May 2018 - Nektar-users - imperial.ac.uk

Incompressible Navier-Stokes solver result problem
by Csaba Biró 15 Jul '18

15 Jul '18

Dear all, I am trying to solve an edge tone configuration related problem. I am using the Incompressible Navier.Stokes Solver to calculate the effects of blowing a jet flow, that impinges a wedge-shaped object. My problems are, that there seems to be an insufficient amount of outflow at the outlet ("u" velocity doesn't change), as if the airflow got stuck and that the errors shown after the calculation is complete are very high. Changing the time step or the number of modes did not decrease these errors considerably. Following the link you can see the used xml file (inc.xml). I attached a picture about the resulting "u" velocity and the geo+mesh. https://drive.google.com/open?id=1PYMXzPJo26wcIEns4OA71oI1lwWMCoYH I am only simulating half of the the problem. (hence the symmetry axis). I think there should be a much greater change in the "u" velocity overall. I double checked my boundary conditions. Please give me your opinion. Best regards, Csaba Biró

2 1

curved elements and BL splitting
by Vishal Saini 15 Jun '18

15 Jun '18

Hi all, This is regarding NekMesh utility. I would like to mesh a pipe which has coarse elements to begin with (inlet cross section shown in figure). In order to make the outer surface smooth I applied "spherigons" which worked fine (see attached). However, when I try to use "boundary layer splitting", NekMesh says: No elements detected to split. The element type on the outer surface are hexahedra. Is it the problem? Can anyone suggest a way out? Cheers, Vishal

2 1

fixed mass flux in channel flow
by Sandeep Pandey 01 Jun '18

01 Jun '18

Hello all, I am a newbie to Nektar++ and I validated turbulent channel flow DNS with the standard DNS database. For now, the flow was driven by the constant pressure gradient. As a next step, I want to drive flow with such forcing so that I have constant mass flux at every time step. The forcing should take the following form to achieve the constant mass flux: *dpdx(time)= dpdx(time-1) -[ 0.3*(massFlowRef (time=0) - 2*massFlow (time) + massFlow(time-1) ) / (Ly*Lz*dt)]* where; *massflow *should be volume integrated value. Does anyone has already implemented such forcing? Or does anyone any idea? I am familiar with ./library/SolverUtils/Forcing, but any hint in the specific direction will be very useful to me. Thanks and have a nice weekend! Sandeep

2 2

History points and reynoldstress files
by Amitvikram Dutta 23 May '18

23 May '18

Hi all, I'm trying to use the Reynolds Stresses time - averaged filter and the History Points filter on my current case. The appropriate section of the .xml file looks like this <FILTER TYPE="ReynoldsStresses"> <PARAM NAME="OutputFile">AverageField</PARAM> <PARAM NAME="RestartFile">MyRestartAvg.fld</PARAM> <PARAM NAME="OutputFrequency">100</PARAM> <PARAM NAME="SampleFrequency"> 10 </PARAM> </FILTER> <FILTER TYPE="HistoryPoints"> <PARAM NAME="OutputFile">TimeValues</PARAM> <PARAM NAME="OutputFrequency">10</PARAM> <PARAM NAME="Points"> 1 0 10 2 0 10 3 0 10 4 0 10 5 0 10 6 0 10 7 0 10 </PARAM> </FILTER> The simulation runs fine over 10000 time steps. However, no .his files or .fld files for reynolds stresses are generated. I'm in a bit of a quandry as to how to resolve this issue. PFA the full .xml file as well. -- *Amitvikram Dutta* Graduate Research Assistant Fluid Mechanics Research Lab Multi-Physics Interaction Lab University of Waterloo

3 2

Adapting SVV Kernel for use in compressible flow solver
by Amitvikram Dutta 20 May '18

20 May '18

Hi all, Our lab is trying to adopt Nektar++ as its solver of choice for both compressible and incompressible problems in the turbulent and transitional regimes of flow. Therefore, we have been exploring the use of the SVV kernel and have found it to be quite suitable for incompressible flows. However, we understand that SVV was never implemented for the compressible flow solver. We'd like to examine the effort it would take to actually carry out this implementation. To this end could you put me in touch with anyone who has done any work on this direction in the past? Sincerely, -- *Amitvikram Dutta* Graduate Research Assistant Fluid Mechanics Research Lab Multi-Physics Interaction Lab University of Waterloo

2 1

MKL_LAPACK-NOTFOUND
by Müller, Tobias 15 May '18

15 May '18

Hi guys, at the moment I'm trying to build nektar++ with mkl, but unfortunately ccmake doesn't find MKL_LAPACK. Nevertheless the other parameters like MKL_CORE, MKL_INCLUDE_DIR, MKL_INTEL and MKL_SEQUENTIAL are set right. Furthermore, the building prozess seems to work properly. Now I have the question whether there is an easy way to confirm that really mkl lapack is used or not? If somone had an similar issue I would be glad for any suggestions. with best regards Tobias Müller

