8 replies to this topic

[Lawrence Craig]

Hi there.

I was wondering if someone could possible help me out with a process I am trying to design on Hysys.

I have 30% MEA absorbing CO2 from the flue gas of a 500MW coal power plant. The aim is to recover 90% of the CO2 for storage. Now I have modelled the absorber in Hysys which works perfectly (95% CO2 is loaded onto the MEA).

Its the stripping column thats causing the problem. I am using a distillation column with a condenser and a reboiler. The MEA-CO2-H2O solution enters the column at 80/90C and having staged off the equilibrium diagram at 120C which produces 31 stages non-ideal. The column operates at 1atm. I try to specify the component recovery of 0.9CO2 in the condenser and set the reboiler temperature at 120C (reboiler duty was calulated as 300MW using hand calcs but not entered into the simulation). However for some reason the simulation is failing to converge. I have looked at numerous papers on Amine Absorption/Stripping simulation however none have helped me to get the column to converge. I have even tried using the different solvers and also using adaptive dampening.

If anyone could possibly help me or point me in the right direction, it would be much appreciated.

Thank you,
Lawrence

[Art Montemayor]

Lawrence:

I don’t know why you have set some far-out parameters on your simulation:

1) No one will design an MEA system with 30% wt solution and actually expect successful operation.. Even 20% is impractical to control in my opinion. I have always standardized in no more than 15% and have even run systems at 11-12%. The reason 30% is impractical is because of the degradation that will take place and the subsequent corrosion will dissolve your reboiler, stripper, and MEA exchangers.

2) You cannot be using a “distillation” column. Here I have to come down hard on your semantics. An amine process uses a STRIPPER – not a distillation column – to flash off and remove the CO2 from the amine solution. Most of the CO2 will flash off in the first 3 or 4 top trays. As I remember my hand calculations and designs, the Stripper only requires approximately 6 trays (actual) for a 15% wt. MEA solution, working at approximately 5 psig. I can’t imagine where you got 31 stages, but that is unreal and totally out of the park.

3) You don’t “set” the amount of CO2 produced in the Stripper. The Stripper is going to “strip” whatever CO2 was picked up in the absorber by converting the Rich MEA to Lean MEA. The CO2 has to come out at the Stripper overheads. It has no where else to go and it can’t stay in the solution.

4) At 120 oC, your stripper/reboiler are operating too hot; The Stripper should be at 105 oC. Something is wacky here. All the MEA reboiler needs is 35 psig steam – so that temperature and the fact that you say you are operating the Stripper at atmospheric pressure means that you MUST be at something closer to 101 oC in the Stripper – never 120 oC.

I hope these empirical field data are of some help to you.

[Lawrence Craig]

I understand 30% MEA is very corrosive and not practical in real solutions. However, I am just doing a simulation to compare various types of carbon capture. Papers by experts in the field (i.e. Gary Rochelle and others) have suggested using 30% MEA. I can understand your practical reasons why not to use it although as part of the critical analysis, I will mention the use of inhibitors may be required.

I understand that an absorption process does not use a distillation column. However HYSYS does not have a desorber with both a condenser and a reboiler, hence it must be modelled using a distillation column in HYSYS. I thought a distillation column can be used as long as its modelled using only the stripping section (i.e. by using equillibrium data).

One of the problems with the current process is that it the stripping column is not stripping enough of the CO2 from the feed steam. Is it possible that it is modelled on hysys wrong? I have modelled it so that the feed stream (containing water-co2-mea) enters the column at the top, the reboiler heats up the water and then the steam acts counter-currently to the feed hence removing the CO2. The tops product is the condensed with the water returning to produce near pure CO2. It could be that my stripper is too hot, however I have found equilibrium data at 120C, ill try lowering the reboiler temperature.

Thanks for your help. If you have any more suggestions (or would like to correct me on any of the above statements) please fire them at me. I appreciate it.

Lawrence

[Annaland]

Check the thermodynamic model you have chosen or used.

[Lawrence Craig]

Im using the Amine Kent Eisenberg Model

[Padmakar Katre]

Dear,
Please refer to the tutorial which gives idea how to model the Amine Sweetening Unit. Now as far as your modelling is concerned make sure that you are using the fluid package which covers all the components and Temp/Press Range. In your case amine package will suffice the need. The regenerator or the Stripper is modelled as a distillation column in Hysys. As there is only standard tempelate which could give the right approach is distillation column only. Now I would like to suggest you here that instead to give the Reboiler temperature as the column spec you try with another specs. As far as a distillation Column is concerned you will get the 2 degrees of freedom so give 2 specs as the CO2 molfraction in bottom stream(it will give an idea about how efficient your stripping is there) and the reflux ratio or some other spec. Normally unconvergence comes in any simulator when you have specs which overrides each other or there is basic problem with the modelling. So just check your specs and rerun the column. Waiting for your reply.

[Rudra]

I had also recently done simulation for absorber-stripeer.

I have used the specifications of reboier temperature and reflux ratio.
For absorber, amine model worked well in Pro-II. But stripper was not converging.
When i went back with NRTL , the simulation converged.

Is it okay to use NRTL for stripeer. I adopted it because in stripper CO2 is just being stripped off amine solution and amine property will not play a major role there. Please correct me if I am wrong

[Padmakar Katre]

I had also recently done simulation for absorber-stripeer.

I have used the specifications of reboier temperature and reflux ratio.
For absorber, amine model worked well in Pro-II. But stripper was not converging.
When i went back with NRTL , the simulation converged.

Is it okay to use NRTL for stripeer. I adopted it because in stripper CO2 is just being stripped off amine solution and amine property will not play a major role there. Please correct me if I am wrong

Dear,
NRTL is an activity coeffiient and I doubt it could solve your regenerator. If it at all solves the results are going to be weired/absurd. One time you can try with EOS like PR or SRK it will help but NRTL I doubt.

[Jojo]

Hi,

Try to use other method for your convergence, instead of newton you can use Broyden or Secant.

some words to this process

it's possible to use 30 wt% MEA, but with additive. Stripper and Absorber is for simulation the same unit, which can be solvented with MESH equation. because stripper is also equilibrium unit as absorber.