Vacuum Tower Design

Download as pdf or txt
Download as pdf or txt
You are on page 1of 136

Training Services

Vacuum Unit

Also called
“Vacuum Distillation Unit”
(VDU)

EDS 2004/VDU-1

1
Outline

„ Introduction
„ Design Topics
„ Design Examples
„ Operating Tips
„ Optimization, Revamps

EDS 2004/VDU-2

2
Introduction

„ Vacuum Unit - Why Do We Need One?


„ What Is A Vacuum Unit?
„ How Does It Work?
„ How Is It Designed?

EDS 2004/VDU-3

3
Crude Oil Mixture

Boiling Point
F C

< 30 <0 Gas & LPG

100-390 38-200 Gasoline & Naphtha

390-540 200-280 Kerosene & Jet Fuel

540-625 280-330 Diesel

625-700 330-370 Atmospheric Gas Oil


----------------------------------
700-1000 370-540 Vacuum Gas Oil

1000+ 540+ Resid

EDS 2004/VDU-4

The above sketch graphically depicts the boiling ranges of products as


compared with their position in the crude assay. As was discussed in the
Crude Breakup session, the crude assay depicts the boiling ranges of products
as compared with their position in the crude assay at an Atmospheric
Equivalent Boiling Point. As was discussed in the Crude Breakup session, at
atmospheric conditions hydrocarbons begin cracking between 660-725°F.
Thus, in order to vaporize and distill hydrocarbons heavier than this, we need
to reduce the pressure in order to reduce the temperature needed for
vaporization by operating the distillation column under vacuum.

4
Why Do We Need a Vacuum Unit?

„ As discussed in the Crude Breakup and Crude


Unit sessions, at atmospheric pressure
hydrocarbons begin to crack between 660-725°F.
„ If the pressure is reduced, we can still vaporize
hydrocarbon compounds which can be separated
from an even heavier residue - we can separate
Vacuum Gas Oil (VGO) from Vacuum Residue
(VR) and process these 2 streams in separate
units and increase the revenue for the refinery.

EDS 2004/VDU-5

With only a atmospheric crude distillation column, the atmospheric residue


(often 40-50% of the total crude) was often sent to fuel oil. Fuel oil markets
are restricted and often sell at a loss for the refinery.

By sending atmospheric residue to a vacuum column, a low impurity distillate


product (Vacuum Gas Oil, or VGO) can be separated from a much heavier
Vacuum Residue (VR). The vacuum residue is now a much smaller volume
than the Atmospheric Residue.

Depending on the refinery configuration and the products produced, there are
many processing options for the VGO and VR that greatly increase refinery
income.

5
What Is a Vacuum Unit?

„ Fuels type
– Atmospheric Residue Charge
– Flash Zone
— 30-50 mmHg(a), 750ºF (400ºC)
– Products
— FLASHED gas oils
— Vacuum Residue (asphalt)
„ Lubes type
– Requires significant fractionation between
cuts, much tighter fractionation control than a
Fuels type Vacuum Column
– We will not discuss lubes type further
EDS 2004/VDU-6

During this session, we will focus exclusively on the fuels type vacuum unit.
By “fuels” type, we mean a unit that is producing VGO and VR for
processing primarily into transportation fuels (gasoline, kerosene, diesel).

Lube type vacuum towers produce lube base stocks. There are typically 3 or
4 distillate cuts that are produced to much tighter specifications (usually to
viscosity and viscosity index specifications). There needs to be reflux and
fractionation between the various cuts in a lube type vacuum tower. In a
fuels type tower, there is often only reflux in the wash zone, with no reflux
between the 2 gas oil cuts.

6
Crude and Vacuum Unit

EDS 2004/VDU-7
IRP-R03-95

The sketch above typicals a “typical” configuration, showing the feed (AR) to
the vacuum unit coming directly (hot) from the bottom of the crude column.
The other option is for the Atmospheric Residue (AR) to be cooled, sent to
storage, and then sent cold to the vacuum unit. This would require separate
exchange trains for the crude and vacuum unit.

The other feature shown here is that there are 3 sidedraw products, and one
Vacuum Residue (VR) product. Often the Light Vacuum Gas Oil (LVGO)
and Heavy Vacuum Gas Oil (HVGO) are recombined into a Vacuum Gas Oil
(VGO) product. LVGO and HVGO are produced in the vacuum unit for
heat recovery purposes (more heat can be recovered by splitting the VGO into
2 products, even if they are recombined again). Slop wax (also called
overflash by some refiners) can either be recycled to the vacuum heater, or
blended into the Vacuum Residue product.

Therefore, the net effect is to have 2 products from the vacuum unit, VGO
and VR.

7
Vacuum Column

EDS 2004/VDU-8
IRP-R03-95a

This sketch depicts a couple of the features discussed in the last slide. This
sketch shows the Slop Wax routed to the Vacuum Residue. The Slop Wax
can also be routed to the heater inlet or back to the column flash zone.

Some of the HVGO is sent as wash oil to the top of the wash zone packing.
Some HVGO is net product, while the rest of it is cooled and sent to the
HVGO heat removal bed in the column.

Similarly, some of the LVGO is taken as a net product, and the rest sent to the
LVGO heat removal bed in the column.

A vacuum ejector system is used to generate the vacuum in the column.

8
Operating Criterion

„ Past
– No product thermal decomposition (lower flash
zone temperatures)
„ Present
– Maximum gas oil yield (close to cracking) - but
maintain good quality gasoil
– Cracking Can
• Reduce vacuum (overload ejectors)
• Lead to coking problems

EDS 2004/VDU-9

Designs >30 years ago were generally not interested in maximum VGO yield.
The economics of refinery operation at the time did not demand it, and
operating the heater and flash zone of the vacuum unit at lower temperatures
would minimize coking, leading to long run lengths with minimal operator
attention.

The current economic climate demands maximum VGO yield. This pushes
design and operations to the limit of cracking. Greater care is needed in
design and operation to still yield long run lengths.

9
Cracking Depends On:

EDS 2004/VDU-10
EDS-R01-3719

A key point about cracking and coke laydown. Cracking is a time AND
temperature relationship.

10
Operating Conditions

„ Flash Zone Pressure


– 30 – 50 mmHg absolute

„ Flash Zone Temperature


– Approximately 750ºF (400°C)

EDS 2004/VDU-11

2 of the key parameters in vacuum unit operation are the flash zone pressure
and temperature. This two points have the greatest influence on VGO yield.
They are limited by cracking and coke laydown.

11
Vacuum Unit Design

EDS 2004/VDU-12

12
Design Considerations

„ Wet vs. Dry Design


„ Level of Vacuum, Pressure Drop
„ Number of Packed Beds
„ Heater Transfer Line
„ Flash Zone and Bottoms Section
„ HVGO and Heat Removal Sections
„ LVGO and Overhead Section

EDS 2004/VDU-13

13
Design Considerations

„ Wet vs. Dry Design


– Stripping Steam
– Heater Coil Steam
– With or without Precondenser
„ Level of Vacuum
– 3 versus 4 Stage Ejectors
– With or without Precondenser
– Column pressure drop
„ Number of Packed Beds
– Benefits of Intermediate Cut
– Fractionation between HVGO and LVGO

EDS 2004/VDU-14

14
Wet vs Dry

„ Major Design Decision - once set, cannot be


changed
„ There really are 3 types - Dry, Wet with
Precondenser, and Wet without Precondenser
(which some call “Damp”)

EDS 2004/VDU-15

A key design consideration is whether the tower will be dry (no steam input at
all) or wet. Steam can be injected in either/both of two locations - into the
heater coils (to decrease heater coking) and in the bottom boot (to strip lighter
materials from the VR).

If steam is injected, another major decision is whether or not to include a


precondenser in the design.

