7 replies to this topic

[praveen.bajaj]

I am currently working on a project for an offshore gas platform. The platform will be used to collect gas from various sources and then metered prior to delivery to sales pipeline. I am trying to calculate the flowrate of gas when the pipes will undergo depressurization due to fire scenario on the platform or due to maintenance reasons. The client wants the pipe segments to be depressurized to 100 psig in 15 minutes from MAOP (maximum allowable operating pressure).

The maximum gas (75 vol. % Methane, 11 vol. % Ethane, 2 vol. % CO2, 1 vol. %N2, 3 vol. % Propane and rest heavy components) pressure is 3,000 psig at 75°F. When conducting depressurization calculation using Aspen HYSYS dynamic depressurization utility, I came across some facts which I can not validate and would need engineering guidance.

HYSYS is stating the vapor fraction of supercritical fluid as 0 at 55°F and vapor fraction 1 for 75°F at 3000 psig. The maximum gas flowrate that would need to be depressurized is 2.5 times higher when HYSYS assumes the fluid to have vapor fraction 0 compared to when vapor fraction is 1. If I want to be conservative, I will use the higher flowrate but should I go with that reasoning and decide to size my flare header to handle high flowrate?

Hope I will get some feedback and engineering input into this problem.

Thanks
Praveen

[kkala]

A very simplified version of the case (also representing pressure reduction of total holdup from 3000 to 100 psig) was run by WinSim's Design-II.

Results can be found in attached depress.xls (with notes / comments), which is hopefully useful to some extent, even though my understanding on the matter is not complete. "Depress.xls" deals with global quantities (not flow rates). But indicates only a small decrease in vapor fraction, when high pressure temperature goes from 75 to 55 oF. It is expected to be so during depressuring.

Attached Files

•  depress.xls   21.5KB   356 downloads

[daryon]

Hi Praveen,

I put your composition into HYSYS using the Peng-Robinson EOS, I modelled C4+ as i-butane. Your starting conditions at 55°F (12.8°C) & 3000 psig (206.8 barg) is below the criconderm which HYSYS is predicting at 62.2°F (16.8°C), you are in the supercritical region but HYSYS is reporting the fluid as liquid. I do notice however that HYSYS does not report a latent heat of vaporisation. I have attached the HYSYS generated phase envelope to this post.

I suspect that when you run the depressurisation utility HYSYS is assuming that the heat input (fire scenario) is vaporising liquid during the blowdown. The additional vapour generation - which will be rapid due to basically no latent heat - is resulting in a higher peak rate during depressurisation. - This is just a thought, I will try and confirm with Aspentech.

When your starting conditions are 75°F (23.9°C) & 3000 psig (206.8 barg) there is no vaporisation and the peak rate is lower. By the way with my composition I have to raise the temperature to 130 F (54°C) before I get a vapour fraction of 1.000. Interestingly when I switch to the soave redlich kwong EOS the temperature reduces to 95 °F (35°C) when the vapour fraction is equal to 1.000.

In reality you are dealing with a supercritcal fluid where there is no gas/liquid boundary and the fluid properties are somewhere between liquid and gas. But HYSYS only reports liquid, vapour, and aqueous phases. In this region I don't think Peng-Robinson or SKR EOS are valid and reliable. I don't think it is valid for HYSYS to assume a liquid fraction of 1.000 for the depressurisation. I would suggest that you experiment with some other property packages and also run the depressurisation calculations without any heat input - see what happens. I'll get back to you if I get an answer from AspenTech. Good Luck.

Attached Files

•  Phase diagram PR EOS.xls   34.5KB   198 downloads

Edited by daryon, 22 September 2010 - 08:17 PM.

[daryon]

Hi again Praveen,

Further to my last email on this subjected I have talked to AspenTech, I have summarised the salient points from the discussion below;

1. The phase fraction values of supercritical stream (dense phase) in HYSYS are of no specific physical meaning since in reality there is no distinct between liquid and vapour phase separation in a supercritical region. Please see attached technical solution for reason why 0 or 1 is designated and why liquid or vapour correlations are used correspondingly. (I downloaded this from AspenTech Support Site - I believe it is acceptable to post this in the forum without infringing copyright/license laws sorry if not).

2.Despite the vapour phase fraction being meaningless, 0 or 1 will be designated based on the compressibility factor (Z) and the isothermal compressibility factor (beta) of the stream, and either vapour correlations or liquid correlations will be applied in the flash calculations correspondingly. This is why the higher peak flowrate results when the depressurisation facility is run with a vapour fraction of 0.

3.Aspentech advised that 'Peng-Robinson has no problem handling supercritical conditions. However, any potential concerns might come from the fact that there is some light gas (non-condensable) in the stream. Besides PR, you may want to try with NRTL as well. However, please be aware that for some physical property methods, there are certain physical condition constrains for validity. Please double-check if the supercritical conditions in your case have exceeded any of these validity ranges. Also, I see there are some “heavy components” as well. If they are created as hypo-components, certain physical property packages might not be able to handle them.'

4. As always Aspentech advise that the use of any thermodynamic model be validated with field data, which obviously is not practical here.

I hope this is helpful.

[fra.telli]

Last time I've seen this argument tereated was on a magazine.
"Designing for pressure safety valves in supercritical service"
Hydrocarbon Processing - January 2010, pag 63-67
www.hydrocarbonprocessing.com

hope this can help you!

[Zauberberg]

See attached:

Attached Files

•  Designing for Pressure Safety Valves in Supercritical Service.pdf   201.73KB   496 downloads
•  High Pressure Safety Relief Valves in Ethylene Service.pdf   455.26KB   353 downloads

[sheiko]

The following link discusses the first attached article by Zauber: http://webwormcpt.bl...ercritical.html

[praveen.bajaj]

I thank you all for very valuable comments. It has guided me and has made my thought process clear to how to handle supercritical fluid depressurization. Thanks once again for all your support and guidance to this engineering problem.

Best Regards
Praveen

Edited by praveen.bajaj, 28 September 2010 - 10:31 AM.