%0 Journal Article
%T Evaluation of Behavior Factors for Steel Moment Frames under Critical Consecutive Earthquakes using Artificial Neural Network
%J Amirkabir Journal of Civil Engineering
%I Amirkabir University of Technology
%Z 2588-297X
%A Rouzrokh, Sahar
%A Rajabi, Elham
%A Ghodrati Amiri, Gholamreza
%D 2021
%\ 10/23/2021
%V 53
%N 8
%P 3517-3534
%! Evaluation of Behavior Factors for Steel Moment Frames under Critical Consecutive Earthquakes using Artificial Neural Network
%K Critical Successive Earthquakes
%K behavior factor
%K Steel Moment Frame
%K Incremental dynamic analysis
%K Artificial Neural Networks
%R 10.22060/ceej.2020.18011.6737
%X Structures that are located in seismic active regions are often subjected to successive earthquakes which occurred with significant PGA in a short time after each other. Studies about different responses of the structures under seismic sequence phenomena, such as structural damage, ductility, displacement, and behavioral factor indicate that the successive earthquakes, depending on their severity, have significant effects on the different demands of structures. For instance, the behavior factor (R factor) is one of the significant parameters in the study of structural response that decreases the lateral forces induced by earthquakes. Therefore, the structure with non-elastic deformations absorbs a great amount of earthquake energy, thus the earthquake energy decreases considerably. Regarding the potential loss of successive earthquakes and the importance of behavioral factors, this paper calculates and estimates this parameter for steel moment frames under critical successive earthquakes. Thus, three steel moment frames with 3, 7, and 11 stories are designed according to Iranian seismic codes (standard No. 2800) and modeled in OpenSEES software. After the design of these frames, critical seismic scenarios with/without successive shocks, are selected and the R factors of steel moment frames are calculated from the results of incremental dynamic analysis (IDA(, time history, and nonlinear static analysis (pushover). The results showed about a 12% reduction in the R factor and, also an increment of damages under successive earthquakes comparing to the individual one. Finally, to estimate the R factor, artificial neural networks are designed using frame properties, successive earthquakes, and extracted behavior factors. The comparison of predicted behavior factors with real values indicated the ability of networks for the estimation of results.
%U https://ceej.aut.ac.ir/article_4047_f59f7dbb550a7779064d09a70a763873.pdf