基于BP神经网络的偏压隧道拱顶沉降预测

Vault settlement prediction of the bias tunnel based on BP neural network

采用BP神经网络,以香山隧道拱顶沉降监测数据为样本进行训练,得到了相应的学习曲线,并采用所建立的神经网络预测模型,对隧道拱顶沉降进行了预测,结果表明:建立的BP神经网络模型能够很好的描述既有训练样本曲线变化特征,且预测精度与既有监测数据相关,亦与预测长度有关,预测长度较长时预测结果可信度降低。

Based on the monitoring vault settlement data of Xiangshan tunnel, BP neural network is adopt to obtain the learning principal curve. With the prediction model of BP neural network established in this paper, it carries out tunnel vault settlement prediction. The resuh indicates that, the BP neural network model established here could well describe characteristic of the training sample curve. Prediction accuracy are related to both the monitoring data and the forecasting length, and credibility of the predicted results would be reduced with increasement of the forecas- ting length.