Centrifugal Compressor --- Suction and interstage flange rating

[keeplrng]

For one of the projects recently for Centrifugal compressor, vendor has proposed that suction and interstage nozzles of compressor shall be suitable for stage discharge pressure or settleout pressure which ever is higher. Vendor had taken exception to API 617 requirement of clause 4.6.4.1.1 which is approved by Consultant.

In this case there are 2 stages, so process connection nozzles are 150#/150#/150# and final discharge is 300#.
This is a fixed speed machine.
There are 3 compressors working in parallel. However the compressors are not discharging to common header directly. Downstream of compressors there are scrubbers and heat exchanger and other equipment.

1) As this is fixed speed machine can there be any scenario when the casing can see the maximum discharge pressure and will there be any concerns for mechanical integrity failure.

2) Why does API 617 require all nozzles to be suitable for MAWP. ? Is it considering the typical arrangement where centrifugal compressors are with variable speed arrangement so the discharge pressure can increase beyond rated discharge pressure? What would be premise of API to specify all nozzles to be suitable for MAWP?

3) Can the machine interstage nozzles be subjected to maximum discharge pressure incase of surge? (For this query vendor mentioned this is not possible).

Views of members are welcome.

 

[1503-44]

In certain cases, during recycling, or a possible recycling valve failure, suction flange may be subject to final stage discharge pressure.

 

[keeplrng]

Thanks for response. We had posed this question to vendor about suction side getting exposed to discharge pressure however vendor clarified that there is antisurge system provided. The cases you mentioned only seems possible scenario and i suppose probably during recycling the suction flange will be subject to final discharge pressure.

Failure of antisurge control and the blockage of discharge cannot be scenario for HAZOP as it will be double jeopardy situation.

 

[1503-44]

Pipeline compressors, or any group of compressors in series, may also experience high inlet pressures on shutdown, especially when upstream units do not react quickly.

1) Beware of the "double jeopardy" excuses for not making your design sufficiently robust. Its proven to be rubbish. Accident reports have observed time and time again that the most severe accidents almost always occur as the result of at least two or more failures, making an even greater improbability of the disasterous event. Why? Because the designers were sure that something of that magnitude of improbability could not possibly happen and were thus caught totally unprepared for the consequences. Fat tail events are more common than their simple mathematics suggests.
2) Most provisions of APIs are there because of a long history of "need to be there". Ignore them at your own risk.
3) So, you're spending millions? buying compressors and you cant afford a couple of high pressure flanges and you want to hit the vendor up to change the API recommended configuration? What's wrong there?

 

[keeplrng]

When something is specified in API absolutely there will be some past experiences. Thanks for inputs. We have informed vendor to provide the flanges of same rating. Whoever I talked to had same question what are you saving with flanges with reduced rating some dollars against a million dollar compressor package.

 

[1503-44]

Great. It potentially saves you that low setting on a recycle PSV and vent piping, so in that case strong flanges would definitely be cheaper.