Yasin Fahjan

Effective disaster management depends on rapidly understanding earthquake damage, yet traditional methods struggle to operate at scale and rely on expert inspections that become difficult when access is limited or time is critical. Satellite-based damage detection also faces limitations, particularly under adverse weather conditions and delays associated with satellite overpass schedules. This study introduces a machine learning-based approach to assess post-earthquake building damage using real observations collected after the event. The aim is to develop fast and reliable estimation techniques that can be deployed immediately after the mainshock by integrating structural, seismic, and geographic data. Three machine learning models—Random Forest, Histogram Gradient Boosting, and Bagging Classifier—are evaluated across both reinforced concrete and masonry buildings and across multiple spatial levels, including building, district, and city scales. Damage is categorized using practical three-class (traffic light) and detailed four-class systems. The models generally perform better in simpler classifications, with the Bagging Classifier offering the most consistent results across different scales. Although detecting severely damaged buildings remains challenging in some cases, the three-class system proves especially effective for supporting rapid decision-making during emergency response. Overall, this study demonstrates how machine learning can provide faster, scalable, and practical earthquake damage assessments that benefit emergency teams and urban planners.

The capacity design philosophy primarily focuses on limiting the seismic shear force that will affect the structural system and has been widely accepted in the structural earthquake engineering community. However, several sources in the literature indicate that there are still some problematic points in this design philosophy. The main problem is that the proposed solution in this philosophy considers the first fundamental mode of the structural system but overlooks the effects of higher modes. This problem may lead to increased seismic shear force demand that is attempted to be limited by the capacity design philosophy. A similar problem with cantilever walls was first identified in the mid-1970s, and several solutions, which consider both dynamic effects and capacity design principles, were proposed in the relevant literature. Regarding Eccentrically Braced Frames (EBFs) with short link beam, no specific study on the dynamic amplification of shear force demand has been observed in the literature. However, other studies in different contexts have alluded to the possibility of such a phenomenon in EBFs and other steel frames. This study focuses on the dynamic amplification of shear force demands and determines its range using incremental dynamic analysis and multi-modal push-over analyses. Consequently, it has been demonstrated that limiting the shear force demand of the system, as recommended by the capacity design principles in all steel structure design specifications, is not possible. The use of a dynamic amplification factor is suggested to address non-ductile failure modes and enhance structural reliability. In this context, the current study examines the previously mentioned issue for eccentrically braced frames with short link beam in 4-story and 8-story buildings, representing low-rise and mid-rise buildings respectively. As a result, the seismic shear demands for EBFs found in Incremental Dynamic Analysis are significantly higher than those calculated by using capacity design principles for both 4-story and 8-story buildings. Additionally, the results from Incremental Dynamic Analysis have been comparatively examined with multi-modal push-over analyses. The internal force demands for elements of EBFs, especially in braces, increased due to the higher base shear demands, excluding the link beam.

Seismic hazard assessment relies upon a complete earthquake catalog, necessitating a declustering procedure to filter out dependent aftershocks or foreshocks from the earthquake catalog. Traditional declustering approaches, such as the Gardner-Knopoff approach, utilize magnitude-dependent temporal and spatial windows to distinguish dependent aftershocks/foreshocks from independent mainshocks. However, these approaches may overlook regional seismic characteristics, which leads to missing out primary independent events occurring within the time and space window following a strong earthquake, particularly when the background rate is elevated. This study proposes a modified declustering approach that employs seismic zonation as an additional parameter in the declustering procedure. The proposed methodology investigates adjustments to the magnitude term within the filtering window criteria based on the correlation of the seismic zone of a designated event with other events. A comparative analysis with the classical Gardner-Knopoff approach is conducted to evaluate the effectiveness of the proposed approach. The findings show that using seismic zonation in declustering improves the depiction of the distribution of seismic activity and reduces the likelihood of filtering out primary independent mainshocks.

This paper introduces a comprehensive cooperative game theory framework to measure the significance of location and neighborhood relations in conjunction with the magnitude of players/parties. The significances of these relations are measured over the EU geography. In this case, there are (i) the test of availability of a core solution that satisfies all associated parties/players; (ii) the measurement of players’/parties’ rational minimal and maximal return expectations from the grand coalition regarding their all individual and sub-group strategies and associated return rationalities; (iii) the determination of the critical players/parties in the grand coalition. The study’s main contributions are the provision of a methodology that identifies spatially/geographically critical players/parties and the design of an algorithm for handling exponentially growing strategies alongside increasing numbers of players/parties. In sum, a comprehensive cooperative game theory framework is introduced to measure the significance of location and neighborhood relations in conjunction with the magnitude of the players/parties. The case of the EU has revealed the union’s geographically critical countries, with Germany being found to be the most influential.

