From nek5000-users at lists.mcs.anl.gov Fri Dec 1 07:09:26 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 1 Dec 2017 13:09:26 +0000 Subject: [Nek5000-users] avg_all averaging Message-ID: Dear Ketan, Many thanks for your reply. Basically I want to get the averaged velocity, rms velocities and shear stress where I got two questions. 1. Once the subroutines were writen in usr file, Is calling avg_all still needed? In other words, the post-processing is done after longtime running, and usr file is changed instead of calling avg_all, isn't it? 2. After trying the subroutines you provided, there isnt no files generated in the directory. Can you give me more infomation about the averaging process. Best wishes, ZJ -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Mon Dec 4 13:31:18 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 4 Dec 2017 14:31:18 -0500 Subject: [Nek5000-users] Calculating very small Lewis numbers in Nekton Message-ID: Hello Neks, I am solving a Reaction Advection Diffusion equation, where with my non-dimensionalization, the Lewis Number(Le) is the parameter multiplying the Laplacian. That is, Le is the non-dimensional Diffusion co-efficient of the species. Currently I can do Le = 0.1, but to reduce it further down to 0.01, or 0.001, the spatial resolution has to be increased. I tried doing that increasing the GLL points in SIZE file by increasing the lx1, lxd, but till now I have been unsuccessful. I am aware there is a special filtering that can be used( Fischer and Mullen, 2001). My question is how to implement that in my .usr file. Is there a routine? Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech. -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Mon Dec 4 13:34:44 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 4 Dec 2017 19:34:44 +0000 Subject: [Nek5000-users] Calculating very small Lewis numbers in Nekton In-Reply-To: References: Message-ID: Salkat, could you please be more specific when you write ? I have not been successful?? Do you get an error message? Marco On Dec 4, 2017, at 2:31 PM, nek5000-users at lists.mcs.anl.gov wrote: Hello Neks, I am solving a Reaction Advection Diffusion equation, where with my non-dimensionalization, the Lewis Number(Le) is the parameter multiplying the Laplacian. That is, Le is the non-dimensional Diffusion co-efficient of the species. Currently I can do Le = 0.1, but to reduce it further down to 0.01, or 0.001, the spatial resolution has to be increased. I tried doing that increasing the GLL points in SIZE file by increasing the lx1, lxd, but till now I have been unsuccessful. I am aware there is a special filtering that can be used( Fischer and Mullen, 2001). My question is how to implement that in my .usr file. Is there a routine? Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech. _______________________________________________ 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 Dec 4 13:53:08 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 04 Dec 2017 19:53:08 +0000 Subject: [Nek5000-users] Compiling large problem sizes Message-ID: Hi all, I'm trying to compile a 3D mesh with mesh size of 6000 ekements. Unfortunately I keep running into compilation errors. To be clear I can compile problems upto 3000 elements large. But setting lelt value larger than 3000 results in an error. I have realised this might be a memory issue and have tried to compile with increasing amounts of available memory. This does not seem to resolve the issue. To be clear my SIZE looks like this parameter (ldim=3) ! domain dimension (2 or 3) parameter (lx1=6) ! p-order (avoid uneven and values <6) parameter (lxd=10) ! p-order for over-integration (dealiasing) parameter (lx2=lx1-2) ! p-order for pressure (lx1 or lx1-2) parameter (lelg=6000) ! max total number of elements parameter (lpmin=1) ! min MPI ranks parameter (lpmax=1024) ! max MPI ranks parameter (ldimt=1) ! max auxiliary fields (temperature + scalars) Sincerely, -- *Amitvikram Dutta* Graduate Research Assistant Fluid Mechanics Research Lab University of Waterloo -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Mon Dec 4 13:57:59 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 4 Dec 2017 19:57:59 +0000 Subject: [Nek5000-users] Compiling large problem sizes In-Reply-To: References: Message-ID: What do you do to increase the amount of available memory? In your current setting, Nek5000 compiles with one processor since lpmin is set to 1. You have to increase it to successfully compile. Marco On Dec 4, 2017, at 2:53 PM, nek5000-users at lists.mcs.anl.gov wrote: Hi all, I'm trying to compile a 3D mesh with mesh size of 6000 ekements. Unfortunately I keep running into compilation errors. To be clear I can compile problems upto 3000 elements large. But setting lelt value larger than 3000 results in an error. I have realised this might be a memory issue and have tried to compile with increasing amounts of available memory. This does not seem to resolve the issue. To be clear my SIZE looks like this parameter (ldim=3) ! domain dimension (2 or 3) parameter (lx1=6) ! p-order (avoid uneven and values <6) parameter (lxd=10) ! p-order for over-integration (dealiasing) parameter (lx2=lx1-2) ! p-order for pressure (lx1 or lx1-2) parameter (lelg=6000) ! max total number of elements parameter (lpmin=1) ! min MPI ranks parameter (lpmax=1024) ! max MPI ranks parameter (ldimt=1) ! max auxiliary fields (temperature + scalars) Sincerely, -- Amitvikram Dutta Graduate Research Assistant Fluid Mechanics Research Lab University of Waterloo _______________________________________________ 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 Dec 4 14:08:07 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 4 Dec 2017 15:08:07 -0500 Subject: [Nek5000-users] Calculating very small Lewis numbers in Nekton In-Reply-To: References: Message-ID: Hey Marco, The error message is "Failed in HMHOLTZ" for the scalar and NaN value is being returned. Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech. On Mon, Dec 4, 2017 at 2:34 PM, wrote: > Salkat, > > could you please be more specific when you write ? I have not been > successful?? Do you get an error message? > > Marco > > On Dec 4, 2017, at 2:31 PM, nek5000-users at lists.mcs.anl.gov wrote: > > Hello Neks, > > I am solving a Reaction Advection Diffusion equation, where with my > non-dimensionalization, the Lewis Number(Le) is the parameter multiplying > the Laplacian. That is, Le is the non-dimensional Diffusion co-efficient > of the species. Currently I can do Le = 0.1, but to reduce it further down > to 0.01, or 0.001, the spatial resolution has to be increased. I tried > doing that increasing the GLL points in SIZE file by increasing the lx1, > lxd, but till now I have been unsuccessful. > > I am aware there is a special filtering that can be used( Fischer and > Mullen, 2001). My question is how to implement that in my .usr file. Is > there a routine? > > Thanks, > Saikat > > > Saikat Mukherjee, > PhD Student, > Paul Research Group - http://www.me.vt.edu/mpaul/ > Engineering Science and Mechanics, > Virginia Tech. > _______________________________________________ > 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 > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Mon Dec 4 14:09:35 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 04 Dec 2017 20:09:35 +0000 Subject: [Nek5000-users] Compiling large problem sizes In-Reply-To: References: Message-ID: Ah thanks Marco, changing the lpmin did the trick. On Mon, Dec 4, 2017 at 2:58 PM wrote: > What do you do to increase the amount of available memory? > > In your current setting, Nek5000 compiles with one processor since lpmin > is set to 1. You have to increase it to successfully compile. > > Marco > > On Dec 4, 2017, at 2:53 PM, nek5000-users at lists.mcs.anl.gov wrote: > > Hi all, > > I'm trying to compile a 3D mesh with mesh size of 6000 ekements. > Unfortunately I keep running into compilation errors. To be clear I can > compile problems upto 3000 elements large. But setting lelt value larger > than 3000 results in an error. > > I have realised this might be a memory issue and have tried to compile > with increasing amounts of available memory. This does not seem to resolve > the issue. > > To be clear my SIZE looks like this > > parameter (ldim=3) ! domain dimension (2 or 3) > parameter (lx1=6) ! p-order (avoid uneven and values > <6) > parameter (lxd=10) ! p-order for over-integration > (dealiasing) > parameter (lx2=lx1-2) ! p-order for pressure (lx1 or > lx1-2) > > parameter (lelg=6000) ! max total number of elements > parameter (lpmin=1) ! min MPI ranks > parameter (lpmax=1024) ! max MPI ranks > parameter (ldimt=1) ! max auxiliary fields (temperature > + scalars) > > Sincerely, > > -- > > *Amitvikram Dutta* > > Graduate Research Assistant > > Fluid Mechanics Research Lab > > University of Waterloo > _______________________________________________ > 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 > -- *Amitvikram Dutta* Graduate Research Assistant Fluid Mechanics Research Lab University of Waterloo -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Mon Dec 4 14:13:19 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 4 Dec 2017 20:13:19 +0000 Subject: [Nek5000-users] Calculating very small Lewis numbers in Nekton In-Reply-To: References: Message-ID: How do you set the Lewis number in the advection-diffusion equation? Marco On Dec 4, 2017, at 3:08 PM, nek5000-users at lists.mcs.anl.gov wrote: Hey Marco, The error message is "Failed in HMHOLTZ" for the scalar and NaN value is being returned. Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech. On Mon, Dec 4, 2017 at 2:34 PM, > wrote: Salkat, could you please be more specific when you write ? I have not been successful?? Do you get an error message? Marco On Dec 4, 2017, at 2:31 PM, nek5000-users at lists.mcs.anl.gov wrote: Hello Neks, I am solving a Reaction Advection Diffusion equation, where with my non-dimensionalization, the Lewis Number(Le) is the parameter multiplying the Laplacian. That is, Le is the non-dimensional Diffusion co-efficient of the species. Currently I can do Le = 0.1, but to reduce it further down to 0.01, or 0.001, the spatial resolution has to be increased. I tried doing that increasing the GLL points in SIZE file by increasing the lx1, lxd, but till now I have been unsuccessful. I am aware there is a special filtering that can be used( Fischer and Mullen, 2001). My question is how to implement that in my .usr file. Is there a routine? Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech. _______________________________________________ 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 _______________________________________________ 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 Dec 4 14:15:38 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 4 Dec 2017 21:15:38 +0100 Subject: [Nek5000-users] Calculating very small Lewis numbers in Nekton In-Reply-To: References: Message-ID: Can you please post your logfile? On 4 Dec 2017, at 21:08, "nek5000-users at lists.mcs.anl.gov " > wrote: Hey Marco, The error message is "Failed in HMHOLTZ" for the scalar and NaN value is being returned. Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group -?http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech. On Mon, Dec 4, 2017 at 2:34 PM, > wrote: Salkat, could you please be more specific when you write ? I have not been successful?? Do you get an error message?? Marco On Dec 4, 2017, at 2:31 PM, nek5000-users at lists.mcs.anl.gov wrote: Hello Neks, I am solving a Reaction Advection Diffusion equation, where with my non-dimensionalization, the Lewis Number(Le) is the parameter multiplying the Laplacian. That is, Le is the? non-dimensional Diffusion co-efficient of the species. Currently I can do Le = 0.1, but to reduce it further down to 0.01, or 0.001, the spatial resolution has to be increased. I tried doing that increasing the GLL points in SIZE file by increasing the lx1, lxd, but till now I have been unsuccessful. I am aware there is a special filtering that can be used( Fischer and Mullen, 2001). My question is how to implement that in my .usr file. Is there a routine? Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group -?http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech. _______________________________________________ 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 _______________________________________________ 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 Dec 5 08:54:08 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 5 Dec 2017 15:54:08 +0100 Subject: [Nek5000-users] hpts for vorticity field Message-ID: Hello Neks, Does anyone knows how to output the vorticity probes using call hpts() When I use the function "call hpts()", it outputs only the probes for native arrays like velocity, prssure (and scalar fields, if they exist). I compute the vorticity using the function comp_vort3(vort,w1,w2,vx,vy,vz). How can I output the values of the vorticity probes in the flow field using hpts? i will be grateful for your answers. Thank you in advance. Cheers, Sandeep -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Tue Dec 5 08:57:20 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 5 Dec 2017 15:57:20 +0100 Subject: [Nek5000-users] hpts for vorticity field In-Reply-To: References: Message-ID: poor man`s solution: Store your vorticity as an additional scalar field. -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Tuesday 5th December 2017 15:55 > To: nek5000-users at lists.mcs.anl.gov > Subject: [Nek5000-users] hpts for vorticity field > > Hello Neks, > Does anyone knows how to output the vorticity probes using call hpts() > When I use the function "call?hpts()", it outputs only the probes for native arrays like velocity, prssure?(and scalar fields, if they exist). > I compute the vorticity using the function?comp_vort3(vort,w1,w2,vx,vy,vz). How can I output the values of the vorticity probes in the flow field using hpts? > i will be grateful for your answers. > > Thank you in advance. > > Cheers, > Sandeep > _______________________________________________ > 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 Dec 5 10:08:40 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 5 Dec 2017 16:08:40 +0000 Subject: [Nek5000-users] hpts for vorticity field In-Reply-To: References: , Message-ID: Dear Sandeep, Please take a look at hemi.usr in the nekexamples repo. There is a routine interp_v() that returns the interpolated velocity field at a user-prescribed set of points. You could easily modify this routine to pass in vorticity and interpolate that (or any other scalar or vector field). Please note that you should not duplicate interrogation points on multiple processors (i.e., each processor should have a unique set of points, xyz(); one possible configuration is that MPI rank 0 has a list of points of length n and the other ranks have a list of length 0; this approach is valid if the length of the list is not too large). hth, Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Tuesday, December 5, 2017 8:57:20 AM To: nek5000-users at lists.mcs.anl.gov Subject: Re: [Nek5000-users] hpts for vorticity field poor man`s solution: Store your vorticity as an additional scalar field. -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Tuesday 5th December 2017 15:55 > To: nek5000-users at lists.mcs.anl.gov > Subject: [Nek5000-users] hpts for vorticity field > > Hello Neks, > Does anyone knows how to output the vorticity probes using call hpts() > When I use the function "call hpts()", it outputs only the probes for native arrays like velocity, prssure (and scalar fields, if they exist). > I compute the vorticity using the function comp_vort3(vort,w1,w2,vx,vy,vz). How can I output the values of the vorticity probes in the flow field using hpts? > i will be grateful for your answers. > > Thank you in advance. > > Cheers, > Sandeep > _______________________________________________ > 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 -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Tue Dec 5 13:31:27 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 5 Dec 2017 14:31:27 -0500 Subject: [Nek5000-users] Calculating very small Lewis numbers in Nekton In-Reply-To: References: Message-ID: Hello all, Thanks for the replies, I think my log-file was too big even after compression and stuck in moderator approval for the list. But here is the part of the log-file where the problem starts. I have attached a google drive link which has the log file and .rea file, the Lewis Number is set in line 133 of the .rea file after parameter 128 (p128). Link - https://drive.google.com/open?id=1ozsy0IgVKuxDiplGl6CfYDn3GuhDWoWI Solving for fluid F T T 2071 Hmholtz TEMP: 1 1.1353E+00 5.5267E+00 7.0000E+00 2071 Hmholtz PS 1: 13 5.5318E+00 3.8255E+21 7.0000E+00 2071 100 **ERROR**: Failed in HMHOLTZ: PS 2 6.5359E+15 3.8255E+21 7.0000E+00 2071 Hmholtz PS 3: 13 5.5318E+00 3.8255E+21 7.0000E+00 2071 1.3076E+03 2.0528E+00 Heat done 2071 Hmholtz VELX: 38 9.8896E-11 4.9469E+02 1.0000E-10 2071 Hmholtz VELY: 38 9.2808E-11 4.4808E+02 1.0000E-10 2071 Hmholtz VELZ: 38 8.0264E-11 4.6649E+02 1.0000E-10 2071 reorthogo: 0 2071 6 alpha: 9.1687E-01 4.2049E-02 6.7213E-03 2.4412E-03 1.0813E-03 1.7375E-03 2071 6 1.7386E-01 1.3111E-03 1.3260E+02 alph12 2071 U-PRES gmres: 5 7.5724E-09 1.0000E-08 4.0064E-08 7.3007E-02 3.7297E-01 2071 DNORM, DIVEX 5.430060732380546E-007 7.572385434962068E-009 2071 1.3076E+03 2.4498E+00 Fluid done 2071 1 flx 1.307563E+03 5.633556E-02 1.121772E+04 1.174017E+04 5.092129E+02 2071 2 flx 1.307563E+03 5.633556E-02 -1.122518E+04 -1.174681E+04 4.949962E+02 Step 2072, t= 1.3075634E+03, DT= 5.6082644E-05, C= 0.500 6.5792E+03 5.2920E+00 Solving for heat Solving for fluid F T T 2072 Hmholtz TEMP: 1 1.1351E+00 5.5225E+00 7.0000E+00 2072 100 **ERROR**: Failed in HMHOLTZ: PS 1 3.2982E+26 3.0368E+40 7.0000E+00 2072 100 **ERROR**: Failed in HMHOLTZ: PS 2 4.0048E+36 3.0365E+40 7.0000E+00 2072 100 **ERROR**: Failed in HMHOLTZ: PS 3 3.2982E+26 3.0368E+40 7.0000E+00 2072 1.3076E+03 4.8085E+00 Heat done 2072 Hmholtz VELX: 38 9.9602E-11 4.9398E+02 1.0000E-10 2072 Hmholtz VELY: 38 9.3511E-11 4.4755E+02 1.0000E-10 2072 Hmholtz VELZ: 38 8.0881E-11 4.6601E+02 1.0000E-10 2072 reorthogo: 0 2072 7 alpha: 9.1610E-01 4.8479E-02 7.7574E-03 5.1508E-03 1.8164E-03 -4.1065E-04 -4.3921E-04 2072 7 1.7378E-01 1.1572E-03 1.5018E+02 alph12 2072 U-PRES gmres: 4 8.4386E-09 1.0000E-08 3.5390E-08 4.9622E-02 3.2245E-01 2072 DNORM, DIVEX 5.419045247755792E-007 8.438599514879914E-009 2072 1.3076E+03 2.5401E+00 Fluid done Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech. On Tue, Dec 5, 2017 at 1:38 PM, Saikat Mukherjee wrote: > Hello all, > > Thanks for the replies, I think my log-file was too big and stuck in > approval for the list. But here is the part of the log-file where the > problem starts. I have attached the .rea file, the Lewis Number is set in > line 133 of the .rea file after parameter 128 (p128). > > Thanks, > Saikat > > > > > Solving for fluid F T T > 2071 Hmholtz TEMP: 1 1.1353E+00 5.5267E+00 7.0000E+00 > 2071 Hmholtz PS 1: 13 5.5318E+00 3.8255E+21 7.0000E+00 > 2071 100 **ERROR**: Failed in HMHOLTZ: PS 2 6.5359E+15 3.8255E+21 > 7.0000E+00 > 2071 Hmholtz PS 3: 13 5.5318E+00 3.8255E+21 7.0000E+00 > 2071 1.3076E+03 2.0528E+00 Heat done > 2071 Hmholtz VELX: 38 9.8896E-11 4.9469E+02 1.0000E-10 > 2071 Hmholtz VELY: 38 9.2808E-11 4.4808E+02 1.0000E-10 > 2071 Hmholtz VELZ: 38 8.0264E-11 4.6649E+02 1.0000E-10 > 2071 reorthogo: 0 > 2071 6 alpha: 9.1687E-01 4.2049E-02 6.7213E-03 2.4412E-03 > 1.0813E-03 1.7375E-03 > 2071 6 1.7386E-01 1.3111E-03 1.3260E+02 alph12 > 2071 U-PRES gmres: 5 7.5724E-09 1.0000E-08 4.0064E-08 > 7.3007E-02 3.7297E-01 > 2071 DNORM, DIVEX 5.430060732380546E-007 7.572385434962068E-009 > 2071 1.3076E+03 2.4498E+00 Fluid done > 2071 1 flx 1.307563E+03 5.633556E-02 1.121772E+04 1.174017E+04 > 5.092129E+02 > 2071 2 flx 1.307563E+03 5.633556E-02 -1.122518E+04 -1.174681E+04 > 4.949962E+02 > Step 2072, t= 1.3075634E+03, DT= 5.6082644E-05, C= 0.500 6.5792E+03 > 5.2920E+00 > Solving for heat > Solving for fluid F T T > 2072 Hmholtz TEMP: 1 1.1351E+00 5.5225E+00 7.0000E+00 > 2072 100 **ERROR**: Failed in HMHOLTZ: PS 1 3.2982E+26 3.0368E+40 > 7.0000E+00 > 2072 100 **ERROR**: Failed in HMHOLTZ: PS 2 4.0048E+36 3.0365E+40 > 7.0000E+00 > 2072 100 **ERROR**: Failed in HMHOLTZ: PS 3 3.2982E+26 3.0368E+40 > 7.0000E+00 > 2072 1.3076E+03 4.8085E+00 Heat done > 2072 Hmholtz VELX: 38 9.9602E-11 4.9398E+02 1.0000E-10 > 2072 Hmholtz VELY: 38 9.3511E-11 4.4755E+02 1.0000E-10 > 2072 Hmholtz VELZ: 38 8.0881E-11 4.6601E+02 1.0000E-10 > 2072 reorthogo: 0 > 2072 7 alpha: 9.1610E-01 4.8479E-02 7.7574E-03 5.1508E-03 > 1.8164E-03 -4.1065E-04 -4.3921E-04 > 2072 7 1.7378E-01 1.1572E-03 1.5018E+02 alph12 > 2072 U-PRES gmres: 4 8.4386E-09 1.0000E-08 3.5390E-08 > 4.9622E-02 3.2245E-01 > 2072 DNORM, DIVEX 5.419045247755792E-007 8.438599514879914E-009 > 2072 1.3076E+03 2.5401E+00 Fluid done > > > > > > > > > > > > > > > > > > > > > Saikat Mukherjee, > PhD Student, > Paul Research Group - http://www.me.vt.edu/mpaul/ > Engineering Science and Mechanics, > Virginia Tech. > > On Mon, Dec 4, 2017 at 3:52 PM, Saikat Mukherjee wrote: > >> Hey all, >> >> Thanks for the replies. Attached are the log file and the .rea file where >> I set my Lewis Number. The Lewis Number is set after parameter 128 (p128) >> at line 133 of the .rea file. >> >> Saikat >> >> Saikat Mukherjee, >> PhD Student, >> Paul Research Group - http://www.me.vt.edu/mpaul/ >> Engineering Science and Mechanics, >> Virginia Tech. >> >> On Mon, Dec 4, 2017 at 3:15 PM, wrote: >> >>> Can you please post your logfile >>> >>> On 4 Dec 2017, at 21:08, "nek5000-users at lists.mcs.anl.gov" < >>> nek5000-users at lists.mcs.anl.gov> wrote: >>> >>> Hey Marco, >>> >>> The error message is "Failed in HMHOLTZ" for the scalar and NaN value is >>> being returned. >>> >>> Thanks, >>> Saikat >>> >>> Saikat Mukherjee, >>> PhD Student, >>> Paul Research Group - http://www.me.vt.edu/mpaul/ >>> Engineering Science and Mechanics, >>> Virginia Tech. >>> >>> On Mon, Dec 4, 2017 at 2:34 PM, wrote: >>> >>>> Salkat, >>>> >>>> could you please be more specific when you write ? I have not been >>>> successful?? Do you get an error message? >>>> >>>> Marco >>>> >>>> On Dec 4, 2017, at 2:31 PM, nek5000-users at lists.mcs.anl.gov wrote: >>>> >>>> Hello Neks, >>>> >>>> I am solving a Reaction Advection Diffusion equation, where with my >>>> non-dimensionalization, the Lewis Number(Le) is the parameter multiplying >>>> the Laplacian. That is, Le is the non-dimensional Diffusion co-efficient >>>> of the species. Currently I can do Le = 0.1, but to reduce it further down >>>> to 0.01, or 0.001, the spatial resolution has to be increased. I tried >>>> doing that increasing the GLL points in SIZE file by increasing the lx1, >>>> lxd, but till now I have been unsuccessful. >>>> >>>> I am aware there is a special filtering that can be used( Fischer and >>>> Mullen, 2001). My question is how to implement that in my .usr file. Is >>>> there a routine? >>>> >>>> Thanks, >>>> Saikat >>>> >>>> >>>> Saikat Mukherjee, >>>> PhD Student, >>>> Paul Research Group - http://www.me.vt.edu/mpaul/ >>>> Engineering Science and Mechanics, >>>> Virginia Tech. >>>> _______________________________________________ >>>> 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 >>>> >>>> >>> _______________________________________________ >>> >>> 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 >>> >>> >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 6 09:23:34 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 6 Dec 2017 16:23:34 +0100 Subject: [Nek5000-users] Deformed geometry after Post-Processing Message-ID: Hi Neks, Now I am doing post-processing with my results for turbulence flow. I have the mesh with 11264 elements. First of all, I want to find a mean value using the post-processing of NEK5000, but after I get the deformed geometry that i observe witin VisIt (the computational domain is not complete). This kind of job worked well with the mesh with 8960 elements. I run my calculation on one processor. Do you have any idea what is wrong with my case? Thanks, Andrew -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 6 13:47:15 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 6 Dec 2017 13:47:15 -0600 Subject: [Nek5000-users] Outputting to 2d VTK using Postnek Message-ID: Hi, all: I have run into problems using postnek/postex to convert simulation output to VTK format for 2d simulations (3d simulations work fine). I have tested this using the eddy_uv example from the NekExamples directory. When using postex, I use the option "DUMP vtk ALL", which produces the attached VTK output. It seems that the grid point position information is incorrect. Also, I have looked through the postnek functions in trap.f and it seems that VTK cell information (CELLS, CELL_TYPES) is omitted for 2d cases. Is there a way to output this information for 2d results? Perhaps there is something I am overlooking? Any help is greatly appreciated. Thanks, Taiyo -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: eddy_uv.vtk Type: application/octet-stream Size: 74319 bytes Desc: not available URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 6 23:55:05 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 6 Dec 2017 23:55:05 -0600 Subject: [Nek5000-users] avg_all averaging Message-ID: Hi ZJ, See the attached tar file for an example of how to use the avg_all files. Read the README file to understand the contents of this example. It has two cases, the first one runs for 5 steps and generates some avg*, rms* and rm2* files. The second one uses these files to calculate the actual rms velocities using the avgp10.f0* and rmsp10.f0* files output by the first case. Hope this helps! Ketan -------------- next part -------------- A