ABSTRACT
Model-based condition monitoring (MBCM) solves the inverse problem of inferring a systems state, including possible faults, from sensor observations. Constructing these models in a knowledge-based manner following the laws of physics is hard due to the inverse nature of the problem and unknown fault types. As a result, it has become more attractive to build a model solely from past observations via machine learning (ML). Although highly promising, shortcomings of ML in the scientific domain, including physically inconsistent results and lack of interpretability, became apparent. This led to recent efforts to enhance machine learning with scientific knowledge including a combination of knowledge-based and data-driven modelling, often referred to as hybrid models. The main contributions of this work are: (1) a link of shortcomings of machine learning in CM to a lack of knowledge; (2) a categorization of unique approaches with respect to required knowledge and mechanism of incorporation that have either been applied in condition monitoring or show potential from their application to scientific problems; (3) derivation of promising research directions uncovered as vacant spaces in the categorization.
- A. Abdul-Aziz, J. Lekki, M. R. Woike, and G. Y. Baakilini. 2010. Health monitoring of a rotating disk using a combined analytical-experimental approach. In Materials Evolution, vol. 68, no. 3A reference.Google Scholar
- H. Aguiar and R. Filho. 2001. Neural network and hybrid model: A discussion about different modeling techniques to predict pulping degree with industrial data. In Chemical Engineering Science - CHEM ENG SCI,vol. 56, pp. 565–570.Google ScholarCross Ref
- Amarasinghe, Kasun & Kenney, Kevin & Manic, Milos. 2018. Toward Explainable Deep Neural Network Based Anomaly Detection. 10.1109/HSI.2018.8430788.Google Scholar
- D. An, N. H. Kim, and J.-H. Choi. 2015. Practical options for selecting data driven or physics-based prognostics algorithms with reviews. In Reliability Engineering & System Safety, vol. 133, pp. 223–236.Google ScholarCross Ref
- H. Braake, H. Can, and H. Verbruggen. 1996. Semi-physical modeling of chemical processes with neural networks. In IFAC Proceedings Volumes, vol. 29, pp. 6119–6124.Google Scholar
- S. L. Brunton, J. L. Proctor, and J. N. Kutz. 2016. Discovering governing equations from data by sparse identification of nonlinear dynamical systems. In Proceedings of the National Academy of Sciences, vol. 113,pp. 3932–3937.Google ScholarCross Ref
- Byington, Carl & Mackos, N. & Argenna, G. & Palladino, A. & Reimann, J. & Schmitigal, J.. 2012. Application of symbolic regression to electrochemical impedance spectroscopy data for lubricating oil health evaluation. Proceedings of the Annual Conference of the Prognostics and Health Management Society 2012, PHM 2012. 111-121.Google Scholar
- Cai, Weili & Zhang, Wenjuan & Hu, Xiaofeng & Liu, Yingchao. 2020. A hybrid information model based on long short-term memory network for tool condition monitoring. Journal of Intelligent Manufacturing. 31. 10.1007/s10845-019-01526-4.Google Scholar
- M. Cao, K. W. Wang, Y. Fujii, and W. E. Tobler. 2004. A hybrid neural network approach for the development of friction component dynamic model. In Journal of Dynamic Systems, Measurement, and Control, vol. 126, no. 1, p. 144.Google ScholarCross Ref
- D. Castelvecchi. 2016. Can we open the black box of AI?. In Nature, vol. 538,pp. 20–23.Google ScholarCross Ref
- F. Doshi-Velez and B. Kim. 2017. “Towards a rigorous science of interpretable machine learning,” arXiv.Google Scholar
- M. Fast, M. Assadi, and S. De. 2008. Condition based maintenance of gas turbines using simulation data and artificial neural network: A demonstration of feasibility.Google Scholar
- V. Galvanauskas, R. Simutis, and A. Luebbert. 2004. Hybrid process models for process optimisation, monitoring and control. In Bioprocess and Biosystems Engineering, vol. 26, pp. 393–400.Google ScholarCross Ref
- Goebel, Kai & Eklund, N. & Bonanni, P. 2006 Fusing competing prediction algorithms for prognostics. IEEE Aerospace Conference Proceedings. 2006. 10 pp.. 10.1109/AERO.2006.1656116.Google Scholar
- C. Gugg, M. Harker, P. O'Leary, and G. Rath. 2014. An algebraic framework for the real-time solution of inverse problems on embedded systems. 2014.Google Scholar
- Guo, Yuming & Wang, Fuli & Jia, Mingxing & Zhang, Shuning. 2017. Parallel hybrid model for mechanical draft counter flow wet-cooling tower. Applied Thermal Engineering. 125. 10.1016/j.applthermaleng.2017.07.138.Google Scholar
- M. Harker and P. O'Leary. 2014. Sylvester equations and the numerical solution of partial fractional differential equations. In Journal of Computational Physics, vol. 293.Google Scholar
- D. Hendrycks and K. Gimpel. 2016. A baseline for detecting misclassified and out-of-distribution examples in neural networks. In CoRR,vol. abs/1610.02136.Google Scholar
- P. A. Higgs, R. Parkin, M. Jackson, A. Al-Habaibeh, F. Zorriassatine, and J. Coy. 2004. A survey on condition monitoring systems in industry. In Volume 3, ASME.Google Scholar
- T.-M. Hwang, H. Oh, Y.-J. Choi, S.-H. Nam, S. Lee, and Y.-K. Choung. 2009. Development of a statistical and mathematical hybrid model to predict membrane fouling and performance. In Desalination, vol. 247, pp. 210–221.Google ScholarCross Ref
- H. Ibeid, S. Meng, O. Dobon, L. Olson, and W. Gropp. 2018. Learning with analytical models.Google Scholar
- Isermann, Rolf. 2005. Model-based fault-detection and diagnosis - Status and applications. Annual Reviews in Control. 29. 71-85.Google ScholarCross Ref
- A. Jagtap and G. Karniadakis. 2019. Adaptive activation functions accelerate convergence in deep and physics-informed neural networks.Google Scholar
- A. Jardine, D. Lin, and D. Banjevic. 2006. A review on machinery diagnostics and prognostics implementing condition-based maintenance. In Mechanical Systems and Signal Processing, vol. 20, pp. 1483–1510.Google ScholarCross Ref
- K. Javed, R. Gouriveau, and N. Zerhouni. 2017. State of the art and taxonomy of prognostics approaches, trends of prognostics applications and open issues towards maturity at different technology readiness levels. In Mechanical Systems and Signal Processing, vol. 94, pp. 214–236.Google ScholarCross Ref
- H. Jeong, S. Park, S. Woo, and S. Lee. 2016. Rotating machinery diagnostics using deep learning on orbit plot images. In Procedia Manufacturing, vol. 5, pp. 1107–1118.Google ScholarCross Ref
- A. Karpatne, G. Atluri, J. H. Faghmous, M. Steinbach, A. Banerjee,A. Ganguly, S. Shekhar, N. Samatova, and V. Kumar. 2017. Theory-guided data science: A new paradigm for scientific discovery from data. In IEEE Transactions on Knowledge and Data Engineering, vol. 29, pp. 2318–2331.Google ScholarDigital Library
- A. Karpatne, W. Watkins, J. Read, and V. Kumar. 2017. Physics-guided neural networks (pgnn): An application in lake temperature modelling.Google Scholar
- Klein, Patrick & Bergmann, Ralph. 2019. Generation of Complex Data for AI-based Predictive Maintenance Research with a Physical Factory Model. 40-50. 10.5220/0007830700400050.Google Scholar
- Klimasauskas, C.C. 1998.. Hybrid modeling for robust nonlinear multivariable control. In ISA Transactions 37, 291 – 297.Google ScholarCross Ref
- Kollment, Werner & O'Leary, Paul & Harker, Matthew & Osberger, Uwe & Eck, Sven. 2016. Towards condition monitoring of railway points: Instrumentation, measurement and signal processing. 1-6. 10.1109/I2MTC.2016.7520434.Google Scholar
- Kollment, Werner & O'Leary, Paul & Ritt, Roland & Klünsner, Thomas. 2017. Force based tool wear detection using Shannon entropy and phase plane. 1-6. 10.1109/I2MTC.2017.7969765.Google Scholar
- U. Leturiondo, M. Mishra, D. Galar, and O. Salgado. 2015. Synthetic data generation in hybrid modelling of rolling element bearings. In Insight -Non-Destructive Testing and Condition Monitoring, vol. 57, pp. 395–400.Google ScholarCross Ref
- X. Li, S. Dong, and K. Patri. 2000. Hybrid learning for tool wear monitoring. In International Journal of Advanced Manufacturing Technology, vol. 16,pp. 303–307.Google ScholarCross Ref
- L. Liao and F. Köttig. 2016. A hybrid framework combining data-driven and model-based methods for system remaining useful life prediction. In Applied Soft Computing, vol. 44.Google Scholar
- H. Maderbacher, T. Christiner, H.-P. Gänser, M. Riedler, M. Stoschka, and W. Eichlseder. 2010. Lifetime evaluation of hot forged aerospace components by linking microstructural evolution and fatigue behaviour. In Procedia Engineering, vol. 2, pp. 2269–2276.Google ScholarCross Ref
