Valve Stiction Detection Using Multitimescale Feature Consistent Constraint for Time-Series Data
The DOI matches its Crossref record; 33 of the 33 references it cites were found in a registry or an index.
Abstract
Using neural networks to build a reliable fault detection model is an attractive topic in industrial processes but remains challenging due to the lack of labeled data. We propose a feature learning approach for industrial time-series data based on self-supervised contrastive learning to tackle this challenge. The proposed approach consists of two components: data transformation and representation learning. The data transformation converts the raw time-series to temporal distance matrices capable of storing temporal and spatial information. The representation learning component uses a convolution-based encoder to encode the temporal distance matrices to embedding representations. The encoder is trained using a new constraint called multitimescale feature consistent constraint. Finally, a fault detection framework for the valve stiction detection task is developed based on the feature learning method. The proposed framework is evaluated not only on an industrial benchmark dataset but also on a hardware experimental system and real industrial environments.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“MTFCC [130] is another view generation method based on multi-scale characteristics, which samples time series samples at multiple scales and considers that the views from the same sample have similar representations, even if their scales are different.”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [130]
Citations
Cites
A simple framework for contrastive learning of visual representations
Chen, T., Kornblith, S., Norouzi, M., & Hinton, G. (2020) · arXiv · doi:10.48550/arxiv.2002.05709 · Cited by 10,857 · cited 1×
“As a potential technique, self-supervised learning learns representations from unlabeled data and is widely used in audio processing and computer vision domains [17].”
Method(2011) · Cited by 6,993
A new convolutional neural network-based data-driven fault diagnosis method
Wen, L., Li, X., Gao, L., & Zhang, Y. (2018) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2017.2774777 · Cited by 2,222 · cited 2×
“The data transformation is also a useful component for other fault detection tasks, such as signal-to-image conversion method [18], multiple sensors stack method [19], and 2-D compressed construction method [20].”
BackgroundMethodTS2Vec: Towards Universal Representation of Time Series
Yue, Z., Wang, Y., Duan, J., Yang, T., Huang, C., Tong, Y., & Xu, B. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i8.20881 · Cited by 850 · cited 1×
“This term is also consistent with the assumption widely used in time-series data [23], [24], [25], i.e., given a window as an anchor, the windows adjacent to it are more likely to be similar, and the windows further away should be less similar.”
ResultFault diagnosis for rotating machinery using multiple sensors and convolutional neural networks
Xia, M., Li, T., Xu, L., Liu, L., & de Silva, C. W. (2018) · IEEE/ASME Transactions on Mechatronics · doi:10.1109/tmech.2017.2728371 · Cited by 723 · cited 3×
“Moreover, CNNs are widely used in industrial fault detection tasks, such as identification of complex power quality disturbances [26], bearing fault diagnosis [18], rotating machinery diagnosis [19], and gearbox fault diagnosis [27].”
BackgroundMethodTime series classification with ensembles of elastic distance measures
Lines, J., & Bagnall, A. (2015) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-014-0361-2 · Cited by 522 · cited 2×
“Based on the experimental results in [22], the DTW distance is used as the measurement method because it is relatively simple, and no significant performance difference is observed in the above elastic measures.”
MethodChoudhury, M. A. A. S., Thornhill, N. F., & Shah, S. L. (2004) · Control Engineering Practice · doi:10.1016/j.conengprac.2004.05.005 · Cited by 363 · cited 2×
“The final output of this data-driven model is again converted back to a mA signal using a lookup table based on the valve characteristics [5], [29].”
BackgroundMethodA review on distance based time series classification
Abanda, A., Mori, U., & Lozano, J. A. (2019) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-018-0596-4 · Cited by 296 · cited 1×
“Lock-stepmeasurereferstothosedis-tancesthatcomparethe i th point of one series to the i th point of another, while elastic measure aims to create a nonlinear mapping in order to align the series and allow comparison of one-to-many points [21].”
BackgroundDiagnosis of poor control-loop performance using higher-order statistics
Choudhury, M. A. A. S., Shah, S. L., & Thornhill, N. F. (2004) · Automatica · doi:10.1016/j.automatica.2004.03.022 · Cited by 283 · cited 1×
“Accordingtothereview[1],thetraditionaldetectionmeth-odscanbebroadlyclassifiedintofourcategories,cross-correlation-function-based[4],limit-cycle-patterns-based[5], [6],nonlinearity-detection-based[7],andwaveform-shape-based[8],[9].”
