[azlan]

Hi

I am beginner in FLOW 3D,and I wish to simulate linear waves in a tank.
Unfortunately,I found the free surface elevation is not linear as a theory. If I choose "outflow" as the boundary condition,the free surface elevation will reduce in a steady rate. Same thing happened when I change the boundary condition into "wall" or "Symmetry", the free surface elevation is still not linear.
So I really wish to know why does it happen and how to solve this problem, can any experienced user can help me out?
Thanks.
Azlan

[Allenlin]

Hi, azlan

The simple way is extend you computing domain, to reduce the outflow boundary influence. If the phenomenon that you focus is not include the downstream side.

[azlan]

tq allenlin..

i already extend the domain.. unfortunately, the wave is still nonlinear.. did you have any example of linear wave propagation using flow 3d?

hope you can help me..

Thank you very much..

[Allenlin]

To azlan

From your post, the free surface elevation will reduce. I assume that the simulated wave shape is similar with theory. It looks like the numerical diffusion error is not under control. Can you try to increasing your mesh resolution ?

[Ricky Dada]

I'm trying to simulate actions of a piston type wavemaker in a numerical wavetank using GMO for prescribed movement of the piston. The fluid is inviscid & incompressible with a free surface. The numerical domain is 2D, with x as lengthwise axis & z as axis along vertical direction. The start of the wave tank has the piston position at its mouth. The bottom of the tank is closed while the top is open to atmosphere. Boundary conditions are as follows
Xmin-Wall
XMax- Outflow
Zmin-Symmetry
Zmax- specified pressure(Stagnation pressure is turned on with fluid fraction pecified as 0. Pressure specified is atmospheric).

Intial conditions:
Total depth of tank is 0.8 m. Water filled upto water depth i.e. 0.6 m.Void pressure is atmospheric.

Prescribed Conditions on Piston GMO:
It is desired to impose a sinusoidal motion of period 1 sec on the piston. However in order to start motion in a smooth manner, i've imposed velocity for first 4 secs using the table format. Later i use a restart simulation to simulate for period between 4-30 secs using the results of previous simulation as described before. Therein i use inbuilt sinusoidal prescribed motions on the piston with appropriately calculated initial phase.

My Problem is first, since flow-3d uses immersed boundary method, is specifying Xmin as wall type boundary right?

Secondly, all wave kinematics calculated analytically do not match with flow 3d calculations. I've seen a Phd thesis from one Eoghan Maguire who has reported good results for simulation upto 20 secs. However i'm unable replicate his results. Any help??

[Ricky Dada]

I'm trying to simulate actions of a piston type wavemaker in a numerical wavetank using GMO for prescribed movement of the piston. The fluid is inviscid & incompressible with a free surface. The numerical domain is 2D, with x as lengthwise axis & z as axis along vertical direction. The start of the wave tank has the piston position at its mouth. The bottom of the tank is closed while the top is open to atmosphere. Boundary conditions are as follows
Xmin-Wall
XMax- Outflow
Zmin-Symmetry
Zmax- specified pressure(Stagnation pressure is turned on with fluid fraction pecified as 0. Pressure specified is atmospheric).

Intial conditions:
Total depth of tank is 0.8 m. Water filled upto water depth i.e. 0.6 m.Void pressure is atmospheric.

Prescribed Conditions on Piston GMO:
It is desired to impose a sinusoidal motion of period 1 sec on the piston. However in order to start motion in a smooth manner, i've imposed velocity for first 4 secs using the table format. Later i use a restart simulation to simulate for period between 4-30 secs using the results of previous simulation as described before. Therein i use inbuilt sinusoidal prescribed motions on the piston with appropriately calculated initial phase.

My Problem is first, since flow-3d uses immersed boundary method, is specifying Xmin as wall type boundary right?

Secondly, all wave kinematics calculated analytically do not match with flow 3d calculations. I've seen a Phd thesis from one Eoghan Maguire who has reported good results for simulation upto 20 secs. However i'm unable replicate his results. Any help??