Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Vacuum Unit Design For High Metals Crudes

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Vacuum unit design for high

metals crudes
Vacuum unit design can influence vacuum gas yield, product quality and run
length. A wet vacuum unit with residue stripping achieves the highest cutpoint
Scott W Golden, Tony Barletta and Steve White Process Consulting Services

everal coker projects are in the planning


phase or being built to process heavy sour
crudes from Canada, Mexico, Venezuela
and the deepwater Gulf of Mexico. These projects
include new or revamped vacuum units.
Although large heavy sour-sweet crude price
differentials are the major economic driver
today, vacuum gas oil (VGO) yield also influences refinery profitability. Feeding VGO to a
coker reduces refinery liquid yields because it
converts 1015% of this material into coke, with
the total value of the remaining coker products
worth considerably less than the VGO feed.
Processing recoverable VGO through the coker
may be necessary due to poor vacuum unit
design, but the assumption that VGO flows
through the coker unchanged is wrong.
Increasing the vacuum unit cutpoint from
1000F (538C) or less to 1050F (566C)
(Figure 1) or higher increases the
refinery liquid volume yield by at
least 0.60.7 vol%, depending on

conversion unit performance.

4"0TEMPERATURE o&

Previous heavy oil projects

to the mid-1980s, refiners outside the producing


countries had little experience of processing
these crudes in high blend percentages.
Consequently, few process engineers had any
experience of designing vacuum units for these
heavy and extra-heavy crudes. Furthermore, the
gas oil cutpoints for the producing countries
vacuum units were less than 1000F (538C),
and whole crude 10001150F (538621C) TBP
boiling-range metals distribution data were
generally not available. Hence, volatile nickel
and vanadiums influence on gas oil product
quality was poorly understood. With all these
unknowns, early heavy crude projects had problems meeting their design goals other than the
crude charge rate.
Low product yield, high metals gas oil product
and short heater run lengths were common. At
least one project has been operating with a gas


In the last 20 years, grassroots and
 o&CUTPOINT
revamped crude unit heavy oil

projects have been designed for
 o&CUTPOINT
Venezuelan heavy and extra-heavy

Maya crudes. Canadian heavy sour
has been processed by some refin
ers, but few have consistently run
high blend percentages. The few

that did operated low cutpoint


vacuum units, feeding 68 vol% of
$ISTILLED VOL
whole crude gas oil boiling-range
material to their coker unit. Prior Figure 1 TBP cutpoint Maya crude

www.digitalrefining.com/article/1000116





PTQ Q1 2007 1

Fired
heater

Fired
heater

Wash zone

Wash zone

Reduced
crude

Reduced
crude

Flash
zone

Flash
zone
Coil
steam

Vacuum
residue

Vacuum
residue

Figure 2 Dry unit no steam

Figure 3 Wet with coil steam only

oil product cutpoint 100F (56C) lower than


design since startup at a loss of 6 vol% in gas oil
yield, resulting in increased coke production and
lower refinery liquid volume yields. Another unit
was designed for a heater outlet temperature of
800F (427C), but actual temperature had to be
reduced to 775F (404C) to increase the run
length to two years between decoking. Nearly
every vacuum column operating above a 730
740F (388-393C) flash zone temperature has
coked the wash section packing in less than a
four-year run. In several instances, the run
length has been less than a year. The lessons
learned from these earlier projects should be
incorporated into new projects.
Unfortunately, many heavy oil projects today
are focusing mainly on the large heavy soursweet price differentials, with minimal concern
for refinery liquid yields. Yet, during most of the

past 20 years, the heavy sour-light crude price


differential has been much lower, making refinery liquid yields a major factor in refinery
profitability. Achieving a high gas oil product
yield on heavy crudes requires the right type of
vacuum unit and correctly designed the equipment. Low-cost dry-vacuum unit designs that
work well for Arab Light or other easy- to-refine
crudes should not be used for heavy crudes. The
refiners decision to use a dry (Figure 2), wet
with no stripping (Figure 3) or wet with stripping (Figure 4) vacuum unit will determine
whether these future heavy oil projects meet all
their goals. Fundamentally sound equipment
design is critical, because deep-cut vacuum units
must operate at a high temperature throughout
the run to achieve yield. In todays environment,
operators are requiring four- to six-year crude
runs.

