Prediction of Total Amount of River Flow Rate on Upper District of Dam for Hydro Power Plant by Using Radar Echo Data

摘要

This paper describes an application of neural network for prediction of total amount of river flow rate from the meteorological radar echo data. A neural network system is developed through a case study on a dam for hydropower plant located the upper district of the Yahagi River in Central Japan. We report the two different type prediction systems. First, the 2step-type prediction system is reported that is constituted of estimation system of rainfall distribution as first step and prediction system of total amount of river flow rate as second step. Second, the 1-step type prediction system is reported that has 6 input nodes corresponding to the rainfall amounts from radar echo data, four ground rainfall gauges and the base flow rate. The output from the system is the predicted total amount of river flow. It is found from our investigations that predictions of total amount of river flow rate are improved by utilization the radar echo data.