基于BP神经网络下的深长隧道突水预测分析与数值模拟研究

发布时间：2018-01-14 13:35

本文关键词：基于BP神经网络下的深长隧道突水预测分析与数值模拟研究　出处：《中南大学》2014年硕士论文　论文类型：学位论文

【摘要】：隧道突水突泥一直以来都是隧道施工过程中的重大难题,突水突泥可能会导致隧道工期延误,造成巨大的财产损失,甚至还可能成为隧道塌方的诱因,导致人员伤亡,施工机械损坏,对整个社会造成不利的影响。过往的研究表明,地质构造、地下水源和开挖扰动成为深长隧道突水突泥灾害发生的三大原因。为了保障隧道施工的安全性,减少不必要的财产损失,加强对隧道突水突泥前兆信息的研究是十分必要的。在隧道施工过程中,综合分析隧道突水突泥灾害诱因,预测突水突泥发生的可能性,采取针对性的控制措施,从而保障隧道的顺利施工和人员的安全。本文以江西钟家山隧道为依托工程,通过收集工程中突水突泥灾害发生时的现场数据,运用BP神经网络对深长隧道的突水突泥前兆信息进行预测,并采用数值模拟的方法对隧道突水过程进行了分析,论文主要完成了以下几方面的工作： (1)根据国内外有关深长隧道突水突泥涌水量预测的研究,探讨了深长隧道涌水预测过程中的影响因素,结合钟家山隧道现场的施工情况,确定了涌水量预测的六个变量参数。 (2)依据钟家山隧道施工过程中采集的数据,利用BP神经网络建立突水突泥的灾害预测模型。通过对现场观测点涌水量的预测,验证了该预测模型的有效性和正确性。 (3)运用MIDAS和FLAC3D数值模拟软件对深长隧道的突水情况进行数值模拟分析,研究了隧道突水前兆信息的演化规律,分析了隧道开挖过程中渗流速度、渗透水压力、孔隙水压力的变化情况。
[Abstract]:Water and mud inrush has always been a major problem in tunnel construction, which may lead to the delay of tunnel construction period, cause huge property losses, and even may become the inducement of tunnel collapse. Causes casualties, construction machinery damage, to the entire society has a negative impact. Previous studies show that the geological structure. The underground water source and excavation disturbance are the three main causes of water inrush and mud inrush in deep tunnel. In order to ensure the safety of tunnel construction and reduce unnecessary property losses. It is necessary to strengthen the study on the precursor information of water and mud inrush in tunnel. In the process of tunnel construction, the inducement of water and mud inrush in tunnel is analyzed synthetically, and the possibility of water inrush is predicted. Take targeted control measures to ensure the smooth construction of the tunnel and the safety of personnel. Based on the project of Zhongjiashan Tunnel in Jiangxi Province, this paper uses BP neural network to predict the precursor information of water inrush mud in deep long tunnel by collecting the field data of water inrush and mud inrush disaster. The water inrush process of tunnel is analyzed by numerical simulation. The main work of this paper is as follows: 1) based on the domestic and foreign research on the prediction of water inrush and mud inrush from deep long tunnel, the influencing factors in the process of water inrush prediction in deep long tunnel are discussed, combined with the construction situation of Zhongjiashan tunnel site. Six variable parameters of water inflow prediction are determined. 2) based on the data collected during the construction of Zhongjiashan Tunnel, a disaster prediction model of water inrush and mud inrush is established by using BP neural network. The validity and correctness of the prediction model are verified. 3) using MIDAS and FLAC3D software to simulate the water inrush of the deep tunnel, and study the evolution rule of the precursor information of water inrush in the tunnel. The variation of seepage velocity, seepage water pressure and pore water pressure during tunnel excavation are analyzed.
【学位授予单位】：中南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U456.3

【参考文献】