Serpent-compatible ACE libraries with NJOY

Questions and discussion about applications, input, output and general user topics
andy
Posts: 15
Joined: Fri May 31, 2013 4:56 pm
Security question 1: No
Security question 2: 92

Serpent-compatible ACE libraries with NJOY

Post by andy » Sat Oct 19, 2013 12:30 am

Dear Jaakko,

I'm trying to produce some ACE libraries from ENDF-B/VII files using NJOY for use with Serpent, so as a cross check I am first trying to reproduce the existing ACE file (94239ENDF7.ace) for Pu-239 (@300K only for now). I looked at the NJOY parameters table in the sss_acelib117u.pdf document to get a handle on how to set up the NJOY input file, and while the input file is giving me a similar output ACE file to 94239ENDF7.ace, some differences remain. Here is the NJOY input I am using, please could you advise on what else I need to change so that my output is identical?

Regards,
Andy

Code: Select all

 
moder
 20 -21/ Load ENDF into binary in mem
 reconr 
 -21 -22/ BROADR needs RECONR produced PENDF file
 'pendf file'/ 
 9437 1/ 
 0.01 0. 7 0.01 5E-7/ 1% linearization
 '<pendf>-file'/ 
 0/
 broadr
 -21 -22 -23/
 9437 1/
 .01 1 0.01 5E-7/
 300.0/
 0/
 heatr
 -21 -23 -24/
 9437 0 0 0 1/
 gaspr
 -21 -24 -25/
 purr
 -21 -25 -26/
 9437 1 1 20 64/
 300.0
 1.e10
 0/
 acer
 -21 -26 0 31 32
 1 1 1/ fast data, list all, type 1, suffix .00
'Pu239 data from ENDF-B/VII'/
 9437 300.0/ MAT and temp
 1 1/
 /
 stop
EOF

User avatar
Jaakko Leppänen
Site Admin
Posts: 2447
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Serpent-compatible ACE libraries with NJOY

Post by Jaakko Leppänen » Mon Oct 21, 2013 9:20 pm

The cross section libraries in Serpent 1 distribution were generated using NJOY version 99.259. Which version did you use?

By differences, do you mean differences in the Serpent result or differences between the two ACE files?
- Jaakko

andy
Posts: 15
Joined: Fri May 31, 2013 4:56 pm
Security question 1: No
Security question 2: 92

Re: Serpent-compatible ACE libraries with NJOY

Post by andy » Tue Oct 22, 2013 12:46 am

Hi Jaakko, I am using njoy 99.364, I refer to differences in the ACE files themselves. The most obvious differences are in the misc. information and NXS/JXS block in the first 12 lines, e.g.

94239ENDF7.ace

Code: Select all

94239.03c  236.998600  2.5852E-08   09/05/09
ENDFB-VII Data for Serpent 1.1.7 (94239 at 300.00K)                    mat9437
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
   381038    94239    27653       47       45       14        0        6
        0        0        0        0        0        0        0        0
        1   138266   149177   149224   149271   149318   149365   215669
   215715   275010   275055   350592   378245   378259   378273   378778
   378792   378792   378806   381035   149552   381038   332469   340945
   340956   340998   341004        0        0        0        0        0
NJOY output

Code: Select all

 94239.00c  236.998600  2.5852E-08   10/18/13
Pu239 data from ENDF-B/VII                                               mat9437
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
   224750    94239     8122       47       45       14        0        6
        0        0        0        0        0        0        0        0
        1    40611    51522    51569    51616    51663    51710    78912
    78958   138253   138298   213835   221957   221971   221985   222490
   222504   222504   222518   224747    51897   224750   195712   204188
   204199   204241   204247        0        0        0        0        0
so the Serpent ACE file seems to contain a lot more data. Any thoughts?

Thanks,
Andy

User avatar
Jaakko Leppänen
Site Admin
Posts: 2447
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Serpent-compatible ACE libraries with NJOY

Post by Jaakko Leppänen » Tue Oct 22, 2013 1:31 am

The third value in NXS is the number of energy points, which is clearly different (27653 vs. 8122). The number of reactions is the same, so the difference is probably in the thinning parameters.
- Jaakko

User avatar
Tuomas Viitanen
Posts: 65
Joined: Thu Mar 25, 2010 1:50 pm
Security question 2: 0
Location: VTT Technical Research Centre of Finland, Espoo

Re: Serpent-compatible ACE libraries with NJOY

Post by Tuomas Viitanen » Tue Oct 22, 2013 11:28 am

Hi,

I just noticed that there is a small but in this case significant error in the NJOY parameter listing of sss_acelib117u.pdf. In this file it is said that the reconstruction tolerance parameters used in RECONR and BROADR modules are
  • err = 0.01
    errint = 5E-7
    errmax = 0.01
As err = errmax, this would effectively disable the integral thinning property. This is, however, not the case and in reality the integral thinning has been used when generating the Serpent 1.1.7 ACE libraries. The correct parameters are:
  • err = 0.01
    errint = 5E-7
    errmax = 0.1
In addition to this obvious difference, I think you have another error in the RECONR input: at least the fourth card of RECONR does not correspond to the format specified in the NJOY Userinp.259 file. This might explain the large difference in the number of energy grid points. Try changing the inputs of RECONR and BROADR modules to

Code: Select all

reconr
-21 -22/ BROADR needs RECONR produced PENDF file
'pendf file'/
9437 1/
0.01 0. 0.1 5E-7/ 
'<pendf>-file'/
0/
broadr
-21 -22 -23/
9437 1/
.01 1 0.1 5E-7/
300.0/
0/
Hope this helps!

