5

Distillation, Absorption and Stripping

in the Petroleum Industry
(." r;s/;a T/ "/a k

/'(/,,,,1'(// Hullfil ....

('fa IIdl' .Uu r/y

After having reviewed the general principles that govern the operations of dis-
tillation, absorption and stripping, we are now going to examine a number of
applications in the petroleum industry.
We will discuss the following topics:
atmospheric distillation of crude oil, including the important desalting
operation;
vacuum distillation of the atmospheric residue;
gasoline distillation and gas fractionation.
Finally in the last part we will present column internals and their use in the
various petroleum processes.

5.1 Atmospheric Distillation of Crude Oil (Topping)

Atmospheric distillation is the basic unit in crude oil refining, historically
speaking it appeared at the end of the 19th century. Because of its position
upstream from all the other units, it is the one that handles the largest prod-
uct quantity and plays a key role in proper refinery operation. The unit’s
capacity ranges from 300 000 t/year to over 10 million tons a year and its cost
was in the neighborhood of $120 to 150 million in 1994 for a 10 Mt/year unit
(200 000 BPSD).
Generally speaking, the unit produces the following cuts:
A wide range gasoline cut which is feed for the gasoline and gas fraction-
ation section (see Section 5.3), usually after hydrotreatment.
Sometimes a naphtha cut for the petrochemical industry (see Vol. 1,
Chapter 6, Section 6.2).
222 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY

A kerosine cut which is designed for use in producing aviation fuel. It can
also be used to produce solvents, space heating oil and automotive diesel
oil.
One or two gas oil cuts which will be used to make space heating oil and
automotive diesel oil.
An atmospheric residue which is sent as feed to the vacuum distillation
unit (see Section 5.2).
A distillate cut is sometimes drawn off between heavy gas oil and the flash
zone. Depending on its quality it is mixed back in the atmospheric residue or
used as a fluxing agent for heavy fuel oil.
The gas oil and kerosine cuts almost always need a finishing treatment in
another unit before they are incorporated in commercial products. These cuts
are obtained by simple distillation and they may contain a variable amount of
unwanted products depending on the type of crude. From the standpoint of
current legislation sulfur and aromatic compounds are the main undesirables.
It is also possible to draw off a light gasoline cut in the distillation unit. The
naphtha is fractionated, a light gasoline cut (containing gases) is recovered at
the top of the main column and a heavy naphtha cut is drawn off laterally via
a stripper. This separation is less effective than what is obtained in a separate
column (see Section 5.3). Additionally, there is a risk of water condensation at
the top of the tower, which limits the use of the process to light gasolines
whose end point is sufficiently high.
When the topping unit is designed it is usually sized to allow crudes of dif-
ferent characteristics to be processed:
a base crude, representative of the refinery’s average oil supply;
a lighter crude, which sizes the top of the scheme and furnace zones;
a heavier crude, which sets the dimensions of the bottom of the column
and the exchanger train.
In these three cases, processing capacity is not identical so as to minimize
overdesign. With the same furnace, processing capacity is higher for a heavy
crude.
The unit is also designed so as to allow variations in the cut points amount-
ing to about twenty degrees Celsius on TBP (summer/winter motor fuel mar-
ket, for example). Finally the unit needs to be able to operate properly at
approximately 60% of its nominal capacity.

5.1.1 Description of the Process

Figure 5.1 shows the flow diagram of a crude oil topping unit.
Separation is generally effected in a single column, working at a pressure
of 1 to 3 bar rel., usually at the lowest possible pressure. Products are drawn
off as side streams by means of strippers. These stripping columns are
reboiled (when a dry product is desired) or more generally steam stripped.
Figure
5.1 Flow diagram of a crude oil topping unit (Source: ENSPM-H).
224 Chapter 5. DISTILLAT/ON, ABSORPTION
AND STRIPPING IN THE PETROLEUM INDUSTRY

The revaporized light fractions are sent back to the main column above the liq-
uid offtake going to the stripper.
This column, which in fact behaves like a reflux absorber, is equipped with
one to three pumparounds that allow heat to be recovered at advantageous
thermal levels.
The overflash, which is the internal reflux of the first tray above the feed
inlet, is (seldom) drawn off or sent back to the bottom of the tower. It mea-
sures the excess vaporization of the feed and guarantees the heavy gas oil
quality.
Overhead condensation is total or partial depending on the temperature of
the cooling fluid and the type of crude. When it is partial, the residual gas frac-
tion is usually taken up by a compressor and sent to the refinery's fuel gas net-
work.
The main column is some fifty meters high and is usually equipped with 30
to 50 conventional valve trays. Side strippers are equipped with 4 to 10 trays
of the same type (see Section 5.4).
The crude is preheated in an exchanger train by recovering heat from the
pumparounds and products to a temperature of between 120 and 160°C at
which it is desalted. The operation takes place at a sufficiently high pressure
(approximately ten bars) so that the crude plus water mixture remains liquid
at this temperature.
The desalted crude is then heated in a second exchanger train, then in a
furnace to a temperature of approximately 330 to 390°C, to be let into the main
column in a partially vaporized state.
The cuts are sometimes dried:
either at a temperature of about 150°C in vacuum driers stacked up with
the side strippers,
or at the draw off temperature by means of simple coalescers or salt fil-
ters.

