Intelligent Prediction of Separated Flow Dynamics using Machine Learning

Document Type : Regular Article

Authors

1 Research laboratory of applied and fundamental physic /Blida 1 University BP 270 Route Soumâa, Blida, Algeria

2 Research Laboratory of energetic, flow and transfers /AMC BP 48 Cherchell terre 42006, Tipaza, Algeria

3 LAMIH, UMR-CNRS 8201, Department of Mechanical Engineering, University of Valenciennes and Hainaut-Cambresis, Valenciennes 59300, France

10.47176/jafm.18.2.2910

Abstract

Understanding separated flow dynamics is crucial for implementing effective flow control techniques. These techniques help mitigate adverse effects on vehicle performance and environmental pollution. This research aims to improve flow control strategies by predicting separated flow dynamics solely through wall pressure measurements using artificial intelligence and numerical data. Initially, we identify numerical models that accurately replicate separated flow dynamics. Notably, the Detached Eddy Simulation (DES) model strongly agrees with experimental data, particularly in the turbulent regime at Reh= 89100, downstream of backward facing steps (BFS). Subsequently we conducted a correlational analysis that revealed a significant relationship between various wall pressure points and the velocity field, leading to the adoption of deep learning techniques such as Recurrent Neural Networks with Long Short-Term Memory (LSTM). These neural networks, tailored for time-dependent data, demonstrate high accuracy of low MSE of 13.48% using ten wall pressure points in predicting velocity magnitude contour over (BFS). To enhance predictions, Proper Orthogonal Decomposition (POD) is utilized to reduce system complexity while retaining essential dynamics, resulting in a lower MSE of 5.07%. Additionally, we identify the ideal wall pressure measurement region that accurately captures the entire dynamic behavior, achieving an acceptable MSE of 23.48% for predicting low order vorticity, with only three wall pressure points. This research aids in developing efficient flow control strategies with limited pressure data and offers valuable insights for closed-loop flow control applications.

