Changing Compressor Arrangement from Parallel to Series

[Siciliano376]

Hello Folks,

I am currently looking at the following plant design Issue:
I have been tasked to optimize buster compression design for a gas treating plant (300 MSCFD of wet gas), during the lifetime of the facility the pressure will drop from 1500 psia to 750 psia(at year 8) and then further down to 340 psia (at year 16). At year 16 the flow-rate will also drop by 50%.
In order to save CAPEX, I would like to install at year 8 2 x 5O% centrifugal compressors at year 8, these to compressors would be working in parallel to cater for the full flow, as the pressure and flow-rate further decrease at year 16 (150mSCFD at 340 psia), I would like to re-arrange the compressors from parallel to series. This would enable me to accommodate the minimum changing conditions with minimal additional CAPEX (Some re-piping, additional scrubbers)

Can this be done?
I am guessing the internals of the compressors would need to be changed, can someone direct me to a technical resource so that I could get better technical understanding of work at hand?
Has someone done this in the past? What are the issues most commonly associated with this approach?

Thanks,

Corrado

 

[BigInch]

Compressors don't lend themselves to adapting to radical changes in flow conditions.
The future units might be radically smaller too, so much so that the internals of the existing units could not be adapted. They are highly dependent on inlet pressure conditions.
Power is dependent on pressure ratio, discharge pressure will drop a lot.
Existing power drivers will be significantly oversized if left unchanged. Driver efficiencies may be severely compromised and VFD will not be efficient at 50% load.

Space and downtime could be the main issues as,
You might consider leaving extra space for new, separate, smaller compressor units, perhaps adjacent to the full size units.
Locate some blind flanges on the inlet and outlet headers to attach to later.
You could keep the full size units in service while you build out the smaller units and have them ready to go by the time you reach the new flow conditions.

 

[zdas04]

Another point is that you've selected the equipment with the absolute smallest and least forgiving performance envelope of any technology. If you get outside the choke line or the surge line the dang things just stop working (sometimes the "just" is kind of violent, but you get what I mean, they don't squash gas). If you were talking about single stage recips or flooded screws (PD machines) then you might just take an efficiency hit and get on with your life. With dynamic equipment they stop working outside the performance envelope.

[bold]David Simpson, PE[/bold]
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. Galileo Galilei, Italian Physicist

 

[georgeverghese]

Retrofits of this kind are popular on compression trains. You may be right that in year 16 the compressor bundles may need rewheeling to cater for higher compression head at lower flow. Put these operating cases down on datasheets and send them out to compressor vendors to see if they can do this for you.

You'll need to include sufficient fouling margin on the GT driver for each of these machines to ensure that they deliver max required power well past year 16.

Once you have these curves from the vendors, carefully work out design conditions, system setup, cooler duties and location of recycle and antisurge lines for each phase of operation. Also check that all system components can turndown on capacity / duty as required for each phase.

 

[EmmanuelTop]

You intend to use one of the 2x50% compressors from Year 8 (compressing total 0.3 MMSCFD from 750 to 1,500 psia) into a booster machine compressing 50% of flow (0.15 MMSCFD) from 340 to 750 psia in Year 16. The compressor will process the same mass flow of gas in both years (equivalent to 0.15 MMSCFD), but different actual volumes.

The flows are quite small, resulting in compression power around 10 kW (Year 8) and 6 kW (Year 16) - this is like a small size pump. Driver size is governed by Year 8, max volumetric flow and compressor head by Year 16 when the compressor is moved to booster service. Suction flow and compressor head variations are within 15-20% of the design case, so re-wheeling would likely be possible.

As others said above, there is a lot of potential drawbacks if you intend to proceed in this direction and all this should definitely be verified by the equipment supplier. Given the size of the machines/drivers, you might be better off with simply buying a dedicated booster machine and keeping the 2x50% compressors (that will turn into 2x100% configuration) as is. If the compressors were huge and additional cost of a new booster would be significant then yes, perhaps - it would would be worth investigating the reconfiguration/re-wheeling options. With pocket size compressors, this seems rather unnecessary.



Dejan IVANOVIC
Process Engineer, MSChE

 

[LittleInch]

It's not clear to me what exactly your varying inlet and outlet conditions are that you're trying to achieve in terms of flow and pressure, so that's the first thing to work out / explain, i.e. what happens between years 1-8? What pressure do you want at y8 when your inlet is 750?

I think you need to plot the proposed flows and pressure for every year to see where any particular set-up runs into difficulty. It may not be the one you initially think.

Then it's time to send those varying demands to the vendors and see what they suggest.

It might be that the way to go is a multi-stage compressor with initially some of the stages removed or not fitted. Then as conditions change beyond the range of the current set-up, the machine is re configured by adding stages or different stages.

Also be aware of changing inlet conditions on the seals. A unit designed for a low inlet pressure which is then fed a much higher one in a series configuration won't like it.

Working in series you might also need an intercooler to avoid high temp on the discharge which it wasn't originally designed for.

