Question regarding transient calculations

60 views
Skip to first unread message

Ferraro, Diego Ernesto (INR)

unread,
Aug 23, 2018, 2:49:49 AM8/23/18
to was...@seamplex.com

Dear milongers & wasorers:

  I’m dealing with some transient calculations using Serpent MC code (which also includes some simplified TH feedback). I’ve done some results comparisons with point kinetics results (obtained from a wasora PK model using parameters obtained by Serpent) with some reasonable agreement. But the model I’m dealing with is a minicore, thus I suppose that leakages are too important (my PK models results predict higher power).

Thus I would like to move to a further step and compare with a Milonga 3d diffusion model . I’ve read the SP-MI-17-93D3 document (“One and two dimensional diffusion kinetics benchmarks”) and I understand that some of this stuff is already available. Thus my questions are basically:

a.       I understand that 3-D diffusion with kinetics is available in Milonga. It is that correct ? (I’ve seen only 2D or 1D cases).

b.       Is there available input examples of this? I’ve check (at a glance) at repository, but I couldn’t find anything.

So, any help is appreciated! Thanks in advance,

Best

Diego

 

César Pablo Camusso

unread,
Aug 23, 2018, 7:24:23 AM8/23/18
to was...@seamplex.com
Dear Diego,
We were working in it last year but it is not included in the official milonga repository because I have to make some improvements.
Tha main one is that I built the equations systems with the precursor concentrations in implicit scheme, so with 2 energy groups and 6 precursor groups, the number of equations are 3 times more (6/2) than in the static case (2 energy gruops). This in 3D makes the calculation too slow. The improvement will be to set the precursor equations in explicit mode.
However, if you have a powerfull computer in:
iepale / milo-tra — Bitbucket
is the repository and it sould work.
There are examples in the examples directory. There, there are all the cases shown in the report you have cited.
For example B143d.mil is in 3D.
Probably milonga is going to send a error mesage regarding to the MESH MAIN because that improvement in wasora hapended after I wrote the input files. In that case please write to the list and we will correct it.

Best regards.



--
You received this message because you are subscribed to the Google Groups "wasora" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wasora+un...@seamplex.com.
To post to this group, send email to was...@seamplex.com.
Visit this group at https://groups.google.com/a/seamplex.com/group/wasora/.
To view this discussion on the web visit https://groups.google.com/a/seamplex.com/d/msgid/wasora/5547c40f7c744e66bd14246fb57b8e67%40kit-msx-25.kit.edu.
For more options, visit https://groups.google.com/a/seamplex.com/d/optout.

Ferraro, Diego Ernesto (INR)

unread,
Aug 23, 2018, 7:30:54 AM8/23/18
to was...@seamplex.com

Thank you very much Pablo!

I’ll download that and make some tests based on your commit. When I’ve some results I’ll tell you.

Thanks again!

Diego

César Pablo Camusso

unread,
Aug 23, 2018, 7:33:06 AM8/23/18
to was...@seamplex.com
I forgot to tell you that the transients were done only for diffusion and finite elements scheme.
You are welcome.

Hector Lestani

unread,
Aug 23, 2018, 7:50:13 AM8/23/18
to was...@seamplex.com
Dear Dieguito,
I reply between the lines as I am reading...


On 23/08/18 03:49, Ferraro, Diego Ernesto (INR) wrote:

Dear milongers & wasorers:

  I’m dealing with some transient calculations using Serpent MC code (which also includes some simplified TH feedback). I’ve done some results comparisons with point kinetics results (obtained from a wasora PK model using parameters obtained by Serpent) with some reasonable agreement. But the model I’m dealing with is a minicore, thus I suppose that leakages are too important (my PK models results predict higher power).

I am not sure that the existence of leakages leads to overestimation in Point Kinetics (PK). Leakages keep the same proportion to production and absorption regardless the power. The only exception being reactivities insertion with high flux profile modifications (i.e. big control rod insertion), in which case, over or under estimation would depend on the flux being flattened or peakened. Am I clear? Do you agree?


