Application of a DNN model for seismic performance prediction of structures retrofitted with steel dampers

The intensity and frequency of earthquakes in Korea have increased in the past few years. Thus, the need for seismic retrofit of many middle- or low-rise buildings has increased which were designed without seismic design provisions. Nonlinear time history analyses need to be performed for accurate seismic performance evaluation and for appropriate retrofit of structures. The NLTH analyses, however, requires significant computational time and modelling efforts. It is possible to predict the nonlinear response of structures subjected to an earthquake by constructing a database of the seismic response of structures equipped with various dampers. The database can be used to train a deep learning algorithm. In this study, a deep learning algorithm that can predict the seismic response of a retrofitted structure will be proposed based on the results of a large number of numerical analyses of structures equipped with steel dampers.

Faculty Supervisor:

Oh-Sung Kwon

Student:

Partner:

Kyungpook National University

Discipline:

Engineering

Sector:

Construction; Environmental Science and Technology; Sustainability & the Environment

University:

University of Toronto

Program: