NEURAL NETWORK METHOD OF PREDICTIVE CONTROL IN MICROGRIDS WITH MECHATRONIC WIND-GENERATOR SYSTEM
Abstract
The influence of various factors on the accuracy of forecasting wind turbine generator (WTG) generation is considered. The optimal set of input parameters (day, month, time, wind speed, air temperature, atmospheric pressure and estimated power output of wind turbine) for forecasting is determined, and the methods of their processing are substantiated. The influence of influencing factors on the accuracy of forecasting the generated power of wind turbines was investigated. Profiles of input data for forecasting the power generation of wind power plants are constructed. The peculiarities of meteorological conditions for a year are considered, frequently occurring wind speed values are determined, etc., for the selection of an optimal wind turbine. It is shown that the meteorological conditions meet the passport requirements of the WTG selected for the region under consideration. Neural network (NN) models for forecasting the power generation of wind turbines are considered, the optimal NN is selected, the structure is built and the algorithm of NN for forecasting the generated power of wind turbines is developed. The developed mathematical model of wind power generation is aimed at improving accuracy and adaptability by taking into account key dynamic factors (wind speed and change in wind direction, air temperature and density, etc.). The combined wind turbine generation control method (MPPT + Pitch) is chosen to ensure a balance between efficiency and safety. The combined method of controlling the wind turbine generation (MPPT + Pitch) is chosen, which provides a balance between efficiency and safety. Based on the estimated generated power of wind turbines, and meteorological conditions at the location, the neural network model showed high accuracy in predicting the power of wind turbines. It is shown that the selected type of wind turbine combines technological reliability, cost-effectiveness and compliance with modern trends in wind energy. The NN model allows maintaining a balance between generated and consumed electricity, and, consequently, increases efficiency, reduces parasitic losses in the microgrid, and reduces wear and tear of equipment.
References
1. Shepelev A.O., Artamonova E.Yu. Klassifikatsiya sovremennykh vetroenergeticheskikh ustanovok po moshchnosti [Classification of modern wind power plants by power], Molodoy uchenyy [Young Scien-tist], 2016, No. 17, pp. 92-96.
2. Fazylova A.R. Razrabotka blokov upravleniya dlya vetrogeneratorov: spetsial'nost' 47.04.01. «Filosofi-ya»: diss. … d-ra filos. nauk [Development of control units for wind generators: specialty 47.04.01. "Philosophy": dr. of philos. sc.] Almaty: Kazakhskiy natsional'nyy issledovatel'skiy tekhnicheskiy uni-versitet imeni K.I. Satpaeva 2024, 121 p.
3. Orlov D.V. Razrabotka i issledovanie metodov prediktivnogo upravleniya moshchnost'yu vetro-elektrostantsiy v elektroenergeticheskikh sistemakh: spetsial'nost' 2.4.3 «Elektroenergetika»: diss. … kand. tekhn. nauk [Development and research of methods for predictive control of wind power plants in electric power systems: specialty 2.4.3 "Electric Power Engineering": cand. of eng. sc. diss.]. Novosi-birsk: Novosibirskiy gosudarstvennyy tekhnicheskiy universitet, 2023, 224 p.
4. Özkan, Mehmet & Karagoz, Pinar. (2019). Data Mining Based Upscaling Approach for Regional Wind Power Forecasting: Regional Statistical Hybrid Wind Power Forecast Technique (RegionalSHWIP), IEEE Access, pp. 1-1. 10.1109/ACCESS.2019.2956203.
5. Gorshenin A.Yu., Denisova L.A. Prognozirovanie vyrabotki elektroenergii vetroelektrostantsiey s prime-neniem rekurrentnoy neyronnoy seti [Forecasting the electricity generation of a wind power plant using a recurrent neural network], Izvestiya TulGU. Tekhnicheskie nauki [Bulletin of Tula State University. Technical sciences], 2023, No. 4. Available at: https://cyberleninka.ru/article/n/prognozirovanie-vyrabotki-elektroenergii-vetroelektrostantsiey-s-primeneniem-rekurrentnoy-neyronnoy-seti (accessed 20 February 2025).
6. Zubova N.V. Prognozirovanie vyrabatyvaemoy moshchnosti vetroenergeticheskoy ustanovki s pomoshch'yu nechetkikh neyronnykh setey [Forecasting the generated power of a wind turbine using fuzzy neural networks], Problemy nauki [Problemy nauki], 2016, No. 32 (74). Available at: https://cyberleninka.ru/article/n/prognozirovanie-vyrabatyvaemoy-moschnosti-vetroenergeticheskoy-ustanovki-s-pomoschyu-nechetkih-neyronnyh-setey (accessed 22 February 2025).
7. Li Huajin. Short-Term Wind Power Prediction via Spatial Temporal Analysis and Deep Residual Net-works, Frontiers in Energy Research, 2022. 10. 920407. 10.3389/fenrg.2022.920407.
8. Jamii Jannet & Mansouri Majdi & Trabelsi Mohamed & Mimouni Mohamed & Shatanawi Wasfi. Ef-fective artificial neural network-based wind power generation and load demand forecasting for optimum energy management, Frontiers in Energy Research, 2022. 10. 898413. 10.3389/fenrg.2022.898413.
