- Nektar-users - imperial.ac.uk

Error "gen_bndry: Illegal B.C. function definition" in Hidden Fluid Mechanics platform.
by Jean Schuster 10 May '21

10 May '21

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hi everybody. I hope everybody is doing good. My name is Jean Schuster, from Brazil. I'm starting my studies in neural networks and machine learning, when I found a subject of interest in a paper named Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations <https://science.sciencemag.org/content/367/6481/1026>. In that papar, Dr. Raissi, Dr. Yazdani and Dr. Karniadakis addressed a fluid mechanics problem of flow visualization, like smoke or dye visualization for qualitative characterization of flow fields, "injecting" a passive escalar (say smoke) through a transport equation together with the Navier-Stokes equations, developing a framework that they called hidden fluid mechanics (HFM), a physics-informed deep-learning framework capable of encoding the Navier-Stokes equations into the neural networks while being agnostic to the geometry or the initial and boundary conditions. That framework uses Nektar, that I suppose they had modified to insert the passive scalar transport equation, to run the CFD simulations for the example problems they demonstrate in the papar. It can be downloaded under a GNU general public license at: https://github.com/alirezayazdani1/HFM After some days I managed to compile it and its dependencies in a 64bits Ubuntu 20.04.2 LTS machine. I attached a file called log_compiling.txt that contains the terminal output during the compiling process. Everything seems fine, but when I try to run, for example the classic problem of a flow over a cylinder that is included in the framework, I get an error saying: gen_bndry: Illegal B.C. function definition. I'm using the following command to run the example: ./nektar -V -chk -S -n4 ./RUN/Cylinder2D/cylinder2d.rea 2>&1 | tee log_run_cyl2d.txt It returns: jean@Prandtl:~/HFM/SourceCode$ ./nektar -V -chk -S -n4 ./RUN/Cylinder2D/cylinder2d.rea 2>&1 | tee log_run_cyl2d.txt rcm bandwidth (u) : 18080 [204031 (0)] Input File : /RUN/Cylinder2D/cylinder2d Concat files : 0 Re (U=1, L=1) : 100 Time step : 0.0008 Integration order : 2 Number of modes : 4 Number of elements : 45856 Number of Families : 45856 Equation type : Navier-Stokes (convective) Integration time : 120.008, or 150010 steps I/O time for saves : 40, or 50000 steps [checkpoint] NS Scheme Type : Splitting scheme Solver Type : Direct RCM gen_bndry: Illegal B.C. function definition Follow attached the cylinder2d.rea file. I would try to use some of the tutorials available here https://www.nektar.info/community/tutorials/ with the compiled code, but it returns me an error saying "The requested URL was not found on this server" in every link in the tutorials page. I'm pretty sure that this is not exactly the right place for this question, but I don't know where else I could get help. So I hope I'm not bothering any of you. Thanks in advance. Best regards, Jean.

1 0

Parallel 3DH2D simulation
by Xiao, Chengnian 09 May '21

09 May '21

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Dear all, I would like to know if the latest nektar++ version supports parallel 3DH2D simulation, i.e. 3D configuration based on 1D mesh and Fourier expansions in y,z directions? According to the user-guide, there are ---npz, -npy,--npx options available. However, whenever I run the 3DH2D test case provided in the sample session file ChanFlow_3DH2D_FFT.xml on multiple processors with mpirun -np >1, the solver would crash after initialization, no matter if npx, npy or npz are specified to be greater than 1 or not. The case does work fine however on a single processor. Best regards, Cheng

2 1

Apply filer to a specified domain range
by Stanisław Gepner 03 May '21

03 May '21

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hi, I was wandering if it is possible to program into a Filter a possibility to be applied onto a selected range of the domain. Much like what is possible using FieldConvert -r option. I looked at the code, and there is the DomainRange, but I am not sure there is a way to use it on ExpList through which I access coordinates, physical values from within the Filter. One solution I see is to just go through what is available through GetCoords() (at initialization) and create a mask to filter out values from outside the specified range. The filter I am working on is based on quantities integrated from the fields, So I think I could create the mask at initialization and than use it at updates to zero-out values outside. But might it be, that something like this is already implemented? Any suggestions are welcome. Cheers! Stan -- Best regards, Stanisław Gepner, Ph.D. *Faculty of Power and Aeronautical Engineering* Nowowiejska 21/25 Str. 00-665 Warsaw, Poland phone +48 (22) 234 51 70 <https://www.pw.edu.pl/engpw>

