基于改进BP网络的盾构隧道开挖地表沉降预测

盾构隧道施工中引起的地表沉降是衡量开挖方式是否合适的关键指标。文中在介绍BP神经网络及盾构施工引起变形情况的基础上,对基于BP神经网络的盾构隧道开挖引起的地表沉降预测进行了研究,考虑了训练样本中奇异数据的剔除,采用变步长的方法,并选取适当的动量项系数,综合考虑各种影响因素,建立了盾构隧道开挖引起的地表沉降预测的BP网络模型,并对广州地铁二号线进行了具体的预测分析。分析结果表明:理论计算结果与工程实际情况一致,误差小于5%,所建立的预测模型是令人满意的。

Predicting the Surface Settlement by Shield Tunneling Based on Modified BP Network

Prediction of surface settlements due to shield tunneling is a key to the evaluation of excavation way.Based on the introduction of the BP neural network and the surface settlements by shield tunnel construction,this paper researches the intelligent prediction of shield construction deformation,and the singular data in exercise sample are excluded.Considering all the factors,a modified BP network model on the surface settlement of shield tunnel is put forward,which is exercised via varied study step and appropriate momentum coefficient,then the model was applied to predict the surface settlement of Guangzhou Metro Line 2.The results of theoretical calculation and the actual situation are in good agreement and the error is less than 5%,so the proposed model is satisfactory.