Page 1 of 1

Simulating a neutron beam?

Posted: Fri Aug 09, 2013 9:26 pm
by eugene171
How might one simulate a neutron beam? I have MCNP code for the beam in question:

mode n
nps 10000000 $ Number of histories to be run
imp:n 1 1 0 $ importance of materials
sdef erg=2 $ monoenergetic neutrons at 16.5 MeV
sur = 6
dir = 1 $ cosine of angle
pos -.75 0 0
rad = d1
vec = 1 0 0
si1 .499
sp1 -21 1
c

But I'm having trouble finding in the manual how to make neutron sources other than those diffuse in the material

Re: Simulating a neutron beam?

Posted: Sat Aug 10, 2013 12:50 pm
by Jaakko Leppänen
What kind of surface do you have in the mcnp input. The problem with Serpent is that it can only handle spherical and cylindrical surfaces (Sec. 9.2.1 in the Manual), and doesn't have the option to apply cosine distributions (other than mono-directinal and isotropic). One option is to generate a source file (Sec. 9.2.4) consisting of discrete points with position, energy and direction given separately for each point.

Re: Simulating a neutron beam?

Posted: Tue Aug 13, 2013 12:05 am
by eugene171
The surface in question is a rectangular cell, bounded by six planes (2 px, 2 py, 2 pz).

Before trying to do anything too complicated with the source, I decided to try using

src 1 sd 0 0 1 se 2

for isotropic neutrons at 2 MeV coming from the Z direction.

I am getting the following error:

Code: Select all

Starting the transport calculation cycle...


***** Mon Aug 12 15:52:44 2013 (seed = 1376340764)

Fatal error in function ExternalSource:

Coordinate sampling routine missing

Simulation aborted.
Is this because of the geometry of the target?

Re: Simulating a neutron beam?

Posted: Tue Aug 13, 2013 9:26 am
by Jaakko Leppänen
You get the error because you are missing the spatial distribution (see Sec. 9.2.1). In fact, you should be able to define the source planes using the sx, sy and sz entries (just set the width of the volume small enough to approximate a planar surface).

To get 6 planar sources, you need 6 separate src cards.

Re: Simulating a neutron beam?

Posted: Thu Aug 22, 2013 12:32 am
by eugene171
I seem to be doing things wrong despite trying to follow the manual.

I keep getting the error:
Fatal error in function ExternalSource:

Coordinate sampling routine missing

Simulation aborted.

Here is my input file thus far; can you give some input on what I'm doing incorrectly?

Code: Select all

%Eugene O'Donnell
% --- Test Thorium file for a 1.5x1x1 cm plate -------------------

set title "Thoriumed"


% --- Plate bounds 

surf 1 cuboid -0.75 0.75 -0.5 0.5 -0.5 0.5
surf 2  sph  0 0 0 50
surf 3  sph  0 0 1.5 0.5  %eventually to be used as source

% --- Cell definitions 

cell  1  0  thorium    -1 2    % inside the plate
cell  2  0  air  1 -2 -3
cell  3  0  air  -3  %is sphere, r of 0.5
cell 99  0  outside  2 

% --- Fuel materials:

mat thorium -11.7  
90232.06c   1  %.06c from jeff311u

mat air -0.001184 %at 25C
8016.06c 0.2 %.06c from jeff311u
7014.06c 0.8 %.06c from jeff311u

% --- Cross section data library file path:

set acelib "/xs/sss_jeff311u.xsdata" %not sure if jeff311u ought to be used

% --- black boundary condition assumed: 

set bc 1

% --- group constant generation:

% universe = 0 (homogenization over all space) assumed, commented out
% symmetry = 0 no symmetry assumed, not input
% number of energy groups commented out. I don't think it's needed here

set gcu  0 
set sym  0
% set nfg  4  0.625E-6  5.5E-3  0.821

% --- Neutron source


%src 1 sx -0.75 0.75 sy -0.5 0.5 sz -0.5 0.5 sd 0 0 1 se 2
src 2 %most basic case to test functionality


% ------------------------------------------------------------[code]
[/code]

Re: Simulating a neutron beam?

Posted: Mon Aug 26, 2013 9:38 am
by Jaakko Leppänen
It looks like you have some geometry errors (overlapping cells) in the input. I think your cell definitions should be:

Code: Select all

cell  1  0  thorium    -1     % inside the plate
cell  2  0  air  1 -2 3
cell  3  0  air  -3  %is sphere, r of 0.5
cell 99  0  outside  2
You can get rid of the error message by adding "sc 1" in the source definition (sample source points in cell 1). But then, because cell 1 is small compared to the geometry dimensions, you'll get another error about low sampling efficiency. This problem is solved by putting x-, y- an z-boundaries around the source cell, as you have done with your other source card. I got the calculation going with definition:

Code: Select all

src 1 sc 1 sx -0.75 0.75 sy -0.5 0.5 sz -0.5 0.5 sd 0 0 1 se 2
For a source in cell 3, try:

Code: Select all

src 1 sc 3 sx -0.5 0.5 sy -0.5 0.5 sz 1.0 2.0 sd 0 0 1 se 2