Physically-informed neural networks for estimation of atmospheric optical turbulence intensity
Many neural networks are based on the «black box» principle, when the structure of the model does not provide opportunities for physical interpretation. In this paper, we demonstrate a methodological basis for overcoming this problem by constructing a deep neural network taking into physical restrictions imposed on the loss function. In particular, a physically informed deep neural network is proposed and analyzed that predicts variations of three-minute intensities of optical turbulence within the atmospheric surface layer at the Large Solar Vacuum Telescope site. The neural network model allows estimating the intensity of optical turbulence based on average meteorological characteristics, and also takes into account variations in time-smoothed values of the outer scale of dynamic turbulence and vertical turbulent heat flux.
Keywords
optical turbulence,
machine learning,
physically-informed neural networksAuthors
| Shikhovtsev Artem Yu. | Institute of Solar-Terrestrial Physics of the Siberian Branch of the Russian Academy of Science | Ashikhovtsev@iszf.irk.ru |
| Kiselev Alexander V. | Institute of Solar-Terrestrial Physics of the Siberian Branch of the Russian Academy of Science | kiselev@iszf.irk.ru |
| Kovadlo Pavel G. | Institute of Solar-Terrestrial Physics of the Siberian Branch of the Russian Academy of Science | kovadlo2006@rambler.ru |
Всего: 3
References
Zhivetiev I.V., Yasyukevich Yu.V. // Atmosphere. - 2022. - V. 13. - Art. 1333. - DOI: 10.3390/atmos13081333.
Wang Y., Basu S. // Opt. Lett. - 2016. - V. 41. - Iss. 10. - P. 2334 - 2337. - DOI: 10.1364/OL.41.002334.
Хабарова Т.А., Землянуха П.М., Домбек Е.М. и др. // Астрофизический бюллетень. - 2024. - Т. 79. - № 2. - С. 350-360.
Shikhovtsev A.Y., Kiselev A.V., Kovadlo P.G., et al. // Atmosphere. - 2024. - V. 15. - Art. 38. - DOI: 10.3390/atmos15010038.
Shikhovtsev A.Y., Kovadlo P.G., Kiselev A.V., et al. // Publ. Astron. Soc. Pac. - 2023. - V. 135. - Art. 014503. - DOI: 10.1088/1538-3873/acb384.
Zhou S., Shi R., Yu H., et al. // JGR Atmospheres. - 2024. - V. 129. - Iss. 17. - Art. e2023JD040603. - DOI: 10.1029/2023JD040603.
Pettit C.L., Wilson D.K. // Proc. Meetings Acoust. - 2020. - V. 42. - Iss. 1. - Art. 022002. - DOI: 10.1121/2.0001383.
Soto A.M., Cervantes A., Soler M. // Open Res. Eur. - 2024. - V. 4. - Art. 99. - DOI: 10.12688/openreseurope.17388.1.
Varey J., Ruprecht J.D., Tierney M., Sullenberger R. // IEEE Aerospace Conference Proceedings. - 2024. - P. 1-8. - DOI: 10.1109/aero58975.2024.10521414.
Badura G.P., Valenta C.R. // J. Astronautical Sci. - 2024. - V. 71. - No. 5. - P. 41. - DOI: 10.12688/openreseurope.17388.1.
Avila R., Zúñiga C.A., Tapia-Rodríguez J.J., et al. // Publ. Astron. Soc. Pac. - 2016. - V. 128. - Art. 105001. - DOI: 10.1088/1538-3873/128/968/105001.
Simakhin V.A., Potekaev A.I., Cherepanov O.S., Shamanaeva L.G. // Appl. Sci. - 2023. - V. 13. - Art. 6116. - DOI: 10.3390/app13106116.
Шиховцев А.Ю. // Оптика атмосферы и океана. - 2022. - Т. 35. - № 1. - С. 74-80. - DOI: 10.15372/AOO20220111.
Laidlaw D.J., Osborn J., Morris T.J., et al. // Mon. Not. R. Astron. Soc. - 2019. - V. 483. - Iss. 4. - P. 4341-4353.- DOI: 10.1093/mnras/sty3285.
Wang Y., Xu H., Li D., et al. // Sci. Rep. - 2018. - V. 8. - Art. 1124. - DOI: 10.1038/s41598-018-19559-9.
Gladkikh V.A., Makienko A.E. // Pribory. - 2009. - V. 7. - P. 21-25.
Nosov V., Lukin V., Nosov E., et al. // Atmosphere. - 2019. - V. 10. - Art. 460. - DOI: 10.3390/atmos10080460.
Shikhovtsev A., Kovadlo P., Lukin V., et al. // Atmosphere. - 2019. - V. 10. - Art. 661. - DOI: 10.3390/atmos10110661.
Shikhovtsev A.Yu. // Publ. Astron. Soc. Jpn. - 2024. - V. 76. - Iss. 3. - P. 538-549. - DOI: 10.1093/pasj/psae031.
Xu M., Shao S., Weng N., et al. // Remote Sens. - 2022. - V. 14. - Art. 3085. - DOI: 10.3390/rs14133085.
Molozhnikova Ye.V., Shikhovtsev M.Yu., Netsvetaeva O.G., et al. // Appl. Sci. - 2023. - V. 13. - Art. 8171. - DOI: 10.3390/app13148171.