non-text attachment was scrubbed... Name: pipeavg.tar Type: application/x-tar Size: 3082752 bytes Desc: not available URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 7 11:35:52 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 7 Dec 2017 17:35:52 +0000 Subject: [Nek5000-users] Reproducing rea file in conj_ht example Message-ID: Dear Neks, Looking through the example files of conj_ht, I am unable to reproduce the conj_ht.rea file from the conj_ht.box file by invoking genbox. The appropriate file with parameters, specified in the first line of box file (cht01.rea) is missing. However, if I copy the conj_ht.rea file to cht01.rea and run genbox, the resulting box.rea file is not the same as conj_ht.rea file. The NELV numbers differ, the fluid boundary conditions additionally specify conditions for the solid domain and maybe there are some other differences too that I have not noticed. How was the conj_ht.rea file produced? Was it tweaked manually? In my own cases, I use pretex in combination with the linux utility expect, to automatically merge rea files for fluid and solid domains (generated with genbox). However, pretex needs the display to open the graphical interface. Is there another way to produce a merged file without the need for a display, pretext or user interaction? Thank you in advance. Best regards, Jure Oder From nek5000-users at lists.mcs.anl.gov Thu Dec 7 14:15:48 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 7 Dec 2017 20:15:48 +0000 Subject: [Nek5000-users] Reproducing rea file in conj_ht example In-Reply-To: References: Message-ID: Hi Jure, we are looking into documenting this issue but the short answer is that you need 2 meshes with compliant interface (e.g. two .box boxes with the same element surface) run through prenek/pretex's option of 3 CONJ. HEAT TRANSFER MERGE Aleks ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Thursday, December 7, 2017 11:35:52 AM To: nek5000-users at lists.mcs.anl.gov Subject: [Nek5000-users] Reproducing rea file in conj_ht example Dear Neks, Looking through the example files of conj_ht, I am unable to reproduce the conj_ht.rea file from the conj_ht.box file by invoking genbox. The appropriate file with parameters, specified in the first line of box file (cht01.rea) is missing. However, if I copy the conj_ht.rea file to cht01.rea and run genbox, the resulting box.rea file is not the same as conj_ht.rea file. The NELV numbers differ, the fluid boundary conditions additionally specify conditions for the solid domain and maybe there are some other differences too that I have not noticed. How was the conj_ht.rea file produced? Was it tweaked manually? In my own cases, I use pretex in combination with the linux utility expect, to automatically merge rea files for fluid and solid domains (generated with genbox). However, pretex needs the display to open the graphical interface. Is there another way to produce a merged file without the need for a display, pretext or user interaction? Thank you in advance. Best regards, Jure Oder _______________________________________________ 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 Thu Dec 7 14:35:30 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 7 Dec 2017 20:35:30 +0000 Subject: [Nek5000-users] avg_all averaging Message-ID: Dear Ketan, Thanks for your attachment. I tried your code 'pp.usr' provided last email but failed in complilation. Can you provide subroutines fgslib_gs_op and fgslib_gs_setup used in subroutines my_z_avg and set_gs_zavg_hndl, respectively. Best wishes, ZJ -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 7 14:44:01 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 7 Dec 2017 15:44:01 -0500 Subject: [Nek5000-users] avg_all averaging In-Reply-To: References: Message-ID: Hi ZJ, From what I understand the "fgslib_? is the updated syntax for the gather/scatter routines, which is now a separate library. You should be able to remove the fgslib_ from the name of the subroutine, or update Nek to the latest repo. Regards, Ravon > On Dec 7, 2017, at 3:35 PM, nek5000-users at lists.mcs.anl.gov wrote: > > Dear Ketan, > > Thanks for your attachment. I tried your code 'pp.usr' provided last email but failed in complilation. Can you provide subroutines fgslib_gs_op and fgslib_gs_setup used in subroutines my_z_avg and set_gs_zavg_hndl, respectively. > > Best wishes, > > ZJ > _______________________________________________ > 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 Dec 7 20:28:29 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 8 Dec 2017 10:28:29 +0800 (CST) Subject: [Nek5000-users] A meshing generation problem for complex geometric structure Message-ID: Dear experts in Nek5000 : Thank you for reading my email. I am a freshman in using Nek5000 and I am willing to simulate unsteady flow phenomenon in bottom head of PWR with LES method through parallel processes. And Nek5000 is the best solver for resolving my problem and I hope to reach my goal with the outstanding SEM-CFD tools. But it is really a matter for me to generate a mesh file for that structure. So could you please show me a proper way to reach the .rea mesh file for the bottom head for the geometry there is really complex and I am afraid I can not achieve the goal with editing .rea file directly. Thank you for your supporting me and I am waiting for your earlier reply eagerly. Best regards Haoran JU -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Fri Dec 8 07:03:42 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 8 Dec 2017 13:03:42 +0000 Subject: [Nek5000-users] A meshing generation problem for complex geometric structure In-Reply-To: References: Message-ID: Dear Horan, the best approach I know to import a mesh in Nek5000, is to export your mesh in a exodus format with HEX27 elements (Cubit has this capabilities) and use the exo2nek script to create a binary file .re2. The exo2nek script can be found in the tools directory and has to be installed. Please read the associated README file for installation instruction. Once it is installed, I suggest you start with a simple geometry (cube or pipe) to run the script, and make sure you have all of the steps right. Note that boundary conditions will have to be set in usrdat2 subroutine of the .usr file: do iel=1,nelt do ifc=1,2*ndim idss = bc(5,ifc,iel,1) if (idss.eq.1) cbc(ifc,iel,1)='v ? if (idss.eq.2) cbc(ifc,iel,1)=?O ? if (idss.eq.3) cbc(ifc,iel,1)=?W ' enddo enddo assuming that you have side set ids 1, 2 and 3 in your mesh. exo2nek stores the side set id in the variable ?bc(5,inc,iel,1)? for the velocity field. If you really need the rea file, a rea file can be generated with the function call gen_rea(2) in usrdat2() but after setting the boundary conditions. Hope this help, Marco On Dec 7, 2017, at 9:28 PM, nek5000-users at lists.mcs.anl.gov wrote: Dear experts in Nek5000 : Thank you for reading my email. I am a freshman in using Nek5000 and I am willing to simulate unsteady flow phenomenon in bottom head of PWR with LES method through parallel processes. And Nek5000 is the best solver for resolving my problem and I hope to reach my goal with the outstanding SEM-CFD tools. But it is really a matter for me to generate a mesh file for that structure. So could you please show me a proper way to reach the .rea mesh file for the bottom head for the geometry there is really complex and I am afraid I can not achieve the goal with editing .rea file directly. Thank you for your supporting me and I am waiting for your earlier reply eagerly. Best regards Haoran JU _______________________________________________ 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 Sun Dec 10 08:55:16 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Sun, 10 Dec 2017 09:55:16 -0500 Subject: [Nek5000-users] Laminar pipe flow in Nek5000 Message-ID: Hello Neks, I am new to Nek5000 and struggling to understand it. At this moment, I want to simulate laminar flow in a pipe. please advice me, with which example, should I start and what modifications in various file should I be doing. Thanks, Devshi Ziyad -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Sun Dec 10 11:40:04 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Sun, 10 Dec 2017 17:40:04 +0000 Subject: [Nek5000-users] Laminar pipe flow in Nek5000 In-Reply-To: References: Message-ID: Hi Devshi, There is a "pipe" example in nekexamples. If you (say) turn off the geometry modification in the helix case, which is found is subroutine usrdat2 in helix.usr, you would end up with a straight pipe. hth. Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Sunday, December 10, 2017 8:55:16 AM To: nek5000-users at lists.mcs.anl.gov Subject: [Nek5000-users] Laminar pipe flow in Nek5000 Hello Neks, I am new to Nek5000 and struggling to understand it. At this moment, I want to simulate laminar flow in a pipe. please advice me, with which example, should I start and what modifications in various file should I be doing. Thanks, Devshi Ziyad -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Mon Dec 11 12:25:57 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 11 Dec 2017 18:25:57 +0000 Subject: [Nek5000-users] pipe turbulence Message-ID: Hi Nek, For the single phase pipe flow, I set bulk reynolds number equals to 25000, but still no turbulence in the pipe after running for over 10,000 timesteps. Is there anyone know what's going on the pipe. Best wishes, ZJ -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Mon Dec 11 12:31:03 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 11 Dec 2017 18:31:03 +0000 Subject: [Nek5000-users] pipe turbulence In-Reply-To: References: Message-ID: ZJ, did you look at the pipe example in Nek5000? I believe you have to add an initial perturbation to trigger the turbulent flow. What type of boundary conditions are you using? Marco > On Dec 11, 2017, at 1:25 PM, nek5000-users at lists.mcs.anl.gov wrote: > > Hi Nek, > > For the single phase pipe flow, I set bulk reynolds number equals to 25000, but still no turbulence in the pipe after running for over 10,000 timesteps. Is there anyone know what's going on the pipe. > > Best wishes, > > ZJ > _______________________________________________ > 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 Mon Dec 11 12:48:51 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 11 Dec 2017 19:48:51 +0100 Subject: [Nek5000-users] pipe turbulence In-Reply-To: References: Message-ID: actually, I would recommend a tripping as described in Schlatter and Orlu (JFM 2012) and Vinuesa et al. (J. Turbulence 2014). Alternatives are of course random initial conditions (but be careful with the choice of axial/azimuthal scales!), or potentially the synthetic eddy method. Single phase turbulent pipe flow is possible in Nek5000, see e.g. the paper El Khoury et al. (Flow Turbulence and Combustion, 2013). Philipp On 2017-12-11 19:31, nek5000-users at lists.mcs.anl.gov wrote: > ZJ, > > did you look at the pipe example in Nek5000? I believe you have to add an initial perturbation to trigger the turbulent flow. What type of boundary conditions are you using? > > Marco > >> On Dec 11, 2017, at 1:25 PM, nek5000-users at lists.mcs.anl.gov wrote: >> >> Hi Nek, >> >> For the single phase pipe flow, I set bulk reynolds number equals to 25000, but still no turbulence in the pipe after running for over 10,000 timesteps. Is there anyone know what's going on the pipe. >> >> Best wishes, >> >> ZJ >> _______________________________________________ >> 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 > From nek5000-users at lists.mcs.anl.gov Thu Dec 14 05:16:40 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 14 Dec 2017 11:16:40 +0000 Subject: [Nek5000-users] pipe turbulence Message-ID: Dear Marco and Phillip, I had a look at the pipe example in nek5000 where there exsit three cases, but I dont know how it generates turbulence quickly. In terms of tripping papers by Phillip Schlatter and Vinuesa et al, I am sorry to say I am not professional with nek and it's hard for me to find the key information in those papers. Would you please check what's wrong with the attached rea file and usr file. Thanks so much! ? turbPipe.rea ? Best wishes, ZJ -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: turbPipe.usr Type: application/octet-stream Size: 3804 bytes Desc: not available URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 14 05:24:20 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 14 Dec 2017 12:24:20 +0100 Subject: [Nek5000-users] pipe turbulence In-Reply-To: References: Message-ID: Hi, the rea files do not include the tripping; this needs to be implemented separately. There should be an example with the random noise though. Best regards, Philipp On 2017-12-14 12:16, nek5000-users at lists.mcs.anl.gov wrote: > Dear Marco and Phillip, > > I had a look at the pipe example in nek5000 where there exsit three > cases, but I dont know how it generates turbulence quickly. In terms of > tripping