How to make Adsorption Dehydration Performane Monitoring

[nong]

My plants have sorbead(Alumino-silica gel) to desorb water from nat. gas stream from 45 to 0.5 lb/MMscfd (gas flowrate 127 MMscfd) and we wonder that we can extend life for 1-2 years if we can find that the Mass Transfer Zone is enough. Some methods are using Gamma ray to determine bed density or using neutron backscatter (which I don't know how they work). Any suggestion 'lb be very appreciated.

Thank you

 

[TD2K]

I've no experience with those methods. You don't say precisely what your design is like or how you operate so the following may or may not apply.

We had a lead/guard bed of mol sieve in an ethylene plant and we monitored the time in lead, gas flow, dP and time until breakthrough. This was plotted, on a normalized flow basis, to give us an idea what the deterioration in bed's performance was like. One difference here is that it doesn't sound like you have a guard bed. Without that, running to breakthrough versus a fixed cycle does entail a risk to the plant.

Also, we had a fixed inlet temperature and so we didn't bother correcting for inlet water content. If your inlet temperature varies, that can increase the water load to the bed significantly.

 

Sir,
Please elaborate your dehydration plant working,i.e intermitent operation with regeneration cycle or contineous operation ,procedure of regeneration (if any), change of any physical parameter which is being monitored,any equipment or process downstream of dehydration unit for further treatment etc.Do you use cryogenic process downstream of this unit.

 

[ggordonstewart]

I assume that you are switching towers on a fixed time cycle. If you monitor the water content of the effluent gas at the end of the cycle, you will see when the maximum allowable water conent is approached. At this time, you should cook your beds. Normally, you only heat your beds to 400 to 450 F during regeneration. In order to cook your beds, reduce your throughput to the regeneration flow rate. Take control of the heating/cooling and tower switching and manually heat the regen tower until you raise the outlet temperature to 500 F. This will drive off some of the compunds that were reducing the efficiency of your bed. Continue this procedure until you have cooked all your beds. Now you can increase your flow rate to normal and return to the original cycle. You should see a marked improvement in performance. You can continue to rejuvenate your beds until the interval becomes too short and the lost income incurred as you cook the beds justifies changing out the bed. The guard bed mentioned is to prevent any free water in the inlet gas from shattering the desiccant at the top of the bed. Another reason for bed change-out is that the pressure drop across the bed has become too high due to the desiccant breaking down physically. G. Gordon Stewart, P.Eng.
Gas & Oil Process Engineering Consultant

http://www.ggordonstewart.com/

ggstewar@telusplanet.net

 

[TD2K]

The guard bed was in my post. We had a 3 bed system, 2 in adsorption in series, 1 in regeneration/standby. When the lead adsorption bed reached breakthrough, it was taken off line for regeneration. The previous guard bed was put into lead position and the previous regenerated/standby bed was put into guard position and the cycle continued.

There was no 'guard bed' to catch carryover or slugs of liquid from the upstream equipment (though there were separators of course).

 

[amitava]

Is your adsorbent bed crumbling to finer powder due to regular use? This will lead to pressure drop increase and chanelling of gas flow. If it is possible you should reduce the flowrate of gas such at lower velocity of gas through the bed there is less attrition and hence less crumbling.But this will reduce the plant capacity.

You may ask the supplier of the adsorbent material about the characteristics and recommended operating parameters like maximum velocity, maximum allowable depth of bed, etc. and compare against your actual operating parameters. You will then know if you are violating any of the norms set by the supplier of the adsorbent bed. You can then take corrrective action. If you find that your actual operating parameters are well within the specified norms, then the supplier has to explain why his material is giving short life.

