REVAMPS

Field Data, New Design Correct

Faulty FCC Tower Revamp
Scott W. Golden
Process Consulting Services, Inc., Houston, Texas

*Karl D. Schmidt
Lyondell Petrochemical Co., Houston, Texas

Gary R. Martin
Process Consulting Services, Inc., Bedford, Texas

I
n 1987, Lyondell Petrochemical Co. After the revamp, at
revamped a fluid catalytic cracking 92,000 b/d charge rate,
unit (FCCU) main fractionator by the gasoline true boiling
TOP PUMPAROUND
replacing trays with structured packing. point (TBP) endpoint was
This revamp did not achieve its design consistently 550ºF or
objectives; therefore, a second revamp higher. The endpoint did
GASOLINE/LCO FRAC.
was performed in 1992. Packing large not change with increased
diameter main fractionators can increase fractionator reflux or LCO PRODUCT
unit capacity and decrease pressure drop decreased unit feed rate. LCO/HCO FRAC.
while meeting fractionation objectives. High gasoline endpoint
STRUCTURED
However, a packed main fractionator resulted in 7,000 b/d of PACKING
revamp is less forgiving than a trayed col- heavy gasoline being HCO PUMPAROUND

umn revamp and must take into consider- blended to the middle dis-
ation proper design and inspection proce- tillate pool and 2,400 b/d WASH
dures. Lyondell’s experience illustrates gasoline lost to LCO prod-
the approach needed to have a successful uct. Several modifications
structured packing revamp. to the packed column SLURRY PUMPAROUND
GRID
internals did not improve
Background gasoline quality. The col-
The 79,000 b/d FCCU main fractionator at umn internals eventually
Lyondell Petrochemical Co.’s Houston were modified by a sec-
refinery was first revamped from trays to ond revamp. This revamp
structured packing in 1987 (Figure 1). The met the 92,000 b/d capac- Figure 1 Initial Column Revamp
justification for the revamp was a capaci- ity, but also the design
ty increase to 92,000 b/d. An ultimate pressure drop and fractionation. Packed losses included 7,000 b/d of heavy gaso-
capacity of 100,000 b/d was anticipated, main column internals must be properly line, which had to be blended to the mid-
at the same gasoline cutpoint as the designed, installed, and inspected before dle distillate pool year-round, and 2,400
trayed column, and at reduced column start-up to ensure good performance. b/d gasoline loss to LCO product. The
pressure drop. The revamp design gaso- Many packed columns have failed remainder of the 550ºF endpoint heavy
line D86 endpoint was 445ºF. because of poor design practices or faulty gasoline could be blended to the refinery
Reactor-regenerator pressure balance is installation. gasoline pool because of the masking
affected by main column pressure drop. effects of the light refinery gasoline blend-
(Figure 2) Before the revamp, unit capaci- Problem Definition ing components.
ty and conversion had been limited by The project justification was increased The approximate FCC gasoline material
low cat-to-oil ratio. Reducing pressure charge rate, higher gasoline production, balance after the revamp was:
drop would allow lower reactor operating and lower pressure drop. The main frac-
pressure, which permits higher catalyst tionator raw gasoline (liquid product from • Light gasoline (gasoline splitter):
circulation and increased conversion. The overhead receiver) had a high endpoint 17,600 b/d
revamped unit pressure-balance also per- and the light cycle oil (LCO) product con-
• Light gasoline (cat. naphtha
mitted lower regenerator pressure.1, 2 The tained 2,400 b/d of recoverable gasoline.
fractionator): 4,800 b/d
revamp included a new regenerator air The gas plant fractionates full range gaso-
grid design that would generate higher line into light, middle (heart-cut), and • FCC gasoline splitter bottoms:
pressure drop; therefore, the regenerator heavy gasoline prior to blending or repro- 6,000 b/d
pressure had to be lowered to maintain cessing (Figure 3). • Heart-cut: 9,950 b/d
the air blower discharge pressure at the In addition to high gasoline endpoint, the
• Heavy gasoline: 12,000 b/d
required level to meet the regenerator air LCO product consistently had a TBP 20
requirements. vol % point of 430ºF or less. Gasoline • Total FCC gasoline: 50,350 b/d