papers by Phillip Schlatter and Vinuesa et al, I am sorry to > say I am not professional with nek and it's hard for me to find the key > information in those papers. Would you please check what's wrong with > the attached rea file and usr file. Thanks so much! > > ? > turbPipe.rea > > ? > > Best wishes, > > ZJ > > > _______________________________________________ > 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 Dec 14 06:54:23 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 14 Dec 2017 12:54:23 +0000 Subject: [Nek5000-users] pipe turbulence In-Reply-To: References: Message-ID: ZJ, please find attached an example of a turbulent flow in a pipe with periodic boundary conditions. It should be quite simple to understand, but do not hesitate if you have questions. Marco -------------- next part -------------- A non-text attachment was scrubbed... Name: dns-turbulent-flow-in-pipe.zip Type: application/zip Size: 1004049 bytes Desc: dns-turbulent-flow-in-pipe.zip URL: -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: ATT00001.txt URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 14 10:19:23 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 14 Dec 2017 16:19:23 +0000 Subject: [Nek5000-users] pipe turbulence Message-ID: Dear Marco and Phillip, Thank Marco for your turbulent pipe cases which is working for me. Regards to the files, may I ask you how long will it take to get fully developed, and how can I increase the elements number. Do you have a 2D mesh rea file, if so I could do more elements in prenek. Kind regards, ZJ -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 14 12:14:03 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 14 Dec 2017 18:14:03 +0000 Subject: [Nek5000-users] pipe turbulence In-Reply-To: References: Message-ID: I attached a script that will generate a mesh for a pipe. You can change some of the parameters to generate a finer mesh. The boundary conditions can be changed at the end of the script. To get a full developed turbulent flow it will depend on the Reynolds number. The files I sent you before should give you a turbulent flow. Marco -------------- next part -------------- A non-text attachment was scrubbed... Name: mkpipe Type: application/octet-stream Size: 1278 bytes Desc: mkpipe URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: one2d.rea Type: application/octet-stream Size: 6293 bytes Desc: one2d.rea URL: -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: ATT00001.txt URL: From nek5000-users at lists.mcs.anl.gov Mon Dec 18 01:57:28 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 18 Dec 2017 08:57:28 +0100 Subject: [Nek5000-users] Announcing Nek5000 17.0 Message-ID: We are proud to announce that Nek5000 version 17.0 is now available for download at: https://nek5000.mcs.anl.gov Please read the Release Notes carefully before using the code. Note, we strongly recommend to use the latest release tarball in a production environment instead of the lastest master branch on GitHub. None of this incredible work would be possible without the support and contributions of our wonderful members and community. Thank you all for your hard work! Enjoy, The Nek Development Team From nek5000-users at lists.mcs.anl.gov Mon Dec 18 12:43:39 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 18 Dec 2017 18:43:39 +0000 Subject: [Nek5000-users] MAXNEL Message-ID: Hi NEK, Is there anyone know where can I change the default MAXNEL value. My case has more than 300000 elements which causes genmap unsuccessful. Best wishes, Z -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Mon Dec 18 12:54:34 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 18 Dec 2017 18:54:34 +0000 Subject: [Nek5000-users] MAXNEL In-Reply-To: References: Message-ID: Hi, You can change this by uncommenting the following line in tools/maketools: #MAXNEL=150000 and setting it to (say): MAXNEL=350000 Regards, Thilina ________________________________ From: Nek5000-users [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: Monday, December 18, 2017 12:43 PM To: nek5000-users at lists.mcs.anl.gov Subject: [Nek5000-users] MAXNEL Hi NEK, Is there anyone know where can I change the default MAXNEL value. My case has more than 300000 elements which causes genmap unsuccessful. Best wishes, Z -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Mon Dec 18 13:17:54 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 18 Dec 2017 20:17:54 +0100 Subject: [Nek5000-users] MAXNEL In-Reply-To: References: Message-ID: Just edit maketools, uncommonly MAXNEL and set it to your preferred value. On 18 Dec 2017, at 19:44, "nek5000-users at lists.mcs.anl.gov " > wrote: Hi NEK, Is there anyone know where can I change the default MAXNEL value. My case has more than 300000 elements which causes genmap unsuccessful. Best wishes, Z _______________________________________________ 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 Dec 18 12:50:27 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Mon, 18 Dec 2017 12:50:27 -0600 Subject: [Nek5000-users] MAXNEL In-Reply-To: <89b9ebf20b574d70adbdbd232b20a0a7@CITESHT4.ad.uillinois.edu> References: <89b9ebf20b574d70adbdbd232b20a0a7@CITESHT4.ad.uillinois.edu> Message-ID: Hi, You can change this by uncommenting the following line in tools/maketools: #MAXNEL=150000 and setting it to (say): MAXNEL=350000 Regards, Thilina On Mon, Dec 18, 2017 at 12:43 PM, nek5000-users at lists.mcs.anl.gov < nek5000-users at lists.mcs.anl.gov> wrote: > Hi NEK, > > Is there anyone know where can I change the default MAXNEL value. My case > has more than 300000 elements which causes genmap unsuccessful. > > Best wishes, > > Z > -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 20 05:08:09 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 20 Dec 2017 12:08:09 +0100 Subject: [Nek5000-users] How to use AMG instead of XXT with Nek5000 v17.0 Message-ID: This has changed, see Release Notes. Workflow: 1) set p40=1 in your .rea or preconditioner=semg_amg in .par 2) run Nek5000 (this will create the amgdump files in case they don't exist) 3) run the amg_hypre tool (this will create the input files required for AMG) 4) run Nek5000 Note: Step (2) and (3) are only needed once. From nek5000-users at lists.mcs.anl.gov Wed Dec 20 07:20:53 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 20 Dec 2017 13:20:53 +0000 Subject: [Nek5000-users] makenek issue Message-ID: Hello, I was suddenly having problems running makenek in the latest nek version. Does anyone have ideas about how to compile it completely. /home/Nek5000/core/gmres.f:(.text+0xd48f): relocation truncated to fit: R_X86_64_PC32 against symbol `fast1dsem_' defined in COMMON section in obj/fast3d.o ... /home/Nek5000/core/gmres.f:(.text+0xd507): additional relocation overflows omitted from the output obj/byte_mpi.o: In function `byte_sync_mpi_': /home/Nek5000/core/byte_mpi.f:(.text+0x8): undefined reference to `mpi_file_sync_' obj/byte_mpi.o: In function `byte_open_mpi_': ... /home/Nek5000/core/byte_mpi.f:(.text+0x8): undefined reference to `mpi_file_sync_' obj/byte_mpi.o: In function `byte_open_mpi_': ... /home/ufaserv1_i/mnsli/Nek5000/3rd_party/gslib/lib/libgs.a(gs.o): In function `pw_exec': ... Best regards, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 20 09:52:30 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 20 Dec 2017 15:52:30 +0000 Subject: [Nek5000-users] makenek issue In-Reply-To: References: Message-ID: Dear Jian, Can you please send me your case files, including SIZE, off-list, and I'll take a look. Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Wednesday, December 20, 2017 7:20:53 AM To: nek5000-users at lists.mcs.anl.gov Subject: [Nek5000-users] makenek issue Hello, I was suddenly having problems running makenek in the latest nek version. Does anyone have ideas about how to compile it completely. /home/Nek5000/core/gmres.f:(.text+0xd48f): relocation truncated to fit: R_X86_64_PC32 against symbol `fast1dsem_' defined in COMMON section in obj/fast3d.o ... /home/Nek5000/core/gmres.f:(.text+0xd507): additional relocation overflows omitted from the output obj/byte_mpi.o: In function `byte_sync_mpi_': /home/Nek5000/core/byte_mpi.f:(.text+0x8): undefined reference to `mpi_file_sync_' obj/byte_mpi.o: In function `byte_open_mpi_': ... /home/Nek5000/core/byte_mpi.f:(.text+0x8): undefined reference to `mpi_file_sync_' obj/byte_mpi.o: In function `byte_open_mpi_': ... /home/ufaserv1_i/mnsli/Nek5000/3rd_party/gslib/lib/libgs.a(gs.o): In function `pw_exec': ... Best regards, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 20 10:08:34 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 20 Dec 2017 16:08:34 +0000 Subject: [Nek5000-users] makenek issue In-Reply-To: References: , Message-ID: Hi Jian, it sounds like your nek5000 executable becomes larger than 2GB so try increasing lpmin to fit (or decreasing lelt if using old SIZE format). Aleks ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Wednesday, December 20, 2017 9:52:30 AM To: nek5000-users at lists.mcs.anl.gov Subject: Re: [Nek5000-users] makenek issue Dear Jian, Can you please send me your case files, including SIZE, off-list, and I'll take a look. Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Wednesday, December 20, 2017 7:20:53 AM To: nek5000-users at lists.mcs.anl.gov Subject: [Nek5000-users] makenek issue Hello, I was suddenly having problems running makenek in the latest nek version. Does anyone have ideas about how to compile it completely. /home/Nek5000/core/gmres.f:(.text+0xd48f): relocation truncated to fit: R_X86_64_PC32 against symbol `fast1dsem_' defined in COMMON section in obj/fast3d.o ... /home/Nek5000/core/gmres.f:(.text+0xd507): additional relocation overflows omitted from the output obj/byte_mpi.o: In function `byte_sync_mpi_': /home/Nek5000/core/byte_mpi.f:(.text+0x8): undefined reference to `mpi_file_sync_' obj/byte_mpi.o: In function `byte_open_mpi_': ... /home/Nek5000/core/byte_mpi.f:(.text+0x8): undefined reference to `mpi_file_sync_' obj/byte_mpi.o: In function `byte_open_mpi_': ... /home/ufaserv1_i/mnsli/Nek5000/3rd_party/gslib/lib/libgs.a(gs.o): In function `pw_exec': ... Best regards, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 20 10:12:14 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 20 Dec 2017 17:12:14 +0100 Subject: [Nek5000-users] makenek issue Message-ID: I don't think this is related to the new version. The compiler tries to tell you that the global or static data exceeded the first 2GB. This is typically related to a large lelt (legacy SIZE) or small lpmin (new SIZE file). Please check this values if they are meaningful. If you think they are you can compile Nek with FFLAGS="-mcmodel=medium". -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Wednesday 20th December 2017 16:53 > To: nek5000-users at lists.mcs.anl.gov > Subject: Re: [Nek5000-users] makenek issue > > > Dear Jian, > > Can you please send me your case files, including SIZE, > off-list, and I'll take a look. > > Paul > > ----------- > From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov > Sent: Wednesday, December 20, 2017 7:20:53 AM > To: nek5000-users at lists.mcs.anl.gov > Subject: [Nek5000-users] makenek issue > Hello, > > I was suddenly having problems running makenek in the latest nek version. Does anyone have ideas about how to compile it completely. > > /home/Nek5000/core/gmres.f:(.text0xd48f): relocation truncated to fit: R_X86_64_PC32 against symbol `fast1dsem_' defined in COMMON section in obj/fast3d.o > ... > /home/Nek5000/core/gmres.f:(.text0xd507): additional relocation overflows omitted from the output > obj/byte_mpi.o: In function `byte_sync_mpi_': > /home/Nek5000/core/byte_mpi.f:(.text0x8): undefined reference to `mpi_file_sync_' > obj/byte_mpi.o: In function `byte_open_mpi_': > ... > /home/Nek5000/core/byte_mpi.f:(.text0x8): undefined reference to `mpi_file_sync_' > obj/byte_mpi.o: In function `byte_open_mpi_': > ... > /home/ufaserv1_i/mnsli/Nek5000/3rd_party/gslib/lib/libgs.a(gs.o): In function `pw_exec': > ... > > Best regards, > > Jian > _______________________________________________ > 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 Dec 20 11:44:40 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 20 Dec 2017 17:44:40 +0000 Subject: [Nek5000-users] Viewing Mesh without solution files Message-ID: Hi all, I'm building some large meshes using genbox and need to know if I can actually view the meshes on visit/paraview without actually launching the simulation or even compiling the problem first. That is, I want to view the mesh after genbox has run. Is this possible? Sincerely, -- *Amitvikram Dutta* Graduate Research Assistant Fluid Mechanics Research Lab University of Waterloo -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 20 12:14:00 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 20 Dec 2017 18:14:00 +0000 Subject: [Nek5000-users] makenek issue Message-ID: Hi Paul, I used the same rea file in previous nek, which was