M.Sc. Tezi Görüntüleme

Student: Banu YILMAZ
Supervisor: Assoc. Prof. Dr. Egemen ARAS
Department: İnşaat Mühendisliği
Institution: Graduate School of Natural and Applied Sciences
University: Karadeniz Technical University Turkey
Title of the Thesis: MODELING OF SUSPENDED SEDIMENT LOAD CARRIED IN ÇORUH RIVER BASIN BY USING DIFFERENT ARTIFICIAL INTELLIGENCE METHODS
Level: M.Sc.
Acceptance Date: 2/6/2016
Number of Pages: 128
Registration Number: i3038
Summary:

      Accurate prediction of the suspended sediment load in rivers is very important for water resources and management. Although direct measurement is most reliable method for sediment, it is very expensive and time-consuming. In additon, sediment transport equations are requires many various parametres about flow and sediment characteristics. Sediment rating curves which is widely used show deficiencies in several points. For this reason, we need other methods which more reliable result. In this study, many methods were developed about suspended sediment estimation, at three stations Altınsu, İnanlı and Karşıköy, on Çoruh River, in Çoruh Basin. In addition to the sediment rating curve, different regression and artificial neural networks were used and comparative analyzes were conducted. For each station 7 methods were applied. They were regression analysis, multivariate adaptive regression splines, artifical bee colony, teaching-learning-based optimization algorithm, multilayer artifical neural network, artificial neural network training using artificial bee colony and artificial neural network training using teaching-learning based optimization algorithm. The results was evaluated according to the criteria of root mean square error, mean absolute error and coefficient of determination. İn addition to the methods which generated with artificial neural networks, multivariate adaptive regression splines model has achieved successful results about estimation of suspended sediment load.

      Key Words: Estimation of suspended sediment load, Artificial neural networks, Artificial bee colony, Teaching-learning based optimization algorithm, Multivariate adaptive regression splines.