水库实时优化调度余留库容的云决策方法研究

Cloud decision-making method for the remaining capacity of reservoir in real-time optimal operation

针对水库(群)水资源系统实时优化调度中水库余留库容的决策问题,讨论了来水预报信息与水库余留库容的模糊性和随机性,并应用人工智能领域中的云模型给予综合描述;通过挖掘历史沉积的知识(不确定推理规则),构建并给出了基于预报期和余留期预报来水量的水库余留库容云模型决策模型及云模型决策方法。将该决策模型和方法应用于龙羊峡水库2000年水库余留库容的决策,结果表明,应以77.3亿m3的余留库容来确定2000年该水库实时优化调度的经济蓄水线,将多产生1.48亿m3的发电效益,优于实际调度的计算结果。

Aiming at the reservoir remaining capacity decision-making in real-time optimal operation of water resources system ,inflow forecasting information and the fuzzy and random characteristics of remaining capacity are discussed and described fully by using cloud model which is used in artifical intelligence field. By digging into history knowledge （uncertainty reasoning rules）,cloud decision-making model for remaining capacity is established, and cloud decision-making method is given according to inflow which is forecasted based on forecasting period and remaining period. The model and method are applied to the remaining capacity decision-making of Longyangxia reservoir in 2000,which indicates that 77.3）〈 108 m^3 remaining capacity of reservoir should be used to confirm economical storage line of real-time optimal operation in 2000,this will produce 1.48×10^8 m^3 more power benefits than that in real operation.