Notes for the CC2 example simulation

Luyben2 reports the steady-state results in Figure 9.2 in his publication. These results form the basis for his dynamic studies. We can replicate these results only with fundamental model changes. Compared to Turton’s1 original design (in the 4th edition), Luyben’s results are based on different numbers for the process conditions and the reactor design:

• The recycle pump’s outlet pressure is 25 bar instead of 30 bar.

• The reactor geometry counts 342 tubes instead of 234.

• The reactor inlet temperature is 360°C instead of 350°C.

• The reactor cold-side temperature is 360°C instead of 250°C.

• The reactor outlet temperature is 427°C instead of 350°C.

If we adjust the flowsheet in Process mode according to the preceding design changes, we can match Luyben's reported cumene production and the reactor outlet temperature (the most prominent numbers to look at) only if we apply a different catalyst mass density of 640 kg/m3 instead of the intended 2000 kg/m3 (Turton specified 1600 kg/m3). We can reverse engineer this number by applying the factor 0.5*1600/2000 = 0.4 to the catalyst density (resulting in 1600 kg/m3*0.4 = 640 kg/m3, with e = 0.5).

Luyben’s control scheme is based on the assumption of a reactor cold-side temperature that is directly manipulated through a temperature controller. Our simulation models the reactor cold side as a high-pressure steam boiler powered by the reaction heat. This affects the dynamics.

Another key variable is the reactor’s overall heat transfer coefficient for the hot reactor gases on the hot side and the convective boiling on the cold side. Both Turton and Luyben suggest U = 65 W/m2-K. This confirms the cold-side temperature of 250°C that the simulation calculates for this U value, which agrees with the cold-side temperature reported by Turton.

If we use U = 65 W/m2-K, the simulation reports a pronounced temperature spike of ±10°C for scenarios 1 and 2. This contrasts Luyben’s results, which suggest an almost unnoticeable spike of ±0.1°C. If we make a 10-fold change to the U value such that U = 650 W/m2-K and we reduce the HPS valve (PV101) pressure drop to be DP = 2 bar, the temperature response to scenario 1 reduces to a maximum increase of 0.6°C. This indicates that the entire process model is very sensitive to the assumptions made for the reactor cooling mechanism. This aspect deserves attention.