基于数据分解重构和AM-CRU-MLR模型的基坑变形研究与应用

打开文本图片集

WANG Andong， ZHANG Xuegang， NING Bo（Shaanxi Railway Institute，Weinan 7140OO， China）

Abstract：Inorder toefectivelygraspthedeformationcharacteristicsoffoundationpits，basedonthedisplacementand deformation monitoringresultsoffoundationpits，and basedontheanalysisofthecurentdeformationcharacteristicsof foundationpits，the integratedempiricalmodeandsampleentropyareusedtodecomposeandreconstructthedeformationdata， accuratelydecomposingitintoreal deformation termsanduncertaindeformation terms.Furthermore，CRUneural networkand multiplelinearregresionareusedtopreprocessdiferentdeformationterms，achievinghigh-precisionpredictionoffoundation pitdeformation，andusingtheresults tograspthedeformationcharacteristicsoffoundationpits.Theresultsshowsthatthrough datadecompositionandreconstructionprocesing，thedeformationdataoffoundationpitscanbeefectivelydecomposed，andits superiorityisobvious compared totraditional decomposition methods；TheAM-CRU-MLR model showedrelativelybeter predictionacuracy，fullyverifying theefectivenessoftepredictionappoachandproviding theoreticalsupportfortheanalysis of deformation characteristics of foundation pits.

Keywords：foundation pit；deformationdata；decompositiontreatment；deformationprediction；deformationcharacteristics

0 引言

随着市政工程的快速发展，基坑工程数量日益增加，在带来巨大经济效益的同时，也引发了一定的工程问题，如基坑多处于建、构筑物密集区，基坑失稳会引发一系列连续破坏；为确保基坑安全施工，变形控制显得格外重要。（剩余6868字）