Online Program

Constructing a Statistical Upper Bound for Defect Sizes in Aerospace Parts in the Presence of Censoring, Spatial Autocorrelation, and Batching (309970)

Keywords: batch effects, autocorrelation, censoring, manufacturing, aerospace

In a manufacturing setting, it is important to understand the quality of outgoing products to ensure that requirements are met. It is not uncommon for manufacturing quality data to be of limited sample size and measurements are frequently censored due to limitations in measurement methods. Characterizing defects in these contexts which often also entail moderate part-to-part or batch-to-batch variation can prove challenging for standard statistical methods used in industrial quality control systems. In this presentation, we discuss a statistical method for determining the upper bound on defect sizes in a population of products in the presence of censoring, spatial autocorrelation, and batch effects.