Application-oriented testing and advancement of numerical models for the simulation of flow and transport processes granite rocks

The presented study has been carried out against the background of problems and questions, which are discussed within long-term safety analysis with regard to the deposition of radionuclide waste within a deep repository in fractured rock. The main research consists in the analysis of - general transport processes within a fractured aquifer by means of 2d and 3d model calculations of different scales, with the objective of determining sensitive parameters, - the influence of heat-fields, with regard to groundwater flow and mass transport within an fractured aquifer, - the new modules of the FEFLOW 5.2 program version, called ''Multi- Species Transport'' and ''Discrete Feature Elements'' for the implementation of radionuclide decay chains, as well as the implementation of explicit fractures in local and regional scale. Additionally it should be asked whether and to what extent the hydrogeological programme FEFLOW provides an opportunity to deliver additional information to the one-dimensional far-field CHETMAD module of the EMOS programme. The models have been tested with increasing complexity as far as possible by using field data from the Itatskij investigation area in the Nishnekanskij Granitoidmassif, near the city of Krasnojarsk in Siberia. The model results show that it is possible to demonstrate the general behaviour of flow and transport processes for different scales and dimensions by using the FEFLOW programme and to detect the essential parameters on the basis of simplified fracture systems. In addition to the Darcy velocity of the solid stone matrix, the various fractures and their characteristics, such as aperture, amount, connectivity and orientation, with regard to the flow-field, are the most important factors. The radionuclide transport also depends to a high degree on the solubility and sorption coefficients of radionuclides on the different rock components. In addition it becomes obvious, that the heat-fields, independent of their occurence as natural geothermical temperature gradient or artificial heat-field over a few hundred years generated by the radioactive decay of waste, have almost no consequences regarding groundwater flow and radionuclide transport within the fractured rock. Furthermore, it is shown that the new programme modules are appropriate to implementing radionuclide decay chains and explicit fractures on a local and regional scale. It is important that there are restrictions for their implementation concerning the amount of fractures and their intersection angles due to the necessity to create as many equilateral grid elements as possible to ensure numerical stability, as well as the fact that fracture implementation in 3d is bound to the availability of geological borders in the form of model slices. Unfortunately, it was not possible to reproduce a radionuclide transport calculation in detail, which was previously done with the ''r³t'' programme. Although the principle transport phenomenon for the mother nuclide and its concentration are very similar between the different programme results, for the daughter nuclide there are differences in advanced calculation time. It is presumed, that the solution of the transport equation caused the different model results. However, only a detailed comparison between the mathematical implementation of the relevant equations should permit the differences. The different key aspects of the hydrogeological model, compared to the typical one-dimensional models of long-term safety analysis, show important additional information. If adequate data is provided, then spacious hydrogeological model calculations can lead to a better understanding of the local and regional groundwater flow regime, a more realistic estimation of the transport path length, as well as flow volume. For this reason it provides a more reproducible description of the highly simplified fracture system in one-dimensional safety analysis. Nevertheless, it should be kept in mind, that the models are only able to show one time and spacial realisation of a possible parameter set. Therefore, they can only clarify well defined flow and transport problems in detail. A stochastical parameter variation of rock characteristics, which is necessary for long-term safety analysis, can not be performed.