2 2

Post-processing - Wall Shear Stress
by Paulo henrique Ferreira 03 May '21

03 May '21

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hi, I ran a typical simulation of a circular cylinder at Re = 3900, but now I'm having to deal with a small issue in the post-processing. I'm unable to properly obtain the wall shear stress, some ripples appear along the span, as shown in the screenshot. The C0 projection (FieldConvert module) has not worked and I'm not sure if it should have worked to the surface mesh obtained from the XML file. Does anyone know what I should do to fix this problem or have any other suggestions in this case? Thank you! Regards, Paulo Henrique PhD StudentInstituto Tecnologico de Aeronautica - ITA 12228-900 - Sao Jose dos Campos - SP - BRAZIL

2 1

Set up Help
by Paul Liu 28 Apr '21

28 Apr '21

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hi Everyone, Some really stupid questions. I managed to install Nektar++ from the source. The installation was successful as I managed to run "ctest" under the "build" folder and all the tests are passed. Then I tried to add the path of the solvers that I want "AcousticSolver" and " to my ".bashrc" file so that I can call the function/exectuable. ex: export PATH = "home/paul/nektar++/build/solvers/ADRsolver:$PATH". However when I tried to call the function to run the tutorial on the advection-convection equation I was getting segmentation errors. My tutorial files are located outside of the "build" folder where all the code was built, is that the reason? If not what should I do/try to debug this problem. Sorry I am completely new to linux haha. *YiLin Liu * *403-397-3457* *liuyilin.liu(a)gmail.com <liuyilin.liu(a)gmail.com>* MASc Student at CML <https://cml.mech.ubc.ca/> UBC

1 1

Nektar++ 5.0.2 Released.
by Chris Cantwell 22 Apr '21

22 Apr '21

Dear all, We are pleased to announce the release of Nektar++ v5.0.2. This release includes a number of further bug fixes to the 5.0.1 release: **Library** - Fix bug in StdHexExp FillMode (!1192) **Documentation** - Updated Documentation to include HDF5 Mesh Output (!1230) **CI** - Add Debian Bullseye to CI system (!1181) - Removed Ubuntu Trusty (14.04) from CI and added Focal (20.04) (!1238) **BuildSystem** - Updated third party zlib version to 1.2.9 to resolve OCE source build issue (!1227) - Adding SolverUtils as a core library that is built by default (!1240) As always, if you encounter any problems please feel free to contact us on the mailing list nektar-users(a)imperial.ac.uk. Cheers, Chris -- Chris Cantwell Senior Lecturer in Aeronautics Department of Aeronautics Imperial College London South Kensington Campus London SW7 2AZ Tel: +44 (0)20 759 45050 Email: c.cantwell(a)imperial.ac.uk www.imperial.ac.uk/people/c.cantwell

