[cnsidero]

The next major version of Caelus, 6.04, was released today. We’ll continue to stick to bi-annual releases with 6.10 coming in October. Much of the effort this release was spent overhauling interpolation schemes. Here’s a brief list:

• Restructuring of interpolation schemes (central and upwind) to use consistent limiting
• Overhaul of limiters to ensure TVD compliance
• Updates to pressure and density-based compressible solvers
• New combustion solvers
• New gammaReTheta transition, kOmegaSSTDES and DDES and SA-VLES turbulence models

The other news is there is now a publicly available source repository to enable better community collaboration:

https://bitbucket.org/appliedccm/caelus-contributors

There is a release notes PDF available on the download page with more descriptions. http://www.caelus-cml.com/download/

If you’re not already familiar with Caelus, you can read more about it on the website. Its free and open-source. The file formats are compatible with OpenFOAM but there have been many improvements to it such that it is now moving independently in its own direction.

We are happy to answer questions, technical or otherwise, by email: caelus@appliedccm.com or also follow Caelus on twitter @caelus_cml

Applied CCM is currently the maintainer and main developer of Caelus. If you’re interested in porting or developing your own application to Caelus, we’ll be happy to provide guidance. If you’ve developed a solver or model that you would like to include in Caelus check out the contributors repository. Any and all external contributors copyright will be kept in their contributed source files.

[phsieh2005]

Hi,

Will the linux binary version work on openSuSe 13.2?

Pei-ying

[phsieh2005]

I installed Caelus-6.04 on OpenSUSE 13.2. The serial solver ran, but, got errors when tried to run in parallel.

I am wondering how I can get caelus to use system openmpi. It looks like caelus tried to look for openmpi-1.6.5 under /opt/caelus/caelus-4.10.

Recommendations will be appreciated.

Pei-Ying

[cnsidero]

Pei-Ying,

We don't build/test on SUSE but I have helped someone previously with issues running Caelus on SUSE.

To your issue, it looks like you left the default OpenMPI (the one Caelus provides) when you installed Caelus rather than configure it to use the system MPI. Caelus is complaining because it's looking to the provided MPI libraries and you're using the system MPI executables. To fix it, follow these steps.

In "lib/python2.6/Caelus/conf.py" change to the following:

MPI_BIN = '/usr/lib64/mpi/gcc/openmpi/bin' 
MPI_LIB = '/usr/lib64/mpi/gcc/openmpi/lib64' 
MPI_INC = '/usr/lib64/mpi/gcc/openmpi/include'

in the "etc/caelus.bashrc":

export MPI_INC=/usr/lib64/mpi/gcc/openmpi/include 
export MPI_LIB=/usr/lib64/mpi/gcc/openmpi/lib64 
export MPI_LIB_NAME=mpi

and create a dynamic link like so:

sudo ln -s /usr/lib64/mpi/gcc/openmpi/lib64/libmpi.so /usr/lib64/mpi/gcc/openmpi/lib64/libmpi.so.1

Finally to get the Python wrapping to work in parallel modify the "etc/controlDict" as follows (note the bold terms):

parRunTemplate "mpiexec -n %(NPROCS)d %(PAROPTS)s  --mca btl sm,self  %(APPLICATION)s  %(ARGS)s -parallel";

Let me know if the fixes things.

Perhaps we'll look into including a SUSE distribution as a future supported platform.

-Chris

[phsieh2005]

Thanks a lot Chris!

Everything works perfectly on OpenSUSE 13.2 now.

Now, I just need to figure out how to compile a new solver using scon.

Pei-Ying

[cnsidero]

Pei-Ying,

Glad to hear its working. I replied to your email about using scons.

-Chris

[anirbanmondal93]

hi guys,

I am running the case vofLPTSolver in CAELUS-6.04. It is basically a case relating to simulation of a number of lagrangian bubble particles (lagrangian particle tracking) in continuous water phase (VOF) using LES turbulence model. But i want to simulate the same test case using DNS turbulence modelling inspite of LES approach.
What modifications i have to do in the existing '...tutorials/vofLPTSolver' directory or in any other directories in CAELUS?

Thanks in advance.

regards,
Anirban

[cnsidero]

Anirban,

No modifications to the solver are necessary. To run DNS means no explicit turbulence modelling (LES, RAS, etc). Therefore to disable an explicit LES model, in the constant/LESProperties dictionary change LESModel from Smagorinsky to laminar.

