TELEMAC-MASCARET Forum: Artificial turbulence and drag from head losses (1/1)

Welcome, Guest

TOPIC: Artificial turbulence and drag from head losses

Artificial turbulence and drag from head losses 6 years 6 months ago #29969

NJ

Hello everyone,
I am trying to see how well Telemac 3D simulates the effects of vegetation in a channel. I read that there's no porosity option so I decided to use head losses at certain nodes, seeing this method used elsewhere in the forum. I used head loss equations inspired by ones others used in the forum. The velocities look fine but I'm not seeing much effects from TKE. I'd expect that as velocities decrease from friction and turbulence, the TKE values would increase but I'm not seeing anything!

I must have misunderstood how Telemac applies head losses. I fear this may be something basic I've overlooked but I can't find what it is! Does anyone know why this might be?

Here's the code I used:

CALL OS( 'X=0      ',X=S1U)
CALL OS( 'X=0      ',X=S1V)
CALL OS( 'X=0      ',X=S1W)
S1U%TYPR='Q'
S1V%TYPR='Q'
S1W%TYPR='Q'

DO I = 1,NPOIN2
 IF (ID.EQ.I)THEN
  DO IPLAN = 1,NPLAN !for layers at that node
   I3D = I+NPOIN2*(IPLAN-1) !get the 3D node IDs

!then apply drag at those nodes
 NORM =SQRT(UN3%R(I3D)**2+VN3%R(I3D)**2+WN3%R(I3D)**2)
 S1U%R(I3D)=0.5D0*27*CD2*NORM
 S1V%R(I3D)=0.5D0*27*CD2*NORM
 S1W%R(I3D)=0.5D0*27*CD2*NORM
END DO

And here's images of the nodes with head losses, affecting velocity but not TKE:

The administrator has disabled public write access.

Artificial turbulence and drag from head losses 6 years 6 months ago #29995

aleroy

Hello,

It is hard to tell the effect of your change since you only show the results with the head loss, and not without it...

Anyway, the equation on k is basically:
Dk/Dt = Prod - epsilon + diffusion
where Dk/Dt contains the advection and partial time derivative of k. The term Prod is a linear or quadratic function of the scalar strain rate S, which depends on the velocity gradient. Epsilon follows a similar equation than k. This means that in order to incease k, it is necessary to increase the velocity gradients: if the simulation with head loss addition at some points is not increasing the value of the velocity gradient components, there is no reason why k should increase. Also, keep in mind that the increase of production is always partially balanced by an increase of the dissipation, epsilon, so the effect on k is not linear.

For more details about the k-epsilon model, please consult the Telemac-3D theory guide avilable here (section 1.6.3):
wiki.opentelemac.org/doku.php?id=documentation_v7p3r0

I hope this helps,
Agnès

The administrator has disabled public write access.