基于DIC粗-细搜索方法的单轴压缩砂样的应变分布及应变梯度的实验研究
Experimental studies on strain distribution and gradients for sand specimens under uniaxial compression based on digital image correlation coarse-fine search method
-
摘要: 基于粒子群优化算法及Newton-Raphson迭代方法的数字图像相关粗-细搜索方法,观测了两个砂样在应力控制加载条件下的3种应变场的分布及演变规律,确定了应变在小范围内猛增之前,大范围内的应变非均匀分布区域的纵向(砂样的加载方向)应变梯度的范围.研究发现,当载荷增加到一定程度后,水平线应变的正峰值(正值代表受拉)和垂直线应变的负峰值(负值代表受压)都会出现在未来的局部化带内部,而剪切应变的峰值有时会发生偏离,这一点可由两种线应变场和剪应变场的差别进行解释.前两者呈倾斜的条状,而后者呈团状.3种应变的纵向梯度的范围大致相同,在0.001~0.002 mm-1之间,如果3种应变的纵向梯度比上述范围小,砂样就不会出现局部的应变猛增现象和随后的宏观裂纹.在局部化带之外,存在一定的塑性应变及应变梯度,这一点不能用非经典的弹塑性模型进行解释.Abstract: The distribution and the evolution of three kinds of strain fields are measured for two sand specimens under uniaxial and stress-controlled compression by use of a digital image correlation method with coarse-fine search capability based on the particle swarm optimization algorithm and the Newton-Raphson method. The range of longitudinal strain gradients in a wide non-uniformly deformational zone is determined before strains rise rapidly in a small band with highly localized strain. It is found that the peak values of the positive (or tensile) horizontal linear strain and the negative (or compressive) vertical linear strain fall into the future strain localized band after the loading exceeds a certain value, while the peak of the shear strain can deviate from the band, as can be explained by the differences between linear strain fields and shear strain fields. The former two are inclined and zonal beyond the uniformly deformational stage, while the latter is lumpy. The nearly same range is found for the longitudinal (or vertical) gradients of three kinds of strains, i.e., 0.001~0.002 mm, before the occurrence of the narrow localized band. If the strain gradients are lower than the values mentioned above, rapid strain increases will not occur in a small band within the sand specimen, nor will macroscopic cracks appear. Outside the band, certain plastic strains and gradients exist, as cannot be explained by the nonclassical elastoplastic models.