@inproceedings{Zahra:2022, author = {Zahra, F and Malaga, Chuquitaype C and Jorge, M}, pages = {431--439}, title = {Towards a hazard-consistent predictive model for drifts in steel MRFs}, year = {2022} }
TY - CPAPER AB - Performance-Based Earthquake Engineering (PBEE) approaches have been permitted by nearly every code for almost a century already, and the importance of developing accurate drift predictive models in support of PBEE is widely recognised. For this aim to be fully realised, it is crucial to identify the most influential structural and ground motion parameters that govern the seismic response. This study applies feature selection Machine Learning (ML) techniques to the identification of the best predictors of maximum inter-storey drift of steel Moment Resisting Frames (MRFs). Several ML techniques are applied to a database assembled from extensive results of nonlinear response history analyses on 24 steel MRFs of different structural characteristics. A suite of 596 ground motions is used to cover a wide range of intensities. These ground motions were selected by means of the Conditional Scenario Spectra (CSS) methodology to ensure the hazard consistency of the estimates, a key concept at the heart of the PBEE. Although the identified best predictors are not surprising, interesting conclusions are obtained from the application of the feature selection methods used in this study and the need for a careful interpretation of the results of ML tools is highlighted. AU - Zahra,F AU - Malaga,Chuquitaype C AU - Jorge,M EP - 439 PY - 2022/// SP - 431 TI - Towards a hazard-consistent predictive model for drifts in steel MRFs ER -