Weight management using spectra as a source

Getting Started
1

Dear Monte Carlo users,

I’m facing a problem using spectra as a source. I can’t find a good way to manage the weight using a spectra as a source during my simulations.

Please could you help me find out what i’m doing wrong :slight_smile:

  • I’m using a differential spectra in Mev-1.cm-2.
  • My source is a world sphere using cos biasing. Here the macro :
/gps/particle proton

/gps/pos/type Surface
/gps/pos/shape Sphere 
/gps/pos/radius 110 mm
/gps/pos/centre 0.000e+00  0.000e+00   0.000e+00 cm

/gps/ene/type User
/gps/hist/type energy
/gps/ene/min 0.001 MeV
/gps/ene/max 400.0 MeV
/gps/hist/point 0.001 0.0
/gps/hist/point 0.1	1.73E+14
/gps/hist/point 0.2	8.54E+13
/gps/hist/point 0.3	4.21E+13
/gps/hist/point 0.4	2.33E+13
.
.
.
/gps/hist/point 320.0	2.50E+08
/gps/hist/point 340.0	1.56E+08
/gps/hist/point 360.0	9.96E+07
/gps/hist/point 380.0	6.53E+07
/gps/hist/point 400.0	4.37E+07

/gps/ang/type cos
  • I’m collecting the energy inside a silicon ball of 5mm radius at the center of the world sphere → (I’m assuming here that I have to multiply the energy deposition by the weight of the particle.)
    This detector is inside a sphere of aluminium with parameters (Rmin = 98mm, Rmax = 100mm)

I’m lacking of theory in Monte Carlo simulations. I’m certainly missing a normalisation factor due to the geometry, if that case, please tell me what i’m missing.

Alexandre

2

Hi Alexandre

Some time ago, I flew a spacecraft through a radiation field. Like you, I made a gps sphere:

/gps/pos/type Surface
/gps/pos/shape Sphere
/gps/pos/radius 2 m
/gps/ang/type cos

then a differential spectrum

/gps/ene/type Arb
/gps/hist/type arb
/gps/hist/point 10      1.21E+10
/gps/hist/point 20      3.60E+09
/gps/hist/point 30      1.75E+09
/gps/hist/point 50      3.27E+08
/gps/hist/point 100     1.57E+07
/gps/hist/point 200     7.59E+05
...
/gps/hist/inter Log

As far as I remember, that’s it. Well, you need particle type.

If you generate N particles, this is equivalent to a homogenous, isotropic fluence of N/(pi*R^2), where R is the radius of the cavity. The best explanation I know is something I wrote for a TOPAS example file, attached.

EnvironmentSource.txt (6.4 KB)

Not sure what you mean by the weight of a particle. I just accumulated the energy (dose) in the component of interest.

Hope this helps.

John

3

Hi allison, thank for your help !

If I just accumulate the dose, I obtain a dose value far fewer than expected because only few particle goes through the detector. I feel I’m missing something which can statistically link the weighted spectrum I give as input and the energy deposition

Alexandre

4

Hi Alexandre

It can be a very inefficient process. Only a few of the generated particles hit the component of interest. But it is reliable, i.e., no biases. I just hired hundreds of hours of CPU time.

There are other options. There is something called “Reverse Monte Carlo”, whereby you start close to the component, and estimate what a particle might have originated from. And I think this only works for electromagnetic processes. I’m not sure what the status is. All sounds a bit dodgy - perhaps others can enlighten us.

Good luck
John