open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear all,

I have observed a problem when I use several frictions zones.
I have performed simulations on a simple case with a rectilinear chanel.
When I use one friction coefficient on the bottom + a friction coefficient for the lateral boundaries (LAW OF FRICTION ON LATERAL BOUNDARIES = YES), everything works well (I observe a decrease of the flow velocities near the lateral boundaries).
However, when I use 2 different friction coefficients on the bottom + a friction coefficient for the lateral boundaries, I do not observe a decrease of the velocities near the lateral walls. For this case, I have seen in the Friction_Zones.f subroutine, that the variables KFRO_B (friction law for lateral boundaries) and CHBORD (friction coefficient for lateral boundaries) are equal to zero when the Friction_Calc.f subroutine is called. Thus I think that there is no friction law and no roughness coefficient applied to the wall when several friction zones (for the bottom) are used.
To resolve this problem, I imposed locally the variables KFRO_B and CHBORD in the Friction_Zones.f but it is not a perennial action. Perhaps, there is a better way to solve this problem?
Thank you in advance.
Best regards.

Nicolas

Hello,

After looking at the implementation, this behaviour is "normal", though not enough explained in the documentation. I mean it is intentionnally programmed like this. It is only when a k-epsilon model is used that friction data on the boundaries must be added in the file. I think however that we should do it another way, either by imposing the friction boundary conditions given by keywords, or by asking in all cases the data on boundaries in the friction data file, any preference ?

Regards,

JMH