Abstract
Accurate extraction of features in fluid flows is of importance due to the presence in many natural and technological systems. Recently, methods based on machine learning have emerged as an alternative to traditional Eulerian-based methods to extract features in fluid flows. One broad category in ML is the convolution operation-based methods. The precision of feature extraction in convolution operation-based methods increases by constraining the measurement box, such as dividing the input data into small patches and using them as input boxes in convolutional neural networks. In this work, we propose a method that transforms each cell of the computational domain into a detection pixel (measurement box) to perform the task of feature extraction at the smallest possible computational level. To demonstrate the performance, we extract the vortical structures in a benchmark two-dimensional lid-driven cavity flow employing a symmetric, fully convolutional network. The number of convolution and deconvolution blocks in the network’s structure is studied to obtain the highest accuracy and yet to avoid the degradation problem. Different parameters, such as the Reynolds number and velocity boundary values, are considered in the complete and clipped cavity cases to create the training and test datasets. The semantic segmentation metrics, including Jaccard and Dice, yield values close to 1 for the test set on complete and clipped cavity cases with varying the Reynolds number or the velocity boundary values.
Graphic Abstract
Similar content being viewed by others
References
Bhatnagar S, Afshar Y, Pan S, Duraisamy K, Kaushik S (2019) Prediction of aerodynamic flow fields using convolutional neural networks. Comput Mech 64:525–545
Bishop CM (2006) Pattern recognition and machine learning. Springer, Berlin
Bushnell DM (2004) Shock wave drag reduction. Annu Rev Fluid Mech 36:81–96
Buxton ORH, Breda M, Chen X (2017) Invariants of the velocity gradient tensor in a spatially developing inhomogeneous turbulent flow. J Fluid Mech 817:1–20
Chong MS, Perry AE, Cantwell BJ (1990) A general classification of three-dimensional flow fields. Phys Fluids A 2:765–777
Colvert B, Alsalman M, Kanso E (2018) Classifying vortex wakes using neural networks. Bioinspir. Biomim 13:025003(1)–025003(12)
Csurka G, Larlus D, Perronnin F, Meylan F (2013) What is a good evaluation measure for semantic segmentation. Comput. BMVC 2013:27
Deng L, Wang Y, Liu Y, Wang F, Li S, Liu J (2019) A CNN-based vortex identification method. J. Visualization 22:65–78
Dubief Y, Delcayre Y (2000) On coherent-vortex identification in turbulence. J Turbul 1:1–22
Esmaeilbeig Z, Khobahi S (2020) Deep-RLS: a model-inspired deep learning approach to nonlinear PCA. arXiv:2011.07458
Gao H, Ji S (2017) Efficient and invariant neural networks for dense prediction. arXiv:1606.05336
Glorot X, Bordes A, Bengio Y (2011) Deep sparse rectifier neural networks. In: Proceedings of the 14th international conference on artificial intelligence and statistics (AISTATS) 2011, Fort Lauderdale, FL, USA
Hahnloser R, Sarpeshkar R, Mahowald MA, Douglas RJ, Seung HS (2000) Digital selection and analogue amplification coexist in a cortex-inspired silicon circuit. Nature 405:947–951
Harlow FH, Welsh JE (1965) Numerical calculation of time-dependent viscous incompressible flow of fluid with free surface. Phys Fluids 8:2182–2189
Holmes P, Lumley J, Berkooz G (1996) Turbulence, coherent structures, dynamical systems and symmetry. Cambridge University Press, Cambridge
Hunt JCR (1987) Vorticity and vortex dynamics in complex turbulent flows. Trans Can Soc Mech Eng 11:21–35
Hunt JCR, Wray AA, Moin P (1988) Eddies, stream and convergence zones in turbulent flows. Center for Turbulence Research Report No. CTRS88, 193
Jeong J, Hussain F (1995) On the identification of a vortex. J Fluid Mech 289:69–94