Appeared in May 31, 1993 issue of Oil and Gas Journal ®, pgs. 54-60.
 REVAMPS

WET GAS
16.7 10.0 LIQUID
15.0 COLLECTOR/REDISTRIBUTOR

4.0
DISTRIBUTOR HOLES

APPARENT
0.5-IN.CRIMP
11 FT-2 IN. HETP > 20 FT
STRUCTURED PACKING

13.7

COLUMN P = 3.7 PSI

Figure 4 Gasoline/LCO Fractionation

equivalent of a theoretical plate (HETP)

PRESSURE, PSIG
of greater than 20 feet. Refinery main frac-
tionator beds are short (3-8 theoretical
AIR
stages) and the columns are typically
large diameter. If one cross-section of the
Figure 2 FCC Pressure Balance (Trayed Column) bed has an L/V ratio much lower than
another section, the result will be high
Field Troubleshooting no effect on fractionation. endpoint material leaving the low L/V sec-
Field troubleshooting was performed after • The measured column pressure drop tion of the column. Lyondell concluded
initial attempts to correct the problem was near the design value. The that there was poor liquid distribution and
failed. Several field observations were individual bed pressure drops could the consensus was that structured pack-
made: not be measured because no ing does not work in large diameter FCC
• The column overhead vapor instrument taps were installed main fractionators. A test run was
temperature did not respond to (and low pressure drop is difficult to planned to gather additional field data.
increased reflux. At times, the measure accurately.)
overhead temperature actually Field testing showed poor fractionation Initial Revamp Justification
increased with higher reflux. throughout the column. Poor initial distri- The original column design used trays,
• The gasoline endpoint and the front bution and/or little or no remixing of inter- which limited FCC feed to 79,000 b/d
end of the LCO did not change nal liquid often cause poor fractionation. (Figure 5). The measured trayed column
materially with large reflux rate Poor liquid distribution causes variations pressure drop was 3.7 psi. At 79,000 b/d,
changes. in liquid/vapor ratios across the tower 75% conversion, and test run heat and
• The fractionation between the cross sectional area, which result in com- material balance conditions, the calculat-
unstabilized gasoline and the LCO position gradients. Figure 4 is a schemat- ed pressure drop using the column inter-
product was the same at 60,000 b/d ic of the gasoline/LCO fractionation sec- nal loadings was 3.6 psi. The stated
or 92,000 b/d feed rate. At 60,000 b/d, tion. This bed had approximately 0.5 the- revamp objective was a 15% increase in
large reflux rate changes had little or oretical stages, or an apparent height feed rate while maintaining fractionation
comparable or better than the trayed col-
umn. Minimizing column pressure drop
FCC UNIT REFORMER UNIT
consistent with these stated objectives
MIXED BUTANES
LIGHT CAT LIGHT CAT had product conversion benefits.
GASOLINE GASOLINE
Reduced pressure drop allowed adjust-
TO ALKY TO STORAGE TO STORAGE
ments in the reactor/regenerator pressure
profile. These adjustments enabled
GASOLINE CAT. NAPHTHA Lyondell to increase the catalyst-to-oil
DEBUTANIZER SPLITTER FRACTIONATOR ratio with a corresponding increase in unit
conversion and selectivity. The regenera-
tor pressure was lowered, permitting high-
HEART-CUT
FROM er air grid pressure drop without reducing
DEPROPANIZER TO REFORMER air blower capacity.
Table 1 shows the raw gasoline and LCO
product distillations during a test run of
the trayed column. The top circulating
reflux operated at about 126 MMBtu/hr
HEAVY CAT heat removal, which equates to about
GASOLINE
CAT NAPHTHA 58,000 b/d of internal reflux. Table 2
SPL. BOTTOMS TO STORAGE
shows the calculated percent flood and
TO STORAGE
observed fractionation section efficien-
cies. The tower performance had approxi-
Figure 3 FCC Gasoline Fractionation System mately three theoretical stages between