Yu, J., & Zhou, X. (2020) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2020.2966326 · Cited by 235 · cited 1×
“Moreover, CNNs are widely used in industrial fault detection tasks, such as identification of complex power quality disturbances [26], bearing fault diagnosis [18], rotating machinery diagnosis [19], and gearbox fault diagnosis [27].”
MethodA simple method for detection of stiction in control valves
Horch, A. (1999) · Control Engineering Practice · doi:10.1016/s0967-0661(99)00100-8 · Cited by 226 · cited 2×
“According to the review [1], the traditional detection methods can be broadly classified into four categories, crosscorrelation-function-based [4], limit-cycle-patterns-based [5], [6], nonlinearity-detection-based [7], and waveform-shapebased [8], [9].”
MethodData-Driven Fault Diagnosis Method Based on Compressed Sensing and Improved Multiscale Network
Hu, Z.-X., Wang, Y., Ge, M.-F., & Liu, J. (2020) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2019.2912763 · Cited by 220 · cited 1×
“D compressed construction method [20].”
A Curve Fitting Method for Detecting Valve Stiction in Oscillating Control Loops
He, Q. P., Wang, J., Pottmann, M., & Qin, S. J. (2007) · Industrial & Engineering Chemistry Research · doi:10.1021/ie061219a · Cited by 170 · cited 3×
“ManyMFE-baseddetection methodshavebeendeveloped,suchasthefeaturesderivedfrom time-seriesdatacharacterization[4],[10],[11],andthefeatures derivedfromtheshapeofthecontrolleroutput(OP)andprocess variable(PV)[6],[9],[12].”
BackgroundSelf-citationPractical Model and Detection Algorithm for Valve Stiction
Kano, M., Maruta, H., Kugemoto, H., & Shimizu, K. (2004) · IFAC Proceedings Volumes · doi:10.1016/s1474-6670(17)31917-1 · Cited by 170 · cited 2×
“The final output of this data-driven model is again converted back to a mA signal using a lookup table based on the valve characteristics [5], [29].”
MethodState of AI-Based Monitoring in Smart Manufacturing and Introduction to Focused Section
Ding, H., Gao, R. X., Isaksson, A. J., Landers, R. G., Parisini, T., & Yuan, Y. (2020) · IEEE/ASME Transactions on Mechatronics · doi:10.1109/tmech.2020.3022983 · Cited by 151 · cited 1×
“In recent years, intelligent factories have received increased attention, and it is easier to collect industrial data than ever before [3].”
BackgroundQiu, W., Tang, Q., Liu, J., & Yao, W. (2020) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2019.2920689 · Cited by 142 · cited 1×
“CNNs are widely used in industrial fault detection tasks, such as identification of complex power quality disturbances [26], bearing fault diagnosis [18], rotating machinery diagnosis [19],”
BackgroundA simple method for detecting valve stiction in oscillating control loops
Singhal, A. K., & Salsbury, T. I. (2004) · Journal of Process Control · doi:10.1016/j.jprocont.2004.10.001 · Cited by 141 · cited 1×
“ManyMFE-baseddetection methodshavebeendeveloped,suchasthefeaturesderivedfrom time-seriesdatacharacterization[4],[10],[11],andthefeatures derivedfromtheshapeofthecontrolleroutput(OP)andprocess variable(PV)[6],[9],[12].”
An automatic method for detection of valve stiction in process control loops
Yamashita, Y. (2005) · Control Engineering Practice · doi:10.1016/j.conengprac.2005.03.004 · Cited by 137 · cited 2×
“ManyMFE-baseddetection methodshavebeendeveloped,suchasthefeaturesderivedfrom time-seriesdatacharacterization[4],[10],[11],andthefeatures derivedfromtheshapeofthecontrolleroutput(OP)andprocess variable(PV)[6],[9],[12].”
Detection and Diagnosis of Stiction in Control Loops
Jelali, M., Horch, A., & Scali, C. (2009) · Advances in industrial control · doi:10.1007/978-1-84882-775-2 · Cited by 126
Temporally-Weighted Hierarchical Clustering for Unsupervised Action Segmentation
Sarfraz, M. S., Murray, N., Sharma, V., Diba, A., Gool, L. V., & Stiefelhagen, R. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.01107 · Cited by 125
Rossi, M., & Scali, C. (2004) · Journal of Process Control · doi:10.1016/j.jprocont.2004.11.003 · Cited by 106 · cited 1×
“Accordingtothereview[1],thetraditionaldetectionmeth-odscanbebroadlyclassifiedintofourcategories,cross-correlation-function-based[4],limit-cycle-patterns-based[5], [6],nonlinearity-detection-based[7],andwaveform-shape-based[8],[9].”