15
Typical Vacuum Distillation Unit - No Precondenser

EDS 2004/VDU-16
EDS-R04-3709

16
Typical Vacuum Distillation Unit - With Precondenser

EDS 2004/VDU-17
EDS-R04-3709

17
Wet vs. Dry

„ Dry Advantages: Much lower utility consumption


(no steam injected and less motive steam for the
ejectors)
„ Dry Disadvantages:
– With no steam, distillate yield is limited by absolute
pressure
– With no steam in the heater, coking is more of a
concern
– Steam is often needed in the stripping section to
produce proper asphalt

EDS 2004/VDU-18

With no precondenser, the overhead vapor from the column is sent directly to
the 1st stage ejector. Therefore, the load on this ejector is high.

18
Wet vs. Dry

„ Wet Disadvantages:
– Much higher utility consumption (steam injection
and more ejector motive steam)
– Greater chance for water damage on startup
„ Wet Advantages:
– With steam, distillate yield is greater because steam
lowers hydrocarbon partial pressure, thereby
enhancing vaporization and ultimate gas oil yield
– With steam in the heater, coking is less of a concern
– Steam is often needed in the stripping section to
produce proper asphalt

EDS 2004/VDU-19

19
Precondenser vs. No Precondenser

„ As the sketch shows, a precondenser will condense most


of the steam that was injected into the heater and/or
stripping section. However - this comes at a price.
„ The condensed water in the precondenser exerts its own
vapor pressure (at condensing temperature) - this
means the overhead pressure cannot go lower than this
pressure.
„ At 95°F (35°C) water exhibits 41 mmHg abs pressure -
to this must be added the inert gas pressure.
„ Without a precondenser, overhead pressure can be any
value (although for a practical matter it is usually 20-25
mmHg at least).
EDS 2004/VDU-20

With a precondenser in the design, the great majority of the steam coming
overhead from the column is condensed before the 1st stage ejector. This
significantly reduces the load on the 1st stage ejector.

However, there is a water phase vapor/liquid equilibrium that exists at the


precondenser. This exerts a vapor phase water partial pressure at the
temperature that it has been condensed at. This effectively means the
overhead pressure must be at least this value (actually, to this water partial
pressure must be added the hydrocarbon and air partial pressures). This
means that a design with a precondenser will have a higher top pressure, and
thus a higher flash zone pressure as well. The only way to make up for these
higher pressures is to increase the amount of steam injected in order to lower
the hydrocarbon partial pressure (which partially reduces the benefit of
having a precondenser).

20
Precondenser vs. No Precondenser

„ End Result - For maximum lift (deep cut) do


not use a precondenser.
„ Precondensers will be more of a problem in
hot climates - if cold cooling water is available
most of the year, then the effect (while still
present) will be small.

EDS 2004/VDU-21

Thus, if cold cooling water is available, the problems of having a


precondenser are significantly reduced compared to hot climates.

21
Pressure Drop, Level of Vacuum

„ Pressure Drop
– From the column top to the flash zone, typical pressure
drop for a packed column ~10 mmHg
– Each accumulator tray ~1 mmHg
– Each packed bed ~1-2 mmHg
– Each tray 3-5 mmHg (that’s why new columns only have
packing - BIG pressure drop for old columns
„ Level of Vacuum
– Pressures vary greatly
– Wet/No Precondenser UOP designs 35 mmHg flash zone
– Wet/Precondenser often 45-65 mmHg flash zone
– Dry often 20-30 mmHg in flash zone
EDS 2004/VDU-22

22
Typical Vacuum
LVGO
Section Column
Dimensions
HVGO
Section
How do we determine
column diameter?
Wash
Section

Flash
Zone

Bottom
Boot

EDS 2004/VDU-23
EDS-R02-3708

Some typical column dimensions are shown.

The height of sections is determined from spacing requirements for spray


headers. For packed sections, the calculation methods will be shown.

We will now show how to calculate the diameter.

23
Vacuum Column Sizing
Glitsch Method

ρv
C = VS
ρl −ρv

ρl −ρv
VS = C
ρv

EDS 2004/VDU-24

The “C” factor is central to the calculation of vapor loading in distillation.

For vacuum service, we use a design value (0.35 ft/sec) in order set the
column diameter.

For an operating column, it is important to calculate the C factor to determine


how heavily vapor loaded the column is. It is a very good method to
determine if more vapor can be put through the column.

24
Vacuum Column Sizing
Glitsch

where: c = 0.35(1) ft/s


VS = superficial vapor velocity, ft/s
ρv = vapor density, lb/ft3
ρl = liquid density, lb/ft3

Note: (1) For new designs using grid and


limited by re-entrainment, not flooding.

EDS 2004/VDU-25

25
Vacuum Column Sizing
Old UOP Method

760
PV = 2.9
P
where: PV = permissible vapor velocity, ft/s
P = pressure at a given point in
column, mmHga

EDS 2004/VDU-26

This is an older criteria that UOP has used in the past. With a constant of
2.9, this results in larger diameters than the 0.35 C factor.

For the rest of this session, only the C factor method will be used (the PV
method was shown to let you know that there is more than one sizing
method).

26
„ Heater Transfer Line

„ Flash Zone

„ Bottoms

EDS 2004/VDU-27

27
EDS 2004/VDU-28
EDS-R05-3709a

28
Heater Transfer Line

EDS 2004/VDU-29

The heater transfer line is another critical piece of equipment.

From a process perspective, the vacuum heater transfer line sets the pressure
drop between the furnace outlet and the flash zone. This is important
because the real limit for the unit is the furnace outlet vaporization. A high
pressure drop between the furnace outlet and the flash zone will result in a
lower level of VGO yield.

UOP typically designs for approximately 125 mmHg pressure drop between
the furnace outlet and the flash zone. However, in reality this is a detail
design layout issue, and the ultimate pressure drop is a function of the furnace
type and header layout, and the plot plan location determines the length and
number of bends in the transfer line.

29
Heater Transfer Line

„ To Size Line
– Use 70% of sonic velocity at column inlet conditions
– Maximum 350 ft/s
„ Transfer line pressure drop will impact flash zone
temperature and distillate recovery (temperature
will drop from heater outlet to flash zone inlet).
We want to minimize the pressure drop.
„ Recommend “telescope” expanding diameter design
for deep cut designs.

EDS 2004/VDU-30

This line is typically a very large diameter because of the need to stay away
from sonic velocity. This is because, for example, if the flash zone is at 35
mmHg, the furnace outlet may be 160 mmHg. This means the volume of the
gas will increase by a factor of about 4! This is a huge increase in absolute
pressure terms.

This is also why a telescoping design is recommended, both to save on


materials and maintain a constant velocity. .

30
Heater Transfer Line

 Cp  T 
Vs = 223   
 Cv  M 

where: Vs = sonic velocity, ft/s


Cp/Cv = ratio of specific heats (use 1.0)
T = vapor temperature, ºR
M = vapor mol weight

EDS 2004/VDU-31

This is the equation for sonic velocity.

31
Feed Distribution

„ The Vapor-Liquid Separation is not an


Equilibrium-Flash separation in reality - with
the vapor velocity at 50-100% of sonic
„ There is a large amount of entrainment
„ Many different designs have been used in order
to minimize entrainment
– Tangential
– Box
– Many others

EDS 2004/VDU-32

Once the feed gets to the flash zone, the feed must be introduced to the
column. The ideal is to get a pure vapor liquid separation. However,
because of the very high velocities, and the very heavy materials we are
dealing with, there is a large amount of entrainment that occurs.

Thus, there has been a lot of interest in the vapor/liquid separation devices
that can provide improved process performance. The industry has moved
toward tangential vapor horns of various styles (90° vs. 360° for example).
UOP has traditionally designed a “box” distributor. There are other
proprietary designs available.

32
Vacuum Column
Tangential Feed Distributor

EDS 2004/VDU-33
EDS-R00-3730

This is an illustration of a tangential vapor horn. The theory is that the the
centrifugal force will help to separate liquid and vapor. The top is closed as
is the side. Most distributors have closed ends. This forces all of the vapor
and liquid to exit the distributor from the bottom, and thus forces the vapor to
make a 180 degree turn to go up the column.

