A physical-data-driven combined transfer learning method for gas turbine performance estimation
-
Ran Ao
,
Yujia Ma
Abstract
Gas turbine performance estimation methods rely on historical data from the same-type gas turbine. However, due to individual differences, this knowledge cannot be directly shared. In this paper, we propose a physical-data-driven combined transfer learning method. This method enables knowledge sharing between a gas turbine with complete washing cycle data and a newly commissioned gas turbine. The physical-driven models are constructed utilizing particle swarm optimization to acquire the labels used for pretraining. A data-driven model is developed through a convolutional neural network that employs labels to extract degradation features. Then, the labels for fine-tuning the network. Historical degradation knowledge is transferred to the newly commissioned gas turbine. Experiments results demonstrate that the accuracy of this method is improved by at least about 25 % compared to other methods. The computation time is reduced by at least about 98 % compared to physical-driven methods. This method enables effective performance estimation of gas turbines.
Acknowledgments
This work presented in this paper was supported by the National Science and Technology Major Project (2019-I-003-0004).
-
Research ethics: Not applicable.
-
Informed consent: Informed consent was obtained from all individuals included in this study, or their legal guardians or wards.
-
Author contributions: R.A: Conceptualization, Methodology, Review and Editing, Writing – review & editing; L.C: Data curation; Y.P.C: Funding acquisition, Supervision, Resources; Y.J.M: Program, Computation, Validation, and, Analysis; S.Y.L: Funding acquisition, Supervision, Resources.
-
Use of Large Language Models, AI and Machine Learning Tools: None declared.
-
Conflict of interest: The authors state no conflict of interest.
-
Research funding: This work presented in this paper was supported by the National Science and Technology Major Project (2019-I-003-0004).
-
Data availability: Not applicable.
References
1. Dai, S, Zhang, XY, Luo, MY. A novel data-driven approach for predicting the performance degradation of a gas turbine. Energies 2024;17:1–17. https://doi.org/10.3390/en17040781.Suche in Google Scholar
2. Luan, JQ, Cao, YP, Ao, R, Han, XY, Li, SY. An overhaul cycle performance degradation modeling method for marine gas turbines. ISA Trans 2025;156:374–88. https://doi.org/10.1016/j.isatra.2024.11.004.Suche in Google Scholar PubMed
3. Sun, JZ, Zuo, HF, Wang, WB, Pecht, MG. Application of a state space modeling technique to system prognostics based on a health index for condition-based maintenance. Mech Syst Signal Process 2012;28:585–96. https://doi.org/10.1016/j.ymssp.2011.09.029.Suche in Google Scholar
4. Palman, M, Leizeronok, B, Cukurel, B. Comparative study of numerical approaches to adaptive gas turbine cycle analysis. Int J Turbo Jet Engines 2023;40:425–36. https://doi.org/10.1515/tjj-2021-0021.Suche in Google Scholar
5. Lu, F, Huang, JQ, Ji, CS, Zhang, DD, Jiao, HB. Gas path on-line fault diagnostics using a nonlinear integrated model for gas turbine engines. Int J Turbo Jet Engines 2014;31:261–75. https://doi.org/10.1515/tjj-2014-0001.Suche in Google Scholar
6. Kong, C, Ki, J. Study on component map identification from gas turbine performance deck data using hybrid method. Int J Turbo Jet Engines 2007;24:171–81. https://doi.org/10.1515/tjj.2007.24.3-4.171.Suche in Google Scholar
7. Loboda, I, Yepifanov, S, Feldshteyn, Y. An integrated approach to gas turbine monitoring and diagnostics. Int J Turbo Jet Engines 2009;26:111–26. https://doi.org/10.1515/tjj.2009.26.2.111.Suche in Google Scholar
8. Elmdoost-gashti, M, Shafiee, M, Bozorgi-Amiri, A. Enhancing resilience in marine propulsion systems by adopting machine learning technology for predicting failures and prioritising maintenance activities. J Mar Eng Technol 2024;23:18–32. https://doi.org/10.1080/20464177.2023.2243748.Suche in Google Scholar
9. Caner, M, Gedik, E, Keçebas, A. Investigation on thermal performance calculation of two type solar air collectors using artificial neural network. Expert Syst Appl 2011;38:1668–74.10.1016/j.eswa.2010.07.090Suche in Google Scholar
10. Pawelczyk, M, Fulara, S, Sepe, M, De Luca, A, Badora, M. Industrial gas turbine operating parameters monitoring and data-driven prediction. Eksploatacja I Niezawodnosc-Maintenance Reliability 2020;22:391–9. https://doi.org/10.17531/ein.2020.3.2.Suche in Google Scholar
11. Jin, YF, Liu, C, Tian, X, Huang, HZ, Deng, GF, Guan, YL, et al.. A hybrid model of LSTM neural networks with a thermodynamic model for condition-based maintenance of compressor fouling. Meas Sci Technol 2021;32:1–13. https://doi.org/10.1088/1361-6501/ac026f.Suche in Google Scholar
12. Li, WH, Huang, RY, Li, JP, Liao, YX, Chen, ZY, He, GL, et al.. A perspective survey on deep transfer learning for fault diagnosis in industrial scenarios: theories, applications and challenges. Mech Syst Signal Process 2022;167:1–30. https://doi.org/10.1016/j.ymssp.2021.108487.Suche in Google Scholar
13. Huo, CR, Jiang, QS, Shen, YH, Qian, CH, Zhang, QK. New transfer learning fault diagnosis method of rolling bearing based on ADC-CNN and LATL under variable conditions. Measurement 2022;188:1–14. https://doi.org/10.1016/j.measurement.2021.110587.Suche in Google Scholar
14. Yang, XS, Bai, ML, Liu, JF, Liu, J, Yu, DR. Gas path fault diagnosis for gas turbine group based on deep transfer learning. Measurement 2021;181:1–20. https://doi.org/10.1016/j.measurement.2021.109631.Suche in Google Scholar
15. Li, YG. Gas turbine performance and health status estimation using adaptive gas path analysis. J Eng Gas Turbine Power Trans ASME 2010;132:1–9. https://doi.org/10.1115/1.3159378.Suche in Google Scholar
16. Cao, YP, Chen, L, Du, JW, Yu, F, Yang, QC, Wu, MH. The degradation simulation of compressor salt fog fouling for marine gas turbine. In: Proceedings of the ASME Turbo Expo 2017: turbomachinery technical conference and exposition. Volume 6: ceramics; controls, diagnostics and instrumentation; education; manufacturing materials and metallurgy. Charlotte, North Carolina, USA; 2017.10.1115/GT2017-64464Suche in Google Scholar
17. Lu, SW, Zhou, WX, Huang, JQ, Wang, B. Research on a component characteristic adaptive correction method for variable cycle engines. Intl J Turbo Jet Engines 2021;40:399–410.10.1515/tjj-2021-0026Suche in Google Scholar
18. Chen, M, Chen, BY, Zhang, HB. The aerothermodynamic cycle optimal design of a turbofan engine. Int J Turbo Jet Engines 2024;40:s3–16. https://doi.org/10.1515/tjj-2021-0045.Suche in Google Scholar
19. Li, J, Zhang, G, Ying, Y. Gas turbine gas path fault diagnosis based on adaptive nonlinear steady-state thermodynamic model. Int J Perform Eng 2018;14:751.10.23940/ijpe.18.04.p18.751764Suche in Google Scholar
20. Wang, ZT, Zhang, JK, Gao, CM, Ming, L. Effect of air properties on a twin-shaft turbofan engine performance during start-up. Appl Therm Eng 2023;218:1–20. https://doi.org/10.1016/j.applthermaleng.2022.119387.Suche in Google Scholar
