April 2015 - Nektar-users - imperial.ac.uk

Mesh order on curved boundary
by Zhicheng Wang 30 Apr '15

30 Apr '15

Hi there, I'd like to use nektar++ to simulate 3d flow over a circular cylinder. but I'm feeling a little confused about the mesh order of curved boundaries. My puzzle is as follow: If SEM basis is 4-order polynomial, and the mesh, generated by Gmsh, is 3-order. it means the order of base function is higher than mesh order of mesh. is there any inconsistency in this case ? zhicheng

2 1

FieldConvert error from calculating vorticity and Qcriterion
by Zhicheng Wang 30 Apr '15

30 Apr '15

Hi nektar++ users and developers, I 'm using the latest nektar++ version. However when running the TurbChFl_3DH1D.xml example and using FieldConvert calculating vorticity or Qcriterion, I got errors as "ERROR: Element u_33 is 0 from dgetrf". I have checked the velocity and pressure field, it is correct, so if there is some inconsistency in FieldConvert for 1D homogeneous problem?

3 9

New collections functionality
by David Moxey 23 Apr '15

23 Apr '15

Dear all, This morning we merged some new functionality into master that you should be aware of. There is a new library, called Collections, which essentially has been formulated to apply operators such as BwdTrans and IProductWRTBase across multiple elements at once, leading to significant performance boosts in some cases. This is like the global optimisation parameters, but more configurable, easier to use and across many elements at once. Collections are enabled by default, but in their default mode will just replicate the functionality that you see in ExpList, so your simulation results should not change. However, if in your session file, you add the tag: <COLLECTIONS DEFAULT="auto" /> inside the main NEKTAR tag, then the collections library will attempt to figure out what the quickest scheme is at runtime. If you use the '-v' command line option, you will see the auto-tuning results, which is a table of the timings for each strategy. This takes 10-60 seconds to run, depending on problem size. You can do some more advanced setup and manual selection of schemes in the XML file, which is documented in the user guide. Note that certain schemes require increased memory usage. Also, this isn't magic: some solvers, particularly those that are heavily solver-bound such as the incompressible NS equations, aren't likely to see huge benefits. However, DG-based solvers such as the APESolver and CompressibleFlowSolver can observe quite large speedups in certain cases. Let us know via this mailing list if you run into problems or encounter any errors. Cheers, Dave -- David Moxey (Research Associate) d.moxey(a)imperial.ac.uk | www.imperial.ac.uk/people/d.moxey Room 363, Department of Aeronautics, Imperial College London, London, SW7 2AZ, UK.

1 0

Nektar++ workshop invitation, 7-8th July 2015 at ICL
by David Moxey 22 Apr '15

22 Apr '15

Dear all, This email is being sent to invite you to a workshop that we are organising for users and developers of Nektar++ (www.nektar.info) as well as anyone that is interested in seeing some of the applications of our high-order framework in a variety of subject areas. The workshop is being held across two days, Tues 7th and Wed 8th July 2015, here at the South Kensington campus of Imperial College London. Further details about exact times and locations will be sent out closer to the date. Tentatively, the schedule is: - Tues 7th July: a series of casual, short-format presentations from users and developers, with dinner in the evening. - Wed 8th July: a general group session for discussing long-term goals, new features and implementation challenges/bugs that you are finding within the current structure. If you're interested in attending, I'd request that you fill in the doodle poll that we have set up for the event so that we know the number of people to expect: http://doodle.com/iyigwkek7bhddu4b You can choose to attend either or both days, depending on what you find most useful. Please feel free to send this email across to anyone else you think might be interested in attending the workshop. Also, we will be looking for volunteers for presentations on applications involving Nektar++, so please email me if you feel that you will be able to present! They will be around 15-20 minutes with questions, depending on the number of presenters we have on the day. Finally, if you are currently a user or developer of Nektar++ and have not registered for our users mailing list, through which you can get some support for the framework, I'd encourage you to do so here: https://mailman.ic.ac.uk/mailman/listinfo/nektar-users Many thanks, Dave -- David Moxey (Research Associate) d.moxey(a)imperial.ac.uk | www.imperial.ac.uk/people/d.moxey Room 363, Department of Aeronautics, Imperial College London, London, SW7 2AZ, UK.

2 1

Womersley velocity profile
by Kamil ÖZDEN 19 Apr '15

19 Apr '15

Dear All, I want to implement the Womersley solution below for inlet velocity boundary condition. where A is the amplitude of pulsation and is the non-dimensional Womersley number. These are equal to 0.667 and 7.5 respectively. Is there any way to implement this profile in Nektar++ (any extension code etc.) ? Regards, Kamil

1 0

Generating curved edge in ICEM CFD for Nektar++
by Kamil ÖZDEN 03 Apr '15

03 Apr '15

Dear All, I found such a way to generate my meshes to use in Nektar++: 1) First I create my mesh in ICEM CFD. 2) Then I export that mesh from ICEM CFD as a Nastran file (with *.dat extension). 3) After that I open that *.dat file in Gmsh and save it as a *.msh file. 4) Finally I generate my *.xml file from that *.msh file. This works in Nektar++ but I have a problem about generating curved edges in ICEM CFD. There is no such option or I couldn't find how to succeed this. Is there any one who use this method to generate mesh for Nektar or can inform me about generating high order edge in ICEM CFD? NOTE: I've tried to generate curved edge in Gmsh with the imported *.dat file but it didn't work. Regards, Kamil

3 6