Fired
heater

Design challenges
Wash zone

Reduced
crude
Flash
zone
Coil
steam
Steam

Stripping
trays

Vacuum
residue

Figure 4 Wet with coil and stripping

2 PTQ Q1 2007

Heavy crudes are difficult to vapourise, contain


large amounts of volatile metals and their properties vary. Many conventional heavy and oil
sands-based crudes are blends of condensate or
upgrader products and very heavy conventional
crude or bitumen, or they consist of several
crudes such as Western Canadian Select (WCS).
Evaluating critical operating variables over a
range of feedstock qualities avoids surprises
when the unit begins operation. Capital investment needs to be balanced against the reduced
profits from a short run length, low product yield
or high metal in the product.
Limited crude assay information is common
for some Canadian heavy sour Dilbit or Synbit
crudes that have limited or no current production. A few Canadian oils sands producers

www.digitalrefining.com/article/1000116

4"0TEMPERATURE o&

currently operate vacuum units


with cutpoints of around 900F

(482C), but refiners have little

experience of processing these

crudes at high blend percent
ages.
Without
commercial
4"0

experience, laboratory data
4"0

have to be used to evaluate
crude properties. For instance,

cracking in the laboratory at

low temperatures may not












reflect the true cracking
$ISTILLED VOL
tendencies in the vacuum
heater, but these data do show
directionally important stabil- Figure 5 Maya crude assay two TBP distillation curves
ity
differences
between
conventional heavy crude oils and some of the project over its lifetime. The wrong decision
oils sands-based crudes such as Cold Lake. results in low gas oil product yield, high metals,
Ultimately, the process designer will need to microcarbon and short run length.
consider this information when selecting the
Dry units use no steam in the heater or
type of vacuum unit or equipment design to meet column. Dry designs need to be operated at a
the realities of crude variability, thermal stability 1520 mmHg flash zone pressure to achieve
problems or other peculiarities of these new reasonable gas oil product cutpoints. Even so,
crudes.
maximum cutpoints are approximately 1000F
(538C), because a well-designed dry heater is
Vacuum unit type
limited to a 760785F (404418C) outlet
The three broad categories of vacuum unit are temperature without periodic decokings. Higher
dry, wet (damp) without stripping and wet with rates of coking occur because the oil residence
stripping. Dry units are the least expensive, but time is 6090 seconds or more in the radiant
they are the most prone to short heater run section. Wet units using coil steam lower the oil
lengths or low heater outlet temperatures, result- residence time to 2030 seconds, depending on
ing in low gas oil product yields. Wet units with the amount of steam and radiant section surface
coil steam only allow higher heater outlet area. Double-fired designs with 1015 seconds of
temperatures, while maintaining four- to six-year oil residence time should be used with oil sandsrun lengths (Figure 2) and a somewhat higher based crude oils.
gas oil yield than a dry unit, but they are more
Of the three types, wet vacuum units with resicostly. Wet units with stripping produce the best due stripping produce the highest gas oil
quality gas oil at a given TBP cutpoint and cutpoint, the lowest metals gas oil product and
permit the lowest heater outlet temperature for a the lowest heater outlet temperature. Product
given gas oil cutpoint. Wet designs with residue metals are lower at a given cutpoint because the
stripping produce the highest revenue by far, but distillation 95%-to-endpoint (EP) tail is lower. A
they are also the most costly to build. Coil and well-designed wet vacuum unit with stripping
stripping steam must be optimised according to will produce VGO with 5070 % of the metals of
capital constraints, product values and operating a wet unit without stripping. When processing
costs. Once vacuum unit type is selected, equip- Maya crude, this can be the difference between
ment design considerations determine the 7.010.0 wt ppm metals in the gas oil product
product yield, product quality and run length. and 3.07.0 wt ppm at the same cutpoint. This
For example, unstable crudes require coil steam has a significant influence on the cost of the
to reduce the oil residence time and double-fired downstream hydrotreater/FCC complex or
heaters to minimise the oil film temperature.
hydro-cracker. Dry vacuum units should only be
The type of vacuum unit is the single most used if capital cost is the sole factor driving
critical choice that will influence a heavy oil selection.