working well. But the error occurs in the new code. I am wondering whether it's the compilation. Attached is the case files. ? pipe.tar ? Kind regards, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 20 12:45:08 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 20 Dec 2017 18:45:08 +0000 Subject: [Nek5000-users] makenek issue In-Reply-To: References: Message-ID: Hi Jian, Is there a reason you have lelx=30, lely=lelx,lelz=lelx? I would set lelx=1. The features that use lelx (there are no good ones) are not scalable. Please let me know if this remedies your problem. Best, Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Wednesday, December 20, 2017 12:14:00 PM To: nek5000-users at lists.mcs.anl.gov Subject: Re: [Nek5000-users] makenek issue Hi Paul, I used the same rea file in previous nek, which was working well. But the error occurs in the new code. I am wondering whether it's the compilation. Attached is the case files. ? [https://ssl.gstatic.com/docs/doclist/images/icon_10_generic_list.png] pipe.tar[X] ? Kind regards, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 20 13:11:36 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 20 Dec 2017 19:11:36 +0000 Subject: [Nek5000-users] makenek issue In-Reply-To: References: , Message-ID: PS - I see that setting lelx=1 is not sufficient (but still a good idea). I added the following to the makenek script: # optional compiler flags FFLAGS="-mcmodel=medium" CFLAGS="-mcmodel=medium" # optional linking flags USR_LFLAGS+="-mcmodel=medium" which did get the code to link and compile, after "makenek clean" Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Wednesday, December 20, 2017 12:45:08 PM To: nek5000-users at lists.mcs.anl.gov Subject: Re: [Nek5000-users] makenek issue Hi Jian, Is there a reason you have lelx=30, lely=lelx,lelz=lelx? I would set lelx=1. The features that use lelx (there are no good ones) are not scalable. Please let me know if this remedies your problem. Best, Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Wednesday, December 20, 2017 12:14:00 PM To: nek5000-users at lists.mcs.anl.gov Subject: Re: [Nek5000-users] makenek issue Hi Paul, I used the same rea file in previous nek, which was working well. But the error occurs in the new code. I am wondering whether it's the compilation. Attached is the case files. ? [https://ssl.gstatic.com/docs/doclist/images/icon_10_generic_list.png] pipe.tar[X] ? Kind regards, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 20 19:40:39 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 20 Dec 2017 22:40:39 -0300 Subject: [Nek5000-users] [Re] Viewing Mesh without solution files Message-ID: Dear Amitvikram, I would like to know that also. What I do is run a simulation and manually finish it after the first time step. Something that I think may work but (I haven't tried yet) is to make a script to run a simulation and setting everything off, (IFFLOW=false, IFTRANSIENT=false, IFHEAT=false,...etc) so the code will do nothing but deliver a mesh. It's just an idea. Regards, JP. -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 20 19:46:13 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 01:46:13 +0000 Subject: [Nek5000-users] [Re] Viewing Mesh without solution files In-Reply-To: References: Message-ID: If you run with nsteps=0 and fintime=0 then the code dumps the initial conditions without setting up the solvers. With the initial conditions comes the mesh. If you turn off all the outputs (ifvo=.false., ifto=.false. etc.) then you would just get the mesh. Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Wednesday, December 20, 2017 7:40:39 PM To: nek5000-users at lists.mcs.anl.gov Subject: [Nek5000-users] [Re] Viewing Mesh without solution files Dear Amitvikram, I would like to know that also. What I do is run a simulation and manually finish it after the first time step. Something that I think may work but (I haven't tried yet) is to make a script to run a simulation and setting everything off, (IFFLOW=false, IFTRANSIENT=false, IFHEAT=false,...etc) so the code will do nothing but deliver a mesh. It's just an idea. Regards, JP. -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 20 20:30:26 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 02:30:26 +0000 Subject: [Nek5000-users] [Re] Viewing Mesh without solution files In-Reply-To: References: Message-ID: Thanks all. Solves a long standing problem for me. On Wed, Dec 20, 2017, 8:46 PM wrote: > > > > If you run with nsteps=0 and fintime=0 then the code dumps the initial > conditions without setting up the solvers. > > > With the initial conditions comes the mesh. > > > If you turn off all the outputs (ifvo=.false., ifto=.false. etc.) then you > would just get the mesh. > > > Paul > > > ------------------------------ > *From:* Nek5000-users on behalf > of nek5000-users at lists.mcs.anl.gov > *Sent:* Wednesday, December 20, 2017 7:40:39 PM > *To:* nek5000-users at lists.mcs.anl.gov > *Subject:* [Nek5000-users] [Re] Viewing Mesh without solution files > > Dear Amitvikram, I would like to know that also. > > What I do is run a simulation and manually finish it after the first time > step. > > Something that I think may work but (I haven't tried yet) is to make a > script to run a simulation and setting everything off, (IFFLOW=false, > IFTRANSIENT=false, IFHEAT=false,...etc) so the code will do nothing but > deliver a mesh. > > It's just an idea. > > Regards, > JP. > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > -- *Amitvikram Dutta* Graduate Research Assistant Fluid Mechanics Research Lab University of Waterloo -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 21 01:21:43 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 02:21:43 -0500 Subject: [Nek5000-users] Error compiling new Nek5000 Message-ID: Hello all, I am having errors compiling the new Nek5000. I am trying to run nekton in a unix cluster. I get the following error, the log file being attached - src/comm.h(65): catastrophic error: cannot open source file "mpi.h" #include Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech. -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: logfile Type: application/octet-stream Size: 16570 bytes Desc: not available URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 21 01:46:30 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 08:46:30 +0100 Subject: [Nek5000-users] Error compiling new Nek5000 In-Reply-To: References: Message-ID: Your compiler does not find the include file mpi.h Have you tried to use the MPI wrappers like mpif77/mpicc instead of ifort/icc. -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Thursday 21st December 2017 8:27 > To: nek5000-users > Subject: [Nek5000-users] Error compiling new Nek5000 > > Hello all, > > I am having errors compiling the new Nek5000. I am trying to run nekton in a unix cluster. I get the following error, the log file being attached -? > > src/comm.h(65): catastrophic error: cannot open source file "mpi.h" > ? #include > > Thanks, > > Saikat > > Saikat Mukherjee, > PhD Student, > Paul Research Group -?http://www.me.vt.edu/mpaul/ > Engineering Science and Mechanics, > Virginia Tech. > _______________________________________________ > 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 Dec 21 02:13:43 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 03:13:43 -0500 Subject: [Nek5000-users] Error compiling new Nek5000 In-Reply-To: References: Message-ID: I think I am using mpif77. Here's my compiler file attached. Thanks, Saikat Saikat Mukherjee, PhD Student, Paul Research Group - http://www.me.vt.edu/mpaul/ Engineering Science and Mechanics, Virginia Tech. On Thu, Dec 21, 2017 at 2:46 AM, wrote: > Your compiler does not find the include file mpi.h > Have you tried to use the MPI wrappers like mpif77/mpicc instead of > ifort/icc. > > -----Original message----- > > From:nek5000-users at lists.mcs.anl.gov > > Sent: Thursday 21st December 2017 8:27 > > To: nek5000-users > > Subject: [Nek5000-users] Error compiling new Nek5000 > > > > Hello all, > > > > I am having errors compiling the new Nek5000. I am trying to run nekton > in a unix cluster. I get the following error, the log file being attached - > > > > src/comm.h(65): catastrophic error: cannot open source file "mpi.h" > > #include > > > > Thanks, > > > > Saikat > > > > Saikat Mukherjee, > > PhD Student, > > Paul Research Group - http://www.me.vt.edu/mpaul/ < > http://www.me.vt.edu/mpaul/> > > Engineering Science and Mechanics, > > Virginia Tech. > > _______________________________________________ > > 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 > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: makenek Type: application/octet-stream Size: 1958 bytes Desc: not available URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 21 02:42:06 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 09:42:06 +0100 Subject: [Nek5000-users] Error compiling new Nek5000 In-Reply-To: References: Message-ID: Somehow gslib gets compiled with icc instead of mpicc. Can you try the following: export CC=mpicc export FC=mpif77 makenek clean (clean 3rd party as well) makenek -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Thursday 21st December 2017 9:14 > To: nek5000-users > Subject: Re: [Nek5000-users] Error compiling new Nek5000 > > I think I am using mpif77. Heres my compiler file attached. > > Thanks, > Saikat >
Saikat Mukherjee, > PhD Student, > Paul Research Group -?http://www.me.vt.edu/mpaul/ > Engineering Science and Mechanics, > Virginia Tech. > > On Thu, Dec 21, 2017 at 2:46 AM, > wrote: > Your compiler does not find the include file mpi.h > Have you tried to use the MPI wrappers like mpif77/mpicc instead of ifort/icc. > > -----Original message----- > > From:nek5000-users at lists.mcs.anl.gov > > > Sent: Thursday 21st December 2017 8:27 > > To: nek5000-users > > > Subject: [Nek5000-users] Error compiling new Nek5000 > > > > Hello all, > > > > I am having errors compiling the new Nek5000. I am trying to run nekton in a unix cluster. I get the following error, the log file being attached -? > > > > src/comm.h(65): catastrophic error: cannot open source file "mpi.h" > > ? #include > > > > Thanks, > > > > Saikat > > > > Saikat Mukherjee, > > PhD Student, > > Paul Research Group -?http://www.me.vt.edu/mpaul/ > > > Engineering Science and Mechanics, > > Virginia Tech. > > _______________________________________________ > > 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 > > _______________________________________________ > 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 Dec 21 02:56:41 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 09:56:41 +0100 Subject: [Nek5000-users] Error compiling new Nek5000 In-Reply-To: References: Message-ID: Have you tried doing "makenek realclean" and then compiling with "makenek". NN. On Thursday 21 December 2017 09:42 AM, nek5000-users at lists.mcs.anl.gov wrote: > Somehow gslib gets compiled with icc instead of mpicc. Can you try the following: > > export CC=mpicc > export FC=mpif77 > makenek clean (clean 3rd party as well) > makenek > > -----Original message----- >> From:nek5000-users at lists.mcs.anl.gov >> Sent: Thursday 21st December 2017 9:14 >> To: nek5000-users >> Subject: Re: [Nek5000-users] Error compiling new Nek5000 >> >> I think I am using mpif77. Heres my compiler file attached. >> >> Thanks, >> Saikat >>
Saikat Mukherjee, >> PhD Student, >> Paul Research Group - http://www.me.vt.edu/mpaul/ >> Engineering Science and Mechanics, >> Virginia Tech. >> >> On Thu, Dec 21, 2017 at 2:46 AM, > wrote: >> Your compiler does not find the include file mpi.h >> Have you tried to use the MPI wrappers like mpif77/mpicc instead of ifort/icc. >> >> -----Original message----- >>> From:nek5000-users at lists.mcs.anl.gov > >>> Sent: Thursday 21st December 2017 8:27 >>> To: nek5000-users > >>> Subject: [Nek5000-users] Error compiling new Nek5000 >>> >>> Hello all, >>> >>> I am having errors compiling the new Nek5000. I am trying to run nekton in a unix cluster. I get the following error, the log file being attached - >>> >>> src/comm.h(65): catastrophic error: cannot open source file "mpi.h" >>> #include >>> >>> Thanks, >>> >>> Saikat >>> >>> Saikat Mukherjee, >>> PhD Student, >>> Paul Research Group - http://www.me.vt.edu/mpaul/ > >>> Engineering Science and Mechanics, >>> Virginia Tech. >>> _______________________________________________ >>> 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 >> >> _______________________________________________ >> 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 > From nek5000-users at lists.mcs.anl.gov Thu Dec 21 02:58:58 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 09:58:58 +0100 Subject: [Nek5000-users] Error compiling new Nek5000 In-Reply-To: References: Message-ID: The realclean option was deleted in 17.0. There is just clean. -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Thursday 21st December 2017 9:57 > To: nek5000-users at lists.mcs.anl.gov > Subject: Re: [Nek5000-users] Error compiling new Nek5000 > > Have you tried doing "makenek realclean" and then compiling > with "makenek". > > NN. > > > On Thursday 21 December 2017 09:42 AM, nek5000-users at lists.mcs.anl.gov > wrote: > > Somehow gslib gets compiled with icc instead of mpicc. Can you try the following: > > > > export CC=mpicc > > export FC=mpif77 > > makenek clean (clean 3rd party as well) > > makenek > > > > -----Original message----- > >> From:nek5000-users at lists.mcs.anl.gov > >> Sent: Thursday 21st December 2017 9:14 > >> To: nek5000-users > >> Subject: Re: [Nek5000-users] Error compiling new Nek5000 > >> > >> I think I am using mpif77. Heres my compiler file attached. > >> > >> Thanks, > >> Saikat > >>
