swak4Foam - New release
 

A new release of swak4Foam can now be downloaded through the usual channels (SVN).

New features are:
- global variables
- functionObjects that allow the inclusion of Python-Code
- a variant of the coded-functionObject that cooperates with swak (only for OF 2.)
- functionObjects that solve the laplacian and the transport equation

Further details on these new features can be found in the 2nd part of the presentation that is found here: http://openfoamwiki.net/index.php/Fi...Leoben2011.pdf

Bernhard

Forgot to mention it here (notifications already reached other channels - had so much fun playing with the new Social-media button on the network): a new release of swak4Foam ist out

Major improvements are:

• extension of the Field-parser to allow surface-fields and tensors
• conditional function objects (execute function objects on demand)
• improved Allwmake (catches common compilation problems)
• finiteArea now also supports the PDF-functionObjects
• and several other improvements

There are separate branches for the 1.x and the 2.x line of OpenFOAM (yes. 2.1 is supported)

PS: I was notified that this is going to be my 2001st post on the message board. In the honor of this this will be the first swak4Foam-release with a release name: "Kubrick"

Hi Bernhard,

concerning the conditional functionsObjects I a question.
Is it also possible to created a condition that calls the functionsObjects at a specific position of the solver, for example in the outer pressure correction loop of the pimple based solvers?

Martin

No. The problem is that functionObjects are only called at the end of a time-loop when the time is increased.

I THINK something like you want to could be achieved by specifying a second functionObjectList and calling it in the appropriate places. But you'd have to modify the solver code for that and it could lead to all kinds of weird behaviour for some function objects as they sometimes can't deal with being called twice for the same time-value

Hi Bernhard,

thanks for the reply. Could you please give me a hint what needs to be done in order to define a second functionObjectList. How is the "normal" functionObjectList defined in the code and how is this list executed from the solver?

Martin

The normal functionObjectList is defined in the Time-class/object. The constructor for FOL (one of them) assumes that he gets a dictionary with one entry "functions" from which he constructs the list. An example of the "non standard" usage of that have a look in the conditional functionObjects-part of the simpleFunctionObjects

Bernhard

Hi Bernhard,

thanks for your answer. I wil have a look át the conditional FunctionObjects and try to understand how it works.

Martin

Last friday (the 13th) a new release of swak4Foam hit the servers. One of the big innovations of this release is that I implemented a system that allows identifying consecutive releases by a system of ascending numbers. This system should make it easier to identify releases and find out whether the current release is the newest. I will call these numbers "version numbers" (not yet trademarked). I think this system is very useful in a lot of contexts. Feel free to use it for your own projects (reference to the place you saw it first would be nice). ;)

The new release will receive the number 0.2.0 (previous releases have been numbered with 0.1.x)

Other new features are
- temporal operators ddt and oldTime. Other operators from fvc are added
- functions for location of minimum and maximum
- Field-parser now supports tensors and pointFields
- New variables for Python-integration
- boundary conditions groovyBCDirection and groovyBCJump
- fully parallel compilation
(see README for a full list).

Downloading is described at http://openfoamwiki.net/index.php/Co...am#Downloading

bug report
 

Hej,

I wanted to use the bug report on the sourceforge website, but somehow it is not accepting my login (I used the openID login) and when I go to the bug report website you give in openfoamwiki contrib I get redirected and it does not fulfill my wishes, therefore I would like to report the bug here :-)

anyway, the pow() function has a problem with the parser, the second variable can not be set with a alphanumerical variable but has to be a value.

That was "by design". The original parser (then only called funkySetFields) was constructed from the table of functions in the ProgrammersGuide and that said that the exponent can not be field. The Programmers Guide has not been updated for a long time and I don't want to go back through the sources to check if that restriction was really there at the time.

I'll make a note and hope that I don't forget (BTW: Sourceforge seems to have fixed their server-problem so maybe you can log in now)

PS: there is a part in me that says "don't do it. This is madness. Remember all the times people 'reported bugs' but in reality where using the pow-function wrong. And this will become 'easier' if the exponent is calculated too" but I guess the part that wants a 'complete' grammar is going to be stronger

Thank you. Hope that the source forge works again, but I had problems before.

