17 replies to this topic

[ankur2061]

Dear All,

I am working on a project for natural gas compression station. The NG is essentially a lean gas with water content less than 7 lb/MMSCF (pipeline quality). The project document has given a PFD which shows the configuration of the major equipments in the compressor package. However, the given PFD (scheme) in my is not to my satisfaction. I have proposed an alternative scheme for the compressor station arrangement based on my understanding of operational & maintenance flexibility.

I am looking for some opinion in the form of merits & demerits of the attached schemes from the knowledgable members of the forum.

Hope to get some incisive comments.

The two schemes are attached herewith.

Regards,
Ankur.

Attached Files

•  Centrifugal_Comp_Configurations.xls   35KB   408 downloads

[Zauberberg]

Ankur,

Your configuration certainly has benefits in terms of higher availability, but it is more costly as well (3 separators, 3 individual air-cooler banks). The question I would ask to myself if I were in your position is: can we expect any condensation downstream of the compressors, considering worst conditions (gas contaminated with heavy hydrocarbons, low ambient temperature)? If the answer is YES, then your configuration (separator at compressor discharge) is a must, regardless of the additional cost.

As for the individual suction scrubbers, I do not think they are mandatory. One common vessel should be more than sufficient, if it is sized correctly. Three separators add certain complexity in unit control.

Best regards,

[Chem01]

I would go with configuration-1. A suction knockout at each compressor seems promising when the end user of each compressor is a different plant or unit. For the same end user a well designed single suction scrubber can do the purpose.
With same logic a single exchanger at discharge is sufficient and the knockout at the discharge really depends on d/s process requirement. Like at some plants steam is mixed with NG for further reaction and a discharge knockout is of no use.
Your PID's will be more attractive to review, covering lot of details.

[Art Montemayor]

Ankur:

I've just now read your thread and find myself going out the front door in order to spend a weekend entertaining my granddaughter who is spending the summer with us.

I have some comments I can add to this thread that may shed some light on the options you are considering. I have done in the past what you propose: individual separators on each compressor suction. So I believe I have a good idea of what you confront.

I'll try to get you a response next week when I get back - around Monday.

[ankur2061]

Zauberberg/Chem01,

Thanks for your comments. I am also looking forward to what Art has to say about this.

The basic premise for my proposed configuration were the following:

1. Greater availability and flexibility (as Zauberberg points out)

2. More flexible time scheule for planned maintenance and shut-down of a set of suction scrubber - compressor - aircooled HE

3. Uncertain performances of common scrubber and Aircooled HE at high turn-downs

4. Client's inherent lack of knowledege in handling such large compression stations & lack of infrastructure for timely and planned maintenance of such sophisticated equipment.

5. Unreliability of provided gas composition and P & T conditions of the incoming gas to the station.

The only over-riding factor in the client's proposed scheme is first cost. Today's engineering demands the concept of "Life Cycle Cost" for the entire life (25-30 years) of the project.

The above factors led to my proposing an alternate scheme from what the client has proposed.

Hope to hear from others too.

Regards,
Ankur.

[ankur2061]

Art,

I am awaiting your comments on my post as`promised by you.

Regards,
Ankur.

[Art Montemayor]

Ankur:

I have worked with this type of natural gas compressor station process and I opted for the same type of individual suction separator design on each centrifugal compressor. My compressors were multi-stage, but of a similar capacity and the idea is still basically the same. My main concerns were, in some respects, much like yours. Our process thinking and experience is obviously very similar. The additional needs and concerns I employed to justify the individual separators were:

• My anti-surge control valve was also designed to furnish some capacity control by recycling discharge gas to suction. In order to ensure smooth operation, I have always routed the recycle back to the suction separator and never to the suction piping. In order to ensure specific anti-surge control to each compressor it was essential to use individual suction separators.
• If you consider the anti-surge flow requirement(s) you will note that you can’t furnish this requirement to each individual machine if you employ one common suction drum.
• Shut down and startup requirements for each machine must be designed on their own merits and needs. That means that any machine could startup or go down while others are operating. Each machine requires proper design and installation to handle the settle-out pressure. The use of an individual separator makes this need a benign one since each machine can continue to operate on total recycle and shutdown or start up on minimum load. This is an important operational feature.
• The flexibility issue cannot be stressed enough. The additional cost of the suction separators are “chump change” when compared to the ease and flexibility of operation that the production crew has. And this affects the maintenance people as well, since they can shutdown and startup with ease and with total flexibility.

I hope these points that I have retained in memory are of some service to you. Keep up the progressive and forward-looking design mentality.

