Metaheuristic and Machine Learning Models for TFE-731-2, PW4056, and JT8D-9 Cruise Thrust
Abstract
The requirement for an accurate engine thrust model has a major antecedence in airline fuel saving programs, assessment of environmental effects of fuel consumption, emissions reduction studies, and air traffic management applications. In this study, utilizing engine manufacturers’ real data, a metaheuristic model based on genetic algorithms (GAs) and a machine learning model based on neural networks (NNs) trained with Levenberg-Marquardt (LM), delta-bar-delta (DBD), and conjugate gradient (CG) algorithms were accomplished to incorporate the effect of both flight altitude and Mach number in the estimation of thrust. For the GA model, the analysis of population size impact on the model’s accuracy and effect of number of data on model coefficients were also performed. For the NN model, design of optimum topology was searched for one- and two-hidden-layer networks. Predicted thrust values presented a close agreement with real thrust data for both models, among which LM trained NNs gave the best accuracies.
Nomenclature
- Abbreviations
- ATM
Air traffic management
- BADA
Base of Aircraft Data
- CG
Conjugate gradient
- CV
Cross-validation
- DBD
Delta-bar-delta
- FNN
Feed-forward neural network
- GA
Genetic algorithm
- HL
Hidden layer
- IP
Intellectual Property
- LM
Levenberg-Marquardt
- MAPE
Mean absolute percentage error
- MSE
Mean squared error
- NN
Neural network
- PE
Processing elements
- Symbols
- ai, bi
Genetic algorithm thrust model coefficients (i=1–3)
- ei
Output error for every input pattern of the network
- E
Error function of the network
- h
Flight altitude
- I
Identity matrix
- J
Jacobian matrix
- M
Mach number
- r
Linear correlation coefficient
- t
Time
- T
Engine cruise thrust
- V
Sum of squared error function of the network
- w
Weight vector of the neural network
- α
Learning rate
- αk
Variable learning rate
- θ
Smoothing factor
- κ
Additive constant
- λ
Non-negative gain
- ϕ
Multiplicative constant
References
1. Cavcar A, Cavcar M. Impact of aircraft performance differences on fuel consumption of aircraft in air traffic management environment. Aircr Eng Aerosp Tech 2004;76:502–15.10.1108/00022660410555176Search in Google Scholar
2. Cavcar M, Cavcar A. Aero-propulsive modeling of transport aircraft for air traffic management applications, in: AIAA Guidance, Navigation, and Control Conference and Exhibit, AIAA Paper 2004–4792, Providence, Rhode Island, 16–19 Aug 2004.10.2514/6.2004-4792Search in Google Scholar
3. Gong C, Chan WN. Using flight manual data to derive aero-propulsive models for predicting aircraft trajectories, in: AIAA Aircraft Technology, Integration and Operations Conference, Los Angeles, 2002.10.2514/6.2002-5844Search in Google Scholar
4. Collins B. Estimation of aircraft fuel consumption. J Aircraft 1982;19:969–75. doi:10.2514/3.44799. also AIAA Paper 81-0789R.Search in Google Scholar
5. Schilling GD. Modeling Aircraft Fuel Consumption with a Neural Network, Master of Science Thesis, Virginia Polytechnic Institute and State University, Blacksburg, 1997.Search in Google Scholar
6. Trani A, Wing-Ho F, Schilling G, Baik H, Seshadri A. A neural network model to estimate aircraft fuel consumption, in: 4th AIAA Aviation Technology, Integration and Operations (ATIO) Forum, AIAA Paper 2004–6401, Chicago, 20–22 Sept 2004.10.2514/6.2004-6401Search in Google Scholar
7. User Manual for the Base of Aircraft Data (BADA), Revision 3.12, 2014. Available at: https://www.eurocontrol.int/sites/default/files/field_tabs/content/documents/sesar/user-manual-bada-3-12.pdf. Accessed 27 Oct 2015.Search in Google Scholar
8. Gen M, Cheng R. Genetic algorithms and engineering design. New York: John Wiley & Sons, Inc, 1997:2, 5.10.1002/9780470172254Search in Google Scholar
9. Bukkapatnam STS, Sadananda K. A genetic algorithm for unified approach-based predictive modeling of fatigue crack growth. Int J Fatigue 2005;27:1354–9. doi:10.1016/j.ijfatigue.2005.07.025.Search in Google Scholar
10. De S, Ghosh A, Pal SK. Fitness evaluation in genetic algorithms with ancestors’ influence. In: SK. Pal, PP. Wang, editor. Genetic algorithms for pattern recognition. Boca Raton, FL: CRC Press, 1996:1, 2, 6, 7.Search in Google Scholar
11. Cantu-Paz E. Efficient and accurate parallel genetic algorithms. Norwell, MA: Kluwer Academic Publishers, 2001:2–4.10.1007/978-1-4615-4369-5Search in Google Scholar
12. Sakawa M.. Genetic algorithms and fuzzy multiobjective optimization. Norwell, MA: Kluwer Academic Publishers, 2002:14, 20.10.1007/978-1-4615-1519-7Search in Google Scholar