Depremler tüm dünyayı etkileyen ve büyük kayıplara sebep olan doğal felaketlerdir. Doğası gereği önlenemezler, durdurulamazlar, çok kısa bir sürede çok büyük kayıplara ve yıkımlara neden olurlar. Meydana gelen depremler sonrasında yaşanan tecrübe, bilgi birikimi ve gelişen teknoloji ile orantılı olarak inşaat ve deprem mühendisliğinde yaşanan gelişmeler; hazırlık yapma, kayıpları önleme ve etkin afet yönetimi disiplinlerinde büyük ilerlemelere kapı açmıştır. Deprem Ön Hasar Tahmin ya da Deprem Kayıp Tahmin Sistemleri de bu ilerlemelerin en önemli ürünlerinden biridir. Bu sistemler ile özellikle sismik olarak aktif bölgelerde, bölgeyi etkileyebilecek tüm senaryolar dikkate alınarak, olası afetlerin sonuçları hakkında yönetim birimleri ve yetkili merciler bilgilendirilebilmektedirler. Böylelikle afetlere karşı gerekli hazırlıklar yapılabilmekte, önlemler alınabilmekte ve etkin bir afet yönetim planı hazırlanabilmektedir. Ayrıca bu sistemler olası bir deprem afeti ile eş-zamanlı çalışarak depremin hemen ardından etkilenen bölgeleri tahmin edebilmekte, etkili bir afet ve acil durum yönetiminin yapılabilmesini sağlamaktadır. Küresel ölçekte, belirli bir bölge ya da belirli bir ülke için geliştirilmiş birçok deprem kayıp sistemi bulunmaktadır. Bu sistemler kapasitelerine göre çevrimiçi (online) ya da manuel (senaryo modunda) çalışabilmekte, sadece tehlike, sadece risk, tehlike ve risk birlikte tahmin edebilecek şekilde geliştirilmişlerdir. Bu çalışma kapsamında küresel ölçekte kabul görmüş USGS ShakeMap, PAGER, FEMA HAZUS, GEM OPENQUAKE, AFAD-RED ve REDAS yazılımlarının çalışma prensipleri, incelenmiştir. Daha sonra REDAS yazılımı kullanılarak 06 Şubat 2023 (Mw7.7) Pazarcık depremi senaryo deprem olarak çalışılmış, depremin hemen ardından lokasyon ve büyüklük verileri kullanılarak başlatılan tehlike analizleri, sonraki aşamada gerçek deprem verileri ile güncellenmiş ve sonuçlar karşılaştırmalı sunularak, afet yönetimi için çevrimiçi veri almanın önemi sunulmuştur.

Bu çalışma, yer hareketi tahmin modellerinin (GMPE), Deprem Ön Hasar Tahmin Sistemleri'nin (DÖHTS) performansına etkilerini incelemektedir. DÖHTS, afet yönetiminde kritik role sahip sistemlerdir. Çalışmada, 6 Şubat Kahramanmaraş (Pazarcık) depreminin verileri kullanılarak sismik tehlike ve risk analizleri gerçekleştirilmiş, analiz sonuçları kuvvetli yer hareketi istasyonları verileri ile güncellenmiştir. Analizlerde küresel ölçekte kabul gören dört farklı GMPE ve HAZUS sınıflandırma sistemi ve hasargörebilirlik eğrileri tercih edilmiştir. Analiz sonuçları, GMPE’lerin seçimi ve güncellenmiş veri kullanımının sonuçlar üzerindeki etkileri olmak üzere iki farklı açıdan incelenmiştir. GMPE seçimi, sismik tehlike ve risk sonuçlarını doğrudan etkilemektedir. GMPE modellerinin ortalaması alınarak yapılan karşılaştırmada, güncellenmiş veri kullanımı ile acil müdahale gerektiren ağır hasarlı ve göçen bina sayılarının sırasıyla yaklaşık %10 ve %15 daha iyi tahmin edilebildiği belirlenmiştir.

Southeastern Europe faces a significant earthquake threat, endangering lives, property, and infrastructure thus jeopardizing sustainable development. The development of a Rapid Earthquake Damage Assessment System (REDAS) designed to deliver crucial earthquake damage information for scenario planning, real-time response, and bolstering public awareness and preparedness is presented. In doing so, REDAS enhances community resilience and safeguards sustainability. REDAS comprises a Rapid Earthquake Damage Assessment platform (REDA.p), a smartphone application, and an Educational Hub (Edu.Hub). REDA.p provides both scenario-based and near real-time seismic damage evaluation of structures, gas pipelines, and geotechnical failures, based on harmonized Ground Motion Prediction Equations and a comprehensive building taxonomy scheme covering the area under investigation. To assess regional landslide hazards, the Infinite Slope Model and a statistics-based model have been implemented, alongside a statistical model for liquefaction probability assessment. Validated against historical data, REDA.p integrates real-time input from key earthquake monitoring networks in the region, covering cross-border areas as well, while in designated urban zones, the system is enhanced by real-time data from a dense earthquake monitoring network deployed in selected school buildings. The smartphone app and Edu.Hub disseminate critical information, guidelines, and tools to improve public prevention, preparedness, and response capacities, thereby enhancing societal resilience.

Türkiye’nin mevcut yapı stokunun önemli bir kısmı, depreme maruz kalmış ama kullanıma devam edilmekte olan orta katlı betonarme binalardan oluşmaktadır. Hasar görmüş binaların tekrar depreme maruz kaldıklarında gösterecekleri performansın belirlenmesi afet ve risk yönetimi açısından oldukça önemlidir. Bu çalışmada, depremde orta derecede hasar görmüş orta katlı betonarme bir binanın hasar görmeden önceki ve sonraki durumlarının hasar olasılıkları değerlendirilmiştir. Bu kapsamda her iki durum için bina üç boyutlu modellenmiş, doğrusal olmayan statik itme analizine tabi tutulmuş, göreli kat ötelemeleri ile hasargörebilirlik olasılık eğrileri elde edilmiştir. Sonuç olarak her iki durumdaki olası hasar oranları elde edilmiş ve risk yönetimi açısından değerlendirilmiştir.