code-mphi / ecogen

A CFD open source code dedicated to multiphase compressible flows

ECOGEN

ECOGEN is a CFD plateform written in C++ object oriented programming langage. It is dedicated to numerical simulation of compressible multiphase flows. It has the vocation to share academics researches in the multiphase flow field in direction to ohter academics but also for industrials, students, etc.

• Official webSite: https://code-mphi.github.io/ECOGEN/
• Official documentation: https://code-mphi.github.io/ECOGEN/docs/sphinx_docs/index.html
• API documentation: https://code-mphi.github.io/ECOGEN/docs/doxygen_docs/index.html

What can I do with ?

With ECOGEN you can solve:

• Interface problems between pure or multicomponents fluids and mixtures of multiphase flows.
• Flow problems involving surface tension, heat and mass transfers for evaporating and condensing flow, cavitation.
• Wave propagation problems in strongly unsteady situations using a specific Adaptative Mesh Reffinement.
• complex flow problems on unstructured grids.

Visit the gallery for examples: https://code-mphi.github.io/ECOGEN/gallery/

Start with ECOGEN

The better way to start with ECOGEN is to follow installation instructions on the official documentation webSite: https://code-mphi.github.io/ECOGEN/docs/sphinx_docs/Chap2_installation_Chapter.html

Then you can follow the dedicated tutorial to run your first test: https://code-mphi.github.io/ECOGEN/docs/sphinx_docs/Chap4_1startWithECOGEN.html

License

ECOGEN is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. ECOGEN is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details http://www.gnu.org/licenses.

Citation

To cite ECOGEN, please use :

Schmidmayer, K., Petitpas, F., Le Martelot, S. & Daniel, E. (2020). ECOGEN, an open-source tool dedicated to multiphase compressible multiphysics flows. Computer Physics Communications.

@article{schmidmayer2020ecogen,
  doi = {10.1016/j.cpc.2019.107093},
  url = {https://doi.org/10.1016/j.cpc.2019.107093},
  title={{ECOGEN}: {A}n open-source tool for multiphase, compressible, multiphysics flows},
  author={Schmidmayer, Kevin and Petitpas, Fabien and Le Martelot, S{\'e}bastien and Daniel, {\'E}ric},
  journal={Computer Physics Communications},
  volume={251},
  pages={107093},
  year={2020}
}

ecogen's Issues

Can I give a injection boundary condition at specific part of boundary line or surface?

Hello,

Is there possible to state a injection boundary condition at specific part of boundary line or surface?

Muticomponent or n phase simulation

Hi,

I have been trying to explore the capabilities of ECOGEN and wanted to use a mutlicomponent gas and/or 3 phases. I have not found an example that addresses this topic. I was wondering how this will work with the multiphase VOF solver because that range is 0 to 1.

If numPhases = 3 (2 gas and 1 liquid) how does one go about defining alpha or is there another tracker ?

Thanks.

Restarting runs for an Unstructured Grid not working

I tried to restart a simulation on an Unstructured Grid and it doesnot run and exits before the first time-step. Can we restart simulations on unstructured grids for Ecogen runs?

Output data only for a part of the domain

Dear all,

I am searching a way to extract the data of only a part of the computational domain (being large) and post-process. It seems that the "cuts" markup is for the same, but I am not able to understand how to implement it from the existing code. Could someone please help? Even extracting a plane section is okay.

Also, an unrelated question: Can ECOGEN be used to simulate bubble chains?

Regards,
Niranjan

Initialisation of pressure in liquid domain

Dear all,

Greetings! I am facing an issue with the initial condition while simulating a bubble collapse problem. I understand that the default non-spherical bubble collapse test case in ECOGEN happens to be the same test case as given in this article. The initial pressure field is set according to a particular relation where pressure varies inversely with distance from the bubble centre outside the bubble. It has been mentioned in the article that this has been done being a realistic model, as well as avoid spurious pressure waves.

However in the test case, it seems to me that the liquid domain is uniformly initialised with pressure Pinf = 50.6 * 10^5 Pa, hence there is a discontinuity in terms of pressure across the interface. Apparently, this is generating a pressure wave as well, shown in the picture below:

Could someone please confirm whether this is the case? If yes, what changes should I make so as to initialise in accordance with the article?

