Hybrid data‐driven hazard‐consistent drift models for SMRF

Abstract

The seismic design and assessment of steel moment resisting frames (SMRFs) rely heavily on drifts. It is unsurprising, therefore, that several simplified methods have been proposed to predict lateral deformations in SMRFs, ranging from the purely mechanics‐based to the wholly data‐driven, which aim to alleviate the structural engineer's burden of conducting detailed nonlinear analyses either as part of preliminary design iterations or during regional seismic assessments. While many of these methods have been incorporated in design codes or are commonly used in research, they all suffer from a lack of consideration of the causal link between the seismic hazard level and the ground‐motion suite used for their formulation. In this paper, we propose hybrid data‐driven models that preserve this critical relationship of hazard‐consistency. To this end, we assemble a large database of non‐linear response history analyses (NRHA) on 24 SMRFs of different structural characteristics. These structural models are subjected to 816 ground‐motion records whose occurrence rates and spectral shapes are selected to ensure the hazard consistency of our outputs. Two sites with different seismic hazards are examined to enable comparisons under different seismic demands. An initial examination of the resulting drift hazard curves allows us to re‐visit the influence of salient structural modelling assumptions such as plastic resistance, geometric configurations and joint deterioration modelling. This is followed by a machine learning (ML)‐guided feature selection process that considers structural and seismic parameters as well as key static response features, hence the hybrid nature of our models. New models for inter‐storey drift and roof displacements are then developed. A comparison with currently available formulations highlights the significant levels of overestimation associated with previously proposed non‐hazard consistent models.

A Hazard-consistent Residual Drift Model for SMRF

“…4(c) shows five feature subsets of two features with the highest performance score based on exhaustive feature selection, and this method also shows a high importance of the θ max . It can also be seen that the performance of the models (R 2 ) only rises up to 0.7, which is lower than the performance of the θ max model found in [18] that can reach 0.83, confirming the higher variability of θ res compared to θ max . The filter and embedded methods results (figures not shown in this paper) also indicate a very high importance of θ max in θ res predictions.…”

“…The CSS has been used to assess reinforced concrete structures in [14,16,17]. However, to date, CSS application to steel structures has only been performed for peak transient drifts [18], and no study has used CSS to establish hazard-consistent estimations of permanent residual drifts.…”

“…In the following text, frames will be referenced with their number of storeys and bays; e.g., Frame 24 (12S-6B) means Frame 24 has 12 storeys and 6 bays. Other structural details of the frames are not discussed in this paper for brevity, and they can be found in Zahra et al [18]. The frames were analysed using the open-source software OpenSees [22] on the highperformance computing facility at Imperial College London [23].…”

“…For this purpose, several Machine Learning (ML) algorithms are used, and they can be classified into filter, wrapper, and embedded types. Further details of the process employed for each method can be found in [18]. The filter method, e.g.…”

“…The θ max values needed in our model and FEMA P-58's model [1] were taken from the same hazard-consistent θ max demand Figure 5 Observed θ res from NRHA vs θ res from the predictive model. model proposed by Zahra et al [18] to facilitate comparisons. The mean period, T m , needed for Bravo-Haro and Elghazouli's model [12] was estimated using the most recent GMPE developed by Du [47].…”