open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I am trying to re-use subroutines that i have personalized in a projet made last year with t2d_V5P8 .
I have NOW (not last year for some obscure reasons) a problem to impose a tidal signal at my open-ocean boundary. Indeed to be applied the limit condition parameter has to fullfil "LIHBOR=KENT" but in my case KENT=1 which prevents from going into the loop and therefore imposing a tidal signal.

Any clearance relative to this parameter KENT is welcomed (I can't really find where it is defined in the source).

Thank you in advance.

Alexis

Hello Alexis,

KENT is a convention for "imposed value", generally at an entrance. It is defined as a parameter in declarations_telemac (library bief). LIHBOR is taken in the boundary conditions file. 1 corresponds to an incident wave boundary condition which is seldom used, so I am surprised that you used this.

You should also check arguments in subroutine bord.f, e.g. an argument NPTFR2 has been added after argument NPTFR, between version 5.8 and 5.9, this is perhaps the reason of your problem.

With best regards,

Jean-Michel Hervouet

Hello Jean-Michel,

How could KENT be set to 1 in my case as I am not doing any coupling with waves? Do I have to simply set KENT to 5 in my subroutine as my LIHBOR is 2 at the wall, 4 at the river boundaries and 5 at the open ocean boundary.

I have check the role of NPTFR2 in the V5P9 bord subroutine: it is listed in the setting as a integer but it is never involved whatsoever in any calculation.

Any clearance would be very much appreciated.

Best regards,

Alexis

Hello Alexis,

Normally LIHBOR is taken in the boundary conditions file (first number on every line) but it can be changed to 1 (KINC) by PROPIN_TELEMAC2D.F when you have a boundary with velocity and depth free (KSORT=4) and a Froude number less than 1, which is not a well-posed case. This is the only possibility I see of changing LIHBOR, and this could be your problem though you seem to mention a value of 5 changed into 1, which I can't explain. Anyway you can look at this subroutine and cancel the tests involving the Froude number.

With best regards,

Jean-Michel

Jean Michel,

I have purposely created a conlim file to impose the water depth but have let velocities free. I have no wave coming into my domain so I have no reason to set LIHBOR to 1.
I will try understand this arising issue with KENT.

Thank you anyway for your help.

Best regards,

Alexis

Jean Michel,

In declarations_telemac.f KENT effectively equals 5, it has been changed to the value of KINT somewhere I ignore.

By the way, what is the procedure to recompile declarations_telemac.f (to update declarations_telemac.mod in bief\win).

Thank you in advance if you have any idea where KENT could have been modified to KINT and for the hint to compile (I am a fortran dummy).

With best regards,

Alexis

Hello Alexis,

I see no other possibility of changing LIHBOR from 5 (KENT) to KINC (1) than in propin_telemac2d.f, so this is a mystery for me if it is not this, unless it is due to what is said in the next paragraph...

Changing declarations_telemac2d.f is very dangerous if you do not recompile the whole system, because some routines will see the memory arrangement of the old version, so they will get wrong values if it does not crash. declarations_telemac.mod is the result of compiling, I'm surprised that you can edit it and you should not do it.

With best regards,

Jean-Michel

Jean Michel,

(i) I can't edit .mod file.

(ii) Maybe i did not make myself clear. I want KENT to be equal to 5 (as usual) but now it is equal to 1 (when i run my cas) and therefore as I set in my conlim file:

LIHBOR(ocean)=5
LIHBOR(river)=4
LIHBOR(wall)=2

hence the condition IF LIHBOR(K)=KENT THEN HBOR(K)=.. is never fullfilled!

I need to know why KENT is changed not LIHBOR (I made it).

Maybe i should have speak in french I would have make me understood better.

It is something even more puzzling as I did not encounter this issue when I ran it in Wales last year.


Sorry for bothering you with this technical issue.

All the best,

Alexis

Sorry Alexis, it was so unbelievable that I did not fancy that it was KENT itself that changed.

So it is very clear that only a bug can do that, probably a problem of arguments shift in your old 5.8 subroutines adapted for the new version. As a matter of fact KENT is given as a parameter and cannot be changed. An argument missing like NPTFR2 in bord.f could easily in this routine change KENT into the last value of NUMLIQ which could be 1.

If you do not find, put your fortran file in a post and we'll have a look at it.

With best regards,

Jean-Michel

Sincere apologies Jean Michel,

Now I have changed NPTFR into NPTFR2 where it was needed, KENT equals 5.

I am very sorry for all the unnecessary posts.

Best regards and merry christmas,

Alexis