Wind and Structures

  • 제36권6호
  • /
  • Pages.393-404
  • /
  • 2023
  • /
  • 1226-6116(pISSN)
  • /
  • 1598-6225(eISSN)
테크노프레스 (Techno-Press)
권호 표지
DOI QR코드

DOI QR Code

Prediction of skewness and kurtosis of pressure coefficients on a low-rise building by deep learning

  • Youqin Huang (Research Centre for Wind Engineering and Engineering Vibration, Guangzhou University) ;
  • Guanheng Ou (Research Centre for Wind Engineering and Engineering Vibration, Guangzhou University) ;
  • Jiyang Fu (Research Centre for Wind Engineering and Engineering Vibration, Guangzhou University) ;
  • Huifan Wu (Research Centre for Wind Engineering and Engineering Vibration, Guangzhou University)
  • 투고 : 2022.06.27
  • 심사 : 2023.01.31
  • 발행 : 2023.06.25
⟨ 이전 논문 다음 논문 ⟩

초록

Skewness and kurtosis are important higher-order statistics for simulating non-Gaussian wind pressure series on low-rise buildings, but their predictions are less studied in comparison with those of the low order statistics as mean and rms. The distribution gradients of skewness and kurtosis on roofs are evidently higher than those of mean and rms, which increases their prediction difficulty. The conventional artificial neural networks (ANNs) used for predicting mean and rms show unsatisfactory accuracy in predicting skewness and kurtosis owing to the limited capacity of shallow learning of ANNs. In this work, the deep neural networks (DNNs) model with the ability of deep learning is introduced to predict the skewness and kurtosis on a low-rise building. For obtaining the optimal generalization of the DNNs model, the hyper parameters are automatically determined by Bayesian Optimization (BO). Moreover, for providing a benchmark for future studies on predicting higher order statistics, the data sets for training and testing the DNNs model are extracted from the internationally open NIST-UWO database, and the prediction errors of all taps are comprehensively quantified by various error metrices. The results show that the prediction accuracy in this study is apparently better than that in the literature, since the correlation coefficient between the predicted and experimental results is 0.99 and 0.75 in this paper and the literature respectively. In the untrained cornering wind direction, the distributions of skewness and kurtosis are well captured by DNNs on the whole building including the roof corner with strong non-normality, and the correlation coefficients between the predicted and experimental results are 0.99 and 0.95 for skewness and kurtosis respectively.

키워드

과제정보

This work was financially supported by the "111" Project (No. D21021), National Natural Science Foundation Project (No. 51925802) and Guangzhou Municipal Science and Technology Bureau Project (Nos. 201904010307, 20212200004) of China. Specially thank the NIST-UWO database for providing the aerodynamic datasets on low-rise buildings.

