摘要
工程勘察中准确获取土层的剪切波速是一项重要的工作。基于BP神经网络可以逼近任意非线性函数的原理,以某深度处土样的土工试验参数为输入变量,以剪切波速为输出变量,构建了BP神经网络预测模型,对该深度处土层的剪切波速进行预测。经过实际工程的检验,该模型预测的剪切波速值与实测值间的最大误差小于20%,且预测结果明显优于经验公式等其他估算值,证明预测结果能满足工程要求,预测模型在镇江地区典型土层中具有实用价值。
It is an important work to obtain the shear wave velocity of soil layer accurately in engineering investigation.Based on the principle that BP neural network can approximate any nonlinear function,the geotechnical test parameters of soil samples at a certain depth are taken as input variables,the shear wave velocity is taken as the output variable,constructed a BP neural network prediction model,the model predicts the shear wave velocity at this depth.After the actual project test,the maximum error between the shear velocity predicted by the model and the measured value is less than 20%,the prediction results are obviously better than other estimates.It is proved that the prediction results can meet the engineering requirements,the prediction model has practical value in the typical soil layer in Zhenjiang area.
作者
徐为海
Xu Weihai(Zhenjiang Urban Investigation and Surveying Institute Co.,Ltd.,Zhenjiang Jiangsu 212000,China)
出处
《山西建筑》
2023年第18期77-80,共4页
Shanxi Architecture
关键词
BP神经网络
预测
剪切波速
土工试验参数
BP neural network
forecast
shear wave velocity
geotechnical test parameters
作者简介
徐为海(1980-),男,高级工程师,从事岩土工程勘察工作。