Saikat Mukherjee, > >> PhD Student, > >> Paul Research Group - http://www.me.vt.edu/mpaul/ > >> Engineering Science and Mechanics, > >> Virginia Tech. > >> > >> On Thu, Dec 21, 2017 at 2:46 AM, > wrote: > >> Your compiler does not find the include file mpi.h > >> Have you tried to use the MPI wrappers like mpif77/mpicc instead of ifort/icc. > >> > >> -----Original message----- > >>> From:nek5000-users at lists.mcs.anl.gov > > >>> Sent: Thursday 21st December 2017 8:27 > >>> To: nek5000-users > > >>> Subject: [Nek5000-users] Error compiling new Nek5000 > >>> > >>> Hello all, > >>> > >>> I am having errors compiling the new Nek5000. I am trying to run nekton in a unix cluster. I get the following error, the log file being attached - > >>> > >>> src/comm.h(65): catastrophic error: cannot open source file "mpi.h" > >>> #include > >>> > >>> Thanks, > >>> > >>> Saikat > >>> > >>> Saikat Mukherjee, > >>> PhD Student, > >>> Paul Research Group - http://www.me.vt.edu/mpaul/ > > >>> Engineering Science and Mechanics, > >>> Virginia Tech. > >>> _______________________________________________ > >>> 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 > >> > >> _______________________________________________ > >> 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 > > > > _______________________________________________ > 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 Dec 21 03:01:31 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 10:01:31 +0100 Subject: [Nek5000-users] Error compiling new Nek5000 In-Reply-To: References: Message-ID: Thanks for letting me know. I have the latest version on my machine but have not used it yet. NN. On Thursday 21 December 2017 09:58 AM, nek5000-users at lists.mcs.anl.gov wrote: > The realclean option was deleted in 17.0. There is just clean. > > -----Original message----- >> From:nek5000-users at lists.mcs.anl.gov >> Sent: Thursday 21st December 2017 9:57 >> To: nek5000-users at lists.mcs.anl.gov >> Subject: Re: [Nek5000-users] Error compiling new Nek5000 >> >> Have you tried doing "makenek realclean" and then compiling >> with "makenek". >> >> NN. >> >> >> On Thursday 21 December 2017 09:42 AM, nek5000-users at lists.mcs.anl.gov >> wrote: >>> Somehow gslib gets compiled with icc instead of mpicc. Can you try the following: >>> >>> export CC=mpicc >>> export FC=mpif77 >>> makenek clean (clean 3rd party as well) >>> makenek >>> >>> -----Original message----- >>>> From:nek5000-users at lists.mcs.anl.gov >>>> Sent: Thursday 21st December 2017 9:14 >>>> To: nek5000-users >>>> Subject: Re: [Nek5000-users] Error compiling new Nek5000 >>>> >>>> I think I am using mpif77. Heres my compiler file attached. >>>> >>>> Thanks, >>>> Saikat >>>>
Saikat Mukherjee, >>>> PhD Student, >>>> Paul Research Group - http://www.me.vt.edu/mpaul/ >>>> Engineering Science and Mechanics, >>>> Virginia Tech. >>>> >>>> On Thu, Dec 21, 2017 at 2:46 AM, > wrote: >>>> Your compiler does not find the include file mpi.h >>>> Have you tried to use the MPI wrappers like mpif77/mpicc instead of ifort/icc. >>>> >>>> -----Original message----- >>>>> From:nek5000-users at lists.mcs.anl.gov > >>>>> Sent: Thursday 21st December 2017 8:27 >>>>> To: nek5000-users > >>>>> Subject: [Nek5000-users] Error compiling new Nek5000 >>>>> >>>>> Hello all, >>>>> >>>>> I am having errors compiling the new Nek5000. I am trying to run nekton in a unix cluster. I get the following error, the log file being attached - >>>>> >>>>> src/comm.h(65): catastrophic error: cannot open source file "mpi.h" >>>>> #include >>>>> >>>>> Thanks, >>>>> >>>>> Saikat >>>>> >>>>> Saikat Mukherjee, >>>>> PhD Student, >>>>> Paul Research Group - http://www.me.vt.edu/mpaul/ > >>>>> Engineering Science and Mechanics, >>>>> Virginia Tech. >>>>> _______________________________________________ >>>>> 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 >>>> >>>> _______________________________________________ >>>> 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 >>> >> _______________________________________________ >> 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 From nek5000-users at lists.mcs.anl.gov Thu Dec 21 07:59:32 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 13:59:32 +0000 Subject: [Nek5000-users] DIV(V)-QTL too large! Message-ID: Hi all, There is always DIV(V)-QTL too large for my flow case until step 42 after which it shows correct. I don't think my mesh needs more resolution (elements number more than 36000)and time step is very short in the first timestep which is DT= 1.0000000E-05. Do you think the results are reliable? Kind regards, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 21 09:51:25 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 16:51:25 +0100 Subject: [Nek5000-users] DIV(V)-QTL too large! Message-ID: You'll get this warning because the error in divergence is larger than 1e-2. Typically the error depends strongly on the spatial resolution but also on other things e.g. step size dt, solver tolerances, boundary conditions ... In your case it looks like the error decays rapidly in time. I guess this is because of your your initial condition? What error levels do you see after a couple of hundred steps? -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Thursday 21st December 2017 15:00 > To: nek5000-users at lists.mcs.anl.gov > Subject: [Nek5000-users] DIV(V)-QTL too large! > > Hi all, > > There is always?DIV(V)-QTL too large for my flow case until step 42 after which it shows correct. I dont think my mesh needs more resolution (elements number more than 36000)and time step is very short in the first timestep which is?DT= 1.0000000E-05.? > > Do you think the results are reliable? > > Kind regards, > > Jian > _______________________________________________ > 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 Dec 21 16:13:47 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 21 Dec 2017 22:13:47 +0000 Subject: [Nek5000-users] DIV(V)-QTL too large! Message-ID: Hi, The error in first time step is L1/L2 DIV(V) : 3.0955E-17 7.4868E+00 L1/L2 QTL : 0.0000E+00 0.0000E+00 L1/L2 DIV(V)-QTL: 3.0955E-17 7.4868E+00 WARNING: DIV(V)-QTL too large! After 41 steps, it shows L1/L2 DIV(V) : 4.1552E-17 9.8458E-02 L1/L2 QTL : 0.0000E+00 0.0000E+00 L1/L2 DIV(V)-QTL: 4.1552E-17 9.8458E-02 Do you think the flow running now is correct for the future running? Kind regards, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Fri Dec 22 02:56:36 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 09:56:36 +0100 Subject: [Nek5000-users] DIV(V)-QTL too large! In-Reply-To: References: Message-ID: What are the error levels after a couple of hundred steps? -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Thursday 21st December 2017 23:14 > To: nek5000-users at lists.mcs.anl.gov > Subject: Re: [Nek5000-users] DIV(V)-QTL too large! > > Hi, > > The error in first time step is? > ? ? ? ? ? ? ? ?L1/L2 DIV(V)? ? :? ?3.0955E-17? ?7.4868E+00 > ? ? ? ? ? ? ? ?L1/L2 QTL? ? ? ?:? ?0.0000E+00? ?0.0000E+00 > ? ? ? ? ? ? ? ?L1/L2 DIV(V)-QTL:? ?3.0955E-17? ?7.4868E+00 > ? ? ? ? ? ? ? ?WARNING: DIV(V)-QTL too large! > After 41 steps, it shows > ? ? ? ? ? ? ? ?L1/L2 DIV(V)? ? :? ?4.1552E-17? ?9.8458E-02 > ? ? ? ? ? ? ? ?L1/L2 QTL? ? ? ?:? ?0.0000E+00? ?0.0000E+00 > ? ? ? ? ? ? ? ?L1/L2 DIV(V)-QTL:? ?4.1552E-17? ?9.8458E-02 > > Do you think the flow running now is correct for the future running? > > Kind regards, > > Jian > _______________________________________________ > 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 Dec 22 06:41:07 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 20:41:07 +0800 (GMT+08:00) Subject: [Nek5000-users] =?utf-8?q?Reynolds_number_setting_in_the_turbChan?= =?utf-8?q?nel_example?= Message-ID: Hello Neks I am a new user of Nek5000 and my long term objective is to develop and implement some LES models in NEK5000. Now I am working with the turbChannel example. I want to make sure that the original parameter setting in the example is corresponding to Re_tau = 590 ? When the Reynolds number (p02) is set to 10935, my results show that the Re_tau is lower than 590. The second question is that, if I want to simulate Re_tau = 180 case, what the Reynolds number (p02) should be? My understanding is that since the U_bar in the turbChannel is equal to 1, the the Reynolds number (p02) should be 2800, which is the Re base on bulk mean velocity in J Kim, P Moin and R Moser 1986 JFM paper. Any responds is greatly appreciated. Regards Zhenrong JING PhD ?cole centrale de Nantes -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Fri Dec 22 07:14:17 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 13:14:17 +0000 Subject: [Nek5000-users] Reynolds number setting in the turbChannel example In-Reply-To: References: Message-ID: Dear Zhenrong, I use one of the following to make the conversions: Re_tau = .166 Re_b^0.88 Re_tau = 0.173 Re_b^0.875 per the attached. I apologize that I don't have the references for these formulas -- probably in Schlicting but I cannot recall at the moment. Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Friday, December 22, 2017 6:41:07 AM To: nek5000-users at lists.mcs.anl.gov Subject: [Nek5000-users] Reynolds number setting in the turbChannel example Hello Neks I am a new user of Nek5000 and my long term objective is to develop and implement some LES models in NEK5000. Now I am working with the turbChannel example. I want to make sure that the original parameter setting in the example is corresponding to Re_tau = 590 ? When the Reynolds number (p02) is set to 10935, my results show that the Re_tau is lower than 590. The second question is that, if I want to simulate Re_tau = 180 case, what the Reynolds number (p02) should be? My understanding is that since the U_bar in the turbChannel is equal to 1, the the Reynolds number (p02) should be 2800, which is the Re base on bulk mean velocity in J Kim, P Moin and R Moser 1986 JFM paper. Any responds is greatly appreciated. Regards Zhenrong JING PhD ?cole centrale de Nantes -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: re_tau2.pdf Type: application/pdf Size: 58662 bytes Desc: re_tau2.pdf URL: From nek5000-users at lists.mcs.anl.gov Fri Dec 22 07:26:14 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 13:26:14 +0000 Subject: [Nek5000-users] DIV(V)-QTL too large! Message-ID: Hi, It shows no warning after 41 steps and runs as follows: Step 131, t= 2.5248861E-01, DT= 2.4021447E-03, C= 0.500 1.1525E+03 7.3667E+00 Solving for fluid T T T 131 PRES alph1n 9.0124E-01 4.3599E-02 2.0671E+01 5 131 PRES halpha 5 7.5261E-03 5.5946E-03 2.3052E-03 2.1843E-03 8.0764E-04 131 PRES gmres: 7 8.8748E-07 1.0000E-06 3.0075E-05 1.5138E+00 2.1487E+00 F 131 1 5.4502E-01 5.4502E-01 1.0000E+00 0 alph1x 131 halpha 0 131 Hmholtz VELX: 5 2.3325E-08 5.4502E-01 1.0000E-07 131 2 2.1461E-01 2.1461E-01 1.0000E+00 0 alph1x 131 halpha 0 131 Hmholtz VELY: 5 1.8128E-08 2.1461E-01 1.0000E-07 131 3 3.4137E-01 3.4137E-01 1.0000E+00 0 alph1x 131 halpha 0 131 Hmholtz VELZ: 5 3.7031E-08 3.4137E-01 1.0000E-07 L1/L2 DIV(V) : 3.7079E-17 9.8516E-03 L1/L2 QTL : 0.0000E+00 0.0000E+00 L1/L2 DIV(V)-QTL: 3.7079E-17 9.8516E-03 131 PRES gmres: 1 2.6792E-16 1.0000E-06 7.4460E-16 2.1183E-01 3.5189E-01 F 131 Hmholtz VELX: 1 3.4546E-17 1.9046E-15 1.0000E-07 131 Hmholtz VELY: 1 3.4815E-17 1.9137E-15 1.0000E-07 131 Hmholtz VELZ: 5 6.4869E-08 9.9916E-01 1.0000E-07 131 1.02688E-03 1.25000E+01 1.00000E+00 basflow Z 131 0.2524886E+00 9.05577E-03 9.29920E-06 7.85390E-01 7.85399E-01 volflow Z 131 2.5249E-01 7.0023E+00 Fluid done Thnaks, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Fri Dec 22 07:31:24 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 14:31:24 +0100 Subject: [Nek5000-users] DIV(V)-QTL too large! In-Reply-To: References: Message-ID: Ok it looks like it decays in time. What errors do you see say after 500 steps? -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Friday 22nd December 2017 14:27 > To: nek5000-users at lists.mcs.anl.gov > Subject: Re: [Nek5000-users] DIV(V)-QTL too large! > > Hi, > > It shows no warning after 41 steps and runs as follows: > Step? ? 131, t= 2.5248861E-01, DT= 2.4021447E-03, C=? 0.500 1.1525E+03 7.3667E+00 > ?Solving for fluid T T T > ? ? ? ? 