Stress Analysis of Vibrating Pipework

[CSGWB]

I am currently reviewing a process pipe system used to blow a slurry from the bottom of one vessel to the top of another. The system historically has had a problem with vibration primarily caused by the density variations in the slurry caused by the motivating fluid - steam.

Does anyone have an idea of how this can be modelled in AutoPIPE (or similar) to ensure that the line is not overtsressed and has a reasonable design life?

I have considered applying a force acting perpendicular to the pipe at the mod point of the bend, however I am unsure of what magnitude this should be. Is this the correct approach?

Any assistance is appreciated.

Regards

 

[LSThill]

Possibility is using BOS FLUID for more information on the videos clip go to boss_steam_hammer

http://www.paulin.com/

(download), (left side VIDEOS)

 

[StressGuy]

Your best bet is going to be to get field measurements of the magnitude of displacement at a few points of the line as it is shaking.

From there, you can check the stress value due to the displacement and see how many cycles that stress level is good for. You want the displacement to be small enough so that your cyclic stresses are low enough to allow "infinite" life.

Quite frankly, I can't tell you the details of how to make the check. Other guys in my department are fluent in the specifics and I've not had the opportunity yet work such a problem. I would recommend posting this to the Caesar II forum over at the COADE website to get some pointers to references to consult. Messrs Breen, Luf, Becht, and others can no doubt point you in the right direction.

Now, all that said, if you can clamp it down to damp out the vibrations w/o imposing excessive thermal stresses, that will be the way to go.

Edward L. Klein
Pipe Stress Engineer
Houston, Texas

"All the world is a Spring"

All opinions expressed here are my own and not my company's.

 

[stanier]

In Algor's pipepak you can apply a time history load to nodes. In order to determine the load you would need to do a transient analysis or some other calculation.

Initially the software undertakes a modal analysis to determine the natural frequency of the pipe for different frequencies. The time history load is applied and the dynamic response of the pipe is reported graphically or by dynamic image.

The problem comes about in the couplingof the fluid model to the dynamic structual model. There are quite few papers covering fluid structure interaction and the debate that still rages on how to do this accurately.

Look for papers by Wachel on vibration analysis and measurement

http://www.engdyn.com/papers/papers_table.htm