Yapay Sinir Ağları Tabanlı En Büyük Yer Hareketi İvmesi (PGA) Tahmini
Özet
Bu çalışmada, Türkiye'nin sismik açıdan oldukça aktif ve tehlikeli bölgelerinden biri olan Erzurum ve civarında en büyük yer hareketi ivmesini (PGA) tahmin etmek amacıyla Yapay Sinir Ağları (YSA) tabanlı bir yaklaşım geliştirilmiştir. Deprem esnasında mühendislik yapılarının maruz kalacağı sismik yüklerin belirlenmesinde kritik bir parametre olan PGA değerleri; depremin büyüklüğü (M), istasyon ile episantır arasındaki uzaklık (Repi) ve zemin özelliğini temsil eden S dalga hızı (Vs30) değişkenleri kullanılarak modellenmiştir. Araştırma kapsamında AFAD ve Atatürk Üniversitesi Deprem Araştırma Merkezi tarafından işletilen 17 farklı kuvvetli yer hareketi istasyonunda 1994-2022 yılları arasında kaydedilen 791 adet depreme (M ≥ 3.0) ait veri seti analiz edilmiştir. Toplanan verilerin 761 tanesi eğitim, 30 tanesi ise bağımsız test verisi olarak ayrılmış; Çoklu-YSA algoritması ve Levenberg-Marquardt geri yayılım eğitimiyle en uygun YSA mimarisi tespit edilmiştir. Seçilen YSA modelinin regresyon katsayısı eğitim veri setinde R=0.871, bağımsız test setinde ise R=0.926 olarak hesaplanarak yüksek doğruluğa ulaşılmıştır. YSA tahmin sonuçları literatürde yaygın kullanılan Akkar (2014) ampirik denklemi ile karşılaştırıldığında, YSA'nın ölçülen gerçek PGA değerlerine çok daha yakın ve üstün performans gösterdiği gözlemlenmiştir. Ayrıca, M=6.0 büyüklüğündeki olası bir senaryo depremi için Erzurum civarındaki istasyonlarda PGA tahminleri gerçekleştirilmiş ve YSA modellerinin yerel sismik tehlike ve risk değerlendirme çalışmalarına önemli katkılar sunabileceği sonucuna varılmıştır.
In this study, an Artificial Neural Network (ANN)-based approach was developed to estimate peak ground acceleration (PGA) in Erzurum and its surrounding area, which is one of the seismically active and high-hazard regions in Turkey. PGA values, which are a critical parameter in determining the seismic loads that engineering structures will be subjected to during an earthquake, were modeled using earthquake magnitude (M), epicentral distance (Repi), and shear wave velocity representing soil properties (Vs30). Within the scope of the study, a dataset of 791 earthquakes (M ≥ 3.0) recorded between 1994 and 2022 at 17 different strong-ground-motion stations operated by AFAD and Atatürk University Earthquake Research Center was analyzed. Out of the collected data, 761 were separated for training and 30 for independent testing, and the optimal ANN architecture was determined using a Multi-ANN algorithm and Levenberg-Marquardt backpropagation training. The regression coefficient of the selected ANN model was calculated as R=0.871 for the training dataset and R=0.926 for the independent test set, achieving high accuracy. When compared with the empirical equation by Akkar (2014) widely used in the literature, the ANN prediction results were observed to be significantly closer to the actual measured PGA values and showed superior performance. Furthermore, PGA estimates were performed at stations around Erzurum for a potential scenario earthquake of M=6.0, leading to the conclusion that ANN models can make significant contributions to local seismic hazard and risk assessment studies.
