Probabilistic assessment of the contribution of geotechnical factors on observed structural damage in Adapazari and Düzce regions

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Fen Bilimleri Enstitüsü, Türkiye

Tezin Onay Tarihi: 2019

Tezin Dili: İngilizce

Öğrenci: BURAK YILDIZLI

Danışman: Zeynep Gülerce

Özet:

The 1999 Kocaeli and Düzce Earthquakes caused extensive damage to the structures in Adapazarı, Düzce and surrounding cities, leading to several attempts to estimate the damage states of the existing building stock using the data collected from these destructive events. Preliminary damage state prediction models that include the geotechnical earthquake engineering factors were developed for Adapazarı after the 1999 Kocaeli earthquake; however, validation exercises were not performed using the data collected from other cities for these earthquakes. In this study, a joint database of performed subsurface soil investigations and the building stock data including the damage states and structural parameters is compiled for Adapazarı and Düzce cities and the geotechnical earthquake engineering parameters are added to the compiled database by performing site-specific ground response and seismic soil liquefaction initiation analysis. Damage state prediction models that include the number of stories, peak ground acceleration, spectral acceleration, liquefaction severity index, and liquefaction induced settlement as predictive parameters are developed by linear discriminant analysis, multinomial logistics regression, and maximum likelihood analysis methods. Models are developed for the Adapazarı dataset that includes significantly higher number of buildings and tested for the Düzce dataset to evaluate the suitability of the selected regression approach and to analyze the contribution of ground motion and geotechnical parameters to the model performance. Analysis results showed that the number of stories is the parameter with the most significant effect on the predictive performance; while the geotechnical parameters increase the true prediction ratio in each damage state by 10%-15%. The final damage state prediction model proposed in this study estimates the damage states in Adapazarı and Düzce correctly by more than 60% for each damage state.