Project

General

Profile

round off problem in EquivalenceModel.cxx

Added by bidaud adrien almost 8 years ago

Hello,

Francisco Alderete made very simple EquivalenceModel for CANDU fueled with reprocessed Uranium (*): the FP will just simply take all the masse of uranium needed in the FissileVector to operate the reactor.
In many cases, we will not suggest a FertileVector, which is dangerous as some warning may say. In fact it is even more dangerous than expected. Indeed, it happens that because of round off, the mass of fertile material needed could be non 0 and in fact negative (about the machine's precision). Then the EquivalenceModel::Build_Fuel_According_Lambda() method is looking for negative masses !
I have proposed a fix that works. Probably, it would be better to put another fabs() to FertilMassNeeded * 1e-6 in the test to check if enough Fertile material is available ? This is the test that stops the fuel fabrication when FertilMassNeeded is negative ?

A bientôt,

Adrien

(*) see his files for test that go wrong.