Rock-mass partitioning

Hudson, J. A. Andersson, J., 2003. T-H-M-C modelling of rock mass behaviour 2 - the input data and rock mass partitioning. Proceedings of the GEOPROC2003 conference, Stockholm. 

We concluded the Part I paper stating that the full development of T-H-M-C modelling is still at an early stage and it is not evident whether current codes provide the information that is required. This also applies to the input data and rock mass partitioning Part II conclusion we have practical techniques for obtaining the input information for simple codes, but not for the new generation of T-H-M-C codes — which require a more systematic and overall approach to the problems described herein. 

The relative area of mine solid waste tails (per 100,000 M3 of rock mass) is 0.7-0.8 of the total area. On average, the disturbed areas of uranium ore exploration site are partitioned as follows 32.3% of disturbed land is occupied by dumps, 27.2%, by pits, 20.3%, by industrial areas, 13.3%, by tails, and about 10%, by other types of land disturbance. 

The idea is to partition the rock mass into domains, as illustrated in Fig. 1, so that the domains can be considered as having constant properties for the purposes of modelling. 

Example partitioning of a rock mass into domains. 

For engineering purposes, we define a rock domain as a rock volume within which a certain rock property or function of rock properties is statistically homogeneous. An immediate corollary of this definition is that the rock mass could well be partitioned into different domains for different rock properties or different functions of rock properties. 

Using the Domain Index values, the rock mass is partitioned into domains according to the requirements established in (A), and the results presented in an appropriate format (e.g. a map, tables of representative numerical values, via computer-based visualization tools). 

This example, with its generic extrapolation, means that in the numerical modelling consideration must be given to the procedure for dealing with the separate components of the input data. This may involve alternative and sequential rock mass domain partitioning, as described in Hudson and Harrison (2002). 

In this Part II paper, we have concentrated on the input data for T-H-M-C modelling and the associated problem of partitioning the rock mass into domains within each one of which the rock properties are assumed to be constant... 

We presented two illustrative examples of obtaining input information. The first concerned partitioning of a borehole into domains (rock segments lengths with similar characterisitcs). The second concerned a Test Case in which a 6(X) m x 180 m x 120 granitic rock mass was studied to establish how additional information would influence the estimated rock properties and their variation. 

Hudson J A Harrison J P, 2002. The principles of partitioning rock masses into structural domains for modeling and engineering purposes. In Mining and Tunnelling Innovation and Opportunity. Proc. of the NARMS-TAC2002 Symposium, Eds. Hammah R., Bawden W., Curran J., Telesnicki M. Univ. Toronto Press, 623-628. 

There are two principal sources of reliable partitioning data for any trace element glassy volcanic rocks and high temperature experiments. For the reasons outlined above, both sources rely on analytical techniques with high spatial resolution. Typically these are microbeam techniques, such as electron-microprobe (EMPA), laser ablation ICP-MS, ion-microprobe secondary ion mass spectrometry (SIMS) or proton-induced X-ray emission (PIXE). 

Major elements in melts formed from mantle rocks are by definition compatible, and most of them are well buffered by the residual minerals, so that their concentrations usually vary by factors of less than two in the melts. In contrast, trace elements, particularly those having very low partition coefficients, may vary by as many as three orders of magnitude in the melt, depending on the degree of melting. This is easily seen from the mass-balance-derived equation for the equilibrium concentration of a trace element in the melt, Cl, given by (Shaw, 1970)... 

Immobile organic matter is usually attributed directly to rocks or deposits. In this case the mass transfer with the rock organic matter is treated as absorption directly by rock, by analogy with adsorption of mineral components from water. As opposed to adsorption, absorption is a slower process as time is required for a uniform distribution of adsorbates in the organic matter. That is why instead of adsorption partition coefficient for mineral components (see ionic exchange) in this case is used by analogy absorption partition coefficient ior nonpolar compounds K. . 

---