-
特征识别是流体力学的重要研究方向, 然而在中高雷诺数情况下物体的尾流流场复杂, 难以通过传统方法实现特征的提取与识别. 深度学习理论与技术的不断发展为复杂流场特征的识别提供了新方法. 基于流场时程数据的深度学习模型, 本文研究了4种模型对尾流场特征提取与识别的精度, 得到了针对流场时程特征提取的高精度新方法. 结果表明: 所提出的模型能够识别尾流物理时程的不同特征, 并通过流场时程实现了目标的外形识别, 验证了方法的可行性; 同时结果表明基于卷积运算的深度学习模型精度高, 适用于流场时程数据的特征分析; 深度学习网络结构更深、层间结构复杂的残差卷积网络识别精度最高, 是尾流时程分析的高精度算法. 本文所提方法从流场物理量时程的角度对流场特征进行了提取与识别, 证明了深度学习方法具有较高的识别精度, 是研究流场特征的重要途径.Extraction and recognition of the features of flow field is an important research area of fluid mechanics. However, the wake flow field of object immersed in fluid is complicated in the case of medium- and high-Reynolds number, thus it is difficult to extract and recognize the key features by using traditional physical models and mathematical methods. The continuous development of deep learning theory provides us with a new method of recognizing the complex flow features. A new method of extracting the features of the flow time history is proposed based on deep learning in this work. The accuracy of four deep learning model for feature recognition is studied. The results show that the proposed model can identify different characteristics of the wake time history and object shapes accurately. Some conclusions can be obtained below (i) The model based on convolutional layers has higher accuracy and is suitable for analyzing the features of flow time history data. (ii) The residual convolutional network, with a deeper structure and more complex inter-layer structure, has highest accuracy for feature recognition. (iii) The proposed method can extract and recognize the flow features from the perspective of physical quantities time history, which is a high-accuracy method, and it is an important new way to study the features of flow physical quantities.
-
Keywords:
- flow feature extraction /
- deep learning /
- flow time history /
- residual convolution network /
- feature identification
[1] 叶舒然, 张珍, 王一伟, 黄晨光 2021 航空学报 42 185
Google Scholar
Ye S R, Zhang Z, Wang Y W, Huang C G 2021 Acta Aeronaut. Astronaut. Sin. 42 185
Google Scholar
[2] 王义乾, 桂南 2019 水动力学研究与进展(A辑) 34 413
Google Scholar
Wang Y Q, Gui N 2019 J. Hydrodyn. 34 413
Google Scholar
[3] 刘超群 2020 空气动力学学报 38 413
Google Scholar
Liu C Q 2020 Acta Aerodyn. Sin. 38 413
Google Scholar
[4] 王怡星, 韩仁坤, 刘子扬, 张扬, 陈刚 2021 航空学报 42 231
Google Scholar
Wang Y X, Qian R K, Liu Z Y, Zhang Y, Chen G 2021 Acta Aeronaut. Astronaut. Sin. 42 231
Google Scholar
[5] 任峰, 高传强, 唐辉 2021 航空学报 42 152
Google Scholar
Ren F, Gao C Q, Tang H 2021 Acta Aeronaut. Astronaut. Sin. 42 152
Google Scholar
[6] 王年华, 鲁鹏, 常兴华, 张来平 2021 力学学报 53 740
Google Scholar
Wang N H, Lu P, Chang X H, Zhang L P 2021 Chinese Journal of Theoretical and Applied Mechanics 53 740
Google Scholar
[7] Ling J, Kurzawski A, Templeton J 2016 J. Fluid Mech. 807 155
Google Scholar
[8] Maulik R, San O, Jacob J D, Crick C 2019 J. Fluid Mech. 870 784
Google Scholar
[9] Ren F, Wang C, Tang H 2019 Phys. Fluids 31 093601
Google Scholar
[10] Ren F, Wang C, Tang H 2021 Phys. Fluids 33 093602
Google Scholar
[11] Huang J, Liu H, Cai W 2019 J. Fluid Mech 875 R2
Google Scholar
[12] Zhang Y, Azman A N, Xu K W, Kim H B 2020 Exp. Fluids 61 1
Google Scholar
[13] Han J, Tao J, Wang C 2018 IEEE Trans. Visual. Comput. Graphics 26 1732
Google Scholar
[14] Liu Y, Lu Y, Wang Y, Sun D, Deng L, Wang F, Lei Y 2019 Comput. Fluids 184 1
Google Scholar
[15] Zhang Y, Azman A N, Xu K W, Kang C, Kim H B 2020 Experiments in Fluids 61 1
