A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Value of Updating Ensemble Streamflow Prediction in Reservoir Operations
This study proposes a methodology that can update a monthly stochastic optimization model in the middle of a month as a new forecast is available and examines the value of this updating in monthly reservoir operations during a 9-month drawdown period. The proposed optimization model is called SSDP/ESP that uses Sampling Stochastic Dynamic Programming with Ensemble Streamflow Prediction. A monthly ESP forecast is updated in the middle of a month using (1) the observed meteorological (single) series from the first day of the month to the forecast day and (2) the historical meteorological (multiple) scenarios from the forecast day to the end of the month. SSDP/ESP runs and updates their policy whenever the ESP forecast is updated. Applied to the Geum river basin in Korea where two reservoirs are located in series and operated primarily for water supply, the proposed model is compared with other optimization models through 2100 cross validation simulation runs. The simulation results show that the proposed model reduces the annual water shortage by 1 million m3 ( = 4.8% of the total shortage) on average and approaches 97% of the ideal model that assumes perfect foresights if the model is updated three times a month. This study also shows the water shortage reduction between a well-forecasted and a poorly-forecasted years can be up to 20% of the total water shortage in the poorly-forecasted year even if a similar amount of the annual inflow is available in the two years.
Value of Updating Ensemble Streamflow Prediction in Reservoir Operations
This study proposes a methodology that can update a monthly stochastic optimization model in the middle of a month as a new forecast is available and examines the value of this updating in monthly reservoir operations during a 9-month drawdown period. The proposed optimization model is called SSDP/ESP that uses Sampling Stochastic Dynamic Programming with Ensemble Streamflow Prediction. A monthly ESP forecast is updated in the middle of a month using (1) the observed meteorological (single) series from the first day of the month to the forecast day and (2) the historical meteorological (multiple) scenarios from the forecast day to the end of the month. SSDP/ESP runs and updates their policy whenever the ESP forecast is updated. Applied to the Geum river basin in Korea where two reservoirs are located in series and operated primarily for water supply, the proposed model is compared with other optimization models through 2100 cross validation simulation runs. The simulation results show that the proposed model reduces the annual water shortage by 1 million m3 ( = 4.8% of the total shortage) on average and approaches 97% of the ideal model that assumes perfect foresights if the model is updated three times a month. This study also shows the water shortage reduction between a well-forecasted and a poorly-forecasted years can be up to 20% of the total water shortage in the poorly-forecasted year even if a similar amount of the annual inflow is available in the two years.
Value of Updating Ensemble Streamflow Prediction in Reservoir Operations
Kim, Y. O. (author) / Eum, H. I. (author) / Ko, I. H. (author)
Operations Management Conference 2006 ; 2006 ; Sacramento, California, United States
2006-08-03
Conference paper
Electronic Resource
English
Value of Updating Ensemble Streamflow Prediction in Reservoir Operations
| British Library Conference Proceedings | 2006
Decision Support System for Optimizing Reservoir Operations Using Ensemble Streamflow Predictions
| British Library Online Contents | 2011
Decision Support System for Optimizing Reservoir Operations Using Ensemble Streamflow Predictions
| Online Contents | 2011
| British Library Online Contents | 2016