Using machine-learning methods in determination of the pipe line gas turbine plant effective power

The paper considers methods of the gas turbine plant power designed for natural gas transportation and reveals their drawbacks. A program in the Python language was created to study applicability of the machine-learning methods to determine the plant power under operating conditions. Archival gas-dynamic parameters registered by the plant automatic control system were used as the initial data. Forecast quality of the machine-learning models was estimated depending on different sets of the feature parameters. Recommendations on the models use are provided; and the method error was determined. Hypothesis on applicability of models learned based on data of a single engine to estimate the power of the other engines of the same type was refuted. Machine-learning methods could be used to determine the gas turbine plant power even in the absence of part of the initial data, which is the main advantage over traditional methods.

References

[1] Islas J. The gas turbine: a new technological paradigm in electricity generation. Technol. Forecast. Soc. Change, 1999, vol. 60, no. 2, pp. 129–148, doi: https://doi.org/10.1016/S0040-1625(98)00036-5

[2] Benito A., Alonso G. Energy efficiency in air transportation. Butterworth-Heinemann, Oxford, 2018. 236 p.

[3] Saravanamuttoo H.I.H., Rogers G.F.C., Cohen H. Gas turbine theory. Pearson Education, Harlow, 2001. 64 p.

[4] Vasiliev B.U., Mardashov D.V. Industrial and environmental safety and efficiency of gas transportation industry. Indian. J. Sci. Technol., 2019, vol. 12, no. 1, pp. 1–10, doi: https://dx.doi.org/10.17485/ijst/2019/v12i1/97745

[5] Blinov V.L., Deryabin G.A. Technical condition estimation of the gas turbine axial compressor. IOP Conf. Ser.: Earth Environ. Sci., 2022, vol. 990, art. 012037, doi: https://doi.org/10.1088/1755-1315/990/1/012037

[6] Shchurovskiy V.A., Sinitsyn Yu.N., Korneev V.I. et al. PR 51-31323949-43-99. Metodicheskie ukazaniya po provedeniyu teplotekhnicheskikh i gazodinamicheskikh raschetov pri ispytaniyakh gazoturbinnykh gazoperekachivayushchikh agregatov [Methodology guidelines on performing thermotechnical and gas-dynamics calculations at tests on gas turbine compressor units]. Moscow, VNIIGAZ Publ., 1999. 26 p. (In Russ.).

[7] Vanchin A.G. Optimization of the mode of parallel collaboration of gas-distributing units. Neftegazovoe delo [Oil and Gas Business], 2013, no. 1. URL: http://ogbus.ru/files/ogbus/authors/Vanchin/Vanchin_9.pdf (in Russ.).

[8] GOST R 55798–2013. Ustanovki gazoturbinnye. Metody ispytaniy. Priemochnye ispytaniya [State standard GOST R 55798-2013. Gas turbines. Methods of testing. Acceptance tests]. Moscow, Standartinform Publ., 2015. 46 p. (In Russ.).

[9] Blinov V.L., Komarov O.V., Zubkov I.S. [Parametric diagnostics of gas turbine power plants as part of steam-gas units]. Sovremennye problemy teplofiziki i energetiki. Mat. III mezhd. konf. [Current issues of Thermal Physics and Energetics. Proc. III Int. Conf.]. Moscow, Izd-vo MEI Publ., 2020, p. 483. (In Russ.).

[10] Blinov V.L., Komarov O.V., Zaslavskiy E.A. Efficiency assessment of capacity and technical condition of a gas compressor unit. Energeticheskie sistemy [Energy System], 2019, vol. 4, no. 1, pp. 268–274. (In Russ.).

[11] Michelassi V., Allegorico C., Cioncolini S. et al. Machine learning in gas turbines. Mech. Eng., 2018, vol. 140, no. 9, pp. 54–55, doi: https://doi.org/10.1115/1.2018-SEP5

[12] Cuccu G., Danafar S., Cudré-Mauroux P. et al. A data-driven approach to predict NOx-emissions of gas turbines. IEEE Int. Conf. on Big Data, 2017, pp. 1283–1288, doi: https://doi.org/10.1109/BigData.2017.8258056

[13] Goyal V., Xu M., Kapat J. et al. Prediction enhancement of machine learning using time series modeling in gas turbines. ASME Turbo Expo, 2021, vol. 4, paper GT2021-59082, doi: https://doi.org/10.1115/GT2021-59082

[14] Ilichev V.Yu., Yurik E.A. Analysis of data arrays using the PANDAS library for Python. Nauchnoe obozrenie. Tekhnicheskie nauki [Scientific Review. Technical Science], 2020, no. 4, pp. 41–45. (In Russ.).

[15] Pedregosa F., Varoquaux G., Gramfort A. et al. Scikit-learn: machine learning in Python. JMLR, 2011, vol. 12, pp. 2825–2830.

[16] Seber G.A.F., Lee A.J. Linear regression analysis. Wiley, 2012. 582 p.

[17] Weisberg S. Applied linear regression. Wiley, 2005. 352 p.

[18] Ostertagová E. Modelling using polynomial regression. Procedia Eng., 2012, vol. 48, pp. 500–506, doi: https://doi.org/10.1016/j.proeng.2012.09.545

[19] Rifkin R.M., Ross A.L. Notes on regularized least squares. MIT, 2007. 8 p.

[20] Friedman J., Hastie T., Tibshirani R. Regularization paths for generalized linear models via coordinate descent. J. Stat. Softw., 2010, vol. 33, no. 1, pp. 1–22, doi: https://doi.org/10.18637/jss.v033.i01

[21] Efron B., Hastie T., Johnstone I. et al. Least angle regression. Ann. Statist., 2004, vol. 32, no. 2, pp. 407–499, doi: https://doi.org/10.1214/009053604000000067

[22] Belevich A.V., Rukteshel O.S., Kusyak V.A. MAPE criterion usage at the approximate polynoma selection for control range width estimation of the automated transmission actuators. Aktualnye voprosy mashinovedeniya [Topical Issues of Mechanical Engineering], 2017, vol. 6, pp. 76–80. (In Russ.).

[23] Breiman L. Random forests. Mach. Learn., 2001, vol. 45, no. 1, pp. 5–32, doi: https://doi.org/10.1023/A:1010933404324