From nek5000-users at lists.mcs.anl.gov Wed Feb 2 08:35:58 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 02 Feb 2011 09:35:58 -0500 Subject: [Nek5000-users] turbulent channel flow In-Reply-To: <1692489571.89628.1295632846968.JavaMail.root@zimbra.anl.gov> References: <1692489571.89628.1295632846968.JavaMail.root@zimbra.anl.gov> Message-ID: <1296657358.21480.13.camel@alex-desktop> Hi, I am trying to reproduce the results for the turbulent channel flow using DNS. However, instead of using the bulk velocity or the flow rate to drive the flow, I would like to use a mean pressure gradient. Using the non-dimensional equations resulting from selecting the friction velocity (u_tau) as the velocity scale, only the Reynolds number based on u_tau needs to be specified because the mean pressure gradient is equal to -1. Consequently the pressure gradient in the streamwise direction is grad p=-1+grad p* where * means fluctuation. I assumed that this could be done by simply setting: ffx=1.0 ffy=0.0 ffz=0.0 in subroutine userf. All my trials initializing the flow with some kind of synthetic turbulent flow ended up damping the fluctuation and converging towards the laminar solution. Any help?. -- Alexandre Fabregat, PhD. Mail: afabrega at uottawa.ca Phone: 630 562-5800 (ext. 6293) Office: B207 Mechanical Engineering Department University of Ottawa 161 Louis Pasteur, K1N6N5, Ottawa ON From nek5000-users at lists.mcs.anl.gov Wed Feb 2 08:49:19 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 02 Feb 2011 15:49:19 +0100 Subject: [Nek5000-users] turbulent channel flow In-Reply-To: <1296657358.21480.13.camel@alex-desktop> References: <1692489571.89628.1295632846968.JavaMail.root@zimbra.anl.gov> <1296657358.21480.13.camel@alex-desktop> Message-ID: <4D496EEF.3050107@mech.kth.se> Hi Alex, You could do like this: subroutine userf (ix,iy,iz,ieg) include 'SIZE' include 'TOTAL' include 'NEKUSE' real retau c visc=param(2) c retau = 180 ffx = (retau*visc)**2 ffz = 0.0 ffy = 0.0 return end /Johan nek5000-users at lists.mcs.anl.gov wrote: > Hi, > > I am trying to reproduce the results for the turbulent channel flow > using DNS. > > However, instead of using the bulk velocity or the flow rate to drive > the flow, I would like to use a mean pressure gradient. > > Using the non-dimensional equations resulting from selecting the > friction velocity (u_tau) as the velocity scale, only the Reynolds > number based on u_tau needs to be specified because the mean pressure > gradient is equal to -1. > Consequently the pressure gradient in the streamwise direction is grad > p=-1+grad p* where * means fluctuation. > > I assumed that this could be done by simply setting: > ffx=1.0 > ffy=0.0 > ffz=0.0 > in subroutine userf. > > All my trials initializing the flow with some kind of synthetic > turbulent flow ended up damping the fluctuation and converging towards > the laminar solution. > > Any help?. > > -- Johan Malm Department of Mechanics, KTH SE-100 44, Stockholm, Sweden Phone: +46 8 7906876 E-mail: johan at mech.kth.se From nek5000-users at lists.mcs.anl.gov Sun Feb 6 18:28:00 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Sun, 6 Feb 2011 18:28:00 -0600 Subject: [Nek5000-users] Visualize streamlines in Visit Message-ID: Neks, This seems like a simple question, but I searched through the archives and didn't see anything related to this. Are there certain steps one must take (either in Nek or in Visit) to visualize streamlines in Visit? In particular, I am wanting to do streamlines in a 2D slice. If I just add "streamline-velocity" nothing shows up (whether it's a slice or not). Thank you, Josh -- Josh Camp "All that is necessary for the triumph of evil is that good men do nothing" -- Edmund Burke From nek5000-users at lists.mcs.anl.gov Sun Feb 6 20:03:20 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Sun, 6 Feb 2011 20:03:20 -0600 (CST) Subject: [Nek5000-users] Visualize streamlines in Visit In-Reply-To: References: Message-ID: Josh, The way I would do this would be to zero out the velocity component normal to the plane of interest, assuming that's easily done. There are no doubt other options as well. Paul On Sun, 6 Feb 2011, nek5000-users at lists.mcs.anl.gov wrote: > Neks, > > This seems like a simple question, but I searched through the archives > and didn't see anything related to this. Are there certain steps one > must take (either in Nek or in Visit) to visualize streamlines in > Visit? In particular, I am wanting to do streamlines in a 2D slice. > If I just add "streamline-velocity" nothing shows up (whether it's a > slice or not). > > Thank you, > Josh > > -- > Josh Camp > > "All that is necessary for the triumph of evil is that good men do > nothing" -- Edmund Burke > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > From nek5000-users at lists.mcs.anl.gov Sun Feb 6 23:05:21 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Sun, 6 Feb 2011 21:05:21 -0800 Subject: [Nek5000-users] Visualize