Structural Homogeneity Delineation Of Rock Mass And Risk Analysis Of Rock Mass Quality

Engineering rock mass quality plays a very important role to ensure the long-term stability and safety of a repository for high level radioactive waste. This paper takes granite rock mass of Gansu Beishan preselected area as research object, based on field investigation and experimental data of discontinuities, the spatial distribution characteristics of rock mass quality for each measured area of Jijicao block were studied. The main research contents and conclusions of this paper are summarized as follows:(1) Necessary pretreatment for the original sample data of discontinuities has been carried out at early, then, according to the principle of Mille method, the improved ninety-one patches network was used and realized the adaptive analysis and processing for the cases which don’t meet the rule of Lancaster. The final homogeneous region partition results were consistent with the actual geological conditions and the space distribution features.(2) On the basis of summarizing the existing grouping clustering analysis methods, the traditional fast clustering algorithm was optimized:according to the principle of maximum density point, the initial clustering centers were determined effectively; using the concept of membership degree, the attitude matrix was optimized reasonably; combined with Xie-Beni indicators, the optimal number of groups was distinguished scientifically; based on the difference of membership degree, the random discrete fracture was eliminated in time. Above optimal algorithm further reduced the possible error of cluster analysis, and ensured the scientificalness and reliability of attitude grouping results for each measured area in Beishan.(3) According to the characteristics of the joints images, the traditional Canny algorithm was optimized, so as to realize the effective extraction of joint image edge information. Then, according to the fractal dimension calculation principle of the walking divider method, programmed to achieve automated calculation of the fractal dimension of the surface profile curve, and established the quantitative relationship between the fractal dimension and JRC.(4) Based on the Q system classification method, combined with the field measured data of Beishan area, the main parameters affecting the rock mass quality were statistical analysed. Then, on the basis of principle of bayesian network, using Netica software to establish the risk analysis model of the rock mass quality evaluation, and applied and validated to the model.