3 replies to this topic

[Zauberberg]

As a rule of thumb, reliability of thermo-models in process simulation software declines as system conditions approach critical pressure and temperature. Knowing that difference between two phases (liquid and vapor) completely vanishes at critical conditions, modeling of systems which exist near or beyond their Tc and Pc becomes a challenging task. Having said that, I would like to hear opinions from other forum experts regarding these issues, and especially related to this particular case I am working on.

Gas condensate reservoir phase envelope is not available in this project. The only things that are known are reservoir pressure (436.3bara), temperature (96.5C) and light ends (C1-C5 composition, including N2, CO2 and H2S). C6+ fraction is present in relatively small amount, ranging from 5-20% weight (depending on bottom-hole sample location) and only bulk properties of this fraction are known (molecular weight and liquid density). The fact is that reservoir conditions are already in retrograde region (sample is 100% dense phase vapor), and when sample pressure is reduced isothermally, condensation takes place all the way down to ~275bara, when system is passing from retrograde to normal behavior and re-evaporation of liquid phase takes place. This has been confirmed by the set of laboratory analyses.

I have tried to characterize this heavy-end portion of gas condensate in 3 different ways, using available and extended procedures in Aspen 2006. The results I obtained are somewhat strange, and I would like to share these results with ChE forum members. I am attaching 3 phase envelopes generated in HYSYS, based on different C6+ characterization. My questions are:

1. Is it possible that fluid has only dew-point curve, without critical point and bubble point curve? The results (laboratory vs. software output data) obtained this way are very close to experimental data, and retrograde behavior pattern is the same. However, I find myself being very suspicious to believe in these results which predict, for example, that multicomponent fluid at 600bara will undergo phase transformation from 100% saturated liquid to 100% vapor within only one degree Celsius.

2. The most strange result is phase envelope in Case 3, where critical temperature is calculated to be higher than actual reservoir temperature, which means that fluid follows normal behavior when pressure is reduced at constant temperature. However, what is most important, fluid behavior within the range of 1-200bara pressure shows extremely good compliance with laboratory measured data. Multistage flash experiment performed at 96.5C and pressures from 200bara to 85bara almost completely matches with software predictions, inspite of the fact that this particular model predicts non-retrograde region at reservoir conditions.

For this particular project, my area of interest is only 1-150bara pressure region and 0-200C temperature region, because within these values future Gas Plant operating envelope is defined. And I have this doubt, whether I should follow original curve (Case 1 or Case 2 - slightly modified), or I should base my future work on the model which accurately reflects fluid behavior within the range of Gas Plant operating envelope, in spite of its innaccuracy in the high pressure and high temperature regions.

It is very difficult to explain all details of this particular case, and if you need any additional data please let me know.
Best regards,

[attachment=756:Phase_En...___Aspen.xls]

[joerd]

Interesting, but not uncommon, I'm afraid. I would not bother too much about the left part of the curve in your case 1 - that is definitely an artefact from the EOS. I have observed that even small amounts of certain hypos will create a curve like that - failure to calculate a bubble point and critical point. Sometimes this behavior occurs when you have Helium in your composition.

I would say that you need to focus on the region of interest, i.e. 0-200 C and 1-150 bara, and get that right, including the quality lines if you can. This will size your gas plant equipment, and that is what you need. Also, you will probably not get the sample composition as feed gas composition when the field starts up, so make sure that your separators and heat exchangers are big enough. Do some sensitivity analysis.

That said, nothing is better than to obtain better data - as Campbell said, "One cannot calculate reliably what one cannot measure reliably."

[Zauberberg]

Thanks for feedback,

Actually I was quite confused by observing the fact that C6+ characterization has significant influence on shape of the phase envelope, inspite of relatively low content of C6+ in gas condensate. I always prefer to create my own hypothetical components using HYSYS Oil Manager and providing such distribution shape that will result in smooth density and distillation curves of liquid products in multistage flash drums. I have noticed long time ago that hypotheticals created in Basis Environment (as Hypo group) always give poor liquid density and heavy-end properties predictions.

I think I found a good way to utilize available experimental data; however, I don't have any basis for model comparison outside of this laboratory report. Basically, there are only isothermal multistage flash experiments at 97C and constant volume depletion experiment at the same temperature. For the purpose of Gas Plant design two extreme cases (compositions) will be considered and I think this can give us a good starting point for exploring various plant configurations and equipment sizes.

Anyway, as you said, by playing with the model without further input (measured) data - everything in this project would fall in speculations domain, and that is exactly what I am trying to avoid.

Best regards,

[mishra.anand72@gmail.com]

Another resignation on behalf of Aspen.