ФІЗИЧНО-ІНФОРМОВАНЕ МАШИННЕ НАВЧАННЯ ДЛЯ МОДЕЛЮВАННЯ ЯВИЩ ПЕРЕНОСУ НАНОРОЗМІРНИХ СИСТЕМ:ВИКЛИКИ, ПІДХОДИ ТА ПЕРС-ПЕКТИВИ

##plugins.themes.bootstrap3.article.main##

М. В. ГЕРКАЛЮК
А. В. ГІЛЬЧУК

Анотація

Моделювання явищ переносу в нанорозмірних системах є критично складним завданням, в якому класичні рівняння континууму виявляються неефективними, а високоточні обчислювальні системи є надто дорогими. Фізично-інформоване машинне навчання (PIML) стало революційним підходом до вирішення цієї дилеми шляхом синергетичного поєднання розрізнених експериментальних даних з основними законами транспортних моделей першого принципу. Цей огляд надає вичерпну інформацію про те, як PIML — особливо фізично-інформовані нейронні мережі (PINN), методи навчання операторів та підхід пожднання моделей різної точності — прискорює аналіз нанорозмірного транспорту від фононного транспорту на основі BTE до балістично-дифузійного теплоперенесення та ефектів випромінювання в ближньому полі. Ми розглядаємо постійні проблеми з даними в дослідженнях наноматеріалів, включаючи шумні вимірювання та формулювання високорозмірних диференціальних рівнянь з частинними похідними (PDE), і представляємо передові стратегії, такі як декомпозиція доменів та гібридні механістичні методи машинного навчання (ML), щоб підвищити гнучкість та масштабованість цих нових підходів. Нарешті, ми окреслюємо поточні прогалини в цій галузі, від кількісної оцінки невизначеності до розробки цифрових двійників у реальному часі, та окреслюємо майбутні напрямки досліджень, спрямовані на об'єднання квантових симуляцій, експериментальної метрології та глибокого навчання. Вбудовуючи фізичні обмеження безпосередньо в робочий процес навчання, ці фізично обґрунтовані методи пропонують трансформаційний шлях для оптимізації нанорозмірного транспорту, та сприятиме вдосконаленню методів дослідження наноматеріалів. Додатково ми формуємо практичну «дорожню карту» інтеграції PIML з високопродуктивними та диференційовними солверами (BTE/MC/FEM) для швидких параметричних досліджень і калібрування інтерфейсної теплопровідності. Запропоновано базові метрики, протоколи валідації та бенчмарки (TDTR, SThM, грейтингові структури) для відтворюваного порівняння PINN/операторних моделей із першим-принципним еталоном. Окрему увагу приділено стратегіям UQ (баєсівські PINN, ансамблі, багатовірогідні втрати), що є ключем до надійного впровадження цифрових двійників у реальному часі в нанотеплотехніці та наноелектроніці.

##plugins.themes.bootstrap3.article.details##

Розділ
КОМП'ЮТЕРНЕ МОДЕЛЮВАННЯ У ФІЗИЦІ
Біографії авторів

М. В. ГЕРКАЛЮК, аспірант, Фізико-технічний інститут, Національний технічний університет України «Київський політехнічний інститут імені Ігоря Сікорського», Київ, Україна

аспірант, Фізико-технічний інститут, Національний технічний університет України «Київський політехнічний інститут імені Ігоря Сікорського», Київ, Україна

А. В. ГІЛЬЧУК, кандидат фіз.-мат. наук, доцент, доцент кафедри прикладної фізики, Фізико-технічний інститут, Національний технічний університет України «Київський політехнічний інститут імені Ігоря Сікорського», Київ, Україна

кандидат фіз.-мат. наук, доцент, доцент кафедри прикладної фізики, Фізико-технічний інститут, Національний технічний університет України «Київський політехнічний інститут імені Ігоря Сікорського», Київ, Україна

