Meta databases of steel frame buildings for surrogate modelling and machine learning-based feature importance analysis

Abstract

Traditionally, nonlinear time history analysis (NLTHA) is used to assess the performance of structures under fu-ture hazards which is necessary to develop effective disaster risk management strategies. However, this method is computationally intensive and not suitable for analyzing a large number of structures on a city-wide scale. Surrogate models offer an efficient and reliable alternative and facilitate evaluating the performance of multiple structures under different hazard scenarios. However, creating a comprehensive database for surrogate mod-elling at the city level presents challenges. To overcome this, the present study proposes meta databases and a general framework for surrogate modelling of steel structures. The dataset includes 30,000 steel moment-resisting frame buildings, representing low-rise, mid-rise and high-rise buildings, with criteria for connections, beams, and columns. Pushover analysis is performed and structural parameters are extracted, and finally, incorporating two different machine learning algorithms, random forest and Shapley additive explanations, sensitivity and explain-ability analyses of the structural parameters are performed to identify the most significant factors in designing steel moment resisting frames. The framework and databases can be used as a validated source of surrogate modelling of steel frame structures in order for disaster risk management.

Keywords

Surrogate models / Meta database / Pushover analysis / Steel moment resisting frames Sensitivity and explainability analyses Machine learning

References

[1] Ghobarah A. Performance-based design in earthquake engineering: state of develop-ment. Eng Struct 2001. doi: 10.1016/S0141-0296(01)00036-0.

[2] Porter KA. An overview of PEER’s performance-based earthquake engineering methodology. 9th Int Conf Appl Stat Probab Civ Eng 2003;273:973-80.

[3] Bertero RD, Bertero VV. Performance-based seismic engineering: the need for a reli-able conceptual comprehensive approach. Earthq Eng Struct Dyn 2002;31:627-52. doi: 10.1002/eqe.146.

[4] Montuori R, Nastri E, Piluso V, Todisco P. A simplified performance based ap-proach for the evaluation of seismic performances of steel frames. Eng Struct 2020;224:111222.

[5] Lin YY, Miranda E. Estimation of maximum roof displacement demands in regular multistorey buildings. J Eng Mech 2010;136(1):1-11.

[6] Applied Technology Council. Quantification of building seismic performance fac-torsTechnical report FEMA 695. Washington, DC: Federal Emergency Management Agency; 2009.

[7] Applied Technology Council. Seismic performance assessment of buildings volume 1 -methodologyTechnical report FEMA P58-1. Washington, DC: Federal Emergency Management Agency; 2012.

[8] Applied Technology Council. Seismic performance assessment of buildings volume 2 -implementation guideTechnical report FEMA P58-2. Washington, DC: Federal Emergency Management Agency; 2012.

[9] Zhong K, Navarro JG, Govindjee S, Deierlein GG. Surrogate modelling of structural seismic response using probabilistic learning on manifolds. Earthq Eng Struct Dyn 2023;52:2407-28. doi: 10.1002/eqe.3839.

[10] Sudret B, Chu VM. Computing seismic fragility curves using polynomial chaos expan-sions. 11th International Conference on Structural Safety and Reliability (ICOSSAR 2013); 2013. Eidgenössische Technische Hochschule Zürich.

[11] Gidaris I, Alexandros AT, Mavroeidis GP. Kriging metamodelling in seismic risk assessment based on stochastic ground motion models. Earthq Eng Struct Dyn 2015;44(14):2377-99.

[12] Gidaris I, et al. Multi-objective risk-informed design of floor isolation systems. Earthq Eng Struct Dyn 2016;45(8):1293-313.

[13] Du A, Padgett JE. Investigation of multivariate seismic surrogate demand mod-elling for multi-response structural systems. Eng Struct 2020;207:110210. doi: 10.1016/j.engstruct.2020.110210.

[14] Vaseghiamiri S, Mahsuli M, Ghannad MA, Zareian F. Surrogate SDOF models for probabilistic performance assessment of multistorey buildings: methodology and application for steel special moment frames. Eng Struct 2020;212:110276. doi: 10.1016/j.engstruct.2020.110276.

