一种基于GSI弱化的应变软化模型

Mohr-Coulomb和Hoek-Brown破坏模型是目前运用最广泛的两种岩体破坏模型。为了能够直观地描述围岩高应力条件下的脆性破坏,众多学者提出基于这两种破坏模型的岩体参数取值方法,主要包括基于Mohr-Coulomb模型的CWFS模型以及基于Hoek-Brown模型的DISL模型和BDP模型。上述模型在表征岩石的脆性破坏方面均有一定的适应性,但是由于存在高围压条件下的硬化现象以及参数取值物理意义不明确等问题,其在实际工程中的应用受到一定限制。在上述模型的基础之上进一步研究Hoek-Brown破坏模型的参数取值规律,在分析地质强度指标GSI值在岩石压缩变形过程中的变化规律的基础上,通过构建地质强度指标GSI值与围压以及塑性应变的函数关系式,建立一种新的基于GSI弱化的应变软化模型——GSI弱化应变软化模型；然后通过锦屏二级水电站白山组大理岩以及Tennessee大理岩三轴压缩试验数值模拟对该模型进行验证。分析表明：该应变软化模型能够较好地模拟大理岩的三轴力学特性。最后运用该模型评价锦屏二级水电站引水隧洞开挖松弛范围,可为岩体开挖支护提供一定参考。

A strain-softening model based on GSI softening

The Mohr-Coulomb and Hoek-Brown failure criteria are the two most widely used ones at present. To simulate brittle failure of rocks in deep tunnels, the models based on these two failure criteria have been proposed, including the CWFS (cohesion weakening and friction strengthening) model, the DISL (damage initiation and spalling limit) model, and the BDP (brittle ductile plastic) model. These models have been used to simulate brittle failure of hard rocks. However, because of the issues of strain hardening under high confinement and large ambiguity in model parameter determination, it is very challenging to apply these models in practical engineering application. Based on the variation of the GSI (geological strength index) value during compression of rocks and by defining the GSI value as a function of plastic strain and confinement, a strain-dependent GSI-softening model, which is based on the Hoek-Brown failure criterion, is proposed in this study. This model is implemented in FLAC^3D to simulate the triaxial compression tests on T_2b marble in Jinping-II Hydropower Station and the Tennessee marble. It is found that the proposed model is able to simulate the mechanical behaviors of the brittle-ductile transition observed in the triaxial compression tests on Jinping and Tennessee marbles. Finally, this model is used to evaluate the relaxation depth of a headrace tunnel in Jinping-II Hydropower Station, and the calculated depths are in good agreement with the field observations.