Hostname: page-component-76c49bb84f-qtd2s Total loading time: 0 Render date: 2025-07-05T00:51:55.389Z Has data issue: false hasContentIssue false
  • English
  • Français

A supervised machine-learning approach for turboshaft engine dynamic modeling under real flight conditions

Published online by Cambridge University Press:  03 July 2025

D. Paniccia Open the ORCID record for D. Paniccia [Opens in a new window] ,
F.A. Tucci ,
J. Guerrero Open the ORCID record for J. Guerrero [Opens in a new window] ,
L. Capone ,
N. Sanguini ,
T. Benacchio  and
L. Bottasso
D. Paniccia*
Affiliation:
Leonardo Labs, Leonardo S.p.A., Genoa 16151, Italy
F.A. Tucci
Affiliation:
Leonardo Labs, Leonardo S.p.A., Genoa 16151, Italy
J. Guerrero
Affiliation:
Leonardo Labs, Leonardo S.p.A., Genoa 16151, Italy
L. Capone
Affiliation:
Leonardo Labs, Leonardo S.p.A., Genoa 16151, Italy
N. Sanguini
Affiliation:
Leonardo Labs, Leonardo S.p.A., Samarate 21017, Italy
T. Benacchio
Affiliation:
Leonardo Labs, Leonardo S.p.A., Samarate 21017, Italy
L. Bottasso
Affiliation:
Leonardo Labs, Leonardo S.p.A., Samarate 21017, Italy
*
Corresponding author: Damiano Paniccia; Email: damiano.paniccia@leonardo.com
Get access Rights & Permissions [Opens in a new window]

Abstract

Rotorcraft engines are highly complex, nonlinear thermodynamic systems operating under varying environmental and flight conditions. Simulating their dynamics is crucial for design, fault diagnostics and deterioration control, requiring robust control systems to estimate performance throughout the flight envelope. Numerical simulations provide accurate assessments in both steady and unsteady scenarios through physics-based and mathematical models, although their development is challenging due to the engine’s complex physics and strong dependencies on environmental conditions. In this context, data-driven machine-learning techniques have gained significant interest for their ability to capture nonlinear dynamics and enable online performance estimation with competitive accuracy. This work explores different neural network architectures to model the turboshaft engine of Leonardo’s AW189P4 prototype, aiming to predict engine torque. The models are trained on a large database of real flight tests, covering a variety of operational manoeuvers under different conditions, thus offering a comprehensive performance representation. Additionally, sparse identification of nonlinear dynamics (SINDy) is applied to derive a low-dimensional model from the available data, capturing the relationship between fuel flow and engine torque. The resulting model highlights SINDy’s ability to recover underlying engine physics and suggests its potential for further investigations into engine complexity. The paper details the development and prediction results of each model, demonstrating that data-driven approaches can exploit a broader range of parameters compared to standard transfer function-based methods, enabling the use of trained schemes to simulate nonlinear effects in different engines and helicopters.

Keywords

machine learningrotorcraft engine modelingsparse regression

Information

Type
Research Article
Information
The Aeronautical Journal , First View , pp. 1 - 25
Check for updates
Copyright
© Leonardo S.p.A., 2025. Published by Cambridge University Press on behalf of Royal Aeronautical Society

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Article purchase

Temporarily unavailable

Footnotes

†

Now at: Weather Research, Danish Meteorological Institute, Sankt Kjelds Plads 11, 2100 Copenhagen Denmark

