Report

Priority list from Code_Bright Consortium (CBC)/SKB.

CBC01) Errors in solutions using 2D axisymmetry.

• Some equations in axisymmetric conditions contain terms proportional to (1/r) where r is the radius (x coordinate in GiD). This is a singularity at the axis of axi-symmetry. 
• The easiest way to improve the solutions, mitigating numerical errors, is to include a boundary condition in the axisymmetry axis, restraining normal displacements.

CBC02) Improved output.

• We have divided this petition in two petitions (CBC02a and CBC02b).
• CBC02a: More output variables: dry density, water content, void ratio.
  • Dry density has been included in the post-process.
  • The ‘select output’ tab has been improved in usability.
  • Void ratio can be calculated directly from porosity.
  • Water content cannot be calculated because it combines element-wise and node-wise variables.
  • In addition, volumetric water content can be calculated in GiD with the “create result” tool by the product of porosity and degree of saturation.
• CBC02b: Water inflow integrated over all boundaries.
  • GiD allows the possibility of integrate results on a surface with the “integrate scalar” tool, which can be used together with the “create result” tool, allowing interesting combinations.
  • The Code_Bright Team will present some examples on the use of the ‘integrate results’ tool.

CBC03) Improved accessibility of convergence statistics.

• A new file fort.396 is written to analyze convergence.
• Few changes in fort.396 file.

CBC04) Non-porous/Rigid materials.

• This petition is similar to POS01.
• It is possible to uncouple volumetric strain effects on hydraulics (M → H), to improve numerical stability (especially relevant on excavation or in the canister). There is a document about this on the website –(UN)COUPLED_2D.pdf– in the ‘Documents of interest’ section, into ‘Downloads’.
• It is also possible to uncouple the hydraulic effects on mechanics (H → M) using Biot’s coefficient. In this case, we would need to input a Biot value closer to zero but different to zero. It has been documented in the User’s Guide.

CBC05) Improved parallelisation.

• We think it is a better solution, in terms of efficiency, to improve the preconditioner.
• Two loops of the iterative solver have been parallelized (before only one loop).
• Parallelization is now user-defined in the ‘solution strategy’ tab. It works only for iterative solver.

CBC06) Improve the numerical solution scheme.

• Since this part very much governs what the numerical tool can accomplish, this item is still very much prioritized for SKB. They do however acknowledge the efforts that the Code_Bright team have done so far regarding this issue and encourage them to keep up the good work.
• Document available on the use of elastic parameters including temperature and suction terms to model compression and swelling of clay-based materials. It is not a complete linearization of the problem, but it improves the numerical performance.
• New meshing possibilities by combination of tetrahedra and triangular prisms. Optimization of number of nodes in the rock. Regular elements in the rock. Better representation of gradients in the buffer.
• It is possible to uncouple volumetric strain effects on hydraulics (M → H), to improve numerical stability (especially relevant on excavation or in the canister). There is a document about this on the website –(UN)COUPLED_2D.pdf– in the ‘Documents of interest’ section, into ‘Downloads’.
• It is also possible to uncouple the hydraulic effects on mechanics (H → M) using Biot’s coefficient. In this case, we would need to input a Biot value closer to zero but different to zero. It has been documented in the User’s Guide.
• We will try to improve the preconditioner.

CBC07) Develop a large strain formulation.

• SKB states that this might be a large undertaking, but the fact is that in many of the systems they investigate, large strains develop. So, in order to claim that they represent the system accurately, a large strain formulation should be used.
• This petition is very similar to POS04.
• The updated Lagrangian method, already implemented in Code_bright, is so far the only tool available to take into account large displacements.
• We would need a PhD student working on this.

CBC08) GiD and Code_Bright.

 

 

Priority list from POSIVA:

POS01) Non-porous materials.

• Posiva would like the possibility of disabling the solution of the balance equation in non-porous materials like metals to avoid numerical problems.
• Posiva states that when the metal is “saturated”, the effective stresses are activated.
• This petition is similar to CBC04.
• It is possible to uncouple volumetric strain effects on hydraulics (M → H), to improve numerical stability (especially relevant on excavation or in the canister). There is a document about this on the website –(UN)COUPLED_2D.pdf– in the ‘Documents of interest’ section, into ‘Downloads’.
• It is also possible to uncouple the hydraulic effects on mechanics (H → M) using Biot’s coefficient. In this case, we would need to input a Biot value closer to zero but different to zero. It has been documented in the User’s Guide.