Thanking you in advance,
Niranjan

Handling output files

Dear all,

Greetings. I use Paraview to visualise the data generated by ECOGEN simulations. While doing even coarse 3D simulations, I am finding it difficult to handle the data as even for 20 timestamps, the total folder size comes out to be 18.7 GB. It is taking a huge amount of time to load each timestep and for finer meshes, the software simply crashes. Is there any way to overcome this issue?

Regards,
Niranjan K. Sabu

How to install ECOGEN on HPC servers

Dear Admin,
I am trying to install ECOGEN on an HPC server. When I tried to install using the instructions given in the documentation file, many issues have appeared and I could not found any solution to the problem yet. Any help on this issue is highly appreciated. Thank you.

How to incorporate mass transfer

Hello ECOGEN team,
I would like to incorporate mass transfer into the ECOGEN since I am interested in the evaporation of the liquid phase. It would be very helpful if you could give some suggestions on how to incorporate the mass transfer into the additional physics term of ECOGEN.
Thank you.

Cartesian mesh

Dear all,

I am an inexperienced college student learning ECOGEN software for a project. For simulating a 3D near-wall bubble collapse, I am trying to mesh a 1mm * 2mm * 2mm cubical domain with a highly refined 0.6mm * 0.6mm * 0.6mm cubical region adjacent to one of the domain's face centres. It involves stretching the mesh along both positive directions. I tried using the meshV5.xml attached with this issue.
meshV5.txt
Unfortunately, I am getting the following error: "Exception on XML file: error on stretching definition." My idea was to stretch the mesh along all 5 directions from the smaller cubical region by a factor of 1.01. Is it possible in ECOGEN? If yes what edits should I make in the above code? Or should it be rewritten completely?

Thanking you in advance

Niranjan K. Sabu

Axisymmetry

Hello everyone,

I am looking towards implementing axisymmetry in a bubble collapse problem to reduce it to a 2D simulation. I saw that there are some test cases already available on the same but everywhere, only the term "symmetry" is being used in the initialconditions.xml file. I am not able to distinguish between axisymmetric cases and planar cases, and in some cases, is confused whether this symmetry refers to mere reflection about a side/face only. I asked about this to a senior friend of mine but even he was helpless. Please help me.

Thanking you in advance,

Niranjan K. Sabu

Add time factor in simulation

Dear All,
Please help me out in following coding stuff. I am working on problem - flow of air bubble inside water.In coding by ECOGEN, I need to add the air bubble after the water/fluid reached steady state flow. Can ECOGEN allows us to introduce that bubble after some time, for example 5 sec.
Kindly suggest me the changes which I need to do.
I am attaching my codes below.

Thanks & Regards.
Aman Tripathi
geo.zip

Ph.D. Scholar,
IISER Bhopal, India

ECOGEN unknown error

The issue that I am encountering is an error output that I cannot back trace -

terminate called after throwing an instance of 'std::length_error'
what(): vector::_M_default_append

The error occurs upon call to 3D simulations, with or without MPI

I expected that there would be a timestep zero but the error is not one I am familiar with, could someone please assist?

Regarding the issue of outputting probes data

Dear all,

I need to output the subtle physical quantity changes on a certain line in the 2D example, so I set up a series of probes and fixed the output time frequency. However, when I was processing the output results, I found that the array length of each point record seemed to be different, which made it impossible for me to transpose the matrix. This phenomenon only occurs in my recent examples, and I speculate that it may be on an approximate order of magnitude with the time interval between the calculation time step I set and the output point?

Can you tell me how to solve the problem of inconsistent array length when outputting results from probes at different positions? Or other methods can directly output data on a custom location line.

