use of the UCL value to apply alert limits for various defects

Sinéad

Hi All

I'm relatively new to JMP and navigating my way with great help from JMP tutorials etc

ive being tasked to apply alert limits to various defects within our department.

so i have variable data across 30 or more batches i used Laney P to generate the graph and I tabulate the mean of my UCL and this is the value that I apply as my alert limit for a specific defect type. then after a year these values will be recalculated based on that years data collected for these defects.

am i correct in my approach in the use of the UCL value (as this would account for over and under dispersion of my data)

the average value would not be used to set my alert limit ? please can someone explain exactly why i wouldn't use the average(my thinking is that the average is what we lost across the 30 batches from a business perspective and that would be what we expect to loose, and that the UCL is applied to ensure our process is in control 

 

 

P_Bartell

This is going to be philosophical rather than methodological,,,,but my comments are at the heart of your basic question, which I believe is, "Where should I set control limits?" If one goes back to the fundamental purpose of control limits, they are limits which provide a signal for economically balancing the risk of making two mistakes:

 

1. Searching for assignable cause variation when in fact none is present.

2. Not searching for assignable cause variation when in fact it is present.

 

The key phrase above is '...economically balancing...'. There is nothing magical about any mathematical algorithm to calculate control limits using good old fashioned Shewhart derived limits or otherwise. Shewhart himself said his process seemed to 'work' practically in a wide variety of settings. So if you can justify 'where' you set your limits from an economic balance point of view...then nobody can tell you your limits are inappropriate. From your narrative it sounds like you are thinking along those lines...but only you and your process knowledge can appropriately apply this economic balance thinking.