Advertisement

Neural Computing and Applications

, Volume 26, Issue 3, pp 581–588 | Cite as

Large, high-dimensional, heterogeneous multi-sensor data analysis approach for process yield optimization in polymer film industry

  • Michael KohlertEmail author
  • Andreas König
Advances in Intelligent Data Processing and Analysis

Abstract

Today’s advanced complex manufacturing processes from microelectronics to pharmaceutical industries provide huge datasets (big data) from heterogeneous multi-sensory monitoring. Analytical tools for such high-dimensional and large datasets are available to support manufacturers in quality and yield optimization, but latent problems efficiently linking process data to these tools exist, the available methods are not fully satisfactory, and the real-time support is very limited. In this work, one particular industrial process from polymer industry was investigated as a research vehicle for the development of methods for efficient process interfacing, process status classification and prediction, as well as online result visualization. Novelty filtering, anomaly detection, and one-class classification (OCC) methods have been in the focus of the investigation. These methods are of particular importance as expert knowledge from the production lines discloses that the process shows unexpected states and behavior during production runs, which lead to low yield and high cost. Causes for this abnormal behavior are assumed sensor faults, environmental influences, and unexpected raw material properties’ deviations. Thus, novelty filters, anomaly detection, and OCC methods are particularly suitable for such cases. A process line from standard production with approximately 160 sensory channels has been monitored for 3 months in 1-min intervals. This provided 21,900 process data mostly normal condition samples of about 160 dimensions each for the conducted experiments. Accuracies of 99 % in OCC were achieved, and a first prototype of an advanced novelty detection GUI for process monitoring was conceived, which supports the process monitoring by the responsible staff. The achieved results allow predicting process yield problems within 2 h in advance of occurrence despite unknown objects. Future work will focus on exploiting this for advanced process control and yield optimization.

Keywords

One-class classification (OCC) Novelty detection Multi-sensor data analysis Process yield prediction and optimization Process state visualization 

References

  1. 1.
    de Ridder D, Tax D, Duin R (1998) An experimental comparison of one-class classification methods. In: Proceedings of the 4th annual conference of the advanced school for computing and imagingGoogle Scholar
  2. 2.
    Martin W (2008) Trends in data mining. BeyeNETWORK. http://www.beyenetwork.de/view/8192. Accessed 28 Feb 2014
  3. 3.
    Steinwart I, Hush D, Scovel C (2005) A classification framework for anomaly detection. J Mach Learn Res, MIT Press, MIT University, 211–232Google Scholar
  4. 4.
    Tax D (2001) One-class classification: concept-learning in the absence of counterexamples. Doctoral Dissertation, University of Delft, The NetherlandsGoogle Scholar
  5. 5.
    Khan S, Madden MG (2010) A survey of recent trends in one class classification. Artif Intell Cognit Sci Lect Notes Comput Sci 6206:188–197CrossRefGoogle Scholar
  6. 6.
    Smets K, Vreeken J (2011) The odd one out: Identifying and characterising anomalies. In: Proceedings of the 11th SIAM international conference on data mining (SDM), Mesa, AZ, pp 804–815Google Scholar
  7. 7.
    Bartkowiak A (2010) Anomaly, novelty, one-class classification: a short introduction. In: Proceedings of computer information systems and industrial management applications 1–6 IEEE Xplore. doi: 10.1109/CISIM.2010.5643699
  8. 8.
    König A, Raschhofer RJ, Glesner M (1994) A novel method for the design of radial-basis-function-networks and its implication for knowledge extraction IEEE world congress on computational intelligence. doi: 10.1109/ICNN.1994.374430
  9. 9.
    König A, Windirsch P, Gasteier M, Glesner M (1995) Visual inspection in industrial manufacturing. IEEE Micro 15:26–31CrossRefGoogle Scholar
  10. 10.
    König A, Gratz A (2005) Advanced methods for the analysis of semiconductor manufacturing process data. In: Pal NR, Jain LC (eds) Advanced techniques in knowledge discovery and data mining. Springer, Berlin, pp 27–74CrossRefGoogle Scholar
  11. 11.
    Gneuss Kunststofftechnik GmbH. http://www.gneuss.de/. Accessed 28 Feb 2014
  12. 12.
    Hodge V, Austin J (2004) A survey of outlier detection methodologies. Artif Intell Rev 22(2):85–126CrossRefzbMATHGoogle Scholar
  13. 13.
    Shin HJ, Eom D-W, Kim S-S (2005) One-class support vector machines: an application in machine fault detection and classification. Comput Ind Eng 48(2):395–408CrossRefGoogle Scholar
  14. 14.
    Kohlert M, König A (2012) Advanced polymeric film production data analysis and process optimization by clustering and classification methods. In: KES, pp 1953–1961 Google Scholar
  15. 15.
    Moyne J, Del Castillo E, Hurwitz A (2001) Run to run control in semiconductor manufacturing. CRC Press, Boca Raton, FL Google Scholar
  16. 16.
    König A (2001) Dimensionality reduction techniques for interactive visualisation, exploratory data analysis and classification. In: Pal NR (ed) Pattern recognition in soft computing paradigm. World Scientific, FLSI Soft Computing Series, vol. 2, ISBN 981-02-4491-6, pp 1–37Google Scholar
  17. 17.
    Sun D, Tran Q, Duan H, Zhang G (2005) A novel method for Chinese spam detection based on one-class support vector machine. J Inf Comput Sci 2(1):109–114Google Scholar
  18. 18.
    Tax DMJ, Duin RPW (2004) Support vector data description. Mach Learn 54:45–66CrossRefzbMATHGoogle Scholar
  19. 19.
    Duin RPW, et al (2007) A matlab toolbox for pattern recognition. PRTools. http://prtools.org/. Accessed 28 Feb 2014
  20. 20.
    Tax DMJ (2013) Data description toolbox. Pattern Recognition Laboratory. http://prlab.tudelft.nl/david-tax/dd_tools.html. Accessed 28 Feb 2014
  21. 21.
    König A, Glesner M (1992) VLSI implementation of associative memory systems for neural information processing. In: Proceedings of the third Int’l Workshop on VLS/for neural networks and artificial intelligenceGoogle Scholar
  22. 22.
    The MathWorks Inc. http://www.mathworks.de/. Accessed 28 Feb 2014

Copyright information

© The Natural Computing Applications Forum 2014

Authors and Affiliations

  1. 1.Department of Electrical and Computer Engineering, Institute of Integrated Sensor SystemsUniversity of KaiserslauternKaiserslauternGermany
  2. 2.Production DepartmentsMondi Gronau GmbHGronauGermany

Personalised recommendations