Kashir B, Tabejamaat S, BaigMohammadi M (2012) An experimental study of the stability of natural gas and propane turbulent non-premixed flame under diluting condition. Therm Sci 16:1055–1065
Kashir B, Tabejamaat S, Jalalatian N (2015a) The impact of hydrogen enrichment and bluff-body lip thickness on characteristics of blended propane/hydrogen bluff-body stabilized turbulent diffusion flames. Energ Convers Manag 103:1–13
Kashir B, Tabejamaat S, Jalalatian N (2015b) A numerical study on combustion characteristics of blended methane-hydrogen bluff-body stabilized swirl diffusion flames. Int J Hydrog Energ 40:6243–6258
Kashir B, Ragone M, Yurkiv V, Mashayek F (2019) Data-driven prediction of vortical structures in turbulent flows employing deep learning techniques. 72nd annual meeting of the American physical society, division of fluid dynamics, Seattle, Washington, 2019
Kingma DP, Ba J (2014) Adam: A method for stochastic optimization. arXiv:1412.6980
Krizhevsky A, Sutskever I, Hinton GE (2017) ImageNet classification with deep convolutional neural networks. Commun ACM 60:84–90
Li D, Komperda J, Ghiasi Z, Peyvan A, Mashayek F (2019) Compressibility effects on the transition to turbulence in a spatially developing plane free shear layer. Theor Comput Fluid Dyn 33:577–602
Liu Y, Lu Y, Wang Y, Sun D, Deng L, Wan Y, Wang F (2019a) Key time steps selection for CFD data based on deep metric learning. Comput Fluids 195:104318
Liu Y, Lu Y, Wang Y, Sun D, Deng L, Wang F, Lei Y (2019b) A CNN-based shock detection method in flow visualization. Comput Fluids 184:1–9
Liu X, Deng Z, Yang Y (2019c) Recent progress in semantic image segmentation. Artif Intell Rev 52:1089–1106
Long J, Shelhamer E, Darrell T (2015) Fully convolutional networks for semantic segmentation. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 3431–3440
Monfort M, Luciani T, Komperda J, Ziebart B, Mashayek F, Marai GE (2017) A deep learning approach to identifying shock locations in turbulent combustion tensor fields. In: Modeling, analysis, and visualization of anisotropy. Springer, Berlin
Morgan NR, Tokareva S, Xiaodong L, Morgan AD (2020) A machine learning approach for detecting shocks with high-order hydrodynamic methods. AIAA Scitech 2020 Forum, Orlando, Florida. https://doi.org/10.2514/6.2020-2024
Oishi A, Yagawa G (2020) A surface to surface contact search method enhanced by deep learning. Comput Mech 65:1125–1147
Raghu M, Poole B, Kleinberg J, Ganguli S, Sohl-Dickstein J (2017) On the expressive power of deep neural networks. arXiv:1606.05336
Ronneberger O, Fischer P, Brox T (2015) U-net: convolutional networks for biomedical image segmentation. In: Navab N, Hornegger J, Wells W, Frangi A (eds) Medical image computing and computer-assisted intervention—MICCAI 2015. Lecture Notes in Computer Science, vol 9351. Springer, Cham
Strofer CM, Wu J, Xiao H, Paterson E (2019) Data-driven physics-based feature extraction from fluid flow fields. Commun. Comput. Phys. 25:625–650
Wu H, Liu X, An W, Chen S, Lyu H (2020) A deep learning approach for efficiently and accurately evaluating the flow field of supercritical airfoils. Comput Fluids 198:104393(1)–104393(17)
Zhang L, Deng Q, Machiraju R, Rangarajan A, Thompson D, Walters DK, Shen H-W (2014) Boosting techniques for physics-based vortex detection. Comput. Graph. Forum 33:282–293
Acknowledgements
The authors acknowledge the Advanced Cyberinfrastructure for Education and Research (ACER) group at the University of Illinois at Chicago for providing high-performance computing (HPC) resource used for the simulations reported in this paper.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Kashir, B., Ragone, M., Ramasubramanian, A. et al. Application of fully convolutional neural networks for feature extraction in fluid flow. J Vis 24, 771–785 (2021). https://doi.org/10.1007/s12650-020-00732-0
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12650-020-00732-0