-Chris

[anirbanmondal93]

Hi Chris,

Greetings
I am running tutorials/vofLPTSolver in Caelus-6.04 changing different values and variables in some files within this directory to simulate lagrangian bubble particles rising in water with VOF free surface capturing using DNS only. But, i am confused whether the injection particles are air bubbles or water after seeing the value of alpha1 in animation.

I am mentioning two cases with the details of the changing stuff i have done.

Case-1:
$Caelus/../ACCM_bubbleCol3D/Allrun.py- line 69: omitting '-parallel': run = subprocess.Popen(['caelus.py', '-l', 'vofLPTSolver'], shell=pltfrm)
$Caelus/../ACCM_bubbleCol3D/system/controlDict- line 22: endTime 10;

Case-2:
same changes as in case-1.
additional changes-

$Caelus/../ACCM_bubbleCol3D/system/controlDict-
line 22: endTime 10;
line 28: writeInterval 1;

$Caelus/../ACCM_bubbleCol3D/bubbleCloudproperties-
line 153: changing threshold value from 0.5 to 0: threshold 0;

$Caelus/../ACCM_bubbleCol3D/LESProperties-
line 15: replacing Smagorinsky with laminar to switch on DNS as per your suggestion: LESModel laminar;


Attaching two case animation separately what i have made.

Case-1: https://www.youtube.com/watch?v=o1Uh...ature=youtu.be

Case-2: https://www.youtube.com/watch?v=WJ1r...ature=youtu.be

[vidyadhar]

Hello Chris,

I want to work with vofLPTSolver to track solid particles in liquid flow.

I have certain basic questions:
1.whether this solver can be run in parallel.
2. Does this solver treats fluid phase as continuum and particles as individual lagrangian identities?
3. Also, does this solver handles solid particles?

Request you to kindly let me know.


Thanks & Regards,
Vidyadhar

Originally Posted by cnsidero

Anirban,

No modifications to the solver are necessary. To run DNS means no explicit turbulence modelling (LES, RAS, etc). Therefore to disable an explicit LES model, in the constant/LESProperties dictionary change LESModel from Smagorinsky to laminar.

-Chris

[cnsidero]

Vidyadhar,

1. Yes, the solver is fully parallel just like all the other solvers.

2. Yes, the fluid is a continuum and the particles are treated as discrete Lagragian objects.

3. The particles behavior is defined by their density and size. With the current implementation the particles do not change size. Therefore, yes, you could specific particle properties for a solid material.

Regards, Chris

[vidyadhar]

Hello Chris,

Thanks for the reply.

Actually, I want to simulate lagrangian particle (bubbles as well as solids) flows in a liquid.

Can you clarify if DPMFoam solver or MPPICFoam solver of openfoam 4.0 can't do this job. Do they solve only solid particle tracking? Because I want to simulate bubble tracking also!

If I have to use caelus, can you help me how to install it on ubuntu 16.04.
Or Is caelus can be installed on ubuntu 14.04 only?

Thanks

[vidyadhar]

Hi,

I just installed caelus6.04.

How to execute commands in terminal to use this software.

Also, how to execute a solver in parallel. Which command is to be used.

Thanks.

[vidyadhar]

Hi Chris,

I have installed caelus 6.04. I tried using vofLPTSolver with the following command, but I got the errors messages as shown below; Can you please help me in using this solver in parallel!!! Thank you.




caelus.py vofLPTSolver -parallel
-------------------------------------------------------
Primary job terminated normally, but 1 process returned
a non-zero exit code.. Per user-direction, the job has been aborted.
-------------------------------------------------------
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL: your MPI job will now abort
[iitmech5:17307] Local abort before MPI_INIT completed successfully; not able to aggregate error messages, and not able to guarantee that all other processes were killed!
--------------------------------------------------------------------------
Sorry! You were supposed to get help about:
opal_init:startup:internal-failure
But I couldn't open the help file:
/opt/caelus/caelus-4.10/external/linux/openmpi-1.6.5/share/openmpi/help-opal-runtime.txt: No such file or directory. Sorry!
--------------------------------------------------------------------------
[iitmech5:17307] [[INVALID],INVALID] ORTE_ERROR_LOG: Error in file runtime/orte_init.c at line 79
--------------------------------------------------------------------------
Sorry! You were supposed to get help about:
mpi_init:startup:internal-failure
But I couldn't open the help file:
/opt/caelus/caelus-4.10/external/linux/openmpi-1.6.5/share/openmpi/help-mpi-runtime: No such file or directory. Sorry!

