Deadtimes of 1 min are included in both column temperature controllers, and relay-feedback tests are run to find controller settings. In TC1 they are KC = 1.15 and tj = 44 min (using a temperature range of 300-400 K and an output range of 0-2). As expected, this control is fairly slow because of the liquid hydraulic lags. In TC2 the controller settings are KC = 0.707 and tj = 9.2 min (using a temperature range of 300-400 K and an output range of 0-5.84 MW).

With this control structure in place, the simulation is run out to a steady state. Figure 8.28 gives the conditions, which are quite close to those found in Chapter 5.

8.4.3 Performance

The first thing to show is that a direct-acting aqueous level controller does not work. Figure 8.29a shows what happens. There is no disturbance. The controller is simply switched from Reverse to Direct. The system shuts down in about 2 h.

In contrast, Figure 8.30a shows the responses of the system when a reverse-acting aqueous level controller is used. The disturbances are positive and negative 20% step changes in the setpoint of the feed flow controller. Stable regulatory control of the highly nonideal distillation system is achieved. Both the ethanol and the water products are kept close to their desired purity levels.

Further evidence of the counterintuitive response of the decanter is given in Figure 8.29b. The aqueous level controller is placed on manual, and its controller output is increased from 53.49% to 55%. The flowrate of the aqueous stream increases. But the aqueous level increases!

Time (h)

Was this article helpful?

0 0

Post a comment