References

Corliss, D. A helicopter handling-qualities study of the effects of engine response characteristics height-control dynamics and excess poweron nap-of-the-earth operations, Tech. rep. U.S. Army Aeromechanics Laboratory, NASA Ames Research Center, 1982.Google Scholar
Mihaloew, D.J., Ballin, M.G. and Ruttledge, G. Rotorcraft flight-propulsion control integration: An eclectic design concept, Technical Report 2815, NASA, 1988.10.4271/872330CrossRefGoogle Scholar
Fuksman, I. and Sirica, S. Real-time execution of a high fidelity aero-thermodynamic turbofan engine simulation, J Eng Gas Turbines Power, 2012, 134, (5), p 054501. issn: 0742-4795. https://asmedigitalcollection.asme.org/gasturbinespower/article-pdf/134/5/054501/5894871/054501_1.pdf. url: https://doi.org/10.1115/1.4004856 CrossRefGoogle Scholar
Volponi, A.J. Gas Turbine parameter corrections, J Eng Gas Turbines Power, 1999, 121, (4), pp 613–621. https://doi.org/10.1115/1.2818516 CrossRefGoogle Scholar
Ballin, M.G. A high fidelity real-time simulation of a small turboshaft engine, Tech. rep. NTRS Research Center, 1988.Google Scholar
Yazar, I. Simulation of a high fidelity turboshaft engine-alternator model for turboelectric propulsion system design and applications, Int J Turbo Jet-Engines, 2018, 36, (3), pp 271–281. issn: 0334-0082, 2191-0332. https://doi.org/10.1515/tjj-2018-0036 Google Scholar
Zheng, Q., et al. A turboshaft engine NMPC scheme for helicopter autorotation recovery Maneuver, Aerosp Sci Technol, 2018, 76, pp 421–432. issn: 12709638. https://doi.org/10.1016/j.ast.2018.01.034 CrossRefGoogle Scholar
Mao, H., et al. Versatile simulation platform for turboshaft engine control system, 2019 Chinese Control Conference (CCC), 2019, pp 7211–7216. https://doi.org/10.23919/ChiCC.2019.8865902 CrossRefGoogle Scholar
Duyar, A., Gu, Z. and Litt, J.S. A simplified dynamic model of the T700 turboshaft engine, J Am Helicopter Soc., 1995, 40, (4), pp 62–70.10.4050/JAHS.40.62CrossRefGoogle Scholar
Litt, J.S. A real-time simulator of a turbofan engine. Vol. 100869. National Aeronautics and Space Administration Technical Memorandum (TM), Lewis Research Center, Cleveland, OH, 1989.Google Scholar
Merrill, W.C., et al. HYTESS: A hypothetical turbofan engine simplified simulation. Tech. rep. 1984.Google Scholar
Duyar, A., et al. A simplified dynamic model of the space shuttle main engine, J Dyn Sys, Meas, Control, 1994, 116, (4), pp 815–819. https://doi.org/10.1115/1.2899286 CrossRefGoogle Scholar
Castiglione, T., et al. Linear model of a turboshaft aero-engine including components degradation for control-oriented applications, Energies, 2023, 16, (6), p 2634. issn: 1996-1073. https://doi.org/10.3390/en16062634 CrossRefGoogle Scholar
Xie, R., et al. Digital twin technologies for turbomachinery in a life cycle perspective: A review, Sustainability, 2021, 13, (5), p 2495.10.3390/su13052495CrossRefGoogle Scholar
Asgari, H., Ory, E., and Lappalainen, J. Recurrent neural network based simulation of a single shaft gas turbine, Proceedings of the 61st SIMS Conference on Simulation and Modelling SIMS. Finland Linko¨ping Electronic Conference Proceedings, 2020, pp 22–24.Google Scholar
De Giorgi, M.G. and Quarta, M. Hybrid multigene genetic programming-artificial neural networks approach for dynamic performance prediction of an aeroengine, Aerosp Sci Technol., 2020, 103, p 105902.10.1016/j.ast.2020.105902CrossRefGoogle Scholar