1 0

Compressible flow solver
by Octavio Morales 17 Apr '21

17 Apr '21

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hi, I've been trying to work with this solver in a high sub-sonic compressible regime (Ma=0.87), however this appears: ======================================================================= EquationType: NavierStokesCFE Session Name: Cilindro Spatial Dim.: 2 Max SEM Exp. Order: 3 Num. Processes: 3 Expansion Dim.: 2 Riemann Solver: HLLC Advection Type: Projection Type: Discontinuous Galerkin Diffusion Type: Advection: explicit AdvectionType: WeakDG Diffusion: explicit Time Step: 1e-05 No. of Steps: 60000 Checkpoints (steps): 400 Integration Type: ClassicalRungeKutta4 ======================================================================= Initial Conditions: - Field rho: rhoInf - Field rhou: rhoInf * uInf - Field rhov: rhoInf * vInf - Field E: pInf / (Gamma - 1) + 0.5 * rhoInf * (uInf * uInf + vInf * vInf) Writing: "Cilindro_0.chk" (0.00527309s, XML) Fatal : Level 0 assertion violation NaN found during time integration. Fatal : Level 0 assertion violation NaN found during time integration. Fatal : Level 0 assertion violation NaN found during time integration. -------------------------------------------------------------------------- Primary job terminated normally, but 1 process returned a non-zero exit code. Per user-direction, the job has been aborted. -------------------------------------------------------------------------- -------------------------------------------------------------------------- mpirun 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: [[28957,1],1] Exit code: 1 -------------------------------------------------------------------------- ************************************************************************************************************************************************************************ This is the input: ************************************************************************************************************************************************************************ <EXPANSIONS> <E COMPOSITE="C[0]" NUMMODES="3" FIELDS="rho,rhou,rhov,E" TYPE="MODIFIED" /> </EXPANSIONS> <CONDITIONS>  <PARAMETERS> TimeStep = 0.00001 FinTime = 0.6 NumSteps = FinTime/TimeStep IO_CheckSteps = 400 IO_InfoSteps = 400 GasConstant = 287.058 Gamma = 1.4 pInf = 22600 rhoInf = 0.364 Mach = 0.87 cInf = sqrt(Gamma * pInf / rhoInf) uInf = Mach*cInf vInf = 0.0 Twall = 258.5 Re = 400000 mu = 0.00001167 Pr = 0.72 </PARAMETERS> <SOLVERINFO> </SOLVERINFO> ************************************************************************************************************************************************************************** Is the problem related to the quality of the mesh or it is related to the Solver info module? Any help is extremely appreciated. Thanks for your time. Morales González Octavio Augusto. U-MOFABI Facultad de Ingeniería, Universidad Nacional Autónoma de México.

2 2

Post-processing - Wall Shear Stress
by paulohferr＠yahoo.com 17 Apr '21

17 Apr '21

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hi, I ran a typical simulation of a circular cylinder at Re = 3900, but now I'm having to deal with a small issue in the post-processing. I'm unable to properly obtain the wall shear stress, some ripples appear along the span, as shown in the screenshot. The C0 projection (FieldConvert module) has not worked and I'm not sure if it should have worked to the surface mesh obtained from the XML file. Does anyone know what I should do to fix this problem or have any other suggestions in this case? Thank you! Regards, Paulo Henrique PhD StudentInstituto Tecnologico de Aeronautica - ITA 12228-900 - Sao Jose dos Campos - SP - BRAZIL

1 0

Rotational arguments for periodic faces
by Vidhi Zala 28 Mar '21

28 Mar '21

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hi everyone, I am trying to implement a 3D advection example on a cube with composite elements (hex+tets+pyramids+quads+tri). I am using periodic boundary conditions. I am getting the following error: "Unable to determine face orientation". I referred to the relevant section of user guide (please see screenshot). I am unable to understand how should I change the rotation argument such that I can simulate a periodic flow. It would really help if I get some more details about this. Many thanks, V ,-~~-.___._. / | ' \ | |--------. Vidhi Zala ( ) 0 | | | School of Computing \_/-, ,----' | | | Scientific Computing and Imaging Institute ==== !_!-v---v--. University of Utah / \-'~; .--------. / __/~| ._-""|| | vidhi.zala(a)utah.edu =( _____|_|____||________| http://sci.utah.edu/people/vidhi.html I support Ruff Haven Crises Shelter for Homeless Pets <https://www.ruffhaven.org/donate>. Consider a donation today :)

2 3

Outflow Boundary Condition for ShallowWaterSolver
by Boyuan Yu 27 Mar '21

27 Mar '21

******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Hello all, I am using ShallowWaterSolver and my application requires an outflow boundary condition. However, there seems no appropriate outflow boundary condition for ShallowWaterSolver (the "HOutflow" and "PressureOutflow" boundary conditions seem to only work for IncNavierStokesSolver). So how could the outflow boundary condition be implemented for ShallowWaterSolver? The "outflow" boundary condition here only requires that the field variables advected out of the domain without reflection using some kind of extrapolation. Thank you. Boyuan Yu

2 1