Dehydration of Low-Pressure Acid Gas

[EmmanuelTop]

Extremely rich acid gas from sweetening unit, recovered at 1.5bar abs, has to be recompressed in a multistage compressor system up to 100 bar before being exported outside of the plant battery limits. The stream is saturated with water and, as such, creates highly corrosive environment in high pressure equipment. In order to aleviate this concern, dehydration of acid gas and reduction of water content down to 40mg/Sm3 is required.

According to my knowledge, TEG units are feasible within pressure range of 30-100bar which means that at least 3 stages of acid gas compression would require special and extremely expensive corrosion resistant alloys before dehydration could be performed.

Are there any recognized and proven technologies for dehydration of low-pressure acid gas (more than 50% H2S and CO2), and what are the possible drawbacks of these applications? The best solution would be if gas dehydration can be executed below 3 or 6 bar abs.

Thanks in advance,

http://antwrp.gsfc.nasa.gov/apod/astropix.html

 

[zdas04]

You didn't say what the temperature of this gas is. If it is elevated (like an engine stack) then the amount of water vapor can be staggering at low pressures (for example, at 1.5 bara and 170C it is over 13,000,000 mg/sm^3).

The alternatives to glycol dehydration are usually a solid product like a mole sieve, membrane, or a deliquescent drier.

With a mole sieve you are required to put in enough heat energy to cook the captured water off during the regen cycle. At 15 J/mg you are looking at a huge heat load to cook off 13 kg/sm^3. If you use heated process gas then the flare will be impressive.

Membranes don't require regeneration, but they have a waste stream that can be pretty large and with this gas you'd have to flare it, again a pretty big flare.

Salt driers are consumable and to get rid of that much water would be a huge effort to keep salt in the driers.

Bottom line on these sequestration projects is to accept that the first couple of stages of compression must be done on saturated gas and provide appropriate materials to the piping, vessels, and compressors. Once you get up around 10 bar and drop the temp to around 20C the water content drops to around 1400 mg/sm^3 (the rest was removed as liquid water in the separators) and you can successfully use either glycol, mole sieve, or membranes. Salt driers can work, but it is still a big maintenance issue.

David

 

[ash9144]

One possibility since you have acid gas maybe you have concentrated sulfuric acid. You could contact the acid gas with 98% H2SO4 and it would remove the water. If you are tied in with an acid plant this would be a simple solution.

 

[dcasto]

I caught the word "acid gas compression" That tells me that the stream is a mixture of CO2 and H2S that is saturated and is being exported to sequestration (AGI), if true here goes.

First off, there is no need to dry the gas unless you just find it fun to do.

As the pressure approaches 1000 psi, the acid gas becomes hydroscopic and is undersaturated, therefore it is a noncorrosive mixture. In a paper presented at the GPA in 2007, an operating company spells out their whole design and compliance with NACE mr0172.

Quickly heres what you do. The compression cylinders are all carbon steel with ss304 valves and trim. The suction piping is all ss304. The discharge piping can be CS, but the interstage coolers will need to be SS304. After the last stage, all the piping can be CS with some exceptions that depend on the exact composition.

You can use the Peng-Robison equation of state to model the system and see the amount of water droped out in each stage and the degree of under saturation at the final pressure.

There have been a few people that have dried this type of stream with TEG successfully at the 300 to 600 psig range. The contactor and still were all SS304.

I recommend a good engineering firm be engauged to get the details. Goggle acid gas injection

 

[zdas04]

Dcasto,
That makes sense as long as there is adequate gas/liquid separation at each pressure change. Overrunning an interstage scrubber on a recip can be really unpleasant and the amount of liquid that can condense really is quite large.

David

David Simpson, PE
MuleShoe Engineering

http://www.muleshoe-eng.com

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

The harder I work, the luckier I seem

 

[dcasto]

zdas04, yes very true. We've had problems with foaming in the interstage seperators too giving false readings and shutdowns. You have to find the right lube oils for the cylinders, like 120 wt or even higher. We then cascade the dumps and on the first stage suction, we end up with a pump that takes the liquids back to the amine overhead scrubber which is dumped to the water disposal system.

 

[EmmanuelTop]

dcasto,

I would gladly give you a huge star, but I need more answers from your side. This is certainly very interesting solution.

Gas flow rate is 2.5 MMSm3/day, and by multistage compression water content is reduced from 150,000 mg/m3 (1st stage suction) down to 5,000 mg/m3 (4th stage suction).

1. Why the water content spec (from the client) is set to be 40 mg/m3 if practice shows there should be no concerns? I think they should be very experienced in this subject since they operate more than 50 of such or similar facilities.

