Application of Multivariate Control Chart Techniques to Identifying Nonconforming Pallets in Automotive Assembly Plants 2020-01-0477
The Hotelling multivariate control chart and the sample generalized variance |S| are used to monitor the mean and dispersion of vehicle build vision data including the pallet information to identify the non-conforming pallets that are used in body shops of FCA US LLC assembly plants. An iterative procedure and the Gaussian mixture model (GMM) are used to rank the non-conforming or bad pallets in the order of severity. The Hotelling multivariate T2 test statistic along with Mason-Tracy-Young (MYT) signal decomposition method is used to identify the features that are affected by the bad pallets. These algorithms were implemented in the Advanced Pallet Analysis module of the FCA US software Body Shop Analysis Toolbox (BSAT). The identified bad pallets are visualized in a scatter plot with a different color for each of the top bad pallets. The run chart of an affected feature confirms the bad pallet by highlighting data points from the bad pallet. The analysis module has been successfully used in the body shops of FCA US plants, including the plants in Canada and Mexico, in identifying the bad pallets. Two examples are presented to demonstrate the capability of the application.
Citation: Huang, M., Wang, Y., Shirinkam, S., Alaeddini, A. et al., "Application of Multivariate Control Chart Techniques to Identifying Nonconforming Pallets in Automotive Assembly Plants," SAE Technical Paper 2020-01-0477, 2020, https://doi.org/10.4271/2020-01-0477. Download Citation
Michael Huang, Yuqin Wang, Sara Shirinkam, Adel Alaeddini, Kai Yang
FCA US LLC, University of the Incarnate Word, University of Texas at San Antonio, Wayne State University