[16] Strfer C A M, Wu J, Xiao H, Paterson E 2018 Commun. Comput. Phys. 25 625
Google Scholar
[17] Murata T, Fukami K, Fukagata K 2020 J. Fluid Mech. 882 A13
Google Scholar
[18] Omata N, Shirayama S 2019 AIP Adv. 9 015006
Google Scholar
[19] Kai F, Nakamura T, Fukagata K 2020 Phys. Fluids 32 095110
Google Scholar
[20] He K, Zhang X, Ren S, Sun J 2016 European Conference on Computer Vision Amsterdam, Netherlands, October 11–14, 2016 630
[21] Szegedy C, Ioffe S, Vanhoucke V, Alemi A 2017 Thirty-First AAAI Conference on Artificial Intelligence San Francisco, California, USA, February 4–9, 2017
[22] 刘芙伶 李伟红 龚卫国 2020 计算机辅助设计与图形学学报 32 150
Google Scholar
Liu F L, Li W H, Gong W G 2020 CAD & CG 32 150
Google Scholar
[23] 郑天韵, 王圣业, 王光学, 邓小刚 2020 69 204701
Google Scholar
Zheng T Y, Wang S Y, W ang G X, Deng X G 2020 Acta Phys. Sin. 69 204701
Google Scholar
[24] Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R 2014 J. Mach. Learn. Res. 15 1929
Google Scholar
[25] Krizhevsky A, Sutskever I, Hinton G E 2017 Commun. ACM 60 84
Google Scholar
[26] Wang Z, Yan W, Oates T 2017 International Joint Conference on Neural Networks (IJCNN) Anchorage, Alaska, USA, May 14–19, 2017 p1578
[27] Ioffe S 2017 Proceedings of the 31st International Conference on Neural Information Processing Systems Long Beach, California, USA, December 4–9, 2017 p1942
[28] He K, Zhang X, Ren S, Sun J 2016 Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Las Vegas, Nevada, USA, June 27–30, 2016 p770
[29] 战庆亮, 周志勇, 葛耀君 2015 哈尔滨工业大学学报 47 75
Google Scholar
Zhan Q L, Zhou Z Y, Ge Y J 2015 J. Harbin Inst. Technol. 47 75
Google Scholar
-
图 6 各形状棱柱的瞬态流场云图 (a) 三棱柱压力云图; (b) 方柱压力云图; (c) 六棱柱压力云图; (d) 三棱柱速度云图; (e) 方柱速度云图; (f) 六棱柱速度云图
Fig. 6. Transient wake contour of prisms with different shapes: (a) Pressure contour of triangular prism; (b) pressure contour of square cylinder; (c) pressure contour of hexagonal prism; (d) velocity contour of triangular prism; (e) velocity contour of square cylinder; (f) velocity contour of hexagonal prism.
图 8 典型测点的流场参数时程 (a)三棱柱; (b) 方柱; (c) 六棱柱
Fig. 8. Time history of flow field parameters at typical measuring points: (a) Triangular prism; (b) square cylinder; (c) hexagonal prism
图 9 训练集的模型损失值 (a) 压力; (b) 速度
Fig. 9. Loss function of different models on training set: (a) Pressure; (b) velocity.
图 10 训练集的模型精度 (a) 压力; (b) 速度
Fig. 10. Accuracy curve of different models on training set: (a) Pressure; (b) velocity.
图 11 验证集的最优模型结果 (a) 模型损失值; (b) 模型准确率
Fig. 11. Summary of best model on the validation set: (a) Model loss value; (b) model accuracy
图 12 压力时程的识别结果散点图 (a1)—(a4) 分别为MLP, TCNN, FCNN, RCNN第一类结果; (b1)—(b4) 分别为MLP, TCNN, FCNN, RCNN第二类结果; (c1)—(c4) 分别为MLP, TCNN, FCNN, RCNN第三类结果
Fig. 12. Identification results of pressure time history: (a1)–(a4) MLP, TCNN, FCNN, RCNN results of class1; (b1)–(b4) MLP, TCNN, FCNN, RCNN results of class2; (c1)–(c4) MLP, TCNN, FCNN, RCNN results of class1.
图 13 速度时程的识别结果散点图 (a1)—(a4) 分别为MLP, TCNN, FCNN, RCNN第一类结果; (b1)—(b4) 分别为MLP, TCNN, FCNN, RCNN第二类结果; (c1)—(c4) 分别为MLP, TCNN, FCNN, RCNN第三类结果
Fig. 13. Identification results of velocity time history: (a1)–(a4) MLP, TCNN, FCNN, RCNN results of class1; (b1)–(b4) MLP, TCNN, FCNN, RCNN results of class2; (c1)–(c4) MLP, TCNN, FCNN, RCNN results of class1.
表 1 一维卷积神经网络模型参数
Table 1. Structural parameters of capillary of different kind of fluid.
