基于优化的DGM(1,1)在桥梁变形预测中的应用  

Application of the optimized DGM( 1,1) model in bridge deformation prediction

作　　者：豆红磊 王景环 杨静 黎付安 程雍 Dou Honglei;Wang Jinghuan;Yang Jing;Li Fu＇an;Cheng Yong(Hunan Key Laboratory of Geomechanics and Engineering Safet,Xiangtan 411105,China;College of Civil Engineering and Mechanics,Xiangtan 4-11105,China)

机构地区：[1]岩土力学与工程安全湖南省重点实验室,湖南湘潭411105 [2]湘潭大学土木工程与力学学院,湖南湘潭411105

出　　处：《工程勘察》2018年第9期39-43,共5页Geotechnical Investigation & Surveying

基　　金：湖南省教育厅科学研究项目(17C1537);岩土力学与工程安全湖南省重点实验室项目(16GES10)

摘　　要：传统DGM(1,1)模型在建模过程中易受到随机扰动的影响导致预测精度低、残值大,为提高模型的预测精度,本文提出一种同时优化原始序列和初始条件的DGM(1,1)模型。首先将卡尔曼滤波和灰色理论结合,使用卡尔曼滤波对原始序列进行降噪,消除建模过程中由于原始序列引起的扰动误差。其次,对建模过程中的迭代初始值增加修正系数,建立优化的DGM(1,1)模型。最后结合工程实例,采用三种模型进行对比分析,研究结果表明,优化的DGM(1,1)模型的拟合预测效果优于传统模型,桥梁沉降监测值与预测值变形一致。The traditional DGM（ 1,1） model is vulnerable to random perturbations in the modeling process,resulting in low prediction accuracy and large residuals. In order to improve the prediction accuracy of the model,this article propose a DGM（ 1,1） model that simultaneously optimizes the original sequence and the initial conditions. Firstly,combining Kalman filter with gray theory,Kalman filter is used to reduce the noise of the original sequence and eliminate the disturbance caused by the original sequence in the modeling process. Secondly,the corrected coefficient is added to the initial value of the iteration in the modeling process to establish an optimized DGM（ 1,1） model. Finally,with the engineering examples,three models are used for comparative analysis. The results show that the fitting prediction effect of the optimized DGM（ 1,1） model is better than the traditional one, and the monitoring value of the bridge settlement is consistent with the predicted value.

关 键 词：DGM(1 1)模型 桥梁 卡尔曼滤波 变形监测 预测 

分 类 号：U446[建筑科学—桥梁与隧道工程]