 |
| Titel |
Upscaling of advective transport in a fractured clay-rich sedimentary formation |
| VerfasserIn |
M. Willmann, W. Kinzelbach, F. Stauffer, B. Lanyon, P. Marschall |
| Konferenz |
EGU General Assembly 2009
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 11 (2009) |
| Datensatznummer |
250026706
|
|
|
|
|
|
| Zusammenfassung |
| Clay-rich sediments appear to be very favorable for deep geological storage of nuclear waste, mainly due to their low permeability and high retention capacity for cationic radionuclides. Performance assessment to test this hypothesis for specific sites relies heavily on numerical modeling of radionuclide transport. The need for Monte Carlo simulation and, subsequently, the costly coupling with chemical reactions make it impossible to run these simulations on the known local scale and upscaling is required. We look here at a generic fractured sedimentary host rock formation where the fracture permeability is on the order of the permeability of the matrix. The sediments are heterogeneous and characterized by large scale structures on the order of the domain size. Tools exist to upscale smaller scale heterogeneities and fracture networks in a nearly impermeable matrix, but it is still unclear how to upscale such a hybrid medium.
Our objective is to find an upscaling method for conservative, advective transport in a sedimentary, clay-rich, fractured formation dominated by large scale heterogeneities.
In a first step only the sediment matrix is considered and the fractures are ignored. 3D numerical simulations are performed at the local scale of a sedimentary host rock dominated by some clay-rich background strata and a single more conductive continuous layer consisting mainly of fractured limestone. Both units are internally highly heterogeneous with correlation lengths on the order of the domain size. The fine scale grid consists of about 15 million nodes. We then gradually upscale the hydraulic conductivity field block-wise, conserving total flux through the domain, and perform particle tracking ignoring local dispersion. We compare the results in terms of travel time and spatial deviation of the particle location relative to the mean flow direction after passing through the domain. Then we reproduce the local scale results for the different block sizes using the proposed upscaling methods. Finally, the fractures are added to the local scale model and the procedure is repeated for the hybrid medium.
For upscaling of the sediments we find that for smaller blocks the use of an advection-dispersion type equation is reasonable. But for larger blocks a multi-porosity model becomes necessary. These large blocks become particularly important if field scale reactive transport is simulated. In any case, to obtain reasonable results, it appears important to keep the binary nature of the host rock intact within the upscaled models. An effective model including all the large scale features seems not reasonable in the given example. Finally, some preliminary results for the additional consideration of the fractures are added. |
|
|
|
|
|
|