Keywords

Main Subjects

References

Almohammadi, K. M. (2020). Assessment of reattachment length using turbulence models on backward facing step (BFS) for turbulent flow with modified general richardson method. Arabian Journal for Science and Engineering, 45(11), 9293–9303. https://doi.org/10.1007/s13369-020-04695-0
Altché, F., & Fortelle, A. D. L. (2017). An LSTM network for highway trajectory prediction. In Proceedings ofthe IEEE 20th International Conference on Intelligent Transportation Systems. Piscataway, NJ: IEEE. https://doi.org/10.48550/arXiv.1801.07962
Antonio, V., & Lacerda De Brederode, S. (n.d.). Three-dimensional effects in nominally two-dimensional flows. https://api.semanticscholar.org/CorpusID:130177294
Bengio, Y. (2009). Learning deep Architectures for AI. Foundations and Trends in Machine Learning, 2(1), 1–127. https://doi.org/10.1561/2200000006
Brown, B., Yu, X., & Garverick, S. (2004). Mixed-mode analog VLSI continuous-time recurrent neural network. In Circuits, Signals, and Systems: IASTED International Conference Proceedings. https://dblp.org/rec/conf/iastedCCS/BrownYG04
Carrio, A., Sampedro, C., Rodriguez-Ramos, A., & Campoy, P. (2017). A review of deep learning methods and applications for unmanned aerial Vehicles. Journal of Sensors, 2, 1–13. https://doi.org/10.1155/2017/3296874
Chen, T. B., & Soo, V. W. (1996). A comparative study of recurrent neural network architectures on learning temporal sequences. In Proceedings of the IEEE International Conference on Neural Networks. Piscataway, NJ: IEEE. https://doi.org/10.1109/ICNN.1996.549199
Chovet, C., Lippert, M., Foucaut, J. M., & Keirsbulck, L. (2017). Dynamical aspects of a backward-facing step flow at large Reynolds numbers. Experiments in Fluids, 58(11). https://doi.org/10.1007/s00348-017-2444-5
Chovet, C., Lippert, M., Keirsbulck, L., & Foucaut, J. M. (2019). Unsteady behavior of a backward-facing step in forced flow. Flow, Turbulence and Combustion, 102(1), 145–165. https://doi.org/10.1007/s10494-018-9944-0
Duriez, T., Brunton, S. L., & Noack, B. R. (n.d.). Fluid Mechanics and Its Applications Machine Learning Control-Taming Nonlinear Dynamics and Turbulence. http://doi.org/10.1007/978-3-319-40624-4
Elman, J. L. (1990). Finding structure in time. Cognitive Science, 14(2), 179–211. https://doi.org/10.1016/0364-0213(90)90002-E
Fadla, F., Alizard, F., Keirsbulck, L., Robinet, J. C., Laval, J. P., Foucaut, J. M., Chovet, C., Alizard, F., & Lippert, M. (2019). Science Arts & Métiers (SAM) investigation of the dynamics in separated turbulent flow. European Journal of Mechanics-B/Fluids, https://doi.org/10.1016/j.euromechflu.2019.01.006
Fadla, F., Graziani, A., Kerherve, F., Mathis, R., Lippert, M., Uystepruyst, D., & Keirsbulck, L. (2016). Electrochemical measurements for real-time stochastic reconstruction of large-scale dynamics of a separated flow. Journal of Fluids Engineering, Transactions of the ASME, 138(12). https://doi.org/10.1115/1.4034198
Fernández, S., Graves, A., & Schmidhuber, J. (2007). An application of recurrent neural networks to discriminative keyword spotting. In Proceedings of the International Conference on Artificial Neural Networks (pp. 220–229). Berlin: Springer. https://doi.org/10.1007/978-3-540-74695-9_23
Fukushima, K. (1980). Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position. Biological Cybernetics, 36(4), 193–202. https://doi.org/10.1007/BF00344251
Giannopoulos, A., & Aider, J. L. (2020). Prediction of the dynamics of a backward-facing step flow using focused time-delay neural networks and particle image velocimetry data-sets. International Journal of Heat and Fluid Flow, 82. https://doi.org/10.1016/j.ijheatfluidflow.2019.108533
Gallagher, J. C., Boddhu, S. K., & Vigraham, S. (2005). A reconfigurable continuous time recurrent neural network for evolvable hardware applications. In Proceedings of the 2005 IEEE Congress on Evolutionary Computation. Piscataway, NJ: IEEE. http://doi.org/10.1109/EH.2005.5
Guo, Y., Liu, Y., Georgiou, T., & Lew, M. S. (2017). A review of semantic segmentation using deep neural networks. International Journal of Multimedia Information Retrieval, 2, 1–7. https://doi.org/10.1007/s13735-017-0141-z
He, T., & Droppo, J. (2016). Exploiting LSTM structure in deep neural networks for speech recognition. In Proceedings ofthe IEEE International Conference on Acoustics, Speech and Signal Processing (pp. 5445–5449). Piscataway, NJ: IEEE. https://doi.org/10.1109/ICASSP.2016.7472718
Hsu, W. N., Zhang, Y., Lee, A., & Glass, J. (2016). Exploiting depth and highway connections in convolutional recurrent deep neural networks for speech recognition. Cell, 50(1), 395–399. https://doi.org/10.21437/Interspeech.2016-515
Ivakhnenko, A. G., & Lapa, V. G. (1965). Cybernetic predicting devices. Sacramento, CA: CCM Information Corporation. https://api.semanticscholar.org/CorpusID:61108232
Ivakhnenko, A. G. (1971). Polynomial theory of complex systems. IEEE Transactions on Systems, Man and Cybernetics, 4, 364–378. http://doi.org/10.1109/TSMC.1971.4308320
Jordan, M. (1986). Attractor dynamics and parallelism in a connectionist sequential machine. In Proceedings of the Annual Conference of the Cognitive Science Society (pp.531–546). Piscataway, NJ: IEEE. https://escholarship.org/uc/item/1fg2j76h
Khan, S., & Yairi, T. (2018). A review on the application of deep learning in system health management. Mechanical Systems and Signal Processing, 107, 241-265. https://doi.org/10.1016/j.ymssp.2017.11.024
Kumar, K. R., & Selvaraj, M. (2023). Novel deep learning model for predicting wind velocity and power estimation in advanced INVELOX wind turbines. Journal of Applied Fluid Mechanics, 16(6), 1256–1268. https://doi.org/10.47176/jafm.16.06.1637
LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278–2324. http://doi.org/10.1109/5.726791
Luo, D. (2019). Numerical simulation of turbulent flow over a backward facing step using partially averaged Navier-Stokes method. Journal of Mechanical Science and Technology, 33(5), 2137–2148. https://doi.org/10.1007/s12206-019-0416-9
Mallinar, N., &Rosset, C. (2018). Deep canonically correlated LSTMs. https://doi.org/10.48550/arXiv.1801.05407
Mehrez, Z., Bouterra, M., Cafsi, A. El, Belghith, A., & Le Quéré, P. (2010). Simulation of the periodically perturbed separated and reattaching flow over a backward-facing step. Journal of Applied Fluid Mechanics, 3(2). https://doi.org/10.36884/jafm.3.02.11883
Ötügen, M. V. (1991). Ex rimeas m l mds expansion ratio effects on the separated shear layer and reattachment downstream of a backward-facing step. Experiments in Fluids, 10.
Palangi, H., Deng, L., Shen, Y., Gao, J., He, X., Chen, J., & Ward, R. (2015). Deep sentence embedding using the long short-term memory network: Analysis and application to information retrieval. IEEE/ACM Transactions on Audio Speech and Language Processing, 24(4), 694–707.  https://doi.org/10.48550/arXiv.1502.06922
Probst, A., Radespiel, R., Wolf, C., Knopp, T., & Schwamborn, D. (2010). A Comparison of Detached-Eddy Simulation and Reynolds-Stress Modelling Applied to the Flow over a Backward-Facing Step and an Airfoil at Stall. https://doi.org/10.2514/6.2010-920
Pearlmutter, B. A. (1989). Learning state space trajectories in recurrent neural networks. Neural Computation, 1(2), 263–269. https://doi.org/10.1162/neco.1989.1.2.263
Qu, Z., Haghani, P., Weinstein, E., & Moreno, P. (2017). Syllable-based acoustic modeling with CTC-SMBR-LSTM. In Proceedings of the IEEE Automatic Speech Recognition and Understanding Workshop (pp. 173–177). Piscataway, NJ: IEEE. https://doi.org/10.1109/ASRU.2017.8268932
Rajabi, E., & Kavianpour, M. R. (2012). Intelligent prediction of turbulent flow over backward-facing step using direct numerical simulation data. Engineering Applications of Computational Fluid Mechanics, 6(4), 490–503. https://doi.org/10.1080/19942060.2012.11015437
Ranzato, M. A., Szlam, A., Bruna, J., Mathieu, M., Collobert, R., & Chopra, S. (2014). Video (language) modeling: A baseline for generative models of natural-videos. https://doi.org/10.48550/arXiv.1412.6604
Rawat, W., &Wang, Z. (2017). Deep convolutional neural Networks for image classification: A comprehensive review. Neural Computation, 29(9), 1–10. https://doi.org/10.1162/NECO_a_00990
Robinson, A. J., & Fallside, F. (1987). The utility driven dynamic error propagation network. Cambridge: University of Cambridge Department of Engineering. https://www.researchgate.net/publication/243683010_The_utility_driven_dynamic_error_propagation_network
Sak, H. I., Senior, A., & Beaufays, F. O. (2014). Long short-term memory based recurrent neural network architectures for large vocabulary speech recognition.  https://doi.org/10.48550/arXiv.1402.1128
Šarić, S., Jakirlić, S., & Tropea, C. (2005). A periodically perturbed backward-facing step flow by means of LES, des and T-RANS: An example of flow separation control. Journal of Fluids Engineering, Transactions of the ASME, 127(5), 879–887. https://doi.org/10.1115/1.2012502
Sharma, P., & Singh, A. (2017). Era of deep neural networks: A review. In Proceedings ofthe 8th International Conference on Computing, Communication and Networking Technologies (pp. 1–5). Piscataway, NJ: IEEE. https://doi.org/10.1109/ICCCNT.2017.8203938
Singh, A. P., Medida, S., & Duraisamy, K. (2017). Machine-learning-augmented predictive modeling of turbulent separated flows over airfoils. AIAA Journal, 55(7), 2215–2227. https://doi.org/10.2514/1.J055595
Smirnov, E. M., Smirnovsky, A. A., Schur, N. A., Zaitsev, D. K., & Smirnov, P. E. (2018). Comparison of RANS and IDDES solutions for turbulent flow and heat transfer past a backward-facing step. Heat and Mass Transfer, 54(8), 2231–2241. https://doi.org/10.1007/s00231-017-2207-0
Sohankar, A., Khodadadi, M., Rangraz, E., & Alam, M. M. (2019). Control of flow and heat transfer over two inline square cylinders. Physics of Fluids, 31(12). https://doi.org/10.1063/1.5128751
Šter, B. (2013). Selective recurrent neural network. Neural Processing Letters, 38(1),1–15 https://doi.org/10.1007/s11063-012-9259-4
Sujar Garrido, P., Moreau, É., Bonnet, J.-P., Benard, N., & Éric Moreau, M. (2014). Active control of the turbulent flow downstream of a backward facing step with dielectric barrier discharge plasma actuators [Doctoral dissertation, Poitiers]. https://api.semanticscholar.org/CorpusID:92790265
 