Appeared in May 31, 1993 issue of Oil and Gas Journal ®, pgs. 54-60.
 REVAMPS

the top pumparound and slurry product gravity specification.

the LCO product draw Heavy cycle oil (HCO) product was sent
#26 stream. The trayed frac- to fuel oil blending to control the slurry
#25 TOP PUMPAROUND
TOP PUMPAROUND #24 tionator was flooding in product to less than -1.0 API gravity.
#23
#22
GASOLINE/LCO one section, although the Slurry is sold as carbon black feedstock. If
#21
#20 FRACTIONATION exact section could not be the slurry pumparound heat removal was
RICH SPONGE OIL #19
#18 determined. Lyondell too high, it was not possible to meet the
RICH OIL
#17
#16
conducted a plant test to slurry product gravity specification
LCO TO STRIPPER #15 LCO/HCO determine whether the because too much light material was con-
#14 FRACTIONATION
LEAN SPONGE OIL #13 flooded section could be densed.
#12
HCO PA #11 HCO PA isolated. The top The first revamp was done to increase
#10
#9 pumparound had the capacity and raise conversion. Lyondell
#8
WASH
highest calculated per- decided to replace the trays with struc-
#7
#6 cent flood. It was postu- tured packing. The baffle trays were
#5
#4
lated that if this section of replaced with grid and the remainder of
SLURRY PA
#2
#3 the column was flooding, the internals with structured packing. The
#1 the fractionation could be packed column hydraulic design was con-
TRAYS 1- 7 improved by reducing the sistent with 100,000 b/d fresh feed at 75%
FEED BAFFLE TRAYS
loading in the raw gaso- conversion, although the unit capacity
line/LCO fractionation was limited to 92,000 b/d because of envi-
section. ronmental permit limitations. A new wet
Field testing was done to flue gas scrubber planned for 1994 would
Figure 5 Trayed Columns, Capacity 79,000 B/D determine where the col- allow operation at as much as 102,000
umn was flooding. The b/d. Table 3 summarizes the fractionation
heat balance was adjusted to lower the bed depth and design packing perform-
load on the top section of the column. ance. The column has two pumparound
Trayed Column Distillation Curves Approximately 17 MMBtu/hr were shifted side draws and one product side draw
from the top pumparound to a previously (Figure 6).
D86 Temperature, ºF
shutdown LCO pumparound system. The A combined liquid collector/redistributor
fractionation became worse, which was selected to decrease the height
Raw Gasoline LCO Product showed that the gasoline/LCO section was required for pumparound/product draws,
IBP -- 265 not flooding. Two theories prevailed. The which allowed maximum packed bed
-- 392 first assumed the slurry pumparound baf- depth. The design column pressure drop
5
fle trays were flooding, resulting in mas- was 0.8 psi. The packed column started
10 92 415 sive entrainment of liquid to the wash up in early 1987.
30 178 492 zone trays. The second theory was that
50 255 545 wash zone trays 8 - 9 were flooded by Revamp Column Performance
high vapor rates caused by reduced slurry The design performance objective was
70 331 604 pumparound heat removal. The symp- improved fractionation in all zones. When
90 399 681 toms of the flooding were rapid loss of col- the column started up, some of the initial
416 715 umn bottoms level and rapid buildup of distillation data showed a gasoline splitter
95
pressure drop in trays 9-17, where a dif- bottoms product endpoint as high as
EP 442 736 ferential pressure recorder was installed. 575ºF, with typical values as shown in
5-95 vol% overlap, 24ºF. Increasing slurry pumparound heat Table 4. The FCC gas plant initially sepa-
removal and raising column pressure rates the full range gasoline in a gasoline
Table 1 eliminated flooding. Increasing the main splitter column, producing light gasoline
fractionator pressure was not a reason- overhead and a bottom product. Part of
able control method because of its the bottom product is routed to a heart-
adverse effect on the pressure balance. cut splitter in the reformer and the
Trayed Column The operators monitored the tray pressure remainder is sent to storage. The heart-cut
Capacity/Efficiency drop and maintained the slurry splitter separates the heavy gasoline into
pumparound duty at about 200 light, heart-cut and heavy gasoline
MMBtu/hr to avoid flooding. The column streams. The light and heavy cuts are
Section % Flood NTS*
was operated with maximum slurry refinery gasoline blendstocks and the
Top pumparound 100 -- pumparound duty consistent with the heart-cut stream is reformer feedstock.
Gasoline/LCO 88 3
LCO/HCO 80 2 Original Design Separation Efficiency
HCO Pumparound 91 --
Fractionation Zone Packing Bed Height, in. Design HETP, in. NTS*
Wash 70 1 Gasoline/LCO 1/2-in. crimp 134 21 6.4
Slurry 71 -- LCO/HCO 1/2-in. crimp 62 21 3.0
*Calculated number of HCO/Slurry 2-in. crimp 42 42 1.0
theoretical stages * Number of theoretical stages
Table 2 Table 3

