|
Activity Number:
|
238
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2003 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Quality & Productivity
|
| Abstract - #300321 |
|
Title:
|
Control Charts for Radial Error
|
|
Author(s):
|
John I. McCool*+
|
|
Companies:
|
Pennsylvania State University-Great Valley
|
|
Address:
|
250 Ridge Pike Apt. 150B, Lafayette Hill, PA, 19444-1930,
|
|
Keywords:
|
control charts ; radial error ; eccentricity ; Weibull ; shape parameter
|
|
Abstract:
|
Radial error is a concern in military applications. Weapons manufacturers care about the distance from a targeted location to where a bullet or bomb strikes. Radial error is also important in processes such as the drilling of holes or the placement of components on a circuit board. Manufacturers must hold radial error within a specified tolerance. When the x and y coordinates of the hole center vary normally and independently about the desired center location with a common variance s^2, the radial error follows a Weibull distribution with shape parameter 2. The mean and variance of the radial error both depend solely on s and control charts may be constructed for individual values of the radial error when s is known. The process is out of control when the mean locations of the x and y errors differ from their desired location. The probability of an out-of-control signal is computed as a function of the radial distance from the desired center to the actual center. Out-of-control signals will also become more numerous if the process variability s increases. It is shown that the Weibull shape parameter estimate may be used to distinguish the two causes for an out-of-control signal.
|
