Supervised probabilistic failure prediction of tuned mass damper-equipped high steel frames using machine learning methods

Abstract

In this study, firstly, the behavior of a high steel frame equipped with tuned mass damper (TMD) due to several seismic records is investigated considering the structural and seismic uncertainties. Then, machine learning methods including artificial neural networks (ANN), decision tree (DT), Naïve Bayes (NB) and support vector machines (SVM) are used to predict the behavior of the structure. Results showed that among the machine learning models, SVM with Gaussian kernel has better performance since it is capable of predicting the drift of stories and the failure probability with R2 value equal to 0.99. Furthermore, results of feature selection algorithms revealed that when using TMD in high steel structures, seismic uncertainties have greater influences on drift of stories in comparison with structural uncertainties. Findings of this study can be used in design and probabilistic analysis of high steel frames equipped with TMDs.

Keywords

• failure analysis
• supervised machine learning
• feature selection
• tuned mass damper