摘要
为了解决平原水库安全评价忽略生态和可持续性的不足,针对平原水库面积大、水深小、坝线长和地基处理差的特点,在健康功能分区的基础上,综合考虑安全、生态特征、可持续性和社会政治环境等因素构建平原水库健康警源指标体系,指标体系包括围坝安全、水库社会效益、水库生态效应和状态持续性4类一级指标和15个二级指标.利用积累的海量数据资料和信息反馈形成的数据库,基于决策树C5.0算法建立大数据环境下平原水库健康预警模型,并以丁东水库健康预警为例进行实证研究.研究结果表明,平原水库健康预警模型精度为91.92%;实证规则集揭示影响水库健康警度分类的因素主要有测压管水位指标、水质安全指标、地下水位动态指标和泥沙变化指标;水库健康状况为健康,为无警级别;平原水库健康预警模型充分利用大数据资源,可操作性强,可在其他水库健康管理中广泛应用;将大数据思维引入水利管理,用以解决水利面临的数据庞杂、相关利益群体需求多样化和区域统筹困难等难题,对水库库区生态系统可持续发展与科学管理具有重要意义.
The ecosystem and sustainability were generally ignored in safety evaluation for plain reservoir. With the consideration of its features such as large area, shallow water depth, long dam length and poor foundation treatment, the health warning indicator system for plain reservoir was established based on the health function regionalization combining with its safety, ecosystem characteristics, persistence, and other social, political and environmental factors. This index system included 4 first-grade indexes, namely reservoir safety, social benefits, ecological effects and health persistence, and 15 second-grade indexes. By using the historical database which contains large amount of data and feedback information, a health early warning model of plain reservoir was designed and established with big data basing on decision tree C5.0 algorithm. The Dingdong Reservoir＇s health status was forecasted as a living example for research. The results show that the accuracy of the early warning model is 91.92%. The empirical rule set reveals that the factors influencing classification of health warning degree are mainly contained with piezometric level, water quality, underground water dynamic level and sediment variation index. The reservoir is in health state The early warning model is feasible and effective be used widely in reservoir health management. making problems such as numerous and diverse of data, , corresponding to no warning level. use of big data resources, and it can Taking advantage of big data, a series of demand diversifications of related interest groups and the difficulty in regional overall plan can be solved in water management. The results are beneficial to the sustainable development and scientific management of plain reservoir.
出处
《应用基础与工程科学学报》
EI
CSCD
北大核心
2017年第5期880-891,共12页
Journal of Basic Science and Engineering
基金
国家科技支撑计划项目(2015BAB07B05)
山东省省级水利科研及技术项目(SDSLKY201305)
山东省农业重大应用技术创新项目(SDNYCX1531963)
关键词
平原水库
大数据
健康预警
数据挖掘
决策树
plain reservoir
big data
health early warning
data mining
decision tree
