Requirements and benefits of flow forecasting for improving hydropower generation

Xiaohua Dong, J.K. Vrijling (Editor), Catarine M. Dohmen-Janssen, E. Ruigh (Editor), Martijn J. Booij, B. Stalenberg (Editor), Suzanne J.M.H. Hulscher, P.H.A.J.M. van Gelder (Editor), M. Verlaan (Editor), A. Zijderveld (Editor), P. Waarts (Editor)

Research output: Contribution to conferencePaperAcademic

23 Downloads (Pure)


This paper presents a methodology to identify the required lead time and accuracy of flow forecasting for improving hydropower generation of a reservoir, by simulating the benefits (in terms of electricity generated) obtained from the forecasting with varying lead times and accuracies. The benefit-lead time relationship was investigated only for perfect inflow forecasts, with a few selected forecasting lead times: 4, 10 days and 1 year. The water level and the release from the reservoir were then optimized. Based on the optimization results, the “threshold” lead time was identified, beyond which, further extension of the forecasting lead time will not benefit significantly. In order to investigate the benefit-accuracy relationship, the forecasting lead time was fixed to be 4 days, and the stochastic nature of the inflow was considered by means of generating noised synthesized inflow series for optimization. Noised inflow series were generated to mimic the flow forecasting with different levels of accuracy. These synthesized flow forecasting series served as input into the optimization model to simulate the benefits. The optimization model consists of two discretized deterministic dynamic programming (DDDP) models, one for long-term (monthly) and one for the short-term (daily) optimization. They were coupled together so that both short-term benefits (in a time horizon of flow forecasting lead time) and long-term benefits (in a time horizon of one year) were considered and balanced. The Qingjiang river in China and a reservoir on its main channel were taken as case study. The results revealed that the “threshold” lead time is about 30 days. A perfect inflow forecasting with 4 days lead time will realize 86% of the theoretical maximum electricity generated in one year. For inflow forecasting with a fixed lead time of 4 days and different forecasting accuracies, the benefits can increase by 3 to 11% (which is quite substantial) compared to the actual operation benefits. It is concluded that the definition of the appropriate lead time will depend mainly on the physical conditions of the basin and on the characteristics of the reservoir. The derived threshold lead time (about 30 days) is not feasible with the present flow forecasting techniques, but gives a theoretical upper limit for the extension of forecasting lead time. Criteria for the appropriate forecasting accuracy for a specific feasible lead-time should be defined from the benefit-accuracy relationship, starting from setting a preferred benefit level, in terms of percentage of the theoretical maximum. Inflow forecasting with a higher accuracy does not always increase the benefits, because these also depend on the operation strategies of the reservoir.
Original languageUndefined
Publication statusPublished - 23 May 2005


  • IR-49945
  • METIS-222961

Cite this