Hello Jean-Michel,
I think I made a mistake this afternoon. I checked the code and in my case line 1730 is indeed the place where the CVDF3D (APPEL DE CVDF3D POUR W) is called in TELEMAC3D.
In the original telemac3d.f, the line number is 1705 but I use the modified file you provided with me (the crash problem of k-epsilon model is fixed in the modified telemac3d.f).So I didn't solve the problem and it still happens in my model when it couples with SISYPHE.
This error occurs in the sequence of OV --> MVSEG --> MATVCT --> MATVEC --> DIRI01 --> DIFF3D --> CVDF3D --> TELEMAC3D.
But it's strange that in ov.f, the error occurs at line 181 which is
CASE('YZ ')
!
DO I=1,NPOIN
X(I) = Y(I) * Z(I)
ENDDO
Does it mean either matrix Y or Z is invalid? I traced back those two variables and I found
Y is MTRA2
Z is T3_03
And in DIRI01, T3_03 is also FBOR which means DIRICHLET BOUNDARY CONDITIONS. But I checked the MODULE DECLARATIONS_TELEMAC3D and found MTRA2 is just a WORKING MATRICES.
Kind regards,
Qilong