Zheng, Q., et al. Aero-engine on-board dynamic adaptive MGD neural network model within a large flight envelope, IEEE Access, 2018, 6, pp 45755–45761.10.1109/ACCESS.2018.2789935CrossRefGoogle Scholar
Kiakojoori, S. and Khorasani, K. Dynamic neural networks for gas turbine engine degradation prediction, health monitoring and prognosis, Neural Comput Appl., 2016, 27, (8), pp 2157–2192.10.1007/s00521-015-1990-0CrossRefGoogle Scholar
Wang, W., et al. Helicopter dynamic modeling and system development for flight simulation, 2021 IEEE 3rd International Conference on Civil Aviation Safety and Information Technology (ICCASIT), 2021, pp 1220–1224. https://doi.org/10.1109/ICCASIT53235.2021. 9633434.CrossRefGoogle Scholar
Gu, Z., et al. An online data-driven LPV modeling method for turbo-shaft engines. Energies, 2022, 15, (4), 1255. issn: 1996-1073. https://doi.org/10.3390/en15041255 CrossRefGoogle Scholar
Badihi, H., Shahriari, A. and Naghsh, A. Artificial neural network application to fuel flow function for demanded jet engine performance, 2009 IEEE Aerospace Conference, IEEE, 2009, pp 1–7.10.1109/AERO.2009.4839656CrossRefGoogle Scholar
Corsini, A., et al. Cascade with sinusoidal leading edges: Identification and quantification of deflection with unsupervised machine learning, Proceedings of the ASME Turbo Expo 2021, Vol. 1, ASME, 2021, pp 1–10.10.1115/GT2021-59277CrossRefGoogle Scholar
Tieghi, L., et al. A machine-learnt wall function for rotating diffusers, J Turbomach., 2021, 143, (8), p 081012.10.1115/1.4050442CrossRefGoogle Scholar
Tayarani-Bathaie, S.S, Vanini, Z.S. and Khorasani, K. Dynamic neural network-based fault diagnosis of gas turbine engines, Neurocomputing, 2014, 125, pp 153–165.10.1016/j.neucom.2012.06.050CrossRefGoogle Scholar
Vatani, A., Khorasani, K. and Meskin, N. Health monitoring and degradation prognostics in gas turbine engines using dynamic neural networks, Turbo Expo: Power for Land, Sea, and Air, Vol. 56758, American Society of Mechanical Engineers, 2015, V006T05A030.Google Scholar
Brunton, S.L., et al. Data-driven aerospace engineering: reframing the industry with machine learning, Aiaa J., 2021, 59, (8), pp 2820–2847.Google Scholar
Schmid, P.J. Dynamic mode decomposition of numerical and experimental data, J Fluid Mech., 2010, 656, pp 5–28.10.1017/S0022112010001217CrossRefGoogle Scholar
Proctor, J.L., Brunton, S.L. and Kutz, J.N. Dynamic mode decomposition with control, SIAM J Appl Dyn Syst., 2016, 15, (1), pp 142–161.10.1137/15M1013857CrossRefGoogle Scholar
Williams, M.O, Kevrekidis, I.G. and Rowley, C.W. A data–driven approximation of the koopman operator: Extending dynamic mode decomposition, J Nonlinear Sci., 2015, 25, pp 1307–1346.10.1007/s00332-015-9258-5CrossRefGoogle Scholar
Brunton, S.L., Proctor, J.L. and Kutz, J.N. Discovering governing equations from data by sparse identification of nonlinear dynamical systems, Proc Natl Acad Sci., 2016, 113, (15), pp 3932–3937.10.1073/pnas.1517384113CrossRefGoogle Scholar
Liu, Y., et al. Data-based modeling and control of the nonlinear aircraft system using extended implicit sparse identification, IEEE Trans Aerosp Electron Syst., 2025.10.1109/TAES.2025.3531345CrossRefGoogle Scholar
