- Nektar-users - imperial.ac.uk

r-adaptation
by 蒋树杰 05 Jan '20

05 Jan '20

******************* This email from 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, Now I want to use Nektar++ to do some transonic simulation (Compressible Solver). I want to increase the resolution for shocks. I find a PDF named 'Mesh adaptation (r & p) for compressible flows in Nektar++', given by Joaquim Peiro et al. They mentioned r-adaptation can be used by clustering nodes near shocks. But I can not find more information about this. Is this function available in Nektar++ 5.0.0? If someone is familiar with this, please let me know. Merry Christmas! Thank you for your help! Shujie University of Southampton

2 1

Field Interpolation during runtime
by Nitish Arya 05 Jan '20

05 Jan '20

******************* This email from 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, I want to perform an aeroacoustic simulation in which I am using APESolver (nektar-4.4.1). I am transferring the source terms from the fluid solver (OpenFoam) to nektar++. I need to interpolate the source terms from the finite volume mesh (which is relatively coarser) ino the spectral mesh which can be used by APESolver. Initially, I used Shepard's algorithm to interpolate the source terms which performed well for serial runs but, unfortunately, there are several numerical errors when I am passing the data using multiple processors. Recently I found out that there is an already available utility in nektar++ for field interpolation ( as post-processing utility). I wanted to inquire if it is possible to use the functionality for interpolating the data during runtime with multiple processors. Happy new Year to all of you!! regards, Nitish Arya Department of Aerospace Engineering IIT Kanpur, India

2 1

Nektar++ question
by Moto Light 20 Dec '19

20 Dec '19

******************* This email from 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 Nektar usaers, Now i have a problem in Nektar, i did a Floquet analysis study using Nektar, it is working fine and the results looks fine, when i am doing this i am just using two modes mode 0 and mode 1, mode 0 is the 2D filed and mode 1 is the 3D expansion, do you have an idea about how i can extract mode 1 from my solved field file?? I tried something but i am not sure it is fine: I used the meanmode module to extract mode zero as a 2D field results from the complete field using this: *FieldConvert -m meanmode file.xml file.fld file-mean.fld* after i used this 2D field to be a 3D by expand it with the same Lz and the same number of modes of my complete case but also putting w=zero After iI used Addition (( to used as a subtraction using scale -1)) *FieldConvert -m addfld:fromfld=modified_mode0.fld:scale=-1 file1.xml complete_field.fld output_field.fld* *and the results like a 3D mode 1 * *do you think this way is ok???* *==============* Thank you in advance, and i wish all of you have a great day. Regards, *Reda El Mansy* PhD candidate Department of Aeronautics UPC Department of Aeronautics

1 0

Nektar++ question
by Josuel Kruppa Rogenski 05 Dec '19

05 Dec '19

******************* This email from josuelkr(a)gmail.com originated outside Imperial. ******************* Dear Nektar users, I am doing some tests using IncNavierStokes solver + steady driver (with default parameters) + periodic boundary conditions. I wonder if someone has to take some extra care if one of the boundaries is similar to a sawtooth wave, as an example. Specifically, I took initially a 2D-Couette flow. After that, I substituted one of the flat walls to a tilted small square step. Results seems to be nonphysical. I played with different meshes and time integration methods but could not draw any conclusion. My .xml file solver info is: <SOLVERINFO> </SOLVERINFO> <EXPANSIONS> <E COMPOSITE="C[0]" NUMMODES="12" TYPE="MODIFIED" FIELDS="u,v,p" /> </EXPANSIONS> Any help would be appreciated. Thank you in advance, J PS: I can provide .xml and .geo files I am working with if necessary.

3 3

Floqueat analysis Mode1 problem
by Moto Light 04 Dec '19

04 Dec '19

******************* This email from redamansyhh(a)gmail.com originated outside Imperial. ******************* Hi All, Now i have a problem in Nektar, i did a Floquet analysis study using Nektar, it is working fine and the results looks fine, when i am doing this i am just using two modes mode 0 and mode 1, mode 0 is the 2D filed and mode 1 is the 3D expansion, do you have an idea about how i can extract mode 1 from my solved field file?? I tried something but i am not sure it is fine: I used the meanmode module to extract mode zero as a 2D field results from the complete field using this: *FieldConvert -m meanmode file.xml file.fld file-mean.fld* after i used this 2D field to be a 3D by expand it with the same Lz and the same number of modes of my complete case but also putting w=zero After iI used Addition (( to used as a subtraction using scale -1)) *FieldConvert -m addfld:fromfld=modified_mode0.fld:scale=-1 file1.xml complete_field.fld output_field.fld* *and the results like a 3D mode 1 * *do you think this way is ok???* *==============* Thank you in advance, and i wish all of you have a great day. Regards, *Reda El Mansy* PhD candidate Department of Aeronautics UPC Department of Aeronautics