POS02) Porosity in intrinsic permeability laws used in BExM.

• The intrinsic permeability, when changes with the porosity, is function only of the macro-porosity in BExM. Posiva would like to have an option where the intrinsic permeability is function of the total porosity.
• An option to use the total porosity has been added: If P10≠0 (ICL=81;ITYCL=1) then total porosity in used instead of macroporosity.
• It has been documented in the User’s Guide.

POS03) kmicro in BExM.

• The first step should be to incorporate suction micro and retention curve micro.
• Jean Vaunat and Matías Alonso will work on this.

POS04) BExM in free swelling.

• We think free swelling is needed only for very special cases (laboratory experiments).

POS05) Large strains setting.

• Posiva states that, due to the large strains that appear in some cases (mainly shearing), the large strains setting is necessary, especially for following the rotations.
• This petition is very similar to CBC07.
• The updated Lagrangian method, already implemented in Code_bright, is so far the only tool available to take into account large displacements.
• We would need a PhD student working on this.

POS06) Random properties in elements.

• It is possible to assign properties in elements (e.g. porosity). Posiva says that it could be interesting to be able to assign random properties at the level of elements in order to reproduce heterogeneity. It could be useful for solving gas problems.
• We have implemented an option to set a heterogeneous porosity field.
• We can make a tutorial.

POS07) Water retention curve dependent of the temperature.

• Surface tension depends of the temperature and modifies the value of P₀. Experimentally, it has seen that the temperature dependence is stronger (Tang, 2005). Posiva would like to have an option for modifying the dependence and for locking the temperature dependence in surface tension as well. They can propose a law.
• Posiva have sent a proposal: to add a factor to P₀ which modifies P as a function of temperature
• It has been implemented.

POS08) Frozen water.

• Frozen water is present in manual but Posiva think it could be good to add an ice problem in the Tutorial.
• Anders Sjöland says that Martin Birgersson have solve the problem of frozen water in bentonite.
• Code_Bright Team has received information about Martin Birgersson’s work.
• We will make a tutorial.

POS09) Orthogonal load or stress to a surface.

• Posiva would like an option for imposing an orthogonal load or stress in a surface.
• Posiva states that the idea behind this is being able to model triaxial tests.
• From an engineering point of view, we think it is enough to use an octagonal (or even more-sided) prism to solve experiments on cylindrical samples. In this case, on each planar face, it is trivial to apply a normal stress.
• We can provide an example of a 24-sided triaxial that we use.

POS10) _bcf file.

• Old -999 method deleted. We recommend using the -99901 method.
• The Code_Bright team would like to have a case describing this issue.
• Solved the issue with the _bcf file.

POS11) Assigning different boundary conditions on the same surface.

• The Code_Bright team would like to have a case describing this issue.
• In principle, the solution is to impose, on that surface, a different boundary condition from the beginning (e.g. using slightly different gammas).

POS12) Post-process problems when the difference in time is small.

• In canister shearing, the saturation process takes some months and the shearing a few minutes.
• A possibility is to use negative time for the saturation process and positive time for shearing.
• The possibility of having different writing frequency in each interval has been implemented.

POS13) Parallel versions.

• Parallelization is now user-defined in the ‘solution strategy’ tab. It works only for iterative solver.

POS14) THMC plans.

• Anna Ramon is working on this.
• A branch version of the coupled system of chemistry in CB has been prepared.
• Once is compiled, very simple examples will be run to test the program operation and performance.
• Compilation of a preliminary THMC version of CODE_BRIGHT. This version is based only on equilibrium of chemical components.

POS15) Viscoelasticity.

• Kelvin model has been implemented.

POS16) Dynamic calculations.

• We would need a PhD student willing to work on this.

POS17) Thermal conductivity of the Argon Gas.

• It has been implemented.

 

 

Priority list from ANDRA:

AND06) Preprocess: Implementation of a nonlinear dependence of Young modulus on stress.

• We may use the Duncan-Chang Hyperbolic model or search in the literature for more alternatives. We would like to know if ANDRA has any specific suggestion.
• A three-linear elastic law has been implemented. It has been documented in the User’s Guide.
• An explanation about how to use the non-linear elasticity model for this purpose has been included in the User’s Guide.

