基于小波神经网络的地铁盾构施工地表沉降预测研究

A Study of the Wavelet-Neural-Network-Based Prediction of the Surface Settlement for Shield Tunneling

地铁盾构施工引起的地表沉降对施工安全影响较大,应加以预防和控制。根据影响地表沉降的主要参数,建立了基于小波神经网络的盾构施工地表沉降预测模型,分析了预测结果的可行性,对比了它在收敛速度、预测精度等方面较传统BP神经网络的优势。结合北京地铁6号线实地测量数据,验证了小波神经网络用于沉降预测的准确性和可行性。

The surface settlement caused by tunneling has to be prevented and controlled because it has a great impact on the construction safety.A wavelet-neural-network-based tunneling settlement prediction model is established according to the main influential parameters affecting the surface settlement in the paper,with the feasibility of the predicted results analyzed,and the advantages of the method in both the convergence speed and precision compared with those of the traditional BP neural-network-based methods.With the measured data of Line 6 of the Beijing Subway as practical examples,both the accuracy and the feasibility of the wavelet-neural-network-based settlement prediction are testified.