참고문헌

  1. Bendat, J.S. and Piersol, A.G. (1986), Random Data: Analysis and Measurement Procedures, John Wiley and Sons, New York, NY, USA.
  2. Chen, Y., Kopp, G.A. and Surry, D. (2002), "Interpolation of wind-induced pressure time series with an artificial neural network", J. Wind Eng. Ind. Aerod., 90(6), 589-615. https://doi.org/10.1016/S0167-6105(02)00155-1.
  3. Chen, Y., Kopp, G.A. and Surry, D. (2003), "Prediction of pressure coefficients on roofs of low buildings using artificial neural networks", J. Wind Eng. Ind. Aerod., 91(3), 423-441. https://doi.org/10.1016/S0167-6105(03)00006-0.
  4. Ding, Z., Zhang, W. and Zhu, D. (2022), "Neural-network based wind pressure prediction for low-rise buildings with genetic algorithm and Bayesian optimization", Eng. Struct., 260, 114203. https://doi.org/10.1016/j.engstruct.2022.114203.
  5. Duchi, J., Hazan, E. and Singer, Y. (2011), "Adaptive subgradient methods for online learning and stochastic optimization", J. Mach. Learn. Res., 12, 2121-2159. https://doi.org/10.1109/TNN.2011.2146788.
  6. Fang, Z.Y., Wang, Z.S. and Li, Z.L. (2020), "Prediction of downburst-induced wind pressure coefficients on high-rise building surfaces using BP neural network", Wind Struct., 30(3), 289-298. https://doi.org/10.12989/was.2020.30.3.289.
  7. Feng, R., Liu, F., Cai, Q., Yang, G. and Leng, J. (2018), "Field measurements of wind pressure on an open roof during Typhoons HaiKui and SuLi", Wind Struct., 26(1), 11-24. https://doi.org/10.12989/was.2018.26.1.011.
  8. Fu, J.Y., Li, Q.S. and Xie, Z.N. (2006), "Prediction of wind loads on a large flat roof using fuzzy neural networks", Eng. Struct., 28(1), 153-161. https://doi.org/10.1016/j.engstruct.2005.08.006.
  9. Fu, J.Y., Liang, S.G. and Li, Q.S. (2007), "Prediction of wind-induced pressures on a large gymnasium roof using artificial neural networks", Comput. Struct., 85(3-4), 179-192. https://doi.org/10.1016/j.compstruc.2006.08.070.
  10. Gavalda, X., Ferrer-Gener, J., Kopp, G.A. and Giralt, F. (2011), "Interpolation of pressure coefficients for low-rise buildings of different plan dimensions and roof slopes using artificial neural networks", J. Wind Eng. Ind. Aerod., 99(5), 658-664. https://doi.org/10.1016/j.jweia.2011.02.008.
  11. Glorot, X., Bordes, A. and Bengio, Y. (2011), "Deep sparse rectifier neural networks", Proceedings of the 14th International Conference on Artificial Intelligence and Statistics, Fort Lauderdale, FL, USA, January.
  12. Hinton, G.E. and Salakhutdinov, R.R. (2006), "Reducing the dimensionality of data with neural networks", Science, 313(5786), 504-507. https://doi.org/10.1126/science.1127647.
  13. Jeong, S.H. (2004), "Simulation of large wind pressures by gusts on a bluff structure", Wind Struct., 7(5), 333-344. https://doi.org/10.12989/was.2004.7.5.333.
  14. Jin, J., Zhang, C., Feng, F., Na, W., Ma, J. and Zhang, Q. J. (2019), "Deep neural network technique for high-dimensional microwave modeling and applications to parameter extraction of microwave filters", IEEE. T. Microw. Theory., 67(10), 4140-4155. https://doi.org/10.1109/TMTT.2019.2932738.
  15. Jin, X.W., Cheng, P., Chen, W.L. and Li, H. (2018), "Prediction model of velocity field around circular cylinder over various Reynolds numbers by fusion convolutional neural networks based on pressure on the cylinder", Phys. Fluids, 30(4), 047105. https://doi.org/10.1063/1.5024595.
  16. Jubayer, C., Romanic, D. and Hangan, H. (2019), "Aerodynamic loading of a typical low-rise building for an experimental stationary and non-Gaussian impinging jet", Wind Struct., 28(5), 315-329. https://doi.org/10.12989/was.2019.28.5.315.
  17. Kingma, D.P. and Ba, J. (2014), "Adam: A method for stochastic optimization", arXiv preprint, arXiv, 1412.6980. https://doi.org/10.48550/arXiv.1412.6980.
  18. Konka, S., Govindray, S.R., Rajasekharan, S.G. and Rao, P.N., (2021), "Estimation of wind pressure coefficients on multi-building configurations using data-driven approach", Wind Struct., 32(2), 127-142. https://doi.org/10.12989/was.2021.32.2.127.
  19. Kumar, K.S. and Stathopoulos, T. (1998), "Fatigue analysis of roof cladding under simulated wind loading", J. Wind Eng. Ind. Aerod., 77, 171-183. https://doi.org/10.1016/S0167-6105(98)00141-X.
  20. Kumar, K.S. and Stathopoulos, T. (1999), "Synthesis of non-gaussian wind pressure time series on low building roofs", Eng. Struct., 21(12), 1086-1100. https://doi.org/10.1016/S0141-0296(98)00069-8.
  21. Lamberti, G. and Gorle, C. (2021), "A multi-fidelity machine learning framework to predict wind loads on buildings", J. Wind Eng. Ind. Aerod., 214, 104647. https://doi.org/10.1016/j.jweia.2021.104647.
  22. LeCun, Y., Bengio, Y. and Hinton, G. (2015), "Deep learning", Nature, 521(7553), 436-444. https://doi.org/10.1038/nature14539.
  23. Li, T., Wu, T., and Liu, Z. (2020), "Nonlinear unsteady bridge aerodynamics: Reduced-order modeling based on deep LSTM networks", J. Wind Eng. Ind. Aerod., 198, 104116. https://doi.org/10.1016/j.jweia.2020.104116.