Hi
First , thanks for groocyBC,
I would like to use groovyBCDirection
Is there any example how to use them ?

Thanks

There is an example case Examples/tests/directionMixedBendPipe. Apart from that there are no examples.

The boundary condition is based on the directionMixed-BC (the same way groovyBC is based on the mixed-BC) so look for documentation on this

New release of 0.2.1 of swak4Foam
 

Hi!

A new release of swak4Foam has hit the repositories (mercurial and SVN)

For general information on this swak-thing see http://openfoamwiki.net/index.php/Contrib/swak4Foam

The most notable changes in this version are:

• Possibility to add functions to the grammars via plugins that are loaded at run-time. This allows adding very special functionality without modifying the core of swak. Examples are:
  • Functions that return the mesh-quality fields (non-orthogonality etc)
  • Functions that allow using differential operators but allow chosing the discretization scheme in the function call
  • Functions that "project" sampledSets and sampledSurfaces into the volume
  • Getting the source terms of a lagrangian cloud that is derived from one of the basic intermediate classes
• Delayed variables: these can be used like regular variables for writing. If they are read (used) they give back the value they had X seconds ago. This allows boundary conditions with delayed reactions
• Function objects that create and evolve clouds of lagrangian particles
• Improved the scripts for compilation. Checking of versions of prerequisites.
• On OpenFOAM 2.x there are now source terms derived from basicSource. So all OpenFOAM-solvers that support runtime-selectable source terms via that class can now use sources that use swak-expressions
• Changes to the grammar to support more operations (especially more tensor operations, but also does the pow-function now support non-uniform exponents)

A full list of the changes (and bugs fixed) in this version can be found in the History section of http://openfoamwiki.net/images/e/e6/...oam_README.pdf (for OpenFOAM 1.x) or http://openfoamwiki.net/images/8/81/...README_2.x.pdf (for OpenFOAM 2.x - contents are slightly different). The list is seven pages long in the PDF so there is more than the above highlights.

Have fun and don't hurt yourself

Bernhard

New Release of 0.2.2 of swak4Foam
 

New version of swak4Foam (0.2.2) is available. As usually it contains bug-fixes (yeah. There were bugs. Unbelievable, but true) and several enhancements. For details see the README that comes with it.

For those who don't want to download it without knowing what is new here is an incomplete list of new features:

- several enhancements to the python-integration. Amongst other things: variables that have fields are now availoable to the python-code as numpy.-arrays
- expressions now can access fields from a different case. This makes it possible to easily use funkySetFields to compare cases
- Patch-expressions now support mappedPatches
- Plugin functions to access fields from the radiation mode or the combustion model (for instance the reaction rates)
- Enhancements to the build-scripts
- The start of a reference-manual (Only the first part. In org-mode. Not proofread)

For information on downloading see:
http://openfoamwiki.net/index.php/Contrib/swak4Foam

Unfortunately the SVN is still not available (everybody say "Thank you, ANSYS") but there are tarballs to download at the above link

Two things about this new release:
- As I found out afterwards the release doesn't compile on 1.6-ext. This is fixed in the development version (mercurial repository on sourceforge. See Wiki-page for details)
- One week after the release OF 2.2 came out. The current release of swak4Foam does not compile with it. But I already fixed it and pushed the fixes to the development repository into a special branch (see https://twitter.com/swakPyFoam/statu...94954107031552). The fixes seem to work OK but I plan to test them and then merge them into the regular 2.x-branch. Updates on this work and info which branch is the "better" one for 2.2 will be posted on the above Twitter-channel

Release 0.2.3
 

The last release 0.2.2 will go down in history as one of the shortest-lived swak-releases: as Version 2.2 of OF came out the week after that and as this release makes it necessary to do some changes on the swak-sources this release was rushed out to allow people to go on swaking on 2.2.

