31 replies to this topic

[Richard525]

I have a research project part of which is simulation of amine process with MDEA + piperazine. I want to use non-equilibrium column model and want a high accuracy of VLE. I wonder what software may be good for my needs?

[Poe]

i have used Promax from Bryan Research Engineering (BRE) there you can use piperazine with is a chemical for treat H2S in natural gas, if you want further information visit the web site:

 

www.bre.com

 

i hope this can help you

Edited by Poe, 23 May 2014 - 07:55 AM.

[AlbertHahn]

You can use HYSYS DBRAmine, or UniSim which is basically the same as HYSYS, or PROII. From my experience, they all have very good VLE and LLE predictions. As I also work in R&D, I had done a bunch of research in thermodynamics and mass transfer in sour gas treating. The thermodynamic models used in the three simulation software are pretty rigorous and they matches literature experimental data and our in-house data very well. I was told by Unisim that they have a comprehensive comparison of simulation data with the experimental data and plant data. The column is non-equilibrium stage model. I think they would be good for your use. Hope this helps. 

[Poe]

You can use HYSYS DBRAmine, or UniSim which is basically the same as HYSYS, or PROII. From my experience, they all have very good VLE and LLE predictions. As I also work in R&D, I had done a bunch of research in thermodynamics and mass transfer in sour gas treating. The thermodynamic models used in the three simulation software are pretty rigorous and they matches literature experimental data and our in-house data very well. I was told by Unisim that they have a comprehensive comparison of simulation data with the experimental data and plant data. The column is non-equilibrium stage model. I think they would be good for your use. Hope this helps. 

Dear AlbertHahn , how you add piperazine in HYSYS.

 

thanks in advance

[Erwin APRIANDI]

Hi Richard525,

 

Two software that was recommended during my conceptual study project (2007-2008) with amine is

- ProMAX and

- ProII

 

Two of this software has been proven to gives the best VLE calculation for amine absorption further their hydraulics calculation for column is also better at that time.

 

However, recently I also known that HYSYS and UNISIM has also improve their Amine Fluid Package, and the one that is recommended for HYSYS is DBRamine, in HYSYS you can add the piperazine to the component list, but I haven't really know how good it is compare to reality.

 

The good part with ProMax is that they usually provide some research results provided in their website and it is proven to be close to reality, based on some paper that they provide.

[AlbertHahn]

 

You can use HYSYS DBRAmine, or UniSim which is basically the same as HYSYS, or PROII. From my experience, they all have very good VLE and LLE predictions. As I also work in R&D, I had done a bunch of research in thermodynamics and mass transfer in sour gas treating. The thermodynamic models used in the three simulation software are pretty rigorous and they matches literature experimental data and our in-house data very well. I was told by Unisim that they have a comprehensive comparison of simulation data with the experimental data and plant data. The column is non-equilibrium stage model. I think they would be good for your use. Hope this helps. 

Dear AlbertHahn , how you add piperazine in HYSYS.

 

thanks in advance

 

in the DBR Amine Package, you can add MDEA, PIPERAZINE in the component list (of course you also need H2O and at least one acid gas --H2S or CO2). Piperazine cannot be used as a single solvent. Please be noted that the DBR Amine Package is different from the new sour gas treating module in recent versions of HYSYS.

[AlbertHahn]

Hi Richard525,

 

Two software that was recommended during my conceptual study project (2007-2008) with amine is

- ProMAX and

- ProII

 

Two of this software has been proven to gives the best VLE calculation for amine absorption further their hydraulics calculation for column is also better at that time.

 

However, recently I also known that HYSYS and UNISIM has also improve their Amine Fluid Package, and the one that is recommended for HYSYS is DBRamine, in HYSYS you can add the piperazine to the component list, but I haven't really know how good it is compare to reality.

 

The good part with ProMax is that they usually provide some research results provided in their website and it is proven to be close to reality, based on some paper that they provide.

