Machine learning-based prediction of seismic limit-state capacity of steel moment-resisting frames considering soil-structure interaction

Kazemi, Farzin; Jankowski, Robert

doi:10.1016/j.compstruc.2022.106886

Cited by 40 publications

(7 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To properly introduce the proposed modeling approach, two low-to mid-rise steel frames with three-, and five-story floor levels were modeled according to Kazemi et al [22,23]. To model the steel frames, it was assumed that the structures had regular plan with five bays in each direction and it was possible to model them as 2D model having the leaning column to consider the p-delta effects (see more detail about modeling process in [24][25][26]).…”

Section: Models Of Structuresmentioning

confidence: 99%

Retrofitting Damaged Buildings Under Seismic ‎mainshock-Aftershock Sequence

Mohebi,

Kazemi,

Asgarkhani

2023

Proceedings of the 8th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering (COM

View full text Add to dashboard Cite

Mainshock-aftershock sequences are a common occurrence in areas with high seismic activity, and they can significantly increase the damage sustained by a building after the initial seismic event. During severe mainshock earthquake, there is a permanent deformation in the floor levels of the building that increases the risk of further damage and potential collapse in the aftershock earthquake. This research aims to introduce a computational method for retrofitting damaged buildings under seismic mainshock-aftershock sequences using Fluid Viscous Dampers (FVDs) and Buckling-Restrained Braces (BRBs). This retrofitting approach aims to mitigate further damage that may be caused by subsequent aftershocks. The use of computational methods enables engineers to accurately simulate the behavior of damaged structures under various seismic loads, which helps in the design and optimization of retrofitting techniques. Due to ability of FVDs and BRBs, two steel structures having four and six-story floor levels were selected to implement the computational method of retrofitting. The results of analysis confirmed that the mainshock-aftershock sequences can highly affect the performance of damaged buildings, and the proposed method can accurately model the damaged buildings with different pre-defined damage states.

show abstract

Section: Models Of Structuresmentioning

confidence: 99%

Retrofitting Damaged Buildings Under Seismic ‎mainshock-Aftershock Sequence

Mohebi,

Kazemi,

Asgarkhani

2023

Proceedings of the 8th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering (COM

View full text Add to dashboard Cite

show abstract

“…Perez-Ramirez et al (2019) succeeded in efficiently predicting structural seismic response time history based on a recurrent neural network. Kazemi and Jankowski (2023) achieved predicting seismic limit-state capacities of steel frames with supervised regression ML algorithms. Noureldin et al (2023) established an explainable neural network for probabilistic prediction of structure seismic response based on a huge NLTHA-generated database.…”

Section: Introductionmentioning

confidence: 99%

“…(2019) succeeded in efficiently predicting structural seismic response time history based on a recurrent neural network. Kazemi and Jankowski (2023) achieved predicting seismic limit‐state capacities of steel frames with supervised regression ML algorithms. Noureldin et al.…”

Section: Introductionmentioning

confidence: 99%

Prior knowledge‐infused neural network for efficient performance assessment of structures through few‐shot incremental learning

Chen,

Feng,

Taciroglu

2024

Computer aided Civil Eng

View full text Add to dashboard Cite

Structural seismic safety assessment is a critical task in maintaining the resilience of existing civil and infrastructures. This task commonly requires accurate predictions of structural responses under stochastic intensive ground accelerations via time‐costly numerical simulations. While numerous studies have attempted to use machine learning (ML) techniques as surrogate models to alleviate this computing burden, a large number of numerical simulations are still required for training ML models. Therefore, this study proposes a prior knowledge‐infused neural network (PKNN) for achieving efficient structural seismic response predictions and seismic safety assessment at a low computation cost. In this approach, first, the prior knowledge inherently within a theoretical dynamic model of a structure is infused into a neural network. Then, by utilizing the few‐shot incremental learning technique, this network would be further fine‐tuned by only a few numerical simulations. The resulting PKNN would be able to accurately predict the seismic response of a structure and facilitate seismic safety assessment. In this study, the PKNN's accuracy and computational efficiency are validated on a typical frame structure. The results revealed that the proposed PKNN could be used for accurately predicting the structural seismic responses and assessing the structural safety under a low computational cost.

show abstract

“…The effects of structural pounding during seismic excitations have been studied for more than three decades now (see, for example, [12][13][14][15][16][17][18]). However, most of the studies have concerned masonry as well as reinforced concrete structures and investigations on steel structures are quite limited (see [19][20][21]). Meanwhile, increased displacements observed in steel structures during ground motions due to their flexibility and low damping properties make them more vulnerable to collisions.…”

Section: Introductionmentioning

confidence: 99%

Experimental study on earthquake‐induced pounding between steel structures

Sołtysik,

Jankowski

2023

ce papers

View full text Add to dashboard Cite

Pounding between adjacent structures has been repeatedly observed during major earthquakes. The phenomenon may cause some local damage at the points of interactions, it may also lead to substantial destruction, permanent deformations or total collapse of colliding structures. The effects of earthquake‐induced structural pounding have been studied for more than three decades now. However, most of the studies concerned masonry as well as reinforced concrete structures and investigations on steel structures are very limited. The aim of the present paper is to show the results of the shaking table experimental investigation focused on earthquake‐induced pounding between three models of steel structures in a row. The models were constructed out of four steel columns made from rectangular box section with additional steel skew bracings so as to prevent transverse and torsional vibrations. To obtain different dynamic characteristics of the models, additional plates were added at the top of each tower. Two configurations were investigated. In the first one, two plates were mounted on the external structures and only one plate at the top of the middle tower. In the second configuration, only one plate was mounted at the top of external towers and two plates on the middle structure. Firstly, the free vibration tests were conducted. Then, the steel towers with different in‐between gap sizes were tested under different earthquakes. The results of the study confirm that interactions between insufficiently separated steel structures may significantly influence the structural behaviour during earthquakes. In most of the cases, pounding has resulted in the increase in the peak responses. In some cases, however, it has played a positive role leading to the decrease in the structural response under seismic excitation. Anyway, even if the response is reduced, interactions should be prevented at the design stage of steel structures since they may result in considerable damage at the points of impacts.

show abstract

Machine learning-based prediction of seismic limit-state capacity of steel moment-resisting frames considering soil-structure interaction

Cited by 40 publications

References 42 publications

Retrofitting Damaged Buildings Under Seismic ‎mainshock-Aftershock Sequence

Retrofitting Damaged Buildings Under Seismic ‎mainshock-Aftershock Sequence

Prior knowledge‐infused neural network for efficient performance assessment of structures through few‐shot incremental learning

Experimental study on earthquake‐induced pounding between steel structures

Contact Info

Product

Resources

About