In addition the compilation on 2.0 and 1.6-ext were fixed (1.7 and 2.1 were working all the time)

Apart from the support for various OF-versions there are not many additional features:
- support for eigenValues and eigenVectors of tensors
- accessing a full row of a tensor with .x, .y and .z
- writing out sampled surfaces and sets that were specified by swak-functionObjects

For details have al look at the README

Bernhard,

Thanks a lot for all the work making swak available for the new release so soon. I was wondering on one particular new feature.

In the readme file it is mentioned that it is now possible to access the internalField from different cases. Does this mean it is limited to the internalField, or is it also possible to access patches from a different case to use in a groovyBC?

I am wondering if it is possible to use the fields at an outlet as an inlet condition for a different case using this framework?

Regards,
Tom

Interesting idea, but: No. Only complete fields I'm afraid

I developed this because I wanted to compare results from different cases (or different time in the same case). It basically is the functionality of the mapFields-utility.

For different cases the mappedPatch-functionality won't help you either I'm afraid. If there is nothing in OF that supports this it is rather unlikely that such a feature will turn up in swak (will have a look at it when I or someone I work for needs it)

Hi Bernhard,

Thank you for the clarification. That will save me some trial and error. Maybe I can find another solution.

Regards,
Tom

I haven't tried that (but would be interested if you succeed): use the sample-utility to write out a sampledSurface type sampledPatch and read that with timeVaryingMappedFixedValue (you'll have to find out which writer-format is working)

Thanks for the suggestion, did not yet consider that option, but it sounds promising. I'll keep you updated once I find a solution.

reaction rate
 

Dear Bernahrd,

Is there any function in the new version of swak4foam to evoke reaction rate (combustion->R(Yi) in YEqn.H) of species? I need to write reaction rate of some soecies in a file.

Thank you in advance for any hint

Hossein

Yes. That is what I was saying with that line.

Just add "libswakChemistryModelFunctionPlugin.so" to the libs-entry in controlDict. Now if you use any swak-thing (function-object or also funkySetFields) that uses the Field-parser for the first time you see an output like this:
which is a list of the additional functions including their signatures. The one you're interested in is psiChem_RR which takes a species-Name as an argument.

All this uses the results from the last run-through of the chemistry-solver. In funkySetFields you've got to trigger this calculation with the psiChem_updateChemistry-function (only use it once):
"psiChem_updateChemistry(0.001)+psiChem_RR(CO)+psi Chem_RR(CO2)"
for Reaction rates of CO and CO2.

Of course you can use it to check whether the time-step influences the reaction rates
"(psiChem_updateChemistry(0.001)+psiChem_RR(CO ))-(psiChem_updateChemistry(0.002)+psiChem_RR(CO))"
but that of course calculates everything twice

reaction rate(2)
 

Thanks Bernhard. It's very useful but when I try implementing the following sentences in my solver, I encounter the following error that states ‘psiChem_RR’ was not declared in this scope.

if (Y[i].name() == "O2")
{
volScalarField wO2 = psiChem_RR(Y[i].name());
wO2.write();
}


I inserted the #include "swakPsiChemistryModelPluginFunction.H" header in my .C file and the address of it in "option" which is in following:

EXE_INC = \
.
.
.
-I$(HOME)/Swak4Foam_2.x_release_0.2.3/Libraries/swak4FoamParsers/lnInclude/ \
-I$(HOME)/Swak4Foam_2.x_release_0.2.3/Libraries/swakSourceFields/lnInclude/ \
-I$(HOME)/Swak4Foam_2.x_release_0.2.3/Libraries/functionPlugins/swakChemistryModelFunctionPlugin/lnInclude

EXE_LIBS = \
.
.
.
-lsampling \
-lmeshTools \
-L$(FOAM_USER_LIBBIN) \
-lswak4FoamParsers \
-lswakChemistryModelFunctionPlugin \
-lswakSourceFields

How can I solve above problem?