- R. Maheswari and R. Umamaheswari. 2017. Trends in non-stationary signal processing techniques applied to vibration analysis of wind turbine drive train – a contemporary survey. In Mechanical Systems and Signal Processing, vol. 85, pp. 296–311.Google ScholarCross Ref
- A. J Majda and J. Harlim. 2012. Physics constrained nonlinear regression models for time series. In Nonlinearity, vol. 26, p. 201.Google ScholarCross Ref
- G. Martius and C. H. Lampert. 2016. Extrapolation and learning equations. In CoRR, vol. abs/1610.02995.Google Scholar
- M. Mishra, U. Leturiondo, O. Salgado, and D. Galar. 2015. Hybrid modelling for failure diagnosis and prognosis in the transport sector. acquired data and synthetic data. In Dyna (Bilbao), vol. 90, pp. 139–145.Google ScholarCross Ref
- C. Molnar. 2018. Interpretable Machine Learning. https://christophm.github.io/interpretable-ml-book/.Google Scholar
- D. Ninevski and P. O'Leary. 2019. Detection of derivative discontinuities in observational data.Google Scholar
- Onchis, Darian. 2020. Should I trust a deep learning condition monitoring prediction?. 10.13140/RG.2.2.12675.53287.Google Scholar
- E. Parish and K. Duraisamy. 2015. A paradigm for data-driven predictive modeling using field inversion and machine learning. In Journal of Computational Physics.Google Scholar
- D. C. Psichogios and L. H. Ungar. 1992. “A hybrid neural network-first principles approach to process modelling.Google Scholar
- C. Rackauckas, Y. Ma, J. Martensen, C. Warner, K. Zubov, R. Supekar,D. Skinner, and A. Ramadhan. 2020. Universal differential equations for scientific machine learning.Google Scholar
- Rai, Rahul & Sahu, Chandan. 2020. Driven by Data or Derived through Physics? A Review of Hybrid Physics Guided Machine Learning Techniques with Cyber-Physical System (CPS) Focus. IEEE Access. PP. 1-1. 10.1109/ACCESS.2020.2987324.Google Scholar
- M. Raissi, P. Perdikaris, and G. E. Karniadakis. 2017. Physics informed deep learning (part i): Data-driven solutions of nonlinear partial differential equations.Google Scholar
- M. Raissi, P. Perdikaris, and G. E. Karniadakis. 2017. Physics informed deep learning (part ii): Data-driven discovery of nonlinear partial differential equations.Google Scholar
- R. Ritt, P. O'Leary, C. J. Rothschedl, and M. Harker. 2018. Advanced symbolic time series analysis in cyber physical systems. In CoRR,vol. abs/1802.00617.Google Scholar
- von Rueden, Laura & Mayer, Sebastian & Garcke, Jochen & Bauckhage, Christian & Schücker, Jannis. 2019. Informed Machine Learning - Towards a Taxonomy of Explicit Integration of Knowledge into Machine Learning.Google Scholar
- M. Samadani, C. Kwuimy, and C. Nataraj. 2014. Model-based fault diagnostics of nonlinear systems using the features of the phase space response. In Communications in Nonlinear Science and Numerical Simulation, vol. 20.Google Scholar
- F. Schrodt, J. Kattge, H. Shan, F. Fazayeli, J. Joswig, A. Banerjee,M. Reichstein, G. Bönisch, S. Diaz, J. Dickie, A. Gillison, A. Karpatne,S. Lavorel, P. Leadley, C. B. Wirth, I. J. Wright, S. J. Wright, and P. B.Reich. 2015. BHPMF - a hierarchical bayesian approach to gap-filling and trait prediction for macro ecology and functional biogeography. In Global Ecology and Biogeography, vol. 24, pp. 1510–1521.Google ScholarCross Ref
- M. Seera, C. P. Lim, S. Nahavandi, and C. K. Loo. 2014. Condition monitoring of induction motors: A review and an application of an ensemble of hybrid intelligent models. In Expert Systems With Applications, vol. 41,no. 10, pp. 4891–4903.Google ScholarCross Ref
- A. P. Singh, R. Matai, A. Mishra, K. Duraisamy, and P. Durbin. 2017. Data driven augmentation of turbulence models for adverse pressure gradient flows.Google Scholar
- D. Solle, B. Hitzmann, C. Herwig, M. P. Remelhe, S. Ulonska,L. Wuerth, A. Prata, and T. Steckenreiter. 2017. Between the poles of data-driven and mechanistic modeling for process operation. In Chemie Ingenieur Technik, vol. 89, pp. 542–561.Google ScholarCross Ref
- R. Stewart and S. Ermon. 2016. Label-free supervision of neural networks with physics and domain knowledge.Google Scholar