As far as I know, the sour gas treating modules DBR Amine Package in ProII, HYSYS and UniSim are essentially the same. The DBR Amine Package has lots of improvements than the old Amine Pkg, though they are both provided by Schlumberger DBR Technology Center (previously DB Robinson & Associates, acquired by Schlumberger. This DBR is the Robinson in the famous Peng-Robinson Equation. They have long history of leading research in thermodynamics, mass transfer, columns, sour gas treating etc.) Many people don't know that DBR Amine Package is the advanced version of Amine Pkg. The Li-Mather model was developed by Robinson's colleagues and is a theoretically rigorous electrolyte thermodynamic model which was derived from Pitzer's work. They have their own lab to obtain their own data which I believe is unique among the commercial sour gas simulation software. Their column solver is a non-equilibrium stage model which is much better than the equilibrium models. The non-equilibrium effect is represented by stage efficiency and you can overwrite it if you are not satisfied with it. They have published lots of articles in academic journals and at conferences ( you may search articles by Alan Mather, Yi-Gui Li, Dan Zhang, Gordon Zhao or Jian Chen). I recently got some articles and data comparison results from them. It is really impressive.

[Richard525]

Thank you guys for all the helpful information.

 

AlbertHahn, can I see some of the comparison results or can you let me know how I can get those results? Thanks.

Edited by Richard525, 29 May 2014 - 11:30 AM.

[Technical Bard]

ProTreat from Optimized Gas Treating is also a very good program for amine simulation with good VLE and rate-based mass transfer calculations.

[RockDock]

I would also recommend ProMax. It is a rate-based simulator perfectly suited for MDEA + piperazine simulation.

[Richard525]

I would also recommend ProMax. It is a rate-based simulator perfectly suited for MDEA + piperazine simulation.

Is it the rigorous rate-based model similar to ProTreat?

[RockDock]

Richard,

 

Yes, ProMax and ProTreat are similar (the only two simulators I would use for sweetening). Both are rigorous rate-based models.

 

The main difference is that ProMax can show the approaches and tie in a SRU. ProMax is also much better at doing case studies. Pro-Treat can be a bit tedious and dated. But both give good results for sweetening units.From what I understand, most amine units are designed using ProMax (roughly 80%, if I had to guess), maybe about 10% with ProTreat, and the other 10% don't work.

 

If anyone gives me a sweetening unit model done in any software other than ProMax or ProTreat, I would not trust it. I would redo it in ProMax. All the talk of DBR and Hysys and Unisim is just a marketing ploy by big companies. If you want a truly good simulator for sweetening, use ProMax. ProTreat would be my second choice just because it is somewhat limited in it's features.

[AlbertHahn]

Richard,

 

Yes, ProMax and ProTreat are similar (the only two simulators I would use for sweetening). Both are rigorous rate-based models.

 

The main difference is that ProMax can show the approaches and tie in a SRU. ProMax is also much better at doing case studies. Pro-Treat can be a bit tedious and dated. But both give good results for sweetening units.From what I understand, most amine units are designed using ProMax (roughly 80%, if I had to guess), maybe about 10% with ProTreat, and the other 10% don't work.

 

If anyone gives me a sweetening unit model done in any software other than ProMax or ProTreat, I would not trust it. I would redo it in ProMax. All the talk of DBR and Hysys and Unisim is just a marketing ploy by big companies. If you want a truly good simulator for sweetening, use ProMax. ProTreat would be my second choice just because it is somewhat limited in it's features.

ProMax uses residence time for non-equilibrium effect, right?

[RockDock]

That's correct. It's a good point to bring up. ProMax uses the residence time (calculated by the column internals and flow rates) to rigorously calculate the progression of the reactions. I like that because there are no fudge factors to deal with. ProTreat uses "enhancement factors" to deal with the non-equilibrium effect. Both approaches work, but since amine sweetening is primarily limited by reaction kinetics, I trust the ProMax approach slightly more.

[AlbertHahn]

That's correct. It's a good point to bring up. ProMax uses the residence time (calculated by the column internals and flow rates) to rigorously calculate the progression of the reactions. I like that because there are no fudge factors to deal with. ProTreat uses "enhancement factors" to deal with the non-equilibrium effect. Both approaches work, but since amine sweetening is primarily limited by reaction kinetics, I trust the ProMax approach slightly more.

