10 replies to this topic

[Zauberberg]

Good morning/afternoon to everyone in the forum,


I need assistance in finding references or articles that may be of some help in modeling reservoir systems, containing high percentage of CO2 - from 3% up to 30%. I am concerned about fluid packages that are to be used within the model.
I have some doubts that PR or PRSV equations of state can be used in such wide range of acid components concentration, especially having in mind that fluid consists of both hydrocarbon (liquid and vapor) and water/steam phases; so as CO2 concentration increases, non-ideality and interraction between highly polar components (CO2 and water/steam) must be taken into account.

I have to consider what fluid package should be used for modeling, in order to calculate (predict) as much as possible accurate phase distribution(s), component concentrations, physical properties etc. One of the main subjects is to estimate heat and material balance required for centrifugal compressor design, which is the next step in this project.

Thanks in advance,

[abhi_agrawa]

Zauberberg,
From your description it seems that it may be a good idea to use PR/SRK for the vapor phase and NRTL for the liquid phase. Such combination is possible in most of the commercially available softwares. However, if possible do get an opinion from a thermodynamics expert. Also if you have some actual data points available, it will be a good idea to simulate and check the accuracy of the prediction.
If the number of components is not too high, I'd suggest that you do a quick check of the Binary Interaction Parameters. I faced a problem once, where ASPEN removed the BIP's of CH4-H2S from their database. So the predictions were quite away from what we were actually getting.
Hope this helps,
abhishek

[shreyas]

Dear All,

This is my first post on the ChE resources.

In reply to the Zauberberg's question, I would like to add my comments.

For selection of any thermo model in the simulation, few key parameters need to consider are:

1. Pressure range
2. Tempt range
3. Availability of binary interaction parameters
4. Composition of the mixture
5. Nature of the properties of interest

Your systems I believe mainly consists of Sour components like CO2, NH3 and SO2. As per me, EOS or special package like GPSWAT/SOUR/GPSA can be applicable.

For your as well as others folks ready reference, I am attaching one .xls sheet which will be helpful in understanding applicable ranges & limitations of various models. Kindly select the desired model keeping in my mind above mentioned criteria's I mentioned.

Do reply later on in the matter.

Regards.

Shreyas Patel
shreyas.patel@ge.com

Attached Files

•  Applicability___limitations_of_thermo_models.xls   143.5KB   221 downloads

[Zauberberg]

Abhishek/Shreyas,

Thanks for the replies. Nevertheless, I am still concerned about applicability of these models within the range of CO2 concentrations, which is huge.

1. Can NRTL handle non-polar hydrocarbon components in liquid phase? The water fraction is quite high, and these hydrocarbons should be considered as immiscible with water.

2. Is there any way to predict at which CO2 concentrations (in vapor and/or liquid phase) thermodynamic correlations should be changed?

Anyway, I was planning to contact AspenTech support service about this issue, and if I get any feedback I will be glad to share this information and raise further discussions.

Regards,

[abhi_agrawa]

Zauberberg,

1) Yes. NRTL was designed to handle the liquid phase non-idealties. It is a good maodel for two-liquid phases also.

2) I personally have no idea about this.

abhishek

[shreyas]

Zauberberg,

I agree with Abhisek on point no. 1 i.e. NRTL is suitable for non-polar and immiscible system of HC and water. Only thing to keep in mind about the system pressure. NRTL is advisable upto 10 bar pressure. For pressure higher than 10 bar, EOS needs to apply for both the phases.

On point 2, If : Partial P(CO2) + Partial P(H2S) < 1200 psia (H2O, H2S, CO2, NH3 components required), better to use GPSWAT method. I think it is difficult to draw a clear-cut boundry line on CO2 concentration based thermo model. Anyway, need to check with experts on it...

Regards.

Shreyas

[abhi_agrawa]

Zauberberg,

Another thing that comes to my mind is the Electrolytic NRTL model. This model assumes that some CO2 that is absorbed in the aqueous phase is actually present in form of h+ and CO3-- ions. The dissociation constants are also built into a example file with ASPEN +. Sometime back i did use it to predict the solubility of CO2 in water. I found the results to be fairly accurate. You may try that also. However, I'd strongly suggest if you have any actual data point, do try to match it. This will give more confidence on the predictions.

