Compressor system shutdown - back flow to LP side

[Bill3752]

I am evaluating a couple of centrifugal compressor systems, where an identified scenario is back flow of gas to the lower stages of a three stage system. Specifically, on unexpected shutdown the high pressure stage will be "closed in", and gas will flow back to the low pressure first stage - for this analysis we are assuming that back flow preventing devices are not operational.
I am looking for guidance on how I might evaluate the velocity head loss for back flowing through the compressor - i.e. the compressor stage turning backwards.
Has any encoountered a similar situation? Any thoughts on doing the dP calcs back through the compressor? Thanks.

 

[don1980]

That question is far too open-ended to be intelligently answered. Fortunately for you, I don't think you need the answer to that question. Instead, I think all you need is the setting out pressure, after the compressor has tripped.

 

[Bill3752]

Don, I had already estimated the settling out pressure, which is greater than the upstream (lower MAWP) side. So that now I know that I can exceed the MAWP I am trying to estimate the flow rate, hence the need for the resistance to flow back through the compressor. Thanks.

 

[maddocks]

This is almost an impossible calculation given the bizarre torturous path inside a compressor. You certainly don't want to rely on the machine spinning backwards, perhaps you can examine a double check valve scenario?

 

[LittleInch]

This is why on shutdown you need to have a blowdown valve on the HP side and a bypass valve to avoid surge, so all the gas flows in the direction it wants to go.

Actually looking at this post again, if you have a closed in blocked flow, then your compressor will instantly go into surge surely - this is the equivalent of a heart attack for a compressor, normally resolved by having a surge bypass valve recircualting gas back to the suction. If you get surge on a compressor it can wreck itself very quiclkly and expensively. This should then allow orderly shutdown of your system. Search compressor surge on this forum or google and you'll find a lot of info.

Your scenario seems to have allowed the compressors to stop spinning and start going backwards. Vendors normally won't allow for this.

At no point can you allow high pressure gas to leak back into a lower pressure rated system. This is just plain wrong. What you seem to need (IMHO) is guidance on how to prevent a shutdown from turning into a disaster. This is the wrong forum to discuss these things. Try re-posting on "Gas compression engineering" as it doesn't have a lot to do with Safety Relief valve eningeering.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Bill3752]

I agree with all of the above regarding keeping system from seeing backflow. Client wishes to not use "two check valve" API 521. Guess no one has tried to determine velocity head for backflow through compressor. Thanks for the responses.

 

[LittleInch]

Tell client he's wrong in writing and you take no responsibility for design in that instance. Often has a significant change of heart by client...

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way