Посилання

References

David G. Cahill, Paul V. Braun, Gang Chen, David R. Clarke, Shanhui Fan, Kenneth E. Goodson, Pawel Keblinski, William P. King, Gerald D. Mahan, Arun Majumdar, Humphrey J. Maris, Simon R. Phillpot, Eric Pop, Li Shi; Nanoscale thermal transport. II. 2003–2012. Appl. Phys. Rev. 1 March 2014; 1 (1): 011305. Retrieved from https://doi.org/10.1063/1.4832615

Pop, E., Sinha, S., & Goodson, K. E. (2006). Heat Generation and Transport in Nanometer-Scale Transistors. Proceedings of the IEEE, 94, 1587–1601. Retrieved from https://doi.org/10.1109/JPROC.2006.879794

Ideue, T., & Iwasa, Y. (2021). Nonlinear Electrical Transport in Van der Waals Nanomaterials. Annual Review of Condensed Matter Physics, 12, 201–223. Retrieved from https://doi.org/10.1146/annurev-conmatphys-060220-100347

Zeng, L., Collins, K. C., Hu, Y., Luckyanova, M. N., Maznev, A. A., et al. (2015). Measuring phonon mean free path distributions by probing quasiballistic phonon transport in grating nanostructures. Scientific Reports, 5, 17131. Retrieved from https://doi.org/10.1038/srep17131

Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021). Physics-informed machine learning. Nature Reviews Physics, 3, 218–229. Retrieved from https://doi.org/10.1038/s42254-021-00314-5

Wang, R., & Yu, R. (2025). Physics-Guided Deep Learning for Dynamical Systems: A Sur-vey. ACM Computing Surveys. Retrieved from https://doi.org/10.1145/3766887

Hao, Z., Liu, S., Zhang, Y., et al. (2023). Physics-Informed Machine Learning: A Survey on Problems, Methods and Applications. Tsinghua University Survey (preprint). Retrieved from https://doi.org/10.48550/arXiv.2211.08064

Farea, A., Yli-Harja, O., & Emmert-Streib, F. (2024). Understanding physics-informed neural networks: Techniques, applications, trends, and challenges. AI, 5(3), 1534–1557. Retrieved from https://doi.org/10.3390/ai5030074

Lukes, J. R., & Zhong, H. (2007). Thermal conductivity of individual single-wall carbon nanotubes. ASME J. Heat Transf., 129(705-715). Retrieved from https://doi.org/10.1115/1.2717242

Li, W., Carrete, J., Katcho, N. A., & Mingo, N. (2014). ShengBTE: A solver of the Boltz-mann transport equation for phonons. Computer Physics Communications, 185(1747-1758). Retrieved from https://doi.org/10.1016/j.cpc.2014.02.015

Majumdar, A. (1993). Microscale heat conduction in dielectric thin films. Journal of Heat Transfer, 115(7). Retrieved from https://doi.org/10.1115/1.2910673

Li, H.-L., Hua, Y.-C., & Cao, B.-Y. (2018). A hybrid phonon Monte Carlo-diffusion method for ballistic-diffusive heat conduction in nano- and micro-structures. International Journal of Heat and Mass Transfer, 127(1014-1022). Retrieved from https://doi.org/10.1016/j.ijheatmasstransfer.2018.06.080

Li, R., Wang, J.-X., Lee, E., & Luo, T. (2022). Physics-informed deep learning for solving phonon Boltzmann transport equation with large temperature non-equilibrium. npj Computa-tional Materials, 8(29). Retrieved from https://doi.org/10.1038/s41524-022-00712-y

Badar, M. S., Shamsi, S., Ahmed, J., & Alam, A. (2020). Molecular dynamics simulations: Concept, methods, and applications. In Advances in Biological Science. Springer. Retrieved from https://doi.org/10.1007/978-3-030-94651-7_7

Bock, F. E., Aydin, R. C., Cyron, C. J., Huber, N., Kalidindi, S. R., & Klusemann, B. (2019). A review of the application of machine learning and data mining approaches in continuum ma-terials mechanics. Frontiers in Materials, 6, 110. Retrieved from https://doi.org/10.3389/fmats.2019.00110