Shreyas,
about the applicability of NRTL above 10 bar, I believe the problem lies with the fact that any activity coefficient model is of no good in the critical and near-critical region. Therefore the recommendation of not using NRTL above 10 bar pressure. If you are away fro the critical pressure, but at pressure more than 10 bar, I see no problem with using NRTL.

abhishek

[Zauberberg]

Very interesting information can be found at Virtual Materials Group website:

http://www.virtualmaterials.com/

Many of these guys originate from Hyprotech (Hysys thermodynamics group), and it looks like thermo-model development has been one of the main concerns of VMG. It is possible to register here, and ask for trial-version of VMGSim software (approx. 100MB), offering some new possibilities in upstream/downstream process modeling, including black oil characterization, hydrate formation, sulfur/amine plants etc.
Shell GSI has developed a special unit operation called C7PLUS accompanied by HYSPPS fluid package, in order to handle heavier components and to provide more accurate heavy oil characterization.

[Zauberberg]

Hello,

I promised I will return to this topic after receiving feedback from Aspen support team. So, from their point of view it is best to use EOS for vapor phase and one of the available activity models for mixed liquid phase (something similar was suggested by Abhishek). But, at the same time, they DO NOT recommend this approach because it is extremely sensitive on interaction parameters - which are not available in the software library by default. Instead of taking too much risk, using Sour PR would be the least erroneous method.

Just for a try, I have used two-models approach but the results were rather discouraging. And furthermore, there are no significant diferences between the results obtained by PR and Sour PR packages.

But, what surprised me the most is the fact that CO2 is four times more dissolved in hydrocarbon liquid (basically it is crude oil) than in aqueous phase - at every pressure and temperature ranging from 0 barg up to 30 barg and between 20C and 200C. Since this is the starting point that will affect downstream facilities design and/or revamp (depending on CO2 loads of every single phase - gas, liquid and production water), I wonder if this is true. I couldn't find any data regarding CO2 solubility in non-polar hydrocarbons, compared to solubility in water.

Any further assistance is much appreciated.
Regards,

[Steve McGahey]

Hello,

I promised I will return to this topic after receiving feedback from Aspen support team. So, from their point of view it is best to use EOS for vapor phase and one of the available activity models for mixed liquid phase (something similar was suggested by Abhishek). But, at the same time, they DO NOT recommend this approach because it is extremely sensitive on interaction parameters - which are not available in the software library by default. Instead of taking too much risk, using Sour PR would be the least erroneous method.

Just for a try, I have used two-models approach but the results were rather discouraging. And furthermore, there are no significant diferences between the results obtained by PR and Sour PR packages.

But, what surprised me the most is the fact that CO2 is four times more dissolved in hydrocarbon liquid (basically it is crude oil) than in aqueous phase - at every pressure and temperature ranging from 0 barg up to 30 barg and between 20C and 200C. Since this is the starting point that will affect downstream facilities design and/or revamp (depending on CO2 loads of every single phase - gas, liquid and production water), I wonder if this is true. I couldn't find any data regarding CO2 solubility in non-polar hydrocarbons, compared to solubility in water.

Any further assistance is much appreciated.
Regards,

Quick side note for you, Zauberberg - since you're talking about ASPEN... if it's HYSYS.Dynamics you're using (I say this because you're talking about compressors, where I do most of my work), then only the Peng-Robinson EOS has really been thoroughly tested for dynamics.

Chatting with my local thermo expert in the past (for a CO2 sequestration project), we came to the conclusion that the various thermo packages weren't that different from PR for the ranges we were dealing with. So, we ended up using PR for the sour gas injection work.

Hope this helps.

- Steve.

[abhi_agrawa]

Zauberberg,
Do you have access to Dechema? If yes then probably you can check your predictions in the following manner:
see if CO2- and heavy hydrocarbon VLE is reported. If yes then try to match it with an EOS. This will allow you to choose an EOS. Similarly see if water-CO2-hydrocarbon VLE/VLLE is reported for any hydrocarbon. If yes try to reproduce the result. Also if these equilibrium data is reported, then the interaction parameters are also quite likely to be reported. I had to resort to ths in the situation I told you earlier, when ASPEN removed CH4-H2S interaction parameters from their databank.

Hope this is of some help,
abhishek