\"Seismic damage assessment is a critical step to enhance community resilience in the wake of an earthquake. This study aims to develop deep learning-based surrogate models for widely used fragility curves to achieve more accurate and rapid assessment in practice. These surrogate models are based on artificial neural networks trained from the labelled ground motions whose resulting damage classes on targeted structures are determined by nonlinear time history analyses. The development of various surrogate models is progressed in four phases. In Phase I, the multilayer perceptron (MLP) is used to develop multivariate seismic classifiers with up to 50 hand-crafted intensity measures (IMs) as inputs when trained for the simultaneous fragility estimation and (both local and global) damage classification of a code-conforming benchmark reinforced concrete (r/c) frame building. In Phase II, a MLP seismic classifier is trained with 6 IMs and 2 structural parameters as inputs to consider both the primary earthquake and structural uncertainties in the portfolio-scale seismic damage assessment of one-story, residential woodframe structures near the New Madrid Seismic Zone. In Phase III, convolutional neural networks (CNNs) with encoded ground-motion images as inputs are trained for the benchmark r/c frame building to automatically extract features (e.g., IMs) of the ground motions and thus avoid the hand-crafted IMs. In Phase IV, new one-dimensional CNNs with raw ground motions as inputs are developed for the benchmark r/c frame building to further improve the computational efficiency of existing CNN-based seismic damage classifications. Based on the above studies, future research directions are identified to mature the developed models for applications in earthquake and wind engineering\"-- Abstract, p. iv
","abstract_html":"<p>"Seismic damage assessment is a critical step to enhance community resilience in the wake of an earthquake. This study aims to develop deep learning-based surrogate models for widely used fragility curves to achieve more accurate and rapid assessment in practice. These surrogate models are based on artificial neural networks trained from the labelled ground motions whose resulting damage classes on targeted structures are determined by nonlinear time history analyses. The development of various surrogate models is progressed in four phases. In Phase I, the multilayer perceptron (MLP) is used to develop multivariate seismic classifiers with up to 50 hand-crafted intensity measures (IMs) as inputs when trained for the simultaneous fragility estimation and (both local and global) damage classification of a code-conforming benchmark reinforced concrete (r/c) frame building. In Phase II, a MLP seismic classifier is trained with 6 IMs and 2 structural parameters as inputs to consider both the primary earthquake and structural uncertainties in the portfolio-scale seismic damage assessment of one-story, residential woodframe structures near the New Madrid Seismic Zone. In Phase III, convolutional neural networks (CNNs) with encoded ground-motion images as inputs are trained for the benchmark r/c frame building to automatically extract features (e.g., IMs) of the ground motions and thus avoid the hand-crafted IMs. In Phase IV, new one-dimensional CNNs with raw ground motions as inputs are developed for the benchmark r/c frame building to further improve the computational efficiency of existing CNN-based seismic damage classifications. Based on the above studies, future research directions are identified to mature the developed models for applications in earthquake and wind engineering"-- Abstract, p. iv</p>","abstract_has_math":false,"creators":["Yuan, Xinzhe"],"institution":"Missouri University of Science and Technology","degree_name":"Ph. D. in Civil Engineering","degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":[],"committee_chairs":[],"committee_members":[],"year":null,"date_issued":"","date_published":null,"updated_at":"2026-07-24T03:18:18Z","subjects":["Deep learning","earthquake damage","surrogate models","Civil Engineering","Computer Sciences","Engineering"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://scholarsmine.mst.edu/doctoral_dissertations/3306","outbound_label":"Repository record","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:creator","label":"Author","values":["Yuan, Xinzhe"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:type","label":"Dc Type","values":["Dissertation - Open Access"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph. D. in Civil Engineering"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["Missouri University of Science and Technology"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Deep learning","earthquake damage","surrogate models","Civil Engineering","Computer Sciences","Engineering"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://scholarsmine.mst.edu/doctoral_dissertations/3306"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["\"Seismic damage assessment is a critical step to enhance community resilience in the wake of an earthquake. This study aims to develop deep learning-based surrogate models for widely used fragility curves to achieve more accurate and rapid assessment in practice. These surrogate models are based on artificial neural networks trained from the labelled ground motions whose resulting damage classes on targeted structures are determined by nonlinear time history analyses. The development of various surrogate models is progressed in four phases. In Phase I, the multilayer perceptron (MLP) is used to develop multivariate seismic classifiers with up to 50 hand-crafted intensity measures (IMs) as inputs when trained for the simultaneous fragility estimation and (both local and global) damage classification of a code-conforming benchmark reinforced concrete (r/c) frame building. In Phase II, a MLP seismic classifier is trained with 6 IMs and 2 structural parameters as inputs to consider both the primary earthquake and structural uncertainties in the portfolio-scale seismic damage assessment of one-story, residential woodframe structures near the New Madrid Seismic Zone. In Phase III, convolutional neural networks (CNNs) with encoded ground-motion images as inputs are trained for the benchmark r/c frame building to automatically extract features (e.g., IMs) of the ground motions and thus avoid the hand-crafted IMs. In Phase IV, new one-dimensional CNNs with raw ground motions as inputs are developed for the benchmark r/c frame building to further improve the computational efficiency of existing CNN-based seismic damage classifications. Based on the above studies, future research directions are identified to mature the developed models for applications in earthquake and wind engineering\"-- Abstract, p. iv
"]},{"key":"dc:title","label":"Title","values":["Deep learning-based surrogate models for post-earthquake damage assessment"]}]}],"canonical_facts":{"dc:creator":["Yuan, Xinzhe"],"dc:description.abstract":["\"Seismic damage assessment is a critical step to enhance community resilience in the wake of an earthquake. This study aims to develop deep learning-based surrogate models for widely used fragility curves to achieve more accurate and rapid assessment in practice. These surrogate models are based on artificial neural networks trained from the labelled ground motions whose resulting damage classes on targeted structures are determined by nonlinear time history analyses. The development of various surrogate models is progressed in four phases. In Phase I, the multilayer perceptron (MLP) is used to develop multivariate seismic classifiers with up to 50 hand-crafted intensity measures (IMs) as inputs when trained for the simultaneous fragility estimation and (both local and global) damage classification of a code-conforming benchmark reinforced concrete (r/c) frame building. In Phase II, a MLP seismic classifier is trained with 6 IMs and 2 structural parameters as inputs to consider both the primary earthquake and structural uncertainties in the portfolio-scale seismic damage assessment of one-story, residential woodframe structures near the New Madrid Seismic Zone. In Phase III, convolutional neural networks (CNNs) with encoded ground-motion images as inputs are trained for the benchmark r/c frame building to automatically extract features (e.g., IMs) of the ground motions and thus avoid the hand-crafted IMs. In Phase IV, new one-dimensional CNNs with raw ground motions as inputs are developed for the benchmark r/c frame building to further improve the computational efficiency of existing CNN-based seismic damage classifications. Based on the above studies, future research directions are identified to mature the developed models for applications in earthquake and wind engineering\"-- Abstract, p. iv
"],"dc:identifier":["https://scholarsmine.mst.edu/doctoral_dissertations/3306"],"dc:subject":["Deep learning","earthquake damage","surrogate models","Civil Engineering","Computer Sciences","Engineering"],"dc:title":["Deep learning-based surrogate models for post-earthquake damage assessment"],"dc:type":["Dissertation - Open Access"],"thesis:degree_name":["Ph. D. in Civil Engineering"],"thesis:institution_name":["Missouri University of Science and Technology"]},"updated_at":"2026-07-24T03:18:18Z"}