Strategy for determination of thermal properties in crystalline rocks
|Location||International Geological Congress,oslo 2008|
|Author||Sundberg, Jan۱; Back, Par-Erik۲; Christiansson, Rolf۳; Ericsson, Lars۴; Wrafter, John۱|
|Holding Date||07 October 2008|
The Swedish Nuclear Fuel and Waste Management Company (SKB) is developing site descriptive models at the Forsmark and Laxemar areas, with the objective of siting a repository for spent nuclear fuel. The spatial variation of thermal conductivity in the rock mass is described in the thermal site descriptive model.
The local temperature field around a canister is of primary concern for the design of a repository. The current design criterion is specified as the maximum temperature allowed in the bentonite buffer outside the canisters. To fulfil the temperature requirement, a low rock thermal conductivity leads to larger distance between canisters than in the case of a high thermal conductivity. Therefore, the spatial variability of the thermal properties of the rock mass is important for the canister spacing.
The strategy for the thermal site descriptive modelling is to produce spatial statistical models of both lithologies and thermal properties and perform stochastic simulations to generate a spatial distribution of thermal properties that is representative of the modelled rock domain. The total variability within a rock domain thus depends on the lithology and the thermal properties of each rock type. Although the thermal conductivity of a single rock type may be close to normally distributed, the statistical distribution of thermal conductivity for the domain as a whole is far from normally distributed. Depending on their fraction of the total volume, the low-conducting rock types may determine the lower tail of the thermal conductivity distribution and influence the canister spacing.
The methodology involves a series of steps, eg. choice of simulation scale, definition of Thermal Rock Classes (TRCs) within the rock domain, stochastic simulation of TRCs (lithology in the domain), stochastic simulation of thermal conductivity within each TRC, and finally merging of the realisations of TRCs (lithology) with the thermal conductivity realisations. The result is a set of 3D realisations describing the spatial variability of thermal conductivity within the rock domain. These realisations can thereafter be upscaled to a desired scale. The described methodology can also be used for other types of rock properties.
The distribution of thermal conductivities comes primarily from TPS-measurements on rock samples in laboratory. In order to describe the spatial structure of thermal conductivity a very large number of samples are needed. A correlation between thermal conductivity and density has been found. The relationship can be explained by the mineral distribution for different rock types and the density and thermal conductivity for the minerals. For acid rock the thermal conductivity decreases when density increases, for basic rock the relationship is opposite. The relationship makes it possible to use density logging to estimate the spatial correlation structure within each rock type, instead of huge amount of thermal conductivity determinations.