[15] Gudipati VK, Cha EJ. Surrogate modelling for structural response prediction of a building class. Struct Saf 2021;89:102041. doi: 10.1016/j.strusafe.2020.102041.

[16] Aristizabal C, Lopez-caballero F. Comparison between two surrogate models for embankment earthquake-liquefaction-induced settlements prediction. 13th Interna-tional Conference on Applications of Statistics and Probability in Civil Engineering (ICASP13); 2019.

[17] Dang-Vu H, Nguyen QD, Chung T, Shin J, Lee K. Frequency-based Data-driven Sur-rogate Model for Efficient Prediction of Irregular Structure’s Seismic Responses. J Earthq Eng 2022;26:7319-36. doi: 10.1080/13632469.2021.1961940.

[18] Cavalagli N, Pepi C, Gioffrè M, Gusella V, Ubertini F. Surrogate models for earthquake-induced damage detection and localization in historic structures using long-term dynamic monitoring data: application to a masonry dome. In: Proceed-ing of the 7th international conference on computing methods structure dynamics earthquake engineering (COMPDYN 2015), Athens: Institute of Structural Analysis and Antiseismic Research School of Civil Engineering National Technical University of Athens (NTUA) Greece; 2019. p. 1329-43. doi: 10.7712/120119.7001.19117.

[19] Micheli L, Alipour A, Laflamme S. Multiple-surrogate models for probabilistic per-formance assessment of wind-excited tall buildings under uncertainties. ASCE-ASME J Risk Uncertainty Eng Syst, Part A: Civil Eng 2020;6(4):04020042.

[20] Micheli L, Alipour A, Laflamme S. Data-driven risk-based assessment of wind-excited tall buildings. Structures congress 2019: Blast, impact loading, and research and education. Reston, VA: American Society of Civil Engineers; 2019.

[21] Micheli L, et al. Surrogate models for high performance control systems in wind-ex-cited tall buildings. Appl Soft Comput 2020;90:106133.

[22] Javidan MM, Kang H, Isobe D, Kim J. Computationally efficient framework for probabilistic collapse analysis of structures under extreme actions. Eng Struct 2018;172:440-52. doi: 10.1016/j.engstruct.2018.06.022.

[23] Zhang R, Wang D, Qu C. Selection and modification of ground motion records using a weighted scaling method based on the Newmark-Hall target spectrum. Structures 2022;44:1546-64. doi: 10.1016/j.istruc.2022.08.088.

[24] Zheng XW, Hong-Nan L, Sh Zhong-Qi. Hybrid AI-Bayesian-based demand models and fragility estimates for tall buildings against multi-hazard of earthquakes and winds. Thin Walled Struct 2023;187:110749.

[25] Xing L, Gardoni P, Zhou Y. Kriging metamodels for the dynamic response of high-rise buildings with outrigger systems and fragility estimates for seismic and wind loads. Resil Cities Struct 2022;1:110-22. doi: 10.1016/j.rcns.2022.04.003.

[26] Zaker Esteghamati M, Flint MM. Developing data-driven surrogate models for holis-tic performance-based assessment of mid-rise RC frame buildings at early design. Eng Struct 2021;245:112971. doi: 10.1016/j.engstruct.2021.112971.

[27] Bhosekar A, Ierapetritou M. Advances in surrogate based modelling, feasibility anal-ysis, and optimization: a review. Comput Chem Eng 2018;108:250-67.

[28] Bruneau M, Uang CM, Sabelli R. Ductile design of steel structures, New York; Toronto: McGraw-Hill; 2011 http://www.myilibrary.com?id=335745 [Accessed 1 Feb 2023].

[29] American Institute of Steel Construction AISC Specification for structural steel buildings. ansi/aisc 360-16. Chicago, IL: American Institute of Steel Construction; 2016.

[30] SAC Joint Venture Technical Report November; 2000.

[31] SAC Joint Venture Technical report; 2000.

[32] Sawada Y, Shimizutani S.How do people cope with natural disasters? Evidence from the great Hanshin-Awaji (Kobe) earthquake in 1995. J Money, Credit Bank 2008;40(2-3):463-88.

