Stress analysis of a piping system in ESD event of Centrifugal Compressor 1

[JustSomeRoark]

Gents,

I am wondering if based on your experience the ESD trip of a centrifugal compressor needs to be investigated by the Stress analysis engineers.

I am a Process Engineer who is working on the dynamic analysis of a compressor. One of the scenarios we are studying is the ESD trip of the machine. In this event, the antisurge (i.e. recycle) valves open to equalize the suction and the discharge pressure of the compressor. Since the valves must open fully very quickly (i.e. less than 2 seconds) and they are quite large, some really high peak values of rhov2 are observed (i.e. between 150000 and 200000 Pa), but only for a very short time (i.e. 2-3 seconds). After that, the suction and discharge pressure equalizes and the flow reduces drastically.

Based on your experience, is this scenario normally studied from a stress analysis point of view? I have seen countless compressor dynamic studies which never looked at this issue, but in my new company they seem to be quite concerned about it, and I am not sure whether this is a bit over the top or if those very short, transient scenarios may actually pose a risk to the piping system.

In case this question is a bit out of topic in this forum, I would appreciate some help to find a more suitable forum.

Regards.

 

[1503-44]

200,000 Pa = 29 psi. Enclose it in a Coca Cola bottle blast protector.
Did you miss some zeros there?


Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[Snickster]

I have done significant amount of stress analysis using Caesar mostly. I have never came across this situation but there is valid reason to check the effects of the transient forces on the piping in a situation such as this. This scenario will be very similar to a pipeline blowdown valve opening or a relief valve dicharge which transient pressure and momentum forces are accounted for in the pipe support design to resist these forces. This assumes that the recycle valve will open just about instantaneously so there is a true imbalance of potentially large forces produced. If it slowly is openened then the forces would likely not be significant.

Considering a quick valve opening:

At the instant of opening of the valve: On the upstream side of the valve there will be a pressure force acting on the upstream elbow but no pressure force on the downstream elbow so the load on the pipe at this instant will be P x Area of pipe.

As flow is established there will be transient forces in the downstream piping. As the flow reaches each downstream elbow there will be a momentum force since the flow will reach the upstream elbow before it reaches the downstream elbow causing an imbalance of momentum forces. Also if the flow is such that sonic transient flow is developed then the sonic wave front of higher pressure will also travel down the pipe and just like the momentum force create an imbalance of force on the upstream elbow.

Finally there could be likely choked sonic flow once the flow reaches the end of a smaller recycle line at the connection to a larger suction line until the pressure decreases to a certain lower value and equalizes. In this case the force on the upstream elbow will be both a momentum and imbalancved pressure force same as calculated at the discharge pipe of a relief valve discharge flowing at sonic velocity.

I have done similar transient analysis of all of the cases above just not on a compressor recycle line but on relief valve discharge lines and pipeline blowdown valve stacks to atmosphere.

 

[1503-44]

Never done that myself, don't know anyone else that has either and that includes design of so many gas transmission pipeline compressor stations operating betweenr 500 to 1460 psig that I could not possibly count them all. Nor have I ever seen momentum forces analyzed for large pump stations where momentum of liquid transients is much higher than gas. Nor is this really the same as at a vent where large unbalanced thrust forces are not well contained and restrained by up and downstream pipe.

Please confirm the pressure.

Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[GD2]

We are all aware of the surge effect of quick closing of a valve or pump trips. In this case, we are talking of a quick opening valve.
The pressure increase of 29 psi may be tolerable in terms of pressure deign of the piping system but the hammer effect has to be evaluated on the piping elbows by the action of thrust force.

GDD
Canada

 

[1503-44]

I'd actually think its all relative to what pipe material and wall thickness you are using. If its 0.5mm wall clear plastic PE tubing, maybe you might have a problem there. Hammer effect of gas is usually not significant, even liquefied gas. Just not enough density there to develop any significant forces, at least as far as typical petroleum and gas pipeline designs are affected.
29psi "compressed" air weighs about 0.15 pcf. You hardly get a pressure spike at all when you slam close a gas pipeline valve with 50 times the pressure. More force hits the pipe from the valve's actuator.

Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[Snickster]

My understanding is that RhoV2 is a factor used to size pressure vessel nozzles and does not relate to presure of 29 psig.

 

[1503-44]

Rho x v^2/2 /g = velocity head (fluid's dead stop pressure load), velocity head, or pressure, is added to the to the steady state pressure to get a total spike pressure.

It's not specifically stated if that is total spike pressure, as I assumed, or if it is only the velocity head. Stating only velocity head does not tell us much, as steady state pressure could be anything. In any case, if he's worried about 29 psi and its only velocity head pressure, the system operating pressure would also probably be something minimal too. If this is a typical design, 29 wouldnt be very concerning, except in unusual circumstances and even then, maybe only to a relief valve. Sure I may be wrong, but I did ask for confirmation of that 200,000 Pa (29 psi) pressure.

Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[JustSomeRoark]

Gents:

Apologies for not coming back to you until today. For some reason I never got any notifications and I failed to come back...

Thanks for your replies. Please be aware that my background is Chem Eng so I am not really well versed in stress analysis etc. Yes, the 200,000 Pa I was mentioning is the peak velocity head of the fluid during the ESD event analysed, lasting for 2-3 seconds (even less). I understand some people are worried about this figure because for example Norsok P-001 considers a max rhov2 of 200000 Pa for all flare lines. I kind of understand their rationale, but before I joined my current company I never saw this result analysed in so much detail after receiving the dynamic simulation reports. We normally just analyzed the operating point trajectories, pressure drops in equipment, etc.

Does any of you have any specific experience with such kind of analysis of compression systems ESD dynamic simulation results?

Thanks a lot in advance for your valuable feedback.

 

[Winnower]

Is there some confusion regarding units here? I don't think rhov2 is a pressure, it is a measure of energy density and has units of energy/unit-volume (J/m3) not Pascals.

The Norsok "limit" of rhov2<=200000 is, I think, to limit a piping systems exposure to high turbulence flow which can set up flow induced vibrations in a line leading to fatigue.

In my opinion, if the duration of the high rhov2 event is only for a few seconds then you should be ok, the flow induced vibrations will not have time to excite the piping before the flow eases. Given that the event is an emergency shut down and of only short duration it's reasonable to assume it will not occur often enough to lead to a failure of the piping.

There may be a "kick" from the momentum change similar to the opening of a PSV, but that is a separate issue to the rhov2 issue you started the thread with.