33
Vacuum
Column
Box Feed
Distributor

EDS 2004/VDU-34
EDS-R02-3715

In this distributor, the top is closed and there are some holes on the side to
allow some vapor to escape, however, as in the vapor horn, most of the vapor
and liquid exit the distributor from the bottom.

34
Wash Section

„ Wash Section is to De-entrain resid material.


Any fractionation is incidental.
„ Column Diameter set by Glitsch C-Factor
limit for packing 0.3 to 0.35 ft/s
„ Overflash or slop wax rate to be controlled by
minimum packing wetting rate.

EDS 2004/VDU-35

The wash section is, for most refiners, a critical operational section. This
section effectively controls the level of contaminants in the HVGO product.
If the destination of the HVGO is a hydrocracker or hydrotreater, then there
will normally be strict controls on concarbon (coke precursors) and metals
(catalyst poisons). Most concarbon and metals are entrained in small liquid
droplets that exit with the vapor from the flash zone (in other words, the flash
zone does not produce a true equilibrium flash as produced by most process
simulators).

(However, it should be noted that when maximizing VGO yield (“deep cut
operation”), the distillate actually can contain significant amounts of
concarbon and metals).

The primary purpose of the wash zone is to de-entrain residue from the vapors
rising up from the flash zone.

35
Grid Bed

EDS 2004/VDU-36

The traditional packing in the wash zone is known as “Glitsch” grid (other
vendors sell similar material). As can be seen in the picture, it is a thick
walled grid with large open area. This has two advantages. The wash zone
is a service highly susceptible to coking. The high level of open space
means that even if some coking occurs, there is a decent chance that the
packing will not close off (which would cause the unit to shutdown). In
addition, the thick walled material also can better handle small amounts of
coking than typical thin walled structured packing.

In recent years, there has been a trend, when going for “deep cut” operation
(maximum VGO), to place a “Glitsch” grid packing in the lower portion of
the wash zone, and a thin walled structured packing at the top of the packing.
This tends to give some fractionation to the wash zone, in addition to the de-
entrainment that grid can give.

36
Grid De-entrainment Efficiencies

Entrainment Removal
Grid Depth, ft-in Efficiency, Percent
2–0 90.0
2–6 95.0
3–0 98.0
3–6 99.0
4–0 99.5
5–0 99.9
6–0 99.99

EDS 2004/VDU-37

In general, if using only grid, it is best to use 6 feet of grid depth to ensure
maximum entrainment removal. Despite the apparent effectiveness of grid at
de-entrainment, some entrainment may still occur due to poor liquid or vapor
distribution, especially at low wash oil rates.

37
Accumulator Tray Details

EDS 2004/VDU-38
EDS-R02-3722

The slop wax (and HVGO and LVGO) is typically withdrawn from an
accumulator tray, a total trap tray. The slop wax withdrawal tray must be
especially designed to minimize liquid residence time since this is heavy
material at high temperature and is prone to coking.

An accumulator tray must also be designed so as to minimize the pressure


drop across the tray. Using the appropriate open cross sectional area will
accomplish this.

38
Wash Section

„ Critical Performance parameter TRUE Wash


Oil Rate in slop wax = minimum of 0.2 gpm/ft2
„ Valuable to determine % entrainment in slop
wax (by performing a concarbon or metals
balance)

EDS 2004/VDU-39

A key performance parameter for wash section performance is to ensure that,


at the bottom of the wash section packing, the flux rate of true wash liquid
(clean HVGO) is at least 0.2 gpm/ft2. This will ensure that the entire cross
section of packing is kept adequately wet and ensure adequate de-
entrainment.

In order to calculate the true wash rate, it is necessary to calculate the


percentage of entrainment that is contained in the slop wax. This can be
done by performing a component balance with either concarbon or metals
analysis in the resid, HVGO and slop wax, and the flow rates of the wash oil
and slop wax.

39
„ Heavy Vacuum Gas Oil Section

„ Heat removal Sections

EDS 2004/VDU-40

40
Heavy Gas Oil Section

„ Primary function is a a condensing/heat removal


section.
„ Note that all withdrawals from the column (slop
wax, HVGO, LVGO) come from Total Trap
Trays - once vapor comes up through the
chimney tray, it can’t go back down the column
(except for wash oil, which is pumped from the
HVGO draw to the slop oil section).

EDS 2004/VDU-41

The primary function of the remaining packed sections of the vacuum tower
are to condense the rising vapors. The packed beds above the HVGO and
LVGO withdrawal trays are contact condensers.

Any vapor that exits the top of the HVGO packed bed will be condensed in
the LVGO bed and withdrawn with the LVGO product.

41
Heat Removal Sections

Q = U “A” (LMTD)

where: “A” = volume of packing

EDS 2004/VDU-42

The heat removal beds are designed in using an equation analogous to heat
exchange design, where instead of and area “A” a packed volume is
substituted.

42
EDS 2004/VDU-43
EDS-R04-3728

43
Heat Removal Sections

U = h o = 4 2 1 C 0 .8 ( g p m A t )
0 .5 8

where: C = capacity factor at bottom of


bed, ft/s
gpm = liquid leaving bed, gal/min
At = column cross sectional area, ft2

EDS 2004/VDU-44

The heat transfer coefficient “U” is calculated as shown above. However, it


should be noted that there is a maximum value for design of 400.

An example is shown in the design example session.

44
GRC Bed - Grid/Ring Combination Bed

EDS 2004/VDU-45
EDS-R02-3726

The heat removal section has been traditionally designed using a “Grid-Ring”
combination bed. The high vapor loaded bottom section uses the same grid
as in the wash zone. Once the vapor load is reduced below a C factor of
0.20-0.23, then random packing (rings) are placed in the top bed. The rings
have greater surface area compared to the grid, but would be overwhelmed if
placed in the bottom of the bed.

Some more recent designs have used structured packing in this service, as
have revamps.

45
GRC Bed

EDS 2004/VDU-46
EDS-R01-3711

46
Grid Layer

EDS 2004/VDU-47

47
Grid Bed

EDS 2004/VDU-48

48
Liquid Distribution

„ As with any packed bed, liquid distribution is


critical to packed bed performance.
„ There are 2 main types - spray nozzles and
gravity fed distributors.
„ Spray nozzles have been the primary device.
„ Gravity fed distributors are gaining more
acceptance with new designs.

EDS 2004/VDU-49

As mentioned above, there are 2 main types of liquid distributors:


pressurized (spray nozzles) or gravity fed (drip type) distributors. Spray
nozzles have been the traditional choice.

Gravity fed devices (drip trays, others) are gaining more acceptance in the
industry. New designs have overcome many of the previous performance
issues. The specific vendors (Koch-Glitsch, Sulzer, etc) should be contacted
to obtain specific recommendations and experience lists.

49
Spray Nozzle
Assembly
Layout

EDS 2004/VDU-50
EDS-R01-3717

When laying out spray nozzles, the nozzle assembly will require a certain
amount of column height. It is important to allow for this height, both for the
nozzle performance and for assembly and maintenance access.

50
Vacuum Column Spray Nozzles
HVGO

(37 Spray Nozzles)


CU-R00-31
EDS 2004/VDU-51

Typically, there are a large number of spray nozzles in any particular layout.

51
Spray Nozzles

Spray
Distributor
Nozzle

Packing
Holddown
Grid

CU-R00-32
EDS 2004/VDU-52

The spray pattern should be set up to have spray pattern overlap between each
nozzle. This is because we need to ensure that ALL of the packing remains
wetted.

52
Liquid Distribution

„ Improper liquid distribution leads to all sorts


of problems.
„ Plugged spray nozzles lead to dry sections of
the packed bed, which (especially in the wash
zone) leads to a coked bed and a shutdown.
„ Normally a filter is installed near the spray
nozzle inlet, and the line after the filter is made
out of stainless steel to prevent rust from
clogging spray nozzles.

