摘要
龙潭隧道是沪蓉国道主干线湖北境内一条主要隧道,它的全长近8 900 m,埋深约500 m,为典型的深埋长隧道.隧道经过地区的地质条件复杂、地应力较大.为了合理地计算围岩稳定性,利用实测位移信息,进行了围岩参数和地应力场的反演.基于实测收敛位移,采用进化神经网络法,提出了一种围岩参数和地应力场联合反演的新方法,并利用该方法对龙潭隧道多个典型断面进行了反演研究.为了验证反演结果,采用反演得到的围岩参数和地应力场进行了数值计算,其计算结果和实际情况吻合.计算结果表明,通过反演,能得到反映整个隧道围岩参数和地应力场的基本参数,反演结果能较好地反映实际情况.
Longtan Tunnel is one key tunnel of Hurong roadway in Hubei Province.This tunnel is a typical long and deep tunnel,whose length and depth are about 8 900 m and 500 m,respectively.For the complicated geological condition and high initial stress,the stability of surrounding rock is calculated,which is a very important work to back analyze the material parameters and initial stress from the measured displacements.Based on evolutionary neural networks proposed by author,one new back analysis method which can get material parameters and initial stress at the same time is proposed.By this new back analysis method,some typical sections of Longtan tunnel are computed.From numerical computation with back analysis results,we can found that the stability of tunnel is in good agreement with engineering practice.And the initial stresses from back analysis is better coincided with the measured ones.So,from the back analysis,suitable parameters of whole tunnel can be obtained conveniently.
出处
《交通科学与工程》
2010年第4期11-16,共6页
Journal of Transport Science and Engineering
基金
国家自然科学基金项目(41072233)
关键词
龙潭隧道
进化神经网络
反演
Longtan Tunnel
evolutionary neural network
back analysis
