421 – Contributed Poster Presentations: Quality and Productivity Section
Statistical Process Control in the Presence of Multiple Batch Effects
Kelsea Cox
Boeing Research & Technology
Lindsay L. W Jones
Boeing Research & Technology
R. Michael Lawton
Boeing Research & Technology
Jay Palma
Boeing Research & Technology
Ensuring the quality of outgoing products is a critical component of most manufacturing processes; however, it is not always cost effective to test or inspect every unit produced. It is often beneficial to implement statistical process control (SPC) in order to reduce the cost associated with testing each product. To implement SPC, a process must first demonstrate its stability and demonstrate that it meets a specific, predefined qualification requirement, often stated in terms of process capability (Cpk). Methods are available to calculate Cpk with no batch effects or a single batch effect. However, many processes consist of multiple levels of batching. Improperly accounting for all batch effects can lead to incorrectly qualifying a process. No previous methods have been published to verify requirements in the presence of such batching structures. Additionally, SPC requires control charts to track changes in the process over time. We will describe a method for verifying Cpk requirements and building control charts for many batching structures. This development fills a technology gap in SPC, reducing the likelihood of mistakes by correctly characterizing process variation.