The (non-dimensional) relative error (which is what tells if your simulation has converged) is invariant to scaling.
- Mario
On Jun 3, 2015, at 11:21 PM, Santana, Mario <msantana_at_slac.stanford.edu<mailto:msantana_at_slac.stanford.edu>> wrote:
Luckily for all of us your math is wrong.
The error associated to an estimator, (which in turn is also an estimator) does not diverge with the sample size (hopefully), but instead it converges (it means it almost freezes around one value) as the sample increases. By the way, the speed of convergence (also an estimator) is roughly the inverse square root of the sample size.
Have a look here
http://en.wikipedia.org/wiki/Moment_%28mathematics%29
I believe that a crash course on statistics and general Monte Carlo principles, followed by an introductory FLUKA course will be more efficient to help you address your studies.
Mario
From: Amit Kumar <amit.live.sal_at_gmail.com<mailto:amit.live.sal_at_gmail.com>>
Date: Wednesday, June 3, 2015 11:07 PM
To: HoSeung Song <genesjp_at_skku.edu<mailto:genesjp_at_skku.edu>>
Cc: Mario Santana <msantana_at_slac.stanford.edu<mailto:msantana_at_slac.stanford.edu>>, fluka-discuss <fluka-discuss_at_fluka.org<mailto:fluka-discuss_at_fluka.org>>
Subject: Re: [fluka-discuss]: Dose calculation (Unit: Gy/min)
Dear HoSeung, Santana
Since there is always some error in the calculated quantities (dose). Therefore, if dose has been calculated by simulating 10e5 number of primaries(electron in this case) and then if we multiply it with 10e15 to get dose corresponding to 10e15 primaries then error will be simply multiplied. But if dose has been calculated by simulating 10e15 primaries then error will be reduced.
That's why I am thinking 10e15 number of primaries should be used in the simulation. I have not seen how much CPS would be appropriate for this simulation but I am expecting it to be difficult.
On Thu, Jun 4, 2015 at 10:28 AM, HoSeung Song <genesjp_at_skku.edu<mailto:genesjp_at_skku.edu>> wrote:
Dear Santana,
Thank you for your reply. As I told you, my interest is to simulate dose in Gy/min.
The target is now fixed yet. It should be a certain object such as bag, box, and etc. But for now lets say the absorbed dose in air. How can I calculate absorbed in Gy/min.
If I simulate a 10E15-16 primary particle number, it seems that simulation time will be so large, is there some way to simulate such kind of large primary numbers?
What if I reduce a number of primaries and do the same simulation. can we believe those results?
Thank you and sorry for your incovenience.
Regards,
Ho-seung Song
On Wed, Jun 3, 2015 at 10:16 PM, Santana, Mario <msantana_at_slac.stanford.edu<mailto:msantana_at_slac.stanford.edu>> wrote:
Hi HoSeung,
Results are normalized per primary (in your case per incident electron).
Thus, to normalize your results per minute you need to multiply those by 7.5E12 x 250 * 60:
Results 7.5E12 e 250 pulse 60 s
----------- x --------------- x --------------- x -----------
e pulse s min
Please note that Gy is the unit for absorbed dose density [J/kg], so you should score it in your target of interest, not in air (unless you want to know the absorbed dose in air).
From: HoSeung Song < genesjp_at_skku.edu<mailto:genesjp_at_skku.edu>>
Date: Wednesday, June 3, 2015 7:26 AM
To: fluka-discuss < fluka-discuss_at_fluka.org<mailto:fluka-discuss_at_fluka.org>>
Subject: [fluka-discuss]: Dose calculation (Unit: Gy/min)
Dear Fluka users,
I am conducting a simulation with 15 MeV electron beam to tunsgen target to generate photon and neutron. After the target, I put conical shape collimator to focus the photon. I want to calculate a photon dose (Gy/min) at the certain distance from the target (example: 1m, 2m, 3m and etc.).
1. For a input card, I used USRBIN (dose-eq), AUXSCORE (photon) to score the photon equivalent dose, and the unit is pSv/primary. Is this true?
The specification of the accelerator is as follows;
- number of electrons per pulse: 7.5*10^12
- pulse frequency: 250Hz
- pulse duration: 6us
With these parameters, I want to calculate a photon-dose (Gy/min) with FLUKA. I am a beginner of this program and this field. Can you tell me how to calculate Gy/min at certain point (or surface) with a little bit detail?
Thank you in advance.
Regards,
Ho-seung SONG
--
Amit Kumar
Scientist
Bhaha Atomic Research Centre
Mumbai
Phone No. : 022-25592841
Mobile Number: 9969799819
Alt. Email Id. amitkumar_at_barc.gov.in<mailto:amitkumar_at_barc.gov.in>
--
Amit Kumar
Scientist
Bhaha Atomic Research Centre
Mumbai
Phone No. : 022-25592841
Mobile Number: 9969799819
Alt. Email Id. amitkumar_at_barc.gov.in<mailto:amitkumar_at_barc.gov.in>
__________________________________________________________________________
You can manage unsubscription from this mailing list at https://www.fluka.org/fluka.php?id=acc_info
Received on Thu Jun 04 2015 - 10:50:20 CEST