首页 | 本学科首页   官方微博 | 高级检索  
     

基于最大熵原理的大坝安全监控组合预测模型研究
引用本文:葛勇坚,伍元,李子阳.基于最大熵原理的大坝安全监控组合预测模型研究[J].黑龙江水专学报,2010,37(1):13-16,24.
作者姓名:葛勇坚  伍元  李子阳
作者单位:1. 义乌市后宅街道农业服务中心,浙江义乌,322008
2. 河海大学水资源高效利用与工程安全国家工程研究中心,南京210098;广州建设工程交易中心,广州510630
3. 南京水利科学研究院,南京,210029
基金项目:国家自然科学基金(50809025)"混凝土坝空间场的安全预警模型研究",江苏省"333高层次人才培养工程科研项目"(2017-B08037)"高拱坝真实工作形态反馈分析研究" 
摘    要:大坝性态受到内外多种不确定因素的影响,使得监测效应量在一定程度上呈现出随机和不确定性的特点,单一的安全监控预测模型不可避免地存在某些缺陷。采用最大熵原理建立大坝安全监控的组合预测模型,该模型以各单一模型的预测结果作为约束信息,利用最大熵原理得到预测结果的分布。实例分析表明该方法能够突出监测量变化的随机性特点,有效地提高了大坝安全监控模型的预测精度。

关 键 词:大坝安全监控  组合预测  最大熵原理

Study of Combination Forecast Model for Dam Safety Monitoring Based on Maximum Entropy Principle
GE Yong-jian,WU Yuan,LI Zi-yang.Study of Combination Forecast Model for Dam Safety Monitoring Based on Maximum Entropy Principle[J].Journal of Heilongjiang Hydraulic Engineering College,2010,37(1):13-16,24.
Authors:GE Yong-jian  WU Yuan  LI Zi-yang
Affiliation:1. Agricultural Service Center of Houzhai Street, Yiwu 322008, Zhejiang, China; 2. National Engineering Research Center of Water Re- sources Efficient Utilization and Engineering Safety, Hohai University,Nanjing 210098, China; 3. Guangzhou Construction Engineering Trade Center, Guangzhou 510630, China; 4. Nanjing Hydraulic Research Institute, Nanjing 210029, China)
Abstract:For dam behavior is affected by many uncertain factors, the monitor data shows the characteristics of random and uncertainty. Therefore, single safety monitoring forecasting models will inevitably has some shortcomings. In this paper, the maximum entropy principle is used to establish a combination fore- casting model of darn safety monitoring. Taking prediction results of each single model as binding informa- tion, the predicted distribution is gotten by using the maximum entropy principle. Case analysis shows that this method can highlight the random characteristics of monitoring data and effectively improve the prediction accuracv of dam safety monitoring model.
Keywords:dam safety monitoring  combination forecast  maximum entropy principle
本文献已被 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号