9. Liu Shu & Wang Lei & Jiang Hongliang & Liu Yan & You Hongyu. Wind Farm Energy Storage System Based on Cat Swarm Optimization–Backpropagation Neural Network Wind Power Prediction, Frontiers in Energy Research, 2022. 10. 10.3389/fenrg.2022.850295.
10. Kharlashkin D.A. Prognozirovanie vyrabotki vetroenergeticheskikh ustanovok na osnove metodov is-kusstvennogo intellekta: spetsial'nost' 1.4.4 «Prikladnaya matematika» [Forecasting the output of wind power plants based on artificial intelligence methods: specialty 1.4.4 "Applied Mathematics"]. Ekaterin-burg: Ural'skiy federal'nyy universitet imeni pervogo Prezidenta Rossii B.N. El'tsina, 2024, 83 p.
11. Gorshenin A.Yu., Vasina D.I. Sravnenie ispol'zuemykh metodov pri prognozirovanii vyrabotki elektro-energii vetroelektrostantsiyami [Comparison of the methods used in forecasting electricity generation by wind power plants], MSiM [MsiM], 2023, No. 3 (67). Available at: https://cyberleninka.ru/ arti-cle/n/sravnenie-ispolzuemyh-metodov-pri-prognozirovanii-vyrabotki-elektroenergii-vetroelektrostantsiyami (accessed 22 February 2025).
12. Manusov V.Z., Kholdonov A.A., Boyko K.N., Khaldarov Sh.K. Povyshenie stepeni integratsii vetroener-geticheskikh stantsiy v energosistemu putem ispol'zovaniya u sistemnogo operatora matematicheskikh modeley prognozirovaniya vetra i elektropotrebleniya [Increasing the degree of integration of wind pow-er stations into the energy system by using mathematical models for forecasting wind and electricity con-sumption by the system operator], Problemy regional'noy energetiki [Problems of regional energy], 2017, No. 3 (35).
13. Lalayan A.G., Bolbakov R.G., Bratus' N.V. Imitatsionnaya model' obrabotki dannykh vetroelektrostantsii na osnove neyronnoy seti [Simulation model for processing wind power plant data based on a neural network], International Journal of Open Information Technologies [International Journal of Open Infor-mation Technologies], 2022, No. 5.
14. Poluyanovich N.K., Dubyago M.N. Otsenka vozdeystvuyushchikh faktorov i prognozirovanie elektropo-trebleniya v regional'noy energosisteme s uchetom rezhima ee ekspluatatsii [Assessment of influencing factors and forecasting of electricity consumption in a regional power system taking into account its op-erating mode], Izvestiya YuFU. Tekhnicheskie nauki [Izvestiya SFedU. Engineering Sciences], 2022, No. 2 (226), pp. 31-46.
15. Dubyago M.N., Poluyanovich N.K. Sovershenstvovanie metodov diagnostiki i prognozirovaniya el-ektroizolyatsionnykh materialov sistem energosnabzheniya: monografiya [Improving methods for diag-nostics and forecasting of electrical insulating materials for power supply systems: monograph]. Rostov-on-Don; Taganrog: Izd-vo YuFU, 2019, 192 p.
16. Poluyanovich N.K., Bur'kov D.V., Dubyago M.N., Neyrosetevoy N.V. Metod prognozirovaniya elektrop-otrebleniya i ego instrumental'naya realizatsiya: monografiya [Method of forecasting electricity consump-tion and its instrumental implementation: monograph]. Rostov-on-Don; Taganrog: Izd-vo YuFU, 2023, 151 p.
17. Kostjukov V.A., Maevskiy A.M., Poluyanovich N.K., Dubyago M.N. Adaptive mechatronic management system of wind-driven powerplant with variable geometry, 18th International Conference of Young Spe-cialists on Micro/Nanotechnologies and Electron Devices EDM 2017: Conference Proceedings, 2017, pp. 460-464.
18. Kostyukov V.A., Medvedev M.Yu., Maevskiy A.M., Poluyanovich N.K., Savchenko V.V. Issledovanie per-spektivnoy vetroenergeticheskoy ustanovki s tipom komponovki "Rotor v rastrube" [Study of a promis-ing wind power plant with a "Rotor in a socket" configuration], Vestnik Donskogo gosudarstvennogo tekhnicheskogo universiteta [Bulletin of the Don State Technical University], 2017, Vol. 17, No. 1 (88), pp. 85-91.
19. Kostyukov V.A., Medvedev M.YU., Maevskiy A.N., Poluyanovich N.K., Savchenko V.V. Optimizatsiya form geometrii rastruba vetroenergeticheskoy ustanovki tipa "Rotor v rastrube" [Optimization of the ge-ometry forms of the bell of a wind turbine of the "Rotor in a bell" type], Vestnik Donskogo gosudar-stvennogo tekhnicheskogo universiteta [Bulletin of the Don State Technical University], 2017, Vol. 17, No. 4 (91), pp. 61-68.
20. Poluyanovich N.K., Azarov N.V., Dubyago M.N. Neyrokomp'yuternoe upravlenie propusknoy sposob-nost'yu kabel'nykh setey posredstvom ucheta i kontrolya ikh parametrov [Neurocomputer control of the bandwidth of cable networks by means of accounting and control of their parameters], Izvestiya YuFU. Tekhnicheskie nauki [Izvestiya SFedU. Engineering Sciences], 2022, No. 3 (227), pp. 84-103.
