Jump to content



Featured Articles

Check out the latest featured articles.

File Library

Check out the latest downloads available in the File Library.

New Article

Product Viscosity vs. Shear

Featured File

Vertical Tank Selection

New Blog Entry

Low Flow in Pipes- posted in Ankur's blog

Wet Gas Polytropic Compression Efficiency Evaluation

polytropic compression natural gas efficiency wet liquid centrifugal temperature pressure

This topic has been archived. This means that you cannot reply to this topic.
12 replies to this topic
Share this topic:
| More

#1 mrbabu

mrbabu

    Gold Member

  • Members
  • 55 posts

Posted 09 October 2012 - 03:16 AM

recently we changed the inlet spec's of our (two stages) centrifugal compressor (natural gas), now we have about 1% fraction of condensate (value estimated with both Peng Robinson or Soave thermo packages) at inlet,
the manufacturer said it's acceptable and we are reevaluating efficiency (and power consumption),
My questions concern the evaluation of polytropic efficiency in case of wet gas,
we have measured pressures, temperatures etc. and calculated the polytropic efficiency with Prode Properties (selecting the procedure for polytropic solution with phase equilibria), at the new operating condition the value is close to the original number (0.76) for dry gas (that could be Ok since we have only 1% condensate on inlet line and dry gas at outlet),

1) to improve the accuracy, should I attempt (and why) to evaluate directly (i.e. measure) the amount of condensate on inlet line?
Actually the liquid fraction is calculated by Prode with Peng Robinson model.

2) the outlet temperature estimated with the correlation for ideal gas tout = tin*(pout/pin)^(k*eff) is very different from the value calculated with Prode Properties (Prode calculates lower temperatures), I suppose it is Ok considering we have a pressure ratio of about 4 and some condensate to evaporate.

Since for this area (wet gas compression) little documentation is available I post this topic hoping to get comments and opinions from experts in this group.

Edited by mrbabu, 09 October 2012 - 03:19 AM.


#2 PaoloPemi

PaoloPemi

    Gold Member

  • Members
  • 549 posts

Posted 09 October 2012 - 04:05 AM

I am not aware of standards discussing the calculation procedure to estimate the polytropic efficiency of a compressor under wet gas (gas + liquid) conditions, Ankur or other experts in this group may give a more thorough comment,
for gas + different amounts of liquid I use the same procedure (Prode) to evaluate the efficiency,
in your case considering the low amount of condensate you may select the Huntington method (which is limited to dry gas) and see how much the results differ (you may be required to alter some value in order to have dry gas at inlet).
About phase equilibria at inlet condition I would presume that Peng Robinson can give reliable results.
Finally, according my experience it is quite usual to get higher temperatures (compared with rigorous procedures) with ideal gas correlation,

consider the case of methane compressed from 300K 10Bar.a to 40Bar.a

with efficiency = 1 (isentropic compression) the ideal gas model gives

cp/cv (tin,pin) = 1.3327 (Soave Redlich Kwong)
(k-1)/(k*eff) = 0.249
tout = tin*(pout/pin)^(k-1)/(k*eff) = 424K

with the rigorous model (see above) we get as tout = 409K

relative error (424-409)/(409-300) = 13%

now if we set a efficiency = 0.5 we get

(k-1)/(k*eff) = 0.499
tout = tin*(pout/pin)^(k-1)/(k*eff) = 599K

with the rigorous model (see above) we get as tout = 525.7K

relative error (599-525.7)/(525.7-300) = 33%

now if we add the contribute of phase equilibria (polytropic solution with phase equilibria model) we can observe large differences in tout-tin mainly due to the contribute of enthalpy of vaporization,

consider the case of the mixture methane 0.8 n-Butane 0.2 (molar fractions) compressed from 10Bar.a to 40Bar.a
with a polytropic efficiency = 0.75 , model Peng-Robinson

this mixture has a dew point of 287.9 at 10Bar.a then at 288K is in vapor phase
with the polytropic solution with phase equilibria starting from 288K @ 10 Bar.a we get as tout = 393K

at 272.47K 10Bar.a the mixture shows a liquid fraction = 0.1
with the polytropic solution with phase equilibria starting from 272.46K @ 10 Bar.a we get as tout = 337.5K

if we compare the values of tout-tin we get 105K in the first case (all gas) and 65.04K in the second case (with 10% liquid molar fraction at inlet condition)
a quite large difference,
in your case the liquid fraction is very small but it has certainly some influence.

Edited by PaoloPemi, 09 October 2012 - 07:31 AM.


#3 frpe

frpe

    Veteran Member

  • Members
  • 47 posts

Posted 09 October 2012 - 10:29 AM

I believe PaoloPemi has given a detailed answer,
in the past we did these calculations (vapor+liquid) by treating each phase individually,
the method requires a long sequence of calculations
and an automatic procedure coded in a software can reduce the risk of errors.
I agree with PaoloPemi that the Vapor/Liquid fraction as calculated by Peng Robinson should be Ok,
if you suspect that the EOS gives wrong results you can measure gas and liquid flows.