Bishara, D., Xie, Y., Liu, W. K., & Li, S. (2022). A state-of-the-art review on machine learn-ing-based multiscale modeling, simulation, homogenization, and design of materials. Archives of Computational Methods in Engineering. Retrieved from https://doi.org/10.1007/s11831-022-09795-8

Greydanus, S., Dzamba, M., & Yosinski, J. (2019). Hamiltonian Neural Networks. ArXiv. Re-trieved from https://doi.org/10.48550/arXiv.1906.01563

Cranmer, M., Greydanus, S., Hoyer, S., Battaglia, P., Spergel, D., & Ho, S. (2020). Lagrangi-an neural networks. arXiv. Retrieved from https://doi.org/10.48550/arXiv.2003.04630

Li, Z., Kovachki, N., Azizzadenesheli, K., Liu, B., Bhattacharya, K., Stuart, A., & Anandku-mar, A. (2021). Fourier neural operator for parametric partial differential equations. Proceed-ings of ICLR 2021. Retrieved from https://doi.org/10.48550/arXiv.2010.08895

Lu, L., Jin, P., Pang, G., Zhang, Z., & Karniadakis, G. E. (2021). Learning nonlinear opera-tors via DeepONet based on the universal approximation theorem of operators. Nature Ma-chine Intelligence, 3(3), 218–229. Retrieved from https://doi.org/10.1038/s42256-021-00302-5

Champion, K., Lusch, B., Kutz, J. N., & Brunton, S. L. (2019). Data-driven discovery of co-ordinates and governing equations. Proceedings of the National Academy of Sciences, 116(45), 22445–22451. Retrieved from https://doi.org/10.1073/pnas.1906995116

Chen, R. T. Q., Rubanova, Y., Bettencourt, J., & Duvenaud, D. (2018). Neural ordinary dif-ferential equations. In Advances in Neural Information Processing Systems 31 (pp. 6572–6583). Curran Associates. Retrieved from https://papers.nips.cc/paper/7892-neural-ordinary-differential-equations

Baddoo, P. J., Herrmann, B., McKeon, B. J., Kutz, J. N., & Brunton, S. L. (2023). Physics-informed dynamic mode decomposition. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 479(2271), 20220576. Retrieved from https://doi.org/10.1098/rspa.2022.0576

Jin, H., Mattheakis, M., & Protopapas, P. (2022). Physics-Informed Neural Networks for Quantum Eigenvalue Problems. 2022 International Joint Conference on Neural Networks (IJCNN), 1-8. Retrieved from https://doi.org/10.48550/arXiv.2203.00451

Baty, H. (2024). A hands-on introduction to Physics-Informed Neural Networks for solving partial differential equations with benchmark tests taken from astrophysics and plasma physics [Preprint]. HAL. Retrieved from https://hal.science/hal-04491808v1

Markidis, S. (2021). The old and the new: Can physics-informed deep-learning replace tradi-tional linear solvers? Frontiers in Big Data, 4(669097). Retrieved from https://doi.org/10.3389/fdata.2021.669097

Costabal, F. S., Pezzuto, S., & Perdikaris, P. (2023). ∆-PINNs: Physics-informed neural net-works on complex geometries. Engineering Applications of Artificial Intelligence, 118, 107324. Retrieved from https://doi.org/10.1016/j.engappai.2023.107324

Chen, S., Liu, Z., Zhang, W., & Yang, J. (2024). A hard-constraint wide-body physics-informed neural network model for solving multiple cases in forward problems for partial dif-ferential equations. Applied Sciences, 14(1), 189. Retrieved from https://doi.org/10.3390/app14010189

Krishnapriyan, A., Gholami, A., Zhe, S., Kirby, R., & Mahoney, M. W. (2021). Characteriz-ing possible failure modes in physics-informed neural networks. In Advances in Neural Infor-mation Processing Systems 34 (NeurIPS 2021). Retrieved from https://proceedings.neurips.cc/paper/2021/file/df438e5206f31600e6ae4af72f2725f1-Paper.pdf

Jagtap, A. D., Kharazmi, E., & Karniadakis, G. E. (2020). Conservative physics-informed neural networks on discrete domains for conservation laws: Applications to forward and in-verse problems. Computer Methods in Applied Mechanics and Engineering, 365, 113028. Re-trieved from https://doi.org/10.1016/j.cma.2020.113028