streamlines in Visit In-Reply-To: References: Message-ID: Hi Josh, If you project the slice to 2D dimensions, it should work. If you leave it in 3D, the vector field is still three dimensional and it will project out of the plane. If you are in 2D space, set me up with your data offline ( hchilds at lbl.gov) and I can take a look. Best, Hank On Sun, Feb 6, 2011 at 4:28 PM, wrote: > Neks, > > This seems like a simple question, but I searched through the archives > and didn't see anything related to this. Are there certain steps one > must take (either in Nek or in Visit) to visualize streamlines in > Visit? In particular, I am wanting to do streamlines in a 2D slice. > If I just add "streamline-velocity" nothing shows up (whether it's a > slice or not). > > Thank you, > Josh > > -- > Josh Camp > > "All that is necessary for the triumph of evil is that good men do > nothing" -- Edmund Burke > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Mon Feb 7 17:50:24 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 7 Feb 2011 17:50:24 -0600 Subject: [Nek5000-users] Visualize streamlines in Visit In-Reply-To: References: Message-ID: Thanks Paul and Hank. It seems the problem was my lack of understanding with the streamline tool. I think I have figured it out now. Thanks, Josh On Sun, Feb 6, 2011 at 11:05 PM, wrote: > Hi Josh, > If you project the slice to 2D dimensions, it should work. ?If you leave it > in 3D, the vector field is still three dimensional and it will project out > of the plane. ?If you are in 2D space, set me up with your data offline > (hchilds at lbl.gov) and I can take a look. > Best, > Hank > > On Sun, Feb 6, 2011 at 4:28 PM, wrote: >> >> Neks, >> >> This seems like a simple question, but I searched through the archives >> and didn't see anything related to this. ?Are there certain steps one >> must take (either in Nek or in Visit) to visualize streamlines in >> Visit? ?In particular, I am wanting to do streamlines in a 2D slice. >> If I just add "streamline-velocity" nothing shows up (whether it's a >> slice or not). >> >> Thank you, >> Josh >> >> -- >> Josh Camp >> >> "All that is necessary for the triumph of evil is that good men do >> nothing" -- Edmund Burke >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > > -- Josh Camp "All that is necessary for the triumph of evil is that good men do nothing" -- Edmund Burke From nek5000-users at lists.mcs.anl.gov Tue Feb 8 16:02:56 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 8 Feb 2011 14:02:56 -0800 Subject: [Nek5000-users] Importing Files into Visit Message-ID: Hello, I increased the number of time steps to 100000, and now the .nek5000 file will not load into Visit after running the "visnek" script. Is there a limit to the number of .fld files or am I making a mistake? Thanks, John From nek5000-users at lists.mcs.anl.gov Tue Feb 8 16:04:20 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 8 Feb 2011 16:04:20 -0600 (CST) Subject: [Nek5000-users] Importing Files into Visit In-Reply-To: References: Message-ID: How many .fld files do you have ? Paul On Tue, 8 Feb 2011, nek5000-users at lists.mcs.anl.gov wrote: > Hello, > > I increased the number of time steps to 100000, and now the .nek5000 > file will not load into Visit after running the "visnek" script. Is > there a limit to the number of .fld files or am I making a mistake? > > Thanks, > > John > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > From nek5000-users at lists.mcs.anl.gov Tue Feb 8 16:05:13 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 8 Feb 2011 14:05:13 -0800 Subject: [Nek5000-users] Importing Files into Visit In-Reply-To: References: Message-ID: Hi John, Not to my knowledge. Can you send me the .nek5000 file (i.e. to hchilds at lbl.gov) so I can take a look? Best, Hank On Tue, Feb 8, 2011 at 2:02 PM, wrote: > Hello, > > I increased the number of time steps to 100000, and now the .nek5000 > file will not load into Visit after running the "visnek" script. Is > there a limit to the number of .fld files or am I making a mistake? > > Thanks, > > John > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Tue Feb 8 22:54:37 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 8 Feb 2011 22:54:37 -0600 (CST) Subject: [Nek5000-users] Importing Files into Visit In-Reply-To: Message-ID: <1565562932.34815.1297227277599.JavaMail.root@zimbra.anl.gov> Hi Josh, The visnek script will not handle 100,000 .f files -- actually the maximum number of .f files generated by nek5000 can not exceed 99,999. Good question whether visnek will create the right VisIt metafile .nek5000 for 100,000 .fld files -- I never tested it at this scale. But in both cases you can modify the metafile to your needs. Let me know if you need more details. Best, Aleks ----- Original Message ----- From: nek5000-users at lists.mcs.anl.gov To: nek5000-users at lists.mcs.anl.gov Sent: Tuesday, February 8, 2011 4:02:56 PM Subject: [Nek5000-users] Importing