Thus I would like to move to a further step and compare with a Milonga 3d diffusion model . I’ve read the SP-MI-17-93D3 document (“One and two dimensional diffusion kinetics benchmarks”) and I understand that some of this stuff is already available. Thus my questions are basically:

a.       I understand that 3-D diffusion with kinetics is available in Milonga. It is that correct ? (I’ve seen only 2D or 1D cases).

I don't know.

b.       Is there available input examples of this? I’ve check (at a glance) at repository, but I couldn’t find anything.

Neither.

So, any help is appreciated! Thanks in advance,

Best

Diego

Good luck with that.
Héctor.

César Pablo Camusso

unread,
Aug 23, 2018, 8:09:10 AM8/23/18
to was...@seamplex.com

Ferraro, Diego Ernesto (INR)

unread,
Aug 23, 2018, 8:13:42 AM8/23/18
to was...@seamplex.com

Thanks Pablo! I’ll check that

 

From: 'César Pablo Camusso' via wasora <was...@seamplex.com>
Sent: Donnerstag, 23. August 2018 14:09
To: was...@seamplex.com
Subject: Re: [wasora] Question regarding transient calculations

 

Diego,
I found more data.

Ferraro, Diego Ernesto (INR)

unread,
Aug 23, 2018, 8:22:20 AM8/23/18
to was...@seamplex.com

Actually maybe I made a long shot, but as far as my problem is a 3-D 3x3 PWR Minicore (where I extract a central CR), I supposed that differences should arise as far as PK shouldn’t be able to model that scenario (spatial effects should be appreciable, I suppose). At least that’s was my explanation for the differences with MC calculations, but maybe it’s just another thing.

Nevertheless the idea is use milonga to solve this J.

Thanks anyway!

Diego

 

 

From: Hector Lestani <lest...@cab.cnea.gov.ar>
Sent: Donnerstag, 23. August 2018 13:48
To: was...@seamplex.com
Subject: Re: [wasora] Question regarding transient calculations

 

Dear Dieguito,

César Pablo Camusso

unread,
Aug 23, 2018, 10:41:56 AM8/23/18
to was...@seamplex.com
Diego,
I have updated the milo-tra repository and it is working.
The B143d.mil should work (disregarding the control rod cusp and other programming mistakes). In B143d.geo you will find the NL variable which is the number of axial levels in the core. This benchmark is similar to the yours.
B6A1.mil is updated to be an example if you see an error like this:
pablo@debian:~/petsc/wasora-suite/milo-tra/examples$ ./milonga B6A2.mil
error: B6A2.mil: 16: undefined keyword 'MESH'
The solution is in B6A1.mil.

Ferraro, Diego Ernesto (INR)

unread,
Aug 24, 2018, 3:20:48 AM8/24/18
to was...@seamplex.com

Ferraro, Diego Ernesto (INR)

unread,
Aug 28, 2018, 7:40:14 AM8/28/18
to was...@seamplex.com

Hi Pablo!

I’ve just cloned the repository in https://bitbucket.org/iepale/milo-tra and tried to compile, but it fails due to some lack of libraries. It looks like it’s using a arpack library (that I don’t have). When I do the make (after config w/o problems) I get:

/usr/bin/ld –cannot find -lparpack
I’ve already installed the recommended libraries (sudo apt-get install libgsl-dev libsundials-serial-dev libreadline-dev petsc-dev slepc-dev).  So I suppose I could get this library from somewhere (but the only think I find with that name is just https://www.caam.rice.edu/software/ARPACK// ). So where should I get this library? Or is there any other think I’m missing? 
Thanks in advance! 
Diego

 

 

 

From: 'César Pablo Camusso' via wasora <was...@seamplex.com>

Sent: Donnerstag, 23. August 2018 16:42
To: was...@seamplex.com

César Pablo Camusso

unread,
Aug 28, 2018, 7:45:12 AM8/28/18
to was...@seamplex.com
Dear Diego,
First install the original milonga because it installs all the things needed.
Once you have it installed, milo-tra should work with
./autogen.sh
./configure
make

Ramiro Vignolo

unread,
Aug 28, 2018, 7:47:13 AM8/28/18
to was...@seamplex.com
Hi Diego!

