研究生: |
陳勁宇 Jin-Yu Chen |
---|---|
論文名稱: |
基於非監督式串流資料分析之錫膏檢測 Unsupervised Streaming Data Analysis for Solder Paste Inspection Process |
指導教授: |
楊朝龍
Chao-Lung Yang |
口試委員: |
歐陽超
Chao Ou-Yang 黃奎隆 Kwei-Long Huang |
學位類別: |
碩士 Master |
系所名稱: |
管理學院 - 工業管理系 Department of Industrial Management |
論文出版年: | 2021 |
畢業學年度: | 109 |
語文別: | 中文 |
論文頁數: | 85 |
中文關鍵詞: | 錫膏檢測 、邊緣運算 、雲端運算 、異常偵測 |
外文關鍵詞: | SPI, Edge Computing, Cloud Computing, Anomaly Detection |
相關次數: | 點閱:174 下載:0 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
錫膏檢測 (Solder Paste Inspection, SPI) 為電子設備製造業中重要的檢測流程之一。透過測量印刷電路板 (Printed Circuit Board, PCB) 上錫膏的尺寸與體積,以針對製程進行後續的品質分析。本研究旨在建立一個非監督式異常檢測系統,對從SPI所收集而來的串流資料進行分析。此串流資料分析系統採用一個結合邊緣運算與雲端運算的串流數據分析框架,藉此解決由表面黏著技術 (Surface Mount Technology, SMT) 產線中取得的SPI檢測資料之數據量過於龐大的問題。本研究分別利用具有自動時間窗格設定且基於局部遞歸率的異常搜索方法 (Local Recurrence Rate based Discord Search with Automatic Time Window, LRRDS-ATW),以及基於遞歸率之具穩健性的K均值分群演算法 (Local Recurrence Rate and Robust K-Means clustering, LRR-RKMeans),進行此數據框架內邊緣運算模型與雲端運算模型的配置。本研究之實驗以實際場域所收集而來的SPI資料進行分析,並配合上述所建立之框架,利用邊緣運算模型在邊緣端所收集的原始時間序列資料中進行初步偵測,並將檢測出的異常時間序列視為代表性的數據,傳至雲端進行進一步的檢測分析。而為了確保在僅傳送邊緣運算模型所偵測出的異常時間序列是否能夠完整代表原始資料,該實驗中使用混淆矩陣 (confusion matrix) 將邊緣端偵測結果與雲端偵測結果進行比較,並確認其相似程度。研究結果顯示出若使用LRRDS-ATW將能夠達到77%的相似程度,而LRR-RKMeans甚至可達到86%的相似程度。此兩種模型所顯示出的結果,表示於邊緣端所偵測出的異常時間序列資料,具有足夠的特徵能夠代表原始資料。
Solder Paste Inspection (SPI) is one of the important inspection processes in electronic device manufacturing. By measuring the size and volume of the solder paste on the Printed Circuit Board (PCB), the quality analysis can be performed on the manufacturing process. In this research, an unsupervised streaming data analysis system was proposed to detect the SPI data. This stream data analysis system used a stream data analysis framework that combined edge computing and cloud computing, and aimed to solve the problem of analyzing the SPI data. This research used Local Recurrence Rate based Discord Search with Automatic Time Window (LRDS-ATW) and Local Recurrence Rate and Robust K-Means clustering (LRR-RKMeans) to configure the edge computing model and cloud computing model within the streaming data frame. The experiments of this research analyzed the SPI data collected from the real field, and cooperated with the framework established above, used the edge computing model to make preliminary detections from the original time series data at the edge site, and then the detected abnormal patterns as the representative data, the abnormal patterns were transmitted to the cloud for further detection and analysis. In order to ensure that the abnormal time series detected at only the edge site can completely represent the original data. In the experiment, a confusion matrix was used to compare the edge results with the cloud results to confirm the similarity. From the results, it can be seen that if LRRDS-ATW is used, it will be able to achieve a similarity of 77%, and LRR-RKMeans even can reach a similarity of 86%. The results of both two methods showed that the abnormal time series data detected from the edge site were well representative enough to present the characteristic of original data.
[1] Y. H. Yoo, U. H. Kim, and J. H. Kim, "Convolutional Recurrent Reconstructive Network for Spatiotemporal Anomaly Detection in Solder Paste Inspection," IEEE Transactions on Cybernetics, pp. 1-13, 2020, doi: 10.1109/TCYB.2020.3033798.
[2] C.-Y. Huang, Y.-H. Lin, K.-C. Ying, and C.-L. Ku, "The solder paste printing process: Critical parameters, defect scenarios, specifications, and cost reduction," Soldering & Surface Mount Technology, vol. 23, pp. 211-223, 09/20 2011, doi: 10.1108/09540911111169057.
