Rock masses are highly characterized by size effect due to the existence of complex discontinuities, which could be reflected by representative volume element (RVE). This study exemplified calculation of RVE in Datengxia sluice rock mass analysis with consideration of anisotropy, which extends the application of traditional RVE featured by three-dimensional (3-D) equal sizes. Specifically, fracture network modeling was conducted to generate fractures in 3-D space. Fracture frequency values, which can comprehensively reflect the fracture density, size and occurrence in different lengths, were used to represent size effect of one-dimensional (1-D) parameters. Subsequently, geostatistic theory was introduced to synthesize substantial fracture frequency data and analyze their relationships in different regions and educe convergence size by analyzing the significant sub-theory of variogram. Linear representative volume elements (LRVE) in different directions were proved to be varied; thus, a modified RVE with orthogonal directions corresponding to individual LRVE values were applied. As RVE takes account on linear anisotropy features of rock masses, size effect can better guide parameter selection and numerical modeling of rock masses.

由于存在复杂的不连续性，岩体具有尺寸效应的高度特征，这可以通过代表性的体积元（RVE）来反映。该研究以各向异性为例，举例说明了大唐峡水闸岩体分析中RVE的计算，从而扩展了以三维（3-D）等尺寸为特征的传统RVE的应用。具体而言，进行裂缝网络建模以在3-D空间中生成裂缝。裂缝频率值可以全面反映裂缝的密度，大小和不同长度的出现，用于表示一维（1-D）参数的大小效应。后来，引入地统计学理论来合成大量裂缝频率数据，并通过分析变异函数的显着子理论来分析它们在不同区域的关系并得出收敛大小。事实证明，不同方向上的线性代表体积元素（LRVE）是变化的；因此，应用了具有与各个LRVE值相对应的正交方向的修改RVE。由于RVE考虑了岩体的线性各向异性特征，因此尺寸效应可以更好地指导岩体的参数选择和数值模拟。