中文摘要：

考虑水电站参与调峰对时段平均出力的限制要求,结合实际电站的典型日出力过程与峰荷分布规律,以水电站的最大出力为随机变量,基于该随机变量服从指数分布的假设,以总调峰效益期望最大为目标,建立了长江三峡-葛洲坝梯级水电站长期优化调度模型,详细分析了三峡-葛洲坝电站单峰调峰、双峰调峰和逐月调峰3种调峰运行方式,计算了相应的分布参数和发电量.计算结果表明：该模型能在长期优化调度中更好地体现电力系统对水电的调峰需求,3种调峰方式运行下三峡-葛洲坝梯级多年平均电量分别是1 080、1 055、1 079亿kW·h,梯级调峰损失电量分别为9.9、34.6、9.5亿kW·h.

英文摘要：

To meet peak shaving demand under the large-scale power grid platform in the long-term operation,hydropower stations＇ average output needs to be controlled.This paper takes the maximum output of hydropower as a random variable,and based on the hypothesis that the variable follows exponential distributions,a long-term optimization model of the maximum peak shaving expectation for the Three Gorges cascade hydropower stations is established.This paper analyzes three peak-shaving modes for hydropower stations：i.e.single-peak curve,double-peak curve and setting the monthly limit value of output.And then,the corresponding distribution parameters are calculated.A case study of the Three Gorges cascade hydropower stations shows that the model of the maximum peak shaving expectation is effective in improving the peak regulation capacity.The hydropower generations of cascade reservoirs in three peak-shaving modes are 108,105.5and 107.9TW·h;while the electricity loss caused by peak-shaving are 990,3460 and 950GW·h respectively.