Advancements in Gas Turbine Fault Detection: A Machine Learning Approach Based on the Temporal Convolutional Network–Autoencoder Model
Source abstract
Abstract
To tackle the complex challenges inherent in gas turbine fault diagnosis, this study uses powerful machine learning (ML) tools. For this purpose, an advanced Temporal Convolutional Network (TCN)–Autoencoder model was presented to detect anomalies in vibration data. By synergizing TCN capabilities and Multi-Head Attention (MHA) mechanisms, this model introduces a new approach that performs anomaly detection with high accuracy. To train and test the proposed model, a bespoke dataset of CA 202 accelerometers installed in the Kirkuk power plant was used. The proposed model not only outperforms traditional GRU–Autoencoder, LSTM–Autoencoder, and VAE models in terms of anomaly detection accuracy, but also shows the Mean Squared Error (MSE = 1.447), Root Mean Squared Error (RMSE = 1.193), and Mean Absolute Error (MAE = 0.712). These results confirm the effectiveness of the TCN–Autoencoder model in increasing predictive maintenance and operational efficiency in power plants.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Engineering Diagnostics and Reliability · Fault Detection and Control Systems · Machine Fault Diagnosis Techniques
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Department of Mechanical Engineering, Academy of Engineering, RUDN University, 6 Miklukho-Maklaya Street, Moscow 117198, Russia
Department of Transport Equipment and Technology, Academy of Engineering, RUDN University, 6 Miklukho-Maklaya Street, Moscow 117198, Russia
References (41)
Show source-provided references
1. An overview of current and future sustainable gas turbine technologies10.1016/j.rser.2004.05.009 ↗
2. Asserting the climate benefits of the coal-to-gas shift across temporal and spatial scales10.1038/s41558-019-0457-1 ↗
3. A transient study on a solar-assisted combined gas power cycle for sustainable multi-generation in hot and cold climates: Case studies of Dubai and Toronto10.1016/j.energy.2023.128423 ↗
4. National Academies of Sciences, Engineering, and Medicine (2020). Advanced Technologies for Gas Turbines, National Academies Press.
5. Meher-Homji, C.B., Chaker, M., and Bromley, A.F. (2009). The Fouling of Axial Flow Compressors: Causes, Effects, Susceptibility, and Sensitivity. Volume 4: Cycle Innovations; Industrial and Cogeneration; Manufacturing Materials and Metallurgy; Marine, Proceedings of the ASME Turbo Expo 2009: Power for Land, Sea, and Air, Orlando, FL, USA, 8–12 June 2009, ASME.10.1115/GT2009-59239 ↗
6. John, E.O. (2011). Condition monitoring and assessment of power plant components. Power Plant Life Management and Performance Improvement, Woodhead Publishing.
7. Failure analysis of gas and wind turbine blades: A review10.1016/j.engfailanal.2023.107107 ↗
8. Waleed, K.M., Reza, K.K., and Ghorbani, S. (2023). Common failures in hydraulic Kaplan turbine blades and practical solutions. Materials, 16.10.3390/ma16093303 ↗
9. A comprehensive review on mechanical failures cause vibration in the gas turbine of combined cycle power plants10.1016/j.engfailanal.2022.106094 ↗
10. Kurz, R., Meher, H.C., Brun, K., Moore, J.J., and Gonzalez, F. (2013, January 1–3). Gas turbine performance and maintenance. Proceedings of the 42nd Turbomachinery Symposium, Houston, TX, USA.
11. Wear failure analysis of ball bearings based on lubricating oil for gas turbine10.1108/ILT-04-2022-0151 ↗
12. Fentaye, A.D., Baheta, A.T., Gilani, S.I., and Kyprianidis, K.G. (2019). A review on gas turbine gas-path diagnostics: State-of-the-art methods, challenges and opportunities. Aerospace, 6.10.3390/aerospace6070083 ↗
13. Gas turbine engine health management: Past, present, and future trends10.1115/1.4026126 ↗
14. Vibration monitoring of gas turbine engines: Machine-learning approaches and their challenges10.3389/fbuil.2017.00054 ↗
15. Temperature measurement techniques for gas and liquid flows using thermographic phosphor tracer particles10.1016/j.pecs.2017.09.001 ↗
16. Non-Destructive Evaluation of Ceramic Gas Turbine Components by X-Rays and Other Methods
17. Lubricating oil conditioning sensors for online machine health monitoring–A review10.1016/j.triboint.2017.01.015 ↗
18. DeSilva, U., Bunce, R.H., Schmitt, J.M., and Claussen, H. (2015). Gas turbine exhaust temperature measurement approach using time-frequency controlled sources. Volume 6: Ceramics; Controls, Diagnostics and Instrumentation; Education; Manufacturing Materials and Metallurgy; Honors and Awards, Proceedings of the ASME Turbo Expo 2015: Turbine Technical Conference and Exposition, Montreal, QC, Canada, 15–19 June 2015, ASME.10.1115/GT2015-42139 ↗
19. Mevissen, F., and Meo, M. (2019). A review of NDT/structural health monitoring techniques for hot gas components in gas turbines. Sensors, 19.10.3390/s19030711 ↗
20. Qaiser, M.T. (2023). Data Analysis and Prediction of Turbine Failures Based on Machine Learning and Deep Learning Techniques. [Master’s Thesis, Norwegian University of Science and Technology].
