加权动态GM(1,1)模型在高层建筑沉降预测中的应用  

Application of Weighted Dynamic GM(1,1) Model in Settlement Prediction of High-Rise Building

作　　者：王景环 程雍 豆红磊 韩以江 WANG Jinghuan;CHENG Yong;DOU Honglei;HAN Yijiang(Key Laboratory of Geotechnical Mechanics and Engineering Safety in Hunan Province,Xiangtan University,Xiangtan 411105,China;College of Civil Engineering and Mechanic,Xiangtan University,Xiangtan 411105,China;Hunan Tianyu Detection Corporation,Loudi 417000,China)

机构地区：[1]湘潭大学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 [3]湖南省天宇检测有限公司,湖南娄底417000

出　　处：《测绘地理信息》2022年第3期43-46,共4页Journal of Geomatics

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：针对传统GM(1,1)模型易受外界环境干扰,易使原始序列发生突变,导致模型精度低、残差大的问题,将相对误差与时间距离结合起来,对原始数据进行加权,建立加权GM(1,1)模型;引入新陈代谢思想,建立加权动态GM(1,1)模型。利用某高层建筑沉降监测数据,将传统模型、加权模型和加权动态模型进行对比分析,并通过MATLAB实现程序化建模。结果表明,加权动态模型能更好地拟合高层建筑沉降监测数据,预测值与实测值变形趋势一致。The traditional GM(1,1)model is susceptible to the interference of external environment,resulting in the abrupt change of original sequences,so the model has low accuracy and large residuals. In view of these problems,we establish a weighted GM(1,1)model by combining relative error with time distance to weigh the original data. Based on this model,we establish a weighted dynamic GM(1,1)model by introducing new data. The settlement monitoring data of high-rise buildings are used to compare the traditional model,weighted model and weighted dynamic model. The programmatic modeling is realized by MATLAB. The results show that the weighted dynamic model can better fit the settlement monitoring data of high-rise buildings,and the predicted values are consistent with the measured values.

关 键 词：加权动态GM(1 1)模型 相对误差 时间距离 沉降预测 

分 类 号：P258[天文地球—测绘科学与技术]