Detection of plan irregularities in earthquake resistant architectural design through machine learning and artificial intelligence integration

Accurately assessing the regularity and irregularity of structures during the design phase is critical to minimizing loss of life and property caused by earthquakes. This thesis aims to detect plan irregularities, such as A1 torsional irregularity and A2 floor discontinuity, defined under the Turkish Earthquake Code, using machine learning and image classification techniques.. In this study, a deep learning model was developed to automatically classify the regularity and irregularity of structural systems. The model was trained to detect A1 and A2 irregularities and was capable of accurately classifying building plans exhibiting these irregularities. The results show that the developed model has a high accuracy rate and can be used as a reliable tool for the early detection of irregularities in structural systems. This thesis demonstrates that the integration of machine learning algorithms into architectural design processes can significantly contribute to reducing earthquake risks. Future studies are recommended to focus on detecting different types of irregularities and improving the overall performance of the model.