[33] Rose A, Lim D. Business interruption losses from natural hazards: conceptual and methodological issues in the case of the Northridge earthquake. Environ Hazards 2002;4(1):1-14.

[34] Engelhardt M, Sabol T. Seismic-resistant steel moment connections: developments since the 1994 Northridge earthquake. Prog Struct Mater Eng 1997;1(1):68-77.

[35] Horton TA, Hajirasouliha I, Davison B, Ozdemir Z, Abuzayed I. Development of more accurate cyclic hysteretic models to represent RBS connections. Eng Struct 2021;245:112899.

[36] Stylianidis P, Bellos J. Survey on the Role of Beam-Column Connections in the Pro-gressive Collapse Resistance of Steel Frame Buildings. Buildings 1696;13(7):2023.

[37] American Institute of Steel Construction, prequalified connections for special and intermediate steel moment frames for seismic applications, ANSI/AISC.358-16.

[38] Guan Xingquan, et al. Seismic drift demand estimation for steel moment frame buildings: from mechanics-based to data-driven models. J Struct Eng 2021;147(6):04021058.

[39] ASCE 7-16 Minimum design loads and associated criteria for buildings and other structures. RestonVA: American Society of Civil Engineers (ASCE); 2016.

[40] Hamburger RO, Malley JO. Seismic design of steel special moment frames. NIST GCR 2009:09-917.

[41] Applied Technology CouncilQuantification of building seismic performance factors, US Department of Homeland Security. FEMA 2009;P695.

[42] American Institute of Steel Construction AISC, Seismic provisions for structural steel buildings, ANSI/AISC. 2016:341-16. Chicago, IL, USA, 2010

[43] Kircher C., Deierlein G., Hooper J., Krawinkler H., Mahin S., Shing B., & Wallace J., Evaluation of the FEMA P-695 methodology for quantification of building seismic performance factors, 2010.

[44] McKenna F, Fenves GL, OpenSees ScottMH. Open system for earthquake engineering Simulation. Pac Earthq Eng Res Center, Univ Calif 2006. http://OpenseesBerkeleyEdu.

[45] Ibarra LF, Medina RA, Krawinkler H. Hysteretic models that incorporate strength and stiffness Deterioration. Earthq Eng Struct Dyn 2005;34(12):1489-511.

[46] ASCE/SEI, ASCE 41-17 Seismic evaluation and retrofit of existing buildings. Reston, Virginia: American Society of Civil Engineers; 2017.

[47] Lignos DG. Sidesway collapse of deteriorating structural systems under seismic ex-citations. Stanford university; 2008.

[48] Lignos DG, Krawinkler H. Deterioration modelling of steel components in support of collapse prediction of steel moment frames under earthquake loading. J Struct Eng-Reston 2011;137(11):1291-302.

[49] Lignos DG, Krawinkler H. A database in support of modelling of component deteri-oration for collapse prediction of steel frame structures. In: Structuralengineering research.Frontiers; 2007. p. 1-12.

[50] Gupta A, Krawinkler H. Seismic demands for the performance evaluation of steel moment resisting frame structures. Berkeley, CA: The John A Blume Earthquake Engineering Center, Stanford University; 1998.

[51] Rou WAkbas B, Shen J. Practical moment-rotation relations of steel shear tab Con-nections. J Constr Steel Res 2013;88:296-308.

[52] Fayaz J, Zareian F. Reliability analysis of steel SMRF and SCBF structures con-sidering the vertical component of near-fault ground motions. J Struct Eng 2019;145(7):04019061.

[53] Ellingwood B, Galambos TV, MacGregor JG, Cornell CA. Development of a probabil-ity based load criterion for American National Standard A58. Washington, DC: US Dept of Commerce; 1980.

[54] FEMA P- 2012 Assessing seismic performance of buildings with configuration irreg-ularities. Redwood City, CA: Applied Technology Council; 2018.

[55] Barbato M, Gu Q, Conte JP. Probabilistic pushover analysis of structural and soil-structure systems. J Struct Eng 2010;136(11):1330-41.

