摘要
针对地下隧道开挖引起的地表下沉问题,运用概率统计理论建立了数学模型,并根据BP神经网络理论,通过对BP神经网络算法的改进,采用反分析方法确定岩体移动变形参数。利用所建模型对隧道开挖引起的地表垂直移动(下沉)进行了具体的计算分析,将理论计算值与实测下沉值进行对比,二者十分吻合。对比结果表明,所给出的数学分析模型及参数确定方法符合工程实际,为解决地下隧道开挖引起地表下沉预计分析问题开辟了新的途径。
The prediction of displacement of rock masses and their surface effects are an important problem of the rock mass mechanics in the excavation activities, especially in the underground engineering excavation in mountain areas. Any tunnel and underground excavation will inevitably disturb the original stress field, which in turn causes displacement and deformation of rock mass which lead to ground surface subsidence. Although there are several empirical and semi-empirical formulae available for predicting ground surface subsidence, most of these do not simultaneously take displacement of rock mass into consideration, resulting in inaccurate predictions. Based on results of the statistical analysis of a large amount of measured data in underground excavation, the fundamental mathematical model of ground surface subsidence is established by using the theory of probability and statistics. In order to improve compution rate and stability of BP neural networks, the coefficient of dynamic-item and varied steps are adopted for back analysis of rock mass displacement due to underground tunnel excavation. The agreement of the theoretical results with the ground surface subsidence by using the numerical analysis methods shows that this model is satisfactory and the formulae obtained are valid and thus effectively for analyzing and predicting the ground surface subsidence or the rock mass displacement and deformation due to underground tunnel excavation.
出处
《岩石力学与工程学报》
EI
CAS
CSCD
北大核心
2004年第Z2期4752-4756,共5页
Chinese Journal of Rock Mechanics and Engineering
关键词
隧道工程
地表下沉
BP神经网络
tunneling engineering, ground surface subsidence, BP neural networks
