This cloud has a radius of 1m. To exctract an energy spectrum with a good resolution i run the simulation with 50000000 events. So in this simulation there were 11.9 events per cm³. But when increasing the cloud to a radius of 10m i would have to run the simulation with 50008773560.059 events for comparable results. This simulation would take approximately 62 days according to FLUKA, which is quite a long time.
Is there a way to decrease the simulation time to (let’s say 7 days) with biasing but still keeping comparable results?
unfortunately the event-by-event scorings like the EVENTBIN you are using and DETECT are only giving meaningful results if the FLUKA simulation is fully analogue (no biasing is applied to the particles).
This would leave only option to speed up your simulation: Using a source routine to direct the primaries toward the detector, and manually correct the results based on the probability of the sampled direction.
However this has another problem. The ISOTOPE source type overrides the direction settings given in a source routine, and FLUKA will always sample the direction from an isotropic distribution. This means you need to recreate the source spectrum is Cs-137 manually. If you are only interested in the main 662 keV peak, then you only need a simple PHOTON primary. But if you need the minor photon peaks or the related beta decays, then you need program those in the source routine as well.
Thanks for your reply. Since i am new to FLUKA, i don’t know exactly how to implement a source routine, like you mentioned. I am using now 8 computers to speed up the process. But there is another thing. When looking at the EVENTBIN file (ASCII) most of the lines it says:
Number of hit cells: 0
In the Eventbin card I activated the option Print: Non-zero cells, but i keep getting those zero cells. Is there a way to get rid of these so it only shows me:
Number of hit cells: 1
Because all these unnecessary Information cost a lot of space of the hard drive. 10000000 Events equals about 12,6 GB of data and therefore the evaluation of the data is really slow.
for implementing a source routine, you can take a look at the new version called source_newgen.f, available in the latest release FLUKA 4.1-0.
It features a simplified way for assigning the beam parameters, and an extensive documentation is available directly in the source file, or in the FLUKA manual under section 13.2.20.bis}.
The Non-Zero Cells printing option means that the printing of the result is skipped, but the header for that event is still printed.
The solution could be the DETECT scoring, which results the energy spectrum in a region directly. The number of event can be calculated, if the energy range is selected correctly, by summing up the values of each bin. See: Question regarding DETECT scoring