2 1

working CFL
by Vishal Saini 10 May '18

10 May '18

hi all, I'm running coarse simulations of fully (not quasi) 3D Taylor-Green vortex case at Re=1600 using SVV in IncNavierStokesSolver. The domain is [-2pi,2pi]^3 with elemental grid 16^3 and NumModes=5 (giving me 64 solution points per direction). The case is defined in the tutorials at: http://doc.nektar.info/tutorials/latest/incns/taylor-green-vortex/incns-tay… I'd like to ask the users a typical working maximum dt or CFL for similar flows with NumModes=5. I could not run the simulation for less than normalised dt=1e-3 which corresponds to: CFL = V0*delT/delX, = 1*1e-3/(2pi/64) ~ 0.01 I tried to run the tutorials (at the link above) with 16^2 grid with NumModes=5 and homogeneous third direction. I could run the simulation up to dt=20e-3. Thus, maximum working dt for fully 3D simulation is at least 20 times smaller. There may be a problem with the boundary since the numerical instability starts at the periodic boundary normal to z-direction. I'm attaching the fully 3D set-up being used. kindly have a look and let me know if I am missing something. Cheers, Vishal

1 0

Initial conditions problem
by Amitvikram Dutta 10 May '18

10 May '18

Hi all, I'm trying to simulate an initially laminar jet flow and getting it to show unsteady behavior is proving a bit problematic. Specifically it arises while adding awgn initial boundary conditions while simataneously trying to run this case in parallel. I have attached my xml file for the case but in brief the problem lies with the initial conditions, which are as follows: <FUNCTION NAME="InitialConditions"> <E VAR="u" VALUE="1 + 0.0001*awgn(1.0)" /> <E VAR="v" VALUE="0.0001*awgn(1.0)" /> <E VAR="p" VALUE="0.0001*awgn(1.0)" /> </FUNCTION> When I try to run this upto 10 cpus, the solver works fine and produces the required output as expected. However, when I try to run this on greater than 10 cores the case crashes instantly. *======================================================================= EquationType: UnsteadyNavierStokes Session Name: 2d_jet_final_Re_4632 Spatial Dim.: 2 Max SEM Exp. Order: 5 Num. Processes: 20 Expansion Dim.: 2 Projection Type: Continuous Galerkin Advection: explicit Diffusion: implicit Time Step: 0.0001 No. of Steps: 150000 Checkpoints (steps): 1000 Integration Type: IMEXOrder2 Splitting Scheme: Velocity correction (strong press. form) Smoothing: SVV (spectral/hp SVV (cut-off = 0.75, diff coeff = 0.10000000000000001)=======================================================================Initial Conditions: - Field u: 1 + 0.0001*awgn(1.0) - Field v: 0.0001*awgn(1.0) - Field p: 0.0001*awgn(1.0)Writing: "2d_jet_final_Re_4632_0.chk" (0.0232423s, XML)CG iterations made = 5001 using tolerance of 1e-06 (error = -nan, rhs_mag = 1.6679)Fatal : Level 0 assertion violationExceeded maximum number of iterations-------------------------------------------------------Primary job terminated normally, but 1 process returneda non-zero exit code. Per user-direction, the job has been aborted.-------------------------------------------------------* It seems as if the solver diverges very quickly. What I cannot understand is why it should fail instantly on running with more processors, particularly as it seems unconditionally stable for 10 processors. If the solution were to blow up shouldnt it do so after at least a few iterations? As you can see, I have also included the SVV option to stabilise the solution but this seems not to make any appreciable difference. I'd appreciate some insight into the matter. -- *Amitvikram Dutta* Graduate Research Assistant Fluid Mechanics Research Lab Multi-Physics Interaction Lab University of Waterloo

1 0

Scaling tests and parallelization
by Amitvikram Dutta 08 May '18

08 May '18

Hi all, I'm trying to run a few scaling tests on the cluster I have acess to. I'm using a mesh with 10,800 elements using an expansion order of 5. The simulation is set to run for 10,000 time steps. The issue that I'm running into is as follows : Doubling the number of processors *increases* the total cpu wall time. *Procs Wall Time* * 1 201 s* *20 209 s* * 40 242 s* I believe this is due to the overhead caused by writing checkpoint files (each parallel stream seems to write a separate checkpoint file). I have reduced the output frequency to the point that only 1 checkpoint file should be written for the entire simulation time, However, this still requires *n *checkpoint files to be written where *n *is the number of processors the case is parallelised on. In all cases I use the mpirun command. For example mpirun -np *n* IncNavierStokesSolver case.xml Could I have some pointers for proceeding further with this issue? Sincerely, -- *Amitvikram Dutta* Graduate Research Assistant Fluid Mechanics Research Lab Multi-Physics Interaction Lab University of Waterloo

2 1

NaN appears in using CompressibleFlowSolver
by 郭先文 07 May '18

07 May '18

Dear all, Recently I'm applying Nektar++ into calculating a 2D case of subsonic flow around a circular cylinder. The flow condition is under Re = 200000 and Ma = 0.75. And NaN in speed of sound keeps appearing within two specific element in the downstream and off-wall region of the cylinder, even after changing of grid size and time step. I wonder if the problem could be solved with the usage of artificial viscosity or smooth shock capture method (Non-smooth didn't work). But the user's guide doesn't cover enough information about those settings in detail. Where could I find some details and settings about those methods? Or is there any other solution to the problem? Thanks very much for your help. Regards, Xianwen. Xianwen Guo College of Engineering Peking University

2 1