Wind and Structures

Techno-Press (테크노프레스)
권호 표지
DOI QR코드

DOI QR Code

Extrapolation of wind pressure for low-rise buildings at different scales using few-shot learning

  • Yanmo Weng (Zachry Department of Civil and Environmental Engineering, Texas A&M University) ;
  • Stephanie G. Paal (Zachry Department of Civil and Environmental Engineering, Texas A&M University)
  • Received : 2022.07.31
  • Accepted : 2023.01.12
  • Published : 2023.06.25
⟨ Previous Next ⟩

Abstract

This study proposes a few-shot learning model for extrapolating the wind pressure of scaled experiments to full-scale measurements. The proposed ML model can use scaled experimental data and a few full-scale tests to accurately predict the remaining full-scale data points (for new specimens). This model focuses on extrapolating the prediction to different scales while existing approaches are not capable of accurately extrapolating from scaled data to full-scale data in the wind engineering domain. Also, the scaling issue observed in wind tunnel tests can be partially resolved via the proposed approach. The proposed model obtained a low mean-squared error and a high coefficient of determination for the mean and standard deviation wind pressure coefficients of the full-scale dataset. A parametric study is carried out to investigate the influence of the number of selected shots. This technique is the first of its kind as it is the first time an ML model has been used in the wind engineering field to deal with extrapolation in wind performance prediction. With the advantages of the few-shot learning model, physical wind tunnel experiments can be reduced to a great extent. The few-shot learning model yields a robust, efficient, and accurate alternative to extrapolating the prediction performance of structures from various model scales to full-scale.

Keywords

Acknowledgement

The authors gratefully acknowledge the support from Dr. Chowdhury's research group at Florida International University. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References