www.digitalrefining.com/article/1000116

PTQ Q1 2007 3

Maya crude vacuum gas oil product yield


Gas oil TBP
cutpoint F (C)
975 (524)
1000 (538)
1025 (552)
1050 (566)
1075 (579)
1100 (593)

Vol% yield on
whole crude
23.7
25.1
26.8
28.3
29.8
31.2

Yield from
975F (524C)

1.4
3.0
4.5
6.0
7.4

Table 1
Maya crude metals and microcarbon distribution
Gas oil TBP
Vanadium, Nickel, Microcarbon,
cutpoint F (C)
wt ppm wt ppm
wt%
9501000 (510538)
1.2
1.3
2.9
10001050 (538566) 4.2
3.8
4.4
10501100 (566593) 22
25
6.0
11001150 (593621) 66
50
7.4

Table 2

Maya crude: yield and quality


Maya crude is used here to present design
concepts, since it is a common heavy crude (21.5
API gravity) containing large amounts of volatile
nickel and vanadium. Total crude unit gas oil
metals will be 3.010.0 wppm, depending on the
product cutpoint. Prior to beginning the design,
accurate assay TBP distillation and metals distribution curves are needed. Vacuum units target
product cutpoints in the 10001100F TBP
cutpoint range. Therefore, it is essential that the
assay TBP distillation is correct in this region of
the curve, but often it is not. Figure 5 shows the
10001100F portions of the TBP curves from
two Maya assays. For the same vacuum unit
flash zone conditions, the flatter slope yields
approximately 2.0 vol% more VGO than the
upper curve. Moreover, volatile nickel and vanadium are highly non-linear in the 10001100F
TBP range. These data are much more costly to
generate and subject to large errors if not done
correctly. In spite of these difficulties, it is necessary to generate accurate metals distribution
data to 1100F (593C) or higher, depending on
the targeted gas oil product cutpoint. The downstream hydrotreating or hydrocracking unit
designs, including reactor volume and catalyst
life, depend on accurate gas oil product metals
prediction.

4 PTQ Q1 2007

The volumes of recoverable gas oil are determined by the TBP distillation curve, while the
metals distribution, vacuum unit type and equipment design dictate the gas oil product quality.
Equipment design determines the run length.
Table 1 shows the 650F (344C)-to-cutpoint gas
oil yields based on the upper curve previously
shown in Figure 4. Maya whole crude has a TBP
slope of approximately 17F (9.5C) TBP per
vol% in the 10001100F TBP range, which is
typical of many heavy crude oils, whereas other
crudes such as tar sands bitumen blends may
have a flatter TBP slope with more recoverable
gas oil in the crude oil.
Refinery liquid volume yields and economics
are dependent on vacuum unit design. Feeding
VGO to a coker increases the coke make per
barrel of crude oil and reduces the conversion
units liquid volume yields. An optimised vacuum
unit design can increase the vacuum gas yield by
4.06.0% on whole crude versus a poorly
designed unit. As gas oil yield increases, so does
the nickel, vanadium and microcarbon in the gas
oil product. The amount depends on distribution, vacuum unit type and equipment design.
The VGO products TBP cutpoint is a measure
of the amount of gas oil product yield, but the
metals depend on more than yield alone. This is
because metals distribution is highly non-linear
and changes rapidly above 1000F (538C)
reducing the 95 vol%-to-EP while maintaining
the yield lowers its metal content. Table 2 shows
the vanadium, nickel and microcarbon distribution in Maya.
A wet vacuum unit with residue stripping
produces a lower metals gas oil product because
the stripping section vapourises the lower boiling-range material compared to the same volume
vapourised in the heater. Another factor is the
wash sections packing efficiency. Five feet of
grid is approximately 0.5 stages, whereas 4.0
5.0 feet of structured packing yields the
equivalent of one stage. Increasing the wash
zones efficiency from 0.5 to 1.0 stages decreases
metals by 30% at the same cutpoint. Combining
residue stripping and higher wash zone efficiency
dramatically reduces product metals.