Trying to design for a wide range of inlet pressure and flows isn't easy and there will be compromises to make on efficiency, sizing and cost of extra stages or re-wheeling. As Dejan says, might be better in the long run to simply buy a different machine at certain stages and fit it into the existing pipework rather than change things all the time.



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[MortenA]

@OP,
You have to remember that compressors dont work in SCF but in actual cubic ft, so when your inlet pressure is half, then roughly your actual volume flow doubles.

I read your suggestion to be that in year 8 your will need to boost the pressure up to 1500 psi for the full flow by installing 2x50% compressors.

Then once your inlet pressure drops again in year 16 your would like to double the head again - but now for half the std flow. But since your pressure is now half the year 8 flow your actual flow rate is almost the same and one of your compressors cant handle that much (acutaul) flow (although the motor could). Maybe it can be re-wheeled - that would require more detailed study. BUT remember that you would only need one more compressor (not 2) in year 16. So in principle you could install an almost identical machine in year 16 and choose to devide the head in two equal chunks. However, in real lige your pressure (and flow) wont drop like that but gradually (when answering this question i get a feeling that its homework) - so your will have a lot of intermediate situations that your system should also be able to handle.

Best regards, Morten

 

[Siciliano376]

Folks,

Thank you for the replies, good discussions. Here are some clarifications: the flow-rate at the facility is around 300 MMSCFD (millions standard cubic feet a day) at year 0 and it will stay constant until year 8.
From year 8 to year 16, the flow-rate will slowly decrease until it hits 150 million standard cubic feet at day, at this point, we would like to change the arrangement of the compressors from 2x50% in parallel to 2x100% installed in series.
At year 8 and at year 16 the Actual Volumetric Flow-rates are practically the same, there is a drop of 5-8% in actual flow-rates.

Thank you,

Corrado

 

[LittleInch]

Ok, but what are your inlet and outlet pressures doing in those periods? No compression up to year 8? then?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[EmmanuelTop]

The flowrates are 1,000 times larger than in the original post, and so are the compressors themselves and all auxiliaries (piping, valves, scrubbers, coolers etc.). Now we are talking about 13,000+ hp driver power for the design case 750 to 1,500 psia and 8,000+ hp driver power for the case 340 to 750 psia at 50% flowrate. This is much more significant size of the (re)work.

It's not really us but the compressor manufacturer you should be talking to, with regards to selection and design of the compressor system - at least at this stage. From my experience, I remember asking for and receiving quotations from OEM's for similar installation (field depletion compression) where similar re-arrangement was considered at a later stage in field life, but never selected as the final option due to all the factors discussed in previous posts.

The compressor for 340 to 750 psia receives higher ACFM at the suction, and needs to develop more head. Again, as mentioned earlier, this appear possible to manage with re-wheeling, given the relatively small differences between the two cases.

My biggest concern here - as mentioned by BigInch, zdas04 and LittleInch - is how much downtime you need to plan for, in order to perform the entire reconfiguration work (and whether that is acceptable or not from production/delivery perspective), and how certain you and reservoir engineers are that the flowrates and pressures in reality will be as indicated in your post (this never happens)? Depending how much the real life goes "off the projected map", the solution with reconfiguration might turn out to be a complete fiasco. This is definitely a solution with so little in-built flexibility, so if I were you I would be looking at performing some sensitivities and see where does that get you and how tolerable it is.

Dejan IVANOVIC
Process Engineer, MSChE

 

[BigInch]

Planning 16 years into the future is a very nice exercise, but rarely comes to any good. There is just too much that can happen in the meantime. It usually isn't much use to make detailed plans for more than 10 years, expecially if production declines are involved. The best option is to leave space for what might happen after 10 years. I've had offshore platforms that were supposed to be for a field life of 20 years pulled up and sold for scrap after 6. Geophysicists are far too optimistic to base very big, long term, plans solely on their predictions. If you have many wells over a number of fields, then maybe there is a better chance. A small number of wells in only one field is a bigger risk of reaching long term production targets. Don't spend too much time on detailed design for that today.

 

[georgeverghese]

If you plan to install waste heat recovery units on these 2 GT drivers, the additional backpressure incurred on the GT exhaust will incur a small derate penalty on hp. Also avoid speed reduction gearboxes if you can between driver and compressor.

Wet gas tells me there may be guey corrosion inhibitor carryover in the gas feed to these machines if you use vane pack type scrubbers.

 

[rotw]

Its unclear from your description what outlet pressure you need to maintain constant, I assumed sort of pipeline pressure ~1500 psi.

Parallel <-> Serial works typically well with production profile decline as you start production with high reservoir pressure, so low head, and high flow (parallel mode). At end of production its high head because reservoir pressure reduced and lower reservoir flow capacity (serial mode). This is very simplistic but should summarize the concept.

What you have is incompatible with the scenario I just described because your low head case is high flow at standard conditions but its low flow at actual suction conditions. So your (pressure, flow) chronological path and the path I described are just mirror one another. Again you need to specify your cases more clearly.

Switching parallel to serial should not require stopping the equipment. Its a controlled case with changeover valves doing the job. That's why you need both modes envelop to overlap (preferably).