Abstract
Multilayered auto-associative neural architectures have widely been used in empirical sensor modeling. Typically, such empirical sensor models are used in sensor calibration and fault monitoring systems. However, simultaneous optimization of related performance metrics, i.e., auto-sensitivity, cross-sensitivity, and fault-detectability, is not a trivial task. Learning procedures for parametric and other relevant non-parametric empirical models are sensitive to optimization and regularization methods. Therefore, there is a need for active learning strategies that can better exploit the underlying statistical structure among input sensors and are simple to regularize and fine-tune. To this end, we investigated the greedy layer-wise learning strategy and denoising-based regularization procedure for sensor model optimization. We further explored the effects of denoising-based regularization hyper-parameters such as noise-type and noise-level on sensor model performance and suggested optimal settings through rigorous experimentation. A visualization procedure was introduced to obtain insight into the internal semantics of the learned model. These visualizations allowed us to suggest an implicit noise-generating process for efficient regularization in higher-order layers. We found that the greedy-learning procedure improved the overall robustness of the sensor model. To keep experimentation unbiased and immune to noise-related artifacts in real sensors, the sensor data were sampled from simulators of a nuclear steam supply system of a pressurized water reactor and a Tennessee Eastman chemical process. Finally, we compared the performance of an optimally regularized sensor model with auto-associative neural network, auto-associative kernel regression, and fuzzy similarity-based sensor models.
Similar content being viewed by others
Abbreviations
- AANN:
-
Auto-associative neural networks
- AAKR:
-
Auto-associative kernel regression
- APWR:
-
Advanced pressurized water reactor
- DAE:
-
Denoising auto-encoder
- FS:
-
Fuzzy similarity
- MSET:
-
Multivariate state estimation technique
- NLPS:
-
Non-linear partial least square
- NSSS:
-
Nuclear steam supply system
- PWR:
-
Pressurized water reactor
- PCA:
-
Principal component analysis
- RBM:
-
Restricted boltzman machine
- TE:
-
Tennessee Eastman
- \(S_{Auto}\) :
-
Auto-sensitivity
- \(S_{Cross}\) :
-
Cross-sensitivity
- R :
-
Robustness
- SO :
-
Spillover
- S :
-
Original non-corrupted sensor readings
- \(\tilde{S}\) :
-
Corrupted sensor readings
- \(\hat{S}\) :
-
Model-predicted sensor readings
References
Baraldi P, Canesi R, Zio E et al (2011) Genetic algorithm-based wrapper approach for grouping condition monitoring signals of nuclear power plant components. Integr Comput Aided Eng 18:221–234. doi:10.3233/ICA-2011-0375
Baraldi P, Di Maio F, Genini D, Zio E (2015) Reconstruction of missing data in multidimensional time series by fuzzy similarity. Appl Soft Comput J 26:1–9. doi:10.1016/j.asoc.2014.09.038
Bengio Y (2009) Learning Deep Architectures for AI. Found Trends\(\textregistered \) Mach Learn 2:1–127. doi:10.1561/2200000006
Bengio Y, Simard P, Frasconi P (1994) Learning long term dependencies with gradient descent is difficult. IEEE Trans Neural Networks 5:157–166. doi:10.1109/72.279181
Bengio Y, Lamblin P (2007) Greedy layer-wise training of deep networks. Adv Neural Inf Process Syst (NIPS) 19:153–160
Coble JB, Meyer RM, Ramuhalli P, et al (2012) A Review of Sensor Calibration Monitoring for Calibration Interval Extension in Nuclear Power Plants. Technical report PNNL-21687, Pacific Northwest Natl Lab Richland, Wash, USA. doi:10.2172/1061413
Davis E, Rubin L, Hussey A (2006) On-line monitoring cost-benefit guide. Final report 1006777,EPRI, Palo Alto,Calif,USA
Di Maio F, Baraldi P, Zio E, Seraoui R (2013) Fault detection in nuclear power plants components by a combination of statistical methods. Reliab IEEE Trans 62:833–845
Downs JJ, Vogel EF (1993) A plant-wide industrial process control problem. Comput Chem Eng 17:245–255. doi:10.1016/0098-1354(93)80018-I
Elnokity O, Mahmoud II, Refai MK, Farahat HM (2012) ANN based sensor faults detection, isolation, and reading estimates—SFDIRE: applied in a nuclear process. Ann Nucl Energy 49:131–142. doi:10.1016/j.anucene.2012.06.003
Erhan D, Courville A, Vincent P (2010) Why does unsupervised pre-training help deep learning? J Mach Learn Res 11:625–660. doi:10.1145/1756006.1756025
