Balanced distillation equipment design

Fouling resistance and efficiency requirements for distillation equipment are

balanced and optimised for reliable unit performance

Soun Ho Lee GTC Technology

F
ouling tendency is a critical issue in crude downgrade column performance and reduce unit
distillation units and should not be over- run length.
looked when designing crude distillation col- This article will discuss common fouling issues
umns. Corrosion tendency can influence fouling associated with crude distillation column design.
issues as well. Since fouling resistance has an Actual retrofits for crude atmospheric columns
inverse relationship to efficiency in distillation are demonstrated through two case studies.
equipment design, optimising equipment design These studies examine how fouling resistance
between fouling resistance and efficiency require- and efficiency requirements for distillation equip-
ments must be precise. Poor application know- ment are balanced and optimised through careful
how as well as poor equipment design often evaluation and design methodologies.

Off gas

Crude
atmospheric
column Unstabilised
naphtha

Steam

Heavy Light kerosene

kerosene
pumparound
Kerosene

Steam

Heavy kerosene
Light diesel
pumparound

Steam

Preflash Light diesel

drum Heavy diesel
pumparound
Desalted
Diesel
crude
Steam

Heavy diesel
Steam
Atmospheric
residue

Figure 1 Case study 1: crude distillation unit configuration

www.digitalrefining.com/article/1001334 PTQ Q1 2017 1

 was designed without a top
pumparound circuit in order to
maximise fractionation between
unstabilised naphtha and kero-
sene at a given column height.1
This crude distillation unit
faced two problems: fouling
and corrosion of the distilla-
tion equipment in the unit were
found during a turnaround
inspection. Valve perforation
hole wearing and corrosion were
found in the trays for naph-
Figure 2 Case study 1: valve/perforation hole wearing and corrosion tha/kerosene fractionation.
Tray fouling was also identified
in the trays for the light kero-
sene/heavy kerosene and heavy
kerosene/light diesel fraction-
ation sections. Kerosene or die-
sel intermediate product yield
limitation was also experienced
when kerosene or diesel boil-
ing range material composition
was increased in the charged
crude slate. Charge crude com-
positions were frequently varied
during operation.

Case study 1: root cause

identification
Figure 3 Case study 1: underside view of fouled trays Figure 2 shows that the naph-
tha/kerosene fractionation trays
Case study 1: crude distillation unit suffered from valve/perforation hole wearing
description and background and corrosion. Some movable valve units dis-
The configuration of the crude distillation unit lodged from the tray deck. Perforation hole sizes
in this case is illustrated in Figure 1. Fractionated on the tray deck were increased by wearing and
light and heavy kerosene streams through the corrosion actions.2 Low column top temperature
crude atmospheric column and the side strippers required for target operation could accelerate
are combined and rundown as a single kerosene hydrochloric acid corrosion and valve/perfora-
intermediate product stream. A diesel interme- tion hole wearing. Significant fractionation effi-
diate product stream is also formed from a com- ciency loss between naphtha and kerosene was
bination of light diesel and heavy diesel streams. not recognised during the operation. The bulky
These crude atmospheric columns and side fractionation nature of crude distillation ser-
strippers were originally designed with conven- vice might result in fractionation efficiency being
tional movable valve trays, traditionally selected insensitive to tray weeping. However, if this
in the past. The exception was the wash sec- valve/perforation hole wearing progresses, signif-
tion which was arranged with structured pack- icant fractionation efficiency loss will be noticed
ing. Three pumparound circuits are arranged at through substantial weeping.
the heavy kerosene, light diesel and heavy diesel Fouled trays located for the light kerosene/
range material locations. The naphtha/kerosene heavy kerosene and heavy kerosene/light die-
fractionation section is positioned as the crude sel fractionation sections are shown in Figure
atmospheric column top section. This column 3. A tar-like substance was discovered around

