MaximaLabs
Simulation gallery

Urea synthesis loop from green ammonia and captured CO2

UreaGreen ammoniaRecycleFertilizer
See it live
Open this exact flowsheet in the workspace, already solved — tweak a spec and re-run.
Launch interactive model

Where green ammonia goes next. Ammonia and CO2 react in a melt at 150 bar and 185 degC; only part of the CO2 converts per pass, so the unconverted reagents are let down, stripped off and recycled, which is the whole reason a urea plant is a loop. Solved: 200 mol/s NH3 and 100 mol/s CO2 give 102.2 mol/s of melt at 96.2 mol% urea, 98.3 mol/s of urea (98% of the CO2 fed, the rest lost with the 2% purge), on a 155 mol/s recycle. The reaction heat is NOT typed in from a handbook sentence: it is computed from formation enthalpies already in this tree -- urea(s) -333.1 kJ/mol and water(l) -285.8 against ammonia(g) -45.6 and CO2(g) -393.5 (CODATA/NIST values bundled with the MIT-licensed chemicals package; this repo's own NBS 1982 row for urea agrees to 410 J/mol, 0.12%) give -134.3 kJ/mol for crystal urea and liquid water, and the tree's own NBS dissolution row for urea (+12.3 kJ/mol at 3 H2O) puts the melt basis the reactor actually produces at -122.0 kJ/mol, which is the value used. HONESTY NOTES, and each is a measurement or a checked absence rather than a hedge: (1) the TWO-STEP mechanism is not resolved -- real synthesis forms ammonium carbamate exothermically and then dehydrates it endothermically, and ammonium carbamate's formation enthalpy is in neither this repo's bundled NBS rows nor the chemicals package (both checked, both return nothing), so splitting the heat between the steps would mean inventing the number that decides it; the overall reaction is exact and the split is not attempted. (2) That lumping has a visible cost: this reactor must be COOLED by 19.2 MW to hold 185 degC, where a real urea reactor is close to adiabatic because the dehydration endotherm absorbs most of the carbamate exotherm. Of that 19.2 MW, 12.0 is the cited reaction heat (98.3 mol/s x 122 kJ/mol) and 7.2 is the streams' own enthalpy change -- Peng-Robinson's model for a pseudocomponent urea in a dense melt, which is screening-grade. The balance closes exactly: duty = F.dh + extent x dH. (3) The per-pass CO2 conversion is a stated plant input, not a prediction. The equilibrium correlations for it (Frejacques, Mavrovic, Gorlovskii) and the 1933 carbamate-urea-water equilibrium data they are fitted to are in paywalled literature this environment cannot read, so nothing here predicts how conversion moves with the N/C ratio, the water content or the temperature. (4) The feed is therefore STOICHIOMETRIC at N/C = 2, where a real plant runs 3 to 4 precisely because excess ammonia drives the equilibrium -- an effect this model cannot show for the reason above. It also has no dedicated ammonia-recovery section, so an excess-ammonia feed simply accumulates until the purge balances it: measured, N/C 3.5 drives the recycle to 9,700 mol/s and the purge to 198 mol/s. (5) The melt concentration is a specified-recovery water separator and NOT a flash, because the cubic gives urea a corresponding-states vapour pressure of 11.5 kPa at 408 K -- above urea's own 406.5 K melting point, where it decomposes rather than boils -- and that correlation is internally inconsistent with the bundled sublimation enthalpy (62.7 against 87.3 kJ/mol over 350-408 K). Flashed on it, two thirds of the urea left in the overhead. The number that would decide a real flash is not available. (6) No biuret, the side product that limits reactor residence time, and the stripper is an equilibrium flash rather than a real stripper and carbamate condenser driven by stripping CO2.

The flowsheet

The solved topology — every unit op's real duty, conversion, or split, read straight off a genuine converged solve.

NH3 FEED
CO2 FEED
MIX
Synth
Letdown
Stripper
Recycle Split
Recycle PUMP
Concentrator
MELT SEP
UREA MELT
Process Vapour
Purge GAS

The stream table

Every stream's flow, temperature, pressure, and composition — real converged numbers, not placeholders.

See it live
Open this exact flowsheet in the workspace, already solved — tweak a spec and re-run.
Launch interactive model

Stop fighting legacy software. Build your first flowsheet in 60 seconds.