 

[Samiran]

We have Drier beds handling almost half a billion gas scfd.These remove the water from the nat gas down to 0.1 ppm of water ( dewpoint -95 C). Downstream we have the liquefaction plant producing LNG. Monitoring Drier beds is extemely important. If you have bed probes at various levels then you can monitor the best. This will tell you how far your MTZ has progressed. However,unfortunately these probes sometimes are overlooked during design. Also look for dp increase across each drying cycle. See if the dp rises even during regen cycle. This will tell you whether you have liq carrying over from upstream or fines are being produced. Sometimes regenerating at higher temp can deteriorate performance - it may lead to cracking if any heavy HC are present. Common problems of mole sieve degradation are due to amines or polar solvents coming to the beds and also free water. These can explode the beads and reduce them to powder. If you're generating fines sometimes you'll catch them in the after filters if you have.
You can also do a breakthough test - this allows you to determine the remaining life of the sieves. This needs expertise.
You can do a gamma scan to determine the health of the beds. Sometimes the beds can be fluidized and displaced which will cause channelling and premature breakthrough.
Regeneration profiles can give clue to what is happening. You'll find a plateau period when all the water is driven off.Monitor the time interval. Increase of this interval will tell you you're getting more liquid. Sometimes heat loss thru insulation can be substantial. Normally expect 15-30 C drop in regen gas inlet temp & outlet temp.

 

[nong]

Sorry for short description.
We have 4 dehydrators (ID 76", H 16.5 ft) for 127 MMscfd natural gas from 45 to 0.5 lb/MMscfd. Desiccant is 97%Silica 3%Alumina. Switching valves to operate 20 hrs cycle time (drying 10 and regen 10 hrs) Two beds always in drying mode about 65F, 657 psig. We use the same gas for cooling and heating beds at 500F 566 psig.

How can I do a breakthrough test or monitor mass transfer zone ? Unfortunately we can't shut down this unit and no bed moisture probes. Does gamma ray provide the reliable results because I'm afraid the difference in density profile between MTZ and active zone is not enough to interpret the moisture content in each zone. Do you know Neutron Backscatter ? Does it provide moisture results directly ?

Thank you very much for every comments.

 

[TD2K]

Samiran, I do not agree that free water destroys molecular sieve or activated alumina. Take a sample of molecular sieve or activated alumina and dump water on it. You can dry it out and still have the original beads or extrudates, it does not shatter or turn to mush. We did quite a bit of testing trying to come up with a way to pre-evalaute the long term stability of different mol sieves and this was one test we tried (with no success). Grace's plant in (I think it was in Switzerland) actually would pour water on their product to show to customers touring the plant it would stand up to slugs of water. They'd also pour water on the mol sieve and then boil it in a beaker and then dry it out to demonstrate its durability.

We did have problems with 'some' molecular sieves prematurely failing but concluded it was due to sodium and calcium being leached from the clay binder during regenerations (we regened countercurrently and had a phenomena called the reflux effect occuring) until essentially the mol sieve bead or extrudate fell apart.

That's not true for silica gel I agree, that will shatter if hit with free water.

 

[Samiran]

TD2K- Mole sieve manufacturers like UOP & Grace would tell you free water is absolute killer.You cannot take a risk in a live plant. In fact any polar substance would as they are attracted easily to the active sites. If the mole sieves do not explode the performance would be reduced anyway.Grace has recommended to put a silica gel layer on top of the mole sieve to act as a guard bed sort of. You must have a KO pot with the lowest temperature achievable (without hydrating) upstream of the beds to separate all free water. Breakthru (of HC)can also be possible if you have HC dew point at the inlet. Beware of this - this is very easily ignored.
nong- Gamma scan will tell you whether the beds are lifted or not . It will not tell anything else.I do not know about neutron back scatter. If you get any info please let me know.
Breakthru test can be done online if you've bed probes - or if you've one probe at the end of the bed ( about 1 m above the vessel bottom). It is better to do offline - that is you have to divert the gas to flare downstream.

 

[TD2K]

We have hit our beds with slugs of water in several plants I worked in during upsets. Of course the beds had to be removed for immediate regeneration but I would dispute strongly 'free is an absolute killer' and I never heard those statements from UOP or Grace. Once we regenerated them, they went back on the same performance (time to breakthrough) as before with no change in the slope of that curve for future cycles.

I suspect this is one of those 'agree to disagree' positions.