Talele, V., Mathew, V. K., Sonawane, N., Sanap, S., Chandak, A., & Nema, A. (2021). CFD and ANN approach to predict the flow pattern around the square and rectangular bluff body for high Reynolds number. Materials Today: Proceedings, 47, 3177–3185. https://doi.org/10.1016/j.matpr.2021.06.285
Werbos, P. J. (1988). Generalization of backpropagation with application to a recurrent gas market model. Neural Networks, 1(4), 339–356. https://doi.org/10.1016/0893-6080(88)90007-X
Weng, J. J., Ahuja, N., & Huang, T. S. (1993, May). Learning recognition and segmentation of 3D objects from 2D images. In Proceedings of the Fourth International Conference on Computer Vision (pp. 121–128). Piscataway, NJ: IEEE. https://doi.org/10.1109/iccv.1993.378228
Williams, R. J. (1989). Complexity of exact gradient computation algorithms for recurrent neural networks (Technical Report NU-CCS-89-27). Boston: Northeastern University, College of Computer Science. https://modeldb.science/citations/107088
Yu, Y., Si, X., Hu, C., & Zhang, J. (2019). A review of recurrent neural networks: LSTM cells and network architectures. Neural computation, 31(7), 1235-1270. https://doi.org/10.1162/neco_a_01199
Journal of Applied Fluid Mechanics
Volume 18, Issue 2 - Serial Number 94
February 2025
Pages 399-418

Files

History

  • Received: 09 June 2024
  • Revised: 18 August 2024
  • Accepted: 05 September 2024
  • Available online: 04 December 2024

Share

How to cite

Statistics