Rohrhofer, F. M., Posch, S., Gößnitzer, C., & Geiger, B. C. (2023). Data vs. Physics: The Apparent Pareto Front of Physics-Informed Neural Networks. IEEE Access, 11, 86252–86261. Retrieved from https://doi.org/10.1109/ACCESS.2023.3302892

Kovachki, N., Li, Z., Liu, B., Azizzadenesheli, K., Bhattacharya, K., Stuart, A., & Anandku-mar, A. (2023). Neural Operator: Learning maps between function spaces. Journal of Ma-chine Learning Research, 24, 4061-4157. Retrieved from https://doi.org/10.5555/3648699.3648788

Pang, G., Lu, L., & Karniadakis, G. E. (2019). fPINNs: fractional physics-informed neural networks. SIAM Journal on Scientific Computing, 41(4), A2603–A2626. Retrieved from https://doi.org/10.1137/18M1229845

He, J., Pal, D., Najafi, A., Abueidda, D., Koric, S., & Jasiuk, I. (2024). Material-Response-Informed DeepONet and its application to polycrystal stress–strain prediction in crystal plastic-ity. JOM, 76(10), 5744–5754. Retrieved from https://doi.org/10.1007/s11837-024-06681-5

Abbasi, J., & Andersen, P. Ø. (2024). Physical activation functions (PAFs): An approach for more efficient induction of physics into physics-informed neural networks (PINNs). Neuro-computing, 608, 128352. Retrieved from https://doi.org/10.1016/j.neucom.2024.128352

Rehman, D., & Lienhard, J. H. (2023). Physics-constrained neural differential equations for learning multi-ionic transport. ICLR 2023 Workshop on Physics for Machine Learning. Re-trieved from https://doi.org/10.48550/arXiv.2303.04594

Cai, S., Wang, Z., Wang, S., Perdikaris, P., & Karniadakis, G. E. (2021). Physics-informed neural networks (PINNs) for heat transfer problems. ASME Journal of Heat Transfer, 143(6). Retrieved from https://doi.org/10.1115/1.4050542

hou, J., Li, R., & Luo, T. (2023). Physics-informed neural networks for solving time-dependent mode-resolved phonon Boltzmann transport equation. npj Computational Materi-als, 9(212). Retrieved from https://doi.org/10.1038/s41524-023-01165-7

Faroughi, S. A., Soltanmohammadi, R., Datta, P., Mahjour, S. K., & Faroughi, S. (2024). Physics-Informed Neural Networks with Periodic Activation Functions for Solute Transport in Heterogeneous Porous Media. Mathematics, 12(63). Retrieved from https://doi.org/10.3390/math12010063

Raissi, M., Perdikaris, P., & Karniadakis, G. E. (2019). Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378(16), 686–707. Retrieved from https://doi.org/10.1016/j.jcp.2018.10.045

Dave, A. J., & Vilim, R. B. (2024). Physics-informed state-space neural networks for transport phenomena. Engineering Applications of Artificial Intelligence, 133(Part C), 108245. Re-trieved from https://doi.org/10.1016/j.engappai.2024.108245

Xiao, T., & Frank, M. (2023). RelaxNet: A structure-preserving neural network to approxi-mate the Boltzmann collision operator. Journal of Computational Physics, 490, 112317. Re-trieved from https://doi.org/10.1016/j.jcp.2023.112317

Kaheman, K., Kaiser, E., Strom, B., Kutz, J. N., & Brunton, S. L. (2019). Learning discrep-ancy models from experimental data. arXiv. Retrieved from https://doi.org/10.48550/arXiv.1909.08574

Fasel, U., Kutz, J. N., Brunton, B. W., & Brunton, S. L. (2022). Ensemble-SINDy: Robust sparse model discovery in the low-data, high-noise limit, with active learning and control. Proceedings of the Royal Society A, 478(20210904). Retrieved from https://doi.org/10.1098/rspa.2021.0904

