Abstract: Stochastic modeling of daily runoff is constructed to analyze the power generation risks of hydropower stations caused by the uncertainty of inflow. Simulated long-term runoff series are used as input data of the model to compare regular scheduling and optimized scheduling. A risk analysis system consisting annual average power generation, power generation stability, water abandonment, power generation guarantee rate, and full storage rate as major risk indexes is established. The Three Gorges Reservoir is taken as an example to compare the risk level of power generation between regular scheduling and optimized scheduling. Results show that the annual mean power generation in optimized scheduling increases by about 5% compared with that of regular scheduling. The calculated entropy values imply that the uncertainty of optimized scheduling model is much smaller and more stable. The abandoned water amount in optimized scheduling is about 50% of that in regular scheduling. In addition, the risk is reduced in optimized scheduling model. The scheduling process derived from the optimized scheduling model in this paper has better performance in economic benefits and risk control.

Key words: runoff stochastic simulation, regular scheduling, optimization scheduling, hydropower generation risk, Three Gorges Reservoir