open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello all,

I am having problems continuing a tomawac model based on a previously computed GLOBAL RESULTS FILE. So I started with the TOMAWAC example Next_Comput and add the options that my model has one at a time.

The option that made the computation fail was setting CONSIDERATION OF A WIND = YES. Here are the entire wind definition options:
But since I had noticed that the GLOBAL RESULTS FILE adds a CURRENT-WIND variable when wind generation is enabled, I run the Shoal case by adding the above options, recreated the GLOBAL file and fed it to Next_Comput. It did not make any difference.
Here is the full listing output:



I am using trunk revision 9915. Any help would be appreciated.

Best Regards,
Costas

Hello Costas,

There was a problem before you next computation issue. In shoal you had a problem to generate a result, for a reason completly different than next computation linked to output variables. It is fixed in rev 10119.

Unfortunately it didn't fix all your problems, there is still a problem with Next Computation

best regards

T.

For your second problem try to replace your routine lecsui.f in tomawac
by this one and i guess it will work, i just succeed to have wind with next computation with the changes i mentioned.

Hello Thierry,

Fantastic! Next_Comput now works absolutely fine with wind enabled. I will put it to test with my real case scenario and get back.

Thank you very much.

Regards,
Costas

I can confirm that NEXT COMPUTATION works in my case too and so far results appear to be as expected. That saves me about a quarter million TOMAWAC timesteps per scenario. :)

Regards,
Costas

Dear Thierry et al,

I too have been struggling with tomawac continuation using v7p2r0.

For example I have run the test case whirl_current generating results file: r2d_whirl_ini.slf

However, I then get a fatal error when attempting a continuation with:

GLOBAL RESULT FILE = 'glo_next.slf'
PUNCTUAL RESULTS FILE = 'tom_whirl.spe'
NEXT COMPUTATION = TRUE
PREVIOUS COMPUTATION FILE = 'r2d_whirl_ini.slf'
PREVIOUS COMPUTATION FILE FORMAT = 'SERAFIN '
2D RESULTS FILE = 'r2d_whirl.slf'

Thinking this might be occurring in LECSUI.f, I removed all the STOPs to yield:

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

READ_MESH_INFO: TITLE= TEST WHIRL CURRENT
NUMBER OF ELEMENTS: 882
NUMBER OF POINTS: 484

SINGLE PRECISION FORMAT (R4)

LECSUI: BAD NUMBER OF PLANES IN THE PREVIOUS
COMPUTATION FILE : 0 FOUND
48 EXPECTED
LECSUI: BAD NUMBER OF POINTS IN THE PREVIOUS
COMPUTATION FILE : 484 FOUND
484 EXPECTED
LECSUI : BAD NUMBER OF FREQUENCIES
IN THE PREVIOUS
COMPUTATION FILE : 10 FOUND
25 EXPECTED

**** FOLLOWING COMPUTATION ****

TITLE OF THE PREVIOUS COMPUTATION :
TEST WHIRL CURRENT SERAFIN

- COMPUTATIONAL RESUMPTION AT TIME 60000.0000000000
ERREUR 1001 LORS DE L APPEL A LECSUI:FIND_VARIABLE
TEXTE DE L'ERROR :
HERMES_VAR_UNKNOWN_ERR

PLANTE: PROGRAM STOPPED AFTER AN ERROR
RETURNING EXIT CODE: 2
_____________
runcode::main:
:
|runCode: Fail to run

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

The resumption time is correct, but something is going wrong in LECSUI.f

I have tried running, but without success, with the updated 2017 version of LECSUI.f included earlier in this Topic.

If I am tripping up on something obvious, it would be most helpful to know since tomawac continuation is becoming an important requirement for me in coupled T2D<>Tom<>Sisphe runs.

Thanks, Alan

Hello Alan,

I am not sure i understand i would need your whole case to be sure, but it seems that you want to do a continuation with a 2d result file but with tomawac you need a global result in your first calculation. It is not like telemac2d because we have 4D, we can't restart from a 2D file.
I would bet that in your first calculation you saved 10 variables and lecsui see them as 10 frequencies.
If it is not that please post your case

Best regards

T.

Thanks for the quick response, Thierry!

I'll check things through again for this standard test.