[Zauberberg]

Art,

Your comments are very much interesting and I'm sure you have witnessed all the advantages that Configuration #1 offers against the other one. I would appreciate if you can elaborate little bit more on these issues:

1. Why do we need suction drum employed in antisurge service? I remember seeing dozens of parallel compressors (or 3x50% arrangements) taking suction from a common suction drum, and having antisurge system (recycle stream tied into the suction line) working perfectly.

2. The same applies for startup/shutdown. Depending on the position of isolation valves, startup/shutdown can be designed as 100% trouble-free without individual suction drums. I remember this particular configuration from both Africa and Middle East plants.

3. Flexibility/availability certainly remains as an advantage, but this could also be achieved by ensuring option for isolation of individual tube banks in cases of tube rupture, so as not to shutdown the entire system for fin-fan repair. Reliability of drums is almost 100% and only in extreme cases there can be a failure of suction drum.

Thanks in advance,

[Chem01]

I was involved in design/comissioning/startup of 2 NG compressors (reciprocating) having a common source for gas but different end users. They had a single knock out vessel, I can share my experience regarding their operationas.

1. Each compressor had a recycle valve, we kept the recycle valve always slightly open. The valve serves as a pressure controller for compressor discharge. There was a flow control valve d/s of each compressor. Whenever one compressor trips or its capacity is reduced, the suction pressure of other compressor increases, along with corresponding increase in recycle valve opening. Never had a problem with tripping of one compressor, the panel operator must ensure the recycle valve is opening accordingly.

2. Complete isolation valves at suction/discharge were provided for maintenance and relatd jobs whenever required, for both compressors.

3. Never observed water/liquid level at suction K.O drum, however water carryover was always observed at one of the compressors suction. Later adequate catchpots were provided at different location of suction piping and problem avoided.

4. Two compressors were very close to each other even sharing a common platform. This gives us problems as 2-3 times operators inadvertantly closed the valve of one compressor, while it was required to close for other compressor.

5. Compressors were covered by a shed, always populated with pigeons. Pigoen droppings affected the house keeping, never able to get rid of these pigeons.

Regrads,

[chemsac2]

I am with Zauberberg and Chem01.

Individual suction KODs:
I may go for individual suction KODs if compression system gets gas from oil-gas separators requiring frequent wire-mesh clean-up. If that is not the case, I may not go for individual suction KODs. In one of my past projects, we had 3 hydrogen reciprocating compressors in parallel with common suction KOD.

Similar to Chem01's scheme we had catchpots in suction piping and had drains on suction/discharge pulsation dampeners. Had not observed any carry-over to compressor at-least during commissioning.

Individual ACHEs:
I am in favour of common ACHE with isolation facility for each bundle as suggested by Zauberberg. VFD on 50% of the fans would also help.

Regards,

Sachin

[Art Montemayor]

Ankur/Zauberberg/chemo1/chemsac2:

I think the title of this thread is "Natural Gas Centrifugal Compressor Station" and, as such, any material or comments made about, or on, a reciprocating compressor have nothing whatsoever to do with the topic. Each is a distinct and different animal.

Therefore, I can't add or comment following on the comments of chemo1 and chemsac2. Our discussion is going to be like a volley ball game: the ball can go anywhere at anytime. We have to stick to the subject at hand if we are to seriously discuss it - and nothing else. Let's agree on sticking to the subject and the substance that forms the core of the subject here because it is a very important subject.

[Zauberberg]

I like Chem01's comment about pigeons

Interesting comments from all participants, in any case. Few points noted.

[ankur2061]

Art/Zauberberg/Chemsac2/Chem01

Thanks for all of your comments. All of them noted. I would reiterate that considering the circumstances of this particular client a system with very high availability and tolerance for longer duration of maintenance turndown is required. It is also important to note that the client has no prior experience of handling such a big (1500 MSCFD) & sophisticated compression station. As I had earlier mentioned the system does not have to be evaluated just based on first cost (CAPEX). As per my evaluation, the configuration I have proposed provides a better life cycle cost.then with the single suction scrubber & single air cooler. I am also in agreement with Art's observation that with individual suction scrubbers the anti-surge recycle control will be simpler compared to having a common suction scrubber. Additionally by having individual coolers I will not require separate recycle gas coolers in the anti-surge recycle line since I can tap the recycle connection from the outlet of the individual coolers.

Regards,
Ankur.