Files into Visit Hello, I increased the number of time steps to 100000, and now the .nek5000 file will not load into Visit after running the "visnek" script. Is there a limit to the number of .fld files or am I making a mistake? Thanks, John _______________________________________________ Nek5000-users mailing list Nek5000-users at lists.mcs.anl.gov https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users From nek5000-users at lists.mcs.anl.gov Wed Feb 9 12:24:32 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 9 Feb 2011 10:24:32 -0800 Subject: [Nek5000-users] Importing Files into Visit Message-ID: I have only 200 .fld files. I had also tried a run with .f files and still could not load the .nek5000 file into Visit. Thanks, John From nek5000-users at lists.mcs.anl.gov Thu Feb 10 05:16:34 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 10 Feb 2011 12:16:34 +0100 Subject: [Nek5000-users] Perturbation mode tutorial? Message-ID: Hi Nek's, I would like to start using the perturbation mode for a fully three-dimensional case, however I have no idea about how it works. What I have understood so far is that I've got to change the lp-parameters in my SIZE file. Apart from these changes, what else do I have to do? Do I specify my input file containing my base flow as in a classical restart procedure or elsewhere? Best regards, -- Jean-Christophe -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Thu Feb 10 10:52:10 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 10 Feb 2011 10:52:10 -0600 (CST) Subject: [Nek5000-users] Controlling BL parameters Message-ID: <1911260317.8183031297356730366.JavaMail.root@neo-mail-3> Hello, I am curious if any one has found a good technique for controlling/specifying desired boundary layer parameters in NEK ? The problem involves turbulent flow over a flat plate with a developing boundary layer in which the displacement thickness, and momentum thickness need to be matched at a specified location. I have not found an efficient way to control this in NEK . Any help or suggestions is greatly appreciated. -Michael M. -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Thu Feb 10 14:48:31 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 10 Feb 2011 12:48:31 -0800 Subject: [Nek5000-users] Visnek Script Message-ID: Hello, I am working on a new machine and saved the link to the visnek script in the bin directory. However, I am now getting an error when trying to execute visnek (it was working on the other machine): bash: /bin/tcsh: bad interpreter: No such file or directory I also tried copying the script into my working directory and deleting the top line, but I was getting more errors. Do you have any suggestions on how to fix this problem? Thanks, John From nek5000-users at lists.mcs.anl.gov Thu Feb 10 14:58:04 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 10 Feb 2011 14:58:04 -0600 (CST) Subject: [Nek5000-users] Visnek Script In-Reply-To: Message-ID: <1065908842.42688.1297371484618.JavaMail.root@zimbra.anl.gov> Hi John, Trying looking for tcsh by running which tcsh and put this path instead. Another alternative is to talk to your system guys to find out where tcsh is. Let me know how it goes. Best, Aleks ----- Original Message ----- From: nek5000-users at lists.mcs.anl.gov To: nek5000-users at lists.mcs.anl.gov Sent: Thursday, February 10, 2011 2:48:31 PM Subject: [Nek5000-users] Visnek Script Hello, I am working on a new machine and saved the link to the visnek script in the bin directory. However, I am now getting an error when trying to execute visnek (it was working on the other machine): bash: /bin/tcsh: bad interpreter: No such file or directory I also tried copying the script into my working directory and deleting the top line, but I was getting more errors. Do you have any suggestions on how to fix this problem? Thanks, John _______________________________________________ Nek5000-users mailing list Nek5000-users at lists.mcs.anl.gov https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users From nek5000-users at lists.mcs.anl.gov Thu Feb 10 15:06:42 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 10 Feb 2011 16:06:42 -0500 Subject: [Nek5000-users] Visnek Script In-Reply-To: References: Message-ID: <1297372003.5697.8.camel@alex-desktop> Hi, You just need to install the cshell (csh). From nek5000-users at lists.mcs.anl.gov Tue Feb 15 11:24:27 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 15 Feb 2011 17:24:27 +0000 Subject: [Nek5000-users] Mesh for a cylinder inside a rectangular domain Message-ID: Hi I am a master student in Linne flow center at KTH University. I am going to simulate a cylinder flow case with Nek5000. The geometry consists of a small circular cylinder inside a rectangular domain. I need to setup a mesh between the cylinder and the rectangle which is clustered near the cylinder. Would you please help me for the mesh? Regards Iman -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Tue