基于小波理论的混凝土坝变形PCA-IPSO-SVM预测模型  

PCA-IPSO-SVM Prediction Model of Concrete Dam Deformation Based on Wavelet Theory

作　　者：柳志坤 周兰庭[1] LIU Zhi-kun;ZHOU Lan-ting(College of Water Conservancy and Hydraulic Engineering,Hohai University,Nanjing 210098,China)

机构地区：[1]河海大学水利水电学院,南京210098

出　　处：《中国农村水利水电》2020年第7期185-189,195,共6页China Rural Water and Hydropower

基　　金：国家自然科学基金项目(51209078)。

摘　　要：混凝土坝变形实测数据往往具有较强的非线性和不确定性。首先通过小波软阈值法对监测数据进行去噪预处理,并根据大坝变形统计模型确定变形的影响因子;然后在SVM的基础上建立变形预测模型,为提高模型的有效性,将模型的输入量进行PCA降维处理,以降低因子间的相关性,并采用IPSO算法对SVM的参数进行寻优,建立了基于PCA-IPSO-SVM的组合预测模型。最后的实例应用表明该模型所得的预测值与实测值拟合较好,并与传统的单一SVM模型和BP神经网络预测结果进行对比,结果表明本文提出的方法具有较高的精度,在处理类似问题上具有较强的优势,为混凝土坝变形预测提供了一种新思路。The measured data of concrete dam deformation are often nonlinear and uncertain.In this paper,the wavelet soft threshold method is used to pre-treat the monitoring data and determine the influencing factors of deformation according to the statistical model of dam deformation.Then,a deformation prediction model is established on the basis of SVM.In order to improve the effectiveness of the model,the input of the model is dimensionalized by PCA to reduce the correlation between factors.Moreover,the IPSO algorithm is used to optimize the parameters of SVM,and a combined prediction model based on PCA-IPSO-SVM is established.The results show that the proposed method has high precision and a strong advantage in dealing with similar issues,providing a new way of thinking for concrete dam deformation forecasting.

关 键 词：混凝土坝 软阈值去噪 主成分分析 粒子群算法 支持向量机 

分 类 号：TV698[水利工程—水利水电工程]