I make an initial run (with probably an unnecessary number of output variables), as follows:

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
FORTRAN FILE = 'tom_whirl.f'
GEOMETRY FILE = 'geo_whirl.slf'
BOUNDARY CONDITIONS FILE = 'geo_whirl.cli'
BINARY CURRENTS FILE = 'current.slf'
CURRENTS FILE FORMAT = 3
/
GLOBAL RESULT FILE = 'glo_next.slf'
PUNCTUAL RESULTS FILE = 'tom_whirl.spe'

2D RESULTS FILE = 'r2d_whirl_ini.slf'
/
PARALLEL PROCESSORS = 4
/

---

/ DIVERS ENTREES-SORTIES
/

---

TITLE = 'TEST WHIRL CURRENT'
PERIOD FOR GRAPHIC PRINTOUTS = 25
VARIABLES FOR 2D GRAPHIC PRINTOUTS =
HM0,DMOY,FMOY,FM01,UX,UY,ZF,WD,TMOY,TPR5,TM02,SPD
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

This initial run ends correctly. I then try to continue as follows:

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

FORTRAN FILE = 'tom_whirl.f'
GEOMETRY FILE = 'geo_whirl.slf'
BOUNDARY CONDITIONS FILE = 'geo_whirl.cli'
BINARY CURRENTS FILE = 'current.slf'
CURRENTS FILE FORMAT = 3
/
GLOBAL RESULT FILE = 'glo_next.slf'
PUNCTUAL RESULTS FILE = 'tom_whirl.spe'
NEXT COMPUTATION = TRUE
PREVIOUS COMPUTATION FILE = 'r2d_whirl_ini.slf'
PREVIOUS COMPUTATION FILE FORMAT = 'SERAFIN '

2D RESULTS FILE = 'r2d_whirl.slf'

/
PARALLEL PROCESSORS = 4
/

---

/ DIVERS ENTREES-SORTIES
/

---

TITLE = 'TEST WHIRL CURRENT'
PERIOD FOR GRAPHIC PRINTOUTS = 25
VARIABLES FOR 2D GRAPHIC PRINTOUTS =
HM0,DMOY,FMOY,FM01,UX,UY,ZF,WD,TMOY,TPR5,TM02,SPD

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

This fails. So there is some slip in the logic. I can send the full details if that helps. But, as I say, I'll run through the steps for this test again first.

Regards, Alan

Hello Alan

That is exactly what i supposed in the second run you try to use a 2d file result (r2d_whirl_ini.slf) when you need a global result
You should put in your second run

PREVIOUS COMPUTATION FILE = 'glo_next.slf'

and give a different name to your global result file (which can be used for another continuation)

In tomawac 2d results are not the unknown of tomawac (the unknown of tomawac is a density of energy in a direction for a frequency at a position) but post treatments of the solution (different integrations on angles and frequencies). So you have less information in a 2d results to be more understandable.

A restart must be with the whole solution for each angle and each frequency that we call "global result file".
Hope that is clear

Best regards

T.

Thanks again, Thierry,

The logic as now clear to me. I have made the change(s) suggested and have run both stages again. But it still fails on continuation with:

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

FORTRAN FILE = 'tom_whirl.f'
GEOMETRY FILE = 'geo_whirl.slf'
BOUNDARY CONDITIONS FILE = 'geo_whirl.cli'
BINARY CURRENTS FILE = 'current.slf'
CURRENTS FILE FORMAT = 3
/
GLOBAL RESULT FILE = 'glo_next_new.slf'
PUNCTUAL RESULTS FILE = 'tom_whirl.spe'
NEXT COMPUTATION = TRUE
PREVIOUS COMPUTATION FILE = 'glo_next.slf'
/PREVIOUS COMPUTATION FILE FORMAT = 'SERAFIN '

2D RESULTS FILE = 'r2d_whirl.slf'
PARALLEL PROCESSORS = 4

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
Error:

*************************************
* END OF MEMORY ORGANIZATION: *
*************************************

40 40
INBIEF (BIEF): NOT A VECTOR MACHINE (ACCORDING TO YOUR DATA)
40 40
STRCHE (BIEF): NO MODIFICATION OF FRICTION

TOM_CORFON : NO MODIFICATION OF BOTTOM

GET_MESH_TITLE: BAD FILE FORMAT: @
PLANTE: PROGRAM STOPPED AFTER AN ERROR
RETURNING EXIT CODE: 2
_____________
runcode::main:
:
|runCode: Fail to run
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

I'll play around with it further. But, if all else fails, would it be too much trouble to get you to run it there. The run takes 4 s !

Regards,

Alan