EDS 2004/VDU-53

53
„ Light Vacuum Gas Oil Section

„ Overhead System

EDS 2004/VDU-54

54
LVGO Process Flow

Vacuum
Column
Ejector
s

PR FR
C

LC

T
I

FRC

LVGO to Blending

CU-R00-30
EDS 2004/VDU-55

The LVGO section is similar to the HVGO section, as you can see.

One major difference is that there is no reflux (called wash oil for the HVGO)
that is sent to a bed below. Actually, some refiners have a fractionation bed
below the LVGO draw, and in this case reflux would be sent below the
LVGO draw to this additional fractionation bed.

Normally, the LVGO is only cooled against a utility (cooling water or air).
This is because the temperature of the LVGO draw is dependent on the
amount of light material slumped in the crude column. If the amount of
crude slump is higher than design, then the LVGO draw temperature will be
lower than expected. If a LVGO exchanged with another process (such as
cold crude) this would mean the LVGO would not be sufficiently cooled.
This would lead to a large quantity of hydrocarbon in the overhead system.

55
Vacuum Column Overhead

„ Consists of multi sets of ejectors, condensers,


and ejector overhead receiver

„ Provides the vacuum atmosphere necessary for


the proper operation of the vacuum column

„ Provides for separation of the non-condensible


gases, liquid hydrocarbon, and water

EDS 2004/VDU-56

56
Typical Vacuum Distillation Unit

EDS 2004/VDU-57
EDS-R04-3709

57
Vacuum Producing Equipment

EDS 2004/VDU-58
EDS-R04-3709b

This is a sketch of the vacuum overhead system. Steam (typically MP steam


is sent at a controlled pressure to the ejectors. The outlet of the ejectors is
sent to an ejector condenser. There is then a large barometric leg between
the condenser outlets and the receiver. The receiver is a
gas/water/hydrocarbon separator.

A major issue in many refineries is what to do with the gas from this receiver.
Traditionally, this gas sent sent to a furnace to be burned. However, this gas
contains a large amount of H2S. Some refineries compress this gas so that it
can be put in the fuel gas system, where the H2S can be removed.

58
Diagram of a Jet Ejector

A E
Steam C

B
Compressed
Vapors
Suction

IRP-R01-67
EDS 2004/VDU-59

The principles behind ejector operation are covered in another session.

59
Ejector Condensate Drum

Liquid from
Condensers
Non Cond. Gas to
Vapor Heater

Oil to Water
Slops Outlet

CU-R00-33
EDS 2004/VDU-60

The ejector condensate drum allows simultaneous separation of oil/water


mixtures. The oil spills over the weir on the left, while water is forced under
the 1st baffle on the right, and over the second.

60
Non-condensables

„ Air leakage
„ Gases produced in the heater and vacuum
column due to thermal cracking
„ Dissolved non-condensable in the feed
„ Water of saturation in the feed

EDS 2004/VDU-61

The non-condensable gases that the ejector system must remove consist of air
leakage, light ends produced by cracking reactions in the furnace, and water
vapor that is in equilibrium with a water phase.

61
Non-condensable in Vacuum Columns

EDS 2004/VDU-62
EDS-R01-3710

This chart is used for design purposes to estimate the load on the ejector
system.

The use of the chart should be self explanatory.

62
Ejector Design
Figure 1: Ejector Performance Curve
1.2
1000
1.1

Correction Factor
1

0.9
Single Stage
Ejector Inlet Pressure (mmHg)

0.8

0.7

100
0.6
0 50 100 150 200 250 300 350 400
Steam Pressure (psig)

2 Stage, Condensing

10
3 Stage, Condensing

4 Stage, Condensing

1
1 10 100
Add 2 % to Steam Consumption for Every
1.0 PSI Increase over 0.5 psig discharge
Steam Ratio (# Steam / # DAE)
EDS 2004/VDU-63

The plot above illustrates the effect of having different ejector stages. For
the same vapor rate and overhead pressure, less steam is needed with more
ejectors.

The charts are based on the ratio of motive steam(ejector steam) to the “dry
air equivalent” (DAE). The dry air equivalent is a calculation of the overhead
vapor rate in terms of an equivalent rate of air

A key point in ejector design and operation is the steam pressure. Ejectors
are sensitive to the motive steam pressure. Once an ejector is designed for a
certain steam level, it should always be operated at that steam level. Ejector
performance degrades very quickly when moving away from the design
pressure.

This is also the reason why there is a pressure let down controller on the
motive steam line (in order to maintain a constant pressure).

63
Vacuum Unit Design
Examples

EDS 2004/VDU-64

64
Vacuum Unit Design Examples

„ Column Sizing at the Flash Zone


„ Column Sizing at the HVGO Draw
„ HVGO Circulation Rate
„ Column Sizing at the LVGO Draw and LVGO
Circulation Rate
„ Bottom Boot Sizing - Stripping Steam Rate
„ Ejector Sizing

EDS 2004/VDU-65

65
Material Balance

ºAPI UOP Mol (106 scfd) Volume Weight


Gravity K Weight bpsd Lb/h Lb -mol/h Percent Percent

Reduced 22.4 12.0 474 39,370 532,602 1124 100.00 100.00


Crude

Estimated Yields

Gas – – 36 (0.381) 1,589 41.9 - 0.3%

Vac Naph 51.7 11.9 131 13 151 1.1 0.03% 0.03%

LVGO 33.5 12.0 289 5,397 67,894 234.8 13.7% 12.8%

HVGO 24.9 12.1 436 20,158 267,087 612.4 51.2% 50.2%

Slop Wax 17.7 12.2 630 1,197 16,631 26.4 3.0% 3.1%

Resid 14.3 12.2 761 12,497 179,405 235.5 31.7% 33.7%

EDS 2004/VDU-66

This is the material balance we will use for the examples that follow.

66
Design
Temperatures

EDS 2004/VDU-67
EDS-R02-3703

These are the temperature estimates that will be used for the examples that
follow.

These temperatures also make a good check on simulation data.

67
Column Size at Flash Zone

750ºF at 40 mmHga (0.774 psia)

Vapors Rising Lb/h Mol/h


Gas 465 15.5
Steam 667 37.0
LVGO 17355 61.3
HVGO 68575 149.1
Slop Wax 4890 8.9

Total 91952 271.8

EDS 2004/VDU-68

This is the detailed data we will use to size the column diameter at the flash
zone. In general, one can do this calculation, for either design or operation,
without having data from a simulation.

68
Column Size at Flash Zone

– Molecular weight (mol wt) of vapors rising:

 91952 lb   h 
 h    = 338.3
   271.8 mol 

EDS 2004/VDU-69

69
Column Size at Flash Zone

– The equation for c-factor in this service is:

ρl −ρv
VS = C
ρv

where: c = 0.35 ft/s


VS = superficial vapor velocity, ft/s
ρv = vapor density, lb/ft3
ρl = liquid density, lb/ft3
EDS 2004/VDU-70

70
Column Size at Flash Zone

– Vapor density (ρv):

( mol wt)( psia )


ρv = = lb ft 3
( 10.73) ( o R )
where:

( psia ) ( ft 3 )
10.73 = Gas Constant ( R ) =
( lb mol) ( o R )

EDS 2004/VDU-71

71
Column Size at Flash Zone

– Therefore, ρv at this point in the column is:

 338.3 lb   0.774 psia   lb mol - R 


o
ρv =     = 0.0202 lb ft 3
 lb mol   ( 750 + 460) R   10.73 psia - ft 
o 3

– The cubic feet per second (CFS) of vapors rising is:

 91952 lb   h   ft 
3
CFS =    3600 s  0.0202 lb  = 1267

 h    

EDS 2004/VDU-72

72
Column Size at Flash Zone

– Therefore, Vs at this point in the column is:

(46.8 - 0.0202)
Vs = 0 .35 = 16.9 ft s
0.0202

EDS 2004/VDU-73

This is the key value is the above calculations. The C factor chosen for
design (in our case 0.35 ft/sec) effectively determines the column diameter.

