Luh Juni Asrini, Dian Retno Sari Dewi, Irene Karijadi
Many modern industrial processes involve multiple quality measures, and using individual control charts for each measure can be misleading if these measures are highly related. This paper proposes a new method for statistically controlling electronic products with multiple, interconnected quality characteristics. The method utilizes a combined model: a multivariate autoregressive (MAR) model with neural networks, to handle the presence of both correlation and autocorrelation in the data. The study compares the effectiveness of MCUSUM, MEWMA, and T2 Hotelling charts in detecting small shifts in the overall process quality. To pinpoint the specific variables causing out-of-control signals in the T2 Hotelling chart, we introduce a novel decomposition technique. This technique allows us to identify which measures are contributing most to these signals. Additionally, the MCUSUM and MEWMA charts demonstrate excellent performance in detecting small quality changes, leading to faster corrective actions. Overall, these findings suggest that our proposed method can significantly improve the reliability and responsiveness of quality control in electronics manufacturing. ©2024 The authors. This article is published by IIETA and is licensed under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).
Department of Industrial Engineering, Widya Mandala Surabaya Catholic University, Surabaya, 60114, Indonesia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock