Abstract
Fault detection and isolation schemes are designed to detect the onset of adverse events during operations of complex systems, such as aircraft, power plants, and industrial processes. In this paper, we combine unsupervised learning techniques with expert knowledge to develop an anomaly detection method to find previously undetected faults from a large database of flight operations data. The unsupervised learning technique combined with a feature extraction scheme applied to the clusters labeled as anomalous facilitates expert analysis in characterizing relevant anomalies and faults in flight operations. We present a case study using a large flight operations data set, and discuss results to demonstrate the effectiveness of our approach. Our method is general, and equally applicable to manufacturing processes and other industrial applications.
Zusammenfassung
Fehlererkennungs- und Fehlerisolierungsschemata sollen das Auftreten unerwünschter Ereignisse während des Betriebs komplexer Systeme wie Flugzeuge, Kraftwerke und Industrieprozesse erkennen. In dieser Arbeit kombinieren wir unüberwachte Lerntechniken mit Expertenwissen, um ein Anomalieerkennungsverfahren zu entwickeln, das in einer großen Datenbank von Flugbetriebsdaten bisher unentdeckte Fehler findet. Die Kombination unüberwachter Lerntechnik mit einem Merkmalsextraktionsschema, das auf die als anomal bezeichneten Cluster angewendet wird, erleichtert die Expertenanalyse bei der Charakterisierung relevanter Anomalien und Fehler im Flugbetrieb. Wir präsentieren eine Fallstudie mit einem großen Flugbetriebsdatensatz und diskutieren Ergebnisse, um die Effektivität unseres Ansatzes zu demonstrieren. Unsere Methode ist allgemeingültig und gleichermaßen auf Herstellungsprozesse und andere industrielle Anwendungen anwendbar.
About the authors
Daniel L. C. Mack received his B.S. degree in Computer Science from the University of Notre Dame, Notre Dame, IN, USA, in 2006, and his M.S. degree in Computer Science from Columbia University, New York, NY, USA, in 2008, with a concentration in machine learning. He received his Ph. D. degree in computer science from Vanderbilt University, Nashville, TN, USA, in 2013, where his dissertation focused on machine learning and anomaly detection. While pursuing his doctorate, he worked as a Research Assistant at the Institute for Software Integrated Systems where he and his research group won the NASA Associate Administrator Award for Technology and Innovation for work combining machine learning with fault diagnosis. He is now the Senior Director of Quantitative Analysis/Amateur Scouting, for the Kansas City Royals. He works closely with the entire Quantitative staff to assist with research and development of analytics in support of all areas of baseball operations with a focus on amateur scouting.
Gautam Biswas received the Ph. D. degree in computer science from Michigan State University, East Lansing, MI, USA. He currently holds a Cornelius Vanderbilt Endowed Chair in Engineering and is a Professor of Computer science and Computer engineering in the Department of Electrical Engineering and Computer Science, and a Senior Research Scientist with the Institute for Software Integrated Systems at Vanderbilt University in Nashville, TN, USA. He is involved in research on intelligent systems with primary interest in hybrid modeling, simulation, and analysis of complex embedded systems, and their applications to diagnosis and fault-adaptive control. He is currently working on data-driven methods for anomaly detection in and diagnosis of complex systems. For this work, he received the NASA 2011 Aeronautics Research Mission Directorate Technology and Innovation Group Award for Vehicle Level Reasoning System and Data Mining methods to improve aircraft diagnostic and prognostic systems. He has over 600 refereed publications and his research has been funded by the AFRL, ARL, NASA, NSF, and DARPA. Prof. Biswas is a Fellow of the IEEE and the PHM society.
Hamed Khorasgani received the B. Sc. degree in electronics and electrical engineering from Isfahan University of Technology, Isfahan, Iran, in 2009, the M. Sc. degree in mechatronics engineering from Amirkabir University of Technology, Tehran, Iran, in 2012, and the PhD in electrical engineering from the Institute for Software Integrated Systems, Vanderbilt University, Nashville, TN, USA, in 2017. During his PhD at the Institute for Software Integrated Systems, he has conducted researches in analysis complex systems and their applications in diagnosis, prognostics, and fault tolerant control. His current research projects in Hitachi Big Data Lab include developing hybrid methodologies and solutions for integrating model-based and data-driven diagnosis methods.
Dinkar Mylaraswamy received the Ph. D. degree from Purdue University, West Lafayette, IN, USA, in 1997. He joined Honeywell Aerospace, Golden Valley, MN, USA, in 1997, after completing his Ph. D. degree. His Ph. D. thesis on blackboard-based architectures was adopted by the Abnormal Situation Management Consortium as the basis of an operator tool for addressing the $16 B loss suffered by the petrochemical industry from abnormal situations and equipment malfunctions. He is the Technology Fellow for condition-based maintenance within Honeywell’s Advanced Technology Organization. His area of expertise is fault diagnosis, process monitoring, modeling and control. In his current role, he is responsible for identifying and maturing strategic health management technologies that cut across multiple products and services, providing inputs for strategic technology investments, and mentoring. He spent the first six years in Honeywell developing and deploying an Early Abnormal Event Detection application at six refinery sites in North America. On the Aerospace side, he was the technical lead for Honeywell’s Predictive Trend Monitoring Program, a web-based application for monitoring aircraft engines. He continues to serve as the technical lead on various health management programs – within Honeywell as well the U.S. Army, NASA, UKMOD, and Navair – to support the Aero services, engines, mechanical components and avionics business within Honeywell. As the Technology Fellow, he routinely works with academic institutes and small businesses, seeking cutting-edge technologies to support the condition-based service business within Honeywell. He has authored over 30 papers and holds 23 patents in the area of fault diagnosis and its applications.
Raj M. Bharadwaj is a Scientist Staff in the Data Science and Vehicle Health Management group at Honeywell Aerospace Advanced Technologies. His work at Honeywell is centered on application of machine learning and prognostic health management. He has worked on system wide safety assurance, vehicle level reasoners, diagnostics, prognostics and adaptive systems. Dr. Bharadwaj was the Honeywell lead for NASA VLRS program. He is also a co-investigator on NASA Anomaly Detection in the National Airspace. He is currently leading many Honeywell internal and U.S. Government funded research including US Army ASTRO-D program which is using machine learning to improve HUMS prognostics for helicopter tail rotor drive systems. Prior to joining Honeywell, Dr Bharadwaj was a Senior Research Scientist at the General Electric Global Research Center, Niskayuna, NY where he focused on health management, system modeling and controls. Dr Bhardwaj received his Ph.D. from Texas A&M University in 2000. Dr Bharadwaj has over twenty co-authored publications, fourteen assigned US patents and many pending patents. He is currently serving on SAE –- HM1 standards committee that is working on several IVHM standards. He is the recipient of the 2006 First Prize Paper award from IEEE Industry Application Society. He also received 2011 NASA Aeronautics Research Mission Directorate Associate Administrators award for technology and innovation for his contributions on the NASA VLRS project. Dr. Bharadwaj presented the keynote on Machine Learning in Prognostics Health Management at the KDD2017 workshop on PHM.
References
1. L. Monostori, Cyber-physical production systems: Roots, expectations and r&d challenges, Procedia CIRP 17 (2014) 9–13.10.1016/j.procir.2014.03.115Search in Google Scholar
2. S. Thiede, M. Juraschek, C. Herrmann, Implementing cyber-physical production systems in learning factories, Procedia CIRP 54 (2016) 7–12.10.1016/j.procir.2016.04.098Search in Google Scholar
3. C. Ohm, M. Bürger, Ausblicke auf Industrie 4.0 und ihr Kybertariat, Das Argument 311 (2015) 16–31.Search in Google Scholar
4. O. Niggemann, G. Biswas, J. S. Kinnebrew, H. Khorasgani, S. Volgmann, A. Bunte, Data-driven monitoring of cyber-physical systems leveraging on big data and the internet-of-things for diagnosis and control., 26th International Workshop on Principles of Diagnosis, Paris, France.Search in Google Scholar
5. D. Chen, S. Heyer, S. Ibbotson, K. Salonitis, J. G. Steingrímsson, S. Thiede, Direct digital manufacturing: definition, evolution, and sustainability implications, Journal of Cleaner Production 107 (2015) 615–625.10.1016/j.jclepro.2015.05.009Search in Google Scholar
6. S. J. Qin, Survey on data-driven industrial process monitoring and diagnosis, Annual Reviews in Control 36(2) (2012) 220–234.10.1016/j.arcontrol.2012.09.004Search in Google Scholar
7. D. L. Mack, G. Biswas, X. D. Koutsoukos, Learning Bayesian network structures to augment aircraft diagnostic reference models, IEEE Transactions on Automation Science and Engineering 14(1) (2017) 358–369.10.1109/TASE.2016.2542186Search in Google Scholar
8. D. L. Mack, Anomaly detection from complex temporal sequences in large data, Ph.D. thesis, Vanderbilt University (2013).Search in Google Scholar
9. C. S. Burrus, R. A. Gopinath, H. Guo, Introduction to wavelets and wavelet transforms: a primer, Prentice-Hall, Inc., 1997.Search in Google Scholar
10. S. C. Johnson, Hierarchical clustering schemes, Psychometrika 32(3) (1967) 241–254.10.1007/BF02289588Search in Google Scholar
11. V. Chandola, A. Banerjee, V. Kumar, Anomaly detection: A survey, CM computing surveys (CSUR) 41(3) (2009) 15–72.10.1145/1541880.1541882Search in Google Scholar
12. V. Venkatasubramanian, R. Rengaswamy, S. N. Kavuri, K. Yin, A review of process fault detection and diagnosis: Part iii: Process history based methods, Computers and chemical engineering 27(3) (2003) 327–346.10.1016/S0098-1354(02)00162-XSearch in Google Scholar
13. C. Bishop, Pattern recognition and machine learning, Springer, 2006.Search in Google Scholar
14. E. Chu, D. Gorinevsky, S. Boyd, Detecting aircraft performance anomalies from cruise flight data, In AIAA Infotech Aerospace Conference, Atlanta, GA, April, 2010.10.2514/6.2010-3307Search in Google Scholar
15. S. Budalakoti, A. N. Srivastava, M. E. Otey, Anomaly detection and diagnosis algorithms for discrete symbol sequences with applications to airline safety, IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews 39(1) (2009) 101–113.10.1109/TSMCC.2008.2007248Search in Google Scholar
16. S. D. Bay, M. Schwabacher, Mining distance-based outliers in near linear time with randomization and a simple pruning rule, In Proceedings of the ninth ACM SIGKDD international con-ference on Knowledge discovery and data mining (2003) 29–38.10.1145/956750.956758Search in Google Scholar
17. D. L. Iverson, Inductive system health monitoring., Proceedings of the 2004 International Conference on Artificial Intelligence. Las Vegas, NV.Search in Google Scholar
18. T. R. Chidester, Understanding normal and atypical operations through analysis of flight data, In Proceedings of the 12th International Symposium on Aviation Psychology, Dayton, OH, (2003) 239–242.Search in Google Scholar