The residence time in ProMax is an interesting and tricky thing. I don't think ProMax is really a rate-base column model. RockDock, do you have some research literature in that? The rigorous model should solve the mass transfer rate using a series of ODEs, like what used in ProTreat. The rigorous model may have problems in convergence. The stage efficiency / enhancement factor is also a rigorous rate-base approach, which is used by the DBR amine. There are tons of research in the rigorous rate-base model and in the enhancement factor model. These two models are really the models that rigorously calculate the reaction kinetics. I personally don't think the DBR amine is a marketing ploy by big companies. On the contrary, I think ProMax is more successful in marketing their products.

Edited by AlbertHahn, 06 June 2014 - 05:41 PM.

[RockDock]

There are two competing phenomena for phase change in the amine absorber: reaction kinetics and mass transfer. For amine sweetening units, most of the time the phase change is limited by kinetics (which is based on reaction time). That, in of itself, makes ProMax rate based. Mass transfer certainly has it's place, but most commonly at higher temperatures and lower pressures (where kinetics are not limiting), like the regeneration column.

 

I would not compare the DBR amine efficiencies with the enhancement factor in a mass transfer model. While both are essentially fudge factors, the enhancement factor is based on mass transfer V/L units. The DBR efficiencies, as I understand, are more rules of thumb. I would not consider DBR to be rate based. I have never come into a plant and seen a Hysys or Unisim model match operating data for an amine unit. Somehow Honeywell and Aspen still manage to sell these companies on using their products for it. I just don't understand it. I've had a great experience with ProMax. ProTreat also matches data quite well. So, those are the only two softwares I can recommend for amine sweetening.

[OwenK]

RockDock and AlbertHahn,

 

From what I am aware, the only rigorous rate-based model is employed by ProTreat, not ProMax, while DBR amine package that is embedded in most of the software vendors’ platform, such as HYSYS, UniSim and Pro/II uses a non-equilibrium stage approach, at least currently.

 

As we all know, the model framework in a typical amine unit consists of Thermodynamic model which takes care of VLE and LLE, kinetic model (mass transfer, etc.) and column solving model. A profound Thermodynamic model is responsible for at least 70% of accuracy of the unit simulation. I would be more interested in seeing if any vendor could offer a comparison between the VLE or LLE calculations and the measured data. Without an evidence of such comparison, any case studies showing the match of plant data mean nothing much.  

 

[AlbertHahn]

Thank you guys for all the helpful information.

 

AlbertHahn, can I see some of the comparison results or can you let me know how I can get those results? Thanks.

I attended their trainings and got results from them. Sorry I cannot give it to you without their permission. I would suggest you contact DBR or UniSim/ProII. Also you can compare your own data or data from public literature to their software predictions. It is impressive to me.

Edited by AlbertHahn, 04 June 2014 - 05:16 PM.

[AlbertHahn]

There are two competing phenomena for phase change in the amine absorber: reaction kinetics and mass transfer. For amine sweetening units, most of the time the phase change is limited by kinetics (which is based on reaction time). That, in of itself, makes ProMax rate based. Mass transfer certainly has it's place, but most commonly at higher temperatures and lower pressures (where kinetics are not limiting), like the regeneration column.

 

I would not compare the DBR amine efficiencies with the enhancement factor in a mass transfer model. While both are essentially fudge factors, the enhancement factor is based on mass transfer V/L units. The DBR efficiencies, as I understand, are more rules of thumb. I would not consider DBR to be rate based. I have never come into a plant and seen a Hysys or Unisim model match operating data for an amine unit. Somehow Honeywell and Aspen still manage to sell these companies on using their products for it. I just don't understand it. I've had a great experience with ProMax. ProTreat also matches data quite well. So, those are the only two softwares I can recommend for amine sweetening.

The non-equilibrium effect is not directly controlled by the reaction, but by mass transfer. It is the reaction that affects the mass transfer, that's how the enhancement factor works. The stage efficiency model is calculated based on the enhancement factor which is determined by the reaction.

