Wind and Structures

  • 제36권6호
  • /
  • Pages.423-434
  • /
  • 2023
  • /
  • 1226-6116(pISSN)
  • /
  • 1598-6225(eISSN)
테크노프레스 (Techno-Press)
권호 표지
DOI QR코드

DOI QR Code

Prediction of aerodynamic coefficients of streamlined bridge decks using artificial neural network based on CFD dataset

  • Severin Tinmitonde (National Engineering Research Center of High-speed Railway Construction Technology, Central South University) ;
  • Xuhui He (National Engineering Research Center of High-speed Railway Construction Technology, Central South University) ;
  • Lei Yan (National Engineering Research Center of High-speed Railway Construction Technology, Central South University) ;
  • Cunming Ma (Department of Bridge Engineering, Southwest Jiaotong University) ;
  • Haizhu Xiao (Major Bridge Reconnaissance & Design Institute Co., Ltd.)
  • 투고 : 2022.05.23
  • 심사 : 2022.12.08
  • 발행 : 2023.06.25
⟨ 이전 논문 다음 논문 ⟩

초록

Aerodynamic force coefficients are generally obtained from traditional wind tunnel tests or computational fluid dynamics (CFD). Unfortunately, the techniques mentioned above can sometimes be cumbersome because of the cost involved, such as the computational cost and the use of heavy equipment, to name only two examples. This study proposed to build a deep neural network model to predict the aerodynamic force coefficients based on data collected from CFD simulations to overcome these drawbacks. Therefore, a series of CFD simulations were conducted using different geometric parameters to obtain the aerodynamic force coefficients, validated with wind tunnel tests. The results obtained from CFD simulations were used to create a dataset to train a multilayer perceptron artificial neural network (ANN) model. The models were obtained using three optimization algorithms: scaled conjugate gradient (SCG), Bayesian regularization (BR), and Levenberg-Marquardt algorithms (LM). Furthermore, the performance of each neural network was verified using two performance metrics, including the mean square error and the R-squared coefficient of determination. Finally, the ANN model proved to be highly accurate in predicting the force coefficients of similar bridge sections, thus circumventing the computational burden associated with CFD simulation and the cost of traditional wind tunnel tests.

키워드

과제정보

This work was supported by the National Natural Science Foundation of China [Grant Numbers 52178516, 51925808], the Science and Technology Research and Development Program of China Railway Group Limited [Grant Number 2021-Special-04-2] and the Tencent Foundation or XPLORER PRIZE. The authors are grateful for resources from the High-Performance Computing Center of the Chinese Academy of Sciences (HPC-CAS-Beijing).

