(GRS 292) Enhancement of the Codes d³f and r³t

Autor: 
GRS, Steinbeis-FZ, Universität Frankfurt, Universität Freiburg, Universität Jena
Jahr: 
2012

With regard to long term safety analyses, the hazardous material which may be released from repositories in deep geological formations is thought to pass three zones before it can affect men, the near field, the far field and the biosphere. The codes d³f and r³t are focused on flow and transport modelling in the far field, i. e. the geosphere. Their development started in a period when rock salt was selected as a potential host rock for radioactive and chemo-toxic waste. The codes were focussed on modelling density-driven flow and nuclide transport in the overburden of a salt dome. For this reason they were restricted to applications in porous media where heat transport could be neglected.

 

Taking into account crystalline rock or mudstone as alternative possible host media, fractures and fracture systems in an otherwise porous medium must now be regarded, too. Furthermore, flow and transport have also to be modelled within the host formation. Here, the maximal warming in the surrounding of the casks is not allowed to exceed approximately 100 centigrade so that the effect of heat on water density and thus on flow and transport cannot be neglected any longer.

 

One difficulty in density-driven flow is to predict the stability of a model. Within this project, a stability number is derived to decide if a flow regime is in a stable or unstable state.

 

By the extensions of d³f and r³t presented in this report the codes are now also empowered to model heat transport. The thermohaline flow problem is described mathematically, and the three field equations to be solved in d³f are realised for two variants, the Boussinesq approximation and the complete equation system.

 

Modelling of porous media is complemented with the explicit modelling of fractures. Here, fractures are represented by lower dimensional structures. The finite volume discretisation is adapted accordingly.

 

The new option to model free surface flow provides the ability to take into account pumping wells and groundwater recharge, too.

 

The evolving equation system was of increased complexity. The solvers within d³f had therefore to be improved and optimized. A higher order finite volume method is introduced to improve accuracy. New filtering algebraic multigrid (FAMG) methods are developed and implemented. Additionally, the UG parallelisation concept is advanced to a flexible tool, the parallel communication layer (PCL), that enables d³f and r³t to use modern parallel computers effectively.

 

Pre and postprocessors are adapted where necessary.