Analysis and selection of mathematical equations of non-stationary gas flow for machine learning of proxy models of main gas pipelines
UDC: 519.63
DOI: -
Authors:
KISLENKO NIKOLAY A.
1,2,
KAZAK ALEXANDER S.2,
OLEYNIKOV ALEXEY V.3,
BELINSKY ALEXANDER V.2,
MALETIN ANDREY V.2
1 Gazprom, St. Petersburg, Russia
2 NIIgazekonomika, Moscow, Russia
3 Gazprom transgaz Tchaikovsky, Tchaikovsky, Russia
Keywords: gas transportation systems, systems of differential equations, numerical methods, neural network models
Annotation:
Changes of energy markets structure and behavior significantly affect the operation of existing gas transportation systems (GTS). Modern conditions of GTS operation are characterized by frequently changing load caused by increasing unevenness of gas consumption, significant difference of current technological modes of gas transportation from design ones as well as increasing variability of modes. Operational management of GTS requires implementation of modern tools for modeling and optimization of gas transportation modes. But if previously the planning of modes was carried out mainly using models of steady-state gas flow, today the transition to non-stationary models is becoming an urgent task. At the same time, despite the great achievements of science in the field of development of mathematical support for solving problems of simulating GTS non-stationary modes, progress in the field of such modes optimization is less significant due to the computational complexity of the corresponding calculation problems. In this regard, research aimed at creating rationally constructed high-speed algorithms for simulating and optimizing GTS non-stationary modes is actively goes on. A promising direction of this research is the use of proxy models (also called surrogate models) based on the application of machine learning methods. Such models allow simulating the processes of non-stationary gas flow in main gas pipelines. Neural network architectures or other types of machine learning models can be used to create proxy models. They are characterized by high computations speed and at the same time have accuracy acceptable for solving many practical problems of gas transportation. Proxy models training is based on pre-prepared training samples – actual or synthesized data obtained by conducting multivariate thermal-hydraulic calculations. The authors of the article present the results of the analysis of several known models of non-stationary gas flow in linear sections of main gas pipelines, which can be used to prepare these training samples. Based on the results of the computational experiments, the recommended model was selected. The main criteria for choosing the model were the adequacy of the results obtained and the speed of calculations.
Bibliography:
1. Kazak A.S., Oleynikov A.V. Primenenie metodov sistemnogo analiza dlya postroeniya tsifrovykh dvoynikov fragmentov gazotransportnykh sistem na osnove neyrosetevogo modelirovaniya (osnovnye polozheniya metodologii analiza i sinteza) // Avtomatizatsiya i informatizatsiya TEK. – 2024. – № 1(606). – S. 5–12.
2. O novom metode tsifrovogo modelirovaniya nestatsionarnykh rezhimov techeniya gaza v magistral'nykh gazoprovodakh s primeneniem neyronnykh operatorov / A.V. Belinskiy, D.V. Gorlov, I.A. Pyatyshev, A.E. Titov // Gazovaya prom-st'. – 2024. – № 5(865). – S. 54–66.
3. Sardanashvili S.A. Raschetnye metody i algoritmy (truboprovodnyy transport gaza). – M.: Izd-vo "Neft' i gaz" RGU nefti i gaza im. I.M. Gubkina, 2005. – 577 s.
4. Metodika rascheta techeniya prirodnogo gaza po trube v programmnom komplekse modelirovaniya gazotransportnykh sistem "Volna" / A.A. Vorob'ev, A.A. Kalinin, G.V. Kaspiev [i dr.] // Mat. modelirovanie. – 2014. – T. 26, № 7. – S. 87–96.
5. Sukharev M.G., Samoylov R.V. Analiz i upravlenie statsionarnymi i nestatsionarnymi rezhimami transporta gaza: monogr. – M.: Izdat. tsentr RGU nefti i gaza im. I.M. Gubkina, 2016. – 397 s.
6. Sukharev M.G., Kosova K.O. Metod identifikatsii parametrov sistem gazosnabzheniya pri nestatsionarnykh rezhimakh techeniya gaza // Avtomatika i telemekhanika. – 2017. – № 5. – S. 141–151.
7. Sukharev M.G., Popov R.V. Novaya metodika modelirovaniya nestatsionarnykh techeniy gaza v sistemakh gazosnabzheniya // Izv. RAN. Energetika. – 2015. – № 2. – S. 150–159.
8. Tsybul'nik V.N., Rubel' V.V. Kompleks modelirovaniya i optimizatsii gazotransportnykh sistem "ASTRA" // Gazovaya prom-st'. – 2006. – № 1. – S. 27–29.
9. Svidetel'stvo o gos. registratsii programmy dlya EVM 2015617924 Ros. Federatsiya. PVK "Vesta-KS" / S.A. Sardanashvili, S.K. Mitichkin, D.G. Leonov [i dr.]; pravoobladatel' OOO Firma "INGOYL". – № 2015613212; zayavl. 21.04.2015; opubl. 20.08.2015.
10. Popov R.V. Matematicheskoe i komp'yuternoe modelirovanie i optimizatsiya nestatsionarnykh rezhimov v krupnomasshtabnykh gazotransportnykh sistemakh: dis. … kand. tekhn. nauk: 05.13.18. – M., 2016. – 151 s.
11. Ot prostogo k slozhnomu: ierarkhicheskiy podkhod k raschetam magistral'nykh gazoprovodov kak slozhnykh sistem / A.S. Kazak, N.A. Kislenko, A.M. Chionov [i dr.]; pod obshch. red. A.S. Kazaka. – M.: NIIgazekonomika, 2021. – 320 s.
Back to issue