Appeared in May 31, 1993 issue of Oil and Gas Journal ®, pgs. 54-60.
 REVAMPS

SPRAY HEADER LIGHT CAT. GASOLINE FROM

GASOLINE SPLITTER
TOP PUMPAROUND STRUCTURED PACKING
COLLECTOR/DISTRIBUTOR
LIGHT CAT. GASOLINE FROM
REFINERY
CAT. NAPH. FRAC.
GASOLINE/LCO FRAC. STRUCTURED PACKING
GASOLINE
GASOLINE SPLITTER BOTTOMS
COLLECTOR/DISTRIBUTOR
FROM FCCU BLENDER

LCO/HCO FRAC. STRUCTURED PACKING

SPRAY HEADER HEAVY FCC GASOLINE FROM
STRUCTURED PACKING CAT. NAPH. FRAC.
HCO PUMPAROUND
COLLECTOR/DISTRIBUTOR

STRUCTURED PACKING
WASH
V-NOTCH DISTRIBUTOR REFINERY
SLURRY PUMPAROUND GRID
MIDDLE-DISTILLATE
7,000 B/D HEAVY GASOLINE TO
FEED MIDDLE-DISTILLATE POOL YEAR BLENDER
ROUND (HIGH ENDPOINT)

Figure 6 Original Packed-Column Internals Figure 7 Blending Problems-High Endpoint FCC Gasoline

blended to the refinery

Heavy Gasoline Distillation middle distillate pool
(Initial Revamp) year-round (Figure 7).
During winter months,

SUMP
the refinery operated
Vol% Temperature, ºF in maximum middle
DRIP TUBES
distillate mode, and
90 420
the 7,000 b/d was typi-
95 460 cally needed for mid- STRUCTURED
VAPOR RISER PACKING
97 490 dle distillate produc- TOP PUMPAROUND LIQUID

99 550 tion. Therefore, during COLLECTOR AND GASOLINE/LCO

FRACTIONATION LIQUID DISTRIBUTOR
gasoline season, 14%
DRIP TUBES
Table 4 (excluding gasoline STRUCTURED
lost to LCO) of the FCC PACKING

