用人工神经网络模拟土体塌陷的可行性分析

Feasibility Study on Soil Collapse Simulation with Artificial Neural Network Model

探讨了利用人工神经网络预测土体塌陷与土体参数和荷载条件之间关系的可行性。BP神经网络是最常用的神经网络之一,通过网络训练,最终确定了6个变量即输入信息、8个隐层和1个输出层的网络结构。在比较了预测值与试验值后,进一步证实了人工神经网络在评估土体塌陷方面效果明显。

The feasibility of using neural networks to model the complex relationship between soil parameters,loading conditions and the collapse potential are investigated.A back propagation(BP) neural network is one of the most common neural nerworks.After network training,the model consisting of six variables,eight hidden nodes and one output data is the most successful.Comparing the trial values with the values predicited from BP,the results show that neural networks are very efficient in assessing the complex behavior of soil collape.