An Assessment Strategy for Masonry Buildings

Date

2025-9

Author

Özkan, Tolgahan

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

136
views

0
downloads

Cite This

Masonry buildings are widely prevalent in Turkey due to their ease of construction and low cost. However, these structures are often built without professional engineering input, making them one of the most vulnerable building types in terms of seismic performance. In seismically active regions such as Turkey, it is crucial to rapidly and reliably assess the seismic safety of such buildings and implement necessary mitigation measures. This thesis proposes a rapid vulnerability assessment strategy for unreinforced masonry buildings by utilizing both a multi-layer artificial neural network (ANN) and a single-layer regression model trained on 61,037 data points derived from real building records and logic-based data augmentation. In addition to vulnerability classification, the models are also employed to predict the fundamental vibration periods of buildings using the same set of parameters. Vulnerability labeling is conducted in accordance with the provisions of the “Regulation on the Identification of Vulnerable Buildings” (RYTEIE, 2019). Model interpretability is achieved through SHAP analysis, revealing the influence of individual parameters on the predictions. A comparative evaluation of the ANN and the linear model is also provided to explore the complexity of the relationship between structural parameters and vulnerability level. The results demonstrate that the proposed models can reliably and efficiently identify the seismic vulnerability level of unreinforced masonry buildings with high accuracy.

Subject Keywords

Unreinforced Masonry Buildings, Seismic Vulnerability Assessment, Multi-layer Artificial Neural Networks, Rapid Evaluation, Vulnerability Classification

URI

https://hdl.handle.net/11511/116156

Collections

Graduate School of Natural and Applied Sciences, Thesis