[djack77494]

To all,
Two rotating equipment specialists that I highly respect have told me to avoid common Suction K.O. Drums (or Scrubbers) for all parallel centrifugal compressor stations. While their explanations were not altogether clear to me, they suggested that a variety of dynamic effects could lead to instabilities in the operation of multiple compressors. Remember that the compressors may (will) not be exactly identical, nor will the suction and discharge piping be identical. Now add a complication such as, "What if a future expansion adds a different type of compressor to the circuit?" Might this be a possibility, or would you rather duck the question? Also, as I look at the Configuration 1 schematic, I see that the anti-surge recycle flow (which is unfortunately not shown) would not be cooled. Could you start the first compressor on total recycle? I'd say "NO". Since your client says they are much more quality sensitive than price sensitive, the thinking process can be concluded. Go with your proposed configuration. I feel quite strongly about this, but there are gaps in my understanding of these systems, as I have stated. So if there are objections to my conclusions, I would be very interested in hearing the reasoning. Thanks and good luck.
Doug

[siddhartha]

Ankur,

First of all let me tender my respects to all of the participants in this discussion. It is only through such forums can we explore, discuss and bounce-off ideas and questions. Everybody of you have made such a positive contribution to the thought process.

In my (over) 40 years of design in the O & G industry, I have learnt one thing, and that is no idea, how silly or alien it might be, should be discounted without evaluation.

The configuration of centrifugal compressors in parallel is one of the least and most understood schemes I have yet to come across. The more I talk and hear about it, the more I find there is something more to take into account.

Your main reason to propose the two schemes was motivated by cost savings. Accepted that cost is the denominator in the fraction called engineering. However my question is saving money at what costs?

To illustrate my point, I would like you to consider the following:

1. What happens if there is a problem with one of the compressors and it trips? You will immediately answer that to prevent gas outgress, we would box-in the system and i.e. settle out. But settle out with what?
Or rather would you box in the total circuit i.e.trip the other parallel compressor as well and lose production maybe not only of the gas, but maybe also the associated oil ($$$$$) with it?.

Suppose we isolate at the rogue compressor's suction and discharge. This may require a higher settle out pressure and with it severe problems of dropout of retrogrades during a rapid blowdown at places where you do not want to dropout. Not only have you not considerably increased the settle-out pressure due to no suction buffer volume, you have increased containment risk.

2. If the Compressor Dynamics requires hot gas bypass, admittedly you will get a quicker response with judiciously placed check valves at suction downstream of this common drum, but you are already increasing the number of valves. Cost is biting back at you.

3. In many parts of the world statutory inspection requirements or ability to isolate a boundary for maintenance as the 'tail' (pardon my expression) wags the 'dog'. Would you not prefer to take one train out in its entirety (say to maintain one specific component) rather than have the entire caboodle taken off altogether. Isolation requirements in Industry are getting more and more stringent through incorporation of double block and bleed valves and removable spools. To avoid enormous cost penalty by incorporating DB&Bs around every maintainable component, there is now a positive move towards isolating a discrete boundary with many more things in place. In such a situation, I am sure the shareholders will not thank you to bring down the entire three compressors to attend to a minor or even a major maintenance function.

4. I completely agree with Art Montemayor and would feel well placed to have the permission to re-quote him:

"The flexibility issue cannot be stressed enough. The additional cost of the suction separators are “chump change” when compared to the ease and flexibility of operation that the production crew has. And this affects the maintenance people as well, since they can shutdown and startup with ease and with total flexibility."

Thank you Art. This is my first post in this ChEResources Forum. I feel privileged to see your writing in this forum.

Ankur, Good luck and do bear in mind MORE IS SOMETIMES LESS!

[ALEXV]

Hello Every One, I had a question about Natural Gas Compressor Package and i taught here is good place to ask, here is my question:

If we want to compress the NG with low volume rates(0.5MSCFD) and this NG is coming from seperators then which of the following is the best option for gas compression station:

1- Having pre-cooler/ pre-knock out drums, compressors(two compressor in parallel/one is back up), intercoolers and after cooler and knock out on one package since the gas flow rate is low.
2- Put the Pre-Cooler and Knock out drum in a separate bid package and the rest on one package?
3- Please introduce me to some of the world known compressor package manufacturers?

Second question:

Does any of you have a sample proposal for these type of projects which includes engineering and management costs?
engineering cost should include the followings:
1- Process Design(data sheet/bid analysis on compressors/ Hydraulic analysis/PFD-P&IDs development,etc)
2- Civil works(foundation and installation) management
3- Mechanical piping
4- Cad works
5- Instrumentation
5- management

Thanks for your time

[ALEXV]

Dear Ankur, Could you please send me that XLS file again? i could n't download that from the web site?

Thanks

Alex V

[ankur2061]

Alex V,

I can download the file. I don't see what the problem is. You need to be logged in for downloading it.

Regards,
Ankur.