April 2017 - Nek5000-users - mailman3.cels.anl.gov

Re: [Nek5000-users] Time derivative of Scalar field
by nek5000-users＠lists.mcs.anl.gov 14 May '17

14 May '17

Hi Saikat, I have attached a usr file that shows how to calculate finite difference based first derivative of the first passive scalar t(1,1,1,1,2) using the current time step and two previous time steps. The coefficients for the derivative are obtained using a call to fd_weights_full, and I am using a common block for storing the values of the scalar at two previous time-steps. Ketan

1 3

Time derivative of Scalar field
by nek5000-users＠lists.mcs.anl.gov 25 Apr '17

25 Apr '17

Hello Neks, I want to find out the time derivative of Scalar field in my Rayleigh Benard calculations. I wan't to know how to go about computing the time derivative during the run, and write that out each time I write my .fld files. Is there any subroutine in the .usr file that I can call? Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech.

1 0

Re: [Nek5000-users] Nek5000-users Digest, Vol 98, Issue 10
by nek5000-users＠lists.mcs.anl.gov 25 Apr '17

25 Apr '17

Hi Sudhakar, thank you for your reply. It seems to me that subroutine BCNEUSC is in charge of reading the "HC" value for all element faces withstanding the "c " boundary condition. Nevertheless there is no info about how to Robin boundary condition is effectively implemented within the solver. May I ask you to give me some more suggestions, in order to implement such BC on a velocity component? Best regards, Francesco 2017-04-22 12:36 GMT+02:00 <nek5000-users-request(a)lists.mcs.anl.gov>: > Send Nek5000-users mailing list submissions to > nek5000-users(a)lists.mcs.anl.gov > > To subscribe or unsubscribe via the World Wide Web, visit > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > or, via email, send a message with subject or body 'help' to > nek5000-users-request(a)lists.mcs.anl.gov > > You can reach the person managing the list at > nek5000-users-owner(a)lists.mcs.anl.gov > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Nek5000-users digest..." > > > Today's Topics: > > 1. Robin boundary condition on velocity > (nek5000-users(a)lists.mcs.anl.gov) > 2. Re: Robin boundary condition on velocity > (nek5000-users(a)lists.mcs.anl.gov) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Fri, 21 Apr 2017 20:35:23 +0200 > From: nek5000-users(a)lists.mcs.anl.gov > To: nek5000-users(a)lists.mcs.anl.gov > Subject: [Nek5000-users] Robin boundary condition on velocity > Message-ID: > <mailman.5414.1492799726.2967.nek5000-users(a)lists.mcs.anl.gov> > Content-Type: text/plain; charset="utf-8" > > Hi Neks > I'm trying to implement a slip-length (Robin) as boundary condition for a > plane channel flow, namely: > (u+du/dy)_{wall}=0 > where x and y are the streamwise and wall normal directions respectively. > > Following the latter post: > http://lists.mcs.anl.gov/pipermail/nek5000-users/2011-April/001313.html > > here there's what I came up with: > > c----------------------------------------------------------------------- > subroutine userchk > include 'SIZE' > include 'TOTAL' > common /my_DEBUG/ rmodulate > common /my_ROBIN/ ROBIN_mask(lx1,ly1,lz1,lelt) > $ , work1(lx1,ly1,lz1,lelt) > $ , work2(lx1,ly1,lz1,lelt) > $ , graduy(lx1,ly1,lz1,lelt) > n = nx1*ny1*nz1*nelv > !!!Build Up MASKS! > if (istep.eq.0) then > do i=1,n > x=xm1(i,1,1,1) > y=ym1(i,1,1,1) > if(x.gt.3.and.x.lt.4)then > ROBIN_mask(i,1,1,1)=1.0 > else > ROBIN_mask(i,1,1,1)=0.0 > endif > enddo > endif > !Compute velocity gradient > call rzero(graduy,n) > call gradm1(work1,graduy,work2,vx) > call dsavg(graduy) > ! call outpost(work1,graduy,work2,pr,t,'GRD') > !Da aggiungere l'eventualit? del 3D... > rmodulate = 0.5*(tanh(0.25*time-10)+1) > if (nid.eq.0) write(6,*)"MODULATE", rmodulate,time > return > end > c----------------------------------------------------------------------- > c----------------------------------------------------------------------- > subroutine userbc (ix,iy,iz,iside,eg) > c NOTE ::: This subroutine MAY NOT be called by every process > include 'SIZE' > include 'TOTAL' > include 'NEKUSE' > integer e, eg > !Imposing Fancy Boundary Conditions... > ![Nek5000-users] Robin boundary conditions for the velocity > common /my_ROBIN/ ROBIN_mask(lx1,ly1,lz1,lelt) > $ , work1(lx1,ly1,lz1,lelt) > $ , work2(lx1,ly1,lz1,lelt) > $ , graduy(lx1,ly1,lz1,lelt) > common /my_DEBUG/ rmodulate > e=gllel(eg) ! global element number to processor-local el. # > ux=1*graduy(ix,iy,iz,e)*rmodulate > uy=0.0 > uz=0.0 > return > end > c----------------------------------------------------------------------- > subroutine useric (ix,iy,iz,ieg) > include 'SIZE' > include 'TOTAL' > include 'NEKUSE' > ux=1.0*y**2 > uy=0.0 > uz=0.0 > temp=0 > return > end > c----------------------------------------------------------------------- > > The code becomes unstable in the gradient computation due to some > point-to-point oscillations. > As you can see I've also tried to use a ramp function in order to make the > onset of Robin condition more gentle, > but at a given "slip-length amplitude" oscillations deteriorates the > solution. > I use ROBIN_mask in order to chose where to have such BC, but it seems that > I'm still missing something in computing the velocity gradient GRD. > > By the way, this is my box.BOX file: > > base.rea > 2 spatial dimension ( < 0 --> generate .rea/.re2 > pair) > 1 number of fields > #======================================================================= > # > # Example of .box file for channel flow > # > # If nelx (y or z) < 0, then genbox automatically generates the > # grid spacing in the x (y or z) direction > # with a geometric ratio given by "ratio". > # ( ratio=1 implies uniform spacing ) > # > # Note that the character bcs _must_ have 3 spaces. > # > #======================================================================= > # > Box > -32 -16 nelx,nely,nelz for Box > 0 6.2832 1. x0,x1,gain > -1 1 1. y0,y1,gain > P ,P ,v ,v , , bc's (3 chars each!) > > > I was wondering if you might be able to give me some advice. > > Best regards, > > Francesco >