Feb 15 11:42:34 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 15 Feb 2011 11:42:34 -0600 (CST) Subject: [Nek5000-users] Mesh for a cylinder inside a rectangular domain In-Reply-To: References: Message-ID: Hi Iman, Is this a 2D or 3D case? What Reynolds number? What diameter? Paul On Tue, 15 Feb 2011, nek5000-users at lists.mcs.anl.gov wrote: > Hi > > I am a master student in Linne flow center at KTH University. I am going to > simulate a cylinder flow case with Nek5000. The geometry consists of a small > circular cylinder inside a rectangular domain. I need to setup a mesh between > the cylinder and the rectangle which is clustered near the cylinder. Would > you please help me for the mesh? > > Regards > Iman > From nek5000-users at lists.mcs.anl.gov Tue Feb 15 11:47:13 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 15 Feb 2011 17:47:13 +0000 Subject: [Nek5000-users] Mesh for a cylinder inside a rectangular domain In-Reply-To: References: , Message-ID: Dear Paul It is a 2D case. The cylinder diameter is equal to 1. The rectangle size is x=50 * y=30. the center of cylinder in placed at (15,15) I would like to work in the range of Reynolds number less than the critical Reynolds number "47". Regards Iman ________________________________________ From: nek5000-users-bounces at lists.mcs.anl.gov [nek5000-users-bounces at lists.mcs.anl.gov] on behalf of nek5000-users at lists.mcs.anl.gov [nek5000-users at lists.mcs.anl.gov] Sent: Tuesday, February 15, 2011 6:42 PM To: nek5000-users at lists.mcs.anl.gov Subject: Re: [Nek5000-users] Mesh for a cylinder inside a rectangular domain Hi Iman, Is this a 2D or 3D case? What Reynolds number? What diameter? Paul On Tue, 15 Feb 2011, nek5000-users at lists.mcs.anl.gov wrote: > Hi > > I am a master student in Linne flow center at KTH University. I am going to > simulate a cylinder flow case with Nek5000. The geometry consists of a small > circular cylinder inside a rectangular domain. I need to setup a mesh between > the cylinder and the rectangle which is clustered near the cylinder. Would > you please help me for the mesh? > > Regards > Iman > _______________________________________________ Nek5000-users mailing list Nek5000-users at lists.mcs.anl.gov https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users From nek5000-users at lists.mcs.anl.gov Tue Feb 15 23:06:16 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 15 Feb 2011 23:06:16 -0600 (CST) Subject: [Nek5000-users] Mesh for a cylinder inside a rectangular domain In-Reply-To: References: Message-ID: Hi, I've added a 2D example for flow past a cylinder to the examples directory, under ext_cyl. Hopefully this will get you started. I'm working on additional documenation for prenek. Paul On Tue, 15 Feb 2011, nek5000-users at lists.mcs.anl.gov wrote: > Hi > > I am a master student in Linne flow center at KTH University. I am going to > simulate a cylinder flow case with Nek5000. The geometry consists of a small > circular cylinder inside a rectangular domain. I need to setup a mesh between > the cylinder and the rectangle which is clustered near the cylinder. Would > you please help me for the mesh? > > Regards > Iman > From nek5000-users at lists.mcs.anl.gov Wed Feb 16 05:18:13 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 16 Feb 2011 12:18:13 +0100 Subject: [Nek5000-users] Perturbation mode Message-ID: Hi Nek's, I am still struggling with the perturbation mode. Cavity_20x20x10_6x6x6_R1000_25D0.f00001 contains my base flow. It is a 2D lid-driven cavity extruded in the third direction (z). Perturbation_20x20x10_6x6x6_25D0.f00001 contains my divergence-free perturbation. It is subject to no-slip conditions at the walls and periodic BC at zmin and zmax. Following is how I set it up. ********************************************** In my blah.rea -1. p31 NPERT: #perturbation modes [...] 2 PRESOLVE/RESTART OPTIONS ***** Cavity_20x20x10_6x6x6_R1000_25D0.f00001 Perturbation_20x20x10_6x6x6_25D0.f00001 ********************************************** In SIZE: parameter (lpelv=lelv,lpelt=lelt,lpert=1) ! perturbation parameter (lpx1=lx1,lpy1=ly1,lpz1=lz1) ! array sizes parameter (lpx2=lx2,lpy2=ly2,lpz2=lz2) ********************************************** I have tried with various initial amplitude for the perturbation, and various time steps. However, I always have a very small CFL and after six timesteps I get: Failed in HMHOLTZ: VELX NAN > Same for vely and velz. I may have missed a key point but being not particularly familiar with Nek numerics yet I'm not sure where I missed it. Best regards, -- Jean-Christophe -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Feb 16 05:40:40 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 16 Feb 2011 12:40:40 +0100 Subject: [Nek5000-users] Perturbation mode In-Reply-To: References: Message-ID: <4D5BB7B8.1020800@mech.kth.se> Hi Jean-Christophe, One thing we found a couple of months ago is that the pertubation mode