 |
| Titel |
Evaluating the uncertainty in geochemical modelling for CO2 storage. The example of Ketzin. |
| VerfasserIn |
Marco De Lucia, Pascal Audigane, Nicolas Jacquemet, Mich. Kühn |
| Konferenz |
EGU General Assembly 2010
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 12 (2010) |
| Datensatznummer |
250043305
|
|
|
|
|
|
| Zusammenfassung |
| Several sources of uncertainty are associated with geochemical modelling of reservoirs
considered for CO2 storage : on one hand only few available data are generally available for a
particular reservoir, which are affected by measure errors, and whose representativity is in
most cases questionable; on the other hand, the phenomenological description itself of the
chemical fluid-rock interactions relies heavily on experimental determination of physical
observables, which are summarized in a thermodynamical and chemical databases
used by the numerical simulators. The latter is for example the case of the CO2
disposal in saline aquifer, where the high ionic strength of the formation fluid requires
a Pitzer ion interaction model to evaluate with sufficient accuracy the activities
of the considered species. Typically, parameters for Pitzer model are discordant
following different authors and data related to a conspicous number of relevant ions are
often unavailable or unreliable -i.e. derived for different P/T conditions or ionic
strength of the solution -, which makes very difficult to estimate the accuracy of the
predictions.
This contribution presents our effort in evaluating the reliability of chemical simulations
in the case of Ketzin on-shore CO2 storage project. Based on available fluid and
mineralogic analysis, an initial equilibrium model was determined, i.e. matching both the
observed fluid composition and the saturation with the mineral phases present in
the sandstone layer of the Stuttgart formation which constitutes the reservoir. A
sensitivity analisys based on small perturbations of such initial model was then
performed, in order to assess the influence of measurement errors and possibly define a
compositional range which can be assumed in spatially variable simulations. This step is
then followed by the sensitivity to the Pitzer parameters that are collected in the
chemical database used throughout this study; both the parameters themselves and
their dependence on temperature are investigated. To reduce the number of degrees
of freedom of all possible simulations while dealing with systems with several
components, some strategies from the theory of Experimental Design have been
applied; all modelling was performed using the chemical equilibrium and speciation
program PHREEQC and a newly developed interface to the statistical environment
R.
Finally, simple 1D reactive transport simulations modelling the injection of pure CO2
were made to highlight the influence of spatially variable initial concentrations, of the
solubility model for CO2 in saline water and the dissolution and precipitation kinetics of
selected minerals. |
|
|
|
|
|
|