摘要
针对传统路基沉降数据典型的“小样本、非线性”特点,提出GM-SVR组合模型,并应用于新疆奎屯乐土驿段铁路线路路基沉降预测。结果表明:组合模型相较于单一灰色模型GM(1,1)和SVR支持向量回归模型,预测精度更高,能更好地体现路基沉降变化趋势。
With respect to the typical characteristics"small sample and nonlinear"of traditional subgrade settlement data,a GM-SVR combined model was proposed and applied to the subgrade settlement prediction of Letuyi Section of Kuitun Railway,Xinjiang.The result shows that,compared with the single grey model GM(1,1)and SVR support vector regression model,the combined model has higher prediction accuracy and can better reflect the change trend of subgrade settlement.
作者
宋军
黄华
张中华
王鹏
孙大伟
李国成
SONG Jun;HUANG Hua;ZHANG Zhonghua;WANG Peng;SUN Dawei;LI Guocheng(Kuitun Public Works Section of China Railway Urumqi Group Co.,Ltd.,Kuitun 833214,Xinjiang,China;Information Technology Institute of China Railway Urumqi Group Co.,Ltd.,Urumqi 830000,China;School of Automation and Electrical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处
《路基工程》
2025年第3期13-18,共6页
Subgrade Engineering
基金
兰州市科技计划项目:企业铁路路基形变监测关键技术研究与应用(2023-RC-7)。
关键词
路基沉降
不良地质
沉降观测
模型预测
精度比选
沉降趋势
沉降量
经济性
subgrade settlement
unfavorable geology
settlement observation
model prediction
precision selection
settlement trend
settlement amount
economy