73
Column Size at Flash Zone

– The required column cross sectional area (CSA) is:

1267 ft 3   s 
CSA =    = 75.0 ft 2
 s  16.9 ft 

– This results in a column diameter (ID) of:

4
ID = 75.0 ft 2   = 9.80 ft (2990 mm )
π 

EDS 2004/VDU-74

We will see in the next calculation (size at HVGO draw) that this in not the
maximum diameter in the column.

74
Size at
HVGO Draw

EDS 2004/VDU-75
EDS-R02-3705

The next area to check for the column diameter is at the HVGO draw.

The envelope show streams into and out of the envelope in question.

75
Column Size at HVGO Draw
665ºF at 37 mmHga (0.716 psia):
ºF lb/h Btu/lb 106 Btu/h
Heat In
Feed to Column 750 138975 459 63.74
Stripping Steam 413 667 1174* 0.78
HVGO Reflux 545 W 271 271(W)

139642 + W 64.52 + (271W)

Heat Out
Resid from Column 710 47690 364 17.35
Net Slop Wax 725 4890 377 1.84
Vapors Rising to HVGO Draw
Steam 665 667 1340* 0.89
Gas 665 465 511 0.24
Net LVGO 665 17355 448 7.78
Net HVGO 665 68575 438 30.04
HVGO Reflux 665 W 438 438(W)

139642 + W 58.14 + (438W)


*Corrected to liquid at 60ºF

EDS 2004/VDU-76

The above table shows the heat into and out of the HVGO envelope.

76
Column Size at HVGO Draw

IN OUT

64.52 ⋅ 10 6 Btu  271 Btu   W lb  58.14 ⋅ 10 6 Btu  438 Btu   W lb 


+  = +  
h  lb   h  h  lb   h 

W = 38204 lb h

EDS 2004/VDU-77

77
Column Size at HVGO Draw

Vapors Rising lb/h Mol/h


Gas 465 15.5
Steam 667 37.0
Net LVGO 17355 61.3
Net HVGO 68575 149.1
HVGO Reflux 38204 83.1

Total 125266 346.0

– Molecular weight of vapors rising:

 125266 lb   h 
 h    = 362.0
   346.0 mol 
EDS 2004/VDU-78

78
Column Size at HVGO Draw

– Vs at this point in the column is:

45.74 − 0.0215
Vs = 0.35 = 16.43 ft s
0.0215

– ρv at this point in the column is:

 362.0 lb   0.716 psia   lb mol - R 


o
ρv =     = 0.0215 lb ft 3
 lb mol   ( 665 + 460) R   10.73 psia - ft 
o 3

EDS 2004/VDU-79

If you compare the Vs at this point to the Vs for the flash zone, you will see
this is a smaller value. This will require a larger cross sectional area than at
the flash zone.

79
Column Size at HVGO Draw

– The CFS of vapors rising is:


 125266 lb   h   ft 
3
CFS =    3600 s  0.0215 lb  = 1621

 h    

– The required column cross sectional area (CSA) is:

1621 ft 3   s 
CSA =    = 98.7 ft 2
 s  16.43 ft 
– This results in a column diameter (ID) of:

4
ID = 98.7 ft 2   = 11.2 ft (3410 mm )
π 
EDS 2004/VDU-80

As can be seen when compared to the flash zone diameter, this is significantly
larger.

In most cases, the size of a vacuum tower will be set at the point just below
the HVGO draw. The flow below the HVGO section is always the largest in
the column because of vaporization of wash oil and the lower pressure
(compared to the flash zone) at this point. Immediately in the HVGO bed,
vapor is condensed using subcooled pumparound liquids, so the vapor rate
begins to go down.

80
HVGO
Circulation

EDS 2004/VDU-81

Another major design activity required is to calculate the quantity of HVGO


needed to condense the net HVGO product.

In modern designs, the required quantity and heat duty is calculated using a
process simulator.

81
HVGO Heat Removal Section

„ Using the previous calculated vacuum column


diameter and other necessary data, calculate
the amount and type of packing required for
the HVGO heat removal section

EDS 2004/VDU-82

What must still be calculated by hand is the height of packing required to


transfer the heat that was removed by the HVGO pumparound.

82
HVGO Heat Removal Section

– 11 feet, 6 inch ID = 103.9 ft2 column cross sectional


area (CSA)

– Calculate the capacity factor (C) below the packing

1621 ft 3    0.0215
C=  2
= 0.339 ft s
 s  103.87 ft  45.74 − 0.0215

EDS 2004/VDU-83

Since we rounded up the diameter from 11.2 to 11.6 ft, the C factor is actually
slightly less than 0.35.

83
HVGO Heat Removal Section

– Calculate the capacity factor (C) above the packing

 444 ft 3    0.01157
C =  2
= 0.064 ft s
 s  122.72 ft  51 . 23 − 0.01157

– Based on the above capacity factor calculations, a


Grid/Ring Combination (GRC) bed is required in
the HVGO section of the column

EDS 2004/VDU-84

The C factor above the bed is also calculated.

As can be seen, a tremendous amount of vapor is condensed in the HVGO


heat removal bed (the majority of the vapor generated in the flash zone).

84
HVGO Heat Removal Section

– Calculate the overall heat transfer coefficient (U) using


Glitsch’s equation:

0.58
 704 
U = ho = 421(0.286 ) = 469 Btu h −ο F − ft 3
0.8
103.87 
 

– Use 400 Btu/h - ºF-ft3 as a maximum practical U for


design.

EDS 2004/VDU-85

The method that UOP uses to size the heat removal bed is the Grid Ring
Combination bed method (GRC).

The key equation is shown above. The result of the equation is a U value
that is based on volume instead of area. This U value can then be used to
calculate a bed volume.

85
HVGO Heat Removal Section

– LMTD is:

665 ⇒ 375

545 ⇐ 325

120 50

80ºF

EDS 2004/VDU-86

Since the HVGO heat removal bed is involved in heat transfer, a temperature
difference is needed as a driving force. A log mean temperature difference
(as is used in heat transfer calculations) is used in the heat transfer calculation.

86
HVGO Heat Removal Section

– Calculate heat removed (Q) in the HVGO section


of the column

 151151 lb   ( 271 − 132) Btu  6


Q=   = 21.01 ⋅ 10 Btu h
 h  lb 

EDS 2004/VDU-87

The amount of heat removal can be from the HVGO that is condensed.

87
HVGO Heat Removal Section

– Calculate volume of packing:

 21.01 ⋅ 106 Btu     h −o F − ft 3 


    = 656.6 ft 3
 h   80o F   400 Btu 

 ft 
656.6 ft 3   = 6.32 ft required bed depth
 103.87 ft 3 

– Use 6 feet, 6 inch actual bed depth.

EDS 2004/VDU-88

88
Bed Composition

Bed Composition

0.4
0.35
C factor, ft/sec

0.3
0.25
0.2 C factor, ft/sec
0.15
0.1
0.05
0
0 1 2 3 4 5 6 7
Bed Height, ft

3 feet, 6 inch grid plus 3 feet, 0 inch rings if operation at turndown is not a prime concern.
3 foot, 0 inch grid plus 3 feet, 6 inch rings if operation at turndown is a prime concern.

EDS 2004/VDU-89

At the bottom of the bed, the C factor is near 0.35. This portion of the bed
requires the use of grid.

As the vapor velocity decreases, random packing rings can be put in the bed
(hence the name Grid-Ring Combination bed). The breakpoint is between a
C factor of 0.20-0.23.

