Message: Replicas vs Navigation for large voxel numbers Not Logged In (login)
 Next-in-Thread Next-in-Thread
 Next-in-Forum Next-in-Forum

None Replicas vs Navigation for large voxel numbers 

Forum: Geometry
Date: 04 Dec, 2009
From: Ben Morgan <Ben Morgan>


I'm setting up a simulation for a liquid argon TPC which has a very fine grained readout over large volumes, meaning that we need to record the energy deposits in the simulated liquid volume on scales of 1-5mm over volumes of 1-10 cubic metres. I've experimented with several ways to set up the geometry of the liquid volume to allow scoring on this scale, but wanted to get some advice on the 'best' way to do this and how this will affect the accuracy of scoring the energy deposits.

As the liquid volume is completely homogeneous, I've tried three approaches (with all other aspects of the simulation, physics especially, the same):

  • Replica volumes - use G4PVReplica to replicate 1-5mm cubic voxels along x, y and z respectively (i.e. replicate voxel along a row, replicate row across a plane, replicate plane through a cube).
  • Using the DICOM example and regular navigation. Seemed o.k. for 1e6 voxels, but started getting bad_alloc exceptions with 1e9 voxels... I think this is due to the voxel->material indexing array
  • 'Virtual' navigation - liquid volume is a homogeneous logical volume, attached sensitive detector takes G4Steps and analytically determines the voxels passed through by the geometric step (voxel arrangement is regular cubic lattice).

Phantom parameterisation doesn't appear to scale - I used the DICOM example as a basis, with all voxels having the same material - seeming because of the large indexing array needed with 1e9 voxels (I don't rule out that I set it up incorrectly though!).

Replica volumes work, but are at least 10x slower than the 'virtual' navigation method. Plus, we may need to use cubic voxels inside a cylindrical volume, and 'virtual' navigation would seem ideal here as no extra geometry needs setting up and it's independent of the volume being scored. It would therefore seem to be a good way to go, but since it only uses the geometric step information, will I need to make corrections to the energy deposit in each voxel due to multiple scattering as is done in G4PSDoseDeposit_RegNav? Is there anything else to watch out for to ensure accurate scoring when only using the geometric step information?

Hopefully, I'm on the right track, but I'd be grateful for any corrections/advice/recommendations on this! It does seem to be a very similar problem to that addressed by regular navigation so perhaps I'm just missing something!



Inline Depth:
 1 1
 All All
Outline Depth:
 1 1
 2 2
 All All
Add message: (add)

1 Feedback: Re: Replicas vs Navigation for large voxel numbers   (Pedro Arce - 07 Dec, 2009)
2 Feedback: Re: Replicas vs Navigation for large voxel numbers   (Pedro Arce - 07 Dec, 2009)
1 None: Re: Replicas vs Navigation for large voxel numbers   (Ben Morgan - 10 Dec, 2009)
 Add Message Add Message
to: "Replicas vs Navigation for large voxel numbers"

 Subscribe Subscribe

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

[ Geant 4 Home | Geant 4 HyperNews | Search | Request New Forum | Feedback ]