Screwing process monitoring using MSPC in large scale smart manufacturing

Abstract

The ability to obtain useful information to support decision-making from big data sets delivered by sensors can significantly contribute to enhance smart manufacturing initiatives. This paper presents the results of a study performed in an automotive electronics assembly line. An approach that uses Multivariate Statistical Process Control based on Principal Component Analysis (MSPC-PCA) was applied to early detect undesirable changes in the screwing processes performance by extracting relevant information from the torque-angle curve data. Since the data of different torque-angle curves are not aligned, the proposed approach includes the linear interpolation of the original data to enable Principal Component Analysis (PCA). PCA proved to be an appropriate technique to obtain significant information from the process variables, which consist of the successive value of the torque at constant angular intervals. Score plots and multivariate control charts were used to detect defective tightening and identify behaviors that represent inefficient tightening. This is a new approach that can be applied to effectively monitor screwing processes in the assembly of different products either periodically or in real-time.