Thanking you in advance
HanWang

Unable to Run 'Kapila' case from ECOGEN 2 version

I created a test case in ECOGEN 2 that ran successfully using the 'Kapila' model for 2 phases with an unstructued mesh. I tried to run the same in ECOGEN 3.1 by modifying the modelV4.xml file as follows

<?xml version = "1.0" encoding = "UTF-8" standalone = "yes"?>
<model>
 <flowModel name="PressureVelocityEq " numberPhases="2" alphaNull="false"/>
 <EOS name="IG_air.xml"/>
 <EOS name="SG_water300K.xml"/>
 <additionalPhysic type="viscosity"/>
</model>

I am now getting the following error:

-------------------------------------------------
Exception on XML file: this portion of code is under development
****************************************
 XML file concerned: '././libTests/myRuns/liquidInject/modelV4.xml'
 infos on exception in code source :
  file: 'src/InputOutput/Input.cpp'
  line: 620
--------------------------------------------------

and ECOGEN causes Segmentation faults. Can someone please help me with Kapila model for ECOGEN 3.1?

Availability of test case

Dear all,

I am learning about bubble collapse near walls. I came across a JFM article by Dr. Trummler. Out of the various crevice geometries tested, is any case directly available among ECOGEN examples, either among the official documentation or somewhere else?

Thanking you
Yours sincerely

Niranjan K. Sabu

Cavitation around a body by the use of unstructured mesh

Hello all,

I want to follow the test case cavitation of 2D foil, but I changed the foil by a rectangle and decrease the ambient pressure. Then the time step was increasing gradually. I don't know why it happens, maybe the pressure is not good? Thank you for your help in advance.

Mesh for 2D axisymmetric gemoetry

Dear all,

I am simulating a 2D axisymmetric problem in ECOGEN and using the default cartesian mesh for the same. In solvers like OpenFOAM or ANSYS, the mesh has a wedge shape with one cell thickness, which is visible in the visuals. In ECOGEN though, such a wedge shape geometry is not visible. Instead, I only see a 2D rectangular slice. Could anyone please clarify what is happening?

Regards
Niranjan

Unstructured Grid In parallel

I have one issue occuring when trying to run the test case
"kapila2DnozzlesInjection". I generate a mesh in gmsh 3.0.6 using command
"gmsh -2 nozzle2D.geo -format msh2" - no errors are output from gmsh for this command. However, once I attempt to run in parallel, I get a segmentation fault listed below.

---

      EXECUTION OF THE TEST CASE NUMBER : 1

---

T1 | Test case : ./libTests/referenceTestCases/kapila/2D/nozzles/injection/
T1 | Number of CPU : 10

0. MESH FILE PRETRAITEMENT nozzles/nozzle2D.msh IN PROGRESS...
   1/Mesh nodes reading ...OK
   2/1D/2D/3D elements reading ...OK

---

MESH GENERAL INFORMATIONS :

---

mesh nodes number : 4397
elements number : 5110

3/Attributing nodes to CPUs ...
0% ...
[apollo:31485] *** Process received signal ***
[apollo:31485] Signal: Segmentation fault (11)
[apollo:31485] Signal code: Address not mapped (1)
[apollo:31485] Failing at address: 0xb7cb1c128
[apollo:31485] [ 0] /lib/x86_64-linux-gnu/libc.so.6(+0x354b0)[0x7f56ca5804b0]
[apollo:31485] [ 1] ECOGEN[0x46dbf6]
[apollo:31485] [ 2] ECOGEN[0x46fb45]
[apollo:31485] [ 3] ECOGEN[0x40a92c]
[apollo:31485] [ 4] ECOGEN[0x406732]
[apollo:31485] [ 5] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf0)[0x7f56ca56b830]
[apollo:31485] [ 6] ECOGEN[0x408389]
[apollo:31485] *** End of error message ***

Am I using the correct command format to generate the mesh? I am using the Unstructured grid, and haven't changed anything from the original test case files.

Thank you.

Help Needed for ECOGEN-3.1 Post-Processing Error

Dear Sir,

I am writing to seek your expertise and assistance in resolving an error that I've encountered during the post-processing of the default Computational Fluid Dynamics (CFD) problem of Shock-Bubble Interaction in 2D using ECOGEN-3.1 solver and post-processing in both ParaView as well as VisIT software.

