双洞隧道施工引起地表移动的多参数反分析研究

Multiparameter back analysis of surface movement induced by double tube tunneling

应用随机介质理论计算隧道开挖引起的地表移动是目前广泛使用的方法,在关键参数的取值上反分析是最有效的手段。在双洞隧道的反分析问题上,通常认为2个隧洞的参数相同而采用双参数反分析,但这种做法不利于反映实际情况,为此文中提出为每个隧洞引入各自计算参数进行多参数反分析。实际算例表明,双洞4参数的分析结果优于双参数的分析结果。但是,4参数反分析问题复杂性大大提高,传统的模式搜索方法不能很好地搜索到最优参数。为克服该问题,采用单纯形混合加速遗传算法作为双洞隧道4参数反分析问题的求解方法。实际的应用及测试表明,该方法能高精度的、稳定的搜索出全局最优参数。

Stochastic medium method（SMM） is widely used in the field of surface movement induced by tunneling. And back analysis is the most useful method for determining key parameters. In the problem of twin tunnel back analysis the parameters of each tunnel is supposed same, so the problem can be solved for dual-parameters. But this way isn＇t access truthfulness. For improvement, multiparameter back analysis method is presented, in which each tunnel should have different parameters. The case study shows that the back analysis result of 4 parameters is better than the dual-parameters. But 4 parameters back-analysis of twin tunnel is much more complex so that traditional pattern search method can＇t solve the problem well. So simplex hybrid accelerating genetic algorithm（SHAGA） is introduced for solving this problem. The test of application shows that SHAGA can solve it accurately and stably.