As far as I know, DBR published lots of papers in which the simulation results and plant data were compared. It appears from their papers the basis of DBR Amine was from the research work by Profs. Li and Mather and their colleagues. These results are from serious experimental and theoretical research. I don't think any one should say that stage efficiency and enhancement factor is fudge theories. I am not aware of any research work using residence time like ProMax does for the non-equilibrium effect (It would be interesting to see if there is any serious articles describing such approach). Users of ProMax have to provide a residence time of their own, and this time comes from experience of simulation, which implies that this time is a parameter for tuning that ProMax uses to match plant data. Is there any experimental measurement for the residence time or theoretical prediction? I doubt it. On the other hand, there are many research on the enhancement factor. This is why I really don’t trust ProMax. I would say there is no merit of ProMax for a rigorous research/development/design because of its not-so-rigorous theoretical basis. One has to give a credit for its marketing success of ProMax though.

On the hand, ProTreat also published some papers and I also can see the rigorousness of their software. I would evaluate simulation software by their background theories validated by experiments, otherwise simulation would be garbage in and garbage out. The ProMax column solver is essentially an equilibrium model. If you read their papers on the website about the inside-out algorithm, you can tell it. One can see the huge discrepancy of their approach with the Aspen/HYSYS inside-out algorithm (ref. Boston's paper, and others. )

Edited by AlbertHahn, 06 June 2014 - 05:44 PM.

[RockDock]

Users of ProMax have to provide a residence time of their own, and this time comes from experience of simulation, which implies that this time is a parameter for tuning that ProMax uses to match plant data. Is there any experimental measurement for the residence time or theoretical prediction? I doubt it. 

 

 

This statement is just not true. In my many years of doing ProMax simulation, I have never had to manually input the residence time (nor is it recommended to do so by BRE). It is calculated by ProMax. If that was the pretense you had to not trust ProMax (as you stated above), hopefully your trust is restored. ProMax calculates the residence time based on the column internals and flow rates. The user does not need to guess any values.

 

I don't know if ProMax is used for research, but I now it is extensively used in industry. I have seen countless papers that match operating data using ProMax. Even better, I have seen how closely it matches for units I actually troubleshoot. Some operating companies will give me their Hysys or Unisim models to help them solve the problem, but they never match. They very well may match in a laboratory, but it is more important to match industry. Almost every one of bre.com's papers show how it matches operating data. Their website also shows a lot of universities use ProMax - so, maybe it is used in research. Perhaps you should ask BRE.

 

The ProMax column solver is not an equilibrium solver, either. Every column shows how far from equilibrium it is operating. So, I don't really know how you can make that statement.

 

OwenK,

 

I would say that ProMax and ProTreat just employ two different methods of rigorous rate-based simulation. ProMax and ProTreat are the only two rate based simulators for amine sweetening. It's not that one does it the "correct" way and one doesn't. They each use different paths to reach essentially the same solution - both utilizing either reaction rates or mass transfer rates. Hysys and Unisim, on the other hand, are normally pretty far from any reasonable solutions for amine sweetening, due to their equilibrium based models.

[AlbertHahn]

 

Users of ProMax have to provide a residence time of their own, and this time comes from experience of simulation, which implies that this time is a parameter for tuning that ProMax uses to match plant data. Is there any experimental measurement for the residence time or theoretical prediction? I doubt it. 

 

 

This statement is just not true. In my many years of doing ProMax simulation, I have never had to manually input the residence time (nor is it recommended to do so by BRE). It is calculated by ProMax. If that was the pretense you had to not trust ProMax (as you stated above), hopefully your trust is restored. ProMax calculates the residence time based on the column internals and flow rates. The user does not need to guess any values.

 

I don't know if ProMax is used for research, but I now it is extensively used in industry. I have seen countless papers that match operating data using ProMax. Even better, I have seen how closely it matches for units I actually troubleshoot. Some operating companies will give me their Hysys or Unisim models to help them solve the problem, but they never match. They very well may match in a laboratory, but it is more important to match industry. Almost every one of bre.com's papers show how it matches operating data. Their website also shows a lot of universities use ProMax - so, maybe it is used in research. Perhaps you should ask BRE.

