open TELEMAC-MASCARET The mathematically superior suite of solvers

Good Morning!

I'm trying to model cohesive sediment transport with SISYPHE internally coupled with TELEMAC2D. I made a steering file and tested the model.

However, I couldn't get result if I set COHESIVE SEDIMENTS to YES. I just got the initial step of SISYPHE result and then nothing else. TELEMAC2D was working normally and can give longer results. There was no error message occurred. But if I set COHESIVE SEDIMENTS to NO and keep everything else the same. SISYPHE could perform calculation normally and give longer results. What has happened when the COHESIVE SEDIMENTS was set to YES?

Another question is that when I got the non-cohesive result, the time of each step in the result were different from those in telemac2D result. In telemac2d the time steps were every 15mins, like 00:00:00, 00:15:00, 00:30:00... In the SISYPHE result the time interval was also 15mins, but it likes 00:14:55, 00:29:55, 00:44:55...What is the reason for this?

Enclosed are the steering files for the model. Please help... Thanks!

Hello:

are you aware that the keyword 'critical erosion shear velocity of the mud' is no longer activated in 6.1. Try instead 'critical erosion shear stress of the mud'and check that your calculated bed shear stress is greater that the critical value. It may just be correct that nothing should happen!

You should also be aware that most keywords are not activated when you work with cohesive sediments (no bedload for example, no equilibrium conc...).

Good luck! Cath

Thanks Cath.

I think I've already used the keyword 'Critical Erosion Shear Stress of the Mud' in the steering file instead of 'Critical Erosion Shear Velocity of the Mud'.

Currently I only got the bed shear stress for non-cohesive case. In some areas of the domain, the bed stress was quite lower than critical shear stress with the magnitude of 1.0E-10 to 1.0E-30. But after several time steps, bed shear stress became larger and most of them was above the critical shear stress although I still got some Nan in the bed shear stress which may be in the dry area.

But if I switched to COHESIVE SEDIMENTS, I can't get the screen printout of the bed shear stress even I've already had the line 'VARIABLES TO BE PRINTED = 'TOB'' in my steering file.

Hi,

I think we have solved this problem. We found that the value of CONCENTRATION PER CLASS AT BOUNDARIES was not consistent with the INITIAL SUSPENSION CONCENTRATIONS. Now we simply canceled the keyword CONCENTRATION PER CLASS AT BOUNDARIES in the steering file and it works.

Hope this could help someone else.

Sorry, false alarm... We didn't solve this problem because I forgot to switch the keyword back to YES. The problem of keyword 'COHESIVE SEDIMENTS' still remains.

I got some more information about this problem by checking the main program 'sisyphe.f'.

In the line 519: IF(CODE(1:7) == 'TELEMAC'.AND.PART==0) THEN
In the line 738: IF(PART==1.OR.PART==-1) THEN

According to the screen printout, the program stops at:

INITIALISATION EN CAS DE COUPLAGE : PART= 0
CONDIM_SISYPHE
END_CONDIM_SISYPHE
INIT_SEDIMENT
END INIT_SEDIMENT
TOB_SISYPHE
END TOB_SISYPHE
INIT_TRANSPORT
QSCL REINITIALISED IN INIT_TRANSPORT
FOR CLASS 1
QS REINITIALISED IN INIT_TRANSPORT
END INIT_TRANSPORT

========================================
ITERATION 0 TIME: 0.0000 S

BIEF_DESIMP
END BIEF_DESIMP
SISYPHE COUPLED WITH: TELEMAC2D
END_INITIALIZATION

In the program the value of PART is 0. I guess this might be the reason why the SISYPHE didn't proceed with the calculation.

I'm going to test what will happen if I change PART=1.

I also found the corresponding value of PART in the main program 'telemac2d.f':

In the line 1105:
CALL SISYPHE(0,LT,LEOPRD,LISPRD,NIT,U,V,H,H,ZF,CF,CF,CHESTR,

In the telemac 2d, it gives the PART of value 0 instead of 1. It is not consistent with SISYPHE.

Hi
I'm not a specialist of the coupling between telemac and sisyphe but it seems to me that the PART flag exist to select the good part of sisyphe code during the execution.
Keep in mind that sisyphe could run alone with an hydrodynamic file instead the telemac result at the time step (In this case PART=-1)
PART=0 indicate the initialisation step of sisyphe when the run is coupled
PART=1 is for the time loop

Hope this helps

Thanks Christophe!

It makes sense that the problem was gone when COHESIVE SEDIMENTS equals to NO. In this case I suppose the variable PART should be fine.

