open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I was running a pretty basic SISYPHE v6p1 simulation, chained with TELEMAC output and the simulation crashed unusually at the first time step, with no feedback to the sortie file. I traced the endpoint of the simulation to CARACT.f (which I understand is related to the method of characteristics, and in this case, advection of sediment concentrations) and the following loop:

DO I=1,U%N
IF(U%ADR(I)%P%ELM.EQ.13) THEN
QUAD = .TRUE.
ENDIF
ENDDO

As I see there have been many changes in this part of the code in the latest version, I decided to try converting my model to v6p2. The simulation still crashed, but this time it gave the feedback:

BEWARE, ALL LAYERS EMPTY, NODE I = 17726
PLANTE: PROGRAM STOPPED AFTER AN ERROR

My guess was that this was somehow related to the initial condition for sediment concentrations (zero, initially) and, sure enough, when I changed the initial condition to a non-zero value, the simulation runs ok. The funny thing is that I have gotten some previous simulations (with a similar setup) to run with a zero initial condition - this is the first time I have encountered this issue.

Could anyone explain why this might be, and also, if there are any changes from v6p1 to v6p2 that affect this issue?

Thanks in advance,

Enda

Dear Enda,
you can take a look to the subroutine condim susp.f, where the initial values of volume concentration for each class of sediment can be imposed.
Do not forget that your initial concentration can be specified in the steering file through the keyword INITIAL SUSPENSION CONCENTRATIONS.

Please let me know,

All the best,

Pablo

Hello Pablo,

Thank you for your response and advice. I understand how to modify/implement the initial conditions. My question was more to try to understand why a zero initial concentration was apparently causing problems (and ultimately, a crash).

Thanks,

Enda

Dear Enda,
below some hints :
* which numerical scheme for the advection are you using? Schemes 4 is recomended (or 14 if tidal flats are present in your simulation)
* do you have tidal flats in your simulation ?

all the best,

Pablo