#4 chemdoc

chemdoc

    Junior Member

  • Members
  • 27 posts

Posted 10 October 2012 - 03:52 AM

for a quick estimate of efficiency with low amounts of condensate you may also consider an extension of the Schultz method where

n = ln(p2/p1) / ln(v1/v2)

and v1 and v2 are the (homogeneous) two phase volumes at inlet and outlet conditions,

differently the ideal gas model

n = (k-1)/k

eff = n * ln(p2/p1)/ln(t2/t1)

doesn't take in account the contribute of liquid fraction

#5 mrbabu

mrbabu

    Gold Member

  • Members
  • 55 posts

Posted 11 October 2012 - 11:59 AM

thanks all for the help,

a) Huntington method gives values (for polytropic efficiency) quite close to those calculated with polytropic solution with phase equilibria, I have tested Peng-Robinson, Soave Redlich Kwong , BWRS and Lee Kesler (Plocker) packages, values of efficiency (as calculated with different thermo options) show little variations (about 1% for different packages), a little more for the two different procedures (Huntington method or polytropic solution with phase equilibria)

B ) yes, in my case values for efficiency calculated with ideal gas method are quite different from those calculated with Huntington or polytropic solution with phase equilibria methods

Edited by mrbabu, 11 October 2012 - 12:00 PM.


#6 pathensey

pathensey

    Brand New Member

  • Inactive Member
  • Pip
  • 9 posts

Posted 14 October 2012 - 10:51 AM

we have a software based on ASME PTC 10, Schultz method,
and another based on GPSA manual,
as far as I understand the documentation both these products do not consider liquid phase,
could you provide some details about the method ?

#7 PaoloPemi

PaoloPemi

    Gold Member

  • Members
  • 549 posts

Posted 15 October 2012 - 09:24 AM

I presume you mean the model which includes phase equilibria,
herebelow a short description of methods included in my copy of Prode Properties with some comments

a ) Huntigton, applicable to gas phase, based on (Maxell) dH=TdS+VdP, works with a simplified model for compressibility integrated from p1 to p2, quite accurate and fast (accuracy depends from correct prediction of gas volumes)

b ) Schultz (Asme PTC 10), applicable to gas phase, based on PV ^ n = C which is exact for real gases only, Schultz includes corrections for real gases, I have read reports that this method gives higher errors than Huntington but it's a reference for industry (as GPSA method)

c) Polytropic solution with phase equilibria, based on solution of phase equilibria along a polytropic path, accurate (possibly most accurate in my tests based on entropy flash), can be very slow (compared to Huntington or Schultz), I use this method when a liquid phase is present.

Edited by PaoloPemi, 15 October 2012 - 11:55 AM.


#8 pathensey

pathensey

    Brand New Member

  • Inactive Member
  • Pip
  • 9 posts

Posted 20 October 2012 - 11:25 AM

thanks for the information,
I did know only ASME PTC 10 method and GPSA method (which averages values),
when you say that for Huntington method "accuracy depends from correct prediction of gas volumes" doesn't that be common to all methods ?

#9 PaoloPemi

PaoloPemi

    Gold Member

  • Members
  • 549 posts

Posted 23 October 2012 - 01:40 AM

pathensey,
in Huntington procedure eff*dH = V*dP , accuracy of resulting values depending only from accuracy of V which is high for a EOS,
differently in Schultz procedure (Asme PTC 10) there is a correction factor f which should minimize the deviation (for real gases) over isentropic path (efficiency = 1), errors are (generally) higher when efficiency < 1
Finally in Prode method the values are the result of database values and EOS corrections which can produce more accurate results.

#10 pathensey

pathensey

    Brand New Member

  • Inactive Member
  • Pip
  • 9 posts

Posted 29 October 2012 - 12:08 PM

I see the point,
details appears to be more important than supposed to be.

#11 marchem

marchem

    Gold Member

  • Members
  • 153 posts

Posted 06 November 2012 - 11:41 AM

that's true,
I have compared Schultz, Huntington and Prode in a range of applications,
Huntington is fast and very accurate for gas (the normal application for compressors)

#12 mattb

mattb

    Brand New Member

  • Members
  • 9 posts

Posted 10 November 2012 - 02:57 AM

mrbabu,
we have a eight impellers compressor on wet gas, with inter-stage condensing / separator system at 4 stage,
the amount of condensate is quite high,
you should carefully verify the process conditions to avoid troubles
(compressors are quite expensive machines).

#13 mrbabu

mrbabu

    Gold Member

  • Members
  • 55 posts

Posted 17 November 2012 - 08:02 AM

mattb, thank you for the information,
our application is probably very different,
it's a two stages compressor originally working with dry gas with a little separator without demister,
now we have about 1% fraction of condensate (Vapor Liquid calculated with Soave Redlich Kwong)
the simulation (see the previous posts) predicts a lower discharging temperature which is true according the tests,
there are still many aspects to evaluate, as the specific maintenance required when working with wet gas.




Similar Topics