TELEMAC-MASCARET Forum: Subroutine bord.f and max iteration (1/1)

Welcome, Guest

TOPIC: Subroutine bord.f and max iteration

Subroutine bord.f and max iteration 13 years 2 months ago #2399

Romain_D
OFFLINE
Fresh Boarder
Posts: 29

File Attachment:

File Name: global.zip
File Size: 888 KB

Hi everyone,

Actually I work on hydrodynamic model with subroutine bord.f and i have some computation problems.

Before I explain you my problem, here is some informations about my model (.cas, .geo, .conlim, .f at the end of the post).

4 boundaries:
North: 5665
South: 5665
West: 5444
East: 2222

For North and South boundaries, I have initialized a table in bord.f to vary in time&space U and V.
For North, South and West I calculated the sinusoïde tide with a different phase between North and South.

For now the model doesn't converge.

Screen:

 ITERATION      150  TEMPS :  5 MN   0.0000 S   (      300.0000 S)
--------------------------------------------------------------------------------

                       ETAPE DE CONVECTION
--------------------------------------------------------------------------------

                   ETAPE DE DIFFUSION-PROPAGATION
 EQUNOR (BIEF) :        9 ITERATIONS, PRECISION RELATIVE:   0.7826496E-04
--------------------------------------------------------------------------------

                        BILAN DE VOLUME D'EAU
     VOLUME DANS LE DOMAINE :   0.3397383E+11 M3
     FLUX FRONTIERE    1 :     0.000000     M3/S  ( >0 : ENTRANT  <0 : SORTANT )

     FLUX FRONTIERE    2 :     0.000000     M3/S  ( >0 : ENTRANT  <0 : SORTANT )

     FLUX FRONTIERE    3 :   -0.2740187-273 M3/S  ( >0 : ENTRANT  <0 : SORTANT )

     ERREUR RELATIVE EN VOLUME A T =        300.0     S :    0.1750116E-02
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.1437
061E-03
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.7163
985E-03
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.1619
366E-02
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.2425
316E-02
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.5373
294E-02
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.1030
613E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.4661
897E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:    1.163
618
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:    4.048
626
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.2372
905E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.1356
976E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.3049
373E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.3380
702E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.4989
989E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.7887
668E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.6392
131E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.4199
243
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:    1.248
601
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.9928
960E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.3454
822E-01
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.3705
168
 EQUNOR (BIEF) : MAX D' ITERATIONS ATTEINT:     100 PRECISION RELATIVE:   0.8222
196E-01

I think the problem come from bord.f but where?

Could you help me?

Regards.

Romain_D

The administrator has disabled public write access.

Re: Subroutine bord.f and max iteration 13 years 2 months ago #2403

jmhervouet

Hello,

Sorry, if you need someone to debug your developments, you should consider taking a maintenance and assistance contract. What I just can say is that you should look at the option :

TREATMENT OF THE LINEAR SYSTEM :

If it is 1 you could try the wave equation (option 2) and in this case move to conjugate gradient (SOLVER : 1).

On what you did in Bord : I just had a glance and found no mistake (except that K=24 is set in 2 different loops, 40 and 41) but your programming is very dangerous : use double precision instead of reals and think that expressions like (K-698)/(722-698) will give 0 because 3/4 in Fortran is 0 for example (integer division), so you are playing with fire if you do not transform your integers into double precision.

With best regards,

Jean-Michel Hervouet

The administrator has disabled public write access.

Re: Subroutine bord.f and max iteration 13 years 2 months ago #2404

Romain_D OFFLINE Fresh Boarder Posts: 29	Hello, Thanks for your advice. I try it rigth now. If I fix my bug I'll post the solution. Regards. Romain_D
	The administrator has disabled public write access.