含不同随机缺陷数目岩样的破坏过程、前兆及全部变形特征

Simulation of failure processes, precursors and overall deformation of rock specimens with different numbers of initially random imperfections

利用拉格朗日元法模拟了平面应变单轴压缩条件下具有初始随机材料缺陷的岩石试样的破坏过程、前兆和宏观力学响应.利用若干FISH函数于岩样内部规定初始缺陷并计算全部变形特征.比密实岩石弱的缺陷在破坏之后经历理想塑性行为密实岩石在破坏之后先是经历线性应变软化行为,然后是理想塑性行为.随着缺陷数目的增加,剪切带间距缩小;岩样的强度下降.在应变软化阶段,最大不平衡力的明显的突增是由于被剪切带切割的试样块体发生了沿剪切带方向的运动.试样内部的缺陷数目越少,应变软化阶段最大不平衡力的峰值越高.无论从单元破坏角度,还是从侧向应变-轴向应变曲线、体积应变-轴向应变曲线、计算得到的泊松比-轴向应变曲线角度,都得到了相同的结论(即,缺陷越多,破坏前兆越明显).采用两条倾斜的扫描线对岩样最终的破坏形态进行了扫描,发现当扫描线间距等于半个至1个单元长度时,初始随机缺陷位置与岩样最终破坏形态具有很好的相关性.

The failure processes, precursors to failure and macroscopically mechanical responses of heterogeneous rock specimens with initially random material imperfections in uniaxial plane strain compression are numerically modeled using FLAC. FISH functions are used to generate initial imperfections and to calculate overall deformational characteristics. Beyond the failure of the imperfection that is weaker than the intact rock element, it undergoes ideal plastic behavior, while the intact rock element exhibits linear strain-softening behavior and then ideal plastic behavior once failure occurs. With an increase of the number of imperfections, the shear band spacing decreases and the strength of the specimen decreases. In the strain-softening stage, the rapid increase in the maximum unbalanced force is due to the movement of rock blocks separated＇by shear hands. The lower the number of imperfections is, the higher the maximum unbalanced force in strain-softening stage is. Both the failure process of the rock specimen and the overall deformation （lateral strain-axial strain curve, volumetric strain-axial strain curves and calculated Poisson＇ s ratio-axial strain curve） show that the precursors to failure become apparent as the number of imperfections increases. Using two scanning lines, the final failure pattern of a specimen is scanned. It is found that when the spacing of scanning lines is in the range of half and one element length, both the initial imperfection distribution and the final failure pattern have a good correlation.