Estimation of Keban dam lake level using fuzzy logic and support vector machines method

Under today's conditions, while the need for water increases with the increase in the world population, unfortunately the water resources have been also decreasing. Therefore, it is very important to protect and direct water basins and especially dam lakes. Managing the presence of water in dam lakes is directly related to the estimation of this amount in the future rather than knowing the amount of water. Proper management of the most important water resource of the region such as Keban dam is of great importance. In this study, by using the variables (The General Directorate of State Hydraulic Works DSI in Turkish acronym) of daily water level, energy purpose, daily total water consumption and total evaporation amount from Keban Dam Lake in Keban district of Elazig between 2013 and 2014, Daily reservoir level estimation is aimed with Fuzzy Logic, Support Vector machines and Multiple Linear Regression Methods. The models were compared with each other according to the three statistical criteria, namely, mean absolute error (MAE), root mean square error (RMSE), and determination coefficient (R2). As a result of the study, it was observed that Fuzzy Logic (BM) and Support Vector Machines (SVM) models were successful in predicting the reservoir level of the dam. Key Words : F u zzy Logic, Support Vector Machines, Regression, Dam Reservoir Level, Keban dam