模糊聚类法在大坝监测资料分析中的应用  

Application of fuzzy clustering method in dam monitoring data analysis

作　　者：廖铖[1] 蔡德所[1] 李苗[2] 张吉燕[1] 王一立 

机构地区：[1]三峡大学水利与环境学院,湖北宜昌443002 [2]华北水利水电大学资源与环境学院,河南郑州450011 [3]三峡大学土木与建筑学院,湖北宜昌443002

出　　处：《人民长江》2015年第13期86-89,99,共5页Yangtze River

摘　　要：为了解模糊聚类法在大坝监测资料分析中应用的适宜性,利用该法对水布垭面板堆石坝2013年面板挠度变形监测数据开展了研究。首先将监测管道划分为若干监测点,运用模糊聚类分析法对监测点进行分类,并采用F统计量评价聚类效果以确定最佳分类,对分类结果进行分析比较,可确定关键的面板变形监控点。随后建立面板挠度变形统计模型,考虑上游水深、温度、时效等因素,利用多元逐步回归分析法对关键点监测数据进行预报拟合。结果表明,利用模糊聚类分析法处理大坝监测数据具有便捷、高效的优点,模型预报结果拟合度较高。In order to understand the applicability of fuzzy clustering method in analysis of dam monitoring data, it is applied to analyze the slab deflection deformation series 2013 of Shuibuya concrete face rockfill dam. Firstly, the monitoring pipeline was divided into several sections to set up monitoring points. The fuzzy clustering method is used to classify the deformation monitoring point, and the best classification is determined with F statistic evaluation. The classification results are analyzed and compared, so the key slab deformation monitoring points were determined. Then a slab deformation statistical model was established by consideration of the upstream water depth, temperature and aging factors, and the monitoring data of key points were predicted and fitted with multiple stepwise regression analysis. The analysis results show that using fuzzy clustering method to deal with dam monitoring data possesses advantages of convenient, efficient, and the fitting degree of the model prediction results is relatively high.

关 键 词：面板变形 模糊聚类 统计模型 多元逐步回归 

分 类 号：TV698[水利工程—水利水电工程]