Asaoka法在软土地基工后沉降量预测中的问题分析  

Problem Analysis of Asaoka Method on the Post-construction Settlement Prediction of Soft Ground

作　　者：周岳雷 刘忠汉 ZHOU Yuelei;LIU Zhonghan(CCCC Fourth Harbor Engineering Institute Co.,Ltd.,Guangzhou 510230,China;CCCC Key Lab of Environmental Protection&Safety in Foundation Engineering of Transportation,Guangzhou 510230,China;The Third Engineering Company of CCCC Fourth Harbor Engineering Co.,Ltd.,Zhanjiang 524005,China)

机构地区：[1]中交四航工程研究院有限公司,广东广州510230 [2]中交交通基础工程环保与安全重点实验室,广东广州510230 [3]中交四航局第三工程有限公司,广东湛江524005

出　　处：《人民珠江》2020年第8期93-97,共5页Pearl River

摘　　要：针对Asaoka法在软土地基工后沉降量预测中存在的不确定性问题,借助BP神经网络的超强非线性拟合能力,将实测数据向等时间间隔进行转化,建立了不同起算点和不同时间间隔下的Asaoka法回归分析模型,对其工后沉降量进行预测,并选用数学方法对预测值进行了稳定性分析。利用某软土地基处理项目的实测资料,并综合分析了预测结果的稳定性。分析结果表明:①BP神经网络的超强非线性拟合能力可有效用于软土地基实测数据的转化;②仅仅通过相关系数的大小来评判Asaoka法预测结果的可靠性是不可取的;③β_1细小的误差就可能引起预测结果的显著误差;④可通过合理加大时间间隔来提高预测结果的可靠性。In view of uncertainty of Asaoka method on the post-construction settlement prediction of soft ground,this paper transforms the measured data at constant time interval through BP neural network with good nonlinear fitting ability,establishes the regression analysis model at different starting points and different time intervals based on Asaoka method,predicts the post-construction settlement by the model and analyze the stability of the predicted value through mathematical methods,as well as comprehensively analyzes the stability of the prediction results with the measured data of a soft foundation treatment project.The analysis results show that:①BP neural network with good nonlinear fitting ability is effectively used in the transformation of the measured data of soft ground;②It is not advisable to judge the reliability of Asaoka method by the correlation coefficient;③The small error ofβ1 can cause significant error to the prediction results;④The reliability of prediction results can be increased by expanding time interval.

关 键 词：软土地基 ASAOKA法 工后沉降量 可靠性 

分 类 号：TU471.8[建筑科学—结构工程]