참고문헌

  1. Abbas, T., Kavrakov, I., Morgenthal, G. and Lahmer, T. (2020), "Prediction of aeroelastic response of bridge decks using artificial neural networks", Comput. Struct., 231, 106198. https://doi.org/10.1016/j.compstruc.2020.106198. 
  2. Akiba, T., Sano, S., Yanase, T., Ohta, T. and Koyama, M. (2019), "Optuna, A next-generation hyperparameter optimization framework", 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Anchorage, AK, USA, July. 
  3. Andersen, M.S. and Brandt, A. (2018), "Aerodynamic instability investigations of a novel, flexible and lightweight triple-box girder design for long-span bridges", J. Bridge Eng., 23(12), 04018095. https://doi.org/10.1061/(asce)be.1943-5592.0001317. 
  4. Bernardini, E., Spence, S.M.J., Wei, D. and Kareem, A. (2015), "Aerodynamic shape optimization of civil structures, A CFD-enabled Kriging-based approach", J. Wind Eng. Ind. Aerodyn., 144, 154-164. https://doi.org/10.1016/j.jweia.2015.03.011. 
  5. Brusiani, F., de Miranda, S., Patruno, L., Ubertini, F. and Vaona, P. (2013), "On the evaluation of bridge deck flutter derivatives using RANS turbulence models", J. Wind Eng. Ind. Aerodyn., 119, 39-47. https://doi.org/10.1016/j.jweia.2013.05.002. 
  6. Chen, C.H. (2003), "Determination of flutter derivatives via a neural network approach", J. Sound Vib., 263(4), 797-813. https://doi.org/10.1016/S0022-460X(02)01279-8. 
  7. Chen, C.H., Wu, J.C. and Chen, J.H. (2008), "Prediction of flutter derivatives by artificial neural networks", J. Wind Eng. Ind. Aerodyn., 96(10-11), 1925-1937. https://doi.org/10.1016/j.jweia.2008.02.044. 
  8. Chen, X.Z. and Kareem, A. (2002), "Advances in modeling of aerodynamic forces on bridge decks", J. Eng. Mech., 128(11), 1193-1205. https://doi.org/10.1061/(asce)0733-9399(2002)128:11(1193). 
  9. Chen, Z.S., Xu, Y.M., Huang, H.L. and Tse, K.T. (2020), "Wind tunnel measurement systems for unsteady aerodynamic forces on Bluff Bodies, review and new perspective", Sensors, 20(16), 4633. https://doi.org/10.3390/s20164633. 
  10. Chobsilprakob, P., Kim, K.D., Suthasupradit, S. and Manovachirasan, A. (2014), "Application of indicial function for the flutter analysis of long span suspension bridge during erection", Int. J. Steel Struct., 14(1), 185-194. https://doi.org/10.1007/s13296-014-1016-2. 
  11. Cid Montoya, M., Nieto, F., Hernandez, S., Kusano, I., Alvarez, A.J. and Jurado, J.A. (2018), "CFD-based aeroelastic characterization of streamlined bridge deck cross-sections subject to shape modifications using surrogate models", J. Wind Eng. Ind. Aerodyn., 177, 405-428. https://doi.org/10.1016/j.jweia.2018.01.014. 
  12. Computational Fluid Dynamics Committee (2002), Guide for the Verification and Validation of Computational Fluid Dynamics Simulations, (AIAA G-077-1998), American Institute of Aeronautics and Astronautics, Reston, VA, United States of America. 
  13. Croux, C. and Dehon, C. (2010), "Influence functions of the Spearman and Kendall correlation measures", Stat. Methods Appl., 19(4), 497-515. https://doi.org/10.1007/s10260-010-0142-z. 
  14. Diana, G., Rocchi, D., Argentini, T. and Muggiasca, S. (2010), "Aerodynamic instability of a bridge deck section model, Linear and nonlinear approach to force modeling", J. Wind Eng. Ind. Aerodyn., 98(6-7), 363-374. https://doi.org/10.1016/j.jweia.2010.01.003. 
  15. Ding, F. and Kareem, A. (2018), "A multi-fidelity shape optimization via surrogate modeling for civil structures", J. Wind Eng. Ind. Aerodyn., 178, 49-56. https://doi.org/10.1016/j.jweia.2018.04.022. 
  16. Dong, J.H, Huang, L., Liao, H.L. and Wang, Q. (2021), "Investigation on suppressing vortex-induced vibrations of the rectangular steel box girder for railway cable-stayed bridges by installing wind fairings", J. Wind Eng. Ind. Aerodyn., 219, 104821. https://doi.org/10.1016/j.jweia.2021.104821. 
  17. Forrester, A., Sobester, A. and Keane, A. (2008), Engineering Design via Surrogate Modelling, A Practical Guide, John Wiley & Sons Publication, Chichester, West Susex, United Kingdom. 
  18. Geman, S., Bienenstock, E. and Doursat, R. (1992), "Neural networks and the Bias/Variance dilemma", Neural Comput., 41(1):1-58. https://doi.org/10.1162/neco.1992.4.1.1. 
  19. Haque, M.N., Katsuchi, H., Yamada, H. and Nishio, M. (2016), "Investigation of edge fairing shaping effects on aerodynamic response of long-span bridge deck by unsteady RANS", Arch. Civil Mech. Eng., 16(4), 888-900. https://doi.org/10.1016/j.acme.2016.06.007. 
  20. He, X.H., Li, H., Wang, H.F., Fang, D.X. and Liu, M.T. (2017), "Effects of geometrical parameters on the aerodynamic characteristics of a streamlined flat box girder", J. Wind Eng. Ind. Aerodyn., 170, 56-67. https://doi.org/10.1016/j.jweia.2017.08.009. 
  21. Hu, L.W., Zhang, J., Xiang, Y. and Wang, W.Y. (2020), "Neural networks-based aerodynamic data modeling, A comprehensive review", IEEE Access, 8, 90805-90823. https://doi.org/10.1109/ACCESS.2020.2993562. 
  22. Huang, D.M., He, S.Q., He, X.H. and Zhu, X. (2017), "Prediction of wind loads on high-rise building using a BP neural network combined with POD", J. Wind Eng. Ind. Aerodyn., 170, 1-17. https://doi.org/10.1016/j.jweia.2017.07.021. 