is unstable with time scheme of 3rd order. Try to switch to 2nd order, change in the rea file line 31 (TORDER) Regards, Antonios nek5000-users at lists.mcs.anl.gov wrote: > Hi Nek's, > > I am still struggling with the perturbation mode. > Cavity_20x20x10_6x6x6_R1000_25D0.f00001 contains my base flow. It is a > 2D lid-driven cavity extruded in the third direction (z). > Perturbation_20x20x10_6x6x6_25D0.f00001 contains my divergence-free > perturbation. It is subject to no-slip conditions at the walls and > periodic BC at zmin and zmax. > > Following is how I set it up. > > ********************************************** > > In my blah.rea > > -1. p31 NPERT: #perturbation modes > > [...] > > 2 PRESOLVE/RESTART OPTIONS ***** > Cavity_20x20x10_6x6x6_R1000_25D0.f00001 > Perturbation_20x20x10_6x6x6_25D0.f00001 > > ********************************************** > > In SIZE: > > parameter (lpelv=lelv,lpelt=lelt,lpert=1) ! perturbation > parameter (lpx1=lx1,lpy1=ly1,lpz1=lz1) ! array sizes > parameter (lpx2=lx2,lpy2=ly2,lpz2=lz2) > > ********************************************** > > I have tried with various initial amplitude for the perturbation, and > various time steps. However, I always have a very small CFL and after > six timesteps I get: > > Failed in HMHOLTZ: VELX NAN > > > Same for vely and velz. I may have missed a key point but being not > particularly familiar with Nek numerics yet I'm not sure where I > missed it. > > Best regards, > -- > Jean-Christophe > ------------------------------------------------------------------------ > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > From nek5000-users at lists.mcs.anl.gov Wed Feb 16 06:08:09 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 16 Feb 2011 13:08:09 +0100 Subject: [Nek5000-users] Perturbation mode In-Reply-To: References: Message-ID: <4D5BBE29.70904@mech.kth.se> Hi Jean-Christophe, I think the restart option does not work properly in the perturbation mode. That means, in the rea-file you can just give the base flow under RESOLVE/RESTART OPTIONS. I usually do the restart of the perturbation manually in the user-file. Another way is to change that in the source code. David nek5000-users at lists.mcs.anl.gov wrote: > Hi Nek's, > > I am still struggling with the perturbation mode. > Cavity_20x20x10_6x6x6_R1000_25D0.f00001 contains my base flow. It is a > 2D lid-driven cavity extruded in the third direction (z). > Perturbation_20x20x10_6x6x6_25D0.f00001 contains my divergence-free > perturbation. It is subject to no-slip conditions at the walls and > periodic BC at zmin and zmax. > > Following is how I set it up. > > ********************************************** > > In my blah.rea > > -1. p31 NPERT: #perturbation modes > > [...] > > 2 PRESOLVE/RESTART OPTIONS ***** > Cavity_20x20x10_6x6x6_R1000_25D0.f00001 > Perturbation_20x20x10_6x6x6_25D0.f00001 > > ********************************************** > > In SIZE: > > parameter (lpelv=lelv,lpelt=lelt,lpert=1) ! perturbation > parameter (lpx1=lx1,lpy1=ly1,lpz1=lz1) ! array sizes > parameter (lpx2=lx2,lpy2=ly2,lpz2=lz2) > > ********************************************** > > I have tried with various initial amplitude for the perturbation, and > various time steps. However, I always have a very small CFL and after > six timesteps I get: > > Failed in HMHOLTZ: VELX NAN > > > Same for vely and velz. I may have missed a key point but being not > particularly familiar with Nek numerics yet I'm not sure where I > missed it. > > Best regards, > -- > Jean-Christophe > ------------------------------------------------------------------------ > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > -- David Tempelmann Linn? Flow Center, Mechanics KTH SE-100 44, Stockholm, Sweden Phone: +46 8 7907161 E-mail: david at mech.kth.se From nek5000-users at lists.mcs.anl.gov Wed Feb 16 06:30:19 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 16 Feb 2011 06:30:19 -0600 (CST) Subject: [Nek5000-users] Perturbation mode In-Reply-To: <4D5BBE29.70904@mech.kth.se> References: <4D5BBE29.70904@mech.kth.se> Message-ID: Hi JC, There may also be an issue with having the projection turned on. (This is a common source of NaN...). I suggest setting p93-95 to 0 in the .rea file. (We can subsequently fix the problem correctly.) Are you trying to run with a steady-state base flow, or an evolving base flow? I'm out of the office today but can hopefully look into this in the next day or two. To date I've not done a lot of perturbation cases, so there are likely a few issues to resolve. Several groups have had some success, however. Regards, Paul On Wed, 16 Feb 2011, nek5000-users at lists.mcs.anl.gov wrote: > Hi Jean-Christophe, > > I think the restart option does not work properly in the perturbation > mode. That means, in the rea-file you can just give the base flow under > RESOLVE/RESTART OPTIONS. > I usually do the restart of the perturbation manually in the user-file. > Another way is to change