1 1

Time derivative of Scalar field
by nek5000-users＠lists.mcs.anl.gov 24 Apr '17

24 Apr '17

Hello Neks, I want to find out the time derivative of Scalar field in my Rayleigh Benard calculations. Can you give me any idea how to go about it? Right now I get the scalar field only along with the primitive variables. But I want to get the time derivative field of the scalars (dc/dt). Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech.

1 4

Reading boundary condition from an external file
by nek5000-users＠lists.mcs.anl.gov 24 Apr '17

24 Apr '17

Hi, I am supposed to carry out a simulation where I have to use the flow profile at the outflow boundary of the first computational domain as the inflow condition for the second domain. The first domain consists of a cylinder and the second domain is empty. Consequently, the meshes in the two situations are different. The study will help us in understanding the far wake characteristics of a bluff body. I checked the Nek archives and found a thread on similar topic. At that time, however, it was concluded that carrying out this exercise on parallel is quite challenging (at least this is what I understood from the thread). I would like to know if there has been any development in this direction? Or is there some trick that I can do to achieve this? Many thanks, N.

1 0

double mesh + solution at boundary + scalar blows up !
by nek5000-users＠lists.mcs.anl.gov 24 Apr '17

24 Apr '17

Dear Neks thank you Paul for your help on my double mesh question !! it works perfectly, and I am now testing quantitatively the results against published data. I am having two issues, one is likely stupid, and the other is likely more complex. stupid issue: When I try to plot the solution at the lower wall of the channel, it appears to be undefined. Is this normal / is there a way to access the solution at the lower wall of the channel ? (The velocity should vanish as I have no slip boundary conditions, but the scalar should not vanish as I have no-flux for the scalar.) complicatd issue: the flow starts from 0, it is forced by an average pressure gradients, has laminar initial condition+30% random fluctuations and it converges nicely. after the turbulent flow becomes stationary, i introduce the scalar, which develops in a nice fluctuating plume, looking reasonable. however, after some times, for no apparent reason, the scalar blows up. it sometimes produces NaN, and sometimes it doesnt. Can anyone think a reason why this happens? Reynolds = Peclet, and the scalar is injected at the inlet through the boundary condition. It does not blow up upon reaching the outlet (where I have outflow BC). I tried moving the source and making it bigger and model it as a Gaussian instead of a step function to avoid Gibbs oscillations. I also tried changing the boundary conditions from no flux to absorption (following https://lists.mcs.anl.gov/mailman/htdig/nek5000-users/2014-July/002961.html) I also halved diffusivity, so Pe = 1/2 Re, but no luck. is there any reason why the scalar should be more resolved than the flow ? is there any known numerical issue that i’m ignoring ? any hint would be fantastic. thanks a bunch ! Agnese > On Jan 21, 2017, at 7:00 PM, nek5000-users-request(a)lists.mcs.anl.gov wrote: > > Send Nek5000-users mailing list submissions to > nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]> > > To subscribe or unsubscribe via the World Wide Web, visit > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users <https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users> > or, via email, send a message with subject or body 'help' to > nek5000-users-request(a)lists.mcs.anl.gov <mailto:[email protected]> > > You can reach the person managing the list at > nek5000-users-owner(a)lists.mcs.anl.gov <mailto:[email protected]> > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Nek5000-users digest..." > Today's Topics: > > 1. Re: double mesh ? (nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]>) > > From: nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]> > Subject: Re: [Nek5000-users] double mesh ? > Date: January 21, 2017 at 1:18:41 PM GMT+1 > To: "nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]>" <nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]>> > Reply-To: nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]> > > > Dear Agnese, > > Attached is an example that shows how to modify the connectivity for the thermal > field to be inflow/outflow. > > If your domain is generated with genbox and your flow is in the x direction then > this should also work for 3D. > > Note that