Xu, C., Cao, B. T., Yuan, Y., & Meschke, G. (2023). Transfer learning based physics-informed neural networks for solving inverse problems in engineering structures under differ-ent loading scenarios. Computer Methods in Applied Mechanics and Engineering, 405, 115852. Retrieved from https://doi.org/10.1016/j.cma.2022.115852

Li, Y., Liu, T., & Xie, Y. (2022). Thermal fluid fields reconstruction for nanofluids convection based on physics-informed deep learning. Scientific Reports, 12, Article 12567. Retrieved from https://doi.org/10.1038/s41598-022-16463-1

Yang, H., Li, L., & Yang, P. (2025). Deep learning-based molecular dynamics simulation re-configuration of efficient heat energy transport of Gra/h-BN heterointerface. Applied Physics Letters, 127(2), 021601. Retrieved from https://doi.org/10.1063/5.0270700

Gomes, D. S., Félix, I. M., Radel, W. F., Dias, A. C., Ribeiro Jr., L. A., & Pereira Jr., M. L. (2025). Computational characterization of the recently synthesized pristine and porous 12-atom-wide armchair graphene nanoribbon. Nano Letters, 25(21), 8596–8603. Retrieved from https://doi.org/10.1021/acs.nanolett.5c01319

Li, J., Knijff, L., Zhang, Z.-Y., Andersson, L., & Zhang, C. (2025). PiNN: Equivariant neural network suite for modeling electrochemical systems. Journal of Chemical Theory and Com-putation, 21(3), 1382–1395. Retrieved from https://doi.org/10.1021/acs.jctc.4c01570

Zeng, L., Collins, K. C., Hu, Y., Luckyanova, M. N., Maznev, A. A., Huberman, S., Chiloyan, V., Zhou, J., Huang, X., Nelson, K. A., & Chen, G. (2015). Measuring phonon mean free path distributions by probing quasiballistic phonon transport in grating nanostruc-tures. Scientific Reports, 5, 17131. Retrieved from https://doi.org/10.1038/srep17131

Wu, Y.-J., Fang, L., & Xu, Y. (2019). Predicting interfacial thermal resistance by machine learning. npj Computational Materials, 5, Article 56. Retrieved from https://doi.org/10.1038/s41524-019-0193-0

Wong, J. C., Gupta, A., Ooi, C. C., Chiu, P.-H., Liu, J., & Ong, Y.-S. (2025). Evolutionary optimization of physics-informed neural networks: Evo-PINN frontiers and opportunities. arXiv. Retrieved from https://doi.org/10.48550/arXiv.2501.06572

Lin, Q., Zhang, C., Meng, X., & Guo, Z. (2025). Monte Carlo physics-informed neural net-works for multiscale heat conduction via phonon Boltzmann transport equation. Journal of Computational Physics, 542, 114364. Retrieved from https://doi.org/10.1016/j.jcp.2025.114364

Ricci, F., Chen, W., Aydemir, U. et al. (2017). An ab initio electronic transport database for inorganic materials. Sci Data 4, 170085. Retrieved from https://doi.org/10.1038/sdata.2017.85

Wu, Y.-J., Zhan, T., Hou, Z., Fang, L., & Xu, Y. (2020). Physical and chemical descriptors for predicting interfacial thermal resistance. Scientific Data, 7, 36. Retrieved from https://doi.org/10.1038/s41597-020-0373-2

Hargreaves, C.J., Gaultois, M.W., Daniels, L.M. et al. (2023). A database of experimentally measured lithium solid electrolyte conductivities evaluated with machine learning. npj Comput Mater 9, 9. Retrieved from https://doi.org/10.1038/s41524-022-00951-z

Petretto, G., Dwaraknath, S., Miranda, H. P. C., Winston, D., Giantomassi, M., van Setten, M. J., Gonze, X., Persson, K. A., Hautier, G., & Rignanese, G.-M. (2018). High-throughput den-sity-functional perturbation theory phonons for inorganic materials. Scientific Data, 5, Article 180065. Retrieved from https://doi.org/10.1038/sdata.2018.65