整合獨立成分分析與統計製程管制圖於產品件內和件間變異監控之應用

Monitoring of Within-Part and Between-Part Variations with Independent Component Analysis and Statistical Process Control

鄭春生(Chuen-Sheng Cheng) ； 黃國格(Kuo-Ko Huang)

《品質學報》 20卷1期 (2013/02) Pp. 137-154

https://doi.org/10.6220/joq.2013.20(1).07

摘要

隨著產品特性之複雜化，我們需要在一件產品的不同位置量測品質特性。因此在監控產品品質之量測值時，由變異來源造成的時間性變異樣式與空間性變異樣式可能會成為觀測到的件內和件間變異。利用傳統管制圖來監控原始觀測值的件內和件間變異雖然可行，但並非是一種有效率或者有效益之方法。此乃因為觀測值是由不同變異來源交互影響之後的結果。如果可以直接監控變異來源之變化，將會是一個更具有效率及效益之製程管制方法。本研究之目的是應用獨立成分分析自原始觀測數據分離出獨立成分（變異來源）後，再使用I-MR管制圖與Hotelling T^2管制圖對獨立成分進行監控。本研究所提出之監控方式將以一個模擬範例及一個物理氣相沉積薄膜的實際製程資料加以驗證。本研究是以監控製程觀測值之I-MR-R/S管制圖做為比較基準，並以平均連串長度作為績效指標。實驗結果顯示，當製程出現平均數偏移以及平均數呈趨勢遞增之製程異常情形時，本研究所提之方法會比傳統管制圖更快偵測到製程平均數的變化狀況。

關鍵字

件內和件間變異 ； 統計製程管制 ； 獨立成分分析

並列摘要

With the complexity of product characteristics, it is now necessary to measure quality characteristic at different locations across a part. In the monitoring of product quality measurements, the temporal pattern and spatial variation pattern caused by a variation source may turn out to be the observed within- and between-part variations. It is feasible to apply traditional control charts to monitoring the within- and between-part variations. However, it may be neither effective nor efficient due to the fact that observed measurements are mixture of several variation sources.The proposed scheme first applies ICA methodology to the process observations to generate the independent components that contain different characteristics of the process. The I-MR control chart and Hotelling T^2 control chart are then used to monitor the independent components for process control. The proposed procedures were implemented via a simulated processes and a case study of the physical vapor deposition process. The experimental results show that the proposed methods can detect faults faster than I-MR-R/S control chart.

並列關鍵字

within-part and between-part variations ； statistical process control ； independent component analysis

參考文獻

Chen, W. S., 2012, Variation pattern identification and fault diagnosis of solder paste deposit by using independent component analysis, Journal of Quality, 19(1), 21-39.

Apley, D. W. and Lee, H. Y., 2003, Identifying spatial variation patterns in multivariate manufacturing processes: a blind separation approach, Technometrics, 45(3), 220-234.

Cheng, H. P. and Cheng, C. S., 2009, Control chart pattern recognition using wavelet analysis and neural networks, Journal of Quality, 16(5), 311-321.

Hsu, C. C., Chen, M. C., and Chen, L. S., 2010a, Integrating independent component analysis and support vector machine for multivariate process monitoring, Computers and Industrial Engineering, 59(1), 145-156.

Hsu, C. C., Chen, M. C., and Chen, L. S., 2010b, Intelligent ICA-SVM fault detector for non-Gaussian multivariate process monitoring, Expert Systems with Applications, 37(4), 3264-3273.