High-resolution reconstruction of the ablative RT instability flow field via convolutional neural networks
-
摘要:
高分辨率流场数据在气象学、航空航天工程、高能物理等领域有着重要的应用价值。实验和数值模拟是两种获取高分辨率流场数据的主要途径。但是高昂的实验成本和仿真计算资源阻碍了研究者对流场演化的具体分析。随着深度学习技术的发展，卷积神经网络被用来实现流场的高分辨率重建。针对烧蚀瑞利-泰勒不稳定性流场重建提出了普通卷积神经网络模型和多重时间路径卷积神经网络模型。这两个模型可以在很短的时间内对流场进行高分辨率重建，极大地丰富了高分辨率重建技术在流体不稳定性研究中的应用。与普通卷积神经网络相比，多重时间路径卷积神经网络模型的误差较小，可以还原流场的更多细节。此外，还讨论了用于获取低分辨率流场的不同池化方法对卷积神经网络模型性能的影响。
-
关键词:
- 卷积神经网络 /
- 烧蚀瑞利-泰勒不稳定性 /
- 高分辨率重建 /
- 多重时间路径 /
- 池化
Abstract:High-resolution flow field data has important applications in meteorology, aerospace engineering, high-energy physics and other fields. Experiments and numerical simulations are two main ways to obtain high-resolution flow field data, while the high experiment cost and computing resources for simulation hinder the specific analysis of flow field evolution. With the development of deep learning technology, convolutional neural networks are used to achieve high-resolution reconstruction of the flow field. In this paper, an ordinary convolutional neural network and a multi-time-path convolutional neural network are established for the ablative Rayleigh-Taylor instability. These two methods can reconstruct the high-resolution flow field in just a few seconds, and further greatly enrich the application of high-resolution reconstruction technology in fluid instability. Compared with the ordinary convolutional neural network, the multi-time-path convolutional neural network model has smaller error and can restore more details of the flow field. The influence of low-resolution flow field data obtained by the two pooling methods on the convolutional neural networks model is also discussed.
-
Figure 6. Comparison of the high-resolution reconstructed density data (weak nonlinear stage with ablation, disturbance wavelength=12 μm)
Figure 7. Comparison of the high-resolution reconstructed density data (classical linear stage, disturbance wavelength=12 μm)
Figure 8. Comparison of the high-resolution reconstructed density data (nonlinear stage with ablation, disturbance wavelength=12 μm)
-
[1] Adrian R J. Twenty years of particle image velocimetry[J]. Exp Fluids, 2005, 39: 159-169. doi: 10.1007/s00348-005-0991-7 [2] Yan Chao, Yu Jian, Xu Jinglei, et al. On the achievements and prospects for the methods of computational fluid dynamics[J]. Adv Mech, 2001, 41(5): 562-589. [3] Kuang Yuanyang, Lu Yan, Lin Zhi, et al. Coupled model analysis of the ablative Rayleigh–Taylor instability[J]. Plasma Sci Technol, 2023, 25: 055201. doi: 10.1088/2058-6272/acac64 [4] Fan Zhengfeng, Luo Jisheng, Ye Wenhua, et al. Weakly nonlinear ablative Rayleigh–Taylor instability at preheated ablation front[J]. Phys Plasmas, 2009, 16: 102104. doi: 10.1063/1.3236746 [5] Ye Wenhua, Wang Lifeng, He Xiantu. Spike deceleration and bubble acceleration in the ablative Rayleigh-Taylor instability[J]. Phys Plasmas, 2010, 17: 122704. doi: 10.1063/1.3497006 [6] LeCun Y, Bengio Y, Hinton G. Deep learning[J]. Nature, 2015, 521(7553): 436-444. doi: 10.1038/nature14539 [7] Zhou Feiyan, Jin Linpeng, Dong Jun. Review of convolution neural network[J]. Chin J Comput, 2017, 40(6): 1229-1251. doi: 10.11897/SP.J.1016.2017.01229 [8] Lecun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition[J]. Proc IEEE, 1998, 86(11): 2278-2324. doi: 