2. Why did you specify reciprocating machine for this service?

3. Do you know how can I find the paper from GPA 2007 you were refering to?

David, thank you for your replies.

http://antwrp.gsfc.nasa.gov/apod/astropix.html

 

[zdas04]

Not to speak for DCASTO (he's quite capable of doing that himself), but my experience with client spec's are that they are quite replete with cut-and-paste entries that either never had a sound basis or have outlived their basis.

People are terrified of acid gas. I see that in CBM all the time--the produced water has a bicarbonate buffer that pushes the pH to 8.0 and everyone is afraid of CO2 becoming carbonic acid.

Often when you take the fear and superstition out of a process many of the specs just go away. I see this with old-line gas transportation companies all the time--many of their specs were written by really good engineers in the '50s and now everyone is afraid to revisit them.

David

 

[dcasto]

1. Thanks zdas04, thats a very likely scerio. When I did the conversion of 40 mg/m^3, I got 2.5 pounds/ MMCF. Thats an extremely low number to achieve with TEG and on hot interstage gas without stahl columns or a drizo process. This tells me that the request is possibily unreasonable.

2. You said it your self, it will take at least 3, I'd recommend 4 or actually 3 stages of compression and then a pump. You don't give volume or mass flow rate, if large enough, you might use centrifugal. But I perfeer the higher effiencies of recriprocating compressors.

3. This will lead you to the paper directly, you may have to navigate to order it.

http://www.gasprocessors.com/product.asp?dept_id=7086&sku=P2007.09

 

[EmmanuelTop]

It is interesting that after 3rd stage of compression and removing the condensed liquid at 50C (intercooler outlet), the fluid already reaches cricondenbar pressure and there could be no condensation regardless of the temperature.

However, I need to take extra care because acid gas is going to be mixed with other gas streams provided for injection purposes. The final mixture will contain hydrocarbons, and the condensation of water (due to new fluid composition or, for example, during equipment blowdons - which has not been considered in our discussion) can become an issue again.

Leaving this way of thinking for a moment... Do you have any information related to minimum operating pressure of solid adsorbent units (activated alumina, mol sieves)?
And what is the corresponding pressure drop for these applications, in general?

If I can apply solid adsorption after 1st or 2nd stage compressor discharge, that will put any potential headache far away from me.

http://antwrp.gsfc.nasa.gov/apod/astropix.html

 

[dcasto]

OK, now you have changed up on me. TEG dehydration has been used on acid gas dehydration many times. As you stated, the maximum water is removed at the 3rd stage or 600 to 900 psi range. At this point 75% of the water will have been removed during compression. I get about 9 pounds/MMSCF or 143 mg/m^3.

Using TEG at the 700 psia would not not be a problen in obtaining a 2 lb/mmscf (40mg/m^3). A detailed run should be made.

Mole sieves have not limits on pressure. They do perform better, ie the mass transfer zone gets smaller at higher pressures. The limits are space velocities, so higher pressures mean smaller diameter vessels AND the higher the pressure, the lower the mass of water to be removed, so less mole sieves to install.

I ran a quick model and the first stage discharge gas will have the same amount of water than the first stage suction. So the best place would be to dehydrate after the third stage at around 600 psia.

The assumed pressure drop I always use for the whole mole sieve dehydration system including regeneration is 25 psi. Typically you can do it in about 15 to 20 psi upon detailed design.

 

[Aggie03]

The design cases that I have seen all have any dehydration being performed after the 3rd stage of compression.

As others have mentioned, the first three stages will be dehydrating the gas themselves. If you look at the saturation curve, the water content of the acid gas reaches a minimum around 700-800psia (for a 50-50 or so mix of CO2/H2S) and actually begins to climb afterwards.

This means that after the 3rd stage of compression (usually around 500-600psia) the gas will no longer be saturated.

Most designers dehydrate at this point as it is the least intensive. Doing so earlier in the compression train would not benefit you much as you're removing water that the process itself would already be doing. You would also need a considerably larger amount of dehydration.

See the Maddox paper "Acid Gas Injection: An Operator's Perspective" presented to SOGAT in 2004. I think you can get it off the Gas-Liquids website

 

[maddocks]

Thanks Aggie - I wrote this paper in 2006 for presentation at the SOGAT conference. Dehydration, while not desirable, is sometimes necessary. It is purely a function of composition and final pressure. If the acid gas was, for example, 75% H2S and 25% CO2, no dehydration would be required. Both H2S and CO2 experience a minima water content allowing the fluid to be undersaturated at injection or sequestration pressures. CO2, while it does exhibit this bahaviour, shows it to a much less degree than pure H2S. However, the addition of hydrocarbons greatly changes the water behaviour of the system. I presented a paper detailing this in SOGAT 2007. Dehydration typically takes place after stage 2 or stage 3 depending on an economical and operating analysis.