89
HVGO Circulation Rate

ºF lb/h Btu/lb 106 Btu/h


Heat In
Feed to Column 750 138975 459 63.74
Stripping Steam 413 667 1174* 0.78
HVGO Circulation 325 W 132 132(W)

139642 + W 64.52 + (132W)

Heat Out
Resid from Column 710 47690 364 17.35
Net Slop Wax 725 4890 377 1.84
Net HVGO 545 68575 271 18.58
HVGO Circulation 545 W 271 271(W)
Vapors Rising to LVGO Draw
Steam 375 667 1202* 0.80
Gas 375 465 315 0.15
Net LVGO 375 17355 276 4.79

139642 + W 43.51 + (271W)

*Corrected to liquid at 60ºF


EDS 2004/VDU-90

90
HVGO Circulation Rate

IN OUT
6 6
64.52 ⋅ 10 Btu  132 Btu   W lb  43.51 ⋅ 10 Btu  271 Btu   W lb 
+  = +  
h  lb   h  h  lb   h 

W = 151151 lb h

EDS 2004/VDU-91

91
LVGO
Circulation

Size at LVGO
Draw

EDS 2004/VDU-92

The vapor load in the LVGO section is much lower than in the HVGO
section.

The diameter of this section must be reduced compared to the HVGO section.
Most column are reduced in external diameter. A few columns keep the
same outside shell diameter, but use an internal column inside the shell to
process the LVGO.

This is why most columns have a swaged section at the top.

92
LVGO Circulation Rate

375ºF at 33 mmHga (0.638 psia):


ºF lb/h Btu/lb 106 Btu/h
Heat In
Vapors Rising to LVGO Draw
Steam 375 667 1202* 0.80
Gas 375 465 315 0.15
Net LVGO 375 17355 276 4.79
LVGO Circulation 110 W 23 23(W)

18487 + W 5.74 + (23W)

Heat Out
Net LVGO 200 17355 70 1.21
LVGO Circulation 200 W 70 70(W)
Vapors to Ejectors
Steam 115 667 1084* 0.72
Gas 115 465 175 0.08

18487 + W 2.01 + (70W)


*Corrected to liquid at 60ºF

EDS 2004/VDU-93

93
LVGO Circulation Rate

IN OUT
6 6
5.74 ⋅ 10 Btu  23 Btu   W lb  2.01 ⋅ 10 Btu  70 Btu   W lb 
+  = +  
h  lb   h  h  lb   h 

W = 79360 lb h

EDS 2004/VDU-94

94
Column Size Above LVGO Draw

Vapors Rising lb/h Mol/h


Gas 465 15.5
Steam 667 37.0
Net LVGO 17355 61.3
Total 18487 113.8

– Molecular weight of vapors rising:

 18487 lb   h 
 h    = 162.5
   113.8 mol 

EDS 2004/VDU-95

95
Column Size Above LVGO Draw

– Vs at this point in the column is:

51.2 − 0.0116
Vs = 0.35 = 23.25 ft s
0.0116

– ρv at this point in the column is:

 162.5 lb   0.638 psia   lb mol - R 


ο
ρv =   ο
  = 0.01157 lb ft 3
 lb mol   ( 375 + 460) R   10.73 psia - ft 
3

EDS 2004/VDU-96

96
Column Size Above LVGO Draw

– The CFS of vapors rising is:

 18487 lb   h   ft 3 
CFS =       = 444
 h   3600 s   0.01157 lb 

– The required column cross sectional area (CSA) is:

 444 ft 3   s 
CSA =    = 19.1 ft
2

 s   23.25 ft 
– This results in a column diameter (ID) of:

4
ID = 19.1 ft 2   = 4.93 ft (1940 mm )
π 

EDS 2004/VDU-97

97
Bottom Boot Size

„ Diameter is selected to give a liquid velocity


of approximately 2 feet/minute based on total
flow (net bottom plus quench)
„ Typical DP cell length used is 60 inches
„ Gives 2.5-5 minutes residence time based on
net product

EDS 2004/VDU-98

The bottom boot typically contains 4-6 trays to assist in stripping the liquid
that flows from the flash zone. There then is also is a boot to contain the
residue before it is pumped out of the column. The residence time is
minimized in order to avoid coking. There is also a cooled residue flow
(quench) that is recirculated into the boot in order to reduce the temperature
and avoid coking. In general the temperature must be quenched below 690F.

98
Bottom Boot Size

– For this example:


 139* gal   ft 3   min 
    = 9.29 ft 2 required
 min   7.48 gal   2 ft 

– This results in a boot diameter of:

4
ID = 9.29 ft 2   = 3.44 ft ( 1050 mm )
π
– The final boot diameter can be set at 3 feet 6 inches ID
(1070 mm).

* Refer to Heat and Weight Balance sheets included in reference material.

EDS 2004/VDU-99

99
Bottom Boot Size

– Bottom Stripping Section utilizes large hole sieve decks


– Boot Diameter based on requirements for sieve tray
design
– Check the residence time across the 60 inch DP cell
based on net resid (e.g.):

 9.62 ft 3   5 ft   min   7.48 gal 


     = 2.93 min
 ft     123* gal   ft 3 

* Refer to Heat and Weight Balance sheets included in reference material.

EDS 2004/VDU-100

Most recent designs utilize large diameter (25mm) hole sieve trays for the
stripping section. Older designs have used shed decks.

100
Stripping Steam Rate

 3,203 bbl Resid   5 lb Steam


 24 h   bbl Resid  = 667 lb/h Steam = 37 mol/h
  

„ For a deep cut design, this level of steam stripping


will result in oversized ejectors and large motive
steam requirements
„ Benefits of bottoms stripping will be reduced by
coil steam injection
„ For deep cut, recommend 2 lb Stm/bbl Resid

 3203 bbl Resid   2 lb Steam 


 24 h   bbl Resid  = 267 lb/h Steam = 15 mol/h
  
EDS 2004/VDU-101

Another design value that must be set is the stripping steam rate.

The values shown above are for a wet-no precondenser design. More steam
would be required for a design with a precondenser.

101
Ejector Sizing Example
Non-condensable

„ Air Rate
– At 37,000 BPD, Air Rate is 72 lb/hr
– Note this calculation is independent of pressure
„ Cracked Gas Rate
– At 37,000 BPD and 750oF, base rate is 720 lb/hr
– Correction for 11.9 UOP K is 1.25
– Net Rate is 900 lb/hr
– Cracked gas Molecular Weight is 36

EDS 2004/VDU-102

The methods described earlier are used to calculate the load of air and cracked
gases going to the ejector.

102
Ejector Performance

„ Ejector steam
– Large portion of unit operating costs
– A function of gas rate and desired pressure
corrected for the following:
— Offgas MW
— Offgas Temperature
— Steam pressure
— Offgas % Non-condensable
„ Curves are based on Dry Air Equivalents,
which must be calculated from charts
„ Method presented is independent of condenser
cooling water temperature
EDS 2004/VDU-103

103
Ejector Design
Figure 1: Ejector Performance Curve
1.2
1000
1.1

Correction Factor
1

0.9
Single Stage
Ejector Inlet Pressure (mmHg)

0.8

0.7

100
0.6
0 50 100 150 200 250 300 350 400
Steam Pressure (psig)

2 Stage, Condensing

10
3 Stage, Condensing

4 Stage, Condensing

1
1 10 100
Add 2 % to Steam Consumption for Every
1.0 PSI Increase over 0.5 psig discharge
Steam Ratio (# Steam / # DAE)
EDS 2004/VDU-104

The above chart allows the user to calculate steam requirements given the
number of ejectors, the required overhead pressure and the mass rate of Dry
Air Equivalent (DAE). The foot note points out that ratios are based on a
discharge pressure of 0.5 psig. Adjustments to the ratio are required if the
discharge pressure is higher or lower.

104
Ejector Design
Figure 2: MW Entrainment Ratio Figure 3: Temp Entrainment Ratio
1.8 1

1.6
0.95
1.4 Air
0.9

Entraiment Ratio
Entraiment Ratio

1.2

1 0.85
Steam
0.8 0.8

0.6
0.75
0.4 Weight of gas
Entrainment Ratio =
Weight of air 0.7 Weight @ Temp
0.2 Entrainment Ratio =
Weight @ 70
0 0.65
0 20 40 60 80 100 120 140 0 100 200 300 400 500 600 700 800 900 1000
Molecular Weight Gas Inlet Temperature (Deg F)

EDS 2004/VDU-105

Dry air equivalent is calculated based on the Molecular weight and


temperature of the overhead gas. Generally the overhead is a mixture of
condensable hydrocarbons, noncondensable gases, and steam. The user can
calculate the dry air equivalents for each of these subcomponents and sum the
total dry air equivalent rate.