[cnsidero]

Vidyadhar,

Let me try to catch up to your questions. Let's diagnose the runtime issues first.

It appears you are using Ubuntu Linux 16.04. Please confirm this.

Which Caelus package did you install? The binary installer or the source-only installer? If it's the later, you will need to compile it using the provided BuildCaelus.py script in the top-level directory.

If you're confident the binaries are there, can you try running the vofLPTSolver tutorial with the AllRun.py script?

Let's start there and I'll answer the earlier questions after.

-Chris

[vidyadhar]

Hello Chris,

Yes I am using ubuntu 16.04.

I have installed the binary installer(Caelus 6.04 source code and binary installer).

I have opened AllRun.py file in ACCM_bubbleCol3D of vofLPTSolver casefile and followed the operations therein: Executing blockMesh, setSet, createPatch, setFields, decomposePar force, vofLPTSolver--- Finally vofLPTSolver.log file has been created in the ACCM_bubbleCol3D directory which has the following messages:




Sorry! You were supposed to get help about:
opal_init:startup:internal-failure
But I couldn't open the help file:
/opt/caelus/caelus-4.10/external/linux/openmpi-1.6.5/share/openmpi/help-opal-runtime.txt: No such file or directory. Sorry!
--------------------------------------------------------------------------
[iitmech5:09519] [[INVALID],INVALID] ORTE_ERROR_LOG: Error in file runtime/orte_init.c at line 79
--------------------------------------------------------------------------
Sorry! You were supposed to get help about:
mpi_init:startup:internal-failure
But I couldn't open the help file:
/opt/caelus/caelus-4.10/external/linux/openmpi-1.6.5/share/openmpi/help-mpi-runtime: No such file or directory. Sorry!
--------------------------------------------------------------------------
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL: your MPI job will now abort
[iitmech5:9519] Local abort before MPI_INIT completed successfully; not able to aggregate error messages, and not able to guarantee that all other processes were killed!
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL: your MPI job will now abort
[iitmech5:9520] Local abort before MPI_INIT completed successfully; not able to aggregate error messages, and not able to guarantee that all other processes were killed!
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL: your MPI job will now abort
[iitmech5:9521] Local abort before MPI_INIT completed successfully; not able to aggregate error messages, and not able to guarantee that all other processes were killed!
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL: your MPI job will now abort
[iitmech5:9522] Local abort before MPI_INIT completed successfully; not able to aggregate error messages, and not able to guarantee that all other processes were killed!
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL: your MPI job will now abort
[iitmech5:9523] Local abort before MPI_INIT completed successfully; not able to aggregate error messages, and not able to guarantee that all other processes were killed!
-------------------------------------------------------
Primary job terminated normally, but 1 process returned
a non-zero exit code.. Per user-direction, the job has been aborted.
-------------------------------------------------------
--------------------------------------------------------------------------
Sorry! You were supposed to get help about:
opal_init:startup:internal-failure
But I couldn't open the help file:
/opt/caelus/caelus-4.10/external/linux/openmpi-1.6.5/share/openmpi/help-opal-runtime.txt: No such file or directory. Sorry!
--------------------------------------------------------------------------
[iitmech5:09526] [[INVALID],INVALID] ORTE_ERROR_LOG: Error in file runtime/orte_init.c at line 79
--------------------------------------------------------------------------
Sorry! You were supposed to get help about:
mpi_init:startup:internal-failure
But I couldn't open the help file:
/opt/caelus/caelus-4.10/external/linux/openmpi-1.6.5/share/openmpi/help-mpi-runtime: No such file or directory. Sorry!
--------------------------------------------------------------------------
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL: your MPI job will now abort
[iitmech5:9526] Local abort before MPI_INIT completed successfully; not able to aggregate error messages, and not able to guarantee that all other processes were killed!
--------------------------------------------------------------------------
mpiexec detected that one or more processes exited with non-zero status, thus causing
the job to be terminated. The first process to do so was:

Process name: [[43169,1],1]
Exit code: 1
--------------------------------------------------------------------------


Thank you for the patient reading.........!!!!!

[cnsidero]

Are the other steps running OK? What do the other log files say? Need to find out if it's one of the other steps failing.

[vidyadhar]

Hello Chris,

I have used the vofLPT solver successfully for the given tutorial case-ACCM_bubblecolumn3D, where the fluid is at rest and bubbles are moving.

But, I want to simulate when the fluid is flowing through the column. (I have modified the column to be full of water.)

The changes made by me are:
In blockMeshDict: bottom is made to patch from wall, as I would like to flow in water through bottom patch.

