【作者】 戴武奎；

【导师】 王清；

【作者基本信息】 吉林大学 ， 地质工程， 2007， 硕士

【摘要】 马鞍山浅埋隧道是浙江省诸永高速公路上一条公路隧道,其修建具有很大的意义。本文以马鞍山隧道为研究对象,通过采用对隧道开挖前的数值模拟分析、现场监控量测、围岩压力的时序预测以及围岩最终稳定性预测等综合方法,对隧道围岩压力在隧道开挖后对隧道的稳定性影响进行预测,本文以现场监控量测为依托,首先通过有限元分析对隧道的开挖进行模拟,得出断面测点的控制压力,后利用监测数据采用人工神经网络的方法建立学习样本并进行学习对隧道的围岩压力进行时序预测,最后采用回归分析的方法研究其变形速率,得到其最终的围岩稳定压力与控制压力进行对比,对隧道的最终稳定性进行预测,判断隧道监测断面在开挖后将会逐渐趋于稳定。更多还原

【Abstract】 The highway tunnel is a type of underground structure ,which is in the rock and soil medium.When excavating tunnel,the surrounding rock often appears the rock break,the rock slip ,the rock falling and the rock plastic deformation because of deformation pressure and loose pressure,at the same time the tunnel support and lining will also generate crack because of big deformation on the inner. In consideration of the important meaning of design and construction and the long time in construction and running,the highway tunnel demands of a great deal of information between construction to guide the design and constrcution for acquiring the superior structure.The shallow tunnel’s embedded depth is shallow and its surrounding rock is breaking ,so the difficulty on the shallow tunnel is greatly.When excavating tunnel,if subjected to violent shake or handle measure not appropriate,not only the arch crown will be appear of large-scale landslide and caving,but also the side wall will be appear to collapse or destabilization.And if the constrcution method is not appropriate ,it may be the cause of the second lining appearing a great deal of crack,influencing the integrity,durable and using of the tunnel.In construction steps,always adopts the departmental or delamination construction.Demand the construction unit finish preliminary support ,sprayed concrete and assemble anchor arm as soon as possible to have enough time to make surrounding rock stabilize. Ma anshan tunnel in this article is a shallow highway tunnel,so comprehensive consider the characteristics of that tunnel,use various means of monitor,measuration and forecast is very necessary to carry on accurate information feedback and dependable forecast under complex construction condition.Surrounding rock pressure is one of the important information during the period of excavating tunnel.Through monitoring the change of the surrounding rock contact pressure,we can understand the stable status of the surrounding rock in time and can also carry on a stability control.So it is important to predominate the information of surrounding rock and use the monitoring data to forecast the change of surrounding rock during excavaling..But the tunnel rock engineering mechanics behavior and its deformation and destroyable mechanics is random and obscure.Therefore,though mechanics is still the necessary method to solve the engineering problem,it is not the only.During the rise of the artificial intelligence academics for example neural network,genetic algorithm etc.,we can resolve this indetermination engineering problem.The neural network have stronger capacity of nonlinear dynamic handling.It can handle nonlinear rock soil body information.A lot of domestic and interational scholars make use of this theory putting forward the new method to analyse the surrounding rock stability.They carry on the knowledge through a great deal of engineering data ,find out the inside relation in the data,then make use of these relations can make a reasonable of judgment through these relations.This article applied BP artificial neural network to built up the trend estimate network model of the surrounding rock pressure .It adopted the actual monitor data to train the network and carried on an time sequence estimate by this trained BP neural network.As a result, this model has very high estimate accuracy.As the same time,combine the theory of regression analysis to estimate the end stability and transform velocity of the surrounding rock.The whole process information of excavating tunnel can attain at that time and can be done a forecast in time.The main content and result of this article are as follows:(1)Begin with the finite element simulation of the Ma anshan sallow tunnel,build up the tunnel excavating model.Imitate the variety of the tunnel during excavating in order to be point to monitor the tunnel.Then contrast the calculation and monitoring result for discovering the problem under construction in time.(2)After excavatng ,we can lay the representation monitor point in time according to the result of the finite element simulation.Then monitor the frequency of surrounding rock pressure according to the requirement.From the result of monitoring,we can give conclusion that the tunnel cross-section is stable.(3)Making use of the BP neural network model in MATLAB and combining the monitoring data of surrounding rock pressure,we can forecast the variety of the surrounding rock pressure.The result has higher accuracy and better credibility.(4)Making use of the regression analysis,we can forecast the end stability of the surrounding rock.The result is that passing by regression analysis is less than calculation by theory.The surrounding rock is stable.In addition we can judge the variety velocity of the surrounding rock pressure is descending continuously.That show the transform of surrounding rock tend in the stability.更多还原