Methods of Creating Functional Models of Gas Turbine Engine Parts by Means of Numerical Modeling and Machine Learning Algorithms Using a Turbine Disk as an Example

Each aircraft gas turbine engine has design, layout and technological features that affect its operational life. To take into account all the factors that affect the operational life of the engine, a large amount of calculations is required. In this regard, it is important to develop functional models that represent virtual images of the main parts of each engine and contain information about their geometry, loading parameters and characteristics. These models should express the relationship between the many parameters measured and calculated during flight with the calculated values of the cyclic durability of the main engine parts. In recent years, machine learning has been used to solve such problems. The essence of this method is in training in the process of analyzing a variety of solutions under different parameters. This paper presents an approach that allows creating functional models of gas turbine engine elements using numerical modeling of the heat-stressed state and machine learning algorithms.

References

[1] Inozemtsev A.A., Sandratskiy V.L. Gazoturbinnyye dvigateli [Gas turbine engines]. Perm, Aviadvigatel’ publ., 2006. 1204 p.

[2] Eliseyev Yu.S., Krymov V.V., Malinovskiy K.A., Popov V.G. Tekhnologiya ekspluatatsii, diagnostiki i remonta gazoturbinnykh dvigateley [Technology of operation, diagnostics and repair of gas turbine engines]. Moscow, Vysshaya shkola publ., 2002. 355 p.

[3] Nozhnitskiy Yu.A. Modern methods for ensuring the strength reliability of aircraft engine parts. V sbornike rabot FGUP «TSIAM im. P.I. Baranova» [In the collection of works of FSUE “TsIAM them. P.I. Baranova”]. Moscow, TORUS PRESS publ., 2010. 456 p.

[4] Reznik S.V. Actual problems of the design, production and testing of space rocket composite structures. Aktual’nyye problemy razvitiya raketno-kosmicheskoy tekhniki i sistem vooruzheniy. Trudy MGTU im. N.E. Baumana [Actual problems of the development of rocket and space technology and weapons systems. Proceedings of BMSTU]. 2013, no. 606, pp. 295–311.

[5] Malinin N.N. Prikladnaya teoriya plastichnosti i polzuchesti [Applied Theory of Plasticity and Creep]. Moscow, Mashinostroyeniye publ., 1975. 400 p.

[6] Lee Y.-L., Barkey M.E., Kang H.-T. Metal fatigue analysis handbook. Practical Problem-Solving Techniques for Computer-Aided Engineering. Boston, Elsevier, 2012. 580 p.

[7] Collins J.A. Failure of Materials in Mechanical Design. New York, John Wiley & Sons, 1981. 622 p.

[8] Formalev V.F., Reviznikov D.L. Chislennyye metody [Numerical methods]. Moscow, Fizmatlit publ., 2004. 400 p.

[9] Johansson R. Numerical Python. New York, Apress, 2015. 481 p.

[10] Birger I.A., Shorr B.F., Iosilevich G.B. Raschet na prochnost’ detaley mashin [Strength calculation of machine parts]. Moscow, Mashinostroyeniye, 1993. 640 p.

[11] Serensen S.V. Issledovaniya malotsiklovoy prochnosti pri vysokikh temperaturakh [High Cycle Strength Studies]. Moscow, Nauka publ., 1975. 128 p.

[12] Haslwanter T. An Introduction to Statistics with Python. Berlin, Springer, 2015. 278 p.

[13] Downey A.B. Think Stats. Sebastopol, O’Reilly Media publ., 2015. 206 p.

[14] McKinney W. Python for Data Analysis. Sebastopol, O’Reilly Media publ., 2015. 482 p.

[15] Andreas C.M., Guido S. Introduction to Machine Learning with Python. Sebastopol, O’Reilly Media publ., 2017. 392 p.

[16] Brink H., Joseph W.R., Fetherolf M. Real-World Machine Learning. Shelter Island, Manning, 2016. 427 p.

[17] Stepnov M.N., Shavrin A.V. Statisticheskiye metody obrabotki rezul’tatov mekhanicheskikh ispytaniy [Statistical methods for processing the results of mechanical tests]. Moscow, Mashinostroyeniye publ., 2005. 399 p.