摘要
为实现基坑变形的高精度预测,以基坑现场变形监测数据为基础,先通过粒子群算法、改进变分模态分解开展变形数据的多尺度分解,并通过信息挖掘构建多类多尺度映射模型;其次,利用改进麻雀搜索算法和相关向量机实现各类多尺度映射模型的预测处理,并通过局部组合权值、全局组合权值构建综合组合权值,实现基坑变形的变权组合预测。分析结果表明,通过PSOIVMD的数据分解处理,所得dnSNR值为24.16,且小波阈值、经验模态分解的dnSNR值分别为33.97和35.16,对比而言,PSO-IVMD具有更优的数据分解能力;同时,在数据分解处理的基础上,通过C4、C9、S3和S7监测点的变权组合预测,得到其REm值范围为2.02%~2.11%,Tt值范围为185.04~194.12 ms,具有较优的预测精度和较快的收敛速度,可验证本文预测思路的合理性,且通过外推预测,基坑后续变形速率较小且具收敛趋势,加之预测后的累计变形介于16.13~25.89 mm,也在变形控制值范围内,说明工作井基坑施工过程中的变形控制效果较优,能够为其安全施工提供理论依据。
To achieve high-precision prediction of foundation pit deformation,based on on-site deformation monitoring data,multi-scale decomposition of the data is carried out through the particle swarm optimization algorithm and improved variational mode decomposition method,and a multi class multi-scale mapping model is constructed through information mining.Secondly,the improved sparrow search algorithm and correlation vector machine are used to achieve prediction processing of various multi-scale mapping models,and comprehensive combination weights are constructed through local and global combination weights.The analysis results show that through the data decomposition processing of PSOIVMD,the obtained dnSNR value is 24.16,and the dnSNR values of wavelet threshold and empirical mode decomposition are 33.97 and 35.16,respectively.In comparison,PSO-IVMD has better data decomposition ability.At the same time,based on data decomposition processing,the variable weight combination prediction of monitoring points C4,C9,S3,and S7 was used to obtain a REm value range of 2.02%~2.11%,and a Tt value range of 185.04~194.12 ms,which has excellent prediction accuracy and fast convergence speed.This verified the rationality of the prediction idea in this article.Moreover,through extrapolation prediction,the subsequent deformation rate of the foundation pit is small and has a convergence trend,and the cumulative deformation after prediction is between 16.13~25.89 mm,which is also within the range of deformation control values,indicating that the deformation control during the construction process of the working well foundation pit is effective,providing a theoretical basis for its safe construction.
作者
郭子飞
Guo Zifei(China Railway First Group Co.,Ltd.,Xi′an 710054,China)
出处
《工程勘察》
2026年第2期76-82,共7页
Geotechnical Investigation & Surveying
关键词
基坑
多尺度信息
数据分解
相关向量机
变形预测
foundation pit
multi scale information
data decomposition
correlation vector machine
deformation prediction