gasoline production
A test run was conducted on Oct. 7, 1987 was lost to the middle
at 93,700 b/d charge. Table 5 shows the distillate pool, which *INITIAL REVAMP
raw gasoline and LCO product D86 distil- had a major impact on
lations. refinery economics. The Figure 8 Blending Problems-High Endpoint FCC Gasoline
The packed main column was producing heaviest portion of the
high endpoint gasoline and the LCO prod- FCC gasoline has a low RVP relative to 6 summarizes the observed HETPs of the
uct contained a large amount of gasoline. average FCC gasoline. The impact on the fractionating beds.
Poor fractionation caused 2,400 b/d of total refinery gasoline pool is thus greater Theoretically, the operation of the com-
gasoline loss directly to LCO product. than the volume loss of 7,000 b/d. The bined collector/distributor will cause poor
However, the main problem was the high two possible solutions were to replace the liquid distribution. Gathering field data to
endpoint gasoline. Of the total FCC gaso- packing with trays or fix the packed col- prove that liquid distribution was poor
line production, about 7,000 b/d had to be umn. was difficult because the column had few
temperature measurements. The column
Troubleshooting overhead and top pumparound draw tem-
Raw Gasoline/LCO Distillation Packed columns do not fractionate where peratures before the revamp were avail-
there is poor liquid distribution, flooding, able. Table 7 shows the recorded temper-
(Initial Revamp)
and/or poor vapor distribution. The initial
revamp used a combined collector/distrib-
Raw Gasoline LCO
utor to maximize packing bed depth. Initial Revamp
Vol% Temp., ºF Vol% Temp., ºF Figure 8 shows the combined top Fractionation Performance
pumparound liquid collector and gasoline
70 316 IBP 282 fractionation section distributor. The top
90 380 5 389 pumparound design rate was about 6,000 Zone Actual NTS Actual HETP, ft.
gpm, while the design internal liquid rate
95 408 10 413 Gasoline/LCO <1 >10
to the fractionation section was about
97 481 20 449 3,500 gpm. A combined pumparound col- LCO/HCO 0 >10
EP 540 30 485 lector and liquid redistributor device has HCO/Slurry 0 >10
never been used successfully in large
Table 5 diameter fractionators. To illustrate, Table Table 6

Appeared in May 31, 1993 issue of Oil and Gas Journal ®, pgs. 54-60.
 REVAMPS

Temperature Comparison,
Trays vs. Packing COLLECTOR/DISTRIBUTOR
SKIN-TEMPERATURE
SURVEY ELEVATION
Packing
Trays (Initial Revamp) PACKED APPARENT
BED 5 FT 2 IN.
NTS ~
~0
Top Temperature 284 276
Top Pumparound Draw Temperature 348 380
Raw Gasoline D86 Distillation, Vol%
520
90 399 393
560
95 416 424
590
100 442 540
600
Table 7
615

ature at about the same top column pres- internal temperatures. 620
sure before and after the revamp. Skin temperatures can
After the revamp, the top pumparound be calibrated against
*BELOW LCO PRODUCT COLLECTOR/DISTRIBUTOR
draw temperature was much higher, indi- actual internal tempera-
cating high endpoint material at this ele- tures by measuring skin
vation in the column. The pumparound temperatures near a col- Figure 10 Skin Temperature Above LCO Draw
draw temperature is its bubblepoint; umn thermowell. If skin
therefore, the higher draw temperature temperatures are measured radially either has one bad distributor and good mixing
implies high endpoint. Previous distilla- above or below a packed bed in the vapor at the lower redistributors eliminates the
tion analyses of the top pumparound space, the magnitude of maldistribution propagation of the composition gradients
draw were not available. The column had can be inferred from the temperature dif- down the column. In an effort to reduce
no thermowells. However, vessel skin ferences of the radial measurements. investment and increase packing bed
temperatures could be measured with a Figure 9 shows a radial survey taken depth, a combined liquid collector and
portable thermocouple to infer column below the top collector/redistributor. The orifice pan distributor was used. An orifice
temperature varied from pan distributor should never be used in
495ºF near the draw large diameter refinery columns. The col-
nozzle to 360ºF directly lector/distributor had a wide sump with a
COLLECTOR/DISTRIBUTOR opposite the pump- draw nozzle at one end. There were no
SKIN-TEMPERATURE
SURVEY ELEVATION around draw nozzle. drip points in the sump, leaving a large
Figure 10 is a radial skin part of the packing without initial distri-
PACKED APPARENT
BED
11 FT 2 IN. NTS ~ 0.5
~
temperature survey bution. The pumparound collector/distrib-
taken above the LCO utor also had a significant liquid gradient
product draw tray. The from the side opposite the pumparound
skin temperature data draw to the draw nozzle. An attempt was
360
380 showed there was severe made to modify the collector/distributors
liquid maldistribution by installing drip tubes in the sump and
throughout the column. using a second nozzle opposite the first.
Compositional gradients These modifications made little or no
caused by poor initial improvement in column performance.
400
liquid distribution were
415
445 never corrected at lower Revamp Two: Fixing The First
475 495
elevations because the Revamp
combined collector/dis- Lyondell decided to perform a revamp of
*BELOW TOP PUMPAROUND COLLECTOR/DISTRIBUTOR
tributor provides no the packed column. A new design basis
Figure 9 Radial Skin-Temperature Below Top Collector remixing. A column that simulation was used to establish product