1 0

(no subject)
by A.P. 04 Dec '19

04 Dec '19

******************* This email from k210(a)list.ru originated outside Imperial. ******************* Hello, is it impossible to use u,v,w in the forcing expressions like this </CONDITIONS> … <FUNCTION NAME="MyForce"> <E VAR="u" VALUE="v+u" /> <E VAR="v" VALUE="1.0" /> </FUNCTION> ... </CONDITIONS> <FORCING> <FORCE TYPE="Body"> <BODYFORCE> MyForce </BODYFORCE> </FORCE> </FORCING> ? Alexander Proskurin

1 0

Combine two .fld files containing time averages
by Moto Light 08 Nov '19

08 Nov '19

*Hi all,* For averaging two average files in Nektar++, the user guide said that we need to use the next order for averaging file1.fld and file2.fld: *FieldConvert -m combineAvg:fromfld=file1.fld file1.xml file2.fld output_file3.fld* *This command works fine using two files as an inputs files.* *________________________________* *How we can take average of three files, I tried to used the same command by adding the third file name after file2.fld to be like:* *FieldConvert -m combineAvg:fromfld=file1.fld file1.xml file2.fld file3.fld output_file3.fld* it works with no error and gives output file, but i tried to corrupt the name of the third file in the command line and it worked fine again, which means there is no consideration for the third file name in the command line???? Any suggestions? for taking the average of three average field files in one command line Thanks, Regards Reda el Masny PhD student in UPC.es

2 1

Product between base flow derivative and the aerodynamic forces
by Daiane Iglesia Dolci 13 Oct '19

13 Oct '19

Dear all, I need to solve a mathematical operation that is shown in the figure attached below. I am simulating an adjoint field and I wanna compute the product between a derivative of the base flow velocity and the aerodynamic force of the adjoint field, at the wall boundary. I know that in the FilterAeroforces.cpp, the forces are computed in each element, then an integral is applied. So I would like to know how I can take the derivative of the base flow in these elements by reading the baseflow.fld. If someone can help me, I thank you. Daiane I. Dolci.

2 1

"True" pressure in periodic BC
by Adriano Takata 13 Oct '19

13 Oct '19

Hi, how are you? (Nektar++ 5.0) I am working in Poiseuille flow 2D with periodic boundary conditions. I get the solutions (u = (1 + y) * (1-y), v = 0, and p = 0) with the solver. The pressure is 0 due to the flowrateforce, but I would like to know if I can get the "true" pressure thought the velocity field. Thank you. Adriano Sueke Takata

2 1

Compressible solver queries
by Vishal Saini 09 Oct '19

09 Oct '19

Hi all, I was giving the compressible flow solver a try and have a couple of queries. 1. Are the history probes set up for this solver? 2. I was looking into the possibility of adding a passive scalar term for some validation. From a first glance, it seems that an additional scalar transport equation is not implemented in the CompressibleFlowSolver (at least in the v4.5.0). Is that correct? If yes, has anyone tried that yet? If not, could you please give me some ideas on how to start about it? I guess, we just need to add convective and diffusive fluxes to the system of equations somehow (in CompressibleFlowSystem.cpp and NavierStokesCFE.cpp?). And I guess the Boundary Conditions can be generically defined for extra variables as in CompressibleFlowSystem::DoOdeProjection function(?) PS. What is the "smooth viscosity model"? 446 // For the smooth viscosity model 447 if (nVariables == m_spacedim+3) 448 { 449 // Add a zero row for the advective fluxes 450 for (j = 0; j < m_spacedim; ++j) 451 { 452 Vmath::Zero(nq, flux[m_spacedim+2][j], 1); 453 } 454 } Thank you in advance for your time. Best Regards, Vishal --- *Vishal SAINI* PhD Student Loughborough University

3 2