AND04) Process: Improvement of error messages.

• The Code_Bright Team encourages the Consortium members to send us specific examples where error messages are needed and we will try to think about solutions in the form of error messages.
• Andra has shown some useful examples of error messages in their presentation in Paris.
• Posiva also thinks that this is an important issue.
• Window with trans.exe log for every procedure to detect errors more easily.
• Error message included in the code about elements with negative volume.
• Error included in case of conflict when using multiple boundary conditions on a particular node.
• Error included related to the use of Biot coefficient for models different from linear elasticity.

AND05) Post-process: Integration of quantities (water inflow, heat flux, etc. through surfaces (3D) or lines (2D)).

• GiD allows the possibility of integrate results on a surface with the “integrate scalar” tool, which can be used together with the “create result” tool, allowing interesting combinations.
• The Code_Bright Team presented some examples on the use of the ‘integrate results’ tool.

AND01) Pre-process: Anisotropy of properties in the elastic model (Young modulus) instead of argillite model.

• Anisotropy has been implemented within the elastic model.
• Plastic anisotropy has also been implemented.
• Biot coefficient can be used with elastic anisotropic model.
• Reading anisotropic angles from interface.

 

 

Priority list from GRS:

GRS01) THMC.

• Status of/further steps in modelling activities to incorporate chemistry in CODE_BRIGHT.
• A specific piece of work is being done by Anna Ramon.
• A simple example for salt concrete corrosion could be delivered by GRS for Spring meeting 2021.
• A branch version of the coupled system of chemistry in CB has been prepared. 
• Once is compiled, very simple examples will be run to test the program operation and performance.
• Compilation of a preliminary THMC version of CODE_BRIGHT. This version is based only on equilibrium of chemical components.

GRS02) Implementation of the entire Biot theory including the hydraulic part (Biot modulus).

• The hydraulic part of Biot’s theory has been implemented.

GRS03) Setting the option of constant fluid parameters.

• GRS states that the option of setting fluid parameters to a constant value should be given, especially the fluid density and the specific heat capacity.
• This option already exists. We have revised default values in each parameter and add an explanation in the User’s Guide. Basically, to set parameters with defaults to zero, a low value but greater than 10^-25 must be input. If the value is lower than 10^-25, then the parameter takes the default value.

GRS04) Integration of the double-structure model (BExM) for bentonite.

• Version developed for Beacon (micro- and macro-porosity with separate retention curves).
• Matías Alonso will work on this.

GRS05) Convergence problem for large suction contrasts

• Especially with bentonite involved. Some work-around or hints for the user.
• We recommend the use of the cubic law for relative permeability.
• In HM problems, we recommend to set a boundary condition P_l = -10 MPa in the tunnel surface.
• In THM problems, use relative humidity as boundary condition, instead liquid pressure.

GRS06) Improvement of the viscoplastic part in crushed salt modelling.

• In cooperation with GRS, that will deliver input for spring meeting 2021.

Finished tasks (we think):

• CBC03) Improved accessibility of convergence statistics.
• CBC05) Improved parallelisation.
• AND01) Pre-process: Anisotropy of properties in the elastic model (Young modulus) instead of argillite model.
• AND06) Preprocess: Implementation of a nonlinear dependence of Young modulus on stress.
• GRS02) Implementation of the entire Biot theory including the hydraulic part (Biot modulus).
• GRS03) Setting the option of constant fluid parameters.
• GRS05) Convergence problem for large suction contrasts.
• POS02) Porosity in intrinsic permeability laws used in BExM.
• POS07) Water retention curve dependent of the temperature.
• POS10) _bcf file
• POS11) Assigning different boundary conditions on the same surface.
• POS12) Post-process problems when the difference in time is small.
• POS13) Parallel versions.
• POS15) Viscoelasticity.

• POS17) Thermal conductivity of the Argon gas.

On-going work to be finished in the following months:

• Tutorial ice problem (POS08).
• THMC documentation and tutorials (POS14, GRS01).
• Continuation of THMC developments (POS14, GRS01)
• Integration of the double-structure model (BExM) for bentonite (GRS04).
• Tutorial for the heterogeneity implementation (POS06).