Attending: Dan, Chris, Stan, Josh, Wesley, Dan, Tim, Ray, Glenn, Matt
The main simulation package for Braidwood will be Reactor Analysis Tool (RAT). Click here for the available RAT documentation. As a subdirectory of the RAT package we will keep a release of the Geant4-based Monte Carlo GLG4sim. GLG4sim will be used as a library to generate Monte Carlo events. However, RAT will be able to process data events as well. Keeping GLG4sim inside the RAT release, rather than a separate package, will ensure that each tagged version of RAT will have its "own" version of GLG4sim known to work with that RAT code.
All Braidwood-specific code will be kept in RAT. This includes any code that overloads classes in GLG4sim. Any changes to the GLG4sim source code will be made available to the GLG4sim maintainers for possible inclusion in their own release. As GLG4sim improves, we will update the Braidwood-maintained version in the RAT release to match the current GLG4sim release. This will be done only after significant testing, to ensure that users only have stable GLG4sim code and do not have to treat GLG4sim as "development" code.
Prebuilt libraries for GLG4sim will not be distributed, rather the make command in RAT will include a configuration and build of the GLG4sim libraries for users' systems. The BWsim package in the CVS repository will be discontinued.
An interesting email from Glenn following up on the neutron capture discussion:
"As far as I can tell, John LoSecco of Double Chooz got hold of the same table as Tim's postdoc got, I'm not sure how. But by all accounts, Hans Peter Wellisch has not responded to direct e-mail requests for these tables from others. In fact, he never responded to my own request on behalf of Braidwood and Double Chooz, which predated LoSecco's mysterious success. I have no idea what's up with that. Probably Wellisch is just very preoccupied with other matters. Another possibility is that this is some sort of off-the-cuff, unchecked dataset that he knows produces results that are inferior to what he could do if he made his best effort, and he wants to make a better effort before this spreads further, and just hasn't had the time.
However, our Gd modelling in Double Chooz has gone beyond simply using H.P.W.'s special tables with the standard Geant4 classes. Some people at Subatech-Nantes (who are interested in counter-proliferation applications as well as Double Chooz) have made several different versions of their own code to model the Gd n-capture fireball. Their first effort in this regard was a class called "GdNeutronHPCapture". Dario Motta mentioned this at MAND-sim: http://neutrino.phys.ksu.edu/MAND-sim/MAND-sim%20talks/DoubleChooz_motta.pdf, slides 34-41. They have made a couple more models since then, and they claim privately that they all match the experimental data better than what you can get using Geant4's G4NeutrinoHPCapture and the unofficial tables from H.P.W. But I don't think they want to butt heads publically with H.P.W. at this point since their work is still a work in progress, especially if H.P.W. will get mad because they weren't supposed to be spreading around and criticizing the results from his code anyway.
My understanding is that Subatech's code works without any special tables, just uses stuff they coded or looked up themselves. If they would agree to put it in GLG4sim or otherwise publish it somehow, it would be really great. I don't know what their schedule for doing so might be."
Matthew Worcester Last modified: Fri Aug 30 17:27 CDT 2005