I suspect you the problem is with petsc or slepc (maybe both), not with milonga. I have never installed those libraries using apt get. What operating system do you have? I can explain you how to install them from their repositories.

Thanks!

Ramiro

Ferraro, Diego Ernesto (INR)

unread,
Aug 28, 2018, 8:12:43 AM8/28/18
to was...@seamplex.com

Hi guys! First of all thanks.

I have tried to install the trunk milonga with a git clone from https://bitbucket.org/seamplex/milonga and I get the same error, so I suppose that it comes from or petsc o slepc as you said

I’m using a VM with Ubuntu 16.04.3 LTS. I’ve installed petsc and slepc  with apt-get install without errors:

 

>> sudo apt-get install libgsl-dev libsundials-serial-dev libreadline-dev petsc-dev slepc-dev

 

libgsl-dev is already the newest version (2.1+dfsg-2).

libreadline-dev is already the newest version (6.3-8ubuntu2).

libsundials-serial-dev is already the newest version (2.5.0-3ubuntu3).

petsc-dev is already the newest version (3.6.2.dfsg1-3ubuntu1).

slepc-dev is already the newest version (3.6.1.dfsg1-2build2).

0 upgraded, 0 newly installed, 0 to remove and 191 not upgraded.

 

And the configure of milonga looks ok:

>> ./configure

## ------------------------------------------------------- ##

## Configuration summary for milonga standalone executable ##

## ------------------------------------------------------- ##

  GSL library (required): yes, version 2.1

 

  IDA library (optional): yes, version unknown

    differential-algebraic systems will be solved

 

  Readline library (opt): yes, version 6.3

    run-time debugging-like capabilities will be provided

 

        PETSc (required): yes, /usr/lib/petsc/

    systems of equations represented by sparse matrices will be solved

 

        SLEPc (required): yes, /usr/lib/slepc ()

    sparse generealized eigenproblems will be solved

 

Now proceed to compile with 'make'

 

If I check the directories /usr/lib/petsc/ and usr/lib/slepc everything looks right (the directories exists and are full of .h)

 

So I suppose I’m missing something, but I cannot figure out where.

Thanks,

Diego

Ramiro Vignolo

unread,
Aug 28, 2018, 8:20:19 AM8/28/18
to was...@seamplex.com
I have checked that the missing library is a "Collection of Fortran77 subroutines designed to solve large scale eigenvalue problems". The problem comes from slepc probably. 

1. Uninstall both libraries.
2. Try to re run ./autogen.sh and ./configure and allow milonga to do the work for you. If this fail, I will give you the detailed instructions on how to build those libraries from their corresponding repositories.

Thanks!

Ramiro

César Pablo Camusso

unread,
Aug 28, 2018, 8:21:45 AM8/28/18
to was...@seamplex.com
Do you have something like:
export PETSC_DIR=/home/pablo/petsc/petsc-3.7.4
export PETSC_ARCH=arch-linux2-c-opt
export SLEPC_DIR=/home/pablo/petsc/slepc-3.7.3
in .bashrc?

Ferraro, Diego Ernesto (INR)

unread,
Aug 28, 2018, 8:25:57 AM8/28/18
to was...@seamplex.com

No, I don’t.

So I suppose I have to do the installation as http://www.mcs.anl.gov/petsc/download/index.html ?

Which should I choose? 3.9 ?

Thanks,

César Pablo Camusso

unread,
Aug 28, 2018, 8:30:45 AM8/28/18
to was...@seamplex.com
I think Ramiro's idea is good.
Milonga used to install all.

Ramiro Vignolo

unread,
Aug 28, 2018, 8:36:13 AM8/28/18
to was...@seamplex.com
But if you installed those libraries using apt get they may be already set. Try:

$ echo $PETSC_DIR

However, I if you are planning to install them yourself from bitbucket just select both petsc and slepc branches as 'master' and both versions will be ok. 

First install petsc and then slepc

Thanks!

