Springs at free surface

[alexchr]

Hi,

Could anyone advice me how we can set springs in a free surface of a fluid80 element so that nodes could move only in vertical directions?

 

[Wecos]

Why not just fix the nodes in the X and Z directions? (assuming Y is vertical)

 

[alexchr]

Thanks Wecos

I thought it could be more complex.

Have you got any idea about the parameter "location of gravity spring" in fluid80 control?

Should we use "all side all elem" or "ele faces at Z=0?

is there any great difference between these two?

 

[Wecos]

From what I understand, the "gravity spring" is a default effect of the elem. type. It basicly just keeps the fluid from floating away as long as gravity effects are included in the model (see ACEL command).

I'm not sure of you model or constraints, but I would think that "all side all elem" would best apply, especially if you have FLUID80 elements with faces other than Z=0

 

[alexchr]

Thanks Wecos.

I have model a tank with shell elements and the internal fluid with fluid80.

1. I have rotate CS to cylindrical.
2. I have the same mesh in tank perimeter. Also coincident nodes in Ux direction. (Uz=vertical)

3. Fluid at the bottom has Ux=Uy=Uz=0.

BUT nodes of fluid in perimeter act like crazy (in Uz) and nodes of fluid in the internal also act like crazy (in Uy,Ux)!

Have you got any idea of what is going wrong?

I think I 've try anything and read help topics again and again..

Thanks for your time

 

[Wecos]

Is the fluid completely enclosded by the tank? Constraining fluid nodes should not be necessary as the fluid elements would already be constrained by the shell elements of the tank.

You could try removing the fixed displacement constraint on the fluid and let it act freely against the tank.

 

[alexchr]

Wecos,

the only load is ACC 1g downwards (z-), but nodes in fluid perimeter move and upwards(z+)!

I really can't find out what it is happening.What else should I try?

Material model for fluid has

Density=10kN/m3
V=1*10^-6
EX=1320kN/m2
?=0.4999

 

[Wecos]

Reverse your acceleration. One of those wierd ansys things. The ACEL command accelerates the model not the environment, so to simulate gravity, the model is accelerated in the positive (Z+) direction.

From the Ansys help menu:

"To simulate gravity (by using inertial effects), accelerate the structure in the direction opposite to gravity. For example, apply a positive ACEL_Y to simulate gravity acting in the negative Y direction."

 

[alexchr]

Wecos,

Yes, I 've simulate gravity effect in positive direction and got no correct results. Beside coincident nodes between shell perimeter and fluid perimeter should i try something else eg any kind of coulped set?

I think I 'll give up and simulate liquid with solid45 elements

 

[alexchr]

Wecos,

I set coincident nodes between fluid perimeter and tank shell and coupled set in uz direction of fluid volume(inculing perimeter) and it worked.

Results seems logical in gravity effect...but performing modal analysis a lot of work has to be done.