Uranium, Mining and Hydrogeology VI
Workshops UMH II
Workshop 2: Reactive transport
Chairs: W.E. Falck and D. Blowes
The discussion focussed more on the modelling of reactive transport and its problems than on conceptual approaches in general.
The participants tried to summarise the state-of-the-art in coupled reactive transport modelling, mainly from the standpoint of view of applicants.
Three mutually dependent complexes can be distinguished: Concepts, Supporting Data, and Mathematical/Numerical Realisations.
The problems surrounding the supporting data, in particular thermodynamic and kinetic data of general nature, have already been highlighted in a workshop organised by SMU in 1996. Other, i.e. case data, should always be subject to sensitivity analyses in order to be able to estimate the influence of analytical errors etc.
The conceptual approach to a problem, i.e. the perception the modeller has, significantly influences the outcome of any model. This has been highlighted in several intercomparison exercises, e.g. the CHEMVAL project, where different modellers came to different conclusions using the same basic data set. Again, sensititivity analyses help to elucidate the criticality of assumptions. Other topics discussed were the problem of up-scaling, i.e. the transferability of data and concepts derived in laboratory experiments to the field, and the problem of heterogeneity of natural systems. Both problems are closely associated.
While the modelling of saturated flow in three dimensions can be considered state-of-the-art, particular aspects, such as dual-porosity systems or multi-phase including unsaturated flow, still poses some conceptual and numerical problems. While the fundamental physics of these processes are well understood, the time-dependent parametrisation of real cases pose a problem.
Static equilibrium modelling of geochemical processes, again, can be considered state-of-the-art. Here the description of the kinetics of heterogeneous processes, such as dissolution/ precipitation, co-precipitation and sorption provide the real challange. Many geochemical system do not reach equilibrium within common observation times (hours...months) and therefore transient and secondary phases may control what is observed. Also, these processes mainly occur on surfaces and may involve only a small fraction of the total solid or dissolved mass, and therefore may escape analyses.
The common strategy to couple flow and chemical reactions is either via solving the respective mass balance equations simulataneously, or by assuming local chemical equilibrium and step-wise iteration between the flow and the chemical system. However, feedback between heterogenous chemical reactions which may constrain or widen flow paths and the hydraulic system is rarely considered.
At present around ten codes for coupled chemical reaction and transport are in general use. It was felt that their number will narrow down to just two or three, as it was the case with the geochemical codes.
The great number of codes makes the verification and quality control of modelling results difficult and respective techniques/strategies should be developed. Lack of documentation for some the codes, which were originally developed mainly as a research tool for in-house use, is one major obstackle for their use in a regulatory context. The most promising way forward appear to be bench-mark exercises, as have been used e.g. in comparing code performance for the CHEMVAL project.
The major constraint in applying coupled chemical reaction and transport models still is computing time. One participant cited an example where the laboratory experiment only took a few minutes but its computer simulation a few hours, even on an advanced computer. Clearly, the target should be the opposite in order to fully utilise the potential power of modelling, a ratio of at least 1:1000 between model execution time and real problem time being desirable.
© A. Berger <email@example.com>, 17.05.2010, http://www.geo.tu-freiberg.de/umh/UMHII_Workshop2e.htm