Abstract: In order to accurately simulate the spatial distribution of rock joints to evaluate the stability of surrounding rock in high level radioactive waste geological disposal repository, at first the probability statistics theory was introduced to simulate the geometric characteristics of joints through the investigation of outcrop joint data, then by using the Monte-Carlo method the spatial distribution characteristics of the joints of the granite rocks in the Tianhu district in Xinjiang were simulated, which is a preliminary area for the disposal of high-level radioactive waste. The authors studied the different depth of joint geometrical characteristics through borehole TV, circled the homogeneous areas, proposed statistical methods and calculation formulas for borehole joint occurrence, density, and size, and finally established a joint 3D network model. The results show that:1) Using the chi-square test of the list table and the overlapping window method, the rock mass of the Tianhu borehole is divided into 0-90,120-240,270-360, and 390-600 m homogeneous zones; 2) Using SPSS software and based on the quantitative relationship between joint diameter, joint trace length, and gap width, the joint occurrence and diameter were in normal and negative exponential distribution respectively; 3) The difference between the model statistics and the field measured joint occurrence is within ±3°, the average diameter is within ±1 m, and the surface density is within ±0.01 m^-2 through data comparison, which verifies the accuracy of the 3D network simulation of the joint established by the Monte-Carlo method.

Key words: rock mass, geometric characteristics of joint, 3D network simulation, borehole TV