19. L. Li, S. Das, R. J. Hansman, R. Palacios, A. N. Srivastava, Analysis of flight data using clustering techniques for detecting abnormal operations, Journal of Aerospace Information Systems 12(9) (2015) 587–598.10.2514/1.I010329Search in Google Scholar
20. S. Das, B. L. Matthews, R. Lawrence, Fleet level anomaly detection of aviation safety data., In IEEE Conference on Prognostics and Health Management (PHM) (2011) 1–10.10.1109/ICPHM.2011.6024356Search in Google Scholar
21. G. Ratsch, S. Mika, B. Scholkopf, K.-R. Muller, Constructing boosting algorithms from svms: an application to one-class classification, IEEE Transactions on Pattern Analysis and Machine Intelligence 24(9) (2002) 1184–1199.10.1109/TPAMI.2002.1033211Search in Google Scholar
22. H. Peng, F. Long, C. Ding, Feature selection based on mutual information criteria of max-dependency, max-relevance, and min-redundancy, IEEE Transactions on pattern analysis and machine intelligence 27(8) 1226–1238.10.1109/TPAMI.2005.159Search in Google Scholar
23. G. Strang, Wavelet transforms versus Fourier transforms, Bulletin of the American Mathematical Society 28(2) (1993) 288–305.10.1090/S0273-0979-1993-00390-2Search in Google Scholar
24. R. S. Stankovic, B. J. Falkowski, The haar wavelet transform: its status and achievements, Computers and Electrical Engineering 29(1) (2003) 25–44.10.1016/S0045-7906(01)00011-8Search in Google Scholar
25. A. K. Jain, R. C. Dubes, Algorithms for clustering data, Prentice-Hall, Inc., 1988.Search in Google Scholar
26. G. Biswas, H. Khorasgani, G. Stanje, A. Dubey, S. Deb, S. Ghoshal, An approach to mode and anomaly detection with spacecraft telemetry data, International Journal of Prognostics and Health Management.Search in Google Scholar
27. F. Murtagh, A survey of recent advances in hierarchical clustering algorithms, The Computer Journal 26(4) (1983) 354–359.10.1093/comjnl/26.4.354Search in Google Scholar
28. A. K. Jain, R. C. Dubes, Algorithms for clustering data., Prentice-Hall, Inc.Search in Google Scholar
29. J. S. Farris, On the cophenetic correlation coefficient, Systematic Biology 18(3) (1969) 279–285.10.2307/2412324Search in Google Scholar
30. G. W. Milligan, M. C. Cooper, An examination of procedures for determining the number of clusters in a data set, Psychometrika 50(2) (1985) 159–179.10.1007/BF02294245Search in Google Scholar
31. M. Yan, Methods of determining the number of clusters in a data set and a new clustering criterion, Ph.D. thesis, Virginia Polytechnic Institute and State University, 2005.Search in Google Scholar
32. T. Calinski, J. Harabasz, A dendrite method for cluster analysis, Communications in Statistics-theory and Methods 3(1) (1974) 1–27.10.1080/03610927408827101Search in Google Scholar
33. W. J. Krzanowski, Y. T. Lai, A criterion for determining the number of groups in a data set using sum of squares clustering, Biometrics 44 (1988) 23–34.10.2307/2531893Search in Google Scholar
34. G. Biswas, G. Simon, N. Mahadevan, S. Narasimhan, J. Ramirez, G. Karsai, A robust method for hybrid diagnosis of complex systems, In Proceedings of the 5th Symposium on Fault Detection, Supervision and Safety for Technical Processes, pp. 1125–1131.Search in Google Scholar
35. W. J. Conover, W. J. Conover, Practical nonparametric statistics., Wiley, NY.Search in Google Scholar
36. J. Faith, Targeted projection pursuit for interactive exploration of high-dimensional data sets, 11th International IEEE Conference on Information Visualization (IV 07): 286–292, 2007.10.1109/IV.2007.107Search in Google Scholar
37. L. Li, S. Das, R. J. Hansman, R. Palacios, A. N. Srivastava, Analysis of flight data using clustering techniques for detecting abnormal operations, Journal of Aerospace Information Systems.Search in Google Scholar
© 2018 Walter de Gruyter GmbH, Berlin/Boston
