Sorry for the third post. I deleted the one at Student, since it's not a student case, it's a real problem for a start-up company. Any comments are appreciated.
Hi all,
I am trying to design a CO2 capture plant. The inlet stream of the absorption process is the output from a 50 MW power plant: either from Gas Turbine (typical value CO2 = 3% molar) or from Direct Fired Steam Turbine (typical value CO2 = 8%), with no H2S. Any thoughts on the values presented is valid, since I don't have practical experience and took the values from hand calculation + numerous papers presenting such values (for example, Gas Turbine with 100% conversion of 93% methane and 40:1 partial pressure of air to methane). We are at first going for the higher CO2 content, which will supposedly make the towers smaller. Please, correct me if I'm wrong.
Now, on the design, I have read many topics here with responses from Art Montemayor, which are extremely valuable. Reading some 4 books of gas sweetening (Ken Arnold Vol2, Kohl, Maddox Vol4, and Perry ChemE Handbook 8th), they all mention the "previous experience" into calculating height or number of stages.
Perry provides the Solubility curve of CO2 in MEA for dilute (up to 1% of CO2) at typical temperatures (40C, 30% MEA, but I am using 20 for simulation), but it is highly difficult to get the values. Mr Art has stated that for low pressures, this curve isn't necessary, which is my case, I want to do absorption at atmospheric. Besides, the fact that it is chemical absorption, makes the use of the curve not simple (although "Perry" also states that for VERY FAST chemical reactions, like MEA/CO2, the problem can be treated just like the physical absorption).
So, the problem statement is: Using Hysys or any other book method, how can I estimate the number of trays?
Via equilibrium curve from Perry 8th ed (14-8) , after struggle some hours to get the values from graph by eye+computer zoom, I got maximum 2 ideal plates, which using 33% efficiency for CO2 would turn to 6 plates and using 50% safety margin would go to 9. Is that reasonable? Any other suggestion on how to calculate this number of trays? The spacing I'm convinced is 30 inches, since diameter is > 6 m (see further).
Another way I calculated was with Colburn equation: Np = LN[(1-1/A)(y1 - y20)/(y2-y20) + 1/A]/LN(A). The data was: y1 = 8%, y2 = 0.8%, y20 = 0 (pure solvent), (Lm/Gm = 1,2 [dy/dx = (0.0792-0.00792)/(0.0598 - 0)], the 0.0598 taken from equilibrium curve of MEA 30% at 40C from Perry, so the Lm/mGm = 1.78 (1.5 Lm/Gm)
Np = 2.75 = 4, 33% CO2 Eff = 8.35 trays = 9 trays.
Objectively:
- Is this algebraic calculation of number of trays correct (colburn equation)?
- Is the use of the equilibrium curve for this system reliable?
- Is 33% plate efficiency reasonable?
More data:
Mass treated: 107 kg/s inlet
Temperature of inlet gas: 40 C
y1 CO2: 8%
y2 CO2: 0.8% (90% capture);
Amine Solution: MEA 15% (weigh);
Calculated mass of Lean Amine (Ken Arnold: Surface Production Operations, Volume 2, pg 186): 381 kg/s -> 15% excess: 438 kg/s. A different result comes from hysys simulation (around 300 kg/s), and it ends up with 5% of CO2 (%w) in the Lean Amine.
Calculated diameter of tower: now, this is tricky. I used Ken Arnold Volume 1 method: d^2 = 5040*T*Z*Qg/P * (rhog/(rhol-rhog)*Cd/dm)^1/2, as an iterative method (assume Cd, calculate Re calculate terminal velocity, calculate Cd again, repeat til converge), using Cd initial as 0.34 (recommended), dm = 150 mm. I arrive in the end at Cd = 6.18, Re = 5.28, Vt = 1.64 and that gives me a diameter of 15.4 meters. With 15% safety: 17 m.
Hysys Equipment design gives me, for valve tray and some other specs, from 10 to 11.5 meters. If I don't iterate Ken's method, just assume Cd = 0.69 (for no reason), it gives me around 11.4 m.
A third method, presented by Maddox on page 43/44 is using the equation: W = C((rhoL - rhoG)/rhoG)^0.5 and C = 350 or 395 for amine absorbers (24 and 30 inches spacing respectively), where W is the gas mass velocity (from which I'm able to calculate diameter by setting W = Q/d and checking which d gives me the closest Qg to my original Qg). That gave me a value of 10.5 m, which is 12 m with 15% safety. Maddox doesn't present a way to estimate number of trays.
Thoughts on that?
I appreciate any comments. Any additional info can be provided.