Comprehensive methodology for detection and diagnosis of oscillatory control loops
Karra, S. R., & Karim, M. N. (2009) · Control Engineering Practice · doi:10.1016/j.conengprac.2009.02.015 · Cited by 88
Review and comparison of techniques of analysis of valve stiction: From modeling to smart diagnosis
di Capaci, R. B., & Scali, C. (2018) · Process Safety and Environmental Protection · doi:10.1016/j.cherd.2017.12.038 · Cited by 88 · cited 1×
“Accordingtothereview[1],thetraditionaldetectionmeth-odscanbebroadlyclassifiedintofourcategories,cross-correlation-function-based[4],limit-cycle-patterns-based[5], [6],nonlinearity-detection-based[7],andwaveform-shape-based[8],[9].”
Valve stiction detection through improved pattern recognition using neural networks
Mohd Amiruddin, A. A. A., Zabiri, H., Jeremiah, S. S., Teh, W. K., & Kamaruddin, B. (2019) · Control Engineering Practice · doi:10.1016/j.conengprac.2019.06.008 · Cited by 56 · cited 7×
“In the experiments, we compare the proposed MTFCC method with other seven methods, including LR, random forest (RF), support vector machine (SVM), extreme gradient boosting(XgBoost),LeNet-5,supervisedconvolutionalnetwork (supervised-Conv)[15],andmodifiedstictiondetectionnetwork (mSDN)[14].”
BackgroundMethodInfluentialAn autonomous valve stiction detection system based on data characterization
Zakharov, A., Zattoni, E., Xie, L., Garcia, O. P., & Jämsä‐Jounela, S. (2013) · Control Engineering Practice · doi:10.1016/j.conengprac.2013.07.004 · Cited by 45 · cited 2×
“ManyMFE-baseddetection methodshavebeendeveloped,suchasthefeaturesderivedfrom time-seriesdatacharacterization[4],[10],[11],andthefeatures derivedfromtheshapeofthecontrolleroutput(OP)andprocess variable(PV)[6],[9],[12].”
BackgroundKamaruddin, B., Zabiri, H., Mohd Amiruddin, A. A. A., Teh, W. K., Ramasamy, M., & Jeremiah, S. S. (2020) · Journal of Process Control · doi:10.1016/j.jprocont.2020.01.001 · Cited by 37 · cited 6×
“Incontrast,ALR-basedmethodsuse learningalgorithmsonthecollecteddatawithouttheinterven-tionofdomainexperts,directlyextractingfeatures[13],[14], [15].”
BackgroundMethodInfluentialData-Based Method To Diagnose Valve Stiction with Variable Reference Signal
Dambros, J. W. V., Farenzena, M., & Trierweiler, J. O. (2016) · Industrial & Engineering Chemistry Research · doi:10.1021/acs.iecr.6b01234 · Cited by 33
Oscillation Detection and Diagnosis in Process Industries by Pattern Recognition Technique
Dambros, J. W. V., Farenzena, M., & Trierweiler, J. O. (2019) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2019.06.078 · Cited by 30 · cited 1×
“DL-based methods have already been developed for valve stiction detection [13], [14], [16], but these approaches require lots”
MethodZhang, K., Liu, Y., Gu, Y., Ruan, X., & Wang, J. (2021) · IEEE/ASME Transactions on Mechatronics · doi:10.1109/tmech.2021.3087503 · Cited by 28 · cited 3×
“In the experiments, we compare the proposed MTFCC method with other seven methods, including LR, random forest (RF), support vector machine (SVM), extreme gradient boosting(XgBoost),LeNet-5,supervisedconvolutionalnetwork (supervised-Conv)[15],andmodifiedstictiondetectionnetwork (mSDN)[14].”
