open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi
I do a simualtion of a channel of with a length of 6 m and a width of 2m, the water depth is 0.02m with a flow of 0.01 m3 / s, slope of 1%, the number of froude in this case is V = Fr / (root (gh)) = 0.56 <1 so it's an uniform flow.

after the simualtion by Telemac 2D I have not found a uniform flow since I see a change in the flow between the different section and the speed?? and here is the table obtained value of the speed and flow??


can you help me to find the problem and generate a uniform flow?

thank you in advance

Hi,
I do not understand exactly what you want to reproduce and what you have tried.
Is your channel rectangular and what is your initial state(velocity and free surface)?

Fr = 0.56 < 1 just means that the flow regime is subcritical but do not imply that the flow is uniform.
The normal depth depends on the friction coefficient (Strickler, manning, Chézy...).

Hope this helps.
Luc

Hi

thank you for your reply, yes it's a rectangular channel( length is 6 m and width is 2m) for the intial condition i take a constant water level egal a 0.02 m for boundry conditions: débit imposé à l'amont = 0.01 m3/s , hauteur imposé à l'aval = 0.02 m. for the friction coefficient i use Strickler i'd like to check the uniform flow with this conditions

Hello,

In case of friction on the lateral walls, to obtain uniform flow (i.e. flow that its velocity field does not depend on the space coordinates), the flow has to be fully developped. This means that starting with a constant velocity profile at the inlet, for instance, the flow needs some time to retrieve a parabolic profile.

In case of no lateral friction on the walls, the uniform flow is obtained directly providing that the initial and boundary conditions are solution of the resolved equations (Saint Venant), otherwise, you need to wait until the steady state is obtained which will give uniform velocity profile.

be careful, you are working with a triangular mesh where nodes are not lined up as cross-sections. In order to see the velocity profile in a cross section, you need absolutely to use a post-processing tool.

With my best regards

Riadh ATA

Hi

thank you very much Mrs Riadh for your answer and explication. for the the velocity profile can I use fudaa prepro ? and if i use the normal depth as an boundry condtion at the outlet can this reproduce an uniform flow ? i will see with this condition also

thank you in advance

Hi,
I am not using Fudaa-prepro, but this should be possible through the keyword VELOCITY PROFILES. The defaut value is 1, which means "constant normal profile". But be careful, this should be defined as a vector (with 2 values if you have 2 boundaries).
As suggested by Riadh, the lateral friction has to be taken in consideration also.
I think this should work to reproduce a uniform flow, but you will have to adapt the velocity profil (especially if the channel is not long enough) according to the choice you do for the lateral friction.

Best Regards,
Luc

HI

Hi
I try to repeat the simulation with experimental data. uniform flow conditions is given with a length 8 m and width 0.4m in a rectanglar flume (Q = 0.022m 3 / s, h = 0.0785 m, average speed = 0.719 m / s, slope = 1%)
I use Telemac 2D with no lateral friction, flow imposed at the upstream and downstream height imposed, but I did not get the uniform conditions
the value of the average speed is around 0.7 m / s, I found an averaged velocity about 0.25 m / s and depend to the spatial variation.
i didn'tunderstand this note :(you need to wait until the steady state is obtained which will give uniform velocity profile) , this i didn't understand, that 's mean that i should change the data until a steady state is given ??? but i can't change neither the depeth nor the flow rate )

thank you in advance

Hello

We can not conclude directly without looking in depth at your model. Here some paths to investigate:
- do you introduce bed friction? the bed friction will dump the velocity field and you will have decelerated flow through the flume.
- do you introduce diffusion of velocity, this will have almost the same effect than the friction, although isotropically, but can alterate the establishment of a uniform velocity profile
- do you reach a steady state or not (at least verify that Q_in is equal to Q_out and that the free surface is stable). If not, transient effects can hide the uniform profile; this is what I meant by old post.
- To get the uniform profile, you have to look far enough from the entry of the flume in order to let flow finishing its development, that means to be far enough from the effect of the (double) boundary layer generated by the entrance.

I hope that this helpful,

with my best regards

Ridh ATA

HI

thank you very much

for the bed friction, yes i use because ( the simualtion is done for a vegetated flume so here i tike it as a roughness = 39.21
for the diffusion yes it's 0.01, i'll try to do a simualtion again without the diffusion
for the steady state, there is a problem in the entry here some result and the simualtion:

BANCS DECOUVRANTS = OUI
LOI DE FROTTEMENT SUR LE FOND = 3
COEFFICIENT DE FROTTEMENT = 39.21
MODELE DE TURBULENCE = 1
COEFFICIENT DE DIFFUSION DES VITESSES = 0.03
/

---

/
/ OPTIONS NUMERIQUES
/

---

/
/
COMPATIBILITE DU GRADIENT DE SURFACE LIBRE : 0.8
TRAITEMENT DU SYSTEME LINEAIRE : 2
/MASS-LUMPING SUR H : 1.
IMPLICITATION POUR LA HAUTEUR : 1
IMPLICITATION POUR LA VITESSE : 1
FORME DE LA CONVECTION : 1;5
OPTION DE SUPG : 1;2
MAXIMUM D'ITERATIONS POUR LE SOLVEUR = 150
PRECISION DU SOLVEUR = 1.E-6
SOLVEUR : 1
OPTION DU SOLVEUR : 1
CLIPPING DE H : NON
/

---

/ CONDITIONS INITIALES
/

---

/
/REMISE A ZERO DU TEMPS : OUI
CONDITIONS INITIALES : 'HAUTEUR CONSTANTE'
HAUTEUR INITIALE : 0.0785
/
/

---

/ CONDITIONS AUX LIMITES
/

---

/
COTES IMPOSEES : 0.0785;0.0
DEBITS IMPOSES : 0.0;0.0226
/
&FIN