基于滑动平均-灰色Verhulst模型的深基坑中长期变形预测  

Prediction for medium and long term deformation of a deep foundation pit based on sliding average-grey verhulst model

作　　者：关占印 刘天成 乔清源 Guan Zhanyin;Liu Tiancheng;Qiao Qingyuan(China Railway Engineering Equipment Group Shield Remanufacturing Co.,Ltd.,Tianjin 300450;the 5 th Engineering Co.,Ltd.of China Railway Wlh Bureau Group,Chongqing 400037;China Construction Tunnel Co..Ltd.,Chongqing 401320)

机构地区：[1]中铁工程装备集团盾构再制造有限公司,天津300450 [2]中铁十一局集团第五工程有限公司,重庆400037 [3]中建隧道建设有限公司,重庆401320

出　　处：《工程建设》2020年第2期62-66,共5页Engineering Construction

摘　　要：准确的深基坑变形预测和控制是当前地下工程监控测量研究领域的热点问题,尤其是深基坑变形的中长期预测,对保障现场人员和财产安全具有重要的预警作用。以某地铁盾构始发竖井深基坑为工程背景,采用滑动平均法(MA)对基坑变形现场实测的地表沉降原始数据进行去噪处理,增加监测数据的权重,然后引入灰色Verhulst模型对去噪后的数据进行中长期拟合预测,最后对比分析灰色GM(1,1)模型、灰色Verhulst模型、滑动平均-灰色Verhulst模型的预测精度。结果表明:滑动平均-灰色Verhulst模型的预测精度较高,适用于基坑中长期变形预测。该结果可为提高类似变形问题的预测精度提供启示作用。Accurate prediction and control of deep foundation pit deformation is a hot issue in the field of monitoring and measurement of underground engineering at present,especially in the medium and long term prediction of deep foundation pit deformation,which plays an important early warning role in guaranteeing the safety of field personnel and property. In this paper,a metro foundation pit is taken as the engineering background. the Moving Average(MA) method is used to denoise the original data of ground settlement measured in the field of foundation pit deformation,and the weight of monitoring data is increased. Then the grey Verhulst model is introduced to fit and predict the denoised data in the medium and long term. Finally,the comparison is made. The prediction accuracy of grey model GM(1,1),grey Verhulst model and Moving Average-grey Verhulst model are analyzed. The results show that the sliding average-grey Verhulst model has a high prediction accuracy and is suitable for long-term deformation prediction of foundation pits. The results can provide references for improving the prediction accuracy of similar deformation problems.

关 键 词：深基坑 中长期变形 滑动平均 VERHULST模型 预测 

分 类 号：U451.2[建筑科学—桥梁与隧道工程]