2 PTQ Q1 2017 www.digitalrefining.com/article/1001334

the periphery of the valve legs.
Phosphates used for crude oil Crude
FT atmospheric
production were suspected as column
FT
the root cause. Boiled phos-
phates may react with kerosene
boiling range material and make LC
fouling deposits.
A dedicated process evalua- Pumparound Side
stripper
tion for kerosene or diesel yield
limitation was conducted. The LT Steam
original column and tray draw-
ing revealed that intermediate Rundown
side product and pumparound product

streams were withdrawn from

fractionating trays directly. The Figure 4 Case study 1: original design side draw configuration
originally designed side draw
configuration is illustrated in Figure 4. Flow As described earlier, light and heavy kero-
from the crude atmospheric column to the side sene streams are combined and rundown as a
stripper relies on gravity flow. If the liquid head single kerosene intermediate product stream.
formed on the collector tray is not high enough Therefore, fractionation performance between
to overcome total friction losses from the crude light kerosene and heavy kerosene streams is
atmospheric column to the side stripper, the not critical. The same rundown configuration of
flow rate can be limited. Moreover, frothy liquid light and heavy diesel streams does not necessi-
withdrawn from the fractionating tray’s active tate sharp fractionation between the two streams.
area can contain vapour. The presence of vapour However, fractionation performance between
can limit this gravity flow. Rigorous pipe line heavy kerosene and light diesel streams affects
hydraulic evaluation revealed that the gravity line rundown kerosene and diesel intermediate
hydraulics could be limited at a maximum target product qualities including the kerosene freez-
draw rate.3 ing point, one of the key specifications for ker-
osene rundown. To predict kerosene freezing
Case study 1: equipment modification point change, dedicated sensitivity analysis was
Based on the aforementioned process evalua- conducted.
tion and root cause analysis, the original mova- The aforementioned high performance tray
ble valve trays were replaced by fixed valve trays. implementation could improve individual tray
This tray type conversion improved equipment efficiency. Nevertheless, extra individual tray effi-
resistance against fouling and valve/perforation ciency improvement was not counted to predict
hole wearing. the retrofit heavy kerosene/light diesel fractiona-
The original fractionating trays at draw loca- tion performance.
tions were converted to chimney trays to increase
the liquid head for gravity flow. This chimney Case study 1: sensitivity analysis
tray conversion also eliminates the chance of For reliable sensitivity analysis, simulation mod-
yield loss and start-up trouble through fixed valve elling was first validated with pertinent unit test
tray implementation and increases draw liquid run conditions. Simulated kerosene freezing point
residence time for vapour disengagement from value was reasonably matched to actual value. The
liquid. However, this conversion resulted in los- tray efficiency and internal vapour/liquid traffic
ing one tray for each fractionation section: light profile for each fractionation section were quan-
kerosene/heavy kerosene, heavy kerosene/light tified through model validation. A constructed
diesel and light diesel/heavy diesel fractiona- kerosene freezing point sensitivity curve per var-
tion. GT-Optim high performance trays with var- ied theoretical stages is plotted in Figure 5. This
ious performance-enhancing features and fixed curve predicted that the freezing point could be
valves were implemented for the rest of the frac- increased by 0.1°F by using a chimney tray con-
tionating trays. version scenario. Figure 6 also shows another ker-

www.digitalrefining.com/article/1001334 PTQ Q1 2017 3

 osene freezing point sensitivity
0.7
curve per heavy kerosene/light

Freezing-point reduction, ƼF

0.6 diesel fractionation section inter-
0.1ºF nal reflux L/V (liquid/vapour)
0.5
ratio. A freezing point increment
0.4 of 0.1°F was predicted at a 3%
Pre-retrofit
0.3
test run lower heavy kerosene/light die-
Retrofit design sel fractionation section inter-
prediction
0.2 nal reflux L/V ratio. Undetected
0.1
kerosene freezing point changes
were anticipated through the sen-
0 sitivity analysis.
Number of theoretical stages gain ∆ for
heavy kerosene − light diesel fractionation section
Case study 1: performance
Figure 5 Case study 1: pre-retrofit sensitivity analysis summary
The pre- and post-retrofit per-
1.4
formances are summarised and
compared in Table 1. Both ret-
Freezing point reduction, ƼF

1.2 rofit test run conditions were

1.0 obtained through the same
operating mode, SOR (start
0.8 0.1ºF
of run) for fair comparison.
3.0% Since the internal vapour/liq-
0.6
Pre-retrofit
test run uid traffic and heat balances of
0.4
Retrofit design the crude atmospheric column
prediction were not measurable operating
0.2
parameters, these values were
0
0 10 20 30 40 50 60 70 80 90 quantified through simulation
Top simulated reflux L/V ∆ for modelling. Like the simulation
heavy kerosene − light diesel fractionation section, % model for the pre-retrofit case,
the simulation model for the
Figure 6 Case study 1: pre-retrofit sensitivity analysis post-retrofit case was also vali-
dated with selected post-retro-
Case study 1: performance summary fit test run conditions. The crude
charge rate was increased dur-
Case Pre-retrofit Post-retrofit
Parameter Test run Test run
ing the post-retrofit test run. The
Yield balance product yield balance reveals
Crude charge, BPD Base +∆12 that the post-retrofit test run
Unstabilised naphtha, LV% Base +∆6.0
Kerosene, LV% Base +∆19.5 crude slate contained more ker-
Diesel, LV% Base +∆12.1 osene boiling range materials
Fractionation performance
Light kerosene 5% - naphtha 95%,1 °F Base +∆1.0 compared to pre-retrofit crude
Light diesel 5% - heavy kerosene 95%,1 °F Base -∆1.0 slate. Kerosene and diesel yield
Kerosene flash point, °F Base ∆0
Kerosene freezing point, °F Base +∆0.2
limitation experienced in the
Heavy kerosene/light diesel internal reflux L/V, weight basis % Base
2,3
-∆15 past was eliminated.
Heat balance The simulated crude atmospheric
Heat removal - overhead condenser, % of total BTU/hr
2
72 62
Heat removal - heavy kerosene pumparound,2 % of total BTU/hr 4 6 column heat balance of the post-
Heat removal - light diesel pumparound,2 % of total BTU/hr 7 10 retrofit case was also shifted
Heat removal - heavy diesel pumparound,2 % of total BTU/hr 16 21
from that of the pre-retrofit case
1. ASTM D86 (LV%) due to a change in the crude
2. Simulated value
3. At the top tray of the section slate composition. Laboratory
test results showed that the
Table 1 post-retrofit kerosene freez-

4 PTQ Q3 2006 www.digitalrefining.com/article/1000133

ing point was relaxed by 0.2°F.
Meanwhile, simulation model- 1.4
ling showed that the post-retrofit

Freezing point reduction, ƼF

1.2
internal reflux L/V ratio for the Post-retrofit
heavy kerosene/light diesel frac- test run
1.0 15%
tionation was reduced compared
0.8 0.2ºF
to the pre-retrofit value.
The heavy kerosene/light die- 0.6
8%
sel fractionation section perfor-
0.4
mance through the post-retrofit Retrofit design Pre-retrofit
prediction test run
test run was evaluated and com- 0.2
pared to the pre-retrofit section
0
performance. 0 10 20 30 40 50 60 70 80 90
The post-retrofit kerosene Top simulated reflux L/V ∆ for
freezing point was plotted and heavy kerosene − light diesel fractionation section, %
compared to the pre-retro-
fit sensitivity curve in Figure 7.
Relaxing the kerosene freez- Figure 7 Case study 1: pre- and post-retrofit sensitivity analysis
ing point by 0.2°F predicted an
internal reflux L/V ratio reduc- Case study 1: fractionation efficiency comparison
tion of 8% on the pre-retrofit
sensitivity curve. A lower inter- Heavy kerosene/light diesel Pre-retrofit Retrofit Post-retrofit
Fractionation section Test run Prediction Test run
nal reflux ratio of 15% was sim- Kerosene freezing point, °F Base +∆0.1 +∆0.2
ulated with the same 0.2°F Internal reflux L/V,1,2 weight basis % Base - ∆3 - ∆15
freezing point relaxation at the Section efficiency, % 1,3
55 50 65