105
Ejector Design
Figure 4: % Condensable Correction
1.8

1.6

1.4
CORRECTION FACTOR

1.2

0.8

0.6

0.4

0.2

0
0 20 40 60 80 100
% NON CONDSENSABLE
EDS 2004/VDU-106

If a high portion of the gas is condensable, ejectors will function better as the
gas rate will decrease as the materials is condensed, improving the motive
steam to offgas ratio. To adjust for condensation, a correction factor can be
determined based on the % noncondensables

106
Ejector Design - Example

Calculate Motive Steam Requirements for the vacuum overhead system.


Given:
Non condensible Rate - 2652 lb/hr MW - 31.3
Condensible Rate - 18 lb/hr MW - 140
Steam Rate - 8336 MW - 18
Vacuum - 25 mmHg Discharge - 1.5 psig
Temp - 120ºF Steam Pressure - 150 psig
Equivalent Air
Non-Cond + Cond MW - 32.8
MW ER (Fig 2) - 1.06 Temp ER (Fig 3) - 0.985
Steam
MW ER (Fig 2) - 0.81 Temp ER (Fig 3) - 0.982

Eq. Air Mass Flow = (2652 + 18) 8336


+
(1.06)(0.985) (0.81)(0.982)
= 13037 lb/hr of Equivalent Air
EDS 2004/VDU-107

As noted, the first step to calculating motive steam requirements is to


determine the off gas rate and convert it into a dry air equivalent. The three
components of offgas are condensable hydrocarbons (referred to as
condensables), non condensable gas, and steam. The temperature and
molecular rate and temperature factor for all these components can be
determined from figures 2 and 3.

107
Ejector Design - Example

Motive Steam Requirements


Two Stage Three Stage
Base Ratio (Fig 1) 6.8 4.5
Steam Pressure Corr* 0.89 0.89
Discharge Modifier 1.02 1.02
Non Cond Corr - 24% (Fig 4) 0.65 0.72
Two Stage Ratio = 6.8*0.89*1.02*0.65 = 4.01 lb steam/lb air
Steam Rate = 4.01*13037 lb air/hr = 52,310 lb/hr of steam
Three Stage Ratio = 4.5*0.89*1.02*0.72 = 2.94 lb steam/lb air
Steam Rate = 2.94*13037 lb air/hr= 38,345 lb/hr of steam
Installing a third stage will save 13,965 lb/hr of steam or $1039 / day

* Based on 140 psig - 10 psi Control valve drop


EDS 2004/VDU-108

Using figure one we can determine that to get to an overhead pressure of 25


mmHg we need 6.8 # motive steam/# DAE for two stages and 4.5 # motive
steam/# DAE for three stages. Corrections for steam pressure, discharge
pressure and non condensible amount also need to be factored. Doing so
gives the benefit for installing a third compressor stage.

There are other methods available for calculating ejector performance, and
most are based on the concept of Dry Air Equivalent ratios.

Do you see any weaknesses of this methodology, especially in calculating


operating ejector performance? Will it adequately reflect changes in
operation from summer to winter?

108
Typical
Vacuum
Column
Dimensions

EDS 2004/VDU-109
EDS-R02-3720

109
Vacuum Unit
Operating Tips

EDS 2004/VDU-110

110
Operating Tips

„ Flash Zone Pressure


„ Flash Zone Temperature
„ Circulating HVGO and LVGO
„ Slop Wax Draw
„ Steam

EDS 2004/VDU-111

111
Pressure

„ Must Maintain Steady Pressure


„ Effect on Gas Oil Yield
„ Effect on Capacity

EDS 2004/VDU-112

As has been stated earlier, most refiners desire to maximize VGO yield. A
low pressure maximizes gas oil yield. However, just as important is that the
pressure in the column remain steady. Without pressure control in the
overhead ejector system, maintaining a steady pressure in the flash zone will
be impossible. A varying pressure in the flash zone often leads to poor
HVGO and VR quality, and a large amount of quality giveaway in order to
compensate.

112
Temperature

„ Increases Gas Oil Yield


„ Sign of Cracking if Temperature Increase
Reduces Vacuum
„ Controlled at Heater Outlet
„ Bottoms to be Quenched to <690°F

EDS 2004/VDU-113

As mentioned earlier, most refiners want to maximize gas oil yield.


Maximum heat input in the charge heater contributes to a maximum gas oil
yield. The temperature limit is cracking of the feed. Cracking will coke up
the furnace, requiring a shutdown, and also generate cracked gases which can
overload the ejector system.

In order to prevent cracking in the bottoms, the bottoms temperature must be


quenched to <690°F.

113
Circulating Gas Oil

„ Find Optimum HVGO Circulation for


Maximum Crude Preheat

„ Find Optimum LVGO Circulation for


Minimum Overhead Temperature

„ Rule of Thumb:
LVGO to HVGO ratio is 1:3 to 1:4

EDS 2004/VDU-114

There is also a tradeoff - both in design and in operation - between the


circulation rate and the delta temperature in the pumparound circuits. The
details for design should be worked out during the pinch analysis of the heat
exchange train. However, a rule of thumb is 150°F temperature drop for a
pumparound circuit.

As mentioned earlier, even if we have only a VGO product, we split the VGO
into LVGO and HVGO (with all of the additional expense that goes along
with that) in order to increase the amount of heat recovered. While 2 or 3
cases are normally looked at for design, a rule of thumb is that the ratio of
LVGO/HVGO product rates is in the 1:3 to 1:4 range.

114
Slop Wax Draw

„ Increase if HVGO Metals and Carbon High


„ Watch Spray Pressure
„ Send to Heater or Bottoms?
„ Recommend external measured draw
„ Actual Wash = Total Wash -
Entrainment -
Leakage and Condensate

EDS 2004/VDU-115

As mentioned earlier, achieving and maintaining clean HVGO product is a


key performance parameter for the vacuum unit. A number of factors come
into play.

As mentioned before, the actual wash oil rate at the BOTTOM of the bed
should be at least 0.2 gpm/ft2. This means the wash oil flux rate at the top of
the bed will have to be much higher than 0.2 gpm/ft2.

The spray headers are sometimes a source of problems. Often the problem
comes from material that plugs individual headers, which can lead to a dry
wash bed and coking of bed sections, which will cause either a high pressure
drop or offspec HVGO. Either will lead to a shutdown, removal of the wash
grid/packing, and replacement.

Another decision is where to send the slop wax. The system is typically
designed to recycle back to the heater inlet. In this way the HVGO contained
in the slop wax can be recovered. An option will exist to blend this back
into the Vacuum Residue. This is used if the heater is at a limit, or if VR
quality demands some light blend component. Some refiners have also put
the slop wax to the top of the stripping trays.

115
Slop Wax Draw

„ An example of entrainment calculation based


on ramscarbon material balance:
HVGO at 0.4 wt% ramscarbon
Vac Resid at 22.7 wt% ramscarbon
Slop Wax at 9.3 wt% ramscarbon

„ By algebraic calculation, x = wt frac of resid


into slop wax
9.3 = (x)(22.7) + (1-x)(0.4)

„ Inferred entrainment x = 0.399


EDS 2004/VDU-116

As mentioned earlier, achieving and maintaining clean HVGO product is a


key performance parameter for the vacuum unit. A number of factors come
into play.

As mentioned before, the actual wash oil rate at the BOTTOM of the bed
should be at least 0.2 gpm/ft2. This means the wash oil flux rate at the top of
the bed will have to be much higher than 0.2 gpm/ft2.

