Estimation of design parameters of a retaining wall with support vector regression approach

Yıl 2024, , 1759 - 1770, 20.05.2024

Ümit Çalık

https://doi.org/10.17341/gazimmfd.994823

Öz

The location of critical sliding surface (αcr), magnitude (Pae) and application point (zae) of seismic active earth thrust are influenced by many parameters related to the soil properties of backfill, the loading conditions and the cross-section geometry of problem for a retaining wall. In recent years, the application of powerful learning algorithms such as Support Vector (Machine) Regression (SVR), to reveal the regression relationships within a large number of input and output variables for engineering problems such as this, has been presented on the estimation of the design factors that should be known without following the complex and complicated calculation steps. In this study, the SVR approach considering 6 different kernel functions with 2 data sampling techniques is applied to find out the best regression relationships between 11 inputs and 3 outputs from a parametrically generated big data set containing 119393 data. Cubic kernel function for Pae and medium gauss kernel function for αcr and zae achieve the best predictive SVR models. The deviations between the predicted and actual values are estimated within a range of ± 20 kN/m2, ± 8º and ± 0.15 m, respectively. Besides, the two types of sampling methods have almost no statistical and practical effects on the performance of the models. The verification of the SVR models is also carried out by statistically comparing the results of another research in the literature using a totally new data set containing 4374 data. Herein, the most successful predictive performance of the SVR models presents for Pae and αcr. However, it is observed that the SVR predictions for zae are relatively weaker but still statistically at an acceptable level.

Anahtar Kelimeler

Support Vector (Machine) Regression, Design parameters of retaining walls, Seismic loads, Cohesive backfill soil, Statistical indicators

Destek vektör regresyonu yaklaşımı ile istinat duvarı tasarım parametrelerinin belirlenmesi

Öz

Bir istinat duvarında kritik kayma yüzeyinin yeri (αcr), sismik aktif toprak itkisinin büyüklüğü (Pae) ve etki mesafesi (zae); arka dolgu zeminin özellikleri, yükleme durumu ve problem geometrisi ile alakalı birçok parametreden etkilenmektedir. Son yıllarda, bunun gibi çok sayıda girdi ve çıktı değişkenine sahip mühendislik problemlerinde, Destek Vektör (Makine) Regresyonu (DVR) gibi güçlü öğrenme algoritmalarının parametreler arasındaki regresyon ilişkilerini ortaya çıkarılmasında kullanılmasıyla kompleks hesap adımlarını izlemeden bilinmesi gereken tasarım parametrelerinin tahmini üzerinde durulmaktadır. Bu çalışmada DVR ile, parametrik olarak üretilen 119393 veri içeren büyük hacimli bir data kümesinden 11 adet girdi parametresi ve 3 adet çıktı parametresi arasındaki en iyi regresyon ilişkileri, 2 farklı örnekleme tekniği, 6 değişik çekirdek fonksiyonunun kullanılmasıyla ile ortaya çıkarılmıştır. Pae için kübik fonksiyon, αcr ve zae için medium gauss fonksiyon en iyi DVR modellerini oluşturmuşlardır. Model tahminlerinin gerçek değerden sapmaları sırasıyla ±20 kN/m2, ±8º ve ±0,15 m gibi bir değişim bandı içinde yer almıştır. Bununla birlikte örnekleme yöntemlerinin modellerin öngörü performansı üzerinde etkisi hemen hemen hiç olmamıştır. DVR modellerin doğrulanması, 4374 adet veri barındıran yeni bir data seti ile literatürdeki başka bir çalışmanın sonuçları ile istatistiksel olarak karşılaştırılması suretiyle gerçekleştirilmiştir. Burada, DVR modeller en başarılı tahmin performansını Pae ve αcr için sergilemişlerdir. Bununla birlikte zae için yapılan DVR öngörülerinin göreceli olarak bir parça zayıf kaldığı gözlense de halen istatistiksel olarak kabul edilebilir seviyededir.

Anahtar Kelimeler

Destek vektör (makine) regresyonu, İstinat duvarlarının tasarım parametreleri, Kohezyonlu arka dolgu zemini, İstatistiksel göstergeler, Sismik yükler, , Sismik yükler