Gas oil product yield


Gas oil product yield is controlled by three operating variables for a given vacuum unit feed
distillation. The columns flash zone temperature

www.digitalrefining.com/article/1000116

and pressure, as well as residue stripping efficiency (Figure 6) determine yield. Maximum
flash zone temperature is dependent on oil
stability, plus heater oil residence time and oil
film temperature. A well-designed heater with an
equal heat flux per pass and a low oil residence
time can operate at higher film temperatures,
and a double-fired heater can further reduce the
film temperature, allowing a higher heater outlet
temperature before the rate of thermal cracking
becomes excessive. Rapid thermal cracking
increases the rate of coke laydown, eventually
requiring decoking even if the ejector systems
are designed to handle high cracked gas production. The column flash zone pressure depends on
the first-stage ejector suction pressure and
column pressure drop. Lower pressure raises the
gas oil yield at the expense of ejector system
capital and utilities cost. Dry vacuum units have
low ejector loads, so the minimum column operating pressure is limited only by the ejectors
technical constraints, whereas the minimum
ejector suction pressure for a wet unit is a matter
of economics.
Heater coil and residue stripping steam dramatically increase the ejector load, requiring a higher
operating pressure to keep the capital and operating costs reasonable. Residue stripping increases
the gas oil yield through lower oil partial pressure
on the trays. Wet units with residue stripping
generally have the highest first-stage ejector
system loads of any vacuum unit type because
they use the most steam. As coil and stripping
steam rates increase, the optimum first-stage ejector suction pressure also increases.
Vacuum column flash zone pressure should be
optimised according to product values and utilities costs. A lower pressure increases yields, but
raises capital and operating costs. Dry vacuum
units can be designed to operate economically at
first-stage ejector suction pressures as low as 4.0
mmHg absolute. The minimum first-stage ejector suction pressure is limited by the maximum
ejector compression ratio of approximately 15 to
1. The motive and process steam must be
condensed against cooling water, which sets the
minimum first-stage ejector discharge pressure
at 6065 mmHg absolute with a cooling water
temperature of 8590F (2932C). Dry vacuum
unit ejector loads are typically low, consisting of
cracked gas, air leakage and condensable
hydrocarbons.

www.digitalrefining.com/article/1000116

Pressure,
mmHg absolute
Reduced
crude
Flash
zone

Steam

Temperature,
F
Stripping
trays

Vacuum
residue

Figure 6 Vacuum column flash zone and residue


stripping

Poor atmospheric column stripping section


performance sometimes significantly increases
the condensable load. Wet units include additional loads from coil and stripping steam,
requiring much more motive steam consumption. Hence, these systems have large first-stage
ejectors and inter-condensers, as well as higher
cooling water rates. Wet units are generally
designed for a first-stage ejector suction pressure
of 1025 mmHg absolute. This pressure depends
on the targeted gas oil cutpoint and oil thermal
stability. Poor oil thermal stability, such as oil
sands-based crudes, requires a lower flash zone
pressure to minimise the heater outlet temperature to avoid rapid coking.
The column flash zone temperature is controlled by the heater outlet temperature and transfer
line pressure drop. A low transfer line pressure
drop reduces the temperature drop from the
heater outlet to the flash zone, raising its temperature. The maximum heater outlet temperature
depends on the rate of thermal cracking in the
heater and on crude oil thermal stability. For
instance, some of the oil sand bitumen-based
crude oils begin to generate significant amounts
of gas at a 2540F lower temperature than Maya
crude. Well-designed low residence time heaters
using coil steam can operate at a higher outlet
temperature because dry heater radiant section
oil residence times might be 6090 seconds
versus 1030 seconds when injecting coil steam.

