Machine learning model for building seismic peak roof drift ratio assessment

Abstract

AbstractThe peak roof drift ratio is one of the most important engineering parameters to describe the expected seismic damage in a building. A predictive model of the drift ratio was developed using a machine learning approach (Gaussian process regression model) on a dataset of approximately 11,800 records from 34 monitored buildings in Japan. Four predictors for ground motion and three predictors for building vulnerability are used in the machine-learning modelling. The residual analysis shows a reduction of 50% compared to the state of the art. The Gaussian process regression model is applied in a second analysis on an original dataset of approximately 4,500 records for 127 monitored buildings in Italy. A satisfactory comparison emerges by comparing the drift ratio prediction map with the observed damage pattern produced by satellite imagery for a test site in central Italy after the 2009 earthquake. The drift ratio map plays an important role in the simulation of an earthquake scenario at regional scale, which is needed by Civil Protection for emergency planning and management activities.

Keywords

• Seismic risk mitigation
• machine learning
• Gaussian process regression model
• building roof drift ratio
• regional scale