[3] W. Huihui, Z. Xianmin, K. Yongcong, and L. Shenglin, "A real-time machine vision system for solder paste inspection," in 2008 IEEE/ASME International Conference on Advanced Intelligent Mechatronics, 2-5 July 2008 2008, pp. 205-210, doi: 10.1109/AIM.2008.4601660.
[4] C.-L. Yang, F. Darwin, and H. Sutrisno, "Local Recurrence Rates with Automatic Time Windows for Discord Search in Multivariate Time Series," Procedia Manufacturing, vol. 39, pp. 1783-1792, 01/01 2019, doi: 10.1016/j.promfg.2020.01.261.
[5] C.-L. Y. H. Sutrisno, "Discovering defective products based on multivariate sensors data using local recurrence rate and robust k-means clustering," presented at the International Conference on Production Research (ICPR 2021), Taichung, Taiwan, July 18-21, 2021, 2021.
[6] C. Yang et al., "Streaming data analysis framework for cyber-physical system of metal machining processes," in 2018 IEEE Industrial Cyber-Physical Systems (ICPS), 15-18 May 2018 2018, pp. 546-551, doi: 10.1109/ICPHYS.2018.8390764.
[7] M. Elsisi, M. Q. Tran, K. Mahmoud, D. E. A. Mansour, M. Lehtonen, and M. M. F. Darwish, "Towards Secured Online Monitoring for Digitalized GIS Against Cyber-Attacks Based on IoT and Machine Learning," IEEE Access, vol. 9, pp. 78415-78427, 2021, doi: 10.1109/ACCESS.2021.3083499.
[8] D. Burr, "Solder paste inspection: process control for defect reduction," in Proceedings International Test Conference 1997, 6-6 Nov. 1997 1997, p. 1036, doi: 10.1109/TEST.1997.639726.
[9] T. Chen, J. Zhang, Y. Zhou, and Y. Murphey, A Smart Machine Vision System for PCB Inspection. 2001, pp. 513-518.
[10] Y. M. Chang, C. C. Wei, J. Chen, and P. Hsieh, "An Implementation of Health Prediction in SMT Solder Joint via Machine Learning," in 2019 IEEE International Conference on Big Data and Smart Computing (BigComp), 27 Feb.-2 March 2019 2019, pp. 1-4, doi: 10.1109/BIGCOMP.2019.8679428.
[11] J. Schmitt, J. Bönig, T. Borggräfe, G. Beitinger, and J. Deuse, "Predictive model-based quality inspection using Machine Learning and Edge Cloud Computing," Advanced Engineering Informatics, vol. 45, p. 101101, 08/01 2020, doi: 10.1016/j.aei.2020.101101.
[12] S. Zakaria, A. Amir, N. Yaakob, and S. Nazemi, "Automated Detection of Printed Circuit Boards (PCB) Defects by Using Machine Learning in Electronic Manufacturing: Current Approaches," IOP Conference Series: Materials Science and Engineering, vol. 767, p. 012064, 03/21 2020, doi: 10.1088/1757-899X/767/1/012064.
[13] Y. Zou, M. Thiel, M. C. Romano, and J. Kurths, "Analytical description of recurrence plots of dynamical systems with nontrivial recurrences," International Journal of Bifurcation and Chaos, vol. 17, no. 12, pp. 4273-4283, 2007.
[14] J.-P. Eckmann, S. Kamphorst, and D. Ruelle, "Recurrence Plots of Dynamical Systems," Europhysics Letters (epl), vol. 4, pp. 973-977, 11/01 1987, doi: 10.1209/0295-5075/4/9/004.
[15] J. B. Gao, Y. Cao, L. Gu, J. G. Harris, and J. C. Principe, "Detection of weak transitions in signal dynamics using recurrence time statistics," Physics Letters A, vol. 317, no. 1, pp. 64-72, 2003/10/13/ 2003, doi: https://doi.org/10.1016/j.physleta.2003.08.018.
[16] C. Webber and J. Zbilut, "Dynamical assessment of physiological systems and states using recurrence plot strategies," Journal of applied physiology (Bethesda, Md. : 1985), vol. 76, pp. 965-73, 03/01 1994, doi: 10.1152/jappl.1994.76.2.965.
[17] J. M. Nichols, S. T. Trickey, and M. Seaver, "Damage detection using multivariate recurrence quantification analysis," Mechanical Systems and Signal Processing, vol. 20, no. 2, pp. 421-437, 2006/02/01/ 2006, doi: https://doi.org/10.1016/j.ymssp.2004.08.007.
[18] W. Luo, M. Gallagher, and J. Wiles, "Parameter-Free Search of Time-Series Discord," Journal of Computer Science and Technology, vol. 28, no. 2, pp. 300-310, 2013/03/01 2013, doi: 10.1007/s11390-013-1330-8.
[19] M. Hu, X. Feng, Z. Ji, K. Yan, and S. Zhou, "A novel computational approach for discord search with local recurrence rates in multivariate time series," Information Sciences, vol. 477, pp. 220-233, 2019/03/01/ 2019, doi: https://doi.org/10.1016/j.ins.2018.10.047.
