I have a hydraulic oil system that is subject to pulse (surge) pressures on failure of a upstream valve. To elevate the problem I use an oil damper that allows 40mm movement in all directions. The damper only comes into play during the shock load and is none operational during normal system operations.
Does anybody use similar products and how could this be simulated in a static analysis.
Th damper is suitable for a max load of 1.5kn and dissipates 40% of the load

"The damper only comes into play during the shock load and is none operational during normal system operations."

Why will there be any need to simulated damper in a static analysis ? Just ignore it.

In Dec'2005 Hyd Processing, you will find an article by S.Saha on this topic.

regards,
sam

Hello,

Pulse dampers are often used in jet fuel piping systems that are (mostly) underground because there is a need for rapid fueling and very abrupt shut-off (especially if the fueller drops the fuel hose while loading fuel into an aircraft). The abrupt shut-off can cause a serious pressure "spike". These are (as Sam points out) of little consequence in the performance of a structural analysis of the piping system. However, early underground fueling systems of this type were subject to failure. Obviously, as many of these systems were placed under many inches of reinforced concrete, their repair was quite expensive. Investigation determined that the failed systems were NPS 6, 8 10 through NPS 16 schedule 10 with longitudinal seam welds. Looking at the cross section of the pipes they were not exactly round, rather (if you will allow a little exaggeration) they were "heart shaped" with the seam weld at the apex of the indentation. When the pressure spike occurred, the pipe "tried to go round" and when the pressure dropped to normal, the pipe assumed its "as manufactured" cross sectional shape. Over some number of such "spikes" the pipe failed due to fatigue at the interface of the unaffected material and the heat affected material.

Nothing here to model in your structural analysis model but certainly something to be aware of if you are designing piping systems that are subject to pressure pulsation or "spikes"

Regards, John.

Sir,

In Gazman's case I thought the failure of upstream valve an upset event of occasional loading of primary stress catagory. If such event occurs recurrently, no doubt fatigue failure may occur, as mentioned by you, due to localized peak stress. One need not consider vibration in static analysis, but neither can he/she ignore the effect of any serious vibration phenomenon on fatigue life of piping & its supporting attachments & may require further dynamic analysis.

But, isn’t it proper to control the cause of pressure surge in source with some kind of relaxation chamber or pulsation bottle in case of the jet fuel piping you mentioned? Other way out could be to make piping thick & strong enough to withstand such fatigue loading resulting from pressure surge.

I remember a main steam steam hammer resulting from 25 mS trip valve closing time created such a situation that we could not design the structural arrangements to design civil structures for rigid struts & mechanical snubbers; we had a trade-off with steam turbine vendor & arrived at 80 mS closing time for a feasible civil/structural design for such shock event. In the first year of operation, several such full load trips occurred; had we not arrived at such a compromise, we could have damaged the plant otherwise.


regards,
sam

PS: When will we have facility to model viscodamper in Caesar-II dynamic analysis? It is risky to wait for plant to coomission & backfit the same with damper; we need to have to consider it in design phase in the piping analysis.


Till then, is there any better way in Caesar-II environment than considering it as like snubber with linear spring stiffness applicable for the dominant frequency we wish to counter with viscodamper and then consult the vendor?

I don't believe you can correctly model a damper in CAESAR II at the present time. A damper and a snubber are totally different animals. Dampers are like sticks, attached to the pipe, that sit in a tub of goo -- so the pipe squishes around while it vibrates.

Our understanding is:

1. The damper should have no effect on the static model.
2. The damper should have no impact on the mode shapes or frequencies.
3. The damper's impact on modal analysis cannot be determined.
4. As far as we can figure, dampers can only be simulated by multiplying the nodal velocity times the damping value (which, by the way is frequency dependent, so the damper has a different value for every mode) to get a time-varying force which then should be added to the force profile.

So how would somebody model this? They would have to estimate its impact as a time dependent force, and include that force profile in the Time History (not an easy matter).

This is a subject we are studying for possible inclusion in the software.

Sir,

I accept that snubbers & dampers are different, but have similarities like inaction against static or quasi-static loading like sustained or thermal expansion & both are used in mitigating the effect of shock loading. Dampers can be effectively used in combatting vibrations also.
Snubber assume zero value in static load & very high linear stiffness ( both expansion & compressive) at high frequency and damper changes its linear stiffness from zero to higher value with the increase in frequency from zero value.

I have objections to your point 2 & 3:

Damped natural frequency of a linear spring-mass system is

wd = wn * sqrt(1 - d^2) where d = damping ratio for d < 1, underdamped system.

Both the natural frequencies & mode shapes differ with the change of damping ratio of a structural system - just check in modal analysis in Caesar-II itself - use 2%, 3%, 5%,10%, 50% & 75% of critical damping. Following web link explains this well.

http://www.engin.brown.edu/courses/en4/notes/Dampedvibes/Dampedvibes.html

We, as users of Caesar-II, feel assured that you are studying the subject for possible inclusion in the software.

regards,

sam

Sam,

Mode shapes and natural frequencies are computed by the well known eigen value problem ( without damping)

[K][X]=[M]W^2[X].

Using damping the eigenvalue problem becomes a quadratic eigenvalue problem.
I don't think CAESAR II or any commercial software solves this eigenproblem.

Regards

Anindya,

You are right. But, in time history analysis, we can use % damping within Caesar- II.

regards,

sam

Sam,

Sure. We do that in Time History analysis. but the eigenproblem solution is still w/o damping.

Richard Ay,Please correct me if I am wrong.

Regards

Anindya is correct. The eigen solution is without damping.

Hi and thank-you to everybody who has responded to this thread.
I was concerned that this was an isolated problem and am pleased to find out that there is no real way to model the damper effect.
We can use our engineering experience to explain to the customer that the damper will work in the one off failure event and can not be simulated by current software.

If one refer Picture-6 of the GERB's article here,
http://www.gerb.com/en/bibliothek/downloads/dokumente/PipeworkDamperSept02.pdf

One can simulate damper by a linear spring just like snubber with this equivalent stiffness in Caesar-II.

So, instead of accepting vibrations in piping, we, too, in process piping field should use vibration dampers like our counterparts in power piping - isn't it ?

regards,
sam