Please use this identifier to cite or link to this item:
https://hdl.handle.net/20.500.13091/4878
Title: | Smart Method Recommendations for The Detection of Post-Earthquake Damages in RC Buildings | Authors: | Dogan, G. Arslan, M.H. Ilki, A. |
Keywords: | Convolutional Neural Networks Damage Deep learning Earthquake Reinforced Concrete Computational methods Concrete buildings Convolutional neural networks Damage detection Decision making Decision support systems Deep learning Earthquake engineering Earthquakes Engineering geology Learning systems Active seismic Convolutional neural network Damage Damage assessments Decision-based Deep learning Earthquake damages Psychological factors Reinforced concrete buildings Seismic zones Reinforced concrete |
Publisher: | National Technical University of Athens | Abstract: | In Turkey, which is located on an active seismic zone, the existing old reinforced concrete (RC) buildings have been extensively and severely damaged in the past earthquakes. In this context, in the event of a possible earthquake such as the 6 February 2023 Kahramanmaras earthquakes (Mw=7.7 and 7.6), it is extremely important to conduct the damage assessment promptly and in an objective/homogenous way. Although there are advanced methods developed for the assessment of post-earthquake damages to buildings in Turkey, there may often be differences/subjectiveness in the damage decisions based on the experience of the assessment staff and the psychological factors in the field. In this respect, introduction of intelligent decision support systems that will accelerate and harmonize the decision-making process of the engineers/technical staff involved in damage assessment activities after earthquakes may be very beneficial. In this study, firstly, information about the preliminary studies on the use of smart systems in structural engineering problems was given. Afterwards, smart software developed to be used in post-earthquake damage assessment of reinforced concrete buildings were compared according to their success/accuracy rates. In the evaluation, it has been seen that pre-trained deep learning models have a very high success in predicting post-earthquake damages in reinforced concrete structures. © 2023 COMPDYN Proceedings. All rights reserved | Description: | 9th ECCOMAS Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering, COMPDYN 2023 -- 12 June 2023 through 14 June 2023 -- 193215 | URI: | https://hdl.handle.net/20.500.13091/4878 | ISSN: | 2623-3347 |
Appears in Collections: | Scopus İndeksli Yayınlar Koleksiyonu / Scopus Indexed Publications Collections |
Show full item record
CORE Recommender
Items in GCRIS Repository are protected by copyright, with all rights reserved, unless otherwise indicated.