Automatic Classification of Earthquake-Induced Building Damages

The scope of this study is the creation of an abstract model based on intelligent methods for the automatic damage assessment of buildings after the occurrence of an earthquake. To develop such a model, an extensive set of seismic data is required. Unfortunately, the availability of natural records is often limited either because of the absence of nearby seismic stations or because the region has low-seismic variations. From the seismic data, seismic intensity parameters truly representative of the expected structural damage are extracted. Structural damage indices are used to quantify the damage potential. This leads to the need to identify suitable parameters to accurately convey structural damage information. This study provides methodologies that circumvent the aforementioned limitations; 1) four techniques for seismic accelerograms generation so as to enrich the seismic data set, 2) a method for seismic intensity parameters extraction strongly related to the damage potential and 3) four intelligent models for automatic assessment of structural damage. The methodologies are tested on a reinforced concrete (RC) frame structure.