A Comparison of Process Variation Estimators for In-Process Dimensional Measurements and Control

Dimensional variation reduction is critical to assure high product quality in discrete-part manufacturing. Recent innovations in sensor technology enable in-process implementation of laseroptical coordinate sensors and continuous monitoring of product dimensional quality. The abundance of measurement data provides an opportunity to develop next generation process control technologies that not only detect process change but also provide guidelines respective of root cause identification. Given continuous product dimensional measurements, a critical step leading to root cause identification is the variance estimation of process variation sources. A few on-line variance estimators are available. The focus of this paper is to study the inter-relationships and properties of the available variance estimators and compare their performance. An OC-curve (Operating Characteristics) is developed as a convenient tool to guide the appropriate use of on-line variance estimators under specific circumstances. The method is illustrated using examples of dimensional control for a panel assembly process.