Ferraro, Diego Ernesto (INR)

unread,
Aug 28, 2018, 8:42:54 AM8/28/18
to was...@seamplex.com

When I uninstall both libraries everything wasora just asks for PETSC after ./configure L.

I’ll try to reinstall using sudo and add the paths to my bashrc.

If  that doesn’t work I will just clone petsc and slepc from git repositories and install by hand.

In any case I’ll tell you how if it works.

Thanks guys!

jeremy theler

unread,
Aug 28, 2018, 9:25:08 AM8/28/18
to was...@seamplex.com
use a newer version of ubuntu and follow the readme


--
--
jeremy theler
www.seamplex.com

Ferraro, Diego Ernesto (INR)

unread,
Aug 29, 2018, 3:38:59 AM8/29/18
to was...@seamplex.com

Hi guys, just for the record:

The sudo apt- … slepc-dev petsc-dev it’s not working in old linux L Ubuntu 16.04.3.

Nevertheless I just eliminated them (sudo apt purge –auto-remove … ) and reinstalled by hand ptsc and slepc:

1. git clone -b maint https://bitbucket.org/petsc/petsc petsc  and so…

2. download http://slepc.upv.es/download/distrib/slepc-3.9.2.tar.gz and so... (warning the git clone version of slepc is development and gives some error when you are trying to compile with petsc).

3. Add paths to .bashrc:  export PETSC_DIR= directory of 1.

                                            export PETSC_ARCH=arch-linux2-c-debug

                                            export SLEPC_DIR= directory of 2.

And voilá! Milonga compiles w/o problems.

Obvioulsly all this stuff could be avoided if a new version of Ubuntu is available.

Thanks everybody for the suggestions!

Diego

Ferraro, Diego Ernesto (INR)

unread,
Aug 29, 2018, 5:44:55 AM8/29/18
to was...@seamplex.com

Hi Pablo:

I’m trying to test your 3D case downloaded from https://bitbucket.org/iepale/milo-tra/src/92fb2f33cc4ac6afb4876eec78128c4484d969b9/examples/?at=master , but when I try to build the .msh I’ve a problem that I suppose comes from the gmsh versions.

If I directly get gmesh from sudo I get the 2.10 version, but it looks like some syntaxes have changed from version to version:

 

Error   : '/home/diego/wasora/milotra/milo-tra/examples/B143d.geo', line 68 : syntax error (Str)

Also I see a geometry that is not plotted correctly (only a partial 2D).

 

I have already tried with different versions of gmsh (looks like wasora does not allow to run with versions above 2.2) without success.

 

So basically, which version of gmsh are you using (or I should use)?

Thanks,

Diego

 

 

From: 'César Pablo Camusso' via wasora <was...@seamplex.com>

Sent: Donnerstag, 23. August 2018 16:42
To: was...@seamplex.com

César Pablo Camusso

unread,
Aug 29, 2018, 7:13:13 AM8/29/18
to was...@seamplex.com
Dear Diego,
I am using one of the lasts versions:
pablo@debian:~$ gmsh -version
--------------------------------------------------------------------------
Petsc Release Version 3.7.5, Jan, 01, 2017
       The PETSc Team
    petsc...@mcs.anl.gov
 http://www.mcs.anl.gov/petsc/
See docs/changes/index.html for recent updates.
See docs/faq.html for problems.
See docs/manualpages/index.html for help.
Libraries linked from /home/geuzaine/src/petsc-3.7.5/complex_mumps_seq/lib
--------------------------------------------------------------------------
3.0.6

Although I see in Gmsh: a three-dimensional finite element mesh generator with built-in pre- and post-processing facilities that there is a version 4.0.0. I ussualy download the already compiled version.
wasora works with versions higher than 2.x.y of gmsh.
Try it:
gmsh -3 B143d.geo
I attached the msh file.



B143d.msh

Ferraro, Diego Ernesto (INR)

unread,
Aug 29, 2018, 7:39:45 AM8/29/18
to was...@seamplex.com

Thanks, with gmsh 3.0.6 works correctly (and wasora does not complain J ).