Kaynakça

• Calik U. (ön baskı), Critical inclination of failure surface and seismic active earth thrust for a broken slope backfill, Teknik Dergi, 33(4), 2022.
• Vapnik V.N., Golowich S.E., Smola A., Support vector method for function approximation, regression estimation, and signal processing, 9th International Conference on Advances in Neural Information Processing Systems, San Mateo-CA, 281-287, 2-5 December, 1996.
• Osowski S., Siwek. K., Markiewicz T., MLP and SVM networks – a comparative study, 6th Nordic Signal Processing Symposium (NORSIG), Espoo-Finland, 37-40, 9-11 June, 2004.
• Samui P., Sitharam T.G., Kurup P.U., OCR prediction using support vector machine based on piezocone data, J. Geotech. Geoenviron. Eng., 134(6), 894-898, 2008.
• Puri N., Prasad H.D., Jain A., Prediction of geotechnical parameters using machine learning techniques, Procedia Computer Science, 125, 509–517, 2018.
• Aboutaleb S., Behnia M., Bagherpour R., Bluekian B., Using non-destructive tests for estimating uniaxial compressive strength and static Young’s modulus of carbonate rocks via some modeling techniques, Bull. Eng. Geol. Environ., 77, 1717-1728, 2018.
• Cruz M., Santos J.M., Cruz N., Using neural networks and support vector regression to relate marchetti dilatometer test parameters and maximum shear modulus, Appl. Intell., 42, 135-146, 2015.
• Kurnaz T.F., Kaya Y., The comparison of the performance of ELM, BRNN, and SVM methods for the prediction of compression index of clays, Arabian Journal of Geosciences, 11(24), 770-784, 2018.
• Samui P., Sitharam T.G., Site characterization model using least-square support vector machine and relevance vector machine based on corrected SPT data (Nc), Int. J. Numer. Anal. Meth. Geomech., 34, 755-770, 2010.
• Günaydın O., Özbeyaz A., Söylemez M., Regression analysis of soil compaction parameters using support vector method, Celal Bayar University Journal of Science, 14(4), 443-447, 2018.
• Debnath P., Dey A.K., Prediction of bearing capacity of geogrid-reinforced stone columns using support vector regression, Int. J. Geomech., 18(2), 1-15, 2018.
• Das M., Dey A.K., Prediction of bearing capacity of stone columns placed in soft clay using SVR model, Arabian Journal for Science and Engineering, 44, 4681-4691, 2019.
• Pal M., Deswal S., Modelling pile capacity using Gaussian process regression, Computers and Geotechnics, 37, 942-947, 2010.
• Kardani N., Zhou A., Nazem M., Shen S., Estimation of bearing capacity of piles in cohesionless soil using optimised machine learning approaches, Geotech. Geol. Eng., 38(9), 2271-2291, 2019.
• Moayedi H., Hayati S., Artificial intelligence design charts for predicting friction capacity of driven pile in clay, Neural Computing and Applications, 31, 7429-7445, 2019.
• Singh T.V., Pal M., Arora V.K., Modeling of oblique load test on batter pile group based on Support Vector Machines and Gaussian Regression, Geotech. Geol. Eng., 36, 1597-1607, 2018.
• Pal M., Support vector machines-based modeling of seismic liquefaction potential, Int. J. Numer. Analyt. Meth. Geomechanics, 30(10), 983-996, 2006.
• Goh A.T.C., Goh S.H., Support vector machines: Their use in geotechnical engineering as illustrated using seismic liquefaction data, Comput. Geotech., 34, 410-421, 2007.
• Lee C., Chern S., Application of a support vector machine for liquefaction assessment, Journal of Marine Science and Technology, 21(3), 318-324, 2013.
• Xue X., Xiao M., Application of genetic algorithm-based support vector machines for prediction of soil liquefaction, Environ. Earth Sci., 75(874), 1-11, 2016.
• Samui P., Support vector machine applied to settlement of shallow foundations on cohesionless soils, Comput. Geotech., 35(3), 419-427, 2008.
• Ocak I., Seker S.E., Calculation of surface settlements caused by EPBM tunneling using artificial neural network, SVM, and Gaussian processes, Environ. Earth Sci., 70, 1263-1276, 2013.
• Oommen T., Baise L.G., Model development and validation for intelligent data collection for lateral spread displacements, Journal of Computing in Civil Engineering, 24(6), 467-477, 2010.
• Samui P., Slope stability analysis: A support vector machine approach, Environmental Geology, 56(2), 255-267, 2008.
• Peng J., Zhu, Y., Derivation of Shukla’s generalized expression for dynamic active thrust by inclined slice element method, Soil Mechanics and Foundation Engineering, 562, 77-81, 2019.
• Nian T., Han J., Analytical solution for Rankine’s seismic active earth pressure in c-ϕ soil with infinite slope, Journal of Geotechnical and Geoenvironmental Engineering, 1399, 1611-1616, 2013.
• Seed H.B., Whitman R.V., Design of earth retaining structures for dynamic loads, ASCE Specialty Conference on Lateral Stresses in the Ground and Design of Earth Retaining Structures, Ithaca-New York, 103-147, 22-24 June, 1970.
• Terzaghi K., Theoretical Soil Mechanics, John Wiley, New York, 1943.
• Das B.M., Sobhan K., Principles of Geotechnical Engineering 9th Edition, Cengage Learning, Boston, 2018.
• Yang J., Lung W.H., Seismic design of retaining walls considering vertical ground acceleration, 4th International Conference on Earthquake Geotechnical Engineering, Thessaloniki-Greece, 25-28 June, 2007.
• Vapnik V.N., The Nature of Statistical Learning Theory, Springer-Verlag, New York, 1995.
• Scholkopf B., Smola A.J., Learning with Kernels: Support Vector Machines, Regularization, Optimization and Beyond, MIT Press, A.B.D., 2002.
• Karal Ö., EKG verilerinin destek vektör regresyon yöntemiyle sıkıştırılması, Journal of the Faculty of Engineering and Architecture of Gazi University, 33(2), 743-755, 2018.
• Yabanova İ., Yumurtacı M., Destek vektör makineleri kullanarak dinamik yumurta ağırlıklarının sınıflandırılması, Journal of the Faculty of Engineering and Architecture of Gazi University, 33(2), 393-402, 2018.
• Cristianini N., Shawe-Taylor J., An Introduction to Support Vector Machine, Cambridge Univ. Press, London, 2000.
• Vapnik V.N., An overview of statistical learning theory, IEEE Transactions on Neural Networks, 10(5), 988-999, 1999.
• Hair J.F., Sarstedt M., Hopkins L., Kuppelwieser V.G., Partial least squares structural equation modeling (PLS-SEM): An emerging tool in business research, European Business Review, 26(2), 106-121, 2014.