Hi,

Finally I think I found the reason why if I set "COHESIVE SEDIMENTS" to YES then it didn't give the result. I looked at the code in the 'telemac2d.f' and I found this code block:

!=======================================================================
!
! COUPLING WITH SISYPHE
!
IF(INCLUS(COUPLING,'SISYPHE')) THEN
!
CALL CONFIG_CODE(2)
WRITE(LU,*) 'I AM HERE AT LINE 2135',SUSP1,CHARR,PERCOU,LT
!
!RK IF(SUSP1.OR.(CHARR.AND.(PERCOU*((LT-1)/PERCOU).EQ.LT-1))) THEN
IF(SUSP1.OR.(CHARR.AND.(PERCOU*(LT/PERCOU).EQ.LT))) THEN
!
IF(DEBUG.GT.0) WRITE(LU,*) 'APPEL DE SISYPHE, CHARRIAGE'
CALL SISYPHE(1,LT,LEOPRD_CHARR,LISPRD,NIT,U,V,H,HN,ZF,
& CF,CF,CHESTR,CONSTFLOW_SIS,NSIS_CFD,SISYPHE_CFD,
& CODE1,PERCOU,U,V,AT,VISC,DT*PERCOU,CHARR,SUSP1,
& FLBOR,SOLSYS,DM1,USIS,VSIS,ZCONV)
IF(DEBUG.GT.0) WRITE(LU,*) 'FIN APPEL SISYPHE, CHARRIAGE'
!
ENDIF
!
IF(SUSP.AND.PERCOU.NE.1) THEN
!
IF(DEBUG.GT.0) WRITE(LU,*) 'APPEL DE SISYPHE, SUSPENSION'
CALL SISYPHE(1,LT,LEOPRD,LISPRD,NIT,U,V,H,HN,ZF,
& CF,CF,CHESTR,CONSTFLOW_SIS,NSIS_CFD,SISYPHE_CFD,
& CODE1,1,U,V,AT,VISC,DT,CHARR_TEL,SUSP,
& FLBOR,SOLSYS,DM1,USIS,VSIS,ZCONV)
IF(DEBUG.GT.0) WRITE(LU,*) 'FIN APPEL DE SISYPHE, SUSPENSION'
!
ENDIF
!
CALL CONFIG_CODE(1)
!
ENDIF
!
!=======================================================================

Take my steering file (attached in the above post) for example.

For the Non-cohesive case, SUSP1=FALSE, so CHARR has to be TRUE which means the BED LOAD=YES, and (PERCOU*(LT/PERCOU).EQ.LT) will always remain TRUE. Then TELEMAC2D will 'APPEL DE SISYPHE, CHARRIAGE'.

For the Cohesive case, keyword 'BED LOAD' has been set to NO automatically, so SUSP1=FALSE, CHARR=FALSE, SUSP=TRUE, but since I chose PERCOU=1 (Coupling Period = 1), the condition (SUSP.AND.PERCOU.NE.1) was not satisfied. SISYPHE will not be called in either way.

So no matter COHESIVE SEDIMENTS was set to YES or NO, SISYPHE will only be called when BED LOAD = YES.

My question is that
1. What's the difference between SUSP1 and SUSP? Why SUSP1 always remains FALSE?
2. In the Cohesive case, why the coupling period (PERCOU) can't be 1?

Thanks in advance!

Hi,

Thank you for adding your steering file to the disussion. It does help to figure out what is going on. There are quite a number of dubious elements in your steering file. Many of those you can find in the manual, or check them by searching in the sisyphev6p1.dico file. Please try to go through those first, to check your settings.

The switch from non-chohesive to cohesive sediment

In v6p1,
CRITICAL EROSION SHEAR STRESS OF THE MUD
is not in the dico file. It should be:
CRITICAL EROSION SHEAR STRESS OF THE MUD PER LAYER
consequently Sisyphe is using
CRITICAL SHEAR VELOCITY FOR MUD DEPOSITION = 1000
with a very high critical shear velocity which propably means that there will be no erosion possible.

You should not be using
BED LOAD = YES
with cohesive sediments. Cohesive sediments are assumed to be transported through suspesion only. I doubt this is causing the error, Sisyphe should be ignoring it, but it irrelevant in any case.

By no means am I a cohesive mud specialist, but I suggest using the Zyserman et Fredsoe profile for reference concentration:
REFERENCE CONCENTRATION FORMULA = 1
Neither are very generic, I believe, but van Rijn's is definetly for sand only. Z&F is a bit more generic.

The entries
EQUILIBRIUM INFLOW CONCENTRATION = YES
CONCENTRATION PER CLASS AT BOUNDARIES = 0
are contradicting each other. I suggest to only use to start with and chaging it only when necassary.
EQUILIBRIUM INFLOW CONCENTRATION = YES

I am unsure about the following two settings in you steering file. I hope somebody else might be able comment on them. I think the flocculation effect is incorporated by changing the settling velocity in the case of cohesive sediments rather than through the sediment diameter and density:
The setting
SEDIMENT DIAMETERS = 0.0001
is confusing. This is a diameter for non-cohesive sediment. It does not define the floc size. The manual states that for cohesive sediments the grain size is generally smaller than 60 micron, suggesting that you shold use the single grain diameter, not a fluc diameter (Anybody, please correct me if I am wrong).
Similarly, I don't think the sediment density should include the flocculation effect
SEDIMENT DENSITY = 1100


Finally, with cohesive sediments, the coupling period can be equal to 1. The if statement just initiaties Sisyphe with a different call.

Good luck with your modelling and I am interested to see if you manage to model your test succesfully.

I hope this helps.

Regards,

Michiel