[20] J. Lei, T. Jiang, K. Wu, H. Du, G. Zhu, and Z. Wang, "Robust K-means algorithm with automatically splitting and merging clusters and its applications for surveillance data," Multimedia Tools and Applications, vol. 75, pp. 12043-12059, 2016.
[21] H. Sakoe and S. Chiba, "Dynamic programming algorithm optimization for spoken word recognition," IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. 26, no. 1, pp. 43-49, 1978, doi: 10.1109/TASSP.1978.1163055.
[22] H. Kagermann, W. Wahlster, and J. Helbig, "Recommendations for Implementing the Strategic Initiative INDUSTRIE 4.0 -- Securing the Future of German Manufacturing Industry," München, 2013. [Online]. Available: http://forschungsunion.de/pdf/industrie_4_0_final_report.pdf
[23] S. Cavalieri, D. D. Stefano, M. G. Salafia, and M. S. Scroppo, "Integration of OPC UA into a web-based platform to enhance interoperability," in 2017 IEEE 26th International Symposium on Industrial Electronics (ISIE), 19-21 June 2017 2017, pp. 1206-1211, doi: 10.1109/ISIE.2017.8001417.
[24] R. Rajkumar, I. Lee, L. Sha, and J. Stankovic, "Cyber-physical systems: The next computing revolution," in Design Automation Conference, 13-18 June 2010 2010, pp. 731-736, doi: 10.1145/1837274.1837461.
[25] E. A. Lee, "Cyber Physical Systems: Design Challenges," in 2008 11th IEEE International Symposium on Object and Component-Oriented Real-Time Distributed Computing (ISORC), 5-7 May 2008 2008, pp. 363-369, doi: 10.1109/ISORC.2008.25.
[26] M. Satyanarayanan, "The Emergence of Edge Computing," Computer, vol. 50, no. 1, pp. 30-39, 2017, doi: 10.1109/MC.2017.9.
[27] R. Huffaker, "Phase Space Reconstruction from Time Series Data: Where History Meets Theory," 2010.
[28] J. S. Iwanski and E. Bradley, "Recurrence plots of experimental data: To embed or not to embed?," (in eng), Chaos, vol. 8, no. 4, pp. 861-871, Dec 1998, doi: 10.1063/1.166372.
[29] J. Theiler, "Spurious dimension from correlation algorithms applied to limited time-series data," Physical Review A, vol. 34, no. 3, pp. 2427-2432, 09/01/ 1986, doi: 10.1103/PhysRevA.34.2427.
[30] E. Keogh, J. Lin, and A. Fu, "HOT SAX: efficiently finding the most unusual time series subsequence," in Fifth IEEE International Conference on Data Mining (ICDM'05), 27-30 Nov. 2005 2005, p. 8 pp., doi: 10.1109/ICDM.2005.79.
[31] R. J. Hyndman and G. Athanasopoulos, Forecasting: Principles and Practice. OTexts, 2018.
[32] Z. Aliniya and S. A. Mirroshandel, "A novel combinatorial merge-split approach for automatic clustering using imperialist competitive algorithm," Expert Systems with Applications, vol. 117, pp. 243-266, 2019/03/01/ 2019, doi: https://doi.org/10.1016/j.eswa.2018.09.050.
[33] C.-L. Yang and H. Sutrisno, "A clustering-based symbiotic organisms search algorithm for high-dimensional optimization problems," Applied Soft Computing, vol. 97, p. 106722, 2020/12/01/ 2020, doi: https://doi.org/10.1016/j.asoc.2020.106722.
[34] G. Xiu and Z. Zhao, "Sustainable Development of Port Economy Based on Intelligent System Dynamics," IEEE Access, vol. 9, pp. 14070-14077, 2021, doi: 10.1109/ACCESS.2021.3051065.
[35] C. A. Ratanamahatana and E. Keogh, "Making Time-series Classification More Accurate Using Learned Constraints," in Proceedings of the 2004 SIAM International Conference on Data Mining (SDM), (Proceedings: Society for Industrial and Applied Mathematics, 2004, pp. 11-22.
[36] M. Shokoohi-Yekta, B. Hu, H. Jin, J. Wang, and E. Keogh, "Generalizing DTW to the multi-dimensional case requires an adaptive approach," Data Mining and Knowledge Discovery, vol. 31, no. 1, pp. 1-31, 2017/01/01 2017, doi: 10.1007/s10618-016-0455-0.
[37] X. Wang, J. Lin, N. Patel, and M. Braun, "Exact variable-length anomaly detection algorithm for univariate and multivariate time series," Data Mining and Knowledge Discovery, vol. 32, no. 6, pp. 1806-1844, 2018/11/01 2018, doi: 10.1007/s10618-018-0569-7.