Process prediction and detection of faults using probabilistic bidirectional recurrent neural networks on real plant data
Abstract
Abstract Attaining Industry 4.0 for manufacturing operations requires advanced monitoring systems and real‐time data analytics of plant data, among other topics. We propose a probabilistic bidirectional recurrent network (PBRN) for industrial process monitoring for the early detection of faults. The model is based on a gated recurrent unit (GRU) neural network that allows the model to retain long‐term dependencies between sensor data along a time horizon, hence learning the dynamic behavior of the process. To reduce the false‐positive detection rate of the model, we compel the model to learn from a highly noisy sensor reading while outputting noise‐free sensor outputs. The performance of the proposed model is compared with other data‐driven statistical process monitoring schemes using real plant data from an industrial air separations unit (ASU) containing noisy sensor readings. We show that the model can learn from noisy data without reducing its performance. Using two different fault cases, we demonstrate the model's ability to carry out early fault detection with average false‐positive rates of 2.9% and 4.9% for both fault cases. The missed detection rates are 0.1% and 0.2%, respectively.
- Authors:
-
[1];
- Rensselaer Polytechnic Inst., Troy, NY (United States)
- Linde, Inc., Tonawanda, NY (United States). Smart Operations, Center of Excellence (COE)
- Publication Date:
- Research Org.:
- Univ. of California, Los Angeles, CA (United States)
- Sponsoring Org.:
- USDOE Office of Energy Efficiency and Renewable Energy (EERE); CESMII
- OSTI Identifier:
- 1868938
- Alternate Identifier(s):
- OSTI ID: 1893184
- Grant/Contract Number:
- EE0007613; 4550-G-WA324
- Resource Type:
- Accepted Manuscript
- Journal Name:
- Journal of Advanced Manufacturing and Processing
- Additional Journal Information:
- Journal Volume: 4; Journal Issue: 4; Journal ID: ISSN 2637-403X
- Publisher:
- American Institute of Chemical Engineers (AIChE), Wiley
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 42 ENGINEERING; 32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATION; Industry 4.0; Probabilistic bidirectional recurrent network; Gated recurrent unit; Air separation unit
Citation Formats
Yerimah, Lucky E., Ghosh, Sambit, Wang, Yajun, Cao, Yanan, Flores‐Cerrillo, Jesus, and Bequette, B. Wayne. Process prediction and detection of faults using probabilistic bidirectional recurrent neural networks on real plant data. United States: N. p., 2022.
Web. doi:10.1002/amp2.10124.
Yerimah, Lucky E., Ghosh, Sambit, Wang, Yajun, Cao, Yanan, Flores‐Cerrillo, Jesus, & Bequette, B. Wayne. Process prediction and detection of faults using probabilistic bidirectional recurrent neural networks on real plant data. United States. https://doi.org/10.1002/amp2.10124
Yerimah, Lucky E., Ghosh, Sambit, Wang, Yajun, Cao, Yanan, Flores‐Cerrillo, Jesus, and Bequette, B. Wayne. Mon .
"Process prediction and detection of faults using probabilistic bidirectional recurrent neural networks on real plant data". United States. https://doi.org/10.1002/amp2.10124. https://www.osti.gov/servlets/purl/1868938.
@article{osti_1868938,
title = {Process prediction and detection of faults using probabilistic bidirectional recurrent neural networks on real plant data},
author = {Yerimah, Lucky E. and Ghosh, Sambit and Wang, Yajun and Cao, Yanan and Flores‐Cerrillo, Jesus and Bequette, B. Wayne},
abstractNote = {Abstract Attaining Industry 4.0 for manufacturing operations requires advanced monitoring systems and real‐time data analytics of plant data, among other topics. We propose a probabilistic bidirectional recurrent network (PBRN) for industrial process monitoring for the early detection of faults. The model is based on a gated recurrent unit (GRU) neural network that allows the model to retain long‐term dependencies between sensor data along a time horizon, hence learning the dynamic behavior of the process. To reduce the false‐positive detection rate of the model, we compel the model to learn from a highly noisy sensor reading while outputting noise‐free sensor outputs. The performance of the proposed model is compared with other data‐driven statistical process monitoring schemes using real plant data from an industrial air separations unit (ASU) containing noisy sensor readings. We show that the model can learn from noisy data without reducing its performance. Using two different fault cases, we demonstrate the model's ability to carry out early fault detection with average false‐positive rates of 2.9% and 4.9% for both fault cases. The missed detection rates are 0.1% and 0.2%, respectively.},
doi = {10.1002/amp2.10124},
journal = {Journal of Advanced Manufacturing and Processing},
number = 4,
volume = 4,
place = {United States},
year = {2022},
month = {5}
}
Free Publicly Available Full Text
Publisher's Version of Record
Other availability
Search WorldCat to find libraries that may hold this journal
Save to My Library
You must Sign In or Create an Account in order to save documents to your library.
Works referenced in this record:
Bidirectional recurrent neural networks
journal, January 1997
- Schuster, M.; Paliwal, K. K.