[56] Kim J, Park J-H, Lee T-H. Sensitivity analysis of steel buildings subjected to column loss. Eng Struct 2011;33:421-32. doi: 10.1016/j.engstruct.2010.10.025.

[57] Zhang D, Lin X, Dong Y, Yu X. Machine-Learning-based uncertainty and sensitivity analysis of reinforced-concrete slabs subjected to fire. Structures 2023;53:581-94. doi: 10.1016/j.istruc.2023.04.030.

[58] Trong Ha N, Xuan Hung D. Sensitivity analysis of the design portal frames of steel industrial buildings. MATEC Web Conf 2018;193:04025. doi: 10.1051/matec-conf/201819304025.

[59] Javidian MM, Kim J. Fuzzy Sensitivity Analysis of Structural Performance. Sustain-ability 2022;14:11974. doi: 10.3390/su141911974.

[60] Rodríguez D, Brunesi E, Nascimbene R. Fragility and sensitivity analysis of steel frames with bolted-angle connections under progressive collapse. Eng Struct 2021;228:111508. doi: 10.1016/j.engstruct.2020.111508.

[61] Tipu Kumar, R Panchal VR, Pandya KS. An ensemble approach to improve BPNN model precision for predicting compressive strength of high-performance concrete. Structures 2022;45:500-8. doi: 10.1016/j.istruc.2022.09.046.

[62] Somala SN, Chanda S, Karthikeyan K, Mangalathu S. Explainable machine learning on New Zealand strong motion for PGV and PGA. Structures 2021;34:4977-85. doi: 10.1016/j.istruc.2021.10.085.

[63] Le HA, Le DA, Le TT, Le HP, Le TH, Hoang HGT, et al. An extreme gradient boosting approach to estimate the shear strength of FRP reinforced concrete beams. Structures 2022;45:1307-21. doi: 10.1016/j.istruc.2022.09.112.

[64] Truong GT, Choi KK, Kim CS. Implementation of boosting algorithms for predic-tion of punching shear strength of RC column footings. Structures 2022;46:521-38. doi: 10.1016/j.istruc.2022.10.085.

[65] Rajbahadur GK, Wang S, Oliva GA, Kamei Y, Hassan AE. The impact of feature im-portance methods on the interpretation of defect classifiers. IEEE Trans Softw Eng 2021;48(7):2245-61.

[66] Saarela M, Jauhiainen S. Comparison of feature importance measures as explana-tions for classification models. SN Appl Sci 2021;3(272).

[67] Covert I, Lundberg S, Lee SI. Understanding global feature contributions with addi-tive importance measures. Adv Neural Inf Process Syst 2020;33:17212-23.

[68] Lundberg SM, Lee SI. A unified approach to interpreting model predictions. In: Pro-ceeding of the advances neural information processing system; 2017. p. 4765-74.

[69] Hooker G, Mentch L, Zhou S. Unrestricted permutation forces extrapolation: variable importance requires at least one more model, or there is no free variable importance. Stat Comput 2021;31:1-16.

[70] Breiman L. Random forests. Mach Learn 2001;45:5-32. doi: 10.1023/A:1010933404324

[71] Gromping U. Variable importance assessment in regression: linear regression versus random forest. Am Statistician 2009;63:308-19. doi: 10.1198/tast.2009.08199.

[72] Parhi SK, Patro SK. Prediction of compressive strength of geopolymer concrete using a hybrid ensemble of grey wolf optimised machine learning estimators. J Build Eng 2023;71:106521. doi: 10.1016/j.jobe.2023.106521.

[73] Mangalathu S, Hwang SH, Jeon JS. Failure mode and effects analysis of RC members based on machine-learning-based SHapley Additive exPlanations (SHAP) approach. Eng Struct 2020;219:110927. doi: 10.1016/j.engstruct.2020.110927.

[74] Lundberg SM, Lee SI. A unified approach to interpreting model predictions. In: Pro-ceedings of the 31st international conference on neural information processing sys-tems; 2017. p. 4768-77.

[75] Cover T, Hart P. Nearest neighbour pattern classification. IEEE Trans Inf Theory 1967;13(1):21-7. doi: 10.1109/TIT.1967.1053964.