131? ? PRES alph1n? 9.0124E-01? 4.3599E-02? 2.0671E+01? ? ?5 > ? ? ? ? 131? ? PRES halpha? ? ?5? 7.5261E-03? 5.5946E-03? 2.3052E-03? 2.1843E-03? 8.0764E-04 > ? ? ? 131 PRES gmres:? ? 7? 8.8748E-07? 1.0000E-06? 3.0075E-05? 1.5138E+00? 2.1487E+00? ? F > ? ? ?131? 1? 5.4502E-01? 5.4502E-01? 1.0000E+00? ?0 alph1x > ? ?131 halpha? ?0 > ? ? ? ? 131? ? Hmholtz VELX:? ? ? 5? ?2.3325E-08? ?5.4502E-01? ?1.0000E-07 > ? ? ?131? 2? 2.1461E-01? 2.1461E-01? 1.0000E+00? ?0 alph1x > ? ?131 halpha? ?0 > ? ? ? ? 131? ? Hmholtz VELY:? ? ? 5? ?1.8128E-08? ?2.1461E-01? ?1.0000E-07 > ? ? ?131? 3? 3.4137E-01? 3.4137E-01? 1.0000E+00? ?0 alph1x > ? ?131 halpha? ?0 > ? ? ? ? 131? ? Hmholtz VELZ:? ? ? 5? ?3.7031E-08? ?3.4137E-01? ?1.0000E-07 > ? ? ? ? ? ? ? ?L1/L2 DIV(V)? ? :? ?3.7079E-17? ?9.8516E-03 > ? ? ? ? ? ? ? ?L1/L2 QTL? ? ? ?:? ?0.0000E+00? ?0.0000E+00 > ? ? ? ? ? ? ? ?L1/L2 DIV(V)-QTL:? ?3.7079E-17? ?9.8516E-03 > ? ? ? 131 PRES gmres:? ? 1? 2.6792E-16? 1.0000E-06? 7.4460E-16? 2.1183E-01? 3.5189E-01? ? F > ? ? ? ? 131? ? Hmholtz VELX:? ? ? 1? ?3.4546E-17? ?1.9046E-15? ?1.0000E-07 > ? ? ? ? 131? ? Hmholtz VELY:? ? ? 1? ?3.4815E-17? ?1.9137E-15? ?1.0000E-07 > ? ? ? ? 131? ? Hmholtz VELZ:? ? ? 5? ?6.4869E-08? ?9.9916E-01? ?1.0000E-07 > ? ? ? 131? 1.02688E-03? 1.25000E+01? 1.00000E+00 basflow Z > ? ? ?131 0.2524886E+00? 9.05577E-03? 9.29920E-06? 7.85390E-01? 7.85399E-01 volflow Z > ? ? ? ? 131? ?2.5249E-01? 7.0023E+00 Fluid done > > Thnaks, > > Jian > _______________________________________________ > 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 Dec 22 07:33:19 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 13:33:19 +0000 Subject: [Nek5000-users] Reynolds number setting in the turbChannel example Message-ID: Hi Zhenrong, I used 2800 for shear Reynolds number 180 in channel flow and got a good agreement with Kim et al. (1987). Hope it work to you. -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Fri Dec 22 07:40:46 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 14:40:46 +0100 Subject: [Nek5000-users] Reynolds number setting in the turbChannel example Message-ID: Re_tau is a simulation result and not known a-priori i.e. the number below holds for a DNS but not for an LES. -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Friday 22nd December 2017 14:33 > To: nek5000-users at lists.mcs.anl.gov > Subject: Re: [Nek5000-users] Reynolds number setting in the turbChannel example > > Hi Zhenrong, > > I used 2800 for shear Reynolds number 180 in channel flow and got a good agreement with?Kim et al. (1987). > > Hope it work to you. > > > _______________________________________________ > 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 Dec 22 08:19:38 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 14:19:38 +0000 Subject: [Nek5000-users] Reynolds number setting in the turbChannel example In-Reply-To: References: Message-ID: Yes -- but presumably if the LES is at all accurate then the mean wall shear should be reasonably close to DNS or it has otherwise missed an important quantity of interest? Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Friday, December 22, 2017 7:40:46 AM To: nek5000-users at lists.mcs.anl.gov Subject: Re: [Nek5000-users] Reynolds number setting in the turbChannel example Re_tau is a simulation result and not known a-priori i.e. the number below holds for a DNS but not for an LES. -----Original message----- > From:nek5000-users at lists.mcs.anl.gov > Sent: Friday 22nd December 2017 14:33 > To: nek5000-users at lists.mcs.anl.gov > Subject: Re: [Nek5000-users] Reynolds number setting in the turbChannel example > > Hi Zhenrong, > > I used 2800 for shear Reynolds number 180 in channel flow and got a good agreement with Kim et al. (1987). > > Hope it work to you. > > > _______________________________________________ > 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 -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Fri Dec 22 08:59:52 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 15:59:52 +0100 Subject: [Nek5000-users] Reynolds number setting in the turbChannel example In-Reply-To: References: Message-ID: Hi, the formula mentioned by Paul can be found in Pope's book (Dean's correlation). Indeed, if Re_tau is off by more than a few percent (say 10) then the quality of the LES is probably not good. Typically, I would expect a dynamic Smagorinsky model to lead to an underprediction of the shear. Philipp On 2017-12-22 15:19, nek5000-users at lists.mcs.anl.gov wrote: > > Yes -- but presumably if the LES is at all accurate then the mean wall > shear should be reasonably close to DNS or it has otherwise missed an > important quantity of interest? > > > Paul > > > ------------------------------------------------------------------------ > *From:* Nek5000-users on > behalf of nek5000-users at lists.mcs.anl.gov > *Sent:* Friday, December 22, 2017 7:40:46 AM > *To:* nek5000-users at lists.mcs.anl.gov > *Subject:* Re: [Nek5000-users] Reynolds number setting in the > turbChannel example > Re_tau is a simulation result and not known a-priori i.e. the number > below holds for a DNS but not for an LES. > > > -----Original message----- >> From:nek5000-users at lists.mcs.anl.gov >> Sent: Friday 22nd December 2017 14:33 >> To: nek5000-users at lists.mcs.anl.gov >> Subject: Re: [Nek5000-users] Reynolds number setting in the turbChannel example >> >> Hi Zhenrong, >> >> I used 2800 for shear Reynolds number 180 in channel flow and got a good agreement with?Kim et al. (1987). >> >> Hope it work to you. >> >> >> _______________________________________________ >> 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 > > > _______________________________________________ > 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 Dec 22 09:18:09 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 15:18:09 +0000 Subject: [Nek5000-users] Reynolds number setting in the turbChannel example In-Reply-To: References: , Message-ID: One additional comment --- I generally find PnPn to do a better job of LES than PnPn-2, presumably because of the better accuracy for pressure. Also, interestingly, you can look both at integrated tau_wall and at mean dp/dx for the LES. Generally, these are in better agreement for PnPn (based on my past experience...) If you are running constant flow rate (p54/55 nonzero) then grep volf logfile > t.t and plot col 3 of t.t vs col 1 of t.t will show the mean pressure drop, assuming that the body force set in userf is == 0. I apologize that the volf line does not also show the time, but only the time step number. (We should correct that...) Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Friday, December 22, 2017 8:59:52 AM To: nek5000-users at lists.mcs.anl.gov Subject: Re: [Nek5000-users] Reynolds number setting in the turbChannel example Hi, the formula mentioned by Paul can be found in Pope's book (Dean's correlation). Indeed, if Re_tau is off by more than a few percent (say 10) then the quality of the LES is probably not good. Typically, I would expect a dynamic Smagorinsky model to lead to an underprediction of the shear. Philipp On 2017-12-22 15:19, nek5000-users at lists.mcs.anl.gov wrote: > > Yes -- but presumably if the LES is at all accurate then the mean wall > shear should be reasonably close to DNS or it has otherwise missed an > important quantity of interest? > > > Paul > > > ------------------------------------------------------------------------ > *From:* Nek5000-users on > behalf of nek5000-users at lists.mcs.anl.gov > *Sent:* Friday, December 22, 2017 7:40:46 AM > *To:* nek5000-users at lists.mcs.anl.gov > *Subject:* Re: [Nek5000-users] Reynolds number setting in the > turbChannel example > Re_tau is a simulation result and not known a-priori i.e. the number > below holds for a DNS but not for an LES. > > > -----Original message----- >> From:nek5000-users at lists.mcs.anl.gov >> Sent: Friday 22nd December 2017 14:33 >> To: nek5000-users at lists.mcs.anl.gov >> Subject: Re: [Nek5000-users] Reynolds number setting in the turbChannel example >> >> Hi Zhenrong, >> >> I used 2800 for shear Reynolds number 180 in channel flow and got a good agreement with Kim et al. (1987). >> >> Hope it work to you. >> >> >> _______________________________________________ >> 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 > > > _______________________________________________ > 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 -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Fri Dec 22 09:21:01 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 15:21:01 +0000 Subject: [Nek5000-users] Reynolds number setting in the turbChannel example In-Reply-To: References: Message-ID: Hi Can on help me to write a matlab code for mortar legendre spectral finite element method in two domains to solve poisson equation. I m tired to imposed the non-matching conditions Thank you Radouen -----Message d'origine----- De?: Nek5000-users [mailto:nek5000-users-bounces at lists.mcs.anl.gov] De la part de nek5000-users at lists.mcs.anl.gov Envoy??: vendredi 22 d?cembre 2017 16:00 ??: nek5000-users at lists.mcs.anl.gov Objet?: Re: [Nek5000-users] Reynolds number setting in the turbChannel example Hi, the formula mentioned by Paul can be found in Pope's book (Dean's correlation). Indeed, if Re_tau is off by more than a few percent (say 10) then the quality of the LES is probably not good. Typically, I would expect a dynamic Smagorinsky model to lead to an underprediction of the shear. Philipp On 2017-12-22 15:19, nek5000-users at lists.mcs.anl.gov wrote: > > Yes -- but presumably if the LES is at all accurate then the mean wall > shear should be reasonably close to DNS or it has otherwise missed an > important quantity of interest? > > > Paul > > > ---------------------------------------------------------------------- > -- > *From:* Nek5000-users on > behalf of nek5000-users at lists.mcs.anl.gov > > *Sent:* Friday, December 22, 2017 7:40:46 AM > *To:* nek5000-users at lists.mcs.anl.gov > *Subject:* Re: [Nek5000-users] Reynolds number setting in the > turbChannel example Re_tau is a simulation result and not known > a-priori i.e. the number below holds for a DNS but not for an LES. > > > -----Original message----- >> From:nek5000-users at lists.mcs.anl.gov >> >> Sent: Friday 22nd December 2017 14:33 >> To: nek5000-users at lists.mcs.anl.gov >> Subject: Re: [Nek5000-users] Reynolds number setting in the >> turbChannel example >> >> Hi Zhenrong, >> >> I used 2800 for shear Reynolds number 180 in channel flow and got a good agreement with?Kim et al. (1987). >> >> Hope it work to you. >> >> >> _______________________________________________ >> 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 > > > _______________________________________________ > 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 From nek5000-users at lists.mcs.anl.gov Fri Dec 22 09:22:02 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 15:22:02 +0000 Subject: [Nek5000-users] Mortar_spectral_element_method Message-ID: Hi Can on help me to write a matlab code for mortar legendre spectral finite element method in two domains to solve poisson equation. I m tired to imposed the non-matching conditions Thank you Radouen From nek5000-users at lists.mcs.anl.gov Fri Dec 22 09:46:56 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 15:46:56 +0000 Subject: [Nek5000-users] DIV(V)-QTL too large! Message-ID: Hi, After 500 and 3000, it shows, Step 517, t= 1.1788322E+00, DT= 2.3973162E-03, C= 0.500 3.9951E+03 6.9817E+00 Solving for fluid T T T 517 PRES alph1n 8.7302E-02 6.4306E-03 1.3576E+01 11 517 PRES halpha 11 1.1966E-03 3.4135E-04 3.5797E-04 3.1090E-04 3.1099E-04 2.3543E-04 1.9043E-04 2.0750E-04 1.7930E-04 1.2367E-04 517 PRES gmres: 4 8.3071E-07 1.0000E-06 4.9308E-06 8.4564E-01 1.1845E+00 F 517 1 1.7008E-01 1.7008E-01 1.0000E+00 0 alph1x 517 halpha 0 517 Hmholtz VELX: 5 6.7578E-09 1.7008E-01 1.0000E-07 517 2 1.1899E-01 1.1899E-01 1.0000E+00 0 alph1x 517 halpha 0 517 Hmholtz VELY: 5 6.5549E-09 1.1899E-01 1.0000E-07 517 3 2.4613E-01 2.4613E-01 1.0000E+00 0 alph1x 517 halpha 0 517 Hmholtz VELZ: 5 2.1073E-08 2.4613E-01 1.0000E-07 L1/L2 DIV(V) : 6.3228E-18 1.6443E-03 L1/L2 QTL : 0.0000E+00 0.0000E+00 L1/L2 DIV(V)-QTL: 6.3228E-18 1.6443E-03 517 PRES gmres: 1 2.6792E-16 1.0000E-06 7.4460E-16 2.1149E-01 3.5070E-01 F 517 Hmholtz VELX: 1 3.4482E-17 1.9046E-15 1.0000E-07 517 Hmholtz VELY: 1 3.4750E-17 1.9137E-15 1.0000E-07 517 Hmholtz VELZ: 5 6.4296E-08 9.9916E-01 1.0000E-07 517 1.02488E-03 1.25000E+01 1.00000E+00 basflow Z 517 0.1178832E+01 6.94484E-03 7.11765E-06 7.85392E-01 7.85399E-01 volflow Z 517 1.1788E+00 6.5287E+00 Fluid done Step 3003, t= 7.0457310E+00, DT= 2.3271524E-03, C= 0.500 2.1994E+04 7.7898E+00 Solving for fluid T T T 3003 PRES alph1n 2.2485E-01 1.2721E-02 1.7676E+01 17 3003 PRES halpha 17 -7.5402E-04 -1.9963E-04 6.1782E-04 7.0527E-04 3.4415E-04 7.8125E-05 6.1111E-04 1.2665E-03 9.2238E-04 1.1935E-03 3003 PRES gmres: 5 8.4347E-07 1.0000E-06 9.5886E-06 1.0585E+00 1.4770E+00 F 3003 1 2.4280E-01 2.4280E-01 1.0000E+00 0 alph1x 3003 halpha 0 3003 Hmholtz VELX: 4 6.7528E-08 2.4280E-01 1.0000E-07 3003 2 9.1891E-02 9.1891E-02 1.0000E+00 0 alph1x 3003 halpha 0 3003 Hmholtz VELY: 4 5.1734E-08 9.1891E-02 1.0000E-07 3003 3 1.9637E-01 1.9637E-01 1.0000E+00 0 alph1x 3003 halpha 0 3003 Hmholtz VELZ: 5 9.5563E-09 1.9637E-01 1.0000E-07 L1/L2 DIV(V) : -2.5507E-17 1.8315E-03 L1/L2 QTL : 0.0000E+00 0.0000E+00 L1/L2 DIV(V)-QTL: -2.5507E-17 1.8315E-03 3003 PRES gmres: 1 2.6792E-16 1.0000E-06 7.4460E-16 2.1105E-01 3.5033E-01 F 3003 Hmholtz VELX: 1 3.3517E-17 1.9046E-15 1.0000E-07 3003 Hmholtz VELY: 1 3.3777E-17 1.9137E-15 1.0000E-07 3003 Hmholtz VELZ: 5 5.6177E-08 9.9916E-01 1.0000E-07 3003 9.95014E-04 1.25000E+01 1.00000E+00 basflow Z 3003 0.7045731E+01 5.43569E-03 5.40859E-06 7.85394E-01 7.85399E-01 volflow Z 3003 7.0457E+00 7.3717E+00 Fluid done Compute statistics ... -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Fri Dec 22 11:36:31 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 17:36:31 +0000 Subject: [Nek5000-users] MAXNEL Message-ID: Hi Thilina, I set maketools file as followed. Then, I did './maketools genmap' even './maketools all' after which all files were updated, but they all didnt work. Do you have any idea? #!