Sorokina, M., Sygletos, S. and Turitsyn, S. Sparse identification for nonlinear optical communication systems: SINO method, Opt Express, 2016, 24, (26), pp 30433–30443.10.1364/OE.24.030433CrossRefGoogle Scholar
Dam, M., et al. Sparse identification of a predator-prey system from simulation data of a convection model, Phys Plasmas, 2017, 24, (2).10.1063/1.4977057CrossRefGoogle Scholar
Narasingam, A. and Kwon, J.S.I. Data-driven identification of interpretable reduced-order models using sparse regression, Comput Chem Eng., 2018, 119, pp 101–111.10.1016/j.compchemeng.2018.08.010CrossRefGoogle Scholar
Hoffmann, M., Fro¨hner, C. and Noe´, F. Reactive SINDy: Discovering governing reactions from concentration data, J Chem Phys., 2019, 150, (2), p 025101. https://doi.org/10.1063/1.5066099. PMID: 30646700.CrossRefGoogle Scholar
Lai, Z. and Nagarajaiah, S. Sparse structural system identification method for nonlinear dynamic systems with hysteresis/inelastic behavior, Mech Syst Signal Process., 2019, 117, pp 813–842.10.1016/j.ymssp.2018.08.033CrossRefGoogle Scholar
Loiseau, J.C., Noack, B.R. and Brunton, S.L. Sparse reduced-order modelling: Sensor-based dynamics to full-state estimation, J Fluid Mech., 2018, 844, pp 459–490.10.1017/jfm.2018.147CrossRefGoogle Scholar
Elsayed, E., El-Badawy, A. and Zometa, P. Data-driven discovery of the equations of motion of twin-tailed fighter aircraft, 2024 28th International Conference on System Theory, Control and Computing (ICSTCC), IEEE, 2024, pp 362–367.10.1109/ICSTCC62912.2024.10744743CrossRefGoogle Scholar
Mukhachev, P., et al. Evaluation of ml algorithms for system dynamics identification of aircraft pressure control system, PHM Society European Conference, 2021, 6, (1), pp 7–7.10.36001/phme.2021.v6i1.2799CrossRefGoogle Scholar
Lee, J.D., et al. Sparse identification of nonlinear dynamics-based model predictive control for multirotor collision avoidance. arXiv preprint arXiv:2412.06388 , 2024.Google Scholar
Yahagi, S., et al. Sparse identification and nonlinear model predictive control for diesel engine air path system, Int J Control Autom Syst., 2025, 23, (2), pp 620–629.10.1007/s12555-024-0452-9CrossRefGoogle Scholar
Goodfellow, I., et al. Deep learning, vol. 1, No. 2, MIT Press, 2016, Cambridge.Google Scholar
Hochreiter, S. Long short-term memory, Neural Computation MIT-Press, 1997.10.1162/neco.1997.9.8.1735CrossRefGoogle Scholar
Staudemeyer, R.C. and Morris, E.R. Understanding LSTM–a tutorial into long short-term memory recurrent neural networks, arXiv preprint arXiv:1909.09586 , 2019.Google Scholar
Paszke, A., et al. Pytorch: An imperative style, high-performance deep learning library, Proc 33rd Int Conf Neural Inf Process Syst., 2019, 721, p 12.Google Scholar
LeCun, Y., Bengio, Y. and Hinton, G. Deep learning, Nature, 2015, 521, (7553), pp 436–444.10.1038/nature14539CrossRefGoogle ScholarPubMed
de Silva, B.M., et al. Pysindy: a python package for the sparse identification of nonlinear dynamics from data, arXiv preprint arXiv:2004.08424 , 2020.Google Scholar
Kaptanoglu, A.A., et al. PySINDy: A comprehensive Python package for robust sparse system identification, arXiv preprint arXiv:2111.08481 , 2021.Google Scholar
Brunton, S.L., Proctor, J.L. and Kutz, J.N. Sparse identification of nonlinear dynamics with control (SINDYc), IFAC-PapersOnLine, 2016, 49, (18), pp 710–715.10.1016/j.ifacol.2016.10.249CrossRefGoogle Scholar