open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all,

My simulation was running smoothly until I tried adding turbines (vertical structure=YES & added a fortran file). Now, Telemac crashes after calling the "propin" subroutine. I receive the error message:

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

I have attached my files, as well as a screenshot of a disclaimer that pops up, which states that "out_dragfo_apr25.exe has stopped working". I'm not quite sure what this means - has this ever come up in the past? I am using the Silverfrost and Plato IDE Fortran compiler, I'm not sure if this might be causing the problem.

Note: I couldn't upload my .liq and .slf liquid boundary and geometry files, which were too big. The geometry file was used to generate the .cli boundary conditions. I've attached a screenshot of part of my .liq file - it shows the first 40 seconds, but continues with a sinusoidal function for 5 days.

I know this error message is common, but I have looked through the other threads and still could not fix my problem. Any help would be greatly appreciated! Thanks.

Best regards,
Alex

Hello,

A segmentation fault is a memory error than is probably a bug triggered by your implementation in subroutine dragfo. I had a quick look at it and find no obvious mistake. Only statements like:

NUMTURBS = 1.0D1

are perhaps dangerous because NUMTURBS is an integer, I hope you get 10 with 1.0D1 but writing NUMTURBS = 10 would be safer, it could be that 1.0D1 eventually gives 9 if the compiler takes function INT instead of NINT to convert the double precision. As you said that you just changed the number of turbines this could be a hint.

With best regards,

Jean-Michel Hervouet

Thank you for your help Jean-Michel. I made the change and am still encountering problems, I'll look some more.

Hi Jean-Michel,

Sorry to bother you again. This follow-up is more a Fortran problem than a Telemac-2D problem..but I realized the bug was that Fortran stores negative values differently, and some of my coordinate points have negative x values.

I've been trying to find a way around this problem but have been unsuccessful so far. Do you have any suggestions by any chance? Thanks!

Best,
Alex

Hello,

Hum, what do you mean by "Fortran stores negative values differently" and where is that a problem is your code ?

JMH