Problem 2069

Summary: I get different dose deposit values in central scorer when scoring with 1 bin vs several bins in depth or several bins in 3D.
Product: Geant4 Reporter: JRSR <ramon_sendon>
Component: processes/scoringAssignee: aso
Status: RESOLVED INVALID    
Severity: critical CC: ramon_sendon
Priority: P4    
Version: 10.3   
Hardware: PC   
OS: Windows   

Description JRSR 2018-07-26 10:52:00 CEST
Source: e- / 6*MeV
Main.cc code:   
  G4ScoringManager *scoringManager = G4ScoringManager::GetScoringManager();
  scoringManager->SetVerboseLevel(1);
  G4ScoringManager::GetScoringManager();
score.mac:
  /score/create/boxMesh vol
  /score/mesh/boxSize xs ys zs mm
  /score/mesh/nBin xb yb zb
  ...
  (scorer is placed inside a wáter volumen, and the source comes from outside)

Different values of dose/energy deposit got in same location if scorer of 1cell, 1D, or 3D:
  1cell: 
        /score/mesh/boxSize 1.0 1.0 0.5 mm
        /score/mesh/nBin 1 1 1
  1D:
        /score/mesh/boxSize 1.0 1.0 50.0 mm
        /score/mesh/nBin 1 1 100
  3D:
        /score/mesh/boxSize 100.0 100.0 50.0 mm
        /score/mesh/nBin 100 100 100

Dose distribution in the 1D scorer is different (and unrelistic) vs the same distribution in the 3D scorer.
This can be tested with Example/b1.
Comment 1 asai 2018-07-26 19:46:23 CEST
Thank you for reporting.
From which direction the primary particle comes in?
Could you also provide your macro commands that define the dose / energy scorers?
Makoto
Comment 2 JRSR 2018-07-30 14:35:32 CEST

The direction is z, going from -z to +z, and coming from outside the main phantom:

 fParticleGun->SetParticleMomentumDirection(G4ThreeVector(0.,0.,1.));
 fParticleGun->SetParticlePosition(G4ThreeVector(0.*cm, 0.*cm, -100.1*cm));

The scorers are inside the main phantom; and the main phantom has its surface at position (0.,0.,0.).

The scorers macro for 3D box are:

 /score/create/boxMesh vol
 /score/mesh/boxSize 50.0 50.0 14.0 mm
 /score/mesh/nBin 50 50 28
 /score/mesh/translate/xyz 0. 0. 14. mm
 /score/quantity/doseDeposit volD
 /score/close
 /score/dumpQuantityToFile vol volD scvol.txt

...for PDD:

 /score/create/boxMesh vol
 /score/mesh/boxSize 1.0 1.0 14.0 mm
 /score/mesh/nBin 1 1 28
 /score/mesh/translate/xyz 0. 0. 14. mm
 /score/quantity/doseDeposit volD
 /score/close
 /score/dumpQuantityToFile vol volD scvol.txt

...and for 1cell:

 /score/create/boxMesh vol
 /score/mesh/boxSize 1.0 1.0 0.5 mm
 /score/mesh/nBin 1 1 1
 /score/mesh/translate/xyz 0. 0. 14. mm
 /score/quantity/doseDeposit volD
 /score/close
 /score/dumpQuantityToFile vol volD scvol.txt



(In reply to asai from comment #1)
> Thank you for reporting.
> From which direction the primary particle comes in?
> Could you also provide your macro commands that define the dose / energy
> scorers?
> Makoto
Comment 3 aso 2018-08-30 02:49:14 CEST
could you give me the number of beamOn?
Also I would like to know how much direrebces you got.
I will try to check with the same condition.
Comment 4 aso 2018-08-30 06:15:19 CEST
Sorry, misspelling.
Also I would like to know how much differences you got in your results.
Thanks.
Comment 5 aso 2019-02-01 01:28:04 CET
Are you still in the problem?

I suppose that the differences of dose values in 1-cell, 1D and 3D were caused by the statistics. You need to increase the number of primary particles.
Comment 6 aso 2019-02-08 09:03:44 CET
I tested the simulation with the command-line scorer and B1 example as reported in the bug report. Since the simulation with 1 mm length cubic voxel requires huge number of histories for getting the results with enough statistics, I used 8mm x 8mm x 2mm voxel size in 1 cell, 1D and 3D segment conditions. The primary particle condition was same as original B1 code except for e-. All dose values in these simulations were consistent. 

Since I could not reproduce the "different dose values" in the simulation and the bug-reporter did not give us the detail simulation parameters, I would like to conclude that the  "different dose values" was not caused by Bugs in Geant4.