open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all,
I'm still working on a 3D-model (Sisyphe+Telemac3D). Now we've a got the new version V7P1. Since I start to impose a sediment Rouse profile I got NaN but a constant profile (constant value of 1.2 g/L) it works.
Does anyone got the same problem?
I'm checking (learning I will say) how the Rouse profil has been implemented, it does not like pretty trivial to me:


FROM BORD3D
IF(IPROF.NE.1) THEN
PROFZ=TRA_PROF_Z(IFRLIQ,NBOR2%I(K),AT,LT,NP,
& INFOGR,IPROF,ITRAC)
IF(IPROF.EQ.2.OR.IPROF.EQ.0) THEN
! Rouse concentrations profiles (IPROF=2) or values given by user (IPROF=0)
TABORL%ADR(ITRAC)%P%R(IBORD)=PROFZ
ELSEIF(IPROF.EQ.3) THEN
! Normalised concentrations profiles (IPROF=3)
TABORL%ADR(ITRAC)%P%R(IBORD)=
& TABORL%ADR(ITRAC)%P%R(IBORD)*PROFZ
ELSE

FROM NOEROD
IF(SETDEP.EQ.1) THEN
!
DO IPOIN =1,NPOIN2
USTAR=MAX(SQRT(UETCAR(IPOIN)),1.D-6)
ROUSE=PRANDTL*WCS(IPOIN)/KARMAN/USTAR
! rouse profile extrapolation up to 1/4 of the first layer
DELTAZ=(Z(IPOIN +NPOIN2)-Z(IPOIN))/FICT
ROUSE_Z=ZREF%R(IPOIN)/(HN%R(IPOIN)-ZREF%R(IPOIN))
& *(HN%R(IPOIN)-DELTAZ)/DELTAZ
CREF%R(IPOIN)=CREF%R(IPOIN)*ROUSE_Z**ROUSE
ENDDO
!
ENDIF

Hello,

I checked on the trunk with the case called Rouse and all the profile options seem to work on this example, so we would need your case to see what happens (there may be a problem but not so obvious). If in doubt you can see with your version if the test case Rouse works with option 2 for the vertical profiles.
By the way there is also an option 4 for the modified Rouse profile, taking molecular viscosity into account, so that the value at z=0 is not infinite, but in that case bord3d.f must be changed:

the line: IF(IPROF.EQ.2.OR.IPROF.EQ.0) THEN

must be: IF(IPROF.EQ.2.OR.IPROF.EQ.4.OR.IPROF.EQ.0) THEN

Regarding your problem it does not seem to be linked with the version 7.1 since in version 7.2 it is the same subroutine, unless the change has not been documented (I have not downloaded all versions where I am, so I cannot check this).

With best regards,

Jean-Michel Hervouet

Hello,
Thanks for you answer. I also checked the case called Rouse but I think this case does not work as the V7P1. I mean, the Rouse’s profile has been implemented in PRERES_TELEMAC3D (! SPECIFIC TO ROUSE'S PROFILE BEGINNING OF ESSAI JMH). Even if I change the PROFILS DES TRACEURS SUR LA VERTICALE : 1;1 to 2;2 I think the soubroutine you’ve made imposed the profile as follow :
ELSE
DELTAZ=MESH3D%Z%R(NODE+(I-1)*NPOIN2)-MESH3D%Z%R(NODE)

I will change the bord3d.f today.
With best regards,
P.N

Hello,

The modification in PRERES_TELEMAC3D is only for the results file, to have the theoretical Rouse profile for comparison, this should not interfere with the rest of the program.

Regards,

JMH

Hello Norsazz,

I tested the Rouse example on v6p3 and the keyword TRACERS VERTICAL PROFILES (or in french PROFILS DES TRACEURS SUR LA VERTICALE, which is set to 1 by default) is taken into account.
The imposed vertical concentration profil is computed in TRA_PROF_Z (module telemac3d) based on the value of this keyword (0, 2 or 3).
Beware that the Rouse profile (if value is set to 2 or 3) depends on the equilibrium concentration (defined by REFERENCE CONCENTRATION FORMULA).

I glanced rapidly through the Rouse validation case and it seems to include only modifications (in Fortran file) for the initialization, erosion flux and output format, but not for the computed vertical sediment concentration profile.

Hope it helps.
With my regards,
Luc

Hello,
the nan value may come from a variable which has not been allocated. For instance if you are dealing with cohesive sediment, you should check how is computed zref.

best

Hi all,

I've done some research about Zref. This elevation corresponds to the interface between bedload and suspended load. In TELEMAC-3D (and also in SISYPHE), there is no bed load for cohesive sediment. That's mean physically (assuming no bed load for cohesive sediment) that Zref = Zbottom for cohesive sediment ?

In erodnc.f, Zref is computed as 2d50 but zref is not defined in erodc.f ?

I've try to declared DIAMETRE MOYEN DES GRAINS = 0.00001 (according to my settling velocity). This doesn't work at all. It didn't call erodnc.f cause I input SEDIMENT COHESIF = OUI

With bests regards,
P.N