基于改进 M5忆鄄主成分模型树的高心墙 堆石坝沉降变形预测.PDFVIP

第46卷第4期 河 海 大 学 学 报 ( 自然 科 学 版 ) Vol.46 No.4 2018年7月 Journal of Hohai University(Natural Sciences) Jul. 2018 DOI:10.3876/ j.issn.1000-1980.2018.04.011 基于改进M5忆鄄主成分模型树的高心墙 堆石坝沉降变形预测 王摇 飞,张宗亮,王佳俊,佟大威,任炳昱 (天津大学水利工程仿真与安全国家重点实验室,天津摇 300072) 摘要:针对高心墙堆石坝沉降变形过程动态非线性特点,建立基于改进M5忆鄄主成分模型树的高心 墙堆石坝沉降变形分析模型,在采用相关性分析甄选沉降变形影响因素和采用主成分分析将高维 影响因素空间进行降维的基础上,利用该全局分段非线性模型对高心墙堆石坝沉降过程进行分析。 通过与沉降量实测值对比,验证了改进M5忆鄄主成分模型树的有效性。 通过绝对差值和均方根误差 2 个指标对比分析改进M5忆鄄主成分模型树与M5忆模型树、多元线性回归模型、主成分回归分析模型 的预测结果,表明改进M5忆鄄主成分模型树预测沉降量具有更高的精度。 关键词:心墙堆石坝;沉降预测;M5忆模型树;相关性分析;主成分回归分析 + 中图分类号:TV641郾4 1摇 摇 摇 文献标志码:A摇 摇 摇 文章编号:1000-1980(2018)04 0353 07 Forecasting of the settlement deformation for high core rock鄄fill dam based on the improved M5忆鄄PCR model tree WANG Fei,ZHANG Zongliang,WANGJiajun,TONG Dawei,REN Bingyu (State Key Laboratory of Hydraulic Engineering Simulation and Safety,Tianjin University,Tianjin300072,China) Abstract:Considering the nonlinear and dynamic characteristics of settlement deformation, a global piecewise nonlinear analysis model of settlement deformation is proposed based on the modified M5忆鄄principal component (M5忆鄄PCR) model tree,toanalyzethesettlementprocessby selectingtheinfluencefactorswithcorrelationanalysis and reducing the dimension of high dimensional influence factors with principal component analysis method. The the improved M5忆鄄PCR model tree is verified by comparing with the measured settlement. By comparing the prediction results of improved M5忆鄄PCR model tree, M5忆 model tree, multiple linear regression, and principal component regression analysis model through the absolute difference and the root mean square error,the results show that imp