no changes are required to the .rea file (you should have "P " for both > velocity and temperature in the .rea/.box file). > > The changes occur in the .usr file. > > Paul > > From: nek5000-users-bounces(a)lists.mcs.anl.gov <mailto:[email protected]> [nek5000-users-bounces(a)lists.mcs.anl.gov <mailto:[email protected]>] on behalf of nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]> [nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]>] > Sent: Monday, January 16, 2017 9:31 AM > To: nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]> > Subject: Re: [Nek5000-users] double mesh ? > > Dear Paul > > thank you so much for your help ! > I am looking forward to see your example, it will be of great help ! > > could you tell me something more about the recycling BC as well ? > > Thank you again for your help > Agnese > >> >> From: nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]> >> Subject: Re: [Nek5000-users] double mesh ? >> Date: January 16, 2017 at 4:05:21 PM GMT+1 >> To: "nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]>" <nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]>> >> Reply-To: nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]> >> >> Dear Agnese, >> >> Sorry for the long delay in answering this question. The issue is that in fact it is possible >> (and used to be a supported but rarely used feature). The challenge, however, is how to >> make it relatively easy in the context of the current setup for a variety of applications. >> >> Nek supports a different gather-scatter handle for each field (velocity, temperature, passive >> scalar, etc.) >> >> We have a handle that allows the user access to the connectivity list, which would give >> you an opportunity to "disconnect" the periodic faces for the temperature array. >> >> I'll try to code up an example for the case of channel flow with a mesh generated by >> genbox and post it shortly. >> >> Incidentally, an alternative approach that we often use for more complex domains >> would be to use a recycling boundary condition. I think it would not be as efficient, >> however, as periodicity for the case that you propose. >> >> Once again, apologies for the delayed response. >> >> Paul >> >> From: nek5000-users-bounces(a)lists.mcs.anl.gov <mailto:[email protected]> [nek5000-users-bounces(a)lists.mcs.anl.gov <mailto:[email protected]>] on behalf of nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]>[nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]>] >> Sent: Thursday, December 22, 2016 4:38 AM >> To: nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]> >> Subject: [Nek5000-users] double mesh ? >> >> Hi >> >> i am quite new, so i’m sorry if i’m getting things wrong. >> >> I am trying to simulate a turbulent channel flow, with a passive scalar. >> in the streamwise direction, i would like to impose periodic boundary conditions for the flow, and non-periodic for the scalar. >> in fact the scalar has a point source near the inlet, and i want to see how that develops downstream. >> >> as far as i understand this is not supported by nek5000, because the last mesh point is identified with the first mesh point when periodic BC are imposed. >> correct ? >> >> my question is : could i define two meshes, one for the scalar and one for the velocity ? >> >> thanks a lot for your help !!! >> agnese >> >> >> Agnese Seminara >> -------------------------------- >> CNRS >> Laboratoire de physique de la matière condensée >> Parc Valrose >> avenue J Vallot >> 06108 Nice, France >> +33 (0) 492 076 775 >> http://sites.unice.fr/site/aseminara/ <http://sites.unice.fr/site/aseminara/> >> >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]> >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users <https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users> > <t.tgz> > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users(a)lists.mcs.anl.gov <mailto:[email protected]> > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users <https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users>

1 0

Exact restart
by nek5000-users＠lists.mcs.anl.gov 24 Apr '17

24 Apr '17

Hi Neks, May be my question is already asked, but I can't find the answer so I asked here: can I restart a simulation in an exact state as if it is run continuously? I have simulation that is extremely sensible to the initial condition and the simulation that is spitted in 2 separate runs give a completely difference solution as if it is run 1 single time. Thank you in advance, Quan

1 1

Outflow reflections - Non reflective BCs
by nek5000-users＠lists.mcs.anl.gov 23 Apr '17

