Dam Water Levels Prediction Using Different Time Series Models

Published in IEEE Xplore: Time series forecasting study achieving 99.32% accuracy for Turkey's largest dams using ARIMA and SARIMAX models.

In this academic study, the water levels of Yusufeli and Deriner Dams, the highest dams in Türkiye, were modeled using advanced time series analysis methods. The study was presented at the 7th International ICHORA Congress and published in the IEEE Xplore library. Methodology and Results: Modeler: Comparative analyses were performed using ARIMA, SARIMA, ARIMAX, and SARIMAX components. High Accuracy: A success score of 99.32% R² was achieved with the ARIMA model for Yusufeli Dam, and 99.05% R² with the SARIMAX model for Deriner Dam. Significance: The study proposes a reliable early warning system model for the sustainability of hydroelectric power generation and flood risk management. Data Analysis: Seasonality and exogenous variables were proven to be preserved as factors affecting the models.