Lattice calculations, Serpent vs. MCNP

Share your results and discuss how Serpent compares to other neutron transport codes or experimental data
User avatar
Jaakko Leppänen
Site Admin
Posts: 2377
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Lattice calculations, Serpent vs. MCNP

Post by Jaakko Leppänen » Wed Mar 24, 2010 9:56 pm

I have used a set of infinite-lattice calculations to validate Serpent against MCNP throughout the project, but so far the only published results are probably found in my doctoral thesis from 2007. I have now been able to set up a more or less automated system for running the tests and generating readable reports without hours of work with Matlab and Latex. The reports are found at:

http://virtual.vtt.fi/virtual/montecarl ... culations/

At the moment there are four comparisons (Serpent 1.1.10 vs. MCNP5 using ENDF/B-VII and JEFF-3.1 based cross section libraries with and without unresolved resonance probability table data). Each comparison consists of 14 test cases:

1. PWR pin-cell mode, with fuel composition corresponding to 1 MWd/kgU burnup
2. PWR pin-cell mode, with fuel composition corresponding to 20 MWd/kgU burnup
3. PWR pin-cell mode, with fuel composition corresponding to 40 MWd/kgU burnup
4. VVER-440 assembly with fresh fuel and 650 ppm boron in coolant
5. Mixed UOX / MOX fuel lattice
6. BWR assembly with fresh fuel and burnable absorber, 25% coolant void fraction
7. BWR assembly with fresh fuel and burnable absorber, 50% coolant void fraction
8. BWR assembly with fresh fuel and burnable absorber, 75% coolant void fraction
9. Conceptual SCWR (Super Critical Water Cooled Reactor) assembly with fresh fuel
10. CANDU fuel cluster with fresh fuel
11. Sodium-cooled fast reactor assembly with fresh fuel
12. Prismatic HTGR fuel block with fresh fuel
13. Prismatic HTGR fuel block with fresh fuel and burnable absorber
14. PBMR fuel pebble with fresh fuel

The parameters calculated for each case are:

1. Infinite multiplication factor
2. Prompt neutron lifetime
3. Overall running time
4. Various homogenized multi-group cross sections

The multi-group cross sections are calculated using a four energy-group structure (thermal, resonance, slowing-down and fast fission). In addition, flux spectra integrated over the entire geometry are compared.

Since Serpent is using the same cross section library format as MCNP, the differences between the codes should be within the range of statistical accuracy when everything works as expected. This set of test cases has indeed proved a convenient way of spotting problems in the methodology, as even small discrepancies can be seen in the results. On the other hand, the comparison is focused on a relatively small number of parameters.

If you have other fuel or geometry types you would like me to include in the comparison, let me know.
- Jaakko

User avatar
Jaakko Leppänen
Site Admin
Posts: 2377
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Lattice calculations, Serpent vs. MCNP

Post by Jaakko Leppänen » Sun Apr 11, 2010 12:05 am

Validation reports for ENDF/B-VI.8, JEF-2.2 and JEFF-3.1.1 based cross section libraries are now found at the same directory. The comparison now covers all libraries distributed with the Serpent installation package.

The CANDU case for JEF-2.2 shows an anomaly in the spectrum just below 1 MeV. This is most likely due to an error in one of the less commonly used ENDF reaction laws. I'll try to find the problem and fix it for the next update.
- Jaakko

User avatar
Jaakko Leppänen
Site Admin
Posts: 2377
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Lattice calculations, Serpent vs. MCNP

Post by Jaakko Leppänen » Thu May 20, 2010 11:29 am

Validation reports for version 1.1.11 are found at the same location. The flux anomaly in the CANDU case with JEF-2.2 data is now fixed, but two new problems emerged. Fission nubar in energy group 3 in the SFR case with JEF-2.2 (ures) is too high. Fast (group 1) capture cross section for the prismatic HTGR calculation using ENDF/B-VI.8 (no ures) is also slightly over-predicted.
- Jaakko

User avatar
Jaakko Leppänen
Site Admin
Posts: 2377
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Lattice calculations, Serpent vs. MCNP

Post by Jaakko Leppänen » Thu Jul 01, 2010 3:48 pm

Validation reports for update 1.1.12 are found at the website. There seems to be no differences compared to the previous version.
- Jaakko

User avatar
Jaakko Leppänen
Site Admin
Posts: 2377
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Lattice calculations, Serpent vs. MCNP

Post by Jaakko Leppänen » Wed Jul 28, 2010 1:20 pm

I added the MCNP and Serpent input files in the directory.
- Jaakko

User avatar
Jaakko Leppänen
Site Admin
Posts: 2377
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Lattice calculations, Serpent vs. MCNP

Post by Jaakko Leppänen » Wed Aug 25, 2010 2:00 pm

Validation reports for version 1.1.13 were added to the website.
- Jaakko

User avatar
Jaakko Leppänen
Site Admin
Posts: 2377
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Lattice calculations, Serpent vs. MCNP

Post by Jaakko Leppänen » Tue Nov 16, 2010 11:56 am

Validation reports for version 1.1.14 were added to the website.
- Jaakko

User avatar
Jaakko Leppänen
Site Admin
Posts: 2377
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Lattice calculations, Serpent vs. MCNP

Post by Jaakko Leppänen » Fri Jun 17, 2011 2:51 pm

Validation reports for version 1.1.15 were added to the website.
- Jaakko

User avatar
Jaakko Leppänen
Site Admin
Posts: 2377
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Lattice calculations, Serpent vs. MCNP

Post by Jaakko Leppänen » Mon Sep 05, 2011 11:03 pm

Validation reports for version 1.1.16 were added to the website.
- Jaakko

User avatar
Jaakko Leppänen
Site Admin
Posts: 2377
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Lattice calculations, Serpent vs. MCNP

Post by Jaakko Leppänen » Mon Mar 05, 2012 2:54 pm

Validation reports for version 1.1.17 were added to the website.
- Jaakko

Post Reply