21. Yan, W., and Yu, L. (2019). On accurate and reliable anomaly detection for gas turbine combustors: A deep learning approach. arXiv.
22. A novel gas turbine fault diagnosis method based on transfer learning with CNN10.1016/j.measurement.2019.01.022 ↗
23. Fault diagnosis of gas turbine based on partly interpretable convolutional neural networks10.1016/j.energy.2020.117467 ↗
24. Rezaeian, N., Gurina, R., Saltykova, O.A., Hezla, L., Nohurov, M., and Reza Kashyzadeh, K. (2024). Novel GA-Based DNN Architecture for Identifying the Failure Mode with High Accuracy and Analyzing Its Effects on the System. Appl. Sci., 14.10.3390/app14083354 ↗
25. Failure mode and effects analysis using an improved pignistic probability transformation function and grey relational projection method10.1007/s40747-023-01268-0 ↗
26. Framework for development of comprehensive diagnostic tool for fault detection and diagnosis of gas turbine engines
27. Vatani, A. (2013). Degradation Prognostics in Gas Turbine Engines Using Neural Networks. [Ph.D. Thesis, Concordia University].
28. Diagnosis for loose blades in gas turbines using wavelet analysis10.1115/1.1772406 ↗
29. Theory and practice in using Fourier transform infrared spectroscopy to detect hydrocarbons in emissions from gas turbine engines10.1016/j.fuel.2005.01.018 ↗
30. RNN-Autoencoder approach for anomaly detection in power plant predictive maintenance systems
31. Farahani, M. (2021). Anomaly Detection on Gas Turbine Time-Series’ Data Using Deep LSTM-Autoencoder. [Master’s Thesis, Department of Computing Science, Faculty of Science and Technology, Umeå University].
32. Zhultriza, F., and Subiantoro, A. (2022, January 16–18). Gas turbine anomaly prediction using hybrid convolutional neural network with LSTM in power plant. Proceedings of the 2022 IEEE International Conference on Cybernetics and Computational Intelligence (CyberneticsCom), Malang, Indonesia.10.1109/CyberneticsCom55287.2022.9865487 ↗
33. April. Anomaly detection for time series using temporal convolutional networks and Gaussian mixture model10.1088/1742-6596/1187/4/042111 ↗
34. Fault detection in the gas turbine of the Kirkuk power plant: An anomaly detection approach using DLSTM-Autoencoder10.1016/j.engfailanal.2024.108213 ↗
35. Meggitt PLC (2023, May 14). CA202 Piezoelectric Accelerometer, Document Reference DS 262-020 Version 9, 15.06.2021. Available online: https://catalogue.meggittsensing.com/wp-content/uploads/2020/09/CA202-piezoelectric-accelerometer-data-sheet-English-.pdf.
36. Moving average and Savitzki-Golay smoothing filters using Mathcad
37. Mean squared error of empirical predictor10.1214/009053604000000201 ↗
38. Conditional variational autoencoders for probabilistic wind turbine blade fatigue estimation using Supervisory, Control, and Data Acquisition data10.1002/we.2621 ↗
39. Reza, K.K., Amiri, N., Ghorbani, S., and Souri, K. (2022). Prediction of concrete compressive strength using a back-propagation neural network optimized by a genetic algorithm and response surface analysis considering the appearance of aggregates and curing conditions. Buildings, 12.10.3390/buildings12040438 ↗
40. New neural network-based algorithm for predicting fatigue life of aluminum alloys in terms of machining parameters10.1016/j.engfailanal.2023.107128 ↗
41. On mean absolute error for deep neural network based vector-to-vector regression10.1109/LSP.2020.3016837 ↗