Referanslar
Abrahamson N.A., Silva W.J., 1997. Empirical Response Spectral Attenuation Relations for Shallow Crustal Earthquakes, Seismological Research Letters. 68, 94-127
AFAD-TADAS, 2022. Türkiye İvme Veri Tabanı ve Analiz Sistemi, Ankara, Erişim adresi: https://tadas.afad.gov.tr
Akkar S., Bommer J.J., 2010. Empirical Equations for the Prediction of PGA, PGV, and Spectral Accelerations in Europe, the Mediterranean Region, and the Middle East, Seismological Research Letters. 81, 195-206
Akkar S., Sandıkkaya M.A., Bommer J.J., 2014. Empirical Ground-Motion Models for Point- And Extended-Source Crustal Earthquake Scenarios İn Europe and the Middle East, Bulletin of Earthquake Engineering. 12, 359–387
Akkar S., Kale Ö., Yakut A., ve diğ., 2018. Ground-Motion Characterization for the Probabilistic Seismic Hazard Assessment in Turkey. Bullettin of Earthquake Engineering 16, 3439–3463
Ambraseys N.N., 1978. Middle East - A Reappraisal of Seismicity, The Quarterly Journal of Engineering Geology. 11(1), 19–32
Ambraseys N.N., Bommer J.J., 1991. The Attenuation of Ground Accelerations in Europe, Earthquake Engineering and Structural Dynamics 20(12), 1179–1202
Askan A., Karımzadeh S., Asten M., Kılıç N., Şişman F.N., Erkmen C., 2015. Assessment of seismic hazard in the Erzincan (Turkey) region: construction of local velocity models and evaluation of potential ground motions, Turkish Journal of Earth Sciences 24 (6), 529-565
Askan A., Yücemen M.S., 2010. Probabilistic methods for the estimation of potential seismic damage: Application to reinforced concrete buildings in Turkey, Structural Safety. 32(4), 262-271
ATAÜNİ-DAM, 2022. Atatürk Üniversitesi, Deprem Araştırma Merkezi, Erzurum, Erişim adresi: https://birimler.atauni.edu.tr/deprem-arastirma-merkezi
Aydan Ö., Sedaki M., Yarar R., 1996. The Seismic Characteristics of Turkish Earthquakes, Eleventh World Conference on Earthquake Engineering, 23-28 June, Acapulco, Mexico,1025
Beyaz T., 2004. Zemin Etkisinden Arındırılmış Deprem Kayıtlarına Göre Türkiye için Yeni Bir Deprem Enerjisi Azalım Bağıntısının Geliştirilmesi, Doktora Tezi, Ankara Üniversitesi, Fen Bilimleri Enstitüsü, Ankara.
Bindi D., Weatherill G., Kotha S.R., Lanzano G., Luzi L., Cotton F., 2019. The Pan-European Engineering Strong Motion (ESM) Fatfile: Consistency Check via Residual Analysis. Bulletin of Earthquake Engineering. 17, 583–602
Boore D.M., Joyner W.B., Fumal T.E., 1997. Equations for Estimating Horizontal Response Spectra and Peak Acceleration from Western North American Earthquakes: A Summary of Recent Work. Seismological Research Letters 68(1), 128-153
Boore D.M., Atkinson G.M., 2008. Ground-motion Prediction Equations for the Average Horizontal Component of PGA, PGV and 5% Damped PSA at Spectral Periods Between 0.01 s and 10 s. Earthquake Spectra 24(1), 99–138
Campbell K.W., 1985. Strong motion attenuation relations: A ten-year perspective, Earthquake Spectra 1(4), 759-804
Chen Z., Maun J.C., 2000. Artificial neural network approach to single-ended fault locator for transmission lines, IEEE Transactions on Power Systems 370-375
Çeken U., 2007. Marmara Bölgesinin Kuvvetli Yer Hareketi Azalim İlişkisi Modeli, Yüksek Lisans Tezi, Sakarya Üniversitesi, Fen Bilimleri Enstitüsü, Sakarya.
Çoban K.H., Sayıl N., 2019. Evaluation of earthquake recurrences with different distribution models in western Anatolia, Journal of Seismology 23, 1405–1422.