 

The ProMax column solver is not an equilibrium solver, either. Every column shows how far from equilibrium it is operating. So, I don't really know how you can make that statement.

 

OwenK,

 

I would say that ProMax and ProTreat just employ two different methods of rigorous rate-based simulation. ProMax and ProTreat are the only two rate based simulators for amine sweetening. It's not that one does it the "correct" way and one doesn't. They each use different paths to reach essentially the same solution - both utilizing either reaction rates or mass transfer rates. Hysys and Unisim, on the other hand, are normally pretty far from any reasonable solutions for amine sweetening, due to their equilibrium based models.

 

RockDock,

1. if you had read ProMax's paper on the inside-out algorithm they developed, you would see that they actually use equilibrium column solver and then put the stage efficiency into the equilibrium column solver. This inside-out algorithm was developed based on their pre-exist column algorithm.

2. I can access papers on academic journals published by both ProTreat and DBR. I can see how rigorous their theory is. ProTreat is the real rate-base one, and DBR is the stage-efficiecy (using enhancement factor) rate-base one. I think your statement that DBR is equilibrium one is obviousely wrong. I did not see rate-base theory of the ProMax type (I would appreciate it if you could pass me some literature of this type.). I never get any academic publication from ProMax. Also you cannot see any theories from those papers published on their website. I cannot trust any simulation software before I can understand their theory basis. As I said, the non-equilibrium effect is caused by mass transfer directly, not the reaction. The effect of reaction affects both the thermodynamic equilibrium and the mass transfer rate.

3.From my experience, I have both experimental data and plant data that can be matched very well by UniSim (DBR Amine). So I really doubt your statement. Users of DBR Amine need a certain level of knowledge to understand and use it. From my understanding, HYSYS/UniSim/ProII are all too big and they actually do not give a good enough education to their users in amine treating. It is true that ProMax is good in marketing.

[RockDock]

I don't think there will be a consensus between you and me. I think from our comments it is obvious that I do not have much experience using DBR and you have not used proMax. Both of us would likely be surprised about what we don't know about the respective models.

 

You believe the mass transfer is limiting equilibrium. I believe it is reaction kinetics. Either way, a predictive model cannot be achieved using tray efficiencies. That is why I don't agree with your statements about DBR/Hysys/Unisim/Pro/II. Perhaps you could change the efficiences of the DBR model to match operating data, but then what's the point? Once any condition changes, those efficiences will not be the same.

 

We also have different definitions of what rate-based is, I suspect. I view it as being time-dependent. I don't quite understand your definition of it. I like how ProMax shows me what the residence time is and how it effects the CO2 and H2S reactions. It helps me design or optimize a unit. I will absolutely call ProMax a true rate-based simulation, especially since I believe kinetics are more limiting than mass transfer.

 

Before our company switched to ProMax, we used Hysys almost exclusively. I have no issues with Hysys for HC processes.

 

I don't have any papers to offer you. I just suggest you contact BRE for them or attend one of their training sessions. I find them to be very knowledgeable and helpful.

[AlbertHahn]

I don't think there will be a consensus between you and me. I think from our comments it is obvious that I do not have much experience using DBR and you have not used proMax. Both of us would likely be surprised about what we don't know about the respective models.

 

You believe the mass transfer is limiting equilibrium. I believe it is reaction kinetics. Either way, a predictive model cannot be achieved using tray efficiencies. That is why I don't agree with your statements about DBR/Hysys/Unisim/Pro/II. Perhaps you could change the efficiences of the DBR model to match operating data, but then what's the point? Once any condition changes, those efficiences will not be the same.

 

We also have different definitions of what rate-based is, I suspect. I view it as being time-dependent. I don't quite understand your definition of it. I like how ProMax shows me what the residence time is and how it effects the CO2 and H2S reactions. It helps me design or optimize a unit. I will absolutely call ProMax a true rate-based simulation, especially since I believe kinetics are more limiting than mass transfer.

