Process Design - Liquid flow from compressor scrubbers

[DoraeS]

Hi, I have a question about where to route my liquid flow from compressor discharge scrubber:
When I have two stage compressor (equipment configuration will be suction cooler, suction scrubber, 1st stg compressor, discharge cooler, discharge scrubber, 2nd stg compressor). Where should I route my liquid collected from 1st stg discharge scrubber?
In my previous jobs, I have seen liquid be routed from discharge scrubber to upstream separator directly, but I have also seen liquid from discharge scrubber be routed back to suction scrubber.
Does anyone know what is the benefit of routing liquid from discharge scrubber back to suction scrubber? Does this help to recover more vapour as the equilibrium changes?
Thanks.

 

[Montemayor]

llc0612:

I consider this a very important, process question that deserves a complete and analytical response. Many engineers out in the field have seen this and even operated compressors under the conditions described – but yet never challenged or questioned it.

I had to evaluate a major compressor station process design package for a client about 4 years ago. This was a BIG compressor station. The natural gas being handled had a 3% CO2 composition and all 3 compressor trains were draining their condensate removal in the 2nd, 3rd, and 4th stages back to the 1st stage suction scrubber. This was a crash-basis, fast-track project and most of the major equipment was already ordered when I got the process package – especially the piping, scrubbers, and compressor intercoolers. Something about the recycle back to suction smelled funny to me so I queried the client. I was told this was to avoid any wastage of dissolved gases in the high pressure stages once the condensate was drained to a lower pressure. The safety prospect of having a high pressure drain valve fail open into a LP scrubber was one thing; another was the unknown effect (if any) of CO2 concentration in the compressor circuit. I discovered that the simulation set up didn’t include the recycle effect of the condensate expansion into the LP scrubber. When I ran the simulation in the manner it was really proposed to operate, I found that the CO2 concentration really increased to an unacceptable level. The carbon steel construction of the piping, scrubbers, and the intercoolers could only last a matters of months – as predicted by the CO2 corrosion program we ran.

Needless to say, the condensate drains were immediately changed to direct drains to a LP drain header. I think this experience answers or addresses your concerns.

 

[dcasto]

I normally cascade the drains backward, from discharge to second and then to fist stages if the gas has a high ethane and heavier concentration. As the amount of heavier hydrocarbons drop, I may only take them all back to the suction scrubber.

Sorry Montemayor, I've never seen unaccepatble CO2 levels be a concern even in CBM at 11% CO2.

 

[DoraeS]

Hi dcasto, could you tell me more about the reasons behind of putting your liq back from discharge to 2nd and 2nd back to suction scrubber? Is there any benefit on doing this? Why can't I drop down my liq from discharge scrubber directly to upstream system(say 1st stg main separator)?

Hi Montemayor, I agreed with you that during the design phase, one should simulate the compression loop with discharge scrubber liq routed back to suction scrubber, as this will definitely affect the suction scrubber sizing and line sizing. But, are you trying to say that based on your experience, routing liq back to suction is not preffered?

 

[Montemayor]

If you recirculate any CO2 + condensed water back to the lower pressure compressor scrubbers and you simulate this correctly with HYSYS (or any other simulation program) you will arrive at the results that show that you have a CO2-saturated solution (a.k.a. carbonic acid) in the bottom section of your LP scrubbers. Additionally, if your LP scrubbers are fabricated of carbon steel, you will quickly find out how corrosive carbonic acid can get.

I have no problems with what dcasto is doing at present. I do have a LOT of problems creating a saturated CO2 water solution within a carbon steel environment. So does NACE for that matter. That's why I brought out the example I had. Common sense tells us all that if you have a CO2 content with a moist natural gas and you compress this gas with subsequent cooling of the pressurized product, you are actually creating a concentrated "Coca-Cola" in the bottom of the HP scrubber. Accepting the previous as a fact of life, one can then quickly visualize what will happen when one recycles the saturated condensate back to a lower pressure scrubber. The obvious is that the dissolved CO2 will flash into the lower pressure gas stream and continue to do so as long as compressions system hasn't reached equilibrium. When the system reaches equilibrium, the final %CO2 will be higher in the recycled loop - and it will be the cause for a maximum of dissolved CO2 in any subsequent water condensate formed.

I think the above is pretty simple to visualize and all that is required is an accurate simulation that models the equilibrium state reached to identify the level of CO2 strength in the water condensate formed.

 

[dcasto]

the reason for a cascade system is to 1) save energy in the recycling of vapours and 2) limit the pressure drop which also limits the temperature drop with associated headaches of freezing.

 

[Montemayor]

dcasto:

I know that. That has always been my incentive in trying to cascade the scrubber draining operations.

However, my point is (& has been) that when you have a moist natural gas with CO2 that is condensing out liquid water at elevated pressures, the creation of carbonic acid is somehow inevitable. Once this corrosive liquid is generated it not only corrodes, but it flashes more CO2 back into the LP suction system of the compressor train.

 

[786392]

In my humble opinion there should be

1)A predictive/foreseeable Gas Chromatography analysis of the gases/vapors mix samples(to be handled)should be run.

2)Considering normal and peak %age(s)loads of any heavier Hydrocarbons, CO2 and water vapors(if any encountered)

3)Thereafter it will be very clear and more formidable design criteria and optimum design may evolve since;

A)Unnecessary recycling is shear waste of energy& additionally the CO2 if recycled along with condensed water heavily corrodes and Assets damages are encountered.

B)If Heavier H-C components are envisaged then local LP draining then a waste of valuable components,possible unsafe conditions to occur.

In this way we may have a better design fulfilling the requirements.
Hope this helps

Best Regards
Qalander(Chem)