'Outside' material within the geometry

Report all good and bad behavior here
Post Reply
pmgorman
Posts: 5
Joined: Tue Jun 04, 2013 12:06 am
Security question 1: No
Security question 2: 92

'Outside' material within the geometry

Post by pmgorman » Mon Sep 16, 2013 8:29 pm

Hello,

So, for a reactor that I was working with, in order to bound the shutdown margin, I tried to use ideal "black" control rods. In MCNP, this was done by setting the neutron importance to 0, but in Serpent 2, I figured I could approximate that by setting the material to "outside." However, I was getting some weird results. The k-eff decreased, more than it did with boron carbide inserted, so I thought I would create a simple case and compare between MCNP6 and Serpent 2. Since I didn't see any post saying whether or not this was a good idea to use this, I thought I'd post about it.

Here's my MCNP6 input file without any black conditions:

Code: Select all

Black Pin Test - PMG
1 1 -10.0 -3 imp:n=0 u=1 $ Fuel - Black pin
2 1 -10.0 -1 imp:n=1 u=2 $ Fuel 
3 2 -6.55  1 -2 imp:n=1 u=2 $ Clad
4 0        2 -3 imp:n=1 u=2 $ Gap
5 3 -0.7   3    imp:n=1 u=2 $ Water
6 3 -0.7   -3 imp:n=1 u=3 $ Water
11 3 -0.7   3 imp:n=1 u=3 $ Water
10 3 -0.7  3    imp:n=1 u=1 $ Water
7 0       -8 9 -10 11 lat=1 u=4 imp:n=1 
      fill=-7:7 -7:7 0:0
      3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
8 0      4:-5:6:-7 imp:n=0 $ outside
9 0      -4 5 -6 7 fill=4 imp:n=1 $ Assembly

c Surfaces
1 cz 0.41
2 cz 0.42
3 cz 0.48
4* px  8.3175
5* px -8.3175
6* py  8.3175
7* py -8.3175
8  px  0.63
9  px -0.63
10 py  0.63
11 py -0.63

c Materials
m1 8016.72c 2.0 92235.72c 0.05 92238.72c 0.95
m2 40090.71c  0.5145
         40091.71c  0.1122
         40092.71c  0.1715
         40094.71c  0.1738
         40096.71c  0.028
m3 8016.71c 1.0 1001.71c 2.0
mt3 lwtr.16t
kcode 10000 1 50 350
ksrc 0.0 0.0 0.0
With black conditions, I changed cell 7 to the following:

Code: Select all

7 0       -8 9 -10 11 lat=1 u=4 imp:n=1 
      fill=-7:7 -7:7 0:0
      3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
      3 1 2 2 2 2 2 2 2 2 2 2 2 1 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
      3 1 2 2 2 2 2 2 2 2 2 2 2 1 3
      3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
The final estimated keff's were 1.42563 +/- 0.00033 and 1.04862 +/- 0.00045, respectively.

My Serpent input without the black rods was as follows:

Code: Select all

set title "Black Rod Serpent2 - MCNP6 Comparison"
% Created by: Phillip Gorman

% --- Fuel Pin definitions:

cell blackPin 1 outside -1
cell water 1 water 1
surf 1 cyl 0.0 0.0 0.48

pin 2 % LEU
fuel1  0.41
void   0.42
clad   0.48
water

pin 3 % Water
water


% --- Lattice definition
lat 4   1  0.0 0.0 15 15   1.26
  3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 3 3 3 3 3 3 3 3 3 3 3 3 3 3

% --- Outer channel (assembly width = 16.635):

surf 2  sqc    0.0     0.0     8.3175
cell assembly 0 fill 4 -2
cell outside  0 outside 2

% --- Materials:

mat fuel1 -10.0 
8016.09c 2.0
92235.09c 0.05
92238.09c 0.95

mat clad -6.55
40090.06c  0.5145
40091.06c  0.1122
40092.06c  0.1715
40094.06c  0.1738
40096.06c  0.028


mat water -0.7 moder lwtr 1001
8016.06c 1.0
1001.06c 2.0

therm lwtr lwe7.12t

% Set parameters

set acelib "/usr/local/SERPENT/xsdata/endfb7/sss_endfb7u.xsdata"
set bc 2
set pop 10000 300 50

plot 3 1000 1000
mesh 3 1000 1000
and I changed the lattice to

Code: Select all

lat 4   1  0.0 0.0 15 15   1.26
  3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 
  3 1 2 2 2 2 2 2 2 2 2 2 2 1 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 2 2 2 2 2 2 2 2 2 2 2 2 2 3
  3 1 2 2 2 2 2 2 2 2 2 2 2 1 3
  3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
The IMP_KEFF's from these were 1.41095 +/- 0.00025 and 1.18183 +/- 0.00040, respectively.