  24. Ling, J., Kurzawski, A. and Templeton, J. (2016), "Reynolds averaged turbulence modelling using deep neural networks with embedded invariance", J. Fluid Mech., 807, 155-166. https://doi.org/10.1017/jfm.2016.615.
  25. Liu, M., Chen, X. and Yang, Q. (2017), "Estimation of peak factor of non-Gaussian wind pressures by improved moment-based Hermite model", J Eng. Mech., 143(7), 06017006. https://doi.org/10.1061/(ASCE)EM.1943-7889.0001233.
  26. Merhi, A. and Letchford, C.W. (2021), "Computational method in database-assisted design for wind engineering with varying performance objectives", Wind Struct., 32(5), 439-452. https://doi.org/10.12989/was.2021.32.5.439.
  27. Miyanawala, T.P. and Jaiman, R.K. (2017), "An efficient deep learning technique for the Navier-Stokes equations: Application to unsteady wake flow dynamics", arXiv preprint, arXiv: 1710.09099. https://doi.org/10.48550/arXiv.1710.09099.
  28. Mockus, J., Tiesis, V. and Zilinskas, A. (1978), "The application of Bayesian methods for seeking the extremum", In: Dixon L. C. W. and Szego G. P., editors, Towards Global Optimization 2, 117-129, North-Holand Publishing Company, New York, USA.
  29. Nielsen, M. (2015), Neural Networks and Deep Learning, Determination Press, Princeton, NJ, USA.
  30. NIST Aerodynamic Database (2003), National Institute Of Standards And Technology Aerodynamic Database, University of Western Ontario, London, Ontario, Canada. https://www.nist.gov/el/materials-and-structural-ystemsdivision-73100/nist-aerodynamic-database
  31. Pratap, A. and Rani, N. (2022), "Review of international wind codes and recent research on mono-slope canopy roof", Wind Struct., 34(4), 371-383. https://doi.org/10.12989/was.2022.34.4.371.
  32. Rizzo, F., Sepe, V., Ricciardelli, F. and Avossa, A. M. (2020), "Wind pressures on a large span canopy roof", Wind Struct., 30(3), 299-316. https://doi.org/10.12989/was.2020.30.3.299.
  33. Robbins, H. and Monro, S. (1951), "A stochastic approximation method", Ann. Math. Stat., 1951, 22(3), 400-407. https://doi.org/10.1007/978-1-4612-5110-1_9.
  34. Rumelhart, D.E., Hinton, G.E. and Williams, R.J. (1986), "Learning representations by back propagating errors", Nature, 323(6088), 533-536. https://doi.org/10.1038/323533a0.
  35. Shahriari, B., Swersky, K., Wang, Z., Adams, R.P. and De Freitas, N. (2016), "Taking the human out of the loop: A review of Bayesian optimization", P. IEEE., 104(1), 148-175. https://doi.org/10.1109/JPROC.2015.2494218.
  36. Snaiki, R. and Wu, T. (2019), "Knowledge-enhanced deep learning for simulation of tropical cyclone boundary-layer winds", J. Wind Eng. Ind. Aerod., 194, 103983. https://doi.org/10.1016/j.jweia.2019.103983.
  37. Su, N., Sun, Y., Yang, L., Wu, Y. and Shen, S.Z. (2016), "Wind load prediction of large-span domes based on generalized regression neural network and its application", J. Build. Struct., 37(6), 101-106. https://doi.org/10.14006/j.jzjgxb.2016.06.013.
  38. Sun, Y., Wu, Y., Lin, Z.X. and Shen, S.Z. (2007), "Non-Gaussian features of fluctuating wind pressures on long span roofs", China Civil Eng. J., 40(4), 1-5, 12. (in Chinese) https://doi.org/10.1016/S1672-6529(07)60007-9.
  39. Tian, J., Gurley, K.R., Diaz, M.T., Fernandez-Caban, P.L., Masters, F.J. and Fang, R. (2020), "Low-rise gable roof buildings pressure prediction using deep neural networks", J. Wind Eng. Ind. Aerod., 196, 104026. https://doi.org/10.1016/j.jweia.2019.104026.
  40. Tieleman, T. and Hinton, G. (2012), "Lecture 6.5-rmsprop: Divide the gradient by a running average of its recent magnitude", COURSERA: Neural Networks Machine Learning, 4(2), 26-31.
  41. Turkkan, N. and Srivastava, N.K. (1995), "Prediction of wind load distribution for air-supported structures using neural networks", Can. J. Civil Eng., 22(3), 453-461. https://doi.org/10.1139/l95-053.
  42. Uematsu, Y. and Tsuruishi, R. (2008), "Wind load evaluation system for the design of roof cladding of spherical domes", J. Wind Eng. Ind. Aerod., 96(10-11), 2054-2066. https://doi.org/10.1016/j.jweia.2008.02.051.
  43. Yang, Q., Chen, X. and Liu, M. (2019), "Bias and sampling errors in estimation of extremes of non-Gaussian wind pressures by moment-based translation process models", J. Wind Eng. Ind. Aerod., 186, 214-233. https://doi.org/10.1016/j.jweia.2019.01.006.
  44. Yang, X., Zhou, Q., Lei, Y., Yang Y. and Li, M. (2022), "Non-Gaussian features of dynamic wind loads on a long-span roof in boundary layer turbulences with different integral-scales", Wind Struct., 34(5), 421-435. https://doi.org/10.12989/was.2022.34.5.421.
  45. Yu, Y. and Liu, F. (2019), "Effective neural network training with a new weighting mechanism-based optimization algorithm", IEEE Access, 7, 72403-72410. https://doi.org/10.1109/ACCESS.2019.2919987.
  46. Zhang, Q.J. and Gupta, K.C. (2000), Neural Networks for RF and Microwave Design, Artech House, Boston, USA.