13. Saxena A, Saad A. Evolving an artificial neural network classifier for condition monitoring of rotating mechanical systems. J Appl Soft Comput 2007;7:441–54.10.1016/j.asoc.2005.10.001Search in Google Scholar
14. Michalewicz Z. Genetic algorithms+data structures=evolution programs, 3rd ed. Berlin Heidelberg, New York: Springer-Verlag, 1999:98, 99.Search in Google Scholar
15. Deb K. Multi-objective optimization using evolutionary algorithms. West Sussex: John Wiley & Sons, Ltd, 2001:85.Search in Google Scholar
16. Saarlas M. Aircraft performance. Hoboken, NJ: John Wiley & Sons, Inc., 2007:267–76.Search in Google Scholar
17. McCormick BW. Aerodynamics, aeronautics and flight mechanics, 2nd ed. New Jersey: John Wiley & Sons, Inc, 1995:345.Search in Google Scholar
18. Musharavati F. Process planning optimization in reconfigurable manufacturing systems. Dissertation.com. Boca Raton, FL, 2010:53.Search in Google Scholar
19. Yeung DS, Cloete I, Shi D, Ng WWY. Sensitivity analysis for neural networks. New York: Springer-Verlag Berlin Heidelberg, 2010:1, 3–5. DOI: 10.1007/978-3-642-02532-7.Search in Google Scholar
20. Tang H, Tan KC, Yi Z. Neural networks: computational models and applications. Berlin Heidelberg, Germany: Springer-Verlag, 2007:9–11.10.1007/978-3-540-69226-3_2Search in Google Scholar
21. He X, Xu S. Process neural networks: theory and applications, advanced topics in science and technology in China. Berlin Heidelberg, Germany: Springer-Verlag, 2010:23. DOI: 10.1007/978-3-540-73762-9.Search in Google Scholar
22. Lynch S. Dynamical systems with applications using mathematica. Boston: Birkhäuser, 2007:389.Search in Google Scholar
23. Basterrech S, Mohammed S, Rubino G, Soliman M. Levenberg-Marquardt training algorithms for random neural networks. Comput J 2011;54:125–35. DOI: 10.1093/comjnl/bxp101.Search in Google Scholar
24. Hagan MT, Menhaj MB. Training feedforward networks with the Marquardt algorithm. IEEE Trans Neural Netw 1994;5:989–93.10.1109/72.329697Search in Google Scholar
25. Costa MA, Braga AP, Menezes BR. Improving generalization of MLPs with sliding mode control and the Levenberg-Marquardt algorithm. Neurocomputing 2007;70:1342–7.10.1016/j.neucom.2006.09.003Search in Google Scholar
26. Zayani R, Bouallegue R, Roviras D. Levenberg-Marquardt learning neural network for adaptive pre-distortion for time-varying HPA with memory in OFDM systems, in: 16th European Signal Processing Conference, Lausanne, Switzerland, 2008.10.1155/2008/132729Search in Google Scholar
27. Haciismailoglu MC, Kucuk I, Derebasi N. Prediction of dynamic hysteresis loops of nano-crystalline cores. Expert Syst Appl 2009;36:2225–7.10.1016/j.eswa.2007.12.051Search in Google Scholar
28. Jacobs RA. Increased rates of convergence through learning rate adaptation. Neural Network 1988;1:295–307.10.1016/0893-6080(88)90003-2Search in Google Scholar
29. Al-Haik MS, Garmestani H, Savran A. Explicit and implicit viscoplastic models for polymeric composite. Int J Plasticity 2004;20:1875–907.10.1016/j.ijplas.2003.11.017Search in Google Scholar
© 2017 Walter de Gruyter GmbH, Berlin/Boston
Articles in the same Issue
- Frontmatter
- Effects of Cavity Configurations on Flameholding and Performances of Kerosene Fueled Scramjet Combustor
- Metaheuristic and Machine Learning Models for TFE-731-2, PW4056, and JT8D-9 Cruise Thrust
- Design Optimization Method for Composite Components Based on Moment Reliability-Sensitivity Criteria
- Use Deflected Trailing Edge to Improve the Aerodynamic Performance and Develop Low Solidity LPT Cascade
- A Co-modeling Method Based on Component Features for Mechatronic Devices in Aero-engines
- Evaluation and Analysis of Curvature-Corrected Filter-based Turbulent Model
- Aircraft Engine Thrust Estimator Design Based on GSA-LSSVM
- Effect of Inner Nozzle Lip Thickness on Co-flow Jet Characteristics
- Comparisons of Two Non-probabilistic Structural Reliability Analysis Methods for Aero-engine Turbine Disk
Articles in the same Issue
- Frontmatter
- Effects of Cavity Configurations on Flameholding and Performances of Kerosene Fueled Scramjet Combustor
- Metaheuristic and Machine Learning Models for TFE-731-2, PW4056, and JT8D-9 Cruise Thrust
- Design Optimization Method for Composite Components Based on Moment Reliability-Sensitivity Criteria
- Use Deflected Trailing Edge to Improve the Aerodynamic Performance and Develop Low Solidity LPT Cascade
- A Co-modeling Method Based on Component Features for Mechatronic Devices in Aero-engines
- Evaluation and Analysis of Curvature-Corrected Filter-based Turbulent Model
- Aircraft Engine Thrust Estimator Design Based on GSA-LSSVM
- Effect of Inner Nozzle Lip Thickness on Co-flow Jet Characteristics
- Comparisons of Two Non-probabilistic Structural Reliability Analysis Methods for Aero-engine Turbine Disk
