Control structures CS1 and CS3 drive both product compositions back close to the desired values at the new steady state. There are significant dynamic departures, particularly in the bottoms composition when using CS3. Consider the effect of the large increase in feed flowrate at time equal to 7 hours. There is a dynamic tenfold increase in the impurity of propane in the bottoms at ^7.5 h. This occurs because the increase in feed flowrate brings more light material into the column, which affects bottoms composition quickly before the corresponding drop in stage 8 temperature can increase reboiler heat input to compensate for the increase in feed flowrate. Remember that the feed is liquid, so it affects the bottoms much more quickly and drastically than the distillate. In addition, there is a one-minute deadtime in the temperature loop. Also keep in mind that the tray selected is in the rectifying section above the feed tray.
This same problem exists in the other control structures, but in CS1 the increase in feed flowrate is accompanied by an immediate increase in reflux flowrate. This quickly affects stage 8 temperature, and reboiler heat input increases in time to limit the peak in the propane impurity in the bottoms to about 1.5 mol%.
The CS3 control structure with its fixed condenser heat removal does not return the product purities to their desired values for feed flowrate changes.
Obviously all three of these control structures could be improved by using feedforward control. In CS1 and CS3 the reboiler heat input could be ratioed to the feed flowrate (with the ratio reset by the temperature controller), and in CS3 the condenser heat removal could be ratioed to the feed flowrate. Figure 8.12 illustrates the improvement that the QR/F ratio provides for CS3 with the large-distillate case.
The solid lines represent CS3 feed rate disturbances with the ratio and the dashed lines, without. The worst-case peak in the propane impurity in the bottoms (xB, C3) is reduced
Was this article helpful?