# Fwd: [fluka2010] Results check

From: Joachim Vollaire <joachim.vollaire_at_cern.ch>
Date: Mon, 6 Dec 2010 14:19:25 +0000

Dear Beatrice,
Cordially
Joachim

Question 1 :

- "To see if neutrons are actually
produced I used a USBRDX card to get the neutron
spectrum out of the target; I saw from the
output file that the number of stars per proton
is nearly 6E-2 and the 95% is reached in terms
of percentage. Can this confirm that the numbers of primaries is enough?"

To verify if the number of primary transported is sufficient check the
quantity of interest you are scoring. If the statistical error is less
than 10 % the result can be considered as reliable (less than 5 % is
better). For a spectrum you will have bins we higher errors but you can
look at the integrated value to have an idea of the quality of the
result. I also suggest to use rather 5 independent runs (minimum)

Question 2

- "What is a reasonable standard deviations
for the results (readable from the usrbin and
usrbdx files produced after prcessing)?"

Same as above (Question 1), quoting the MCNP manual, results are
doubtful above 20 % error and meaningless above 50 %. 10 % results start
to be reliable 5 % is better of course…. For USRBIN/USRBDX this of
course concerns your region/bins of interest…

Question 3

- "For USRBIN fluence graph I normalized
by 1,87*10**15 p/s (being the proton beam
300microA) to get results in terms of n/(cm2
s); For USRBIN energy deposition graph should I
use the same factor 1,87*10**15 p/s to get
results in terms of GeV/(cm3 s)? If I need
deposition on an unit area (per cm2) should I multiply for which length?
For the same graph if I want to convert GeV/(cm3s) into °C/s
I need to divide by 6,27*10**9 (from MeV to W) to divide by concrete density
(2,33g/cm3) and concrete conductivity (0,949 J/(g °C)) and to multiply
by 1,87*10**15 p/s.
This way I obtain 1,35 °C/s. If I consider the target will work
5000hours per year I get 2E7°C.
Is it possible to have such a thermal incre from ambient temperature due

I think there is a mistake in your conversion, 300 microA should be
1.87e12 and not 1.87e15 ? It should make other results more reasonable…

Question 4

"As for the USRBDX graph how do I read the y axis of the plot, the x
axis being [GeV], if I choose as USRBDX Type: phi1,LogE,LinOmega that is:
fluence scoring, one way scoring, logarithmic in energy, linear in
angle? For my specific case
how should I normalize it?"

You should normalize using the factor corresponding to 300 microA
(1.87e12 proton per sec) to obtain the results in cm-2/s-1…. Note that
this is only true if you input the surface between the two areas in your
one way scoring (otherwise there is a factor equal to the surface)….

"Is it correct to say the spectrum is a typical fission spectrum?"

Hard to tell from the plot, if the maximum is in the 1 MeV range and the
average in the 2 MeV range I would say yes it has the characteristics of
a fast neutron induced fission spectrum…..

"I plotted a 1D graph of neutrons flux density through the thickness
(n/(cm2 s) vs cm): are
error bars acceptable? Can I catch the thermal and the fast neutron
behavior separately within the same graph.
And can I superpose similar 1-D graphs from different shielding material
(i.e. different input files)?"

You can do this with flair when selecting the file from which you take
the 1-D plot… Just browse to the folder where you have the files you are
interested in….

- "In order to see fluences in all the shielding I have further applied
biasing;
order not to complicate geometry I have biased
with the usrbin.f subroutine I provide in attachment; however as you can
see from graphics
I did not better so much my analysis (just to clear, I applied two
different resolutions fo
the scoring: 5cm for biasing and 10cm for not biasing).
Is the biasing correct?"

As a beginner I would suggest splitting your geometry even if it adds
some complication and using region based biasing to really see the gain….

From: beatrice pomaro
Date: Wed, 17 Nov 2010 08:48:08 +0100

Dear Fluka users,
thank you for the previous reply on USRBIN
processing, I cope with the problem; I forgot
to mention I am using the 2010 pre-release 0.8-4.

I would be grateful to have a check of my
analysis with you, since I am a Fluka
beginner. My problem/input file is in
attachment (inside the folder: 1E7 biasing). I
have some questions on my results (I give in
attachment also the USRBIN and USBRDX scores
for 1E7 ran particles with and without biasing applied).
My geometry is quite simple: a target made of
UC2 directly impinged by a proton beam
(300microA, 70MeV), within a concrete target bunker.
I am interested in the neutron fluence in the shielding.
I ran 1E7 particles and 3 cycles. I have some
specific questions I have tried to list here below:

- To see if neutrons are actually
produced I used a USBRDX card to get the neutron
spectrum out of the target; I saw from the
output file that the number of stars per proton
is nearly 6E-2 and the 95% is reached in terms
of percentage. Can this confirm that the numbers of primaries is enough?

- What is a reasonable standard deviations
for the results (readable from the usrbin and
usrbdx files produced after prcessing)?

- For USRBIN fluence graph I normalized
by 1,87*10**15 p/s (being the proton beam
300microA) to get results in terms of n/(cm2
s); For USRBIN energy deposition graph should I
use the same factor 1,87*10**15 p/s to get
results in terms of GeV/(cm3 s)? If I need
deposition on an unit area (per cm2) should I multiply for which length?
For the same graph if I want to convert GeV/(cm3
s) into °C/s I need to divide by 6,27*10**9
(from MeV to W) to divide by concrete density
(2,33g/cm3) and concrete conductivity (0,949
J/(g °C)) and to multiply by 1,87*10**15 p/s.
This way I obtain 1,35 °C/s. If I consider the
target will work 5000hours per year I get 2E7°C.
Is it possible to have such a thermal incre
from ambient temperature due to radiation?

As for the USRBDX graph how do I read the y axis
of the plot, the x axis being [GeV], if I choose
as USRBDX Type: phi1,LogE,LinOmega that is:
fluence scoring, one way scoring, logarithmic in
energy, linear in angle? For my specific case
how should I normalize it? Is it correct to say
the spectrum is a typical fission spectrum? I
plotted a 1D graph of neutrons flux density
through the thickness (n/(cm2 s) vs cm): are
error bars acceptable? Can I catch the thermal
and the fast neutron behavior separately within the same graph.
And can I superpose similar 1-D graphs from
different shielding material (i.e. different input files)?

- In order to see fluences in all the
shielding I have further applied biasing;
order not to complicate geometry I have biased
with the usrbin.f subroutine I provide in
attachment; however as you can see from graphics
I did not better so much my analysis (just to
clear, I applied two different resolutions fo
the scoring: 5cm for biasing and 10cm for not biasing).
Is the biasing correct?

I am sorry if my questions are many or can
result trivial: I tried to put all in one mail.