Çoban K.H., Sayıl N., 2020a. Different probabilistic models for earthquake occurrences along the North and East Anatolian fault zones, Arabian Journal of Geosciences 13, 971.
Çoban K.H., Sayıl N., 2020b. Conditional Probabilities of Hellenic Arc Earthquakes Based on Different Distribution Models, Pure Applied Geophysics 177, 5133–5145.
Derakhshani A., Foruzan A.H., 2019. Predicting the Principal Strong Ground Motion Parameters: A Deep Learning Approach, Applied Soft Computing 80,192-201
Dhanya J., Raghukanth S.T.G., 2020. Neural Network-Based Hybrid Ground Motion Prediction Equations for Western Himalayas and North-Eastern India, Acta Geophysica 68, 303–324
Emre Ö., Duman T.Y., Özalp S., Elmacı H., Olgun Ş., Şaroğlu F., 2013. Açıklamalı Türkiye Diri Fay Haritası. Ölçek 1:1.250.000, Maden Tetkik ve Arama Genel Müdürlüğü, Özel Yayın Serisi-30, Ankara-Türkiye. ISBN: 978-605-5310-56-1
Erdik M., Birol Y.A., Onur T., Şeşetyan K., Birgören G., 1999. Assessment of Earthquake Hazard in Turkey and Neighboring Regions, Annals of Geophysics 42, 1125–1138
Esteva L., Villaverde L., 1973. Seismic Risk, Design Spectra and Structural Reliability, Proceedings of Fifth World Conference on Earthquake Engineering 2, 2586-2596
Fichtner A., Saygin E., Taymaz T., Cupillard P., Capdeville Y., Trampert J., 2013. The deep structure of the North Anatolian Fault Zone, Earth and Planetary Science Letters 373, 109-117
Giardini D., Danciu L., Erdik M. ve diğ. 2018. Seismic Hazard Map of the Middle East, Bulletin of Earthquake Engineering 16, 3567–3570
Gülerce Z., Kamai R., Abrahamson N.A., Silva W.J., 2017. Ground Motion Prediction Equations for the Vertical Ground Motion Component Based on the NGA-W2 Database, Earthquake Spectra 33(2), 499-528
Güllü H., Erçelebi E., 2007. A Neural Network Approach for Attenuation Relationships: An Application Using Strong-Ground-Motion Data from Turkey, Engineering Geology 93(3), 65–81
Günaydın K., Günaydın A., 2008. Peak ground acceleration prediction by artificial neural networks for northwestern Turkey, Mathematical Problems in Engineering ID 919420
Idriss I.M., 2008. An NGA Empirical Model for Estimating the Horizontal Spectral Values Generated by Shallow Crustal Earthquakes, Earthquake Spectra 24(1), 217-242
İnan E., Çolakoğlu Z., Koç N., Bayülke N., Çoruh E., 1996. 1976-1996 Yılları Arası İvme Kayıtları Olan Deprem Kataloğu, T.C. Bayındırlık ve İskan Bakanlığı Afet İşleri Genel Müdürlüğü Deprem Araştırma Dairesi Başkanlığı, Ankara, 98 s.