- M. von Stosch, R. Oliveira, J. Peres, and S. F. de Azevedo. 2014. Hybrid semi-parametric modeling in process systems engineering: Past, present and future. In Computers & Chemical Engineering, vol. 60, pp. 86–101.Google ScholarCross Ref
- J. Sun, G.S. Hong, Y.S. Wong, M. Rahman, Z.G. Wang. 2006. Effective training data selection in tool condition monitoring system, International Journal of Machine Tools and Manufacture, Volume 46, Issue 2, Pages 218-224.Google ScholarCross Ref
- W. Sun, S. Shao, R. Zhao, R. Yan, X. Zhang, and X. Chen. 2016. A sparse auto-encoder-based deep neural network approach for induction motor faults classification. In Measurement, vol. 89, pp. 171–178.Google ScholarCross Ref
- B. D. Tracey, K. Duraisamy, and J. J. Alonso. 2015. A machine learning strategy to assist turbulence model development. In 53rd AIAA Aerospace Sciences Meeting, American Institute of Aeronautics and Astronautics.Google Scholar
- M. Twa, S. Parthasarathy, C. Roberts, A. M Mahmoud, T. Raasch, and M. Bullimore. 2006. Automated decision tree classification of corneal shape. In Optometry and vision science: official publication of the American Academy of Optometry, vol. 82, pp. 1038–46.Google Scholar
- S.-M. Udrescu and M. Tegmark. 2019. Ai feynman: a physics-inspired method for symbolic regression.Google Scholar
- Vargas, A., Iqbal, M.R., Erickson, J.S., & Bennett, K.P. 2018. Grey-box Process Control Mining for Anomaly Monitoring and Deconstruction. ArXiv, abs/1808.03709.Google Scholar
- J.-X. Wang, J. Wu, J. Ling, G. Iaccarino, and H. Xiao. 2016. Physics-informed machine learning for predictive turbulence modeling: Toward a complete framework.Google Scholar
- K. Worden, C. Wong, U. Parlitz, A. Hornstein, D. Engster,T. Tjahjowidodo, F. Al-Bender, D. Rizos, and S. Fassois. 2007. Identification of pre-sliding and sliding friction dynamics: Grey box and black-box models. In Mechanical Systems and Signal Processing, vol. 21, pp. 514–534.Google ScholarCross Ref
- T. Wuest, D. Weimer, C. Irgens, and K.-D. Thoben. 2016. Machine learning in manufacturing: advantages, challenges, and applications. In Production& Manufacturing Research, vol. 4, pp. 23–45.Google ScholarCross Ref
- Yang, Zhuo & Eddy, Douglas & Krishnamurty, Sundar & Grosse, Ian & Denno, Peter & Lu, Yan & Witherell, Paul. 2017. Investigating Grey-Box Modeling for Predictive Analytics in Smart Manufacturing. V02BT03A024. 10.1115/DETC2017-67794.Google Scholar
- Yao, Y. & Lu, Yong & Yuan, Zhe & Hu, J. 2004. A Hybrid Model for Tool Condition Monitoring and Optimal Tool Management. Materials Science Forum - MATER SCI FORUM. 471-472. 865-870. 10.4028/www.scientific.net/MSF.471-472.Google Scholar
- F. Q. Yuan. 2016. Critical issues of applying machine learning to condition monitoring for failure diagnosis. In 2016 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM), IEEE.Google ScholarCross Ref
- M. Yuan, Y. Wu, and L. Lin. 2016. Fault diagnosis and remaining useful life estimation of aero engine using LSTM neural network. In 2016 IEEE International Conference on Aircraft Utility Systems (AUS), IEEE.Google Scholar
Index Terms
- Knowledge Incorporation for Machine Learning in Condition Monitoring: A Survey
Recommendations
Integrating machine learning with knowledge acquisition through direct interaction with domain experts
Knowledge elicitation from experts and empirical machine learning are two distinct approaches to knowledge acquisition with differing and mutually complementary capabilities. Learning apprentices have provided environments in which a knowledge engineer ...
An Experimental Evaluation of Integrating Machine Learning with Knowledge Acquisition
Machine learning and knowledge acquisition from experts have distinct capabilities that appear to complement one another. We report a study that demonstrates the integration of these approaches can both improve the accuracy of the developed knowledge base ...
Condition Monitoring using Machine Learning: A Review of Theory, Applications, and Recent Advances
AbstractIn modern industry, the quality of maintenance directly influences equipment’s operational uptime and efficiency. Hence, based on monitoring the condition of the machinery, predictive maintenance can minimize machine downtime and ...
Comments