Best Regards,

Hossein

Plugin-functions are designed to be used in parsers. It could be possible to use them on their own, but the calling would be plain weird. The two best options you have are
- create an instance of FieldValueExpressionParser. Feed it the expression you want to get and get the result from that. I don't explain this in detail because for the purpose of what you want to do this would be like using dynamite to go after the cockroaches in your drain: lot of fun, but the neighbours might think it is a little bit excessive.
- have a look at the source of the plugin and try to lift the parts you need (If I remember it correctly: get from the registry the object-registry the object with the name "thermophysicalProperties". Cast it down to the combustion-thermo and then call the RR there). That might be the easiest (for your purposes)

reaction rate(3)
 

Thanks for quick reply. I looked up the "class swakPsiChemistryModelPluginFunction_RR" in swakPsiChemistryModelPluginFunction.C and found my question.

Thanks again

I was hoping you might find the answer there ;)

Release 0.2.4
 

A new version of swak4Foam has been released. It is the version found on the USB-stick of the 8th OpenFOAM workshop. New features include (but for all see the README which has a complete list):

- functionObjects now store their data in the sub-directory postProcessing according to the new convention OpenFOAM has (sorry: this may break existing scripts relying on the old location)
- slight change in the format written may break the restart of cases where swak-specific restart data was written with older swak-version (should not hit many people)
- Example cases are adapted to run with OF 2.2. This may prohibit them from running with 2.1 or 2.0

But now the good parts:

- build scripts have been adapted to run on more machines, allow tuning of the compilation process, compile on non-gcc compilers and give more information for finding problems
- A new library with additional searchableSurfaces (to be used in snappyHexMesh) that allow boolean operations and coordinate transformations of searchable surfaces
- Function objects to calculate the distributions of variables (and expressions)
- specification of aliases for field names that contain special characters (like thermo:rho)
- functionObject to list all registered objects (basically for finding out what is there)
- functionObjects to handle global variables
- functionObject that executes a list of other functionObjects only if the run is parallel/serial
- A new type of variable (stacked) that allows recording multiple values
- More flexible output control for functionObjects based on simpleFunctionObjects
- swakDataEntry is improved
- other stuff and bug fixes (check the README)

Also to make life easier for the non-Emacs people who have problems converting the (still incomplete ) reference guide to a more readable form it can be downloaded here: http://openfoamwiki.net/images/d/db/...mReference.pdf

OpenFOAM 2.2.1 + Swak4Foam 0.24 ?
 

Hi Bernhard,

Will swak4Foam 0.2.4 install correctly on the new OpenFOAM 2.2.1 ?

Also, when will swak4Foam 0.2.4 be posted to the openfoamwiki.net site?

Thank you very much.

I think so. But the only test is trying to compile it
I like to think that the SVN-download at http://openfoamwiki.net/index.php/Co...am#Downloading is sufficient. Providing the tarballs was only a workaround while the SVN was down but I don't see the point creating/testing/uploading tarballs if there is already a way to get the sources

New Release of 0.3.0 of swak4Foam
 

8 months ago the last release of swak4Foam happened. A week later Edward Snowden blew the whistle on the NSA. Coincidence? Unlikely! (who is back there saying "Correlation does not imply causality"? Obviously someone who doesn't appreciate a good conspiracy theory). Today the next version of swak4Foam is announced and the world is expecting the next Spy-Agency to be exposed. The only question is: will it be China, Russia or Austria.

Most of the features have already leaked to the public through the public development repository and the twitter-acount @swakPyFoam but for all those whose intelligence-branch failed to interpret these informations correctly here is an incomplete list of the new features

