Two Stage Centrifugal Compressor Design

[Sawsan311]

Dear All,

I would like to kindly inquire on the design of two stage centrifugal compressor (variable speed with suction pressure performance controller) which has each of its stages located in the same casing, driven by the same shaft and having common seal gas system. Each stage has its such cooler/gas-liquid scrubber and they are not segregated by shutdown valve, instead two dissimilar non slam check valves are provided at the discharge of the first stage in order to close fast in case of compressor shutdown preventing reverse over speed through the compressor.

My questions are as below:
1) I believe each stage shall have its standalone anti-surge control loop where the anti-surge line is taken from compressor discharge and connected upstream the suction cooler of each stage. This is in order to cover the process dynamics and reduce the surge event duration in case of the various blocked suction/outlet dynamic scenarios.
2) For designing the emergency depressurization system of the compressor and considering no SDV between the two stages and with standalone ASV for each stage,would you recommend having one blowdown valve common for the two stage compressor and located at the discharge of the 2nd stage compressor, what pressure levels we would consider while carrying out the blowdown out of the below:
a) 2nd stage settle out pressure directly to 7 barg considering the total system volume (1st stage+2nd stage) in 15 minutes
b) two segment depressurization taking the maximum flow out of : Segment 1: 2nd stage settle out pressure to 1st stage settle out pressure (with 2nd stage volume only) then segment 2: from 1st stage settle out pressure to 7 barg (total time shall be 15 minutes, RO size is fixed from segment 1).
3) Can we have one common anti-surge valve for both stages since they are located in the same casing and their overall settle out pressure will be used for setting the design pressure of the suction cooler/scrubber of the 1st stage compressor. I think in this case we may compromising surge protection due to the increase in system volume, delay in loop response time. Additionally, the design pressure of the inter stage scrubbers/cooler will have to be set only based on maximum 1st stage discharge pressure.

Thanks

Regards,

 

[rotw]

1) I believe each stage shall have its standalone anti-surge control loop where the anti-surge line is taken from compressor discharge and connected upstream the suction cooler of each stage. This is in order to cover the process dynamics and reduce the surge event duration in case of the various blocked suction/outlet dynamic scenarios.
Not a matter of "belief". A anti-surge loop common for two stages is also doable. There is just pros and cons. I am not going to go through.

2) For designing the emergency depressurization system of the compressor and considering no SDV between the two stages and with standalone ASV for each stage,would you recommend having one blowdown valve common for the two stage compressor and located at the discharge of the 2nd stage compressor, what pressure levels we would consider while carrying out the blowdown out of the below:
a) 2nd stage settle out pressure directly to 7 barg considering the total system volume (1st stage+2nd stage) in 15 minutes
b) two segment depressurization taking the maximum flow out of : Segment 1: 2nd stage settle out pressure to 1st stage settle out pressure (with 2nd stage volume only) then segment 2: from 1st stage settle out pressure to 7 barg (total time shall be 15 minutes, RO size is fixed from segment 1).
For a back to back machine, the pressure in stages will tend to come to a common settle-out pressure because there is no way the two stages do not leak via the balance piston (amongst other things). It takes a while but ultimately the system tends to a common equilibrium pressure.

3) Can we have one common anti-surge valve for both stages since they are located in the same casing and their overall settle out pressure will be used for setting the design pressure of the suction cooler/scrubber of the 1st stage compressor.
It is not a matter of two stages located in the same casing. Even two sections (or more) in two separate casing can have a combined anti-surge valve.

I think in this case we may compromising surge protection due to the increase in system volume, delay in loop response time. Additionally, the design pressure of the inter stage scrubbers/cooler will have to be set only based on maximum 1st stage discharge pressure.
In short, one anti-surge valve is less costly (the valve, the loop/instrumentation, the piping, the installation costs and the controllers) but will squeeze your operating range because you will have to consider the overall map for two stages, depending on your application (heavy gas) the cut of operating range can be severe. And of course, let alone the lower operating flexibility, this option can translate into energy waste because, due to the effect mentioned above of restrictive operating range, you would have to recycle much more flow if you have turn down / low flow cases. So you would save on capital costs but operating costs may suffer. Generally said, it is appropriate (good practice for machine protection for example) to design the system with an anti-surge control loop per each compression stage.

 

[Sawsan311]

Thank you very much Rotw for your feedback,

I understand that the increase in operating cost you are referring to for the case of higher recycle flow rate is due to the larger duty required for the suction air cooler upstream the compressor's first stage and which will need to be designed for the full recycle operation?

