Forum: Processes Involving Optical Photons
Date: 17 Apr, 2009
From: Maxim Gonchar <Maxim Gonchar>

 Hi, Sometimes Geant4 fails to calculate a surface normal. In this case it sets the flag "valid" to false and G4OpBounadryProcess prints an error message, but continues to work. Then it uses this normal in DoReflection(), G4LambertianRand(). The problem is in G4LambertianRand(G4ThreeVector* normal) which falls into the infinite loop when argument is (0.,0.,0.). I think that the magnitude of a surface normal should be checked before calling G4LambertianRand. But what happens to the Lambertian reflection in case when the normal is 0? What should be used instead? ```PS. Code to get normal=(0., 0., 0.): G4Tubs* blk = new G4Tubs("blk",2.249740e+03, 2.252740e+03, 4.990000e02, 0.000000e+00, 2.609116e-01); G4ThreeVector localpoint(-1.59250955522304866463e+03, 1.60239571861107833683e+03, 1.55565646370305239543e+03); G4ThreeVector dir(-2.97231637897997502673e-01, -9.46902667985754731284e-01, -1.22591560859357129321e-01); bool valid; G4ThreeVector norm; double dist=blk->DistanceToOut(localpoint, dir, true, &valid, &norm); printf("dist=%20f, norm(%20f, %20f, %20f), valid=%i\n", dist,norm.x(),norm.y(),norm.z(),valid);``` best regards, Maxim Gonchar

 Inline Depth: Outline Depth: Add message:

1 Re: Inifinite loop in G4OpBounadryProcess::DoReflection()   (Gumplinger Peter - 06 May, 2009)
 to: "Inifinite loop in G4OpBounadryProcess::DoReflection() "
This site runs SLAC HyperNews version 1.11-slac-98, derived from the original HyperNews