The spray headers are sometimes a source of problems. Often the problem
comes from material that plugs individual headers, which can lead to a dry
wash bed and coking of bed sections, which will cause either a high pressure
drop or offspec HVGO. Either will lead to a shutdown, removal of the wash
grid/packing, and replacement.

Another decision is where to send the slop wax. The system is typically
designed to recycle back to the heater inlet. In this way the HVGO contained
in the slop wax can be recovered. An option will exist to blend this back
into the Vacuum Residue. This is used if the heater is at a limit, or if VR
quality demands some light blend component. Some refiners have also put
the slop wax to the top of the stripping trays.

116
Effect of Slop Wax on Yields - Need to Make Wash Oil

EDS 2004/VDU-117
EDS-R04-3729

This slide just gives us a visual reminder that the slop wax that is generated is
the equivalent of overflash in the crude unit - it is extra vapor that must of
generated in the flash zone in order to make the wash oil that is sent back to
the wash zone.

117
Steam

„ Bottom Stripping

„ Steam to Heater?

„ Optimize steam injection against ejector size,


operating costs, and lift

EDS 2004/VDU-118

As mentioned earlier, if steam is injected into the column, it is normally


injected into 2 locations, into the heater coils (to minimize coking) and in the
stripping section (to remove light material from the residue, which, if asphalt
is made assists in asphalt production).

The choice of precondenser/no precondenser, and the levels of steam for


design are normally studied during the design phase.

During operation, the amount of steam used should be optimized. Many


refiners try to minimize steam usage, which ends up causing product quality
problems. Minimum utility usage is not always optimal in terms of total
profitability.

118
Operation Design Considerations

„ Bottoms Rundown (fuel oil production)


– Quench draw to 650oF-690oF
– Inject hot cutter stock to improve resid heat
transfer coefficient
– Blend cold cutters to improve rundown viscosity
– Make allowances for asphalt production
– Avoid finned fans. Use tempered water cooling

EDS 2004/VDU-119

As mentioned earlier, the bottom boot residue must be cooled below 690°F in
order to prevent coking in the boot.

Often the ability to inject a low viscosity cutter stock (such as LCO from a
FCC unit) into the hot residue. If the residue is to be sent to fuel oil and
cutter will be added anyway, the lower viscosity will improve heat transfer in
the residue heat exchangers.

The ability to make asphalt also affects the design. A stripping section is
required for asphalt production.

For residue, tempered water systems should be used for cooling.

119
Operation Design Considerations

„ Non-condensible Off Gas


– Arrange for disposal (typically to crude furnace)
– Preferable to amine treat
– Heat trace offgas line

EDS 2004/VDU-120

120
Operation Design Considerations

„ Packed Sections
– Ensure good vapor and liquid distribution

– Provide filters for reflux lines

– Use stainless steel downstream of filters to


avoid scaling
– Mount pressure points at reflux inlets to
monitor spray pressure drop
„ Overhead pressure control
– Often installed, seldom used

EDS 2004/VDU-121

121
Operation Design Considerations

„ Water Damage
– Water at 40 mmHg explodes to 25,000 times its
volume inside a vacuum column
– Only use superheated steam with traps

– Thoroughly check vessel and piping for low


points where water might collect
– Install hold down grids to keep packing in place

EDS 2004/VDU-122

The biggest operational problem seen in vacuum units is water causing


damage. This cannot be emphasized enough.

122
Vacuum Unit
Monitoring, Optimization,
Troubleshooting,
Revamps and Control

EDS 2004/VDU-123

123
VDU Monitoring

„ Heater
– Monitor Flux

– Calculate velocity profile / optimize steam injection

– Monitor draft

– Monitor transfer line velocity

– Monitor wetgas make / composition

EDS 2004/VDU-124

124
VDU Monitoring

„ Column
– Monitor column pressure profile
– Calculate c-Factors for each bed
– Monitor section heat removal
– Monitor wash rate / slop wax rate
– Monitor spray header pressure drops
– Monitor flash zone entrainment and wash
zone effectiveness (HVGO product quality)
– Monitor top temperature

EDS 2004/VDU-125

125
VDU Monitoring

„ Ejectors
– Monitor ejector/condenser performance

– Monitor skim oil / wet gas make

– Monitor steam consumption

EDS 2004/VDU-126

126
VDU Unit Optimization

„ Optimize Stripping Steam Injection


– Measure ejector performance / determine
optimal injection rate
„ Maximize coil outlet temperature.
– Monitor heater flux / offgas make

EDS 2004/VDU-127

127
VDU Unit Optimization

„ Increase CDU stripping steam


– Reduces VDU skim oil, improves vacuum

„ Optimize LVGO / HVGO recovery


– Increasing HVGO draw will limit load on
LVGO section allowing lower top temp & lower
pressure drop
– Increasing LVGO draw will increase draw
temperatures / heat recovery

EDS 2004/VDU-128

128
VDU Troubleshooting

Problem Symptoms Possible Cause


Loss of Distillate High flash zone a) High column pressure drop
Yield pressure b) Poor Ejector performance (see
Poor Overhead Vacuum)
c) Wash section coking
High overflash rate a) Leaking draw tray
b) Control failure on wash oil
Low Flash Zone a) Poor furnace performance
Temperature b) Open vent/steam out valve

Loss of Section High PA return temp a) Exchanger fouling


recovery b) Open bypass
c) Control problems
Low PA flow a) Hydraulic problems
b) Control problems
Low product flow a) Leaking draw tray
b) Poor liquid/vapor distribution
c) Section flooding
EDS 2004/VDU-129

129
VDU Troubleshooting
Problem Symptoms Possible Cause
Poor overhead High overhead a) High LVGO return temp, poor
vacuum temperature / High exhanger performance
skim oil rate b) Poor LVGO or HVGO section
vapor/ liquid distribution
c) Loss of CDU stripping
Low Ejector a) Fouled condensers
compression b) Blocked or undersized ejector
throats
c) Poor steam quality / low flow
d) Increased gas / stripping steam
rates
High Seal Drum a) Blockage in offgas line
Pressure b) High pressure drop due to high
flow or equipment problems
HVGO quality High metals * / a) Entrainment due to high flow
problems concarbon rates or equipment damage
b) Insufficient wash rate
c) Coking in wash section
* Note that some organo-metallic components may be present
in the HVGO boiling range
EDS 2004/VDU-130

130
VDU - Revamp Opportunities
Increased Capacity - $$$

„ High performance (structured) packing


– Increased heat transfer
– Allows for higher pumparound rates
„ Add Furnace Capacity
– Additional Transfer Line ?
„ Increase CDU Distillate recovery
– Remove more diesel
– Increase AGO draw

EDS 2004/VDU-131

131
VDU - Revamp Opportunities
Improved Yields - $$

„ Increase Vacuum
– Add/Rearrange Ejectors
– Allows for higher pumparound rates
„ Increase Temperature
– Consider run length/cracked gas
– Calculate Flux
„ Decrease pressure drop
„ Reduce entrainment
– Improve wash section
– Upgrade feed separation device

EDS 2004/VDU-132

132
VDU - Revamp Opportunities
Reduce Energy Costs - $

„ Change from HP to MP Steam


„ Replace ejector with Vacuum Pump
„ Recover vacuum naphtha as diesel product
– May require some fractionation
„ Increase number of draws

EDS 2004/VDU-133

133
VDU Process Control

EDS 2004/VDU-134
EDS-R04-3709

134
Control Issues

„ HVGO PA Rate
– Control duty, not Rate
– Use Vap Temp above bed to reset duty
– Set LVGO/HVGO cutpoint
„ LVGO Reflux (fractionation bed)
– Use Vapor Temp above bed to reset flow
– Sets LVGO End Point

EDS 2004/VDU-135

135
Control Issues

„ Column top temp


– Try to minimize
– Control by LVGO pumparound duty
„ Top pressure control
– Limits top vacuum
– Often kept closed

EDS 2004/VDU-136

136

You might also like