| 摘要: |
| 【目的】合理控制龙羊峡水库年末（水文年年末，6月末）运行水位，为安全高效利用龙羊峡水库和保障黄河水资源综合利用效益提供参考。【方法】通过逐步回归分析方法，率定龙羊峡水库年末水位关键影响因子；建立BP神经网络预测模型，通过反推法确定龙羊峡水库年末消落水位下限；引入水库临界弃水概率的概念及计算公式，通过控制汛期临界弃水概率确定龙羊峡水库年末消落水位上限。【结果】龙羊峡水库年末消落水位关键影响因子为当年入库水量、年初水位、次年入库水量和当年黄河干流用水量；当年年末水位不低于死水位且汛期弃水概率在10%以内时，龙羊峡水库合理的年末水位应控制在2 575.0～2 576.5 m。【结论】通过BP神经网络模型对水库年末消落水位的预测控制，并结合水库汛期临界弃水概率指导水库汛期防洪，可在防洪安全的基础上有效地提高水库的发电效益。 |
| 关键词: 龙羊峡水库 BP神经网络 年末水位预测 临界弃水概率 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金项目(51179148，51179149)；水利部公益性行业基金项目(201101043)；陕西省重点实验室项目(11JS077)；教育部新世纪优秀人才支持计划项目(NCET-10-0933) |
|
| Year-end water level control and water discharge in flood season for Longyangxia reservoir |
|
XIE Yang-yang,WANG Yi-min,HUANG Qiang
|
| Abstract: |
| 【Objective】This study was to provide reference for utilizing Longyangxia reservoir safely and efficiently,and guaranteeing the comprehensive benefits of water resources at Yellow river through controlling the year-end water level (water level at the end of June) of Longyangxia reservoir.【Method】The key factors affecting water level of Longyangxia reservoir were gained through stepwise regression analysis method.BP neural network prediction model was built,and the lower limit of Longyangxia reservoir’s year-end water level was found out through back-stepping method.The concept of reservoir’s critical abandoned water probability in flood season was defined,and its formula was also presented.The upper limit of Longyangxia reservoir’s year-end water level was obtained through controlling the critical abandoned water probability in flood season.【Result】The key factors of Longyangxia reservoir’s year-end water level are the incoming water at the first year,the water level at the beginning of the year,the incoming water at the second year and the water consumption in mainstream of Yellow river.In order to ensure the water level at the end of the first year not bellow the dead water level and make the critical abandoned water probability in flood season below 10%,Longyangxia reservoir’s year-end water level should be controlled at 2 575.0－2 576.5 m.【Conclusion】The year-end water level of multi-year regulating storage reservoir,which is predicted and controlled by BP neural network model and the concept of reservoir’s critical abandoned water probability,will help to direct flood control during flood season and improve the dynamoelectric benefit of reservoir efficiently on the basis of flood safety. |
| Key words: Longyangxia reservoir back propagation network prediction of year-end water level critical water discharge probability |
