open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I was doing my first tests on 3D estuary model .
When I launch the test, I obtained

' CALCOT : LES PLANS 1 ET 2
SE CROISENT AU POINT 10883
COTE BASSE : -2.03361890925897
COTE HAUTE : -2.25733009971098
DIFFERENCE : -0.223711190452013
HAUTEUR : -0.677912698339432 '

which correspond in calcot.f to lines :

CALCOT: PLANES ',IPLAN-1,' AND ',IPLAN
WRITE(LU,*) ' INTERCROSS AT POINT ',IPOIN
WRITE(LU,*) ' LOWER POINT : ',ZZ(IPOIN,IPLAN-1)
WRITE(LU,*) ' HIGHER POINT: ',ZZ(IPOIN,IPLAN)
WRITE(LU,*) ' DIFFERENCE : ',ZZ(IPOIN,IPLAN)-
& ZZ(IPOIN,IPLAN-1)
WRITE(LU,*) ' DEPTH : ',HH(IPOIN)

Firstly, I thought, it's a problem with tidal falts, so I put 'TIDAL FLATS : 1' . But I obtain the same results.

Secondly, I increased the water level from 4.5 to 10. and the run starts.

Thirdly, I try to increase the time step from 0.05 to 0.1 and the error message appears once again

Do you have some suggestions ?

Bonjour Philippe,

You have to check the keyword :

TRANSFORMATION DU MAILLAGE

in your steering file, that must be consistant with what you program in subroutine CONDIM (it is used before we go into CONDIM, so we cannot deduce it from what you implement in it).

Also check that in CONDIM, if you give ZSTAR, you have written ZSTAR%R(1)=0.D0 and ZSTAR%R(NPLAN)=1.D0, this lacks in the example given and has already confused users.

Otherwise there was a formula in subroutine calcot.f that was written differently to avoid truncation errors that gave this problem, but in your case if you do not have tidal flats it is unlikely that it is this problem (I remember to have posted a correction on this forum but cannot find it...).

With best regards,

Jean-Michel Hervouet

Hello again,

In fact the main problem may be that you have a negative depth, as says the listing, so this may have to see with your subroutine CONDIM or your initial conditions, though depth is clipped within the standard CONDIM, and also after calling CONDIM, so I hardly understand how this negative depth can survive, or is it that you do a computation continued from a previous computation that would have negative depths in the results file ?

JMH

Bonjour Jean Michel,

Many thanks for your proposals, but :

In my CONDIM, I have the followings :
do iplan = 1,nplan
transf_plane%i(iplan) = 2
enddo
zstar%r(1) = 0.D0
zstar%r(2) = 0.33D0
zstar%r(3) = 0.66D0
zstar%r(4) = 0.75D0
zstar%r(5) = 0.87D0
zstar%r(6) = 1.D0

and in my steering file,:

NOMBRE DE PLANS HORIZONTAUX = 6
TRANSFORMATION DU MAILLAGE = 2

in my initial condition or/and in my boundary condition file, I imposed 10. m as COTE CONSTANTE.

This run is not the followwing of a previous one.

Anything else ?

Best regards

Philippe

Hello ,

Without Salinity, I succeed in starting the run with a time step of 0.1s, but a similar error message appears after 420 iterations (at another point)!

???

Hello,

I'm afraid you have to send your case so that I can see why subroutine calcot receives a negative depth that escapes control.

JMH

Dear Jean Michel,

Before sending you my files, I decided to delete some comments and unused keywords in the steering file and I found a second line with TIDAL FLATS = 0 behind the end of the file.
I deleted it and now the model starts.
I'm sorry to have disturbed for a blunder.
Many thanks for your help !

Best regards

Philippe

Dear all,

I would like to ask your opinion and help regarding Calcot error. I am having the same error as Phillipe had, and have tried every possible methods discussed in this forum to no avail.

I am running a simple channel model to observe the effects of momentum sink. I have been using both Zstar and Zplane transformation both my models and it worked well. Then as I increased the value of my variable that I am observing (e.g from a factor of 2 to 1000s), thats when it I got this dreadful error similar to Phillipe.

here is my vertical planes


DO IPLAN = 1,NPLAN
TRANSF_PLANE%I(IPLAN)=3
ENDDO
ZPLANE%R(2)=20.D0
ZPLANE%R(3)=22.5D0
ZPLANE%R(4)=25.D0
ZPLANE%R(5)=27.5D0
ZPLANE%R(6)=30.D0
ZPLANE%R(7)=32.5D0
ZPLANE%R(8)=35.D0
ZPLANE%R(9)=37.5D0
ZPLANE%R(10)=40.D0

From my observation
1. the error happens at different node for different simulations
2. the model run at first then the error camne up after few hundreds iterations
3. based on my zplane vertical layer, layer 1 should be = 0. Instead, when the simulation is running, layer 1 becomes 2m (instead of 0) while other layers was as coded. this happened both when I used Zstar and Zplane transformation.

the variable that I am investigating is as follows:

FORCEH = -(0.5D0*2.D0/1.5D0)*(UN3(I3D)**2+VN3(I3D)**2)

using the above equation, the models run well. However, when I changed 2.D0 to a higher value, thats when I got the error.

I would appreciate any helps on this matter

Thanks

Here's the selafin file that I used for the model

Hello,

We would need to see your Fortran file, because I do not understand what is the link between FORCEH and the coordinates of the mesh. Calcot also gives a message, including sometimes that the depth is negative. We have secured cases where two planes are superimposed and truncations errors give the impression that they intercross, but maybe you are in another overlooked case (in this case planes intercross by only 1.D-12 or something like this. Another possibility would be that you have dry zones and not TREATMENT OF NEGATIVE DEPTHS = 2.

With best regards,

Jean-Michel Hervouet