Analysis of reservoir outflow using deep learning model

Predicting reservoir outflow is crucial for managing water resources in under extreme flood and drought conditions. Time series study of reservoir outflow relies heavily on previous information on climate and reservoir factors. The Long Short Term Memory model of Deep Learning is applied using rainfall, rainfall intensity, runoff rate, temperature, surface water area, and reservoir outflow to predict reservoir outflow. This study summarizes the parameter setting effect on model performance and analyzes the main factors that affect reservoir outflow prediction. Monthly rainfall, rainfall intensity, runoff rate, temperature, outflow, and surface water area data are used in the multipurpose reservoir prediction model to analyze monthly and yearly water outflow of the reservoir. This system help in water management to reduce the risk of flooding downstream while ensuring sufficient water storage for monthly utilization, i.e., an outflow of a reservoir to the city. This method determines the appropriate amount of water released from the reservoir during the dry season and helps set a relationship with other input variables and outflow. The model has been trained and tested using the obtained data. The result analyzes that combined iterations and neurons of a hidden layer mainly impact manipulating the model precision; computation speed is primarily affected by the batch size of the model. The proposed model can simultaneously predict entire parameters in an accurate and efficient way.