PTQ Q1 2007 5

Maya crude atmospheric and vacuum residue metal


Stream
Nickel, wt ppm
Atmospheric residue
108
Vacuum residue
193

Vanadium, wt ppm
395
704

Table 3

The majority of vacuum units do not have residue stripping. Yet, stripping produces the lowest
metals content VGO for a given TBP cutpoint
and reduces the heater outlet temperature
needed to meet a given cutpoint at a given process steam load. For example, using a wet design
with 1.0 wt% coil steam without residue stripping requires a 40F (22C) higher heater outlet
temperature than the wet unit with residue stripping using the same total amount of coil plus
stripping steam. Vacuum units designed for
heavy crudes at high cutpoints should use residue stripping. Yet, new heavy oil vacuum units
are still being designed without stripping
sections. Worse still, a few use a dry design.

Gas oil product metals


Gas oil product total metals come from atmospheric and vacuum column gas oil products. The
source is both volatile metals in the product boiling range and entrained atmospheric and vacuum
residue. Volatile metals vapourise at the operating
conditions in the atmospheric and vacuum
column flash zones. Thus, there will always be
some present, no matter what type of vacuum
unit or equipment is used. But the amount of
volatile metals depends on the type of vacuum
unit and wash section efficiency. Surprisingly,
many atmospheric columns produce AGO products that are black from entrained atmospheric
residue, increasing the AGO product metals
contents to as high as 1015 wt ppm in some
cases. Poor atmospheric column flash zone, wash
zone and stripping section designs all contribute
to black AGO. Eliminating atmospheric column
entrainment reduces the AGO product volatile
metals to below the detection limits of the inductively
coupled
plasma
atomic-emission
spectrometer (ICP-AES). Entrained vacuum residue is more difficult to control because wash
section internals capacity limits can be exceeded.
And once this occurs, HVGO product metals can
increase to 2040 wt ppm.

PTQ Q1 2007

Maya and other heavy crude atmospheric and


vacuum residues contain large amounts of nickel
and vanadium (Table 3). Maya atmospheric residue contains over 500 wt ppm nickel and
vanadium. Worse still is the influence of
entrained vacuum residue, containing almost
900 wt ppm metals on the HVGO product.
Vacuum unit transfer line critical flow expansions, flash zone vapour horn and wash section
internals determine the amount of entrainment.
Poorly designed transfer lines with high pressure
drop critical flow expansions at the column inlet
nozzle generate fine mists that are difficult to
remove. Yet, entrainment can be almost eliminated through prudent transfer line and column
internal designs, as long as the packing capacity
factor does not exceed 0.40.45 ft/s.
The gas oil product cutpoint alone does not
determine the product metals content. Since
volatile metals are highly non-linear, small
changes in the 95 vol% - EP distillation tail have
a large effect on gas oil product metals. Process
and equipment designs that minimise the distillation tail will reduce metals. Units designed
with residue stripping and improved wash
section efficiency have a lower distillation tail
and reduced metals at a given product TBP
cutpoint. Residue stripping vapourises lower
boiling-range material for the same HVGO product cutpoint than the heater. Wash zone
efficiency must be balanced against the increased
likelihood of coking when the packed bed depth
or packing efficiency are excessive.

