A Hierarchical Matrix Factorization-Based Method for Intelligent Industrial Fault Diagnosis
Abstract
:1. Introduction
- This paper proposes an HMF for fault diagnosis. The presented model consecutively decomposes fault data, cascading for several hierarchies. The middle hierarchies are expected to learn intrinsic characteristics of processes, the final hierarchy is expected to achieve high-level discriminative features of process data.
- Further, this paper extends HMF to the nonlinear version, referring to nonlinear HMF (NHMF). It adopts activation functions as the nonlinear transformer to describe the nonlinearities in practical industrial processes. Unlike traditional kernel-based methods, this strategy allows our model to avoid tricky kernel parameter tuning.
- The experimental results on the Multi-Phase Flow Process verify the effectiveness of the hierarchies and the nonlinear transformations.
2. Preliminary
3. Methodology
3.1. Hierarchical Matrix Factorization
3.2. Nonlinear Hierarchical Matrix Factorization
3.3. Fault Identification
4. Case Study
4.1. Data Description
4.2. Comparison with Shallow Methods
4.3. Comparison with Deep Models
4.4. Time Consumption Analysis
4.5. Effectiveness of Hierarchies
4.6. Parameter Study on Hierarchical Structure
4.7. Convergence Analysis
4.8. Visualization of the Learned Feature Matrices
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Yang, Z.; Chai, Y.; Yin, H.; Tao, S. LPV model based sensor fault diagnosis and isolation for permanent magnet synchronous generator in wind energy conversion systems. Appl. Sci. 2018, 8, 1816. [Google Scholar] [CrossRef]
- Fan, S.K.S.; Cheng, C.W.; Tsai, D.M. Fault Diagnosis of Wafer Acceptance Test and Chip Probing between Front-End-of-Line and Back-End-of-Line Processes. IEEE Trans. Autom. Sci. Eng. 2022, 19, 3068. [Google Scholar] [CrossRef]
- Sun, Z.; Wang, Y.; Sun, G. Fault Diagnosis of Rotating Machinery Based on Local Centroid Mean Local Fisher Discriminant Analysis. J. Vib. Eng. Technol. 2023, 11, 1417–1441. [Google Scholar] [CrossRef]
- Liu, M.; Kong, X.; Luo, Z.Y.L. Novel reduced kernel independent component analysis for process monitoring. Trans. Inst. Meas. Control. 2024, 46, 1374–1387. [Google Scholar] [CrossRef]
- Liu, Z.; Wu, D.; Huang, K.; Yang, C.; Gui, W. Fault Diagnosis of Complex Industrial Systems Based on Multi-Granularity Dictionary Learning and Its Application. IEEE Trans. Autom. Sci. Eng. 2024, 21, 192–203. [Google Scholar] [CrossRef]
- Tian, J.; Zhang, J.; Jiang, Y.; Wu, S.; Luo, H.; Yin, S. A novel generalized source-free domain adaptation approach for cross-domain industrial fault diagnosis. Reliab. Eng. Syst. Saf. 2024, 243, 1–12. [Google Scholar] [CrossRef]
- Zhang, B.; Lin, J.; Miao, Y.; Jiao, J.; Liu, H. Harmonic Sparse Structured Nonnegative Matrix Factorization: A Novel Method for the Separation of Coupled Fault Feature. IEEE Trans. Ind. Inform. 2024, 20, 6209–6221. [Google Scholar] [CrossRef]
- Liang, L.; Ding, X.; Wen, H.; Liu, F. Impulsive components separation using minimum-determinant KL-divergence NMF of bi-variable map for bearing diagnosis. Mech. Syst. Signal Process. 2022, 175, 1–23. [Google Scholar] [CrossRef]
- Zhou, H.; Yin, H.; Chai, Y. Multi-grained mode partition and robust fault diagnosis for multimode industrial processes. Reliab. Eng. Syst. Saf. 2023, 231, 109011. [Google Scholar] [CrossRef]
- Xiu, X.; Fan, J.; Yang, Y.; Liu, W. Fault Detection Using Structured Joint Sparse Nonnegative Matrix Factorization. IEEE Trans. Instrum. Meas. 2021, 70, 1–11. [Google Scholar] [CrossRef]
- Matania, O.; Dattner, I.; Bortman, J.; Kenett, R.S.; Parmet, Y. A systematic literature review of deep learning for vibration-based fault diagnosis of critical rotating machinery: Limitations and challenges. J. Sound Vib. 2024, 590, 118562. [Google Scholar] [CrossRef]
