11 replies to this topic

[Dmitry]

Hello dear colleagues,

I am a young process engineer and I've got a task to consider the most simple option for gas dehydration. I decided to use Eglycol and J-T expansion to dehydrate gas.

I have feed stream with the following data
T= 10 C
P= 54,72 bar.
Flow=4491 sm3/hr

Gas analysys:

Component

mol.fr.,%

CO2

0.771

Nitrogen

6.077

Helium

0.006

Hydrogen

0

Methane

90.085

Ethane

2.415

Propane

0.434

і-Butane

0.051

n-Butane

0.094

neo-Pentane

0.002

i-Pentane

0.021

n-Pentane

0.017

Naphthenes C₅

0.002

i-Hexanes

0.007

n-Hexane

0.003

Naphthenes C₆

0.005

Benzol

0.002

i-Heptanes

0.002

n-Heptane

0.002

Naphthenes С₇

0.002

Toluene

0.002

The sales gas should be in accordance with 3 specifications:
1) water dew point -15 °C
2) HC dew point - 15 °C
3) Hydrate formation temperature -15 °C
4) Pressure of sales gas 25 bar

Please see PFD in attached file.

So, I used Hysys software and Peng-Robinson fluid package for my simulation and I did next:

Due to gas analysis didn't shown water in the feed stream I used Hysys saturation utility to saturate gas with water and as result I got saturated gas (Vapour fraction - 1) and at current working conditions I got:
1) HC dew point - 50 C
2) Water dew point - 10C
3) Hydrate formation temperature 10,97 C

Then I used adjust operation to find mass flowrate of glycol for reduce the hydrate formation temperature from 10,97 С to - 15 C. As result I got 1,4 kg/hr Ethylen glycol injection. Then I started to add glycol till HC dew point and Water dew point became - 15C.
For it I need to inject apr. 6 kg/hr of Ethylen glycol. Then I added J-T valve and specified pressure downstream 25 bar. Temperature of gas was reduced till - 5,5 C. And then I used 3 ph separator for separate rich glycol (Eglycol water) , NGL and sales gas.

Conclusions:

Sales gas is in accordance with design specifications, required quantity of Eglycol is apr. 6 kg/hr.

Dear colleges, Have I done all correctly? Fluid package should be PR? Do you have any comments regarding my case?
Probably you can propose another technology for my case?

Thanks a lot
Dmitry

Attached Files

•  case.jpg   84.16KB   111 downloads

Edited by Art Montemayor, 29 June 2012 - 08:40 AM.

[Technical Bard]

I would use the GLYCOL package in HYSYS 7.3 instead of PR - it better predicts the glycol VLE.

I haven't seen a process like this done before. Reasons include:

• You are using a "single stage" of contact, as opposed to a multi-stage contact column - which means you are very likely using more EG than should be necessary.
• Condensing liquid hydrocarbons with the EG will maximize the collection of heavier hydrocarbons in the EG phase. Notably, benzene and toluene are highly soluble in glycol - which means that when you regen the glycol you will have a problem with BTEX emissions, or will have to add significant scope to recapture the aromatics. See the site processecology.com (I have no affiliation) for more on this subject.
• I'm not sure your gas will be "saturated" with water at the plant inlet conditions. It was probably saturated with water at bottom hole conditions - and if the wells aren't producing liquid water the gas may not be saturated at surface conditions. You really need to know how much water you are producing from the gas wells.
• Also - you need to check the viscosity of EG at the cold separator conditions and see if that has any effect on the separation efficiency in that drum.

[Dmitry]

Tech Bard:

Thank you for comments,
My comments are:

1. When I used the Glycol Package and Hysys calculated 2 kg/hr of glycol for gas dehydration to meet sales gas specification .It is 3 times less if compare with results given by PR fluid package. So PR is not very good for such case?

2. Yes, I used a "single stage" of contact. I know about using a contactor (absorber) but in my case it's not applicable.

3. In rich glycol stream (EG+water) (using Glycol package) Hysys predicted Toluene, Benzene and other HC, so all it will go to the vent. When I use PR package Hysys predicted only CO2 and Nitrogen in rich glycol stream.
4.It depends that gas analysys cann't show water % in the feed stream, but as I know gas will always saturated with water at T and P. (in my case at inlet plant conditions). Tell me if I am wrong.
5. Not exactly understood about "you need to check the viscosity of EG at the cold separator conditions and see if that has any effect on the separation efficiency in that drum." Could you give me more inf. regarding it, please.

Regads,
Dmitry

[Technical Bard]

PR is not very good for glycol systems. THe GLYCOL package was specifically developed using the TST equation of state to do exactly the systems you are dealing with. Is it perfect? No. But it is better than straight PR.