The keff's for the black rods don't agree with each other, although Serpent is clearly treating them differently somehow. For my other (not-simple) case, I got good agreement between Serpent 1 and MCNP6 by setting the control material to be 8 kg/cc pure boron-10, but Serpent 2 has a hard limit of 1 kg/cc for any material. When I set it to use 999 g/cc B-10, it ran for 39 active cycles (at a much slower rate, as well) before I got the following error message:

Code: Select all

***** Fri Sep 13 07:43:40 2013 (seed = 1342136821, MPI task = 0, OMP thread = 0)

Fatal error in function MicroCalc:

Loss rate is zero

Simulation aborted.
I have no idea what that error message was referring to, since the loss rate was hopefully a lot more than 0. The k-eff values before it crashed looked reasonable, though.

orca.blu
Posts: 59
Joined: Wed Apr 20, 2011 1:39 pm

Re: 'Outside' material within the geometry

Post by orca.blu » Wed Sep 18, 2013 7:26 pm

Hi Phillip,
the problem comes from the adoption of delta-tracking.
In MCNP all the neutrons that enter in the "black rods" are lost.
In Serpent, due to DT, many neutrons might enter and exit from the rod,
if a virtual collision is not sampled there.
You just need to add "set dt 0" at the end of your file.
In this way, you will force surface-tracking.

I tried rapidly on my laptop your input (stopping the simulation after few tens of cycles):
First lattice:
Keff = 1.41052 +/- 0.00044
Second lattice:
Keff = 1.18156 +/- 0.00093
Second lattice, "set dt 0"
Keff = 1.04631 +/- 0.00070
Which is not too far from you MCNP result.

You may try with more neutrons.
Of course, this slows downs the simulation.
Manuele Aufiero
LPSC/IN2P3/CNRS Grenoble

pmgorman
Posts: 5
Joined: Tue Jun 04, 2013 12:06 am
Security question 1: No
Security question 2: 92

Re: 'Outside' material within the geometry

Post by pmgorman » Wed Sep 18, 2013 9:29 pm

Oh, excellent! I did not know about that option. Thanks for the tip! It definitely seems to be doing the trick.

-Phil

User avatar
Jaakko Leppänen
Site Admin
Posts: 2387
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: 'Outside' material within the geometry

Post by Jaakko Leppänen » Thu Sep 19, 2013 6:51 pm

There might be other problems as well. When the neutron enters an "outside" zone, the calculation goes to boundary conditions. I've never tried to see what happens, but you might get errors or strange results if you have set bc 2 or 3.
- Jaakko

pmgorman
Posts: 5
Joined: Tue Jun 04, 2013 12:06 am
Security question 1: No
Security question 2: 92

Re: 'Outside' material within the geometry

Post by pmgorman » Fri Sep 20, 2013 1:10 am

Some of the results were indeed strange. I had considered that it might go to the boundary conditions, so I also ran a case where the entire cell (not just the pin) was outside; it still lowered the k-eff value significantly (which wouldn't be expected if it were just reflecting off of the surface).

I think there's something more going on than just the streaming, though, because the outside rods lowered the k-eff more than boron carbide. If the neutrons simply streamed through, then the k-eff wouldn't be impacted much. Regardless of what's happening, though, using surface tracking eliminated the discrepancy.

I think the other error message I received might have something to do with group cross section generation (which is what I was doing when I noticed this discrepancy). I got the same error message when I used surface tracking in the cross section generation case, but it was only after a number of cycles had run. Since the cross sections really didn't matter for black rods, I commented out the "set gcu line" which may have fixed the problem. On the other hand, since it didn't have that error for every cycle, it might have been simply luck of the draw.
Phillip Gorman
PhD Student, University of California - Berkeley

Post Reply