  23. Kareem, A. (2020), "Emerging frontiers in wind engineering, Computing, stochastics, machine learning and beyond", J. Wind Eng. Ind. Aerodyn., 206, 104320. https://doi.org/10.1016/j.jweia.2020.104320. 
  24. Lalonde, E.R., Vischschraper, B., Bitsuamlak, G. and Dai, K.S. (2021), "Comparison of neural network types and architectures for generating a surrogate aerodynamic wind turbine blade model", J. Wind Eng. Ind. Aerodyn., 216, 104696. https://doi.org/10.1016/j.jweia.2021.104696. 
  25. Lampinen, J. and Vehtari, A. (2001), "Bayesian approach for neural networks - Review and case studies", Neural Networks, 14(3), 257-274. https://doi.org/10.1016/S0893-6080(00)00098-8. 
  26. LeCun, Y., Bottou, L., Orr, B.G. and Muller, K.R. (1998), "Efficient BackProp", Neural Networks: Tricks of the Trade Lecture Notes in Computer Science Vol. 7700, Springer, Berlin, Germany. https://doi.org/10.1007/978-3-642-35289-8_3. 
  27. Li, W.J., Laima, S.J. Jin, X.W. Yuan, W.Y. and Li, H. (2020), "A novel long short-term memory neural-network-based self-excited force model of limit cycle oscillations of nonlinear flutter for various aerodynamic configurations", Nonlinear Dyn., 100(3), 2071-2087. https://doi.org/10.1007/s11071-020-05631-5. 
  28. Li, Y.L, Chen, X.Y., Yu, C.J., Togbenou, K., Wang, B. and Zhu, L.D. (2018), "Effects of wind fairing angle on aerodynamic characteristics and dynamic responses of a streamlined trapezoidal box girder", J. Wind Eng. Ind. Aerodyn., 177 , 69-78. https://doi.org/10.1016/j.jweia.2018.04.006. 
  29. Liao, H.L., Mei, H.Y., Hu, G., Wu, B. and Wang, Q. (2021), "Machine learning strategy for predicting flutter performance of streamlined box girders", J. Wind Eng. Ind. Aerodyn., 209, 104493. https://doi.org/10.1016/j.jweia.2020.104493. 
  30. Lin, P.F., Hu, G., Li, C., Li, L.X. Xiao, Y.Q., Tse, K.T. and Kwok, K.C.S. and Kareem, A. (2022), "Machine learning-enabled estimation of crosswind load effect on tall buildings", J. Wind Eng. Ind. Aerodyn., 220, 104860. https://doi.org/10.1016/j.jweia.2021.104860. 
  31. MacKay D.J.C. (1992), "A practical Bayesian framework for back-propagation networks", Neural Comput., 4(3), 448-472. https://doi.org/10.1162/neco.1992.4.3.448. 
  32. Madhiarasan, M. and Deepa, S. N. (2017), "Comparative analysis on hidden neurons estimation in multi-layer perceptron neural networks for wind speed forecasting", Artif. Intell. Rev., 48(4), 449-471. https://doi.org/10.1007/s10462-016-9506-6. 
  33. Marquardt, D.W. (1963), "An algorithm for least-squares estimation of nonlinear parameters", J. Soc. Ind. Appl. Math., 11(2), 431-441. https://doi.org/doi:10.1137/0111030. 
  34. Moller, M.F. (1993), "A scaled conjugate gradient algorithm for fast supervised learning", Neural Netw., 6(4), 525-533. https://doi.org/10.1016/S0893-6080(05)80056-5. 
  35. Nieto, F., Cid Montoya, M., Hernandez, S., Kusano, I., Casteleiro, A., Alvarez, A.J., Jurado, J.A. and Fontan, A. (2020), "Aerodynamic and aeroelastic responses of short gap twin-box decks, Box geometry and gap distance dependent surrogate based design", J. Wind Eng. Ind. Aerodyn., 201, 104147. https://doi.org/10.1016/j.jweia.2020.104147. 
  36. Parr, J.M., Keane, A.J., Forrester, A.I.J. and Holden, C.M.E. (2012), "Infill sampling criteria for surrogate-based optimization with constraint handling", Eng. Optim., 44(10), 1147-1166. https://doi.org/10.1080/0305215X.2011.637556. 
  37. Pearlmutter, B.A. (1994), "Fast exact multiplication by the Hessian", Neural Comput., 6(1), 147-160. https://doi.org/10.1162/neco.1994.6.1.147. 
  38. Reich, Y. and Barai, S.V. (1999), "Evaluating machine learning models for engineering problems", Artif. Intell. Eng., 13(3), 257-272. https://doi.org/10.1016/S0954-1810(98)00021-1. 
  39. Rizzo, F. and Caracoglia, L. (2020), "Artificial neural network model to predict the flutter velocity of suspension bridges", Comput. Struct., 233, 106236. https://doi.org/10.1016/j.compstruc.2020.106236. 
  40. Sharma, S., Sharma, S. and Anidhya, A. (2020), "Activation functions in neural networks", Int. J. Eng. Appl. Sci. Technol., 4(12), 310-316.  https://doi.org/10.33564/IJEAST.2020.v04i12.054
  41. Wilcox, D.C. (2006), Turbulence Modeling for CFD (3rd Edition), DCW Industries, La Canada, CA, United States of America.
  42. Wu, T. and Kareem, A. (2011), "Modeling hysteretic nonlinear behavior of bridge aerodynamics via cellular automata nested neural network", J. Wind Eng. Ind. Aerodyn., 99(4), 378-388. https://doi.org/10.1016/j.jweia.2010.12.011. 
  43. Wu, T., Kareem, A. and Ge, Y.J. (2013), "Linear and nonlinear aeroelastic analysis frameworks for cable-supported bridges", Nonlinear Dyn., 74(3), 487-516. https://doi.org/10.1007/s11071-013-0984-7. 
  44. Xu, G.J., Kareem, A. and Shen, L. (2020), "Surrogate modeling with sequential updating, Applications to bridge deck-wave and bridge deck-wind interactions", J. Comput. Civ. Eng., 34(4), 04020023. https://doi.org/10.1061/(ASCE)CP.1943-5487.0000904. 
  45. Yao, Z.Q., Shen, H.C. and Gao, H. (2013), "A new methodology for the CFD uncertainty analysis", J. Hydrodyn., 25(1), 131-147. https://doi.org/10.1016/S1001-6058(13)60347-9.