Erhan D, Bengio Y, Courville A, Vincent P (2009) Visualizing higher-layer features of a deep network. Departement d’Informatique Rech Operationnelle, Tech Rep 1341:1–13
Fantoni PF, Hoffmann MI, Shankar R, Davis EL (2003) On-line monitoring of instrument channel performance in nuclear power plant using PEANO. Prog Nucl Energy 43:83–89. doi:10.1016/S0149-1970(03)00017-9
Fantoni PF (2005) Experiences and applications of PEANO for monitoring in power plants. Prog Nucl Energy 46:206–225. doi:10.1016/j.pnucene.2005.03.005
Fantoni PF, Mazzola A (1996) A pattern recognition-artificial neural networks based model for signal validation in nuclear power plants. Ann Nucl Energy 23:1069–1076. doi:10.1016/0306-4549(96)84661-5
Garvey J, Garvey D, Seibert R, Hines JW (2007) Validation of on-line monitoring techniques to nuclear plant data. Nucl Eng Technol 39:133–142. doi:10.5516/NET.2007.39.2.133
Glorot X, Bengio Y (2010) Understanding the difficulty of training deep feedforward neural networks. In: Proceedings of 13th international conference artificial intelligence, vol 9. Aistats, pp 249–256
Gribok AV, Hines JW, Urmanov A, Uhrig RE (2000) Regularization of ill-posed surveillance and diagnostic measurements. Power plant surveillance and diagnostics. Appl Res Artif Intell, Springer
Gribok AV, Hines JW, Urmanov A, Uhrig RE (2002) Heuristic, systematic, and informational regularization for process monitoring. Int J Intell Syst 17:723–749. doi:10.1002/int.10047
Gross KC, Singer RM, Wegerich SW, Herzog JP (1997) Application of a model-based fault detection system to nuclear plant signals. In: Proceedings of the intelligent system applications to power systems, ISAP, Seoul, Korea, pp 66–70
Gross KC, Singer RM, Wegerich S, Mott J (1998) Multivariate state estimation technique (MSET) based surveillance system. In: U.S. Patent No. 5764509
Hashemian HM (2011) On-line monitoring applications in nuclear power plants. Prog Nucl Energy 53:167–181. doi:10.1016/j.pnucene.2010.08.003
Heo G-Y (2008) Condition monitoring using empirical models: technical review and prospects for nuclear applications. Nucl Eng Technol 40:49–68. doi:10.5516/NET.2008.40.1.049
Hines JW, Gribok A, Attieh I, Uhrig R (2000) Regularization Methods for Inferential Sensing in Nuclear Power Plants. In: Ruan D (ed) Fuzzy systems and soft computing in nuclear engineering SE - 13. Physica-Verlag HD, pp 285–314
Hines JW, Usynin A (2005) Autoassociative model input variable selection for process monitoring. In: International symposium on the future I&C for nuclear power plants
Hines JW, Uhrig RE, Wrest DJ (1998) Use of autoassociative neural networks for signal validation. J Intell Robot Syst 21:143–154
Hines JW, Garvey DR (2006) Development and application of fault detectability performance metrics for instrument calibration verification and anomaly detection. Pattern Recognit 1:2–15
Hochreiter S, Bengio Y, Frasconi P, Schmidhuber J (2001) Gradient flow in recurrent nets: the difficulty of learning long-term dependencies. A F Guid to Dyn Recurr Networks 237–243: doi:10.1109/9780470544037.ch14
IAEA (2008) On-line monitoring for improving performance of nuclear power plants part 1: instrument channel monitoring. Technical report nuclear energy series NP-T-1.1, International Atomic Energy Agency, Vienna, Austria
Karklin Y, Simoncelli EP (2011) Efficient coding of natural images with a population of noisy linear–nonlinear neurons. Adv Neural Inf Process Syst 24:999–1007
Kim KD, Lee SW, Hwang M et al (2007) Development of a Visual System Analyzer based on reactor system analysis codes. Prog Nucl Energy 49:452–462. doi:10.1016/j.pnucene.2007.07.005
Kramer MA (1991) Nonlinear principal component analysis using autoassociative neural networks. AIChE J 37:233–243
Kramer MA (1992) Autoassociative neural networks. Comput Chem Eng 16:313–328. doi:10.1016/0098-1354(92)80051-A
Larochelle H, Larochelle H, Bengio Y et al (2009) Exploring strategies for training deep neural networks. J Mach Learn Res 10:1–40
Liu E, Zhang D (2003) Diagnosis of component failures in the space shuttlemain engines using Bayesian belief network: a feasibility study. Int J Artif Intell Tools 12:355–374. doi:10.1142/S0218213003001277
Ma J, Jiang J (2011) Applications of fault detection and diagnosis methods in nuclear power plants: a review. Prog Nucl Energy 53:255–266. doi:10.1016/j.pnucene.2010.12.001
Marseguerra M, Zoia A (2005) The autoassociative neural network in signal analysis: II. Application to on-line monitoring of a simulated BWR component. Ann Nucl Energy 32:1207–1223. doi:10.1016/j.anucene.2005.03.005
