改进GM(1,1)模型对地铁开挖沉降的预测  

Prediction of settlement caused by metro excavation based on improved GM(1,1) model

作　　者：戴文亭[1] 王振 王宇放 王琦 DAI Wen-ting;WANG Zhen;WANG Yu-fang;WANG Qi(School of Transportation,Jilin University,Changchun 130022,China)

机构地区：[1]吉林大学交通学院

出　　处：《沈阳工业大学学报》2019年第5期571-577,共7页Journal of Shenyang University of Technology

基　　金：国家自然科学基金资助项目(51178204)

摘　　要：为了提高GM(1,1)模型在地铁施工过程中地表沉降量预测的精度,预防较大沉降或其他危险出现,提出了改进GM(1,1)模型预测方法.通过对比不同原始序列个数建立起来的预测值并确定最佳原始序列个数后,对GM(1,1)模型进行优化,并对构造背景值进行优化.通过缓冲算子对原始序列进行优化,之后再构造背景值进行优化.结果表明,背景值对模型的预测影响较小,缓冲算子在原始序列变化较大、变化不平顺时优化较好.In order to enhance the prediction precision of surface settlement in the process of metro construction with the GM(1,1)model and prevent the occurrence of large settlement or other risks,an improved prediction method based on GM(1,1)model was proposed.By comparing the predictive values established with the different original sequence number and determining the optimal original sequence number,the GM(1,1)model was optimized as follows:the optimization of established background values;the optimization of original sequence with the buffer operator;the optimization of secondly established background values.The results show that background values have small influence on the model prediction.The optimization of buffer operator is satisfactory when the original sequence has great and irregular changes.

关 键 词：地铁 暗挖法 地表沉降预测 灰色理论 GM(1 1)模型 缓冲算子 背景值 优化 

分 类 号：TU[建筑科学]