- Jalal, M.; Khalil, I.; Haq, A. Deep Learning approaches for visual faults diagnosis of photovoltaic systems: State-of-the-art review. Results Eng. 2024, 23, 102622. [Google Scholar] [CrossRef]
- Liu, Y.; Jiang, H.; Yao, R.; Zhu, H. Interpretable data-augmented adversarial variational autoencoder with sequential attention for imbalanced fault diagnosis. J. Manuf. Syst. 2023, 71, 342–359. [Google Scholar] [CrossRef]
- Liu, Y.; Jiang, H.; Yao, R.; Zeng, T. Counterfactual-augmented few-shot contrastive learning for machinery intelligent fault diagnosis with limited samples. Mech. Syst. Signal Process. 2024, 216, 111507. [Google Scholar] [CrossRef]
- Liu, Y.; Jiang, H.; Liu, C.; Yang, W.; Sun, W. Data augmented wavelet capsule generative adversarial network for rolling bearing fault diagnosis. Knowl. Based Syst. 2022, 252, 109439. [Google Scholar] [CrossRef]
- Yao, R.; Jiang, H.; Jiang, W.; Liu, Y.; Dong, Y. Deep discriminative sparse representation learning for machinery fault diagnosis. Eng. Appl. Artif. Intell. 2024, 135, 108836. [Google Scholar] [CrossRef]
- Chen, H.; Jiang, B.; Lu, N.; Mao, Z. Deep PCA based real-time incipient fault detection and diagnosis methodology for electrical drive in high-speed trains. IEEE Trans. Veh. Technol. 2018, 67, 4819–4830. [Google Scholar] [CrossRef]
- Deng, X.; Tian, X.; Chen, S.; Harris, C.J. Deep principal component analysis based on layerwise feature extraction and its application to nonlinear process monitoring. IEEE Trans. Control. Syst. Technol. 2018, 27, 2526–2540. [Google Scholar] [CrossRef]
- Ren, Z.; Zhang, W.; Zhang, Z. A Deep Nonnegative Matrix Factorization Approach via Autoencoder for Nonlinear Fault Detection. IEEE Trans. Ind. Inform. 2019, 16, 5042–5052. [Google Scholar] [CrossRef]
- Lee, D.D.; Seung, H.S. Learning the parts of objects by non-negative matrix factorization. Nature 1999, 401, 788–791. [Google Scholar] [CrossRef]
- Yang, J.; Chen, Y.; Zhang, L. An efficient approach for fault detection, isolation, and data recovery of self-validating multifunctional sensors. IEEE Trans. Instrum. Meas. 2017, 66, 543–558. [Google Scholar] [CrossRef]
- Fan, S.; Jia, Q.; Cheng, W.S. Safety Monitoring by a Graph-Regularized Semi-Supervised Nonnegative Matrix Factorization with Applications to a Vision-Based Marking Process. IEEE Access 2020, 8, 112278–112286. [Google Scholar] [CrossRef]
- Yang, Y.; Ming, A.; Zhang, Y.; Zhu, Y. Discriminative non-negative matrix factorization (DNMF) and its application to the fault diagnosis of diesel engine. Mech. Syst. Signal Process. 2017, 95, 158–171. [Google Scholar] [CrossRef]
- Yi, J.; Wu, L.; Zhou, W.; Wei, Z.; He, H.; Yao, L. A Sparse Dimensionality Reduction Approach Based on False Nearest Neighbors for Nonlinear Fault Detection. IEEE Trans. Syst. Man Cybern. Syst. 2021, 51, 4980–4992. [Google Scholar] [CrossRef]
- Ding, C.H.Q.; Li, T.; Jordan, M.I. Convex and semi-nonnegative matrix factorizations. IEEE Trans. Pattern Anal. Mach. Intell. 2008, 32, 45–55. [Google Scholar] [CrossRef]
- Ruiz-Carcel, C.; Cao, Y.; Mba, D.; Lao, L.; Samuel, R.T. Statistical process monitoring of a multiphase flow facility. Control. Eng. Pract. 2015, 42, 74–88. [Google Scholar] [CrossRef]
- Chiang, L.H.; Russell, E.L.; Braatz, R.D. Fault diagnosis in chemical processes using Fisher discriminant analysis, discriminant partial least squares, and principal component analysis. Chemom. Intell. Lab. Syst. 2000, 50, 243–252. [Google Scholar] [CrossRef]
- Balakrishnama, S.; Ganapathiraju, A. Linear discriminant analysis-a brief tutorial. Inst. Signal Inf. Process. 1998, 18, 1–8. [Google Scholar]
