Buy article PDF
The purchased file will be sent to you
via email after the payment is completed.
US$ 35
|
Steel and Composite Structures Volume 54, Number 4, February 25 2025 , pages 295-312 DOI: https://doi.org/10.12989/scs.2025.54.4.295 |
|
|
|
Machine learning-based safety assessment of steel frames under seismic loadings using nonlinear time-history analysis |
||
Viet-Hung Truong, Sawekchai Tangaramvong, Manh-Cuong Nguyen and Hoang-Anh Pham
|
||
| Abstract | ||
| This research assesses various machine learning (ML) algorithms for predicting steel frame safety during seismic events. The study employs time-history dynamic analysis with plastic-hinge beam-column elements to form a training dataset, consisting of W-steel section properties for input variables and the safety labels based on maximum seismic interstory drift for output. Nine ML algorithms are tested on three different steel frames under earthquake conditions, including Naïve_Bayes, gaussian process, k-nearest neighbors (kNN), support vector machines (SVM), deep learning (DL), random forest (RF), gradient tree boosting (GTB), extreme gradient boosting (XGBoost), and light gradient boosting machines (LightGBM). Findings indicate that in view of accuracy, Naïve_Bayes, SVM, and kNN perform weaker compared to tree-based methods like RF, GTB, XGBoost, and LightGBM, with XGBoost and LightGBM outperforming. Regarding training time, Naïve_Bayes and kNN show the shortest times, whereas Gaussian, GTB, and DL experience substantial increases as the training dataset grows. RF and XGBoost have moderate times, but LightGBM is notably efficient. Overall, LightGBM demonstrates superior performance, followed by XGBoost and RF in predicting steel frame safety during seismic events. | ||
| Key Words | ||
| machine learning classification; seismic loadings; steel frame; safety assessment; time-history dynamic analysis | ||
| Address | ||
| Viet-Hung Truong:1)Faculty of Civil Engineering, Thuyloi University, 175 Tay Son, Dong Da, Hanoi 100000, Vietnam 2)Center of Excellence in Applied Mechanics and Structures, Department of Civil Engineering, Faculty of Engineering, Chulalongkorn University, Bangkok 10330, Thailand Sawekchai Tangaramvong:Center of Excellence in Applied Mechanics and Structures, Department of Civil Engineering, Faculty of Engineering, Chulalongkorn University, Bangkok 10330, Thailand Manh-Cuong Nguyen:Faculty of Civil Engineering, Thuyloi University, 175 Tay Son, Dong Da, Hanoi 100000, Vietnam Hoang-Anh Pham:1)Department of Structural Mechanics, Hanoi University of Civil Engineering, 55 Giai Phong Road, Hanoi 100000, Vietnam 2)Frontier Research Group of Mechanics of Advanced Materials and Structures (MAMS), Hanoi University of Civil Engineering, 55 Giai Phong Road, Hanoi 100000, Vietnam | ||