Estimation of the volume and water level of Alibey Dam using classical and artificial intelligence methods

Accurate estimation of the water level and volume in the dam reservoir in water structures is important for the use, planning and operation of water resources. As a result of increasing climate changes, globalization, rapid population growth and urbanization, the value of water in dams is gaining importance day by day. In this study, multiple linear regression (MLR), interaction multivariate regression (IMR), quadratic multivariate regression (QMR), pure-quadratic multivariate regression (PMDR), artificial neural networks (ANN), adaptive neural fuzzy inference system (ANFIS), fuzzy simple membership functions and fuzzy rule generation technique (Fuzzy SMRGT) methods are used. Alibey Dam located in Sultangazi district of Istanbul was chosen as the study area. In the first stage of this study, 381 monthly measurement data from DSI for the years 1989-2020 were used for monthly water level estimation. For monthly water level (KA) estimation; monthly water amount discharged from the dam (XA), monthly evaporation amount (YA), monthly drinking water amount (ZA), monthly water amount coming into the lake (TA) and monthly water level in the previous month (QA) parameters one month ago were used. 75% of the measured data is divided into training and 25% for testing. Correlation coefficient (R), root mean square error (RMSE) and mean absolute error (MAE) were used as statistical performance evaluation criteria. When the measurement and model estimation results are compared, it has been observed that artificial intelligence methods give better results than classical methods. In the second stage of this study, 2190 daily measurement data from IBB for the years 2015-2020 were used for the daily dam reservoir volume estimation.For the daily dam reservoir volume (GK) estimation; daily evaporation amount (GX), daily precipitation amount (GY), daily used water amount (GZ), daily water amount transferred from Terkos dam to Alibey dam (GT) and daily dam reservoir volüme in the previous day (GQ) the day before were used. Correlation coefficient, root mean square error and mean absolute error were used as statistical performance evaluation criteria. When the measurement and model estimation results are compared, it has been observed that artificial intelligence methods give better results than classical methods.