Message: help needed with variance reduction method - reposting Not Logged In (login)
 Next-in-Thread Next-in-Thread
 Next-in-Forum Next-in-Forum

Question help needed with variance reduction method - reposting 

Forum: Biasing and Scoring
Date: 22 May, 2013
From: Magda Constantin <Magda Constantin>

Dear G4 Colleagues,

I've posted this message on the Electromagnetic Processes Forum, but I would also like to post the same question in this Forum.

I am simulating dose deposited to a water phantom at 1 meter away from a 6MeV medical X-ray source. The beam is transmitted through ~4.5 tenth-value-layer slab (~156 mm of tungsten). The incident radiation intensity decreases by a factor of ~25e-06 after passing through the tungsten slab. Because of this drastic intensity attenuation, I have to simulate a very large (~in the billions) number of incoming 6 MeV electrons on the source target. When I calculate with Geant4 (v9.4.p02 ) the dose is ~10 times larger than the experimental measurement.

I suspect that this disagreement is due to limited statistics on the scoring voxels. My intention is to use variance reduction techniques to increase the number of electrons on the X-ray target without increasing the simulation time too much. Could you please advise which variance reduction methods are the most suitable in my case? Is there a working example that I can look at? How about validation studies and/or publications?

Are there any other strategies to help me lower the simulation time (i.e. range cuts, other Physics List, etc)? I have been using Opt3 with 10 um range cut for the particles. Thank you in advance for your advice.

Magda

 Add Message Add Message
to: "help needed with variance reduction method - reposting"

 Subscribe Subscribe

This site runs SLAC HyperNews version 1.11-slac-98, derived from the original HyperNews