Identification of the deviation from the mean and multimodal distribution in the automotive assembly process: Correspondence analysis approach

Abstract

The dimensional integrity of an automobile is one of the key quality characteristics, since poor fit can impact not only the downstream production processes, but also the customer's perception of quality. Principal component analysis (PCA) has been widely used to identify patterns of dimensional variation in the automotive body assembly process. It enables engineers to find measurement points that have large variances but fails to identify the measurement points that have large deviation from the mean with small variances. Correspondence analysis (CA) was applied to identify the deviation from the mean and bimodal distribution of measurement points in the automotive body assembly process. Two case studies are presented. By applying both CA and PCA, we have a better understanding for the deformation patterns in automotive body assembly process.