that in the source code. > > David > > nek5000-users at lists.mcs.anl.gov wrote: >> Hi Nek's, >> >> I am still struggling with the perturbation mode. >> Cavity_20x20x10_6x6x6_R1000_25D0.f00001 contains my base flow. It is a >> 2D lid-driven cavity extruded in the third direction (z). >> Perturbation_20x20x10_6x6x6_25D0.f00001 contains my divergence-free >> perturbation. It is subject to no-slip conditions at the walls and >> periodic BC at zmin and zmax. >> >> Following is how I set it up. >> >> ********************************************** >> >> In my blah.rea >> >> -1. p31 NPERT: #perturbation modes >> >> [...] >> >> 2 PRESOLVE/RESTART OPTIONS ***** >> Cavity_20x20x10_6x6x6_R1000_25D0.f00001 >> Perturbation_20x20x10_6x6x6_25D0.f00001 >> >> ********************************************** >> >> In SIZE: >> >> parameter (lpelv=lelv,lpelt=lelt,lpert=1) ! perturbation >> parameter (lpx1=lx1,lpy1=ly1,lpz1=lz1) ! array sizes >> parameter (lpx2=lx2,lpy2=ly2,lpz2=lz2) >> >> ********************************************** >> >> I have tried with various initial amplitude for the perturbation, and >> various time steps. However, I always have a very small CFL and after >> six timesteps I get: >> >> Failed in HMHOLTZ: VELX NAN >> >> >> Same for vely and velz. I may have missed a key point but being not >> particularly familiar with Nek numerics yet I'm not sure where I >> missed it. >> >> Best regards, >> -- >> Jean-Christophe >> ------------------------------------------------------------------------ >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >> > > > -- > David Tempelmann > Linn? Flow Center, Mechanics KTH > SE-100 44, Stockholm, Sweden > Phone: +46 8 7907161 > E-mail: david at mech.kth.se > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > From nek5000-users at lists.mcs.anl.gov Wed Feb 16 06:57:00 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 16 Feb 2011 06:57:00 -0600 (CST) Subject: [Nek5000-users] UPDATE: correction to midside node support Message-ID: Dear Group, In the course of preparing some recent runs, I've discovered a flaw in the midside node (curve side defn: 'm') support, that may or may not be an issue for existing cases. In my example, it was manifest only at certain processor counts because of the way that certain arrays interacted. I have implemented a fix that should be sufficient for any existing cases, and will develop a cleaner fix in the next week or two. At present, the error manifests itself as "Vanishing Jacobian" I recommend updating to the current repo version to fix this. Regards, Paul From nek5000-users at lists.mcs.anl.gov Wed Feb 16 07:02:04 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 16 Feb 2011 07:02:04 -0600 (CST) Subject: [Nek5000-users] Perturbation mode In-Reply-To: References: <4D5BBE29.70904@mech.kth.se> Message-ID: PS -- It appears that p93 = 20 (and mxprev=20 in SIZE) p94 = 0 p95 = 5 (say) is a good combination -- this still allows projection of the pressure onto the prior solutions, which gives important savings in many cases. On Wed, 16 Feb 2011, nek5000-users at lists.mcs.anl.gov wrote: > > Hi JC, > > There may also be an issue with having the projection turned on. > (This is a common source of NaN...). I suggest setting p93-95 to 0 > in the .rea file. (We can subsequently fix the problem correctly.) > > Are you trying to run with a steady-state base flow, or an evolving > base flow? > > I'm out of the office today but can hopefully look into this in > the next day or two. To date I've not done a lot of perturbation > cases, so there are likely a few issues to resolve. Several groups > have had some success, however. > > Regards, > > Paul > > > On Wed, 16 Feb 2011, nek5000-users at lists.mcs.anl.gov wrote: > >> Hi Jean-Christophe, >> >> I think the restart option does not work properly in the perturbation >> mode. That means, in the rea-file you can just give the base flow under >> RESOLVE/RESTART OPTIONS. >> I usually do the restart of the perturbation manually in the user-file. >> Another way is to change that in the source code. >> >> David >> >> nek5000-users at lists.mcs.anl.gov wrote: >>> Hi Nek's, >>> >>> I am still struggling with the perturbation mode. >>> Cavity_20x20x10_6x6x6_R1000_25D0.f00001 contains my base flow. It is a >>> 2D lid-driven cavity extruded in the third direction (z). >>> Perturbation_20x20x10_6x6x6_25D0.f00001 contains my divergence-free >>> perturbation. It is subject to no-slip conditions at the walls and >>> periodic BC at zmin and zmax. >>> >>> Following is how I set it up. >>> >>> ********************************************** >>> >>> In my blah.rea >>> >>> -1. p31 NPERT: #perturbation modes >>> >>> [...] >>> >>> 2 PRESOLVE/RESTART OPTIONS ***** >>> Cavity_20x20x10_6x6x6_R1000_25D0.f00001 >>> Perturbation_20x20x10_6x6x6_25D0.f00001 >>> >>> ********************************************** >>> >>> In SIZE: >>> >>> parameter (lpelv=lelv,lpelt=lelt,lpert=1) ! perturbation >>> parameter (lpx1=lx1,lpy1=ly1,lpz1=lz1) ! array sizes >>> parameter (lpx2=lx2,lpy2=ly2,lpz2=lz2) >>> >>> ********************************************** >>> >>> I have tried with various initial amplitude for the perturbation, and >>> various time steps. However, I always have a very small CFL and after >>> six timesteps I get: >>> >>> Failed in HMHOLTZ: VELX NAN >>> >>> >>> Same for vely and velz. I may have missed a key point but being not >>> particularly familiar with Nek numerics yet I'm not sure where I >>> missed it. >>> >>> Best regards, >>> -- >>> Jean-Christophe >>> ------------------------------------------------------------------------ >>> >>> _______________________________________________ >>> Nek5000-users mailing list >>> Nek5000-users at lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>> >> >> >> -- >> David Tempelmann >> Linn? Flow Center, Mechanics KTH >> SE-100 44, Stockholm, Sweden >> Phone: +46 8 7907161 >> E-mail: david at mech.kth.se >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > From nek5000-users at lists.mcs.anl.gov Thu Feb 17 09:20:07 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 17 Feb 2011 09:20:07 -0600 Subject: [Nek5000-users] Shutdown after "Failed in HMHOLTZ" Message-ID: Neks, Is there a simple way to make Nekton quit if the simulation blows up? I know Nek will quit if the CFL gets too high, but it doesn't seem to want to quit something get messed up in the Helmholtz or Pressure iterations (resulting in NaNs). I am submitting my jobs to a cluster, and I may not be able to babysit the simulation all the time (particularly if my turn in the queue comes up late at night). Therefore, hours can get wasted. Any help would be appreciated! -- Josh Camp "All that is necessary for the triumph of evil is that good men do nothing" -- Edmund Burke From nek5000-users at lists.mcs.anl.gov Thu Feb 17 09:30:29 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 17 Feb 2011 09:30:29 -0600 (CST) Subject: [Nek5000-users] =?utf-8?q?Shutdown_after_=22Failed_in_HMHOLTZ=22?= In-Reply-To: References: Message-ID: Hi Johs, It depends on your compiler. For my favorite, pg, I set the flag in makenek: G="-Ktrap=fp -mcmodel=medium" The -Ktrap=fp flag causes the code to stop at the first inf or NaN. I'm not 100% certain what happens in a parallel environment when only one of the cores encounters the NaN... At least that's a start... perhaps some others have other ideas. Paul On Thu, 17 Feb 2011, nek5000-users at lists.mcs.anl.gov wrote: > Neks, > > Is there a simple way to make Nekton quit if the simulation blows up? > I know Nek will quit if the CFL gets too high, but it doesn't seem to > want to quit something get messed up in the Helmholtz or Pressure > iterations (resulting in NaNs). > > I am submitting my jobs to a cluster, and I may not be able to babysit > the simulation all the time (particularly if my turn in the queue > comes up late at night). Therefore, hours can get wasted. > > Any help would be appreciated! > > -- > Josh Camp > > "All that is necessary for the triumph of evil is that good men do > nothing" -- Edmund Burke > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > From nek5000-users at lists.mcs.anl.gov Fri Feb 18 15:22:01 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 18 Feb 2011 13:22:01 -0800 Subject: [Nek5000-users] Element mapping in Grooved Channel Message-ID: Hello, I am trying to adjust my model from what it is now (see attached gn6f.usr) to a model that has the smaller deformed boxes (grooves) grow until about 24 units in the x-direction while maintaining the overall channel length. I am having difficulty mapping the elements to the grooves. My goal is the following: 1. gdx( 0 ) = 0.6, ght( 0 ) = 0.3 2. Maintain gdx( i ) = 2 * ght( i ) 3. Grow the grooves such that gdx( x = 24) = 2.4 and ght( x = 24 ) = 1.2 4. Each groove thereafter is of same dimension. I estimated a growth factor of 1.08, which will apply to the first 18 grooves. I have tried this but it either won't run or it doesn't come out with the correct geometry. Thanks, John -------------- next part -------------- A non-text attachment was scrubbed... Name: gn6f.box Type: application/octet-stream Size: 1150 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: gn6f.usr Type: application/octet-stream Size: 11478 bytes Desc: not available URL: From nek5000-users at lists.mcs.anl.gov Fri Feb 18 21:07:55 2011 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 18 Feb 2011 21:07:55 -0600 (CST) Subject: [Nek5000-users] Help with planar averaging for dyn smag In-Reply-To: References: <8F6FDFEA-EA5F-4A28-889C-B5805543D874@uic.edu> <380897F2-4C0D-44F0-B89A-462DD0322F75@uic.edu> Message-ID: Hi Harish, I've added a function that will homogenize any function on mesh one (lx1,ly1,lz1) in the z direction, assuming you know how many elements you have in the x-y plane, as would be the case if your mesh is generated using n2to3, say. To use it, you must set lelx