Chemical Engineering August 2022 - 24

Tower Doctor
X-Raying The Patient: Gamma Scanning
Vapor Cross-Flow Channeling
Henry Kister shares lessons learned from troubleshooting distillation towers
I
n 1988, the author led a team to
troubleshoot a chemical tower that
was revamped to achieve a small
increase in capacity (up to 4%) by
increasing the tray hole areas from
8.5% to 13% of the active area. In addition,
radii were added at the bottom
of the downcomers to smooth the
exit of liquid and reduce downcomer
backup. Strangely, instead of gaining
capacity, it lost 5% capacity.
Before 1992, the mechanism of
vapor cross-flow channeling (VCFC)
on sieve and valve trays was not recognized
by the industry. The author
inquired around, and everyone was
telling him that the capacity reduction
should not have happened. There
was one lead, as yet considered
unlikely, that suggested that VCFC
could be taking place, a mechanism
well-known in bubble-cap trays. The
author was following this lead.
In VCFC (Figure 1) the hydraulic
gradient on the tray induces preferential
rise of vapor near the outlet
and middle of the tray, and forms a
vapor-deficient region near the inlet
to the tray. The high vapor velocities
near the tray outlet step up entrainment,
while the low vapor velocities
near the tray inlet induce weeping.
Interaction between adjacent trays
accelerates both the outlet entrainment
and the inlet weeping. The
result is excessive entrainment and
premature flooding at the tray middle
and outlet, simultaneous with weeping
from the tray inlet, accompanied
by a loss of efficiency and turndown.
The theory was that increasing the
tray open area on the trays led to the
onset of VCFC and consequent premature
flooding on the trays.
The challenge was to test this
theory. Gamma scanning was
the most promising technique,
but no one had experience with
applying gamma scans for diagnosing
VCFC, nor any idea as to
how to do it. So it was decided
to brainstorm together with TruTec
(now part of Tracerco) who
had extensive gamma scanning
expertise. Understanding the
challenge, Tru-Tec lined up their
three top experts, and together
with the client's top experts
and the author, met in Tru-Tec
LaPorte's office. The meeting
started about 3 p.m. and lasted
for a few tedious hours. A lot
of ideas and stimulating brainstorming
were tossed around,
but no breakthroughs. People
began to wonder if there would
be any resolution.
FIGURE 2. Clear liquid heights derived from three gamma
scan chords along the flow path length are plotted on a toscale
column sketch. Tray liquid head values derived from
quantitative analysis of gamma scans are plotted so that
a liquid head of half the tray spacing is plotted as a point
on the tray above. Zero liquid head is plotted as a point on
the tray floor. Hydraulic gradients can be inferred from the
diagram (a) Unflooded (b) Flooded
Just then, one of the Tru-Tec experts
spoke. " Looks like we have
brainstormed for a few hours and
got nowhere. Maybe we are doing
it all wrong. So let me suggest another
way. " Then he continued, " Not
far from here there is an excellent
Mexican Restaurant called Don Key.
They have fabulous food, but what is
more important, is they have an excellent
beverage called Don'KeyRita.
It is similar to Margarita, except that it
kicks like a donkey. After one or two
of these, we may get more creative
and come up with a winning idea. "
His motion was unanimously and enthusiastically
accepted.
Not only did we enjoy a delicious
meal and Don'KeyRitas, but we left
with a plan that everyone was happy
with - and laid the foundations to
the quantitative scanning that is still
used to troubleshoot channeling.
The winning idea was to scan
FIGURE 1. This diagram depicts vapor cross-flow
channeling (VCFC)
24
three different chords along the flow
path length, both under flooded and
unflooded conditions, and to apply
quantitative analysis to derive
froth heights, froth densities and liquid
heads. From the liquid heads,
hydraulic gradients can be inferred.
The details are described elsewhere
[1,2]. The results are shown in Figure
2, where all the tray dimensions are
drawn to scale. Tray liquid head values
derived from quantitative analysis
of gamma scans are plotted so that a
liquid head of half the tray spacing is
plotted as a point on the tray above.
Zero liquid head is plotted as a point
on the tray floor.
In the unflooded scans (Figure 2a),
hydraulic liquid gradients are flat or
slight, sloping from tray inlet to outlet.
Upon flood initiation (Figure 2b),
the hydraulic liquid gradients on
the odd trays become very steep,
especially between the middle and
outlet. The even trays (trays 2, 4,
and 6), show large uniform hydraulic
gradients stretching from inlet to outlet
at flood initiation.
This intensification of the hydraulic
gradients strongly supported VCFC
as the root cause of the observed
premature flood. From an unlikely
hypothesis, VCFC became the leading
theory. Based on this diagnosis,
minor modifications were made
to tower auxiliaries that permitted
raising tower pressure. With VCFC
inducing a premature entrainment
flood, raising tower pressure reinstated
the lost capacity. Prior to the
CHEMICAL ENGINEERING WWW.CHEMENGONLINE.COM
AUGUST 2022
http://WWW.CHEMENGONLINE.COM
Chemical Engineering August 2022

Table of Contents for the Digital Edition of Chemical Engineering August 2022
