28 ◾  Distillation Control, Optimization, and Tuning
increased from 700 ppm to 1,000 ppm C7, that is, from 99.93
wt% to 99.90 wt% C6, while the light key impurity concentration in the bottoms went down from 700 ppm to 400 ppm
C6. The results were similar to the tray temperature gradient
method (Figure 4.1). The D/F shift gave a maximum temperature change of 12.4% of 17.7°C, or 2.2°C, at stage number 16
from the bottom as being the MRT (most responsive temperature), that is, the most responsive vapor composition point, in
the column when there was a change in D/F. This small 300
ppm shift in material balance at the top and bottom of the column nearly shifted the stage 14 composition up into stage 16.
A shift in D/F, that is, distillate/feed ratio, material balance
split has a greater effect on a small stream than on a large
stream. For example, if the feed contains 10% lights, then the
ratio of distillate/feed ratio would be about 0.10, so the concentration change and variability would be more significant in
the small distillate stream. However, if the feed contains 90%
lights, then the concentration change would be most significant in the small bottoms stream. Consequently, a study with
two simulations should be run with a constant separation
power base and a change of composition in the smaller of the
two streams between the distillate and bottoms.
When the distillate heavy key impurity moved up from
0.0700% to 0.1000%, that is, a shift of 0.0300% C7, and the bottoms light key impurity moved down from 0.0700% to 0.0400%,
that is, 0.0300% C6, the stage number 16 shifted up 12.4% C7.
In this case, the stage 16 temperature moved 413 times more
than the top temperature due to change in composition. Any
significant temperature movement at the top or bottom of the
column would be mostly due to changes in pressure.
The tray temperature gradient method requires only one
steady-state simulation at design conditions, while the shift in
material balance method requires two simulations with different product stream compositions. The maximum tray temperature gradient of 7.6% of 17.7°C or 1.35°C/tray occurs at stages
15 and 16 as shown in Figure 4.1. These observations show