and lely in SIZE such that the product lelx*lely >= nelxy. And you similarly need to define nelx and nely values. Here are some sample lines from my userchk routine: call outpost(vx,vy,vz,pr,t,' ') ! Output the orginal field nelx = 8 ! Pick an nelx,nely pair nely = 8 ! such that nelx*nely matches my domain call z_distribute(vx) ! Homogenize vx in the z-direction call z_distribute(vy) ! Homogenize vy in the z-direction call z_distribute(vz) ! Homogenize vz in the z-direction call z_distribute(pr) ! Homogenize pr (only for Pn-Pn!!) call outpost(vx,vy,vz,pr,t,' ') ! output homogenized field and quit call exitti('quit after z-distribute$',nelx) I've committed "z_distribute" to the svn repo. Paul On Mon, 14 Feb 2011, Harish Kanchi wrote: > Hello Paul, > > I have been working on setting up a case for dyn smag. It is the same case > that ran with Pn/Pn and smaller dt. > > It has NEL = 656, made up for 4 blocks. I build the 3D mesh with genbox. > Here is the genbox file used to produce the mesh. > > In my case the homogeneous direction is the z-direction. Here is the > subroutine I am using for planar averaging. But it does not work!! > > Any help would be appreciated. > > Thanks, > > Harish. > > ************************************************************************* > bfs3d.rea > -3 spatial dimension (if negative dump .re2 file) > 1 number of fields > # > #======================================================================= > # > box_1 > -18 -2 -4 nelx,nely,nelz (negative --> auto spacing) > 0.0 20.0 1.085 x_0 x_1 ratio > -1.0 -0.5 2.6 > 0.0 4.0 1 > W ,O ,W ,E ,P ,P bc's ! west,east,south,north,bottom,top > box_2 > -18 -2 -4 nelx,nely,nelz (negative --> auto spacing) > 0.0 20.0 1.085 x_0 x_1 ratio > -0.5 0.0 0.39 > 0.0 4.0 1 > W ,O ,E ,E ,P ,P bc's ! west,east,south,north,bottom,top > box_3 > -18 -4 -4 nelx,nely,nelz (negative --> auto spacing) > 0.0 20.0 1.085 x_0 x_1 ratio > 0.0 5.0 2.8 > 0.0 4.0 1 > E ,O ,E ,SYM,P ,P bc's ! west,east,south,north,bottom,top > box_4 > -5 -4 -4 nelx,nely,nelz (negative --> auto spacing) > -10.0 0.0 0.58 x_0 x_1 ratio > 0.0 5.0 2.8 > 0.0 4.0 1 > v ,E ,W ,SYM,P ,P bc's ! west,east,south,north,bottom,top > ************************************************************************* > > > > ************************************************************************* > c----------------------------------------------------------------------- > subroutine planar_average_t(ua,u,w1,w2) > c > c Compute r-s planar average of quantity u() > c > include 'SIZE' > include 'GEOM' > include 'PARALLEL' > include 'WZ' > include 'ZPER' > > real ua(nz1,nelz),u(nx1,ny1,nx1,nelv),w1(nz1,nelz),w2(nz1,nelz) > integer e,eg,ex,ey,ez > > nz = nz1*nelz > call rzero(ua,nz) > call rzero(w1,nz) > > blk1 = 144 ! 18*2*4 > blk2 = 288 ! 18*2*4 > blk3 = 576 ! 18*4*4 > blk4 = 656 ! 18*4*4 > > do e=1,nelt > eg = lglel(e) > if(eg .le. blk1) then > nelx = 18 > nely = 2 > nelz = 4 > call get_exyz(ex,ey,ez,eg,nelx,nely,nelz) > elseif(eg .gt. blk1 .and. eg .le. blk2) then > nelx = 18 > nely = 2 > nelz = 4 > call get_exyz(ex,ey,ez,eg,nelx,nely,nelz) > elseif(eg .gt. blk2 .and. eg .le. blk3) then > nelx = 18 > nely = 4 > nelz = 4 > call get_exyz(ex,ey,ez,eg,nelx,nely,nelz) > elseif(eg .gt. blk3 .and. eg .le. blk4) then > nelx = 5 > nely = 4 > nelz = 4 > call get_exyz(ex,ey,ez,eg,nelx,nely,nelz) > endif > > do k=1,nz1 > do j=1,ny1 > do i=1,nx1 > zz = (1.-zgm1(k,3))/2. ! = 1 for i=1, = 0 for k=nx1 > aa = zz*area(i,1,5,e) + (1-zz)*area(i,1,6,e) ! wgtd jacobian > w1(k,ez) = w1(k,ez) + aa > ua(k,ez) = ua(k,ez) + aa*u(i,j,k,e) > enddo > enddo > enddo > enddo > call gop(ua,w2,'+ ',nz) > call gop(ua,w2,'+ ',nz) > > do i=1,nz > ua(i,1) = u(a,i)/w1(i,1) ! Normalize > enddo > > return > end > c----------------------------------------------------------------------- > subroutine planar_fill_t(u,ua) > c > c Fill array u with planar values from ua(). > c For tensor-product array of spectral elements > c > include 'SIZE' > include 'GEOM' > include 'PARALLEL' > include 'WZ' > include 'ZPER' > > real u(nx1,ny1,nz1,nelv),ua(lz1,lelz) > integer e,eg,ex,ey,ez > > blk1 = 144 ! 18*2*4 > blk2 = 288 ! 18*2*4 > blk3 = 576 ! 18*4*4 > blk4 = 656 ! 18*4*4 > > do e=1,nelt > eg = lglel(e) > if(eg .le. blk1) then > nelx = 18 > nely = 2 > nelz = 4 > call get_exyz(ex,ey,ez,eg,nelx,nely,nelz) > elseif(eg .gt. blk1 .and. eg .le. blk2) then > nelx = 18 > nely = 2 > nelz = 4 > call get_exyz(ex,ey,ez,eg,nelx,nely,nelz) > elseif(eg .gt. blk2 .and. eg .le. blk3) then > nelx = 18 > nely = 4 > nelz = 4 > call get_exyz(ex,ey,ez,eg,nelx,nely,nelz) > elseif(eg .gt. blk3 .and. eg .le. blk4) then > nelx = 5 > nely = 4 > nelz = 4 > call get_exyz(ex,ey,ez,eg,nelx,nely,nelz) > endif > > do k=1,nz1 > do j=1,ny1 > do i=1,nx1 > u(i,j,k,e) = ua(k,ez) > enddo > enddo > enddo > enddo > > return > end > c----------------------------------------------------------------------- > ************************************************************************* > >