I am working on the CFD problem and have successfully completed the simulation using ECOGEN-3.1, Pressure and Velocity Equilibrium Model. However, when I attempted to analyze and visualize the results of Vorticity and Baroclinic Vorticity, I encountered an error that I have been unable to resolve. This issue is hindering my progress and preventing me from moving forward with my analysis. Moreover I am also interested in knowing about how to modify the existing code so that I can incorporate the equations of Reynolds Stress, Turbulent Heat flux, Time Averaging and Favre Averaging so that I can explore more on these areas too.

Given your extensive experience in CFD and post-processing, I believe you might be able to help me identify the root cause of this error and provide guidance on how to resolve it. Your insights and expertise would be invaluable in helping me overcome this obstacle and ensure the success of my research work.

Thanks and Regards,

Nithin Krishnan S

Need output files for both XML & GNU formats

Hello ECOGEN developers & users,

I am trying to run an unstructured grid in 3D and need the output datafiles of both GNU & XML (Paraview/ Visit) formats. Right now it seems we can only choose only one format for the output datafiles. Running the same case twice for data generation in two different formats for 3D cases will be really computationally expensive for my case. Is there a way we can specify ECOGEN to generate datafiles in both formats? i.e.:

<outputMode format="XML" binary="false" />
and
<outputMode format="GNU" binary="false" />

Need datafiles from both formats simultaneously for the same simulation.

How to install ECOGEN on Windows system？

The setup of axisymmetric flow and possibility of separated domain for ARM

Describe the bug
Dear Developer of ECOGEN,

Thanks for sharing the source code and detailed tutorial files. I am working on a case for a nozzle with perfect gas emission into the water. I am trying to setup an axisymmetric case. May I ask is it enough to set the axis boundary as symmetry one for the axisymmetric case or I should also include a disc-like setup in sphericalCollapse case?

In the tutorial, is the nozzle case in /ECOGEN/tree/devel/libTests/referenceTestCases/PUEq/2D/nozzles/tank/ essentially 2-D or axisymmetric?

I found the ARM is pretty useful in cartesian mesh. Is it possilble to mesh two domains, one for unstructured mesh, and one for ARM of cartesian mesh or I could program a nozzle shape in a cartesian domain?

Thanks for your time and effort.

Best regards

Wei

Getting physics error

I am trying to run a 2D unstructured mesh generated using the recommended GMSH version. I am able to mesh it for both single and multiple core runs using the partitioning in GMSH. The mesh is being imported properly by ECOGEN. However, I am getting the following errors for it when running on a single core:

3/Linking Geometries -> Physics ...

It throws a Signal: Segmentation fault (11) immediately after that.

Multi core runs simply hangup at that stage.

I tried to trace back the error but I am unable to figure out the problem in the Physics. The geometry is a simple rectangular box, with 1 location (section/ part of the bottom wall) for injecting a different fluid into the domain. Please let me know if I can inject fluids into the domain at particular locations (not the entire boundary) in unstructured meshes?

Testing scaling

Dear all,

I have been trying to test strong scaling of ECOGEN in an HPC. The tutorials used were the spherical collapse and non spherical collapse cases. But the results are unexpectedly bad and I am confused. I went through the main journal article of ECOGEN and found some examples used for scaling but I am not sure whether they are available as such. Could anyone please point out if they are there or provide the input files if not there?

Thanking you
Yours sincerely

Niranjan K. Sabu

Something wrong with reading unstructured grid file

Hello all,

I am a new one using ECOGEN to solve two-phase flows, and very grateful to you all for developing this great solver!

I am making my own unstructured grid by the use of GMSH, but here errors occur when the solver is reading .msh file under multiple-cores condition, but it is OK to be read only by single core.

Coupling to other codes with PreCice by TU Munich

Coupling to other codes with PreCice by TU Munich
https://precice.org/

perhaps a possibility for a wide use of this software with other FEM and CFD Codes like Calculix and more.

APECSS is a new software for cavitation and his acoustics.
https://github.com/polycfd/apecss

Acoular is a software to see sources in measurement.
https://www.acoular.org/

so a coupling can perhaps improve results and workflow here.