Design Product Yields, Second Revamp (BBL/D) Design Heat Balance

Pumparound Duty, MMBTU/hr
Products Initial Revamp* Second Revamp Delta Yield
FCC Gasoline 49,397 51,797** +2,400 Initial Revamp* Second Revamp**
LCO 25,983 23,583 -2,400 Top 203 203
HCO 3,400 3,400 HCO 121 121
Slurry 4,134 4,134 * Design Rate
** Same Internal Reflux Rates
* Actual Performance
** Gasoline TBP Endpoint, 455ºF Table 9
Table 8

Appeared in May 31, 1993 issue of Oil and Gas Journal ®, pgs. 54-60.
 REVAMPS

Figure 11 illustrates the proposed LCO/HCO fractionating bed of an FCC,

SPRAY HEADER second revamp design. The design the liquid rates are approximately 0.5 to
uses separate liquid collectors and 1.5 gpm/sq ft2 of tower area. In the gaso-
COLLECTOR redistributors and less packing. line/LCO section or the light/heavy naph-
LIQUID DISTRIBUTOR
Good liquid distribution to a tha section of an atmospheric pipe still,
GASOLINE/LCO FRAC.
packed bed is important because the liquid rates are 6-8 gpm/sq ft2 of tower
COLLECTOR packing does not redistribute liquid. area.4 The higher liquid rate distributors
LIQUID DISTRIBUTOR
LCO/HCO FRAC.
A major misconception is that col- have very different momentum and hori-
SPRAY HEADER lector/distributor spacing should be zontal velocity design considerations. At
HCO PUMPAROUND sacrificed to increase packed bed low liquid rates, these collectors/redistrib-
LIQUID DISTRIBUTOR depth. Approximately 25% of all utors are easier to design.
refinery large diameter packed main Designing a large diameter distributor
WASH fractionator revamps do not meet requires attention to detail. Liquid enter-
LIQUID DISTRIBUTOR
design objectives, primarily be- ing the parting boxes has momentum and
SLURRY PUMPAROUND
cause of poor liquid collector and the methods of feeding the parting boxes
redistributor designs. In one case, should minimize horizontal velocity.
the liquid collector design caused a Aeration and horizontal velocity will
major product yield loss on a vacu- cause poor liquid distribution; parting
Figure 11 Second Revamp um column.3 Packing HETP in boxes should use some type of calming
large diameter main fractionators is zone to reduce the effects of momentum.
yields and column design internal load- controlled by liquid distribution quality. The liquid level in the parting box should
ings. Good liquid distribution results in low be adequate over the entire operating
Design/Installation HETPs. HETPs in large diameter refinery range so that horizontal velocity is low.
Table 8 shows the new, estimated design fractionators depend on the collector/ Higher liquid rates require more elaborate
yields for the main fractionator. The raw redistributor system designs. When and more costly distributor designs. Each
gasoline design yield was based on a TBP designing a liquid distributor for a 24-foot design uses the same basic principles, but
endpoint of 455ºF. The gasoline product diameter column, the following items the specifics are different. All fractiona-
yield improvements required better frac- need to be evaluated: tion bed liquid distributors in this column
tionation between gasoline and LCO. are fed internally from collector trays
Modifying the collector/distributor would • Liquid rate (Figure 12). This is typical of refinery
increase gasoline/LCO fractionation to • Distributor feed method columns having multiple product draws
four theoretical stages from one half of a - Feed pipe and heat removals. Sections of the collec-
theoretical stage. The second revamp heat - Internal overflow from liquid tor tray and the parting boxes and distrib-
balance is shown in Table 9. collector (such as a pump- utor troughs all form part of the liquid dis-
The product quality comparison between around or product draw) tributor system. The weirs feeding the liq-
actual performance and the second • Mechanical requirements uid from the collector trays, parting boxes,
revamp design is shown in Table 10. - Support and troughs were installed with water lev-
- Installation els to ensure levelness. The distributor
- Leveling parting boxes and troughs were designed
Design Product Qualities Increasing liquid rate makes the dis-
with independent level adjustments to
ensure each part could be properly lev-
TBP Distillations, ºF tributor design more difficult. In a
eled.
lube vacuum column or the
Initial Second
Raw Gasoline, Vol% Revamp Revamp
50 262 269
70 317 325
PACKING
90 393 401
95 424 426
EP 523 455 PUMPAROUND DRAW =
COLLECTOR
6,000 GPM
LCO, Vol %
IBP 263 327 LIQUID TO BED =
3,500 GPM
5 380 428 REDISTRIBUTOR
10 414 452
30 488 508 10 FT STRUCTURED
PACKING
50 540 552
70 601 607
90 673 668
95 701 690 *TOP PUMPAROUND/FRACTIONATION BED

