Hi all,
as a beginner I have trouble wrapping my head around Elmer. This is partly because I get confused on what to set where and how. First I started mit ElmerGUI to create my .sif files. It has all the options for the solvers and general simulation settings. But I find the interface somewhat confusing. So I started to write the .sif directly. Then I realised that there are some differences between the .sif generated by ElmerGUI and the options shown in the ElmerSolver manual. For example Elmer Gui defines the coordinate systems as
Coordinate System = Cartesian
whereas the ElmerSolver Manual reads
Coordinate System = "Cartesian 2D"
Also I read somewhere that options use
option = REAL <number>
and sometimes I read
option = <number>
Now I wonder is there a complete list of all the available options and their default values. I know the ElmerModels manual has a lot of info an the particular models but not on the gnereal solver options and no default values.
Thank you
complete keyword list
-
raback
- Site Admin
- Posts: 4832
- Joined: 22 Aug 2009, 11:57
- Antispam: Yes
- Location: Espoo, Finland
- Contact:
Re: complete keyword list
Hi
There are in some cases different variations of the same keyword bacause we strive at backward compatibility but sometimes development enables better conventions. If you sometimes see .sif files written ~10 years ago they have even quite a different layout. Still they may be used even today. The "Cartesian 2D" is nowadays not needed since the "Cartesian" + the mesh dimension defines the coordinate system.
The keyword list can never be complete. Firstly because the development tends to include some undocumented features. Also the keywords are derived. Imagine that you would have a solver for variable "MyField". Then you would have Dirichlet conditions like "MyField = Real 0.0", you could have condition related to it "MyField Condition = Equals Coordinate 1", or soft limiters "MyField Upper Limit = Real 1.0" etc. All these choices, and many others, become possible when you introduce "MyField".
When the keywords are read their type must either be in the SOLVER.KEYWORDS database of be casted. Each keyword may be Real, Integer, String, File. If the keyword is not in the database and is not casted, the program cannot handle it. It does not matter that it is casted even though it is in the database. Of course conflicts are not allowed.
-Peter
There are in some cases different variations of the same keyword bacause we strive at backward compatibility but sometimes development enables better conventions. If you sometimes see .sif files written ~10 years ago they have even quite a different layout. Still they may be used even today. The "Cartesian 2D" is nowadays not needed since the "Cartesian" + the mesh dimension defines the coordinate system.
The keyword list can never be complete. Firstly because the development tends to include some undocumented features. Also the keywords are derived. Imagine that you would have a solver for variable "MyField". Then you would have Dirichlet conditions like "MyField = Real 0.0", you could have condition related to it "MyField Condition = Equals Coordinate 1", or soft limiters "MyField Upper Limit = Real 1.0" etc. All these choices, and many others, become possible when you introduce "MyField".
When the keywords are read their type must either be in the SOLVER.KEYWORDS database of be casted. Each keyword may be Real, Integer, String, File. If the keyword is not in the database and is not casted, the program cannot handle it. It does not matter that it is casted even though it is in the database. Of course conflicts are not allowed.
-Peter