Research output: Contribution to journal › Article › peer-review
Мeasurement of Dry Spot Features during Boiling Using Neural Network Processing of High-Speed Visualization. / Surtaev, A. S.; Perminov, P. O.; Malakhov, I. P. et al.
In: Thermal Engineering, Vol. 72, No. 6, 27.06.2025, p. 473-482.Research output: Contribution to journal › Article › peer-review
}
TY - JOUR
T1 - Мeasurement of Dry Spot Features during Boiling Using Neural Network Processing of High-Speed Visualization
AU - Surtaev, A. S.
AU - Perminov, P. O.
AU - Malakhov, I. P.
AU - Polovnikov, M. A.
AU - Chernyavskiy, A. N.
N1 - The work was carried out in accordance with the state assignment of the Kutateladze Institute of Thermophysics, Siberian Branch, Russian Academy of Sciences (no. 121031800216-1 January 1, 2021). Мeasurement of Dry Spot Features during Boiling Using Neural Network Processing of High-Speed Visualization / A. S. Surtaev, P. O. Perminov, I. P. Malakhov [et al.] // Thermal Engineering. – 2025. – Vol. 72, No. 6. – P. 473-482. – DOI 10.1134/S004060152570017X.
PY - 2025/6/27
Y1 - 2025/6/27
N2 - Abstract: It is known that dry spots formed under vapor bubbles during the boiling process have a huge impact on both local heat transfer and the development of crisis phenomena. In this study, new experimental information on the evolution of dry spots under vapor bubbles during liquid boiling was obtained using high-speed reflected light imaging, and an algorithm for automatic processing of experimental data based on U-Net convolutional neural networks was developed. It is shown that it is possible using machine learning models and high-precision optical high-speed methods to determine a wide range of characteristics of dry spots during liquid boiling in a short period of time and with high accuracy, including the evolution of the total area and size of dry spots, total number, and the growth rate and lifetimes of dry spots in a wide range of heat fluxes. Based on the analysis of the collected data, it was established that the average total area of dry spots and the nucleation site density during boiling of water increase linearly with increasing heat flux in the studied range. It has been demonstrated that the growth rate of dry spots is constant in the period before the onset of the bubble detachment stage, with the average value of this rate increasing with increasing heat flux. The characteristic maximum size of dry spots turns out to be almost half the capillary length. The results obtained, presented in the article, indicate that there is a huge potential for using artificial intelligence methods, which open up new prospects for studying two-phase systems, modeling heat transfer during boiling, and predicting crisis phenomena associated with uncontrolled growth of dry spots.
AB - Abstract: It is known that dry spots formed under vapor bubbles during the boiling process have a huge impact on both local heat transfer and the development of crisis phenomena. In this study, new experimental information on the evolution of dry spots under vapor bubbles during liquid boiling was obtained using high-speed reflected light imaging, and an algorithm for automatic processing of experimental data based on U-Net convolutional neural networks was developed. It is shown that it is possible using machine learning models and high-precision optical high-speed methods to determine a wide range of characteristics of dry spots during liquid boiling in a short period of time and with high accuracy, including the evolution of the total area and size of dry spots, total number, and the growth rate and lifetimes of dry spots in a wide range of heat fluxes. Based on the analysis of the collected data, it was established that the average total area of dry spots and the nucleation site density during boiling of water increase linearly with increasing heat flux in the studied range. It has been demonstrated that the growth rate of dry spots is constant in the period before the onset of the bubble detachment stage, with the average value of this rate increasing with increasing heat flux. The characteristic maximum size of dry spots turns out to be almost half the capillary length. The results obtained, presented in the article, indicate that there is a huge potential for using artificial intelligence methods, which open up new prospects for studying two-phase systems, modeling heat transfer during boiling, and predicting crisis phenomena associated with uncontrolled growth of dry spots.
KW - U-Net
KW - convolutional neural networks
KW - dry spot evolution
KW - high-speed imaging
KW - liquid boiling
KW - tracking algorithm
UR - https://www.mendeley.com/catalogue/32c81575-0fb6-379e-9829-6106a4d11359/
UR - https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=105009347186&origin=inward
UR - https://www.elibrary.ru/item.asp?id=82550132
U2 - 10.1134/S004060152570017X
DO - 10.1134/S004060152570017X
M3 - Article
VL - 72
SP - 473
EP - 482
JO - Thermal Engineering (English translation of Teploenergetika)
JF - Thermal Engineering (English translation of Teploenergetika)
SN - 0040-6015
IS - 6
ER -
ID: 68294785