00000ctm a22000004i 4500 UP-99796217613975856 Buklod 20220401092257.0 m o u ta 220401s2010 xx d r |||| u| (iLib)UPD-00536372720 DENG eng DMLUC LG 995 2010 E63 R63 Rodero, Augusto O. Hourly streamflow forecasting with ARMA models for maximizing hydropower revenues Augusto O. Rodero. 2010 xvi, 308 leaves ill. (some col.) "October 2010" Thesis (M.S. Civil Engineering)--University of the Philippines Diliman Hydrologic streamflow forecasting has always been a concern for water resources engineers for optimal operations of water systems. The autoregressive - moving average (ARMA) models is a group of stochastic models popularly used by modelers over the past decades as a tool for streamflow forecasting. The necessity for a good hourly streamflow forecasting model has never been more valuable to operators of hydropower plants. With the enactment of the Electric Power Industry Reform Act, 2001 (EPIRA), the operators are required to submit a day-ahead forecast generation to the National Grid Corporation of the Philippines (NGCP) for its next 24-hour generation. Changes for imbalance (over/under) generation are now in effect. This paper has illustrated that with the development of an hourly ARMA forecasting model for the operation of the 16.5MW Sibulan-A hydropower plant (HPP); charges for imbalance generation can be minimized and eventually led to maximized hydropower revenue. A step-by-step procedure of model construction was presented on the hourly data of the mentioned powerplant. It was highlighted that the final models were chosen was based on how sound the models are in 24-hours ahead forecasting applications. It was demonstrated that operating Sibulan-A with ARMA forecasting yielded more revenues at an average of Php 14,000.000. Compared to operating the HPP by simply assuming that past 7-day hourly mean is the forecast for the next 24-hours. Hydrological forecasting. Hydrologic models. ARMA (Autoregressive Moving Average). FI UP UPD DENG-II LG 995 2010 E63 R63 Thesis