摘要
针对沈阳地铁一号线重工街站至启工街站区间隧道开挖引发地面沉降变形的问题,利用现场实测的地表沉降变形数据建立BP神经网络模型,并进行网络训练与预测。预测结果表明,时间序列神经网络模型能够很好地表达地面沉降监测数据序列间的非线性关系。利用BP神经网络建立的预测模型,所得预测值与实测值拟合很好,是预测地铁施工引发地面沉降变形的一种有效方法,能为沈阳地铁隧道的设计及施工提供科学合理的依据。
BP neural network model is established with the use of the field data of fround subsidence caused by subway excavation in the section between Zhonggong Street Station and Qigong Street Station of Shenyang Metro. The network forecasting shows that neural network modeling of time series well expresses the non-linear relationship among series of ground subsidence monitoring data. The predictive values obtained from the prediction model based on BP neural network fit well with the measured values. As an effective method for predicting ground subsidence caused by subway construction, BP neural network can serve as a scientific and reasonable basis for the design and construction of the Shenyang subway tunnel.
出处
《地质灾害与环境保护》
2014年第3期97-102,共6页
Journal of Geological Hazards and Environment Preservation
基金
沈阳市2013年度科技计划项目(No.F13-164-9-00)
关键词
地铁开挖
地面沉降
神经网络
沉降预测
subway excavation
ground subsidence
neural network
subsidence prediction