Best,

César Pablo Camusso

unread,
Aug 29, 2018, 7:40:14 AM8/29/18
to was...@seamplex.com
Diego and German,
I downloaded  the last gmsh version and I got:
pablo@debian:~/petsc/wasora-suite/milo-tra/examples$ ./milonga B143d.mil
error: mesh 'B143d.msh' has an incompatible version '4', only version 2.2 files are supported

By reading http://gmsh.info/CHANGELOG.txt and Gmsh 4.0 I think that the new gmsh format is not supported by wasora yet, as Diego told us.
So, you should use the last 3.x.y gmsh version until it is solved.



Ramiro Vignolo

unread,
Aug 29, 2018, 7:57:52 AM8/29/18
to was...@seamplex.com
Hi,

Happy to hear you have succesfully installed milonga.

Just a minor comment: you have not been able to install slepc from its git repository because you were cheking out the wrong branch. As I previously said, please make sure to checkout the master branch before compiling (in both petsc and slepc).

Thanks!

Ramiro


jeremy theler

unread,
Aug 29, 2018, 1:56:53 PM8/29/18
to was...@seamplex.com
some comments

1. please always use updated software, espcially the operating system and especially regarding nuclear data (remember stuxnet and friends). This a design-basis point, if you do not agree then I recommend not to use wasora/milonga

2. it seems that gmsh 4 saves in msh :v4 by default
2.a. it was released last week and it has been more than a month that I am not working because of an issue with my youngest son health, I did not test it just barely read these emails and the gmsh mailing list
2.b. I said some tike ago "I have to write a v4 msh parser" bit did not have time before, let alone now
2.c. the version of the msh file is not fixed to the gmsh version, you can set the output version of the mesh from the geo file (i.e. tell gmsh 4.0 to write msh2.2), please look up in the gmsh docs how to do it

to sum up, use gmsh4 (and a new ubuntu please!) and ask it to save version 2.2 msh and/or modify src/mesh/gmsh.c in wasora to handle version 4 (I am lacking time and concentration skills now)

another thing, no need to clone petsc/slecp, the tarballs would have done the trick (again, it is better to update the os and use the repos)



César Pablo Camusso

unread,
Aug 29, 2018, 2:53:27 PM8/29/18
to was...@seamplex.com
Dear German and Diego,
Regarding 2.c: in order to obtain msh version 2.2 with gmsh 4 we should write:
Mesh.MshFileVersion = 2.2;
in the geo file.

Ferraro, Diego Ernesto (INR)

unread,
Aug 30, 2018, 3:00:06 AM8/30/18
to was...@seamplex.com

Thanks man. I hope everything is going better.

And no worries, no rush for the gmsh stuff.

Diego

 

BTW, I don’t want to diverge from main topic, but a design-basis question arise for me. There will be a lot of users that probably don’t (and won’t) have sudo attributes (such those running in clusters, cloud servers, institutional infrastructure, etc. etc.). Should wasora/milonga tutorials include some regard in this point?. For example, a short explanation (5-10 lines) in the bitbucket repo for to install milonga w/o sudo privileges (something like à download petsc & slepc à  compileà edit bashrc à download gmesh à  download milongaà  compileà run & have fun) might probably down the entrance barrier for those users without disregard of the main proposed path.

César Pablo Camusso

unread,
Aug 30, 2018, 7:35:54 AM8/30/18
to was...@seamplex.com
If you cannot use sudo, you will have to install the dependences and tools not installed by hand:
m4 make autoconf automake gcc g++ git findutils
libgsl-dev libsundials-serial-dev libreadline-dev petsc-dev slepc-dev

Perhaps make, gcc and g++ will be already installed.
Each dependence will have its trick. So, you will have to go to their repository, git clone it and compile.
Then in wasora and milonga when you execute ./configure you will see what is not installed.



jeremy theler

unread,
Aug 30, 2018, 12:42:03 PM8/30/18
to was...@seamplex.com
yes, ask the admin to install those packages

Reply all
Reply to author
Forward
0 new messages