A numerical reactive transport model for crystalline rocks is developed and evaluated. The model is based on mineral maps generated by X-ray micro computed tomography (X-μCT); the maps used have a resolution of approximately 30 μm and the rock samples are on the cm scale. A computational grid for the intergranular space is generated and a micro-DFN (Discrete Fracture Network) model governs the grid properties. A particle tracking method (Time Domain Random Walk) is used for transport simulations. The basic concept of the model can now be formulated as follows; “when a particle is close to a reactive mineral surface it has a certain probability to get sorbed during a certain time span. Once sorbed it will remain so a certain time”. The model requires a number of input parameters that represent the sorption properties of the reactive minerals. Attempts are made to relate the parameters to traditional distribution parameters. The model is evaluated by comparisons with recent laboratory experimental data. These experiments consider two rock types (veined gneiss and pegmatitic granite) and two radionuclides (cesium and barium). It is concluded that the new reactive transport model can simulate the experimental data in a consistent and realistic way.

建立并评估了结晶岩的数值反应输运模型。该模型基于X射线计算机断层扫描（X-μCT）生成的矿物图；使用的地图的分辨率约为30μm，岩石样品的尺寸为cm。生成用于晶间空间的计算网格，并使用微型DFN（离散断裂网络）模型控制网格属性。粒子跟踪方法（时域随机游走）用于传输模拟。该模型的基本概念现在可以表述如下：“当颗粒靠近反应性矿物表面时，在一定时间范围内有一定的可能性被吸附。一旦被吸收，它将保持一定的时间。” 该模型需要许多输入参数，这些参数代表了活性矿物的吸附特性。试图将参数与传统的分配参数相关联。通过与最新的实验室实验数据进行比较来评估该模型。这些实验考虑了两种岩石类型（脉状片麻岩和斜长花岗岩）和两种放射性核素（铯和钡）。结论是，新的反应性输运模型可以以一致和现实的方式模拟实验数据。