Best regards,
Tuomas

andy
Posts: 15
Joined: Fri May 31, 2013 4:56 pm
Security question 1: No
Security question 2: 92

Re: Serpent-compatible ACE libraries with NJOY

Post by andy » Tue Oct 22, 2013 4:19 pm

Hi Tuomas,

Yeah I noticed the error in my input last night after Jaakko's suggestion, so my script is now giving me a lot more energy points. But I didn't know about the errmax parameter, so I've modified that according to your suggestion but I'm still getting a lot more energy points:

Code: Select all

 94239.00c  236.998600  2.5852E-08   10/22/13
Pu239 data from ENDF-B/VII                                               mat9437
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
   629206    94239    58679       47       45       14        0        6
        0        0        0        0        0        0        0        0
        1   293396   304307   304354   304401   304448   304495   432811
   432857   492152   492197   567734   626413   626427   626441   626946
   626960   626960   626974   629203   304682   629206   549611   558087
   558098   558140   558146        0        0        0        0        0
Any thoughts?

Regards,
Andy

User avatar
Jaakko Leppänen
Site Admin
Posts: 2447
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Serpent-compatible ACE libraries with NJOY

Post by Jaakko Leppänen » Wed Oct 23, 2013 10:24 am

Are you using the same ENDF file? I believe our data is based on ENDF/B-VII.0. At least revision 1 has been published since then.
- Jaakko

andy
Posts: 15
Joined: Fri May 31, 2013 4:56 pm
Security question 1: No
Security question 2: 92

Re: Serpent-compatible ACE libraries with NJOY

Post by andy » Wed Oct 23, 2013 5:08 pm

Hi Jaakko,

I used the Pu-239 neutron library from

http://t2.lanl.gov/nis/data/endf/endfvii-n.html

which, as far as I can tell, is the same version.

Regards,
Andy

User avatar
Tuomas Viitanen
Posts: 65
Joined: Thu Mar 25, 2010 1:50 pm
Security question 2: 0
Location: VTT Technical Research Centre of Finland, Espoo

Re: Serpent-compatible ACE libraries with NJOY

Post by Tuomas Viitanen » Thu Oct 24, 2013 4:06 pm

Hi Andy,
First of all, sorry about the errmax error in the documentation, it was my bad. Now the documentation should be fixed.

I just noticed another error in the input. Unlike in many Monte Carlo codes, in NJOY the basic unit of energy is eV. Thus, the thnmax parameter should be 1E6 instead of 1. I hope this finally solves the problem:

Code: Select all

reconr
-21 -22/ BROADR needs RECONR produced PENDF file
'pendf file'/
9437 1/
0.01 0. 0.1 5E-7/
'<pendf>-file'/
0/
broadr
-21 -22 -23/
9437 1/
.01 1E6 0.1 5E-7/
300.0/
0/
Regards,

Tuomas

andy
Posts: 15
Joined: Fri May 31, 2013 4:56 pm
Security question 1: No
Security question 2: 92

Re: Serpent-compatible ACE libraries with NJOY

Post by andy » Thu Oct 24, 2013 10:40 pm

Hi Tuomas,

No apologies necessary, I really appreciate all the effort you guys have put into Serpent and I understand the difficulties inherent in software development. I tried out your suggestion and now my output is a lot closer (see below). Can you see if there is anything glaringly wrong remaining in my input file (also below)? The small difference could be due to njoy versions or compiler/platform specifics, so at this point I will likely try out the ACE files with Serpent and see how my results look.

Thanks a lot for all your help :-)
Andy

Code: Select all

 94239.00c  236.998600  2.5852E-08   10/24/13
Pu239 data from ENDF-B/VII                                               mat9437
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
      0         0.      0         0.      0         0.      0         0.
   380726    94239    27619       47       45       14        0        6
        0        0        0        0        0        0        0        0
        1   138096   149007   149054   149101   149148   149195   215391
   215437   274732   274777   350314   377933   377947   377961   378466
   378480   378480   378494   380723   149382   380726   332191   340667
   340678   340720   340726        0        0        0        0        0

Code: Select all

 moder
 20 -21/ Load ENDF into binary in mem
 reconr 
 -21 -22/ BROADR needs RECONR produced PENDF file
 'pendf file'/ 
 9437 1/ 
 0.01 0.0 0.1 5E-7/ 1% linearization
 '<pendf>-file'/ 
 0/
 broadr
 -21 -22 -23/
 9437 1/
 .01 1E+6 0.1 5E-7/
 300.0/
 0/
 heatr
 -21 -23 -24/
 9437 0 0 0 1/
 gaspr
 -21 -24 -25/
 purr
 -21 -25 -26/
 9437 1 1 20 64/ Last two are 20 64 for Serpent
 300.0
 1.e10
 0/
 acer
 -21 -26 0 31 32
 1 1 1/ fast data, list all, type 1, suffix .00
'Pu239 data from ENDF-B/VII'/
 9437 300.0/ MAT and temp
 1 1/
 /
stop
EOF

Post Reply