Simulations stochastiques de fractures pour determiner la granulometrie in situ du gisement de cuivre du mont-porphyre, Mines Gaspe, Quebec

We present a geostatistical method to assess the bloc size distribution at an early stage of ore deposit mechanical characterisation. The method is based on a simulation of fractures of different families identified from core-oriented drillhole surveys. The bloc size distribution is obtained by creating local graphs at a number of seed points within the simulated domain. Arcs of the local graph are cut by simulated discs. Starting from the seed, one identifies all nodes still-connected by a path to the seed. The number of connected nodes is proportional to the volume of the bloc. The method can be fast and flexible. It could be made as precise as required for the problem at hand by refining the grid. Computing time remains tractable because of the limited number of seeds necessary to represent the rock mass. Moreover, in addition to bloc volumes, the bloc shapes can simultaneously be obtained. The methodology is applied to the Mon-Porphyre low-grade Cu-deposit in the Murdochville area. The host rocks are intrusive porphyry and associated porcellanite. A stereonet analysis showed the presence of two dominant families plus a random. Fracture simulation was cross validated by creating pseudo-holes and comparing the fracture, distributions with those obtained with known drillholes. The simulated bloc distributions confirm previous qualitative assessment based on low fracture densities and high RQD (close to 99%). (Abstract shortened by UMI.)...