post-retrofit test run conditions. 1. Simulated value

2. At the top tray of the section
Results indicated that the actual 3. Overall efficiency (number of theoretical stages/number of fractionating trays)
post-retrofit heavy kerosene/
light diesel fractionation section Table 2
efficiency was more satisfactory
than the predicted retrofit efficiency value. fractionator excluding the bottom stripping
Table 2 shows simulated heavy kerosene/light section.
diesel fractionation section efficiencies. As one Unstabilised naphtha, kerosene and atmos-
fractionating tray was converted to a chimney pheric gas oil intermediate products were dis-
tray, a lower number of theoretical stages were tillated through the crude atmospheric column
counted for the retrofit performance prediction. and the side strippers. The unstabilised naphtha
Nevertheless, the simulated theoretical stage stream was further separated into LPG (liquefied
count through the post-retrofit test run condi- petroleum gas), light naphtha, and heavy naph-
tion data was maintained at the same level as the tha through the naphtha stabiliser and naphtha
pre-retrofit theoretical stage count, resulting in splitter. The heavy naphtha stream was directed
improved fractionation efficiency between heavy to the reforming unit for aromatic component
kerosene and light diesel. The aforementioned production. The unit utilised two pumparound
GT-Optim high performance tray implementa- circuits as well as an overhead condenser for
tion contributed to the fractionation efficiency heat removal. One circuit was positioned as the
improvement. top pumparound. The other circuit was located
between kerosene and atmospheric gas oil boiling
Case study 2: crude distillation unit description range materials. A second packed bed at the top
and background of the crude atmospheric column had the func-
The second case study also examines a crude dis- tion of fractionating naphtha and kerosene boil-
tillation unit. Figure 8 illustrates the schematic of ing range materials.
the unit. The crude atmospheric column in this A particular unit limitation the refiner faced
case was designed as a fully structured packed was that fractionation performance between

www.digitalrefining.com/article/1001334 PTQ Q1 2017 5

 Off gas
Crude LPG
atmospheric
column

Top Light
naphtha
pumparound

Naphtha/kerosene
fractionation section
Heavy
naphtha

Steam

Bottom Kerosene
pumparound

Preflash
drum
Desalted Steam
crude Atmospheric
gas oil

Steam
Atmospheric
residue

Figure 8 Case study 2: crude distillation unit configuration

naphtha and kerosene was substantially down- migrate to the naphtha/kerosene fractionation
graded after a four-month operation. Substantial bed.
amounts of kerosene boiling range materials Findings showed that the column’s inside
were downgraded to the naphtha stream. This wall cladding using Monel metallurgy was only
downgrading not only limited the kerosene yield applied to the portion where the top pumparound
but also influenced the downstream reform- section was positioned. The column inside wall
ing unit performance. The high rear end distilla- portion of the naphtha/kerosene fractionation
tion point of the heavy naphtha stream adversely section remained as carbon steel. Rusted wall
affected the reforming reactor catalyst activation. pieces from corrosion could accelerate fouling.
A review of the original distillation equipment
Case study 2: root-cause identification drawing revealed that the gravity flow trough-
Inspection during unit turnaround showed that type liquid distributor for the naphtha/kero-
the trough-type liquid distributor for the naph- sene fractionation section was designed with
tha/kerosene fractionation section was fouled. high drip point density and small drip hole size.
Several root causes of the fouling were identified The naphtha/kerosene fractionation packed bed
through rigorous evaluation. was equipped with structured packing with a 1in
Amine-based corrosion inhibitor was injected crimp size and a 45° inclination angle.
into the crude distillation unit. Chloride present This packing size at the given bed height was
in the column overhead may react with the inhib- suitable to achieve target fractionation efficiency.
itor and form ammonium salt, which can foul the However, the liquid distributor drip point density
distributor. Formed ammonium salt particles originally selected was excessive for the 1in crimp
could reside in the top pumparound bed and also size.

