Huge errors in scoring (max value 99%)

Scoring
1

Versions

Please provide the used software versions.

FLUKA: Version 4-5.1
Flair: Version 3.4

Description

Hi everyone,

I am trying to do a simple simulation where a point proton source with custom spectrum irradiate a thick target of natural Boron. The goal is to estimate the number of pB11 fusion reaction and pB10->alpha reaction. I am trying to also evaluate the flux of alphas emitted by the target. I setup different scorer: one to monitor the beam, and two scorer to check the alphas produced inside the target and the ones that goes out of the target. Looking at the output files I have the feeling that something is wrong because alpha are recorded but then the error associated is extremely high. Does anyone had a similar problem?

Input files

input.inp (1.8 KB)

test.flair (1.5 KB)

TP_spectrum.txt (2.7 KB)

source_protons.f (20.8 KB)

2

Are you running multiple cycles? (If you run one single cycle, no statistical error can be evaluated and you will read by construction 99%).

3

Thanks for replying!

I did not run multiple cycles. Only one, with 1e+8 primaries. So, in order to have a meaningful statistical error, I should run multiple cycles. Since I am running all in my laptop, can I just maintain the overall number of events (1e+8) but split them in 10 cycles of 1e+7. This way I could have the same statistics of the one cycle-simulation I did, but with a meaningful error evaluation. Is it right?

4

That’s right. See also a related lecture (pages 15-31).

5

Thanks a lot Francesco. Now it is very clear!