• the version of the release has jumped to 0.3.0 as this release has now a parser for lagrangian particles (these were the last major OF "things" that swak4Foam couldn't access. It is now possible to let swak4Foam do calculations on clouds of particles
• the new release now supports Foam-3 (this is in the 2.x branch). The 1.x branch supports OF 1.7 and 1.6-ext, all newer Versions are supported in the 2.x-branch
• a number of new accumulations have been added (median, quantile and weighted counterparts for most of them)
• Python-integration has been improved (IPython inclusion, stabler etc)
• new utilities: fieldReport to quickly get information about fields, funkyPythonPostproc allows loading OF-data into a Python-workspace and analyze it with the NumPy/SciPy/Matplotlib-ecosystem of scientific software, funkySetLagrangianFields allows setting values on the particles of a lagrangian cloud
• a number of new function objects (calculation of potential flow, recalculation of thermophyiscs etc). The one with the longest name is swakExpressionAverageDistribution (applications for this are for instance the calculation of the average pressure as a function of the channel length)
• several other enhancements to utilities and function objects (all found in the README)
• a number of bugs was mercilessly squashed. Among them a problem with the correct summation of expressions on faceSets/faceZones in parallel runs
• attempts to improve compilation

A few highlights of the release are demonstrated here http://openfoamwiki.net/staticPages/...K_again.html#/

For further information see
- the Wiki page http://openfoamwiki.net/index.php/Contrib/swak4Foam
- the README that comes with the sources
- the Documentation-folder of the sources has an incomplete reference guide. It documents all the expression-grammars

Have fun and look out for the men in the trenchcoats

New Release of 0.3.1 of swak4Foam
 

This release has been sitting on my hard-disk for several weeks but now I've gotten around to releasing it officially (it is technically identical to the version on the USB-sticks distributed at the Workshop and also the version that is pulled when you compile it as one of the ThirdParty-packages with foam-extend-3.1)

This Release supports the latest versions of OpenFOAM (2.3) and Foam-Extend (3.1). It has its fair share of bug-fixes and also some new features:

• Compiling a "private" bison-version. The reason for this is that a growing number of distributions install bison 3.0 which is incompatible with swak4Foam. On such machines a script downloads a compatible version of bison (2.7), compiles it and installs it locally so that it will be only used for the compilation of swak4Foam
• New utility funkyWarpMesh allows deforming the mesh according to an expression
• A number of function-objects that trigger the writing of time-steps according to conditions
• Function object writeOldTimesOnSignal that writes the current state of the simulation when a floating-point-exception etc happens
• Additional library with fvOptions for OF-versions that support it
• Macro expansion in expressions

For a full list see the section about 0.3.1 in the README

Dear Bernhard,

How can I store the reaction rate of species O2 in each time step using psiChem_RR? Which expressions should be added to controlDict file after adding "libswakChemistryModelFunctionPlugin.so" in the libs or do something else?
I'm using swak4Foam_2.x_release_0.2.3 now.

Thank you in advance,

Hossein

Not sure what you mean with "store at each timestep" but maybe you may find the slides 125ff of http://openfoamwiki.net/images/7/7a/...icTraining.pdf helpful

Thanks.
That is what I need. Sorry for another question. For example If I set the deltaT value in controlDict to 1e-07, Does psichem_RR function consider that value to calculate the reaction rate?

Thanks

It reuses the RR that was calculated by the chemistry and used to solve the species transport equations. If you want to check the influence of the timestep there is a separate function for it (it is explained in slides 138ff. Please get into the habit of checking more than just a few pages before asking again. You're one "under-researched" question away from being permanently ignored)

Dear foamers,

I am trying to install swak4Foam on OF2.3, and I already tried the OF2.2x version suggested by Bruno but still there are some errors. I am wondering if there is a package developed for OF2.3. I am looking forward to your answers,

Thanks in advance and Best,
Mahdi

The current 2.x-release should compile with 2.3 (at least it did in June when the release was done). Please provide a link to the release you mean (Bruno is very active so looking for something he might have said takes a long time)

Hi Bernhard,

Thanks a lot for your quick response as always. I tried the 2.x-release downloaded from:

svn checkout svn://svn.code.sf.net/p/openfoam-extend/svn/trunk/Breeder_2.0/libraries/swak4Foam/ swak4Foam_2.x

I attached the make log file output from this compile. Then I also tried this one which was suggested by Bruno (https://github.com/wyldckat/swak4foam/tree/OF22X), but I am not sure if this principally should work for me or not; anyway it did not :)

Thanks and Best,
Mahdi