21. Yan, XA, Yan, WJ, Xu, YD, Yuen, KV. Machinery multi-sensor fault diagnosis based on adaptive multivariate feature mode decomposition and multi-attention fusion residual convolutional neural network. Mech Syst Signal Process 2023;202:1–33. https://doi.org/10.1016/j.ymssp.2023.110664.Suche in Google Scholar
22. Xiang, S, Qin, Y, Luo, J, Wu, F, Gryllias, K. A concise self-adapting deep learning network for machine remaining useful life prediction. Mech Syst Signal Process 2023;191:1–17. https://doi.org/10.1016/j.ymssp.2023.110187.Suche in Google Scholar
23. Chen, ZY, Gryllias, K, Li, WH. Mechanical fault diagnosis using convolutional neural networks and extreme learning machine. Mech Syst Signal Process 2019;133:1–21. https://doi.org/10.1016/j.ymssp.2019.106272.Suche in Google Scholar
24. Rassoulinejad-Mousavi, SM, Al-Hindawi, F, Soori, T, Rokoni, A, Yoon, H, Hu, H, et al.. Deep learning strategies for critical heat flux detection in pool boiling. Appl Therm Eng 2021;190:1–11.10.1016/j.applthermaleng.2021.116849Suche in Google Scholar
25. Dizaji, MS, Mao, Z, Haile, M. A hybrid-attention-convLSTM-based deep learning architecture to extract modal frequencies from limited data using transfer learning. Mech Syst Signal Process 2023;187:1–22. https://doi.org/10.1016/j.ymssp.2022.109949.Suche in Google Scholar
26. Tao, LF, Liu, HF, Ning, GA, Cao, WY, Huang, BH, Lu, C. LLM-based framework for bearing fault diagnosis. Mech Syst Signal Process 2025;224:1–19. https://doi.org/10.1016/j.ymssp.2024.112127.Suche in Google Scholar
27. Wang, Z, Gu, YJ. A steady-state detection method based on Gaussian discriminant analysis for the on-line gas turbine process. Appl Therm Eng 2018;133:1–7. https://doi.org/10.1016/j.applthermaleng.2018.01.025.Suche in Google Scholar
28. Wang, Z, Gu, YJ, Han, XD, Zhu, JJ, Xu, JH. Anomaly detection for heavy power generation gas turbine considering the effect of output power variation. Proc Inst Mech Eng Part A J Power Energy 2020;234:795–803. https://doi.org/10.1177/0957650919879610.Suche in Google Scholar
29. Aretakis, N, Roumeliotis, I, Doumouras, G, Mathioudakis, K. Compressor washing economic analysis and optimization for power generation. Appl Energy 2012;95:77–86. https://doi.org/10.1016/j.apenergy.2012.02.016.Suche in Google Scholar
30. Li, JC, Ying, YL. Gas turbine gas path diagnosis under transient operating conditions: a steady state performance model based local optimization approach. Appl Therm Eng 2020;170:1–14.10.1016/j.applthermaleng.2020.115025Suche in Google Scholar
31. Chatterjee, S, Litt, JS. Online model parameter estimation of jet engine degradation for autonomous propulsion control. In: Proceedings of the AIAA guidance, navigation, and control conference and exhibit. Austin, Texas, USA; 2003.10.2514/6.2003-5425Suche in Google Scholar
32. Loboda, I, Zárate, LAM, Yepifanov, S, Herrera, CM, Ruiz, JLP. Estimation of gas turbine unmeasured variables for an online monitoring system. Int J Turbo Jet Engines 2020;37:413–28.10.1515/tjj-2017-0065Suche in Google Scholar
33. Dávalos, JO, García, JC, Urquiza, G, Huicochea, A, De Santiago, O. Prediction of film cooling effectiveness on a gas turbine blade leading edge using ANN and CFD. Int J Turbo Jet Engines 2018;35:101–11. https://doi.org/10.1515/tjj-2016-0034.Suche in Google Scholar
34. Kang, X, Cheng, JS, Yang, Y, Liu, F. Repetitive transient impact detection and its application in cross-machine fault detection of rolling bearings. Mech Syst Signal Process 2025;228:1–19. https://doi.org/10.1016/j.ymssp.2025.112422.Suche in Google Scholar
© 2025 Walter de Gruyter GmbH, Berlin/Boston
