relay-feedback test with the other two controllers on manual and using the Tyreus-Luyben settings. The controller tuning constants are given in Table 10.2. The only surprising result of this tuning is the very high gain of the TC9 controller.
The response of this control structure for changes in feed flowrate is quite acceptable, as shown in Figure 10.23a. However, when some types of feed composition disturbance are made, the unit shuts down.
For example, as shown in Figure 10.23b, the process shuts down after about 4 h, when the feed composition is changed from 35/35/30 DME/MeOH/H2O to 40/30/30. The reflux flowrate R2 in the rectifier goes to zero because the temperature controller responds to lowering temperatures in the rectifier. This occurs because too much DME has worked its way down the column to the vapor sidestream drawoff tray. Once this happens, the rectifier can do nothing about rejecting the light DME. Instead of using stage 3 for temperature control, stage 12 was tested, but the results were similar.
To overcome these problems, a new control structure was developed. The flowrate of the reflux in the main column is quite small, so instead of manipulating it to control a temperature, it is held constant for a given feed flowrate (a reflux to feed ratio is used). As shown in Figure 10.24, the major change in the control structure is to control the
TABLE 10.2 Rectifier Temperature Controller Parameters
TC4 TC51 TC9
Was this article helpful?