The problem with benzene and toluene in the glycol is that the vent contains these toxic substances. In most jurisdictions it isn't legal to vent them - and it isn't ethical to do so even if the regulations in a jurisidiction don't cover it - eventually they will.

If your inlet gas is in contact with liquid water at plant inlet conditions then yes, it will be saturated with water.

When you size the separator, you need to be careful of the liquid-liquid separation, which is governed by Stokes' Law. If the viscosity of the glycol phase is high it is more likely to entrain other hydrocarbon liquids into the regen section.

[Dmitry]

What Aspen's consultants told me:

"The Glycol package BIP's has been regressed for TEG, Hydrocarbon and Water only. Glycol package contains interaction parameters for a limited component pairs for Ethylene Glycol system. So we recommend customers' to use Peng Robinson package with EG in HYSYS"

So for EG in my case I should use PR fluid package.

Regards

[Technical Bard]

I would check what the folks at processecology.com say - they have spent more time on this issue than anyone else I know, including the people at Aspen.

[Dmitry]

I would check what the folks at processecology.com say - they have spent more time on this issue than anyone else I know, including the people at Aspen.

OK, it will be good to find out

Regards,
Dminry

[Dmitry]

Also, as I know, the dehydration capacity depends of type of contact - "single stage" of contact – glycol injection in the feed stream upstream separator and a “multi-stage” of contact - absorber column. Could you tell me how much is the difference?

Regards,
Dmitry

[jojeecares]

Other portions of the dehydration using EG have already been adressed. However, i may have some comments regarding the use of Glycol Package for simulation. I have carried out a simulation of similar nature using HYSYS V7.3 and have observed slightest of differences when using the two property packages. I also consulted the local HYSYS trainer that advised on using PR for TEG as well. Anyways, I do agree with considering the viscosity factor for sizing of LTS Separator but at the same time have would like to tell you that EG has low viscosity as compared to TEG and so a standard weir arrangement can be used effectively other an interface controller would have to employed. Still, there would have been problems with entrainment. Therefore, generally, a 3-Phase a glycol flash tank is recommended prior to EG regeneration.

Secondly, I am not sure to the pressure profile of your facility but i would recommend considering that factor as well during future production scenarios and then expansion through JT. I did a conceptual study based on input parameters from reservoir department concerning pressure and production profile of hydrocarbons after which I had to drop the idea of JT and go for propane based mechanical refrigeration system.

Thirdly, you are right to assume water as fully saturated at inlet conditions. I have generally seem water production at inlet separator that leaves me to assume the fact that gas is saturated at separator outlet conditions as well. Furthermore, a metastable saturated water with hydrocarbon phase is not easy to model in HYSYS (or at least i dnt know how to).

[vamax]

hi,

you can consider alternative drying solution, using calcium chloride as drying agent. a basic column, with product.

www.newton-s.com

[Rajkumar Chate]

Hello Friend,

Please keep in mind, for water saturation u can use PR but for Contactor i will recommend to use Glycol package only. Connect gas feed stream to two phase separator then add some water in to separator to saturate the gas stream and sepcify the feed temperature to sepratore. After separator break the thermo from PR to glycol (it is important to break the thermo after separator because u will find different water content at same temperature with PR and glycol thermo) this is the way i do simulation for gas dehydration.

Please check the viscosity of glycol, it is important parameter for hydraulics purpose. one more thing u shall check is the gas feed and Lean glycol temperature difference.
I will not recommend PR for gas dehydration.

Regards,
Rajkumar Chate

[Neelakantan]

HI DMITRY

kindly note that any simulation is an indication and not the final word on the actuals - neither on the operations nor the design; what is is your role in the project - are you from the design / contractor side or from the client (operator) side? are you designing the facility including the equipment sizing for project to finalize the equipment? are you going to finalize the operations and safety design review? and who is handling the regeneration- design and operation?

as you mentioned you are a young engineer which means there will be a higher up in the setup; what is his advice or opinion?

if you are going to procure the package from another process company, i would advise you not to put your inputs on the operations or design but let them do it and then you review it to the client's satisfaction;

my questions may sound non-technical but will enable you to understand the design issue in a company;

now coming to the design, any glycol package is non-ideal hydrocarbon system which means that engineering thumbrules are better for sizing and operation rather than a rigorous simulation to trim any design; use general engineering guidelines for sizing the flow of glycol and lean and rich glycol concentration. as mentioned by others, use PR in the general case but use glycol package for the regeneration unit; definitely aromatics will go the vapours and not the regenerated glycol;

what are you doing for the regeneration vapours, normally venting is a cheap solution and may not be acceptable in many enlightened companies and countries; you either incinerate(force flare) or normal flare; when you go for normal flare the regeneration pressure incresases and so normally evacuation units are required

regards
neelakantan