10.1109/5.726791 [9] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[J]. Commun ACM, 2017, 60(6): 84-90. doi: 10.1145/3065386 [10] Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition[DB/OL]. arXiv preprint arXiv: 1409.1556, 2014. [11] Szegedy C, Liu Wei, Jia Yangqing, et al. Going deeper with convolutions[C]//Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2015: 1-9. [12] He Kaiming, Zhang Xiangyu, Ren Shaoqing, et al. Deep residual learning for image recognition[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2016: 770-778. [13] Jin Xiaowei, Cheng Peng, Chen Wenli, et al. Prediction model of velocity field around circular cylinder over various Reynolds numbers by fusion convolutional neural networks based on pressure on the cylinder[J]. Phys Fluids, 2018, 30: 047105. doi: 10.1063/1.5024595 [14] Sekar V, Jiang Qinghua, Shu Chang, et al. Fast flow field prediction over airfoils using deep learning approach[J]. Phys Fluids, 2019, 31: 057103. doi: 10.1063/1.5094943 [15] Fukami K, Fukagata K, Taira K. Super-resolution reconstruction of turbulent flows with machine learning[J]. J Fluid Mech, 2019, 870: 106-120. doi: 10.1017/jfm.2019.238 [16] Deng Zhiwen, He Chuangxin, Liu Yingzheng, et al. Super-resolution reconstruction of turbulent velocity fields using a generative adversarial network-based artificial intelligence framework[J]. Phys Fluids, 2019, 31: 125111. doi: 10.1063/1.5127031 [17] Liu Bo, Tang Jiupeng, Huang Haibo, et al. Deep learning methods for super-resolution reconstruction of turbulent flows[J]. Phys Fluids, 2020, 32: 025105. doi: 10.1063/1.5140772 [18] Zhou Xuhui, McClure J E, Chen Cheng, et al. Neural network–based pore flow field prediction in porous media using super resolution[J]. Phys Rev Fluid, 2022, 7: 074302. doi: 10.1103/PhysRevFluids.7.074302 [19] Jagodinski E, Zhu Xingquan, Verma S. Inverse identification of dynamically important regions in turbulent flows using three-dimensional convolutional neural networks[J]. Phys Rev Fluids, 2023, 8: 094605. doi: 10.1103/PhysRevFluids.8.094605 [20] Gallis M A, Koehler T P, Torczynski J R, et al. Direct simulation Monte Carlo investigation of the Rayleigh-Taylor instability[J]. Phys Rev Fluids, 2016, 1: 043403. doi: 10.1103/PhysRevFluids.1.043403 [21] Fu Chengquan, Zhao Zhiye, Xu Xin, et al. Nonlinear saturation of bubble evolution in a two-dimensional single-mode stratified compressible Rayleigh-Taylor instability[J]. Phys Rev Fluids, 2022, 7: 023902. doi: 10.1103/PhysRevFluids.7.023902 [22] Piriz A R, Cortázar O D, López Cela J J, et al. The Rayleigh-Taylor instability[J]. Am J Phys, 2006, 74(12): 1095-1098. doi: 10.1119/1.2358158 [23] Zhong Baojiang, Lu Zhifang, Ji Jiahuan. Review on image interpolation techniques[J]. J Data Acquis Process, 2016, 31(6): 1083-1096. [24] Shi Wenzhe, Caballero J, Huszár F, et al. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2016: 1874-1883. [25] Wang Mengjie, Yang Xiaomin, Anisetti M, et al. Image super-resolution via enhanced multi-scale residual network[J]. J Parallel Distrib Comput, 2021, 152: 57-66. doi: 10.1016/j.jpdc.2021.02.016 [26] Kingma D P, Ba J. Adam: a method for stochastic optimization[C]//Proceedings of the 3rd International Conference on Learning Representations. 2015. -
下载:
点击查看大图
计量
- 文章访问数: 116
- HTML全文浏览量: 44
- PDF下载量: 16
- 被引次数: 0
