Information Extraction from Industrial Sensor Data Using Time Series Meta-Features
Source abstract
Abstract
In the smart manufacturing sector, analyzing time series data is essential for monitoring plants and machinery to prevent costly failures or shutdowns. In order to gain new insights and make better control decisions, new methods are needed for extracting information and interpreting sensor data from hundreds of systems. In this paper, we present an approach for visualizing and interpreting sensor data from TRIMET Aluminium SE Essen (TAE) using time series meta-features and principal component analysis (PCA). We describe our general approach of generating multiple two-dimensional feature spaces to identify salient and implausible sensor data. Using a set of 20 time series meta-features, we applied our approach to sensor data from TAE which were generated by thermocouples. Each step of the approach was integrated into a dashboard to ensure a user-friendly and approachable interaction in finding salient and implausible sensor data.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Time Series Analysis and Forecasting · Forecasting Techniques and Applications · Anomaly Detection Techniques and Applications
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Institute for Automation/Computer Science, University of Wuppertal, Rainer-Gruenter-Str. 21, 42119 Wuppertal, Germany
WSW Wuppertaler Stadtwerke GmbH, Bromberger Str. 39-41, 42281 Wuppertal, Germany
TRIMET Aluminium SE, Aluminiumallee 1, 45356 Essen, Germany
Alcoa Nederland Holding B.V., Weena 798, 3014 DA Rotterdam, The Netherlands
Funding
European Union and the European Regional Development Fund · Open Access Publication Fund of the University of Wuppertal
References (32)
Show source-provided references
1. Aluminium process fault detection by Multiway Principal Component Analysis10.1016/j.conengprac.2010.12.005 ↗
2. Kremser, R., Grabowski, N., Düssel, R., Kessel, K., and Tutsch, D. (2018, January 2–7). Investigation of Different Measurement Techniques for Individual Anode Currents in Hall-Héroult Cells. Proceedings of the 12th Australasian Aluminium Smelting Technology Conference, Queenstown, New Zealand.
3. Pursche, T., Grabowski, N., Nowitzki, J., Claub, R., Patryarcha, L., Dreisbach, H., and Tibken, B. (2018, January 11–14). Identification of Overtemperature Disturbances in Industrial Food Refrigeration Processes. Proceedings of the 2018 57th Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE), Nara, Japan.10.23919/SICE.2018.8492663 ↗
4. Fulcher, B.D. (2018). Feature Engineering for Machine Learning and Data Analytics, CRC Press/Taylor & Francis Group.
5. Characteristic-Based Clustering for Time Series Data10.1007/s10618-005-0039-x ↗
6. A Perspective View and Survey of Meta-Learning10.1023/A:1019956318069 ↗
7. Hyndman, R. (2023, April 21). Mcomp: Data from the M-Competitions, R package Version 2.8. Available online: https://cran.r-project.org/package=Mcomp.
8. O’Hara-Wild, M., Hyndman, R., Wang, E., and Feasts: Feature Extraction and Statistics for Time Series (2023, April 21). R package Version 0.1.6 (conda-forge). Available online: https://cran.r-project.org/package=feasts.
9. Talagala, T.S., Hyndman, R.J., and Athanasopoulos, G. (2023). Meta-learning how to forecast time series. J. Forecast.10.1002/for.2963 ↗
10. GRATIS: GeneRAting TIme Series with diverse and controllable characteristics10.1002/sam.11461 ↗
11. Grjotheim, K., and Kvande, H. (1993). Introduction to Aluminium Electrolysis: Understanding the Hall-Héroult Process, Alu Media GmbH. [2nd ed.].
12. Düssel, R. (2016). Entwicklung eines Regelungskonzepts für Aluminium-Elektrolysezellen unter Berücksichtigung einer Variablen Stromstärke und eines Regelbaren Wärmeverlusts. [Ph.D. Thesis, Bergische Universität Wuppertal].
13. Depree, N., Düssel, R., Patel, P., and Reek, T. (2016). Light Metals 2016, Springer.
14. Chesonis, C. (2019). Light Metals 2019, Springer.10.1007/978-3-030-05864-7 ↗
15. Hyndman, R.J., Wang, E., and Laptev, N. (2015, January 14–17). Large-Scale Unusual Time Series Detection. Proceedings of the 2015 IEEE International Conference on Data Mining Workshop (ICDMW), Atlantic, NJ, USA.10.1109/ICDMW.2015.104 ↗
16. Grabowski, N., Kremser, R., Düssel, R., Mulder, A., and Tutsch, D. (2018, January 2–7). Using Random Forest Regression for Predicting and Analysing Reduction Cell Behaviour. Proceedings of the 12th Australasian Aluminium Smelting Technology Conference, Queenstown, New Zealand.
17. Kremser, R., Grabowski, N., Düssel, R., Mulder, A., and Tutsch, D. (2020). Anode Effect Prediction in Hall-Héroult Cells Using Time Series Characteristics. Appl. Sci., 10.10.3390/app10249050 ↗
18. catch22: CAnonical Time-series CHaracteristics10.1007/s10618-019-00647-x ↗
19. Garza, F., Gutierrez, K., Challu, C., Moralez, J., Olivares, R., and Mergenthaler, M. (2023, April 21). tsfeatures. Python Package. Available online: https://github.com/Nixtla/tsfeatures.
20. Nanopoulos, A., Alcock, R., and Manolopoulos, Y. (2001). Information Processing and Technology, Nova Science Publishers, Inc.
21. Data Mining For Quality Control of Primary Aluminium Production Process
22. Anomaly Detection in Streaming Nonstationary Temporal Data10.1080/10618600.2019.1617160 ↗
23. Visualising forecasting algorithm performance using time series instance spaces10.1016/j.ijforecast.2016.09.004 ↗
24. The M3-Competition: Results, conclusions and implications10.1016/S0169-2070(00)00057-1 ↗
25. Meta-learning approaches to selecting time series models10.1016/j.neucom.2004.03.008 ↗
26. NOEMON: Design, implementation and performance results of an intelligent assistant for classifier selection
27. Manolescu, P., Duchesne, C., Tessier, J., and Saevarsdottir, G. (2018). Light Metals, Springer.
28. LaJambe, D., Poulin, É, Duchesne, C., and Tessier, J. (2020). Light Metals, Springer.
29. Yao, Y., Cheung, C.Y., Bao, J., Skyllas-Kazacos, M., Welch, B.J., and Akhmetov, S. (2016). Light Metals, Springer.
30. Majid, N.A.A. (2011). Cascade Fault Detection and Diagnosis for the Aluminium Smelting Process using Multivariate Statistical Techniques. [Ph.D. Thesis, University of Auckland].
31. hctsa: A Computational Framework for Automated Time-Series Phenotyping Using Massive Feature Extraction10.1016/j.cels.2017.10.001 ↗
32. The Dash Development Team (2021). Dash, The Dash Development Team. Available online: https://github.com/plotly/dash.