Marseguerra M, Zoia A (2005) The autoassociative neural network in signal analysis: I. The data dimensionality reduction and its geometric interpretation. Ann Nucl Energy 32:1191–1206. doi:10.1016/j.anucene.2005.03.006
Penha R, Hines J (2001) Using principal component analysis modeling to monitor temperature sensors in a nuclear research reactor. In: Maintenance and reliability conference (MARCON 2001)
Petrović I, Baotić M, Perić N (2000) Regularization and validation of neural network models of nonlinear systems. e i. Elektrotech Informationstech 117:24–31. doi:10.1007/BF03161395
Rasmussen B, Hines JW, Uhrig RE (2000) Nonlinear partial least squares modeling for instrument surveillance and calibration verification. In: proceedings of the maintenance and reliability conference (MARCON 2000), Knoxville, TN
Reyes J, Vellasco M, Tanscheit R (2014) Fault detection and measurements correction for multiple sensors using a modified autoassociative neural network. Neural Comput Appl 24:1929–1941. doi:10.1007/s00521-013-1429-4
Rifai S, Muller X (2011) Contractive auto-encoders: explicit invariance during feature extraction. Icml 85:833–840
Şeker S, Ayaz E, Türkcan E (2003) Elman’s recurrent neural network applications to condition monitoring in nuclear power plant and rotating machinery. Eng Appl Artif Intell 16:647–656. doi:10.1016/j.engappai.2003.10.004
Shaheryar A, Yin X-C, Hao H-W et al (2016) A denoising based autoassociative model for robust sensor monitoring in nuclear power plants. Sci Technol Nucl Install 2016:1–17. doi:10.1155/2016/9746948
Urmanov AM, Gribok A V, Hines JW, Uhrig RE (2000) Complexity-penalized model selection for feedwater inferential measurements in nuclear power plants. In: International topical meeting on nuclear plant instrumentation, controls, and human-machine interface technologies (NPIC&HMIT 2000), Washington, DC, November, 2000
Usynin A, Hines Wesley J, Ding J (2004) On-line monitoring robustness measures and comparisons. In: International atomic energy agency technical meeting on increasing instrument calibration intervals through on-line calibration technology, OECD Halden Reactor Project, Halden, Norway
Usynin A, Hines JW (2005) MSET performance optimization through regularization. Nucl Eng Technol 37:159–166
Utgoff PE, Stracuzzi DJ (2002) Many-Layered Learning. Neural Comput 14:2497–2529. doi:10.1162/08997660260293319
Vincent P, Larochelle H, Lajoie I et al (2010) Stacked denoising autoencoders: learning useful representations in a deep network with a local denoising criterion. J Mach Learn Res 11:3371–3408. doi:10.1111/1467-8535.00290
Vincent P, Larochelle H, Bengio Y, Manzagol P-A (2008) Extracting and composing robust features with denoising autoencoders. In: Proceedings of the 25th international conference on Machine learning - ICML ’08. pp 1096–1103
Wrest DJ, Hines JW, Uhrig RE (1996) Instrument surveillance and calibration verification through plant wide monitoring using autoassociative neural networks. In: In the proceedings of the 1996 American nuclear society international topical meeting on nuclear plant instrumentation, Control and human machine interface technologies. pp 6–9
Yin S, Ding SX, Haghani A et al (2012) A comparison study of basic data-driven fault diagnosis and process monitoring methods on the benchmark Tennessee Eastman process. J Process Control 22:1567–1581. doi:10.1016/j.jprocont.2012.06.009
Zavaljevski N, Gross KC (2000) Sensor fault detection in nuclear power plants using multivariate state estimation technique and support vector machines. In: ANS international topical meeting on “advances in reactor physics and mathematics and computation into the next millennium”
Acknowledgements
The authors acknowledge the support of the China Scholarship Council for research funding. We are also thankful to the International Atomic Energy Agency’s (IAEA) simulator development program for providing the nuclear power plant simulator tools, which were primarily used in the preparation of the sensor dataset for this research.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Shaheryar, A., Yin, XC., Hao, HW. et al. Selection of optimal denoising-based regularization hyper-parameters for performance improvement in a sensor validation model. Artif Intell Rev 50, 341–382 (2018). https://doi.org/10.1007/s10462-017-9546-6
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10462-017-9546-6