FLARING GAS RECYCLE SYSTEM

[jeap]

HI ALL!

We are about to design a flaring gas recycling system for a LPG plant. It will include a low suction pressure compressor attached to the flare header downstream KOD and upstream the flare. Also, a water seal liquid is going to be installed before flare in order to avoid backflow at very low flowrates and to improve suction control for the compressor. Pressure suction control is to be done by means a recycle control valve between suction and discharge of the compressor. Discharge of compressor is expected to be returned to LPG plant.

The system is to be designed for handling up to 60,000 Sm3/day of flaring gas, of course it includes only some continuos vents from different pressurized drain vessels. Emergency flaring for the plant is estimated to be 2,000,000 Sm3/day (design conditions for water seal).

Could anybody share the design/operational experiences with such a system?
Do anybody have some suggestions in water seal design in order to avoid pressure fluctuations during low flaring flowrates?

Please see attached pdf.

Thanks in advance.
jeap

 

[tkdhwjd]

Just a reminder... need to consider cases (other than normal) where the compressor can see gases of different composition. One example would be... N2 during the equipment purging, etc. Compressors are often designed to a specific gas composition and not really forgiving to fluctuations in composition.

 

[JoeWong88]

Please consider to include maximum MW composition as one of the design case so that the motor would not trip on high power demand.

You sketch showed that the water seal drum is part of the recovery system. What is reason behind this ? Water seal may contribute pressure drop to the flare system and potential affect the flaring. If possible, consider to make water seal drum as part of the flare vendor scope of supply.

May consider to use VSD motor if the turndown is too low to minimise power consumption...

Subject to pressure, may consider liquid ring compressor instead of centrifugal.

 

[Flareman]

Jeap
The relevance of the compressor sizing is primarily in its relationship to the normal day to day relief into the flare system. The compressor will have a limited turn-down capability and you want to it to cover most of the day to day service (so you need a historical profile). I'm sure you know this already but it doesn't hurt to repeat it for the forum. When the flows are too low, the compressor recycle is working overtime and the cost effectiveness is much reduced.
The impact of the changing compositions is certainly a factor so you should run some alternate scenarios with the extreme compositions. You will probably find that liquid rings are more forgiving than rotary screw, and rotary more forgiving than recips.
That is possibly also true of design capacity in the lower ranges.
Other composition considerations involve final utilization. Are you recovering gases you can't use or which have expensive retreatment issues? Are you (un-necessarily) choosing a destination which forces you to double or triple stage to get the final pressure.
In your specific case, what are the chances that you will finish up with liquid rather than gas? Do you want liquid?
Look at the associated installation costs (eg: foundations) and maintenance (spares and downtime) when picking a compressor. Again, in my view, liquid ring and rotaries have the edge.

Regarding the water seal, it's there (as you say) to isolate the flare from the flare system but that means the flare is not working concurrently, so you need to put the operational purge into the flare downstream of the water seal. that raises all sorts of other questions about whether the resulting flare condition will be operational with, or without a flame (a whole new discussion).

The water seal design depends in large part on how you plan to operate day to day. If your idea is to keep the flare notionally at minimum service unless there is a significant dump then you may be using a high seal depth to maximize the capacity of the header and make it easier on the compressor. In this case, when the water seal breaks, you may get a significant surge of flow which displaces all (or much of) the water and allows a high flow to the flare. Look at the design back pressure of the flare at maximum flow and pro-rate flow to square root of pressure to work out an approximate magnitude of the surge. It could easily exceed the smokeless capacity of the flare so you need to address the potential duration of that relief. Whether this happens depends on the design of the dip leg, the available reservoir of water and the destination of the water which is displaced by the initial surge. In many vertical vessels, it goes straight out of the vessel into the flare line so you might want to consider a horizontal vessel so that you can recapture and recycle the water. Horizontal vessels also allow you to use larger water reservoir volumes.
Are you hoping that the water seal will permit a "natural" and gradual, seamless transition from no flow into the flare, through a tiny amount, to a major dump with no impact on the rest of the system? If so, you may be disappointed. Although the water seal acts as if it is a pressure relief valve, it is a dynamic system using pressure to physically move water from one place to another. As such it has time and mass dependencies associated with the water, drum and flare system volumes which give it second-order oscillation properties which will dominate at some flow or other. If you want a seamless transition I suggest that you consider a small bypass line around the main water dip leg, fitted with a real pressure control valve. Then you can use it to manage the flows (automaticay or manually) into the flare when you are in the transitional stages. You still have the security of the dip leg in case of a sudden emergency. You can also seal the end of the bypass by dipping it into the water, but you don't need a deep cover because the PRV is doing the work and the relative size of the bypass makes it much less likely to react dynamically to water movements.

You have plenty of work. Good luck.

David

 

[dcasto]

You'll need to either send the recycle back to the flare KO drum or install a seperate suction scrubber on the compressor.

Normally I see multiple compressors on a flare header. This way you get a better turndown ratio and flexibility for operations and maintenance.

The impact of gas density on positive displacement compressors is almost nil. The biggest concern is condensation of hydrocarbons after the compressor.

If using flooded screws, condensing and absorption of hydrocarbons in the oil must be looked at carefully.

 

[Jason5000]

For flare gas (or off gas) service I'd recommend looking hard at positive displacement compressors, as suggested above. Positive displacement compressors tend to handle a broader range of molecular weight gas better than centrifugal compressors do, and your flare system won't be nearly as predictable or easy to design for as say a coker gas plant or an FCC wet gas machine. Mostly I've seen both dry and flooded screw, liquid ring, and sliding vane type compressors in this sort of highly variable service.

Good luck.