open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear all,

I would like to get clarification regarding the different options for accounting for the dispersion in sisyphe. Until now, I was using option 1 (OPTDIF = 1) where the keywords DISPERSION ALONG THE FLOW and DISPERSION ACROSS THE FLOW allow to input a value for the dispersion coefficient. However, I wanted to test the second option (OPTDIF = 2) which uses the Elder model to calculate the dispersion coefficients as follows:

Long. disp. coeff = alphal * u* * h
Tran. disp. coeff = alphat * u* * h

where alpahl and alphat are dimensionless dispersion coefficients having the value of 6 m2/s and 0.3 m2/s respectively.

However, if I look at the subroutine in the sisyphe library where the dispersion is calculated i.e. suspension_dispersion.f, the following formula is applied:

IF(OPTDIF == 2) THEN
DO K = 1, NPOIN
UETH = SQRT(TOB%R(K)/XMVE)
T1%R(K) = XKX * UETH * HN%R(K)
T2%R(K) = XKY * UETH * HN%R(K)
ENDDO

My question is what are XKX and XKY ? I could not find where these varaibles were calculated. It is said that they correspond to COEFFICIENT USED FOR COMPUTING THE DISPERSION. Does this mean that they correspond to alphal and alphat as explained in the formula or do they correspond to the values of the keywords DISPERSION ALONG THE FLOW and DISPERSION ACROSS THE FLOW? In that latter case, then the formula in the subroutine is different from the formula mentioned in the sisyphe manual.

Thank you for your clarification.

Jean-Philippe

Dear Jean-Philippe,
the keyword "OPTION FOR THE DISPERSION" allows to choose 3 differents options for the treatment of the dispersion term: (1) affected by a constant value alpha; (2) affected by streamwise and crosswise coefficients XKX and XKY; and (3) computed from t2d.

For the case (2), the coefficients XKX and XKY can be provided by the user via the keywords DISPERSION ALONG THE FLOW and DISPERSION ACROSS THE FLOW, respectively or by using the default, non-dimensional values XKX=1.E-2 and XKY=1.E-2.

The user's manual is being revisited, I hope it will be ready soon.

all the best,

pablo

Ok thank you Pablo

However, where can I set up the value of this alpha constant in the case I want to use option 1 (OPTDIF)? In fact, I am playing with the dispersion term as a way to get my inflow suspended concentrations travelling further within my computational domain. Indeed, I set up an inflow suspended concentration of 5mg/l at an open boundary but in my simulations, this inflow SSC do not progress within the domain but directly deposits near the boundary section. Flow velocities are pretty slow (i.e., max 0.05m/s), so as it is usually the case in tidal systems, suspended transport is rather a diffusive process, this is why I wanted to try different options for the dispersion term in an attempt to have a higher dispersion effect of the sediment transport . Another reason might be related to the settling velocity which seems to be a bit high after I have made a quick estimation of the order of magnitude (10^-2 m/s) based on the Van Rijn formula, but since this parameter is sensitive, it might be difficult to impose the right value using the corresponding keyword.

Thank you for your reply

Dear Jean-Philippe,
* with option (1) you can provide directly the values of the longitudinal and transversal dispersivity coefficients (T1 and T2 respectively) with the keywords DISPERSION ALONG THE FLOW and DISPERSION ACROSS THE FLOW.

* with option (2) the values of the longitudinal and transversal dispersivity coefficients are computed with the expressions T1=alphal u* h and T2=alphat u* h, where the coefficients alphal and alphat can be provided with the keywords DISPERSION ALONG THE FLOW and DISPERSION ACROSS THE FLOW.

This kind of "overriding" operation is valid, but could be confused for the users. We will clarify that in the future.

By the way, I'll take a closer look in the way in which the components of the dispersion tensor are computed and I'll back to you soon.

All the best,

Pablo

Jean-Philippe,

With flow velocities as low as 0.05, you will struggle to get any sediment staying in suspension! Even as small as 0.01 mm, and assuming no cohesion effecs (unlikely for such fine material), the critical velocity to get sediment moving is in the range of 0.2-0.3m/s depending on depth. Playing around with dispersion will not geet you far and it would not be very realistic anyway. flows of 0.05 m/s will lead to clear water.

Michiel