23 Apr '17

Hi Neks, I am running a 2D axisymmetric case of a jet and I am having some problems with the outflow boundary. More specifically, whenever a vortex hits that boundary I get reflections. I have the ON condition along with turb_outflow with rq=10 (as high as it will go without crashing). I've also read something about non-reflective BCs in the archive, but it doesn't seem like they were finally implemented. Is there any way I could find them? If not, is there any other easier way to solve this? Thanks in advance for the help, MViturro

1 0

Robin boundary condition on velocity
by nek5000-users＠lists.mcs.anl.gov 22 Apr '17

22 Apr '17

Hi Neks I'm trying to implement a slip-length (Robin) as boundary condition for a plane channel flow, namely: (u+du/dy)_{wall}=0 where x and y are the streamwise and wall normal directions respectively. Following the latter post: http://lists.mcs.anl.gov/pipermail/nek5000-users/2011-April/001313.html here there's what I came up with: c----------------------------------------------------------------------- subroutine userchk include 'SIZE' include 'TOTAL' common /my_DEBUG/ rmodulate common /my_ROBIN/ ROBIN_mask(lx1,ly1,lz1,lelt) $ , work1(lx1,ly1,lz1,lelt) $ , work2(lx1,ly1,lz1,lelt) $ , graduy(lx1,ly1,lz1,lelt) n = nx1*ny1*nz1*nelv !!!Build Up MASKS! if (istep.eq.0) then do i=1,n x=xm1(i,1,1,1) y=ym1(i,1,1,1) if(x.gt.3.and.x.lt.4)then ROBIN_mask(i,1,1,1)=1.0 else ROBIN_mask(i,1,1,1)=0.0 endif enddo endif !Compute velocity gradient call rzero(graduy,n) call gradm1(work1,graduy,work2,vx) call dsavg(graduy) ! call outpost(work1,graduy,work2,pr,t,'GRD') !Da aggiungere l'eventualità del 3D... rmodulate = 0.5*(tanh(0.25*time-10)+1) if (nid.eq.0) write(6,*)"MODULATE", rmodulate,time return end c----------------------------------------------------------------------- c----------------------------------------------------------------------- subroutine userbc (ix,iy,iz,iside,eg) c NOTE ::: This subroutine MAY NOT be called by every process include 'SIZE' include 'TOTAL' include 'NEKUSE' integer e, eg !Imposing Fancy Boundary Conditions... ![Nek5000-users] Robin boundary conditions for the velocity common /my_ROBIN/ ROBIN_mask(lx1,ly1,lz1,lelt) $ , work1(lx1,ly1,lz1,lelt) $ , work2(lx1,ly1,lz1,lelt) $ , graduy(lx1,ly1,lz1,lelt) common /my_DEBUG/ rmodulate e=gllel(eg) ! global element number to processor-local el. # ux=1*graduy(ix,iy,iz,e)*rmodulate uy=0.0 uz=0.0 return end c----------------------------------------------------------------------- subroutine useric (ix,iy,iz,ieg) include 'SIZE' include 'TOTAL' include 'NEKUSE' ux=1.0*y**2 uy=0.0 uz=0.0 temp=0 return end c----------------------------------------------------------------------- The code becomes unstable in the gradient computation due to some point-to-point oscillations. As you can see I've also tried to use a ramp function in order to make the onset of Robin condition more gentle, but at a given "slip-length amplitude" oscillations deteriorates the solution. I use ROBIN_mask in order to chose where to have such BC, but it seems that I'm still missing something in computing the velocity gradient GRD. By the way, this is my box.BOX file: base.rea 2 spatial dimension ( < 0 --> generate .rea/.re2 pair) 1 number of fields #======================================================================= # # Example of .box file for channel flow # # If nelx (y or z) < 0, then genbox automatically generates the # grid spacing in the x (y or z) direction # with a geometric ratio given by "ratio". # ( ratio=1 implies uniform spacing ) # # Note that the character bcs _must_ have 3 spaces. # #======================================================================= # Box -32 -16 nelx,nely,nelz for Box 0 6.2832 1. x0,x1,gain -1 1 1. y0,y1,gain P ,P ,v ,v , , bc's (3 chars each!) I was wondering if you might be able to give me some advice. Best regards, Francesco

1 1

How to print out time derivatives of scalars in a flowfield
by nek5000-users＠lists.mcs.anl.gov 20 Apr '17

20 Apr '17

Hello Neks, I am working on Rayleigh Benard convection in a cylindrical domain, and I have some reactive scalars advecting with the flowfield. I need to find out a way to print the values of the temporal derivative of these scalars at each time step. At this point, I get .f files which have the information of velocity, temperature and the scalar field (c). I need to find out a way to include the information for the time derivative of the scalar filed (dc/dt). How can this be done? Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech.

1 0