Learning Forum

[ansi]

Good day everyone,nI am simulating a combustion chamber using a reaction mechanism (Finite-Rate, Stiff chemistry solver).nI did get an unconverged solution on my local machine (6 cores) and to accelarte my calculations I want to run it on a cluster using a Journal file. This cluster has 64 cores. As I don't have an ISAT table for that many cores, I need to start without the ISAT table.nFluent stops before calculating the first iteration and hands out following error message:n********** ISAT_ABORT **************n routine = isat_leaf_init n location = 1 n message = svd failed, INFO= n 38 n****** END ISAT_ABORT **************n ell_radii: info, n, lwork, work(1) = 39 40 16800n 1400.00000000000 n sv = n NaN NaN NaN NaN NaNn NaN NaN NaN NaN NaNn NaN NaN NaN NaN NaNn NaN NaN NaN NaN NaNn NaN NaN NaN NaN NaNn NaN NaN NaN NaN NaNn NaN NaN NaN NaN NaNn NaN NaN NaN NaN NaNnI've been using another reaction mechanism and never got this error message. But I could retrieve an ISAT table with 64 cores from these simulations. After loading case and data, I read in that ISAT table and get a different error message:n isat_table_read: mismatch for: nx 40 25n********** ISAT_ABORT **************n routine = isat_table_read n location = 9 n****** END ISAT_ABORT **************nnI deleted the ISAT table before and after reading case and data, I increased max storage of ISAT table but still got this error.nUsing direct Integration is not really an option, it takes too long.nAny idea on how to solve this issue? Is there a way to calculate an ISAT table for 64 cores on my local machine or to convert a 6 cores ISAT table into a 64 cores ISAT table?.

[Rahul Kumar]

Hello, nThis is a failure in chemistry integration. There could be many reasons. Since you have already enabled the stiff chemistry solver : nThe error probably could be because of poor initialization. Can you please solve for the flow first and then turn on the chemistry. nThe mechanism is probably too stiff. Please identify stiff reactions in your mechanism, reduce the pre-exponential factors initially and as the simulation progresses, you could gradually increase to the correct value. nIf it is a transient simulation, please reduce the time step. nTry reducing the ODE tolerances and check if that helps. n