100 785 722

Figure 12 Liquid Collector and Redistributor
Table 10

Appeared in May 31, 1993 issue of Oil and Gas Journal ®, pgs. 54-60.
 REVAMPS

Distillation Curves, Raw Gasoline

First Second Revamp Second Revamp

Trays Revamp Design Basis Actual Data
IBP -82 -78 -78 -79
5 37 28 29 33
10 92 89 91 90
30 178 206 210 183
50 255 262 269 250
70 331 317 325 312
90 399 393 401 402
95 416 424 426 425
EP 442 523 455 446
Tail, 95-EP 26 99 29 21

Photo 1 Thorough Packed Column Internals Inspection Table 11

Field Inspection Performance

A checklist of all the items to be field The column internals were modified in main fractionator with external reflux
inspected was made before the shutdown. early 1992. A thorough test run was from the overhead receiver because the
An experienced engineer should inspect planned to determine the actual packed reflux is metered. If the wash oil rate or
the column internals before the manways bed efficiencies. Before the test run, the another internal reflux stream is meas-
are closed (Photo 1). Many columns fail meters were zeroed and calibrated and ured due to a total draw, then a good
because there is no one person responsi- the material balance checked. Once the check of the column heat balance is pos-
ble for this activity. A column should not material balance was acceptable, a heat sible. During the test run, the top
be inspected by committee or by an inex- balance around the fractionator was per- pumparound duty was approximately 170
perienced engineer because the job is crit- formed. The heat balance was acceptable; MMBtu/hr with a calculated internal
ical.5 Inspections also should not be therefore, a full test run was scheduled. reflux of about 2,500 gpm to the gaso-
made for the sole purpose of identifying The unit was operated stably for 24 hours. line/LCO fractionating bed. The column
that the equipment was manufactured per A full set of stream samples were taken was operated with lower than design
the drawings. Sometimes something is every 8 hours for laboratory analysis. internal reflux because the top
designed incorrectly, but built and Material and energy balances were per- pumparound heat removal was limited by
installed correctly. The inspection is the formed on these three sets of data. The pump circulation and exchanger prob-
last opportunity to catch mistakes that resulting data were then used to run the lems. The top pumparound pumps will be
cause shutdowns (Photo 2). Many computer simulation. Lyondell’s main modified during a future shutdown so
columns have to be shut down for modifi- fractionator has no metered reflux that column internal reflux can be
cations after a revamp, which is much streams; therefore, consistent heat and increased.
more costly than correcting an error material balance data was needed to Table 11 shows the product qualities for
before start-up.6,7 determine the efficiency of the the trayed column, initial revamp, and
gasoline/LCO frac- actual performance of the second revamp.
tionation bed. It is The column has met design objectives
easier to check the (Table 12). The column now responds to
accuracy of the heat operational changes and the product
balance data on a qualities reflect these changes.