Structure of initialConditionsVX.xml file for surface tension

Dear All,
Please help me out in following coding stuff. I want to create a code which includes injection and outflow Boundary conditions along with two wall conditions. In this case, I need to use the Equation of state having surface tension (SG_liquidForSurfaceTensionCases.xml). i made an attempt to write such code but it shows following error:

Exception during reading XML file: file not found or incorrect structure

---

XML file concerned: '././libTests/myTests/BubbleinWater/initialConditionsV4.xml'
infos on exception in code source :
file: 'src/InputOutput/Input.cpp'
line: 861

Segmentation fault (core dumped)

Kindly suggest me the changes which I need to do.

Thanks & Regards.
Aman Tripathi
Ph.D. Scholar,
IISER Bhopal, India

modelV4.txt
meshV5.txt
mainV5.txt
initialConditionsV4.txt

Automatic Pressure values at Inlet and Outlet

Dear All,
Please help me out in following coding stuff. I am working on problem - flow of air bubble inside water.In coding by ECOGEN, I need to put the Neumann boundary condition instead of Dirichlet boundary condition, such that while taking the Boundary condition as Injection (inlet) and outflow(outlet), I don't want to give the value of inlet pressure and outlet pressure manually, I want that these pressures will be taken automatically. Please suggest the changes in code.

mixture_model3.zip
`

Thanks & Regards.
Aman Tripathi

Ph.D. Scholar,
IISER Bhopal, India

Slope limiters

Dear all,

For the PUEqn model, which slope limiter is used, especially for the bubble collapse case? I come across multiple names like THINC, van Leer, MC, etc. when going through papers.

Thanking you,

Niranjan K. Sabu

Unstructured mesh cannot be run in parallel

I used OpenMPI 3.1.4 to compile the code. Every works well but when unstructured mesh are distributed to CPU, I got segmentation fault.

For the test case in kapila/2D/testUnstructured/. I can run the simulation in single process but cannot run on multiple processors.

Unphysical flow field for 3D shock-bubble interaction when lvlMax>2

Hi! I'm using ECOGEN-v2.0, and I find the unphysical result still exists for 3D shock-bubble case when using high refinement level.

The case is

ECOGEN/libTests/referenceTestCases/kapila/3D/shockBubble/heliumAir/

if the lvlMax is larger than 2, the flow field will be unphysical after about 40us when the shockwave reaches the upstream pole of the bubble. Attached please find the density flowfield at 42us when lvlMax = 3.

Please fix this problem. Thanks!

Regards,
Fan E

Unstructured 3D test case didn't work

Hello,

The 3D unstructured test case can not work normally. And we saw the case that water flows in a cylinder in the document. Is that mean the release version can only solve single-phase problem in 3D unstructured domain?

Best wishes

New Person in CFD Simulation

Hi Kevin,

I installed the latest and an older version of ECOGEN on WSL/Windows.
Thanks for your hard work. The code works great. Just, I do not know why all examples using mesh (unstructured) does not work on my system. I get typical following error. Pls, if you have time, let me know your opinion.

Thanks a lot.

---

T1 | Test case: ./libTests/referenceTestCases/euler/2D/HPUnstructured/

A) READING MESH FILE EXTENSION : square/square.msh

Exception sur lecture file XML : file introuvable ou structure incorrecte

---

file XML concerne : 'mesh file not found : ./libMeshes/square/square.msh'
infos sur exception code source :
file : 'src/Meshes/MeshUnStruct/MUSGmsh/MUSGmsh.cpp'
ligne : 54

---

Running multiple parallel jobs at once

In the ECOGEN main folder, cases can be defined in ECOGEN.xml and this file is referred to when running the script for a parallel run. Is it possible to run multiple parallel cases at once without creating multiple instances of the ECOGEN folder?

It is my understanding that if I defined multiple paths within one ECOGEN.xml file, then ECOGEN will run these cases one after another.

Non physical results of cavitation problems in two-dimensional unstructured grids

hello kevin：

When I used a two-dimensional unstructured fan grid to solve the cavitation problem, I found that the radius change of the cavitation seems to be different from the theoretical solution, and this situation does not occur in orthogonal grids.

In short, under the same operating conditions, the time required for bubble collapse will significantly increase with the increase of the computational domain, as shown in the figure. I tried to increase the mesh density, but this situation did not improve. This has greatly troubled me in doing grid independence verification.