6 PTQ Q1 2017 www.digitalrefining.com/article/1001334

 The basic equation used to size
Case study 2: distributor modification summary
gravity liquid distributors is:
2 Naphtha/kerosene fractionation Original Modification
⎛ Lv ⎞
H = ⎜ ⎟ Distributor drip point density, drip point/ft2 16 4.6
⎝ k × N × HA ⎠ Distributor operating range
Drip hole elevation,1 inch
2.5:1
1.5
1.7:1
3
Distributor drip hole diameter, inch 3/16 7/16
H = Liquid height (‘Head’) above
1. From the bottom of the trough
round shaped hole
Lv = Liquid volumetric flow
N = Number of drip holes Table 3
HA = Hole area
K = Orifice coefficient
A minimum liquid head needs 330
Post-retrofit
to be maintained to ensure uni- 320

Temperature, ºF
Pre-retrofit
form liquid distribution. A cer- 310
tain number of drip holes, which
300
indicates ‘drip point density’, is
required for the desired distri- 290

bution quality. However, unnec- 280

essarily high drip point density 270
reduces distributor drip hole size 0 50 100 150 200 250 300 350 400
and increases a chance of foul- Unit run length, days
ing. Distributor operating range
affects hole size because the liq- Figure 9 Case study 2: performance trend – naphtha end point
uid head should be maintained at
the minimum rate for uniform liquid distribution. unit operation in the entire charge range and more
The original liquid distributor (as designed) was efficient overall unit economics can be achieved.
not properly optimised between fouling resistance Distributor drip hole elevation from the bottom of
and liquid distribution quality. the trough was increased to slow down distributor
fouling.
Case study 2: distributor modification The measured naphtha end point and kerosene
The liquid distributor modifications for the naph- flash point trends are plotted in Figure 9 and Figure
tha/kerosene fractionation section are sum- 10 respectively. Plots in red indicate values gath-
marised in Table 3. In order to enlarge liquid ered during pre-modification operating periods
distributor drip hole size, the drip point density while plots in blue represent values achieved after
was reduced in a new design. The new density was the modification. Stable naphtha end points and
carefully selected by considering the commercially kerosene flash points were maintained for more
proven drip point density in the given size pack- than eight months of operation.
ing and application. Distributor operating range These case studies show how fouling resistance
was also adjusted further to increase distributor and efficiency requirements for distillation equip-
drip hole size. The minimum end of the distribu- ment are balanced and optimised for reliable
tor operating range was increased. This adjusted crude distillation unit performance.
distributor operating range does not reduce the
unit operating range. The minimum rate of the liq- This article is an updated version of a presentation given at
uid distributor does not have to be matched to the AIChE’s Spring Meeting Distillation Symposium, 11-14 Apr 2016,
minimum unit charge rate.4 Heat balance shift- in Houston, TX.
ing through pumparound adjustments or increas-
GT-OPTIM is a mark of GTC Technology US LLC.
ing furnace coil outlet temperature can maintain
the required minimum distributor rate during
lower unit charge rate operation. This strategy can References
increase energy consumption during minimum 1 Lee S H, et al, Optimising crude unit design, PTQ, Q2 2009.
charge rate operation. But it can assure efficient 2 Kister H Z, Distillation Operation, McGraw-Hill Company, 1990.

www.digitalrefining.com/article/1001334 PTQ Q1 2017 7

 3 Libermann N P, Process Design for Reliable Operations, Gulf
Publishing Company, 2nd Edition. LINKS
4 Bonilla J A, Don’t neglect liquid distributors, Chemical
More articles from: GTC Technology
Engineering Progress, Mar 1993.
More articles from the following categories:
Soun Ho Lee is Manager of Refining Application for GTC Technology Corrosion and Fouling Control
US, LLC, in Euless, Texas, specialising in process design, Crude and Vacuum Units Mass Transfer
simulation modelling, energy saving design and troubleshooting
for refining and aromatic applications.
Email: [email protected]

8 PTQ Q1 2017 www.digitalrefining.com/article/1001334