/bin/bash # maximum number of elements MAXNEL=350000 # binary path #bin_nek_tools=`cd ../bin && pwd` # Fortran compiler #FC="gfortran" #FFLAGS="" # C compiler #CC="gcc" #CFLAGS="" # linking flags #LDFLAGS="-L$HOME/lib -lm" Best wishes, Z -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Fri Dec 22 14:41:17 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 22 Dec 2017 21:41:17 +0100 Subject: [Nek5000-users] MAXNEL In-Reply-To: References: Message-ID: Did you clean before recompiling? On 22 Dec 2017, at 18:37, "nek5000-users at lists.mcs.anl.gov " > wrote: Hi?Thilina, I set maketools file as followed.?Then, I did './maketools genmap' even './maketools all' after which all files were updated, but they all didnt work. Do you have any idea? #!/bin/bash # maximum number of elements ? ? ? MAXNEL=350000 # binary path #bin_nek_tools=`cd ../bin && pwd` # Fortran compiler #FC="gfortran" #FFLAGS="" # C compiler #CC="gcc" #CFLAGS="" # linking flags #LDFLAGS="-L$HOME/lib -lm" Best wishes, Z _______________________________________________ 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 Dec 26 22:20:29 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 27 Dec 2017 01:20:29 -0300 Subject: [Nek5000-users] Set gradient and/or pressure at outlet Message-ID: Dear Nek users, I've been struggling to program an adecuate outlet BC. I simulate a vertical isothermal surface in an open domain, forming a free convection boundary layer. When there are vortex leaving the domain, the code diverges. I've tried turb_outflow subroutine, stabilized outflow BC (Dong 2014), convective BC (sommerfield radiation condition, thanks to Phillip and Eric, from this comunity), etc. But I can't achieve to make the code run propperly, and I can't go on with my research while I have this problem. There is a paper, from Desrayaud (2013), "Benchmark solutions for natural convection flows in vertical channels submitted to different open boundary conditions". I want to try these BC, which can be stated as: (x: horizontal and y:vertical.) At the top face (outlet) Ux=0 dUy/dy=0 If Uy > 0 (fluid going out) dTEMP/dy=0 pressure=0 If Uy < 0 (fluid going in because there is backflow) TEMP=0 pressure= -0.5Uy^2 or 0 (these are two options in the paper) As far as I know, in Pn-Pn-2 formulation there are not pressure BC, but it is possible to set it using Pn-Pn. Am I right? If so, how can I set BC for pressure in Pn-Pn? And for the velocity gradient and temperature gradient equal cero at the top face, how can it be done? Would it be enough to set dirichlet BC and replicate the last row of values to the last one? Any help is welcomed, thanks in advance, Juan Pablo. -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Tue Dec 26 23:00:13 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 27 Dec 2017 05:00:13 +0000 Subject: [Nek5000-users] Set gradient and/or pressure at outlet In-Reply-To: References: Message-ID: Dear Juan Pablo, If you are doing a Boussinesq approximation with an open boundary you probably need to subtract off the hydrostatic mode from the pressure, given that pressure in that case is far from uniform in the vertical. Please see: http://www.mcs.anl.gov/~fischer/users.pdf eqs. (13) and (17) for an example of how to do this. The key is to have the "pressure" variable be close to uniform at outflow. Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Tuesday, December 26, 2017 10:20:29 PM To: nek5000-users at lists.mcs.anl.gov Subject: [Nek5000-users] Set gradient and/or pressure at outlet Dear Nek users, I've been struggling to program an adecuate outlet BC. I simulate a vertical isothermal surface in an open domain, forming a free convection boundary layer. When there are vortex leaving the domain, the code diverges. I've tried turb_outflow subroutine, stabilized outflow BC (Dong 2014), convective BC (sommerfield radiation condition, thanks to Phillip and Eric, from this comunity), etc. But I can't achieve to make the code run propperly, and I can't go on with my research while I have this problem. There is a paper, from Desrayaud (2013), "Benchmark solutions for natural convection flows in vertical channels submitted to different open boundary conditions". I want to try these BC, which can be stated as: (x: horizontal and y:vertical.) At the top face (outlet) Ux=0 dUy/dy=0 If Uy > 0 (fluid going out) dTEMP/dy=0 pressure=0 If Uy < 0 (fluid going in because there is backflow) TEMP=0 pressure= -0.5Uy^2 or 0 (these are two options in the paper) As far as I know, in Pn-Pn-2 formulation there are not pressure BC, but it is possible to set it using Pn-Pn. Am I right? If so, how can I set BC for pressure in Pn-Pn? And for the velocity gradient and temperature gradient equal cero at the top face, how can it be done? Would it be enough to set dirichlet BC and replicate the last row of values to the last one? Any help is welcomed, thanks in advance, Juan Pablo. -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 27 19:28:25 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 27 Dec 2017 22:28:25 -0300 Subject: [Nek5000-users] Set gradient and/or pressure at outlet Message-ID: Thanks Paul for the answer. Yes, I'm using Boussinesq approximation, but using incompresible equations, there are no stratification in the density, like the equations you mention. But, anyway, I'm more interested in setting the spatial gradient equal to cero in the boundary. ?How can I achieve that? dUy/dy=0 dTEMP/dy=0 Thanks a lot, Juan Pablo. -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 27 19:36:06 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 28 Dec 2017 01:36:06 +0000 Subject: [Nek5000-users] Set gradient and/or pressure at outlet In-Reply-To: References: Message-ID: Hi Juan Pablo, Perhaps I might have misinterpreted the documentation, but wouldnt the SYM boundary condition achieve the zero gradient conditions you are looking for? On Wed, Dec 27, 2017, 8:28 PM wrote: > Thanks Paul for the answer. Yes, I'm using Boussinesq approximation, but > using incompresible equations, there are no stratification in the density, > like the equations you mention. But, anyway, I'm more interested in setting > the spatial gradient equal to cero in the boundary. ?How can I achieve that? > > dUy/dy=0 > > dTEMP/dy=0 > > > Thanks a lot, > Juan Pablo. > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > -- *Amitvikram Dutta* Graduate Research Assistant Fluid Mechanics Research Lab Multi-Physics Interaction Lab University of Waterloo -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Wed Dec 27 20:35:41 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 28 Dec 2017 02:35:41 +0000 Subject: [Nek5000-users] Set gradient and/or pressure at outlet In-Reply-To: References: Message-ID: Dear Juan Pablo, The documentation I pointed you to is for constant density Boussinesq and the change in density to which I refer is the first-order perturbation that sets up the free convection. (Otherwise, you have no natural convection.) If you have outflow conditions you will need to do as suggested in that document. If you don't have outflow conditions then there is something else going on. Since you talk about vortices leaving the domain I was assuming you have outflow, but maybe that's not the case. If you'd like, feel free to send me the case files off-list and I can take a look. Paul ________________________________ From: Nek5000-users on behalf of nek5000-users at lists.mcs.anl.gov Sent: Wednesday, December 27, 2017 7:28:25 PM To: nek5000-users at lists.mcs.anl.gov Subject: Re: [Nek5000-users] Set gradient and/or pressure at outlet Thanks Paul for the answer. Yes, I'm using Boussinesq approximation, but using incompresible equations, there are no stratification in the density, like the equations you mention. But, anyway, I'm more interested in setting the spatial gradient equal to cero in the boundary. ?How can I achieve that? dUy/dy=0 dTEMP/dy=0 Thanks a lot, Juan Pablo. -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 28 11:45:20 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 28 Dec 2017 17:45:20 +0000 Subject: [Nek5000-users] avg_all post processing Message-ID: Hi Nek, I am struggling with a problem in post-processing the files generated by calling avg_all (see https://lists.mcs.anl.gov/mailman/htdig/nek5000-users/2017-December/004832.html). I used the attached usr files to get the time averaged velocity profile and vms. But until then, it worked with wrong results. The usr file generated only one visit file and the number scale of averaged streamwise velocity is not correct. Questions: 1. May I how I can have a velocity profile dat file and rms file by using the avg*0.f000.. rms*0.f000...and rm2*0.f000, rather than a visit file. 2. Would you please help me check the attached usr file whether correct. Kind regards, ZJ -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: pipe.usr Type: application/octet-stream Size: 9713 bytes Desc: not available URL: From nek5000-users at lists.mcs.anl.gov Sun Dec 31 12:34:36 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Sun, 31 Dec 2017 18:34:36 +0000 Subject: [Nek5000-users] Different velocity values in different 'v' inlets Message-ID: Hi all, I'm trying to simulate a shear-layer which involves two different jets emerging at different velocities. The simulation is 2D for now. The width of the first jet extends from y=0.0 to y=1.0 and the second extends from y=1.0 to y=3.0 I generated a mesh using genbox with 'v' BCs for both inlets. In the .usr file I have the following form for usrbc subroutine userbc (ix,iy,iz,iside,ieg) c NOTE ::: This subroutine MAY NOT be called by every process include 'SIZE' include 'TOTAL' include 'NEKUSE' if (ym1(i,j,k,e).lt.1) then ux=10.0 elseif (ym1(i,j,k,e).gt.1) then ux=1.0 endif ux=1.0 uy=0.0 uz=0.0 temp=0.0 return end As you can see I try assigning ux = 10.0 and 1.0 to the different zones. However the results show a uniform value of 1.0 across both the inlets. What am I missing here? Sincerely, -- *Amitvikram Dutta* Graduate Research Assistant Fluid Mechanics Research Lab Multi-Physics Interaction Lab University of Waterloo -------------- next part -------------- An HTML attachment was scrubbed... URL: From nek5000-users at lists.mcs.anl.gov Sun Dec 31 13:18:06 2017 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Sun, 31 Dec 2017 20:18:06 +0100 Subject: [Nek5000-users] Different velocity values in different 'v' inlets In-Reply-To: References: Message-ID: What is the variable e in userbc? On 31 Dec 2017, at 19:35, "nek5000-users at lists.mcs.anl.gov " > wrote: Hi all, I'm trying to simulate a shear-layer which involves two different jets emerging at different velocities. The simulation is 2D for now. The width of the first jet extends from y=0.0 to y=1.0 and the second extends from y=1.0 to y=3.0 I generated a mesh using genbox with 'v' BCs for both inlets. In the .usr file I have the following form for usrbc ?subroutine userbc (ix,iy,iz,iside,ieg) c???? NOTE ::: This subroutine MAY NOT be called by every process ????? include 'SIZE' ????? include 'TOTAL' ????? include 'NEKUSE' ????? if (ym1(i,j,k,e).lt.1) then ????????? ux=10.0 ????? elseif (ym1(i,j,k,e).gt.1) then ??? ??? ? ux=1.0 ????? endif ????? ux=1.0 ????? uy=0.0 ????? uz=0.0 ????? temp=0.0 ????? return ????? end As you can see I try assigning ux = 10.0 and 1.0 to the different zones. However the results show a uniform value of 1.0 across both the inlets. What am I missing here? Sincerely, -- Amitvikram Dutta Graduate Research Assistant Fluid Mechanics Research Lab Multi-Physics Interaction Lab University of Waterloo _______________________________________________ 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: