한국가시화정보학회지 (Journal of the Korean Society of Visualization)

  • 제21권1호
  • /
  • Pages.18-25
  • /
  • 2023
  • /
  • 1598-8430(pISSN)
  • /
  • 2093-808X(eISSN)
한국가시화정보학회 (The Korean Society of Visualization)
권호 표지
DOI QR코드

DOI QR Code

합성곱 신경망 기반 선체 표면 유동 속도의 픽셀 수준 예측

Pixel-level prediction of velocity vectors on hull surface based on convolutional neural network

  • 투고 : 2022.11.04
  • 심사 : 2023.01.14
  • 발행 : 2023.03.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

In these days, high dimensional data prediction technology based on neural network shows compelling results in many different kind of field including engineering. Especially, a lot of variants of convolution neural network are widely utilized to develop pixel level prediction model for high dimensional data such as picture, or physical field value from the sensors. In this study, velocity vector field of ideal flow on ship surface is estimated on pixel level by Unet. First, potential flow analysis was conducted for the set of hull form data which are generated by hull form transformation method. Thereafter, four different neural network with a U-shape structure were conFig.d to train velocity vectors at the node position of pre-processed hull form data. As a result, for the test hull forms, it was confirmed that the network with short skip-connection gives the most accurate prediction results of streamlines and velocity magnitude. And the results also have a good agreement with potential flow analysis results. However, in some cases which don't have nothing in common with training data in terms of speed or shape, the network has relatively high error at the region of large curvature.

키워드

과제정보

이 논문은 2022년도 정부(과학기술정보통신부)의 재원으로 한국 연구재단의 지원을 받아 수행된 연구임(No. 2022R1A2C2010821)

참고문헌

  1. Ronneberger, Olaf, Philipp Fischer, and Thomas Brox. "U-net: Convolutional networks for biomedical image segmentation." International Conference on Medical image computing and computer-assisted intervention. Springer, Cham, 2015.
  2. Thuerey, N., Weissenow, K., Prantl, L., Hu, X., 2020, Deep Learning Methods for Reynolds-Averaged Navier-Stokes Simulations of Airfoil Flows, AIAA
  3. Kim, H. J., Choi, H. J., Chun, H. H., 2007, A Practical Hull Form Optimization Method Using the Parametric Modification Function. Journal of the Society of Naval Architects of Korea. 44(5), 542-550. https://doi.org/10.3744/SNAK.2007.44.5.542
  4. Hoyle, J. W., Cheng, B. H., Hays, B., 1986, A Bulbous Bow Design Methodology for High-Speed Ships. Transactions of Society of Naval Architects and Marine Engineers. 94, 31-56.
  5. Wigley, W.C.S., 1936, The theory of the bulbous bow and its practical application. Transactions of the North East Coast Institution of Engineers and Shipbuilders. 52.
  6. Thomson, G. R., White, G. P., 1969, Model experiments with stern variations of a 0.65 block coefficient form. Transactions of the Royal Institution of Naval Architects. 111, 299-316.
  7. Dawson, J., Thomson, G. R., 1969, Model Experiments with Stern Variations of a 0.80 Block Coefficient Form. Transactions of the Royal Institution of Naval Architects. 111, 507-524.
  8. LeCun, Yann, et al. "Gradient-based learning applied to document recognition." Proceedings of the IEEE 86.11 (1998): 2278-2324. https://doi.org/10.1109/5.726791
  9. Alzubaidi, Laith, et al. "Review of deep learning: Concepts, CNN architectures, challenges, applications, future directions." Journal of big Data 8.1 (2021): 1-74. https://doi.org/10.1186/s40537-021-00444-8
  10. He, Kaiming, et al. "Deep residual learning for image recognition." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.