MethodSelf-citationDetection of stiction in flow control loops based on fuzzy clustering
Daneshwar, M., & Noh, N. M. (2015) · Control Engineering Practice · doi:10.1016/j.conengprac.2015.02.002 · Cited by 24
Garcia, O. P., Zakharov, A., & Jämsä‐Jounela, S. (2016) · IEEE Transactions on Control Systems Technology · doi:10.1109/tcst.2016.2583959 · Cited by 17 · cited 1×
“Many MFE-based detection methods have been developed, such as the features derived from time-series data characterization [4], [10], [11], and the features derived from the shape of the controller output (OP) and process variable (PV) [6], [9], [12].”
MethodNeighborhood Contrastive Learning Applied to Online Patient Monitoring
Yèche, H., Dresdner, G., Locatello, F., Hüser, M., & Rätsch, G. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.05142 · Cited by 16 · cited 1×
“This term is also consistent with the assumption widely used in time-series data [23], [24], [25], i.”
ResultDiagnosis quantification of control valves
Yamashita, Y. (2008) · SICE Annual Conference (SICE) · doi:10.1109/sice.2008.4655010 · Cited by 8
Research works printed but not linked: 1
- Brsio, 2015, Detection of stiction in level control loops, 48, 421
Cited by
Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects
Zhang, K., Wen, Q., Zhang, C., Cai, R., Jin, M., Liu, Y., Zhang, J. Y., Liang, Y., Pang, G., Song, D., et al. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3387317 · Cited by 295
“MTFCC [130] is another view generation method based on multi-scale characteristics, which samples time series samples at multiple scales and considers that the views from the same sample have similar representations, even if their scales are different.”
MethodSelf-citationUnraveling the ‘Anomaly’ in Time Series Anomaly Detection: A Self-supervised Tri-domain Solution
Sun, Y., Pang, G., Ye, G., Chen, T., Hu, X., & Yin, H. (2024) · IEEE International Conference on Data Engineering (ICDE) · doi:10.1109/icde60146.2024.00080 · Cited by 48
“This result verifies the generalizability of our model, as it does not rely on any specifically designed augmentation methods.”
BackgroundLiao, Y., Zhao, W., Feng, J., & Zi-sheng, L. (2024) · IEEE/ASME Transactions on Mechatronics · doi:10.1109/tmech.2023.3343398 · Cited by 30
“Furthermore, reduction of the clearance could cause stiction and wear of the spool and the sleeve [15].”
BackgroundAn Automated Diagnostic Framework for Multiple Oscillations With Median Complementary EEMD
Liu, S., Lang, X., Zhang, Y., Li, P., Xie, L., & Su, H. (2023) · IEEE Transactions on Control Systems Technology · doi:10.1109/tcst.2023.3343601 · Cited by 26
“Quantification approaches for valve stiction, on the other hand, contain categories such as apparent stic-tion [28], Hammerstein model [29], [30], [31], [32], nonlinear process model [33], [34], hybrid model [35], [36], [37], [38], and machine learning [39], [40], [41].”
BackgroundDebiased Contrastive Learning for Time-Series Representation Learning and Fault Detection
Zhang, K., Cai, R., Zhou, C., & Liu, Y. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2024.3359409 · Cited by 24
“Convolutional neural networks (CNNs) are commonly used in data-driven fault detection methods [14], [15], [16], [17], [18].”
MethodSelf-citation
Metadata
- Authors
- Zhang, K., Liu, Y., Gu, Y., Wang, J., & Ruan, X.
- Affiliations
- Zhejiang University (CN)
- Zhejiang Supcon Technology Co. Ltd., Hangzhou, China
- Venue
- IEEE/ASME Transactions on Mechatronics
- Published
- 2023-06
- Volume
- 28
- Issue
- 3
- Pages
- 1488-1499
- ISSN
- 1083-4435 (ISSN-L)
- 1941-014X
- OpenAlex
- W4313229394
- Cited by
- 16
- References
- 33
- Citations by year
- 2026: 2 · 2025: 4 · 2024: 8 · 2023: 2
- FWCI
- 1.41
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE/ASME Transactions on Mechatronics (publishedVersion)
- Funders
- National Key R&D Program of China (2021YFB2012300)
- Topics
- Physical Sciences › Engineering › Control and Systems Engineering › Fault Detection and Control Systems
- Physical Sciences › Engineering › Control and Systems Engineering › Machine Fault Diagnosis Techniques
- Physical Sciences › Computer Science › Signal Processing › Time Series Analysis and Forecasting
- Keywords
- fault detection, self-supervised contrastive learning, representation learning, convolutional encoder, time series, industrial process monitoring, unlabeled data
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.