- IEEE Transactions on Signal Processing, Vol. 45, Issue 11
Fault diagnosis of rolling bearings with recurrent neural network-based autoencoders
journal, June 2018
- Liu, Han; Zhou, Jianzhong; Zheng, Yang
- ISA Transactions, Vol. 77
Product property and production rate control of styrene polymerization
journal, April 2002
- Prasad, Vinay; Schley, Matthias; Russo, Louis P.
- Journal of Process Control, Vol. 12, Issue 3
Hybrid Modeling in the Era of Smart Manufacturing
journal, September 2020
- Yang, Shu; Navarathna, Pranesh; Ghosh, Sambit
- Computers & Chemical Engineering, Vol. 140
A novel dynamic PCA algorithm for dynamic data modeling and process monitoring
journal, July 2018
- Dong, Yining; Qin, S. Joe
- Journal of Process Control, Vol. 67
Dynamic modeling and collocation-based model reduction of cryogenic air separation units
journal, January 2016
- Cao, Yanan; Swartz, Christopher L. E.; Flores-Cerrillo, Jesus
- AIChE Journal, Vol. 62, Issue 5
Dynamic-Inner Partial Least Squares for Dynamic Data Modeling
journal, January 2015
- Dong, Yining; Qin, S. Joe
- IFAC-PapersOnLine, Vol. 48, Issue 8
Nonlinear process monitoring using kernel principal component analysis
journal, January 2004
- Lee, Jong-Min; Yoo, ChangKyoo; Choi, Sang Wook
- Chemical Engineering Science, Vol. 59, Issue 1
A combined canonical variate analysis and Fisher discriminant analysis (CVA–FDA) approach for fault diagnosis
journal, June 2015
- Jiang, Benben; Zhu, Xiaoxiang; Huang, Dexian
- Computers & Chemical Engineering, Vol. 77
Recent Advances in Variational Autoencoders With Representation Learning for Biomedical Informatics: A Survey
journal, January 2021
- Wei, Ruoqi; Mahmood, Ausif
- IEEE Access, Vol. 9
Statistical process monitoring of a multiphase flow facility
journal, September 2015
- Ruiz-Cárcel, C.; Cao, Y.; Mba, D.
- Control Engineering Practice, Vol. 42
Quantum computing assisted deep learning for fault detection and diagnosis in industrial process systems
journal, December 2020
- Ajagekar, Akshay; You, Fengqi
- Computers & Chemical Engineering, Vol. 143
Interpretation and compensation of backlash error data in machine centers for intelligent predictive maintenance using ANNs
journal, May 2015
- Wang, Ke-Sheng; Li, Zhe; Braaten, Jørgen
- Advances in Manufacturing, Vol. 3, Issue 2
Smart Manufacturing
journal, July 2015
- Davis, Jim; Edgar, Thomas; Graybill, Robert
- Annual Review of Chemical and Biomolecular Engineering, Vol. 6, Issue 1
Fault detection of process correlation structure using canonical variate analysis-based correlation features
journal, October 2017
- Jiang, Benben; Braatz, Richard D.
- Journal of Process Control, Vol. 58
Categorization of Anomalies in Smart Manufacturing Systems to Support the Selection of Detection Mechanisms
journal, October 2017
- Lopez, Felipe; Saez, Miguel; Shao, Yuru
- IEEE Robotics and Automation Letters, Vol. 2, Issue 4
Smart manufacturing and energy systems
journal, June 2018
- Edgar, Thomas F.; Pistikopoulos, Efstratios N.
- Computers & Chemical Engineering, Vol. 114
A novel process monitoring approach based on variational recurrent autoencoder
journal, October 2019
- Cheng, Feifan; He, Q. Peter; Zhao, Jinsong
- Computers & Chemical Engineering, Vol. 129
Information concentrated variational auto-encoder for quality-related nonlinear process monitoring
journal, October 2020
- Zhu, Jiazhen; Shi, Hongbo; Song, Bing
- Journal of Process Control, Vol. 94
Long Short-Term Memory
journal, November 1997
- Hochreiter, Sepp; Schmidhuber, Jürgen
- Neural Computation, Vol. 9, Issue 8
Process monitoring using causal graphical models, with application to clogging detection in steel continuous casting
journal, September 2021
- Yang, Shu; Rebmann, Andreas; Tang, Ming
- Journal of Process Control, Vol. 105
Geometric properties of partial least squares for process monitoring
journal, January 2010
- Li, Gang; Qin, S. Joe; Zhou, Donghua
- Automatica, Vol. 46, Issue 1
Statistical process monitoring: basics and beyond
journal, January 2003
- Joe Qin, S.
- Journal of Chemometrics, Vol. 17, Issue 8-9
Backpropagation through time: what it does and how to do it
journal, January 1990
- Werbos, P. J.
- Proceedings of the IEEE, Vol. 78, Issue 10
Deep convolutional neural network model based chemical process fault diagnosis
journal, July 2018
- Wu, Hao; Zhao, Jinsong
- Computers & Chemical Engineering, Vol. 115