be removed from the control limit calculations. Do not exclude points when the cause of the OOC condition is not or cannot be identified. This is useful when analyzing data for a process that has an estab