 

Before our company switched to ProMax, we used Hysys almost exclusively. I have no issues with Hysys for HC processes.

 

I don't have any papers to offer you. I just suggest you contact BRE for them or attend one of their training sessions. I find them to be very knowledgeable and helpful.

I agree with you that it is hard to get a consensus between us.

But for other aspects, it should not be justified by what you believe or what I believe. It should be judged by scientific facts and theories which is not dependent on or influenced by any individuals or companies. That's the point I have been insisting on. I listed so many theories with references and unfortunately I did not see a single reference that you could provide. 

As for the argument of mass transfer/reaction kinetics, I would suggest you check some scientific articles or books, e.g. Gas Treating with chemical solvents, by Astarita, Savage and Bisio.

As for the definition of rate-based, lets see what ProTreat says:

"A model deserves the name rate-based only if it is firmly grounded in mass transfer rate calculations. In other words, it must be a mass and heat transfer rate model. It’s not good enough that it might use chemical reaction kinetics as an element in the overall model—it must directly use mass transfer rates as well. Simulation software that relies on liquid residence time for its calculations on trays or packing is a case in point." (ref. protreat's website under "information" - "contactor" - "vol 4 No 3 Conflicted Meanings of "rate based" ).

BRE has never published any articles related to the fundamental works nor do they disclose the theory basis of their models. But I have read this type of work from other companies. Really cannot understand why. I think it is good for them to disclose the theoretical framework while still keep their trade secret.

Edited by AlbertHahn, 06 June 2014 - 05:38 PM.

[RockDock]

Of course ProTreat is going to say that their method of doing it is the only way of doing it. I don't necessarily appreciate their "holier than thou" approach to marketing. There is always more than one way to skin a cat. Both ProTreat and ProMax are able to match data. ProMax is even able to do it without enhancement factors. While ProTreat is a fine software, I do not agree with their definition of rate-based. It is just too self-serving.

 

I'm not in the research business. My job primarily involves design, troubleshooting and solving operating problems. Perhaps you should use a different model for research, but matching operating data with the least amount of information about the unit is highly valuable. That is one reason ProMax is so popular.

 

I'm familiar with gas treating using chemical solvents. That work also states that mass transfer and kinetics are independent of one another (the argument I am making). ProTreat accounts for the kinetics by using enhancement factors. ProMax does not need to because the kinetics are the limiting step. If kinetics prevent phase behavior, it does not matter what the mass transfer is.

 

I don't know why ProMax does not publish papers on their theory (maybe they do). I recommend you contact them if you are interested.

[AlbertHahn]

Of course ProTreat is going to say that their method of doing it is the only way of doing it. I don't necessarily appreciate their "holier than thou" approach to marketing. There is always more than one way to skin a cat. Both ProTreat and ProMax are able to match data. ProMax is even able to do it without enhancement factors. While ProTreat is a fine software, I do not agree with their definition of rate-based. It is just too self-serving.

 

I'm not in the research business. My job primarily involves design, troubleshooting and solving operating problems. Perhaps you should use a different model for research, but matching operating data with the least amount of information about the unit is highly valuable. That is one reason ProMax is so popular.

 

I'm familiar with gas treating using chemical solvents. That work also states that mass transfer and kinetics are independent of one another (the argument I am making). ProTreat accounts for the kinetics by using enhancement factors. ProMax does not need to because the kinetics are the limiting step. If kinetics prevent phase behavior, it does not matter what the mass transfer is.

 

I don't know why ProMax does not publish papers on their theory (maybe they do). I recommend you contact them if you are interested.

I think it won't help anything by further arguing on each side's opinion, but I have to say, it is not Protreat or any other's definition of rate base, it is the definition in text books and research articles/books. The kinetic or mass transfer, the same thing, it is not any one's individual opinion, it is the widely accepted research results. There are tons of books and articles on both topics. I don't have to give this credit to Protreat. Before you are able to read related books and articles, I think I should stop here after pointing this out again and again.

Edited by AlbertHahn, 15 June 2014 - 09:44 AM.