12 replies to this topic

[andiez]

Hi all,

Im currenly involve in a project which requires the conditioning of offshore gas and fiscally metering it according to the agreed GSAs. I ve made a simulation for this project. Basically this process involves the following eqiupments:-

1. An inlet gas scrubber to remove any entrained liquids in the feed gas.
2. A gas/gas heat exchanger to pre-cool the inlet stream and to superheat the outlet stream with LDHI injected.
3. A Joule-Thomson (J-T) valve to cool the gas by lowering its pressure.
4. Finally a SMSM low temperature separator which separates the liquids (heavier HC and water) condensing out as a result of the pre-cooling in the heat exchanger and JT valves. SMSM separator is used as it has high efficiency since prense of even minute quantity degrate the product quality.

In the simulation, the inlet to the LTS separator from the JT valves is adjsted to obtained a dewpoint of 6 °C at the vapor outlet of the LTS separator. This inlet stream has an operating pressure of 33 bara. Below are the gas composition (mol%) entering the LTS separator:-
N2 - 0.38%
C1 - 90.4%
CO2 - 0.68%
C2 - 4.1%
C3 - 2.6%
C4+ - remaining %

Initially the HYSYS simulation was runned without considering any pressure drop across the vapor outlet for the LTS separator. However, in reality pressure is loss as the vapor pass through the two mist eliminator and the swirldeck. As agreed with customers, liquid SHOULD not be present at the final gas product. The problem is that when i entered a pressure drop of about 3 kPa, liquid phase is observed coming out together with the vapor product.

So my concern is, what really cause the liquid phase to formed just by increasing a few kPa? Generally what is the range of pressure drop across the swirl deck and the mist eliminator? I hope someone can help me in this problem. Thanks.

Regards,
Andy Lum

[Srikrishna Chaitanya]

Dear Andy
i have just checked LTS simulation , with the data given by you
inlet temp 6 degree centigrade
pressure = 33 bara
composition
N2 - 0.38%
C1 - 90.4%
CO2 - 0.68%
C2 - 4.1%
C3 - 2.6%
C4+ - remaining % ( assuming C4= 50% ,C5=50% )
my property package is peng robinson
at these conditions inlet to LTS is super heated vapor , reduction of pressure dosent influence much
since dew pressure is around 195.435 bara .
normally reduction pressure leads to vaporization that to if the feed is at saturated conditions , since your feed is at superheated state at the conditions provided by you reduction in pressure will not cause any vaporization since already its at superheated state .
I hope in this problem your facing a problem with property package your using , once check it, you told water is there in your feed , but in composition you have not mentioned any water percentage .

i hope i understood your problem well and expecting your reply
regards ,
satish

[joerd]

Andy,

What are the operating pressure and temperature upstream the JT valve, and what are the operating pressure and temperature of the LTS?
Also, can you give a more detailed analysis of your C4+ fraction. The heaviest components in that fraction will determine the dewpoint, so it is absolutely necessary to know what the composition is.

[JoeWong]

Andy,
Welcome.

Initially the HYSYS simulation was runned without considering any pressure drop across the vapor outlet for the LTS separator. However, in reality pressure is loss as the vapor pass through the two mist eliminator and the swirldeck. As agreed with customers, liquid SHOULD not be present at the final gas product. The problem is that when i entered a pressure drop of about 3 kPa, liquid phase is observed coming out together with the vapor product.

This is pretty normal. When you allocate pressure drop on the "Vapour outlet" (Pls confirm my guess), HYSYS view it as pressure drop on the gas stream, outlet of VL separator and any liquid generated will be carried over to the gas stream.

So my concern is, what really cause the liquid phase to formed just by increasing a few kPa? Generally what is the range of pressure drop across the swirl deck and the mist eliminator? I hope someone can help me in this problem. Thanks.

JT effect due to pressure drop results liquid form from saturated vapor.

When you start to mentioned SMSM, you probably dealing with SHELL project. Don't you ? If so check out some value in DEP on Gas-Liquid Separator.

Proper simulation could be provide dummy separator (with zero pressure drop on gas stream) and "pump dummy sep. liquid stream" back to the liquid stream of real separator (usptream sep.). The pump here is an artificial pump again. Artificial pump represent gravity force...

Hope above help.

[andiez]

Hi all, thanks for the reply....

Dear Satish,
I am also using peng robinson for my packages... water quantity is quite small about 2e-7mol% at the LTS inlet, which i forgot to mention.