Process modelling
Process flow sheet model assumptions and structure are critical to predicting the VGO yield and
maintaining unit reliability. It is not unusual for
the process engineers model to over-predict gas
oil product yields by 2.04.0 vol% on whole
crude and underpredict the wash rate needed to
prevent coking by 300%. Most engineers assume
the transfer line liquid and vapour entering the
vacuum column flash zone are in equilibrium,
but field observation has proved this assumption
is incorrect. This results in large yield prediction
errors and causes many engineers to incorrectly
select a dry vacuum unit to meet a design 1050F
(566C) cutpoint when processing crudes such as
Maya.
Several refiners have found their new dry
vacuum units actually produced less gas oil and

www.digitalrefining.com/article/1000116

more coker feed by 3.04.0


Calculated gas oil TBP cutpoint, F (C)
vol% on whole crude than the
design. Unfortunately, to meet

Equilibrium model Non-equilibrium model _
the design cutpoint, increasing
HVGO TBP cutpoint, F (C)
1050 (566)
985 (530)
- 65 (36)
the heater outlet temperature
Heater outlet temperature, F (C) 773 (412)
773 (412)

Flash zone pressure, mmHg


15
15

to compensate is not practical


because it needs to be raised by
4055F (2231C) to offset Table 4
the modelling error. Over the
years, numerous field tests
Calculated heater outlet temperature, F (C)
have proved that another
model
structure
(Deepcut

Equilibrium model Non-equilibrium model _
vacuum unit design, PTQ, Q4
HVGO TBP cutpoint, F (C)
1050 (566)
1050 (566)

2005) is essential to selecting


Heater outlet temperature, F (C) 773 (412)
827 (442)
+54 (30)
Flash zone pressure, mmHg
15
15

the correct type of vacuum unit


and meeting design gas oil
yields. Additionally, poorly Table 5
structured models underpredict
wash oil by 200400%, causing many wash section packed
7ASH
RATE
beds to coke, some in less than
7ASH
SECTION
a year.
Based on design reviews of
4RANSFERLINE
VAPOUR
recent projects and discussions
&LASH
&URNACE
with process design engineers,
OUTLET
there continues to be contro&LASH
&LASH 4RANSFER
LINE
versy surrounding vacuum unit
process modelling. Using 100%
&LASH
%NTRAINMENT
/VERFLASH
Maya, both an equilibrium and
non-equilibrium model were
created for a dry vacuum unit
to illustrate the yield or heater
outlet temperature error that
3PLITTER
3TRIPPING
results from the modelling
SECTION
error. The equilibrium model
3TEAM
shows the calculated heater
6ACUUM
outlet temperature (Table 4)
RESIDUE
needed to meet a 1050F
(566C) TBP cutpoint for a
vacuum column flash zone Figure 7 Process model structure
pressure of 15 mmHg absolute
pressure is 773F (412C), assuming a heater achieve a 1050F (566C) TBP cutpoint, which is
outlet pressure of 5.0 psi (258 mmHg absolute). impossible because the dry heater outlet temperHowever, the non-equilibrium model for the ature is limited to approximately 785F (418C).
same feedstock calculates the product cutpoint The gas oil cutpoint is limited to about 1000F
closer to 985F (530C) or 4.0 vol% less gas oil (536C) with a dry vacuum unit. Dry vacuum
product on whole crude for the same heater units are low cost, but they also achieve a lower
outlet temperature and pressure. Once a dry unit gas oil product yield.
is built, increasing the cutpoint requires a higher
An improved method of modelling transfer
heater outlet temperature.
lines and vacuum columns based on field experiThe non-equilibrium model shows the required ence, which better predicts yields and other
heater outlet temperature is 827F (Table 5) to critical operating parameters such as the wash

www.digitalrefining.com/article/1000116

PTQ Q1 2007 7

Dry vacuum units

(EATEROUTLETTEMPERATURE o&

(EATEROUTLETTEMPERATURE o&

Dry vacuum units have the lowest


capital investment, because process
and ejector system steam consump
tion and cooling water requirements
are less than for the other two types

of unit, but the gas oil product yield

is also the lowest. The heater outlet
temperature for a well-designed dry
0RACTICALHEATERLIMITo&

