Re: [fluka-discuss]: Regarding importance biasing

From: George Kharashvili <georgek_at_jlab.org>
Date: Fri, 22 Jun 2018 05:02:09 -0400 (EDT)

Dear Sujoy,

As you correctly noted the default importance of a region is 1. To take advantage of the full allowed range from 0.0001 to 100000 you can set the importance of all regions to 0.0001 with the first BIASING card (from BLKBODY to _at_LASTREG) and then start increasing the importance as you go through the shielding. Please note that this is one of the rare cases when the order of FLUKA cards is relevant: the first one will set all regions to 0.0001 and the subsequent BIASING cards will overwrite importances of the referred regions.

Cheers,
George

----- Original Message -----
From: "Sujoy Chatterjee" <sujoy_at_vecc.gov.in>
To: "George Kharashvili" <georgek_at_jlab.org>
Cc: "FLUKA Discussion List" <fluka-discuss_at_fluka.org>
Sent: Thursday, June 21, 2018 10:36:09 AM
Subject: Re: [fluka-discuss]: Regarding importance biasing

Dear George,
I thank you once again for taking the trouble of replying my queries. You are absolutely right that the absolute value of importance of a particle in a region does not mean anything, but the ratio of the particle importance value of the region and the region immediately after that. As I understand the default importance of a particle in every region is set to 1. So, when we start applying importance biasing from the first region and if the value is less than 1, then RR is played. If it is more than 1, then particles are split depending upon the ratio. To see this, I have made a small input file, attached herewith and find the neutron fluence at the first boundary and at the last boundary. The results are as follows:
1) Without any biasing Neutron fluence at first boundary: 4.113E-06 +/- 5.462% and at last boundary 2.034E-06 +/- 6.465%
2) With Imp=0.1 in Shield region, First boundary: 3.727E-06 +/- 5.6822% and at last boundary: 1.5366E-06 +/- 9.769%
3) With Imp=2 in Shield region, First boundary: 3.992E-06 +/- 4.196 % and at last boundary: 1.9014E-06 +/- 3.894%

So, we see that with same number of histories, applying Imp more than 1 decreases the error, where as applying Imp less than 1 increases the error. So, I guess, I have to start putting the Imp values 2,4 8,....for every region for my simulation job of deep penetration.

__________________________________________________________________________
You can manage unsubscription from this mailing list at https://www.fluka.org/fluka.php?id=acc_info
Received on Fri Jun 22 2018 - 12:33:14 CEST

This archive was generated by hypermail 2.3.0 : Fri Jun 22 2018 - 12:33:19 CEST