Hydrocarbon Processing - June 2021 - 31

Process Optimization
only 273 mm above the top shed deck. At
the intended operating conditions, the
feed mixture was 4.8% vapor by weight.
By volume, the mixture was 89% vapor.
It was entering at a significant feed mixture
velocity of 16.7 m/sec. The ρM
VM
2
at
the 10-in. pipe and nozzle was 26,600 kg/
msec2
. This is about seven times higher
. At
VM
2-which signifies a high
than the common good-practice criterion
of keeping it below 3,700 kg/msec2
this high ρM
kinetic energy-and coming in so close
to the shed deck floor, it is highly unlikely
that the tower would have properly operated.
The high velocity and kinetic energy
would blow the liquid on the top shed
deck (T1) right up the tower.
Two other factors further aggravated
the situation. First, the open area of the
top shed deck was low-approximately
34% of the tower cross-section area.
Alone, this would not have been so bad,
as there was only a small amount of vapor
coming up through the window from underneath.
However, with the feed pipe discharging
so close to the floor, the top shed
deck was holding a very foamy mix, which
would have had difficulty descending via
the small window. Second, based on the
superficial area, the liquid velocity down
the tower (below the feed entry) was 95
m3
/ hm2
correlation1
. At this rate, the Barber and Wijn
predicts a foam height of 1.7
m. Looking at FIG. 2, this foam will almost
reach the bottom valve tray.
The task force concluded that the feed
arrangement and discharge velocities were
the root cause of the dark naphtha events.
Correction strategy. The corrective
action needed was to lower the point of
feed discharge into the tower, and to reduce
the kinetic energy of the discharged
feed, effectively separate the vapor from
the liquid upon entry to prevent foaming,
and prevent the feed from impinging on
the top shed deck liquid. The conceptual
process design of the new feed pipe was
performed by Fluor, and the mechanical
engineering and fabrication were done
by Bazan. Kock-Glitsch joined the task
force as a consultant on optimizing the
feed piping design and on achieving a
sound mechanical design. The mechanical
design was a special challenge due to
the very high kinetic energy of the feed
mixture. Neither of the authors' companies
had any involvement in the original
design of this tower and its internals.
The new feed pipe is shown in FIG. 5.
The 10-in. feed pipe was extended into the
tower, ending in a tee that split the feed
into two 10-in. laterals. The laterals had
93-mm-wide slots pointing downward at
37° to the horizontal (FIG. 6), discharging
the feed onto a new vertical impingement
baffle. This impingement baffle directed
the liquid downward toward shed deck
T2, with the vapor disengaging upward.
To allow the free rise of vapor from the
impingement baffle, two segments of the
top shed deck (T1) were removed. The
segment of T1 at the opposite end of the
tower from T2 was kept and was used to
collect liquid from the trays that descended
from the secondary seal pan via the
4-in. pipe. To distribute this liquid, a 50mm
weir was installed at the end of the unremoved
segment of the top shed deck T1.
Taking out the two segments of T1
was central to the success of the design.
The vapor rises on the inside of the impingement
baffle, so the region at the back
of the impingement baffle (between the
baffle and the tower shell) does not contribute
an open area for vapor rise. The
vapor velocity through the open area is
expressed by the C-factor in Eq. 1:
C = U MρG
where
C = the capacity factor (m/sec) based
on the open area for vapor rise
U = the vapor superficial velocity
based on the same open area, m/sec
ρG = gas density, kg/m3
ρL = liquid density, kg/m3
.
With no T1 panels removed, the C-factor
would have been 0.092 m/sec, which
is too large for a foaming system. With two
panels
removed, the C-factor declined
several-fold to an acceptable 0.02 m/sec.
The new feed entry design faced many
challenges. First, before splitting to the
FIG. 5. New feed entry design: (A) elevation
and (B) plan.
/ (ρL
- ρG
) (1)
tee, the feed mixture velocity in the extended
feed pipe was a high 16.7 m/sec.
The ρM
VM
kg/msec2
. At this
VM
2, which signifies a high
2 at the 10-in. pipe was 26,600
, about seven times higher than
the common good-practice criterion of
keeping it below 3,700 kg/msec2
very high ρM
kinetic energy, there would be intense hydraulic
pounding of the pipe and a large
bending moment. The tee was strongly
anchored to the tower shell. The flanges
on the feed entry pipe and the impingement
plate supports were double-nutted
FIG. 6. The laterals with the slots that directed the feed toward the impingement baffle at 37°
to the horizontal.
Hydrocarbon Processing | JUNE 2021 31

Hydrocarbon Processing - June 2021

Table of Contents for the Digital Edition of Hydrocarbon Processing - June 2021