基于BP神经网络的单桩极限承载力预测

从基桩高应变动测试验的结果中提取特征参数,组成反映基桩竖向承载力的特征向量,利用BP神经网络的非线性映射特性和学习能力,建立了特征向量和单桩竖向极限承载力之间的一种对应关系,实例研究表明,人工神经网络的预测结果比凯司法更接近静载试验结果,因此,此方法有一定的工程实用价值。

Predication of Ultimate Bearing Capacity of Pile by BP Neural Network

Parameters,gathered from the high strain dynamic test of foundation pile,are used to reflect the characteristic vector of vertical bearing capacity of foundation pile.At the same time,the relationship between characteristic vectors and vertical ultimate bearing capacity of pile is established by using BP neural network with non-linear mapping character and learning ability.From the example research,it is shown that the predicted data of neural network model is similar with the static loading test,compared with the CASE.Therefore the prediction of vertical ultimate bearing capacity of single pile by BP neural network is feasible and effective in the foundation pile.