heater is approximately 785F
(418C) on some stable crude oils,

whereas a poorly designed heater

may limit temperature to 760F






(404C) or less for the same crude.
4"0CUTPOINT o&
Figure 8 shows the non-equilibrium
model predicted heater outlet
Figure 8 Dry vacuum unit heater outlet temperatures and TBP cutpoint
temperature versus the cutpoint.
This design assumes a 5.0 psi (258
mmHg absolute) heater outlet pressure and 15 mmHg flash zone

pressure at each cutpoint. When
processing 100% Maya, the gas oil

product cutpoint is limited to
approximately 1000F (537C) at a

heater outlet temperature of 785F
0RACTICALHEATERLIMITo&

(418C).
The most common reason for

using dry units is that they use less
steam and cooling water. Ultimately,

the refinerys liquid volume yield

and utilities costs need to be






balanced to optimise refinery profit4"0CUTPOINT o&
ability. The economics of dry
columns look good on paper because
Figure 9 Damp coil steam only (0.5 wt% steam)
it is common to model the vacuum
unit incorrectly and overpredict the
rate, requires the model to be segmented into a dry vacuum units gas oil product yield.
number of operations prior to the vapour enter- Investments are lower at relatively high product
ing the column wash section (Figure 7). Using yields. When actual plant gas oil product yields
multiple unit operations allows the non-equilib- are used and actual heater decoking intervals are
rium nature of the system to be estimated. This factored into the economic calculations, dry units
technique, although still only an approximation, do not make sense.
more accurately predicts the gas oil product
yield, vacuum residue yields and better estimates Wet vacuum units: coil steam only
the wash flow rate needed to avoid coking the Vacuum units designed with coil steam to reduce
wash section packing. The operating pressure at the heater oil residence time and decrease the
the heater outlet and in the first large section of column flash zone oil partial pressure allow a
the transfer line must also be calculated using a higher gas oil product yield because a welltwo-phase model with critical flow-limiting algo- designed heater outlet can operate at
rithms. Both process and rigorous equipment temperatures up to 800805F (427430C)
system models must be used when designing without rapid heater coking. Higher outlet
vacuum units.
temperatures are possible compared with a dry


8 PTQ Q1 2007

www.digitalrefining.com/article/1000116

heater because the oil residence time is much


lower. A dry heater has a 6090 seconds radiant
section oil residence time. A wet, single-fired
heater designed at an oil mass flux of 400 lb/ft2sec, with an average radiant section heat flux of
10 000 Btu/hr-ft2 and 0.5 wt% steam-on-feed,
will have a radiant section residence time of 30
40 seconds. A double-fired heater with a similar
oil mass flux, an average radiant section heat
flux of 14 000 btu/hr-ft2 and 1.0 wt% steam-onfeed will have a radiant section residence time of
15 seconds. Ultimately, oil thermal stability, oil
residence time and oil film temperature will
determine the rate of thermal cracking, which
sets the maximum heater outlet temperature.
Wet designs have a higher capital investment.
Furthermore, coil steam increases the ejector
systems motive steam consumption and cooling
water requirements. The first-stage ejectors
motive steam consumption is five to ten times
higher than for a dry ejector. The process and
motive steam must be condensed by the intercondensers, so cooling water requirements are
dramatically higher. Utilities savings continue to
be the stated justification for using dry vacuum
units, but simply minimising utilities is not the
purpose of a vacuum unit. Gas oil product and
refinery liquid yields need to be part of the
economic equation.
The coil steam rate is optimised according to
the heater outlet temperature, column operating
pressure and ejector capital constraints. The rate
varies from 0.1 to over 1.0 wt% steam on feed. In
this example, to show the difference in gas oil
product cutpoint versus a dry unit, the steam
injection rate is set arbitrarily at 0.5 wt% of feed.
Figure 9 shows the approximate vacuum heater
outlet temperatures for a 5.0 psi (258 mmHg
absolute) outlet pressure, 22 mmHg flash zone
pressure and 0.5 wt% on feed required at each
cutpoint when processing 100% Maya. It shows
the gas cutpoint is limited to approximately
1020F (557C) at a heater outlet temperature of
800805F (427430C). The flash zone operating pressure is increased by 7.0 mmHg absolute
over the dry design because the ejector inlet
pressure was increased to 10 mmHg absolute
versus a dry design of 4.0 mmHg absolute. The
flash zone pressure should be optimised according to product yield, operating cost and capital
investment.
Increasing the coil steam rate raises the