  1. Akiba, T., Sano, S., Yanase, T., Ohta, T. and Koyama, M. (2019), "Optuna: A next-generation hyperparameter optimization framework", Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, pp. 2623-2631. 
  2. Alajaji, D.A. and Alhichri, H. (2020), "Few shot scene classification in remote sensing using meta-agnostic machine", 2020 6th Conference on Data Science and Machine Learning Applications (CDMA), 77-80. 
  3. ASCE (2016), Minimum Design Loads for Buildings and Other Structures (ASCE Standard), American Society of Civil Engineers. 
  4. Azzi, Z., Habte, F., Elawady, A., Gan Chowdhury, A. and Moravej, M. (2020), "Aerodynamic mitigation of wind uplift on low-rise building roof using large-scale testing", Front. Built Environ. 5. 
  5. Banks, D., Meroney, R.N., Sarkar, P.P., Zhao, Z. and Wu, F. (2000), "Flow visualization of conical vortices on flat roofs with simultaneous surface pressure measurement", J. Wind Eng. Ind. Aerod., 84, 65-85.  https://doi.org/10.1016/S0167-6105(99)00044-6
  6. Banks, D., Sarkar, P.P., Wu, F. and Meroney, R.N. (2001), "A device to mitigate vortex induced rooftop suction", Amer. Conference on Wind Engineering. 
  7. Bergstra, J. and Bengio, Y. (2012), "Random search for hyper-parameter optimization", J. Machine Learning Res., 13(2). 
  8. Bitsuamlak, G.T., Warsido, W., Ledesma, E. and Chowdhury, A.G. (2013), "Aerodynamic mitigation of roof and wall corner suctions using simple architectural elements", J. Eng. Mech., 139, 396-408.  https://doi.org/10.1061/(ASCE)EM.1943-7889.0000505
  9. Chen, Y., Ko, T., Shang, L., Chen, X., Jiang, X. and Li, Q. (2018), "An investigation of few-shot learning in spoken term classification", arXiv preprint arXiv:1812.10233. 
  10. Chen, Y., Liu, Z., Xu, H., Darrell, T. and Wang, X. (2021), "Meta-baseline: Exploring simple meta-learning for few-shot learning", In Proceedings of the IEEE/CVF International Conference on Computer Vision, 9062-9071. 
  11. Chowdhury, A.G., Zisis, I., Irwin, P., Bitsuamlak, G., Pinelli, J.-P., Hajra, B. and Moravej, M. (2017), "Large-scale experimentation using the 12-fan wall of wind to assess and mitigate hurricane wind and rain impacts on buildings and infrastructure systems", J. Struct. Eng., 143(7), 04017053. 
  12. Coffman Bradley, F., Main Joseph, A., Duthinh, D. and Simiu, E. (2010), "Wind effects on low-rise metal buildings: Database-assisted design versus ASCE 7-05 Standard Estimates", J. Struct. Eng., 136, 744-748.  https://doi.org/10.1061/(ASCE)ST.1943-541X.0000140
  13. Dongmei, H., Shiqing, H., Xuhui, H. and Xue, Z. (2017), "Prediction of wind loads on high-rise building using a BP neural network combined with POD", J. Wind Eng. Ind. Aerod., 170, 1-17.  https://doi.org/10.1016/j.jweia.2017.07.021
  14. Fei-Fei, L., Fergus, R. and Perona, P. (2006), "One-shot learning of object categories", IEEE Transact. Pattern Analysis Machine Intell., 28, 594-611.  https://doi.org/10.1109/TPAMI.2006.79
  15. Fernandez-Caban, P.L., Masters, F.J. and Phillips, B.M. (2018), "Predicting roof pressures on a low-rise structure from freestream turbulence using artificial neural networks", Front. Built Environ., 4, 68. 
  16. Finn, C., Abbeel, P. and Levine, S. (2017), "Model-agnostic meta-learning for fast adaptation of deep networks", International Conference on Machine Learning, 1126-1135. 
  17. 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, 658-664.  https://doi.org/10.1016/j.jweia.2011.02.008
  18. Hinton, G.E., Plaut, D.C. (1987), Using fast weights to deblur old memories, In Proceedings of the 9th annual conference of the cognitive science society, 177-186. 
  19. Ho, T.C.E., Surry, D., Morrish, D. and Kopp, G.A. (2005), "The UWO contribution to the NIST aerodynamic database for wind loads on low buildings: Part 1. Archiving format and basic aerodynamic data", J. Wind Eng. Ind. Aerod., 93, 1-30.  https://doi.org/10.1016/j.jweia.2004.07.006
  20. Holmes, J.D. (2007), Wind Loading of Structures. CRC press.
  21. Hu, G., Liu, L., Tao, D., Song, J., Tse, K.T. and Kwok, K.C. (2020), "Deep learning-based investigation of wind pressures on tall building under interference effects", J. Wind Eng. Ind. Aerod., 201, 104138. 
  22. Kopp, G.A., Mans, C. and Surry, D. (2005), "Wind effects of parapets on low buildings: Part 4. Mitigation of corner loads with alternative geometries", J. Wind Eng. Ind. Aerod., 93, 873-888.  https://doi.org/10.1016/j.jweia.2005.08.004
  23. 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. 
  24. Lang, L., Tiancai, L., Shan, A. and Xiangyan, T. (2021), "An improved random forest algorithm and its application to wind pressure prediction", Int. J. Intell. Syst., 36, 4016-4032.  https://doi.org/10.1002/int.22448
  25. Levitan, M.L., Mehta, K.C., Vann, W.P. and Holmes, J.D. (1991), "Field measurements of pressures on the Texas tech building", J. Wind Eng. Ind. Aerod., 38, 227-234.  https://doi.org/10.1016/0167-6105(91)90043-V
