CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FloEFD, FloWorks & FloTHERM (http://www.cfd-online.com/Forums/floefd-floworks-flotherm/)
-   -   shock absorber simulation (http://www.cfd-online.com/Forums/floefd-floworks-flotherm/71242-shock-absorber-simulation.html)

 dik December 21, 2009 09:04

shock absorber simulation

hi all, I have a problem about simulating a shock absorber fluid.

here's the model, a very simplified model as a preliminary design only to test the general flow characteristics. the real model will be more complicated but understanding the basic phenomena is a very important step.

http://i47.tinypic.com/2yllmap.jpg

i've tried setting the Boundary Condition of 2,3,4,5 faces (piston) as downward moving wall -0.2 m/s (minus sign for downward motion on Y-axis), while faces 1,6 and the cylinder wall set as a stationary wall BC. the results for this setting is qualitatively incorrect, indicated by the unreasonably lower velocity at the four orifices (3) where according to continuity it shuold be higher than any other points in the model.

second trial, i set the BC of 6 as inlet velocity of 0.2 m/s (upward on Y-axis). and i set 1 as a pressure opening (default value, about 1 atm). the cylinder wall was a moving wall which velocity is the same with the inlet velocity of 1 (0.2 m/s upward) and the remaining is a stationary real wall. the result was more make sense, at least qualitatively correct at the orifices where it gave more velocity than other points. but the problem is: face 1 and 6 is in fact a wall, not openings. setting 6 as an opening will result an eliminated rebound effect due to the collision between the fluid and the wall. just consider a shock absorber, there's no atmospheric opening.

so, what should I do to make this simulation realistic (not only just seems realistic, but also show the real phenomena indeed).

what should the Boundary Conditions of the faces?

can somebody help me with this problem? you can use the numbers in the picture above to designate the face in your explanations.

Thanks for viewing this thread. I will be glad and highly appreciate if you post some reply. well, once again, thanks folks.

 MuratSuer March 3, 2010 22:05

Quote:
 Originally Posted by dik (Post 240647) hi all, I have a problem about simulating a shock absorber fluid. here's the model, a very simplified model as a preliminary design only to test the general flow characteristics. the real model will be more complicated but understanding the basic phenomena is a very important step. http://i47.tinypic.com/2yllmap.jpg i've tried setting the Boundary Condition of 2,3,4,5 faces (piston) as downward moving wall -0.2 m/s (minus sign for downward motion on Y-axis), while faces 1,6 and the cylinder wall set as a stationary wall BC. the results for this setting is qualitatively incorrect, indicated by the unreasonably lower velocity at the four orifices (3) where according to continuity it shuold be higher than any other points in the model. second trial, i set the BC of 6 as inlet velocity of 0.2 m/s (upward on Y-axis). and i set 1 as a pressure opening (default value, about 1 atm). the cylinder wall was a moving wall which velocity is the same with the inlet velocity of 1 (0.2 m/s upward) and the remaining is a stationary real wall. the result was more make sense, at least qualitatively correct at the orifices where it gave more velocity than other points. but the problem is: face 1 and 6 is in fact a wall, not openings. setting 6 as an opening will result an eliminated rebound effect due to the collision between the fluid and the wall. just consider a shock absorber, there's no atmospheric opening. so, what should I do to make this simulation realistic (not only just seems realistic, but also show the real phenomena indeed). what should the Boundary Conditions of the faces? can somebody help me with this problem? you can use the numbers in the picture above to designate the face in your explanations. Thanks for viewing this thread. I will be glad and highly appreciate if you post some reply. well, once again, thanks folks.
FloEFD does not solve this problem as you explained it is very difficult to describe how to model it

 tam85 May 6, 2010 22:52

simulating shock absorber requires trial and error procedure. you need a lot of tme and money to be able to do this. fidap is a best tool for the design of the shock absorber to be able to test the fluid interaction and to simulate what happens inside the shock absorber.

 alainF May 7, 2010 08:36

The problem you have for this simulation with flowsimulation/efd is that it doesn't take into account solid movement : If youput a velocity on a wall it treat it as a tangential velocity without deplacement.

So the pressure remain constant everywhere and the velocity remains equal to zero.

A way to handle this phenomena could be to make a simulation with a software which will be able to take into account wall movement thank to a mesh deformation algorithme. There is a lot of (fluent, cfx, adina, starCD,...) but the task can be quite complex.

Another way can be to have a simplified approach. Since the deplacement of the piston is at low velocity, you can approximate the pressure inside each chamber as adiabatique comprexion/dilatation. Then you will know the pressure delta between the two chamber, The flow rate can be calculated since the pressure drop in the holes which equal to pressure difference between the chamber.

You can make a numerical solving of these equation by a quite simple explicite temporal integration.

The only thing that you may need to determinate by 3D simulation would be the pressure drop through the hole versus the gas flowrate. of course you can do this with flowsimulation/efd

 All times are GMT -4. The time now is 16:44.