SPC for Particulate Alert/Action Levels

[Cec28007]

Can anyone suggest a good resource to describe how to set alert/action levels for particulate results based on SPC? We have minitab, I'm just having a hard time knowing how to utilize it to establish our internal alert/action levels. Any good articles, webinars, seminars that describes the process and does a good job of it, I'm looking for something that is very specific. Alot of what I'm finding is very general, simply says to 'use SPC' to set levels. Ultimately, I don't know how to, so I need assistance at a basic level.

 

[randomname]

First thing you need to do is study what SPC is all about. The statistical control limits are usually the action levels, although you can use other indications (runs, trends, etc.) if you want. Also need to understand difference between control limits and specifications.

 

[Steve Prevette]

Generally speaking, SPC has its control limits set at three standard deviations from the average of some past, stable data. If you are talking Industrial Hygiene here, what you will be detecting is CHANGE in exposure levels, which may be better to react to the change before you reach regulatory limits.

Dr. Wheeler has a simple, thin book called "Understanding Variation, the Key to Managing Chaos" which is a good buy. Unfortunately, the host of my materials on the internet got taken down a few years ago.

Off to the Chalk River site next week to help them with SPC and metrics. . .

 

[bobdoering]

Shewart charts expect that your variation is independent and random - otherwise it is a special cause. So, you need to determine if that is the type of variation you are experiencing. In order to determine the type of charting that is appropriate, first you need to understand the nature of your variation. What are the variances that affect the particulate level (total variance equation)? Is the particulate level a random variable? Or does it change as a function of time? Have you performed a time-ordered sequence data collection? How much of the variation is systemic, and how much is measurement noise? What is your "gage R&R" of your particulate level measurement technique? This blog may also help...