Joerd,
my operating pressure and temperature at the inlet of JT valve is roughly 10°C and 42 bara. my C+ component includes ibutane, nbutane, 22Mpropane, pentane, Mbutane, Mpentane, nHexane, Mcyclopentane,cyclonehexane, and minute quantity of C7 to C14. operating condition of LTS is about 6 °C and 33 bara.

JoeWang,
i have calculated the pressure drop across the mist elimintator and schoepentoeter which is about 1 -2 kPa (too small??) based on DEP. However pressure drop formula across swirldeck is not given. they did giv a range of 500 to 800 mm process liquid depending on the application.

Based on these few kPa drop across the mist eliminator and swirldeck, very small quantity of liquid is formed and carryover in the vapor stream. In reality does this really happens?

Thanks again for the reply.
Regards,
Andy Lum

[JoeWong]

i have calculated the pressure drop across the mist elimintator and schoepentoeter which is about 1 -2 kPa (too small??) based on DEP. However pressure drop formula across swirldeck is not given. they did giv a range of 500 to 800 mm process liquid depending on the application.

Based on these few kPa drop across the mist eliminator and swirldeck, very small quantity of liquid is formed and carryover in the vapor stream. In reality does this really happens?

Thanks again for the reply.
Regards,
Andy Lum

Normally calculated pressure drop is small...but provide additional pressure drop as margin.

The purpose of mist eliminator and swirldeck are to promote liquid droplet coalescence (entrained liquid droplet in vapor and liquid form due to JT effect) and separation from vapor by mean of centrifugal and gravitation...those liquid should trap in these devices and will not carry over to vapor.

However, there are separation efficiency involve in the trapping process. Different device has different separation efficiency. You may refer to DEP and check with vendor. Somehow do not too optimistic with these device, somehow certain amount of liquid may potentially carry over...Provide margin in your design...There may be some guideline in the DEP.

Hope above help.

[joerd]

Andy,
Reason I was asking for the composition is that I couldn't reproduce your dewpoint, and that you might have operating conditions in the retrograde region. However, with the data you provided, I predict a critical pressure of about 60 bara, so you are below the retrograde region. Additional pressure drop would not have the effect of condensing liquid.
The only other explanations I can come up with is that you specified some carryover in the separator, or that you have a funny hypothetical component in your composition specification that causes the simulator to predict liquids that are not there.
Other than that, without seeing the full stream and basis definition, I can't give you a better answer.

[Zauberberg]

I cannot find any physical reason for liquid condensation at LTS separator vapor outlet, having such small pressure drop. Something has to be wrong with vessel spec, fluid composition or the way how calculation order is configured. Or, unless you operate in the retrograde region, as Sjoerd pointed out.

I don't believe any liquid drop-out will take place in your system. This means that we would have to expect, for example, liquid in compressor suction after vapor passes suction scrubber and encounters pressure drop of, let's say, 5kPa across the outlet nozzle. This will not happen.

[andiez]

Hi all,

i tried to attach the HYSYS Simulation file for ur reference. But it seems that i cannot attach HYSYS file. Is there any other way i can attach it?

Regards,
Andy Lum

[joerd]

Zip it first.

[andiez]

hi all,

attached is my HYSYS simulation. Stream 8 is the mixture inlet to the LT separator, V-600. My main concern is stream 9 which contain some minute quantity of liquid when i specify 10 kPa drop across the vapor outlet. My goal is to get a pure vapor. Hope to get some advice. Thanks.

Regards,
Andy Lum

Attached Files

•  hysys.zip   116.03KB   142 downloads

[joerd]

Andy,

Thanks for posting the simulation.
What you are getting is a small J-T effect due to the pressure drop. You can verify this if you set the pressure drop in the separator to zero, and then put in a valve operation in the vapor line. If you give the vapor valve a pressure drop of 10 kPa, you will get a small liquid fraction due to the adiabatic expansion. This is the way Hysys calculates it; the vapor in the separator is saturated (i.e. on the dewpoint line), and the adiabatic expansion causes it to move into the two-phase region. You can verify it if you plot the phase envelope for the vapor stream (stream 9).
In reality, I doubt if you will get 10 kPa pressure drop. Also, any liquid formed during expansion may be caught by the rest of your swirldeck/mistmat combo. Lastly, if any liquid is formed, it will be a very fine mist and barely noticeable, and it will evaporate as soon as the stream heats up a little. Summarizing: theoretically you will predict liquid due to the pressure drop, practically you will observe vapor.

[andiez]

Thanks Joerd,

i think im much more confidence in my simulation now...
thanks all for the advices

Regards,
Andy Lum