Inconclusive results with USIMBS

Dear Fluka Developers,

I recently started a Fluka - MCNP comparison / result check for my
project. And during this I noticed that the results from Fluka are
inconclusive if I use USIMBS. So I started to run some test to check
it.

It these tests I scored the Flux (Total Response) in BurkK region with USRTRACK.

In the first three test (v1, v2, v3) I didn't use the USIMBS, only
three different region importance biasing setup with BIASING cards.
This three run gave the same result 3,52E-8 1/cm2/primary, and the
differences between them was in the errors range.
v1: 3.5119321E-08 +/- 1.049784%
v2: 3.5208146E-08 +/- 0.2559452%
v3: 3.5244092E-08 +/- 0.5248427%

For the test from v4 to v9 I used different USIMBS codes. For v4, v5,
v6 I tried to make the similar biasing like in the first tree test,
but the results got inconclusive.
The results were:
v4: 2.3961730E-08 +/- 0.5406314%
v5: 3.9323904E-08 +/- 0.5327918%
v6: 6.5616213E-08 +/- 1.557709%

For v7, v8, v9 I used an empty USIMBS.f with a fix FIMP value.
v7: FIMP = 1
v8: FIMP = 1.05
v9: FIMP = 1.2

For v7 I got back the result with no USIMBS, but the other two was off again:
v7: 3.5149231E-08 +/- 0.7727996%
v8: 3.9468961E-08 +/- 0.8326670%
v9: 5.2824475E-08 +/- 0.5843777%

For me it looks that the particle weight calculation can be wrong when
USIMBS is used. Please check what could be the problem.

2011.2.8 version of Fluka was used for the test.

The used input and usimbs.f files and the quoted result files can be
downloaded from http://users.hszk.bme.hu/~hd556/usimbs.tar.gz

Best regards,
David HORVATH
Received on Tue Dec 27 2011 - 15:24:41 CET