Column Performance, NTS

Initial Proposed Second
Section Trays* Revamp* 2nd Revamp Revamp*
Gasoline/LCO 3 <1 4 4
LCO/HCO 2 0 2 2
HCO/Slurry 1 0 1 1
(Wash)

* Simulation of plant data

Photo 2 Final Inspection: Last Opportunity to Correct Mistakes Table 12

Appeared in May 31, 1993 issue of Oil and Gas Journal ®, pgs. 54-60.
 REVAMPS

The Authors References

Scott W. Golden is a chemical engineer with 1. Golden, Scott W., Sloley, Andrew W.
Process Consulting Services, Inc., Houston, Martin, Gary R., Revamping FCC
Texas. Mr. Golden’s focus is applying funda- Main Fractionators Effects on Unit
mental chemical engineering skills and basic Pressure Balance, Hydrocarbon
process equipment knowledge to identify Processing, March 1993, pp 77-81.
low-capital revamps to improve profitability.
Mr. Golden previously worked as a refinery 2. Golden, Scott W., Integrating
process engineer and distillation system theoretical and practical aspects of
troubleshooter. He holds a BS in chemical refinery FCCU fractionator revamps,
engineering and is a registered professional 1989 AIChE Spring National Meeting,
engineer in Texas. He is the author of more Houston, Apr. 2, 1989.
than 75 papers on refinery unit revamps and
troubleshooting. 3. Lieberman, Norman P., and
Lieberman, Elizabeth T., Design,
installation pitfalls appear in vac
*Karl D. Schmidt is process engineer for
tower retrofit, Oil and Gas Journal,
Lyondell Petrochemical Co., Houston, Texas,
Aug. 26, 1991, p. 57.
and is currently responsible for the fluid cat-
alytic cracking unit. He has had previous
4. Golden, Scott W., and Costanzo,
experience in the aromatics recovery and
Stefano, Commercial Performance
sulfur recovery areas. Mr. Schmidt has a BS
Data for Structured Packing in
in chemical engineering from the University
Refinery Atmospheric Crude
of Texas at Austin.
Columns, AIChE 1990 Spring
National Meeting, Orlando, Fla.,
Gary R. Martin is a chemical engineer with March 1990.
Process Consulting Services, Inc., Bedford,
Texas. Mr. Martin specializes in improving 5. Lieberman, Norman P., and
refinery profitability by troubleshooting, Lieberman, Elizabeth T., Inadequate
optimizing and revamping refinery units. He inspection cause of flawed vac
previously worked as a refinery process engi- tower revamp, Oil and Gas Journal,
neer and distillation system troubleshooter. Sept. 14, 1992, p. 33.
Mr. Martin holds a BS degree in chemical
engineering from Oklahoma State 6. Golden, Scott W., and Sloley,
University. He is the author of more than 40 Andrew W., Simple methods solve
revamp and troubleshooting technical vacuum column problems using
papers. plant data, Oil and Gas Journal,
Sept. 14, 1992, p. 74.
* When this article was reprinted
7. Kister, Henry Z., Distillation Operations,
Mr. Schmidt was Manager of the Business
McGraw Hill, New York, 1990.
Optimization Group at CITGO Petroleum
Corporation’s Lake Charles Refinery.

Process Consulting Services, Inc.

3400 Bissonnet
Suite 130
Houston, Texas 77005
U.S.A
☎ [1]-(713)-665-7046
✉ [1]-(713)-665-7246
✉ info@revamps.com
✉ www.revamps.com