  26. Li, L., Jamieson, K., DeSalvo, G., Rostamizadeh, A. and Talwalkar, A. (2017), "Hyperband: A novel bandit-based approach to hyperparameter optimization", J. Machine Learning Res., 18, 6765-6816. 
  27. Li, Y. and Yang, J. (2021), "Meta-learning baselines and database for few-shot classification in agriculture", Comput. Electron. Agriculture, 182, 106055. 
  28. Liu, Y., Qin, H., Zhang, Z., Pei, S., Jiang, Z., Feng, Z. and Zhou, J. (2020), "Probabilistic spatiotemporal wind speed forecasting based on a variational Bayesian deep learning model", Appl. Energy, 260, 114259. 
  29. Lu, Y., Peng, X., Yang, D., Yang, M. and Zhong, W. (2021), "Model-agnostic meta-learning with optimal alternative scaling value and its application to industrial soft sensing", IEEE Transact. Indus. Inform., 17, 8003-8013.  https://doi.org/10.1109/TII.2021.3058426
  30. Luo, X. and Kareem, A. (2020), "Bayesian deep learning with hierarchical prior: Predictions from limited and noisy data", Struct. Safety, 84, 101918. 
  31. Mallick, M., Mohanta, A., Kumar, A. and Charan Patra, K. (2020), "Prediction of wind-induced mean pressure coefficients using GMDH neural network", J. Aeros. Eng., 33, 04019104. 
  32. Marc, L.L. and Kishor, C.M. (1992a), "Texas tech field experiments for wind loads part 1: building and pressure measuring system", J. Wind Eng. Ind. Aerod., 43, 1565-1576.  https://doi.org/10.1016/0167-6105(92)90372-H
  33. Marc, L.L. and Kishor, C.M. (1992b), "Texas tech field experiments for wind loads part II: meteorological instrumentation and terrain parameters", J. Wind Eng. Ind. Aerod., 43, 1577-1588.  https://doi.org/10.1016/0167-6105(92)90373-I
  34. Marc, L.L., Kishor, C.M., Chee, V.C. and Daniel, L.M. (1990), "An overview of Texas Tech's wind engineering field research laboratory", J. Wind Eng. Ind. Aerod., 36, 1037-1046.  https://doi.org/10.1016/0167-6105(90)90100-Q
  35. Moravej, M. (2018), Investigating Scale Effects on Analytical Methods of Predicting Peak Wind Loads on Buildings, Ph.D. Dissertation, Florida International University, Miami. 
  36. Ni, Y. and Li, M. (2016), "Wind pressure data reconstruction using neural network techniques: A comparison between BPNN and GRNN", Measurement, 88, 468-476.  https://doi.org/10.1016/j.measurement.2016.04.049
  37. Nichol, A., Achiam, J. and Schulman, J. (2018), "On first-order meta-learning algorithms", arXiv preprint arXiv:1803.02999. 
  38. Smith, D., Mehta, K. and Morse, S. (2018), Wind Engineering Research Field Laboratory Selected Data Sets for Comparison to Model-Scale, Full-Scale and Computational Fluid Dynamics Simulations, DesignSafe-CI. 
  39. Snell, J., Swersky, K. and Zemel, R.S. (2017), "Prototypical networks for few-shot learning", arXiv preprint arXiv:1703.05175. 
  40. Sung, F., Yang, Y., Zhang, L., Xiang, T., Torr, P.H.S. and Hospedales, T.M. (2018), "Learning to compare: Relation network for few-shot learning", In Proceedings of the IEEE conference on computer vision and pattern recognition, 1199-1208. 
  41. Tieleman, H.W., Akins, R.E. and Sparks, P.R. (1981), "A comparison of wind-tunnel and full-scale wind pressure measurements on low-rise structures", J. Wind Eng. Ind. Aerod., 8, 3-19.  https://doi.org/10.1016/0167-6105(81)90003-9
  42. van de Lindt, J.W. and Dao, T.N. (2009), "Performance-based wind engineering for wood-frame buildings", J. Struct. Eng., 135, 169-177.  https://doi.org/10.1061/(ASCE)0733-9445(2009)135:2(169)
  43. Vinyals, O., Blundell, C., Lillicrap, T. and Wierstra, D. (2016), "Matching networks for one shot learning", Adv. Neural Inform. Processing Syst., 29, 3630-3638. 
  44. Wang, J. and Kopp Gregory, A. (2021), "Comparisons of aerodynamic data with the main wind force-resisting system provisions of ASCE 7-16. I: Low-Rise Buildings", J. Struct. Eng., 147, 04020347. 
  45. Wang, L.X., Huang, K., Ma, T., Gu, Q. and Huang, J. (2021), "Variance-reduced first-order meta-learning for natural language processing tasks", In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 2609-2615. 
  46. Wang, X.J., Li, Q.S. and Yan, B.W. (2020a), "Full-scale measurements of wind pressures on a low-rise building during typhoons and comparison with wind tunnel test results and aerodynamic database", J. Struct, Eng., 146(10), 04020196. 
  47. Wang, Y., Yao, Q., Kwok, J.T. and Ni, L.M.J.A.c.s. (2020b), "Generalizing from a few examples: A survey on few-shot learning", ACM computing surveys (csur), 53(3), 1-34  https://doi.org/10.1145/3386252
  48. Weng, Y. and Paal, S.G. (2022), "Machine learning-based wind pressure prediction of low-rise non-isolated buildings", Eng. Struct., 258, 114148. 
  49. Yang, L., Gurley, K.R. and Prevatt, D.O. (2013), "Probabilistic modeling of wind pressure on low-rise buildings", J. Wind Eng. Ind. Aerod., 114, 18-26. https://doi.org/10.1016/j.jweia.2012.12.014.