- Clemmensen, L.; Hastie, T.; Witten, D.; Ersbll, B. Sparse discriminant analysis. Technometrics 2011, 53, 406–413. [Google Scholar] [CrossRef]
- Lee, J.-M.; Yoo, C.K.; Choi, S.W.; Vanrolleghem, P.A.; Lee, I.-B. Nonlinear process monitoring using kernel principal component analysis. Chem. Eng. Sci. 2004, 59, 223–234. [Google Scholar] [CrossRef]
- Yu, W.; Zhao, C. Sparse exponential discriminant analysis and its application to fault diagnosis. IEEE Trans. Ind. Electron. 2017, 65, 5931–5940. [Google Scholar] [CrossRef]
- Lee, H.; Cichocki, A.; Choi, S. Kernel nonnegative matrix factorization for spectral EEG feature extraction. Neurocomputing 2009, 72, 3182–3190. [Google Scholar] [CrossRef]
Class | Type | Description | #Samples |
---|---|---|---|
#1 | Normal condition | - | 4000 |
#2 | Fault | Air line blockage | 1000 |
#3 | Fault | Water line blockage | 1000 |
#4 | Fault | Top separator input blockage | 1000 |
#5 | Fault | Open direct bypass | 1000 |
Method | avg.TPR | avg.PPV | Method | avg.TPR | avg.PPV | ||
PCA(5) | 71.98% | 79.95% | KPCA(Gaussian) | 87.40% | 87.21% | ||
LDA(4) | 82.23% | 83.21% | SEDA(Default) | 88.40% | 88.05% | ||
Linear | SDA(4) | 85.88% | 86.41% | Nonlinear | KNMF(Gaussian) | 85.70% | 86.05% |
NMF(5) | 82.80% | 83.29% | KNMF(Polynomial) | 86.90% | 88.25% | ||
HMF(19-12-5) | 92.23% | 92.34% | NHMF(19-12-5) | 94.48% | 94.45% |
SAE | DBN | HMF | NHMF | |
---|---|---|---|---|
avg. TPR | 88.49% | 89.88% | 92.23% | 94.48% |
avg. PPV | 88.45% | 89.98% | 92.34% | 94.45% |
NMF | NHMF | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
Shallow (5) | 2-Layer (19-5) | 3-Layer (19-12-5) | 4-Layer (20-15-10-5) | 5-Layer (22-20-15-10-5) | ||||||
TPR | PPV | TPR | PPV | TPR | PPV | TPR | PPV | TPR | PPV | |
#1 | 79.55% | 85.26% | 85.95% | 90.57% | 90.85% | 97.95% | 93.05% | 96.18% | 92.20% | 95.20% |
#2 | 53.60% | 54.03% | 94.00% | 81.74% | 98.00% | 87.50% | 94.80% | 84.79% | 88.20% | 84.48% |
#3 | 100.00% | 74.85% | 72.20% | 68.89% | 94.80% | 81.87% | 91.20% | 91.38% | 95.20% | 89.81% |
#4 | 91.40% | 96.41% | 98.00% | 97.42% | 99.60% | 98.81% | 99.20% | 97.83% | 98.00% | 96.27% |
#5 | 99.20% | 100.00% | 100.00% | 100.00% | 100.00% | 99.60% | 99.80% | 99.80% | 100.00% | 99.60% |
Avg. | 82.80% | 83.29% | 88.50% | 88.79% | 94.48% | 94.95% | 94.65% | 94.81% | 93.78% | 93.87% |
NMF | HMF | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
Shallow (5) | 2-Layer (19-5) | 3-Layer (19-12-5) | 4-Layer (20-15-10-5) | 5-Layer (22-20-15-10-5) | ||||||
TPR | PPV | TPR | PPV | TPR | PPV | TPR | PPV | TPR | PPV | |
#1 | 79.55% | 85.26% | 89.00% | 91.24% | 91.65% | 92.81% | 88.60% | 88.78% | 89.50% | 92.55% |
#2 | 53.60% | 54.03% | 74.00% | 71.43% | 84.80% | 79.25% | 79.20% | 74.86% | 77.80% | 75.68% |
#3 | 100.00% | 74.85% | 96.20% | 89.91% | 90.00% | 93.56% | 79.40% | 80.36% | 97.20% | 87.88% |
#4 | 91.40% | 96.41% | 95.20% | 93.15% | 96.40% | 94.88% | 96.60% | 96.60% | 95.20% | 90.32% |
#5 | 99.20% | 100.00% | 96.80% | 99.79% | 100.00% | 99.80% | 96.00% | 99.79% | 93.40% | 98.94% |
Avg. | 82.80% | 83.29% | 89.78% | 89.90% | 92.23% | 92.34% | 88.20% | 88.34% | 90.20% | 90.38% |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Li, Y.; Zhou, H.; Liu, J.; Tan, X. A Hierarchical Matrix Factorization-Based Method for Intelligent Industrial Fault Diagnosis. Sensors 2024, 24, 5408. https://doi.org/10.3390/s24165408
Li Y, Zhou H, Liu J, Tan X. A Hierarchical Matrix Factorization-Based Method for Intelligent Industrial Fault Diagnosis. Sensors. 2024; 24(16):5408. https://doi.org/10.3390/s24165408
Chicago/Turabian StyleLi, Yanxia, Han Zhou, Jiajia Liu, and Xuemin Tan. 2024. "A Hierarchical Matrix Factorization-Based Method for Intelligent Industrial Fault Diagnosis" Sensors 24, no. 16: 5408. https://doi.org/10.3390/s24165408