In U: bottom is made to fixedValue uniform (0 0.08 0), top as zeroGradient.

In p_rgh: top is made to fixedValue uniform 0.

I tried with various boundary conditions for U and p_rgh, but I could run only till time=3 s. After that Courant number is increasing largely, volume fraction Alpha1 is also becoming very large and I am getting errors as below:

Courant Number mean: 2.387736012e+97 max: 1.998894153e+101
Interface Courant Number mean: 2.488915823e+96 max: 2.311900924e+100
Time = 3.81

Evolving bubbles


Solving 3-D cloud bubbleCloud

--> Cloud: bubbleCloud injector: model1
Added 2 new parcels

Cloud: bubbleCloud
Current number of parcels = 2907
Current mass in system = 0.0001668853125
Linear momentum = (-nan -nan -nan)
|Linear momentum| = -nan
Linear kinetic energy = -nan
Rotational kinetic energy = 0
model1:
number of parcels added = 9119
mass introduced = 0.0005142614062
Parcels absorbed into film = 0
New film detached parcels = 0
Parcel fate (number, mass)
- escape = 147, 0
- stick = 0, 0

MULES: Solving for alpha1
Phase-1 volume fraction = 1.774412665e+87 Min(alpha1) = -1.333103629e+102 Max(alpha1) = 2.357171278e+102
MULES: Solving for alpha1
Phase-1 volume fraction = -4.252960616e+259 Min(alpha1) = -5.124760501e+280 Max(alpha1) = 5.031106494e+280
PIMPLE: iteration 1
DILUPBiCG: Solving for Ux: solution singularity
DILUPBiCG: Solving for Uy: solution singularity
DILUPBiCG: Solving for Uz: solution singularity
GAMG: Solving for p_rgh, Initial residual = nan, Final residual = nan, No Iterations 1000
GAMG: Solving for p_rgh, Initial residual = nan, Final residual = nan, No Iterations 1000
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 4 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------
[2]
[2]
[2] --> CAELUS FATAL IO ERROR:
[2] wrong token type - expected Scalar, found on line 0 the word 'nan'
[2]
[2] file: /home/mech5/vofLPTSolver/vofLPTSolvertestedfor5s/ACCM_20processors-COMPLETEWATER_BC_MODIFIED/processor2/system/data::solverPerformance:_rgh at line 0.
[2]
[2] From function operator>>(Istream&, Scalar&)
[2] in file core/primitives/Scalar/doubleScalar/doubleScalar.cpp at line 98.
[2]
CAELUS parallel run exiting



May I know the correct boundary conditions for the flow of fluid along with bubbles.

Also, can you please brief me about p_rgh & how it is different from p.


Thanks & Regards,
vidyadhar

[vidyadhar]

Originally Posted by cnsidero

Vidyadhar,

1. Yes, the solver is fully parallel just like all the other solvers.

2. Yes, the fluid is a continuum and the particles are treated as discrete Lagragian objects.

3. The particles behavior is defined by their density and size. With the current implementation the particles do not change size. Therefore, yes, you could specific particle properties for a solid material.

Regards, Chris

Hello Chris,

1. I tried using vofLPT solver for bubbles. It was working fine.
But, I have doubt regarding Youngs Modulus and Poissons ratio for bubbles in the bubble cloud properties file. Are they meant for bubbles only? Please clarify.

2.Now I would like to use the same for simulating solid particles. I have changed the density. But, I am facing the problem with viscosity of phase2 in transport properties file under constant directory.

Can you please let me know how to go about this.

Thank you.

[vidyadhar]

Hello chris,
I am using vofLPT solver of caelus 6.04.

In bubbleCloudProperties file:
I would like to understand the meaning of the terms such as parcel and particle. Are they same or different?
Also, how to calculate no.of parcels per second? Does this value depend on the run time of the simulation?


injectionModels
{
model1
{
type patchInjection;
patchName inlet; // Name of patch
SOI 0; // Start Of Injection
flowRateProfile constant 1; // Flow rate profile relative to SOI
massTotal 0.02025; // Total mass injected over injection duration
parcelBasisType mass; // How are the number of particles calculated
duration 7; // Duration of injection. NOTE: set to 1 for steady state
U0 (0 0.08 0); // Initial parcel velocity
parcelsPerSecond 3357; // Number of parcels to introduce per second

sizeDistribution
{
type fixedValue;
fixedValueDistribution
{
value 4e-03;
}
}
}
}

Can you please help me in understanding this. Thank you