名称 特征提取运算 加速收敛方法 网络层数 模型参数个数 MLP 全连接层 Dropout 10 662501 TCNN 卷积层 局部池化 8 961 FCNN 卷积层 归一化层 13 264833 RCNN 卷积层 残差直连层 43 504129 
下载: 导出CSV
-
[1] 叶舒然, 张珍, 王一伟, 黄晨光 2021 航空学报 42 185
Google Scholar
Ye S R, Zhang Z, Wang Y W, Huang C G 2021 Acta Aeronaut. Astronaut. Sin. 42 185
Google Scholar
[2] 王义乾, 桂南 2019 水动力学研究与进展(A辑) 34 413
Google Scholar
Wang Y Q, Gui N 2019 J. Hydrodyn. 34 413
Google Scholar
[3] 刘超群 2020 空气动力学学报 38 413
Google Scholar
Liu C Q 2020 Acta Aerodyn. Sin. 38 413
Google Scholar
[4] 王怡星, 韩仁坤, 刘子扬, 张扬, 陈刚 2021 航空学报 42 231
Google Scholar
Wang Y X, Qian R K, Liu Z Y, Zhang Y, Chen G 2021 Acta Aeronaut. Astronaut. Sin. 42 231
Google Scholar
[5] 任峰, 高传强, 唐辉 2021 航空学报 42 152
Google Scholar
Ren F, Gao C Q, Tang H 2021 Acta Aeronaut. Astronaut. Sin. 42 152
Google Scholar
[6] 王年华, 鲁鹏, 常兴华, 张来平 2021 力学学报 53 740
Google Scholar
Wang N H, Lu P, Chang X H, Zhang L P 2021 Chinese Journal of Theoretical and Applied Mechanics 53 740
Google Scholar
[7] Ling J, Kurzawski A, Templeton J 2016 J. Fluid Mech. 807 155
Google Scholar
[8] Maulik R, San O, Jacob J D, Crick C 2019 J. Fluid Mech. 870 784
Google Scholar
[9] Ren F, Wang C, Tang H 2019 Phys. Fluids 31 093601
Google Scholar
[10] Ren F, Wang C, Tang H 2021 Phys. Fluids 33 093602
Google Scholar
[11] Huang J, Liu H, Cai W 2019 J. Fluid Mech 875 R2
Google Scholar
[12] Zhang Y, Azman A N, Xu K W, Kim H B 2020 Exp. Fluids 61 1
Google Scholar
[13] Han J, Tao J, Wang C 2018 IEEE Trans. Visual. Comput. Graphics 26 1732
Google Scholar
[14] Liu Y, Lu Y, Wang Y, Sun D, Deng L, Wang F, Lei Y 2019 Comput. Fluids 184 1
Google Scholar
[15] Zhang Y, Azman A N, Xu K W, Kang C, Kim H B 2020 Experiments in Fluids 61 1
[16] Strfer C A M, Wu J, Xiao H, Paterson E 2018 Commun. Comput. Phys. 25 625
Google Scholar
[17] Murata T, Fukami K, Fukagata K 2020 J. Fluid Mech. 882 A13
Google Scholar
[18] Omata N, Shirayama S 2019 AIP Adv. 9 015006
Google Scholar
[19] Kai F, Nakamura T, Fukagata K 2020 Phys. Fluids 32 095110
Google Scholar
[20] He K, Zhang X, Ren S, Sun J 2016 European Conference on Computer Vision Amsterdam, Netherlands, October 11–14, 2016 630
[21] Szegedy C, Ioffe S, Vanhoucke V, Alemi A 2017 Thirty-First AAAI Conference on Artificial Intelligence San Francisco, California, USA, February 4–9, 2017
[22] 刘芙伶 李伟红 龚卫国 2020 计算机辅助设计与图形学学报 32 150
Google Scholar
Liu F L, Li W H, Gong W G 2020 CAD & CG 32 150
Google Scholar
[23] 郑天韵, 王圣业, 王光学, 邓小刚 2020 69 204701
Google Scholar
Zheng T Y, Wang S Y, W ang G X, Deng X G 2020 Acta Phys. Sin. 69 204701
Google Scholar
[24] Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R 2014 J. Mach. Learn. Res. 15 1929
Google Scholar
[25] Krizhevsky A, Sutskever I, Hinton G E 2017 Commun. ACM 60 84
Google Scholar
[26] Wang Z, Yan W, Oates T 2017 International Joint Conference on Neural Networks (IJCNN) Anchorage, Alaska, USA, May 14–19, 2017 p1578
[27] Ioffe S 2017 Proceedings of the 31st International Conference on Neural Information Processing Systems Long Beach, California, USA, December 4–9, 2017 p1942
[28] He K, Zhang X, Ren S, Sun J 2016 Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Las Vegas, Nevada, USA, June 27–30, 2016 p770
[29] 战庆亮, 周志勇, 葛耀君 2015 哈尔滨工业大学学报 47 75
Google Scholar
Zhan Q L, Zhou Z Y, Ge Y J 2015 J. Harbin Inst. Technol. 47 75
Google Scholar
下载:
计量
- 文章访问数: 7244
- PDF下载量: 246
- 被引次数: 0