If you want to test it, here is my example file.
o-mesh-novis90-600 50mm.zip

You sincerely han

Inlet Conditions

I have not been able to find injection in the documentation for boundary conditions available. I have noticed that there is an injection of sorts occurring in the test case "euler/2D/nozzles/tankWithShock/". Which boundary should be selected for creating an inlet? Should I reference boundCond type="Injection"?

Can I apply a velocity directly to the injection?

Great code so far, and thank you for the help.

UEq model

Dear all,

Greetings! I am simulating near-wall bubble collapse and am using pressure disequilibrium model with infinite relaxation. I have read here that UEq model with infinte relaxation converges to PUEq model/ Kapila's model. This should mean for a given cell, the liquid pressure and gaseous pressure should be same as both are in equilibrium in terms of pressure. However when I check simulation results, I see that both differ. Am I understanding the model wrong?

Thanking you in advance
Niranjan K. Sabu

ECOGEN is not producing .vtk file

Cavitation flow around a body

Hello all,

I am doing a research on cavitation flow. The solver ECOGEN is a good solver. Thank you!

My simulation issue is a break down when inflow is high enough or ambient pressure is low enough(The 2nd and 3rd figures below are my results which the configuration files are modified on the basis of foil cavitation files in test cases). The minimal pressure is very low compared with saturated vapor pressure. I read the paper by Saurel et al., and they said mass transfer is necessary(marked in green in the first figure below in this e-mail). But I don't know how to add mass transfer into the model. Sorry for my poor knowledge about it. Thanks for your help in advance,

Cheers

Is ECOGEN Able to Simulate Non-Newtonian Fluids?

Hello all,

I am new to ECOGEN and found this project really amazing. I have a question for ECOGEN based on my application: can ECOGEN simulate Non-Newtonian multiphase flows? It seems that the rheology models are not included in the source code.

Thank you in advance.

YBY

Recombination of VTK files

Dear all,

Is there a way to recombine VTK files associated with multiple cores to produce a single VTK file for a given time-step using ECOGEN itself ? If not is there any other way?

Regards,
Niranjan K. Sabu

Conductivity

Is the conductivity module functional?

The axisymmetric source term of viscosity.

Dear ECOGEN developers,

Hi!

When reading the axisymmetric source term of viscosity in APKViscosity.cpp, it seems the divergence of velocity is not properly evaluated in cylindrical coordiantes. Extra terms related to

shall be added to the cylindrical source terms. Please refer to 1.2-7 for more details as the equations are not easily rendered on github..

For example, when accounting for this term, the source term of momentum in x direction on line 353

static_cast<FluxKapila*> (fluxBuff)->m_qdm.setX(muMix * (dvdx + dudy) / r);

is

static_cast<FluxKapila*> (fluxBuff)->m_qdm.setX(muMix * (dvdx + 1.0/3.0 *dudy) / r);

Since other methods can be performed accounting for the axisymmetric assumption, such as modifying the cellInterface area instead of adding extra source terms, I'm not sure if this is a bug.

Regards,
Fan E

Injection of Transport for Surface Tension

I am trying to inject a liquid with surface tension on an unstructured grid with airflow. I am including the surface tension physics using the color transport. However, when I try to run, I get the following error:

Exception sur file XML : attribut introuvable 
****************************************
 file XML concerne : './libTests/myRuns/liquidInjectSurfTens/initialConditionsV4.xml'
 infos sur exception code source :
  file : 'src/BoundConds/BoundCondInj.cpp'
  ligne : 135
 name de l attribut recherche : 'Not enough transport equations in inj BC'

Here is my input for the liquid phase in the initial conditions file:

<boundCond name="jet" type="injection" number="4">
    <dataInjection m0="50000"/>
    <dataFluid EOS="IG_air.xml" density="1." pressure="1.01e5" alpha="0."/>
    <dataFluid EOS="SG_liquidForSurfaceTensionCases.xml" density="1000." pressure="1.01e5" alpha="1."/>
    <transport name="color" value="1."/>
</boundCond>

Can someone please let me know how to include the missing transport equation for injection BC?