Additionally,I just need to confirm that as long as the compressor stages are driven by the same shaft and have the same seal gas system, they will always have a combined settle out pressure or is that the back-to-back impeller arrangement is what mandates combined settle out?

Thank you very much again,

Regards,

 

[rotw]

I understand that the increase in operating cost you are referring to for the case of higher recycle flow rate is due to the larger duty required for the suction air cooler upstream the compressor's first stage and which will need to be designed for the full recycle operation?
Well, not exactly. Consider two stages 1,2 in serial (your case). If your design is based on a dedicated anti-surge valve per each process stage, then each stage has its own surge limit line. If your design however has a combined anti-surge valve for both stages, the operating envelop becomes restricted/locked by whichever section has more stringent surge limit line. Therefore your operating range could be reduced. In other words, if you ever have to operate a point in recycle (any low flow/ turn down case), you would operate the most stringent section at the control line, which is okay, but then the less stringent section will be operated with much more flow then strictly needed for anti-surge protection. How severe is the energy waste, depends on the aero-design of the stages, molecular weight, etc.

Additionally,I just need to confirm that as long as the compressor stages are driven by the same shaft and have the same seal gas system, they will always have a combined settle out pressure or is that the back-to-back impeller arrangement is what mandates combined settle out?

When machine is stopped, each section will undergo a pressure differential which tends to come to an equilibrium (settle-out-pressure). At the same time, the higher pressure compartment will leak to the lower pressure compartment if there is any means for that to happen. The balance piston at the inter-stage and at the shaft end allows for that leakage. Not only back-to-back arrangement have that. For example, a two-stage casing, with straight-through arrangement is subject to same (these configurations are rares by they do exist). In addition, if you do not have each section isolated (block valves), then you would have a combined settle out pressure occurring for the whole system anyway. As a note, check valves are not considered tight for isolation, if I am correct. So of course you need to check this based on your case / process flow diagram. Side note: Combined settle-out pressure could also increase with time as the buffering of the seals, that is to keep containment and positive flow in seals during stand still, will tend to increase the settle out pressure. Especially, if the stand still is for "relatively" long period with no provisions installed to relief/venting the seal gas that accumulates in the casing.

 

[georgeverghese]

The recycle line is typically taken from the compressor stage discharge cooler outlet and returned to upstream of the same stage suction scrubber ( not to upstream of the suction cooler). Else you may explain why you have the arrangement you described.
A 2 step depressurisation is typically implemented only when the flare system handling capacity is less than what would result with a single step depressurisation, or in cases where peak dP/dt exceeds certain limits set by the shaft seals. A 2 step depressurisation also has attendant process safety concerns, so avoid 2 step depressurisation if you can.

 

[rotw]

Not only the shaft seals but the casing in general, and gaskets o'rings depressurization rate shall be verified. Best is to qualify for explosive depressurization.

 

[rotw]

There is an excellent article here about where all these seals can be found in a centrifugal.

 

[Sawsan311]

Thank you very much Rotw and georgeverghese..

 

[Sawsan311]

Hi Mr.Rotw,

Reference to our earlier discussion in June, I would like to re-confirm your opinion with regards to having combined settle out for the following case:
- We have 3 stages compressor. ALL three stages are driven by the same shaft. 1st stage and 2nd stage are in the same casing, they have SDV in between their external piping systems. vendors confirmed that they have a combined settle out due to the leakage through the labyrinth between the stages within the SAME casing.

The third stage is also driven by the same shaft as the 1st and 2nd stage, but it is located in a separate casing with a SDV as well at its inlet.

Now considering the shaft is the same, there will be a balance piston drum at the Driven end and non-driven ends of the shaft.. knowing that there will be also leakage across that drum. Is the settle out pressure should be considered as combined for all three stages despite having the 3rd stage in separate casing.

In other words, combined settleout is driven by:

1- Common shaft with the presence of the balance drum for thrust balancing at the non-driven end and driven end..
2-Common casing with the reasoning that the leakage will occur through the stages labyrinth.

Thanks

Regards,

 

[rotw]

Is there an isolation valve between 2nd stage and 3rd stage? Is this what you mean by 'SDV'?
If this is the case, the loops are isolated and there are individual settle out pressures.
I suppose dry gas seal are used for shaft sealing.
The anti-surge loop tie in on the suction of 3rd stage needs not to be located upstream the isolation valve.