www.digitalrefining.com/article/1000116

cutpoint at the same heater outlet pressure


because it reduces the oil partial pressure. In this
case, the flash zone pressure was arbitrarily
increased to 25 mmHg absolute, as the ejector
load increases would yield a higher optimum
pressure due to the process steam load having
doubled to 1.0 wt% on feed. Figure 10 shows the
vacuum heater outlet temperatures for a 5.0 psi
(258 mmHg absolute) outlet pressure, 25 mmHg
flash zone pressure and 1.0 wt% on feed required
at each cutpoint when processing 100% Maya. It
shows the gas cutpoint is limited to 1035F
(557C) at a heater outlet temperature of 800
805F (427430C). Increasing the coil steam
injection rate raises the gas oil product yield. It
also increases the investment cost.

Wet vacuum units: residue stripping


Wet vacuum units designed with both coil and
stripping have the highest gas oil product
cutpoint and lowest metals gas oil product at a
given cutpoint. Coil and stripping steam rates
need to be optimised so that the load on the
ejector is minimised. In our example, the total
steam rate is 1.0 wt% of feed. The ejector load
would be the same as the previous example.
Moving steam from the coils to the stripping
section affects the heater oil residence time, but
this can be offset by the lower heater outlet
temperature. A wet vacuum unit with residue
stripping using the same 1.0 wt% coil and stripping steam has a 40F (22C) lower heater outlet
temperature than a wet design with the same
amount of steam. Furthermore, the wet design
gas oil product has lower total metals.
Figure 11 shows the vacuum heater outlet
temperatures for a 5.0 psi (258 mmHg absolute)
outlet pressure, 25 mmHg flash zone pressure
and 1.0 wt% coil and stripping steam on feed
required at each cutpoint when processing 100%
Maya. It shows the gas cutpoint is limited to
approximately 1090F (588C) at a heater outlet
temperature of 800805F (427430C). Capital
investment for this design versus a wet unit with
only coil steam is only slightly greater, but the
potential product yield is much higher.
Gas oil product metals depend on product
distillation. Since residue stripping vapourises
lower boiling-range material than the heater, the
gas oil product 95% -EP distillation tail is lower.
The metals are much lower compared to a wet
design with only coil steam. Furthermore, the

PTQ Q1 2007 9

increasing wash section efficiency can further


reduce the gas oil product 95% -EP distillation
tail and metals. Table 6 compares the gas oil
product total metals for each design.

Conclusion
The type of vacuum unit and equipment design
determine product yield, product quality and run
length. A wet vacuum unit with residue stripping
achieves the highest cutpoint and produces a gas
oil product containing the lowest metals content.
All other things being equal, a wet vacuum
column achieves the longest run length.
Scott W Golden is a chemical engineer with Process Consulting
Services Inc in Houston, Texas, USA. He holds a BS degree in

10 PTQ Q1 2007

chemical engineering from the University of Maine.


Email: sgolden@revamps.com
Tony Barletta is a chemical engineer with Process Consulting
Services Inc in Houston, Texas, USA. Email: tbarletta@revamps.com
Steve White is a chemical engineer with Process Consulting
Services Inc in Houston, Texas, USA.

Links
More articles from: Process Consulting Services
More articles from the following categories:
Revamps, Shutdowns and Turnarounds
Crude Vacuum Unit
Process Modelling & Simulation

www.digitalrefining.com/article/1000116

You might also like