Jemei S., Hissel D., Pera M., Kauffmann J.M., 2008. A new modeling approach of embedded fuel-cell power generators based on artificial neural network, IEEE Transactions on Industrial Electronics. 437-447
Joyner W.B., Boore D.M., 1988. Measurement, Characterization, and Prediction of Strong Ground Motion, Earthquake Engineering and Soil Dynamics II, Proceedings American Society Civil Engineering Geotechnical Engineering Division Specialty Conference, 43-102
Kale O., Akkar D.S., Ansari A., Hamzehloo H., 2015. A Ground-Motion Predictive Model for Iran and Turkey for Horizontal PGA, PGV, and 5% Damped Response Spectrum: Investigation of Possible Regional Effects, Bulletin of the Seismological Society of America 105, 963-980
Kalkan E., Gülkan P., 2004. Empirical attenuation equations for vertical ground motion in Turkey, Earthquake Spectra 20(4), 1111-1138
Koçyiğit A., Canoğlu M.C., 2017. Neotectonics and Seismicity of Erzurum Pull-apart Basin, East Turkey, Russian Geology and Geophysics 58, 99-122
Lou H.L., 1995. Implementing the Viterbi algorithm, IEEE Signal processing magazine. 12, 42-52
McQuarrie N., Stock J.M., Verdel C., Wernicke B.P., 2003. Cenozoic evolution of Neotethys and implications for the causes of plate motions, Geophysical Research Letters 30 (20), doi:10.1029/2003GL017992
Nacar S., Hinis M.A., Kankal M., 2018. Forecasting Daily Streamflow Discharges Using Various Neural Network Models and Training Algorithms, Ksce Journal of Civil Engineering 22(9), 3676-3685
Özbey C., Sarı A., Manuel L., Erdik M., Fahjan Y., 2003. Empirical Strong Ground Motion Attenuation Relations for Northwestern Turkey, Fifth National Conference on Earthquake Engineering, Istanbul, Turkey
PEER, 2022. Pacific Earthquake Engineering Research Center (PEER), University of California, Berkeley, Erişim adresi: http://peer.berkeley.edu
Raghucharan M., Somala S.N., Rodina S., 2019. Seismic Attenuation Model Using Artificial Neural Networks, Soil Dynamics and Earthquake Engineering 126.
Sadigh R.K., Egan J.A., 1998. Updated Relationships for Horizontal Peak Ground Velocity and Peak Ground Displacement for Shallow Crustal Earthquakes. In Proceedings of the Sixth U.S. National Conference on Earthquake Engineering.
Sayıl N., 2014. Evaluation of the seismicity for the Marmara region with statistical approaches, Acta Geodaetica et Geophysica 49, 265–281
Selçuk L., Selçuk A.Z., Beyaz T., 2010. Probabilistic Seismic Hazard Assessment for Lake Van Basin, Turkey, Natural Hazards 54, 949–965
Sianko I., Özdemir Z., Khoshkholghi S., ve diğ., 2020. A Practical Probabilistic Earthquake Hazard Analysis Tool: Case Study Marmara Region, Bulletin of Earthquake Engineering 18, 2523–2555
Simao N.M., Nalbant S.S., Sünbül F., Mutlu A.K., 2016. Central and eastern Anatolian crustal deformation rate and velocity fields derived from GPS and earthquake data, Earth and Planetary Science Letters 433, 89–98
Şengör A.M.C.,1979. The North Anatolian transform fault: its age, offset and tectonic significance, Journal of the Geological Society136, 269-282
Şeşetyan K., Demircioğlu M.B., Duman T.Y., ve diğ. 2018. A Probabilistic Seismic Hazard Assessment for the Turkish Territory—part I: the Area Source Model, Bulletin of Earthquake Engineering 16, 3367–3397
Ulusay R., Tuncay E., Sönmez H., Gökceoğlu, C., 2004. An Attenuation Relationship Based on Turkish Strong Motion Data and Iso-Acceleration Map of Turkey, Engineering Geology 74, 265-291.
Withers K.B., Moschetti, M.P., Thompson, E.M., 2020. A Machine Learning Approach to Developing Ground Motion Models from Simulated Ground Motions, Geophysical Research Letters 47
Wuest T., Weimer D., Irgens C., Thoben K.D., 2016. Machine learning in manufacturing: advantages, challenges, and applications, Production & Manufacturing Research 4(1), 23-45
Yaghmaei-Sabegh S., 2018. Earthquake ground-motion duration estimation by using of general regression neural network, Scientia Iranica 25(5), 2425-2439
Yılmaz B., Aras E., Kankal M., Nacar S., 2019. Prediction of Suspended Sediment Loading by Means of Hybrid Artificial Intelligence Approaches, Acta Geophysica 67, 1693–1705