5.1.2 Fractionation Quality

Fractionation quality is often formulated in terms of "gap" or "overlap" of the
relevant products.
If fractionation was perfect, there would be no common compound
between two successive cuts. The TBP end point of the light cut would be the
same as the TBP initial point of the heavy cut. The ASTM curves of the two
products would then exhibit a positive offset called gap, which is representa-
tive of fractionation quality.
When the ASTM curves exhibit a negative offset, it is called overlap.
Since determining the initial and end points is imprecise, the 5%(or 10%)
points and the 95% (or 90%) points are usually taken to qualify the fractiona-
tion quality between products. Figure 5.2 illustrates the concepts of gap and
overlap. Table 5.1 gives a number of typical values.
Chapter 5. DISTILLATION, A6SORPTlON AND STRIPPING IN THE PETROLEUM INDUSTRY 225

Perfect TBP fractionation

Light cut Heavy cut

Low selectivity fractionation

The gap and overlap.

IBP: Initial boiling point. EP: End point.
226 Chapter 5. DISTILLATION, ABSORPTION
AND STRIPPING IN THE PETROLEUM /NDUSTRY

Standard distillation
methods
1 Separation between products

ASTM D 86 Widerange gasoline/kerosine 10 to 20

Light cut 95%point Kerosine/light gas oil 5 to 10
Heavy cut 5%point Light gas oil/heavy gas oil 0 to -5 (overlap)
ASTM D 1160 Vacuum gas oil/
Light cut 90% point vacuum distillate 5
Heavy cut 10%point

I I
Table
5.1 Typical values o f gap and overlap (for the definition o f distillation standards,
see Volume 1).

Other internal specifications in the unit are also commonly controlled:

the flash point of kerosine and gas oils;
the end boiling point ASTM D 86 of heavy naphtha, generally lower than
185"C, so that the end boiling point of reformed gasoline complies with
commercial standards;
the cloud point of heavy gas oil;
the flash point of atmospheric residue.
Appropriate management of the unit allows the cut points and the frac-
tionation quality to be regulated.

5.1.2.1 Modifying Cut Points

The cut points in the unit are controlled by the overhead temperature con-
troller (TRC), which indirectly regulates the flow rate of total naphtha drawn
off, and by the various flow rate controllers (FRC) located at the discharge end
of the pumps at the bottom of side strippers, which are respectively kerosine,
light gas oil and heavy gas oil in a column with three side streams (see Fig. 5.1).
The atmospheric residue flow rate is level controlled at the bottom of the
tower and is therefore automatic, thereby tying up the material balance.
The overhead temperature controller, if located at the top of the column in
single condensation or on the first drum (hot drum) in double condensation,
is representative of the dew point temperature of the total naphtha at its par-
tial pressure in this spot.
The set points of the four controllers mentioned above dictate the col-
umn's material balance.
Modifying the set point of one single controller changes all the cut points
of the heavier products down from the product where the modification was
made. Accordingly, lowering the set point of the kerosine flow rate controller
modifies the kerosine end cut point (lower TBP cut point) but also the initial
and end cut points of gas oils and the initial cut point of the atmospheric
 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM /NOUSTRY 227

residue. The naphtha cut points remain unchanged. The naphtha and gas oil
flow rates do not vary but the flow rate of the atmospheric residue, the inter-
nal reflux and the overflash increase by the fraction of kerosine previously
drawn off. The quality of naphtha does not change, but that of all the other
products does. These products are lighter and their offtake temperature is
lower. It should be noted that this lessens the amount of heat extracted by
pumparounds and therefore acts on the separation quality of the products as
we will see later on.
Modifying the set point inversely makes the kerosine, gas oils and residue
heavier and the offtake temperature higher. Offtake flow rates are unchanged,
except for the kerosine flow rate which increases and the atmospheric residue
flow rate which decreases. Overflash and internal reflux flow rates also rise.
The naphtha flow rate is unchanged.
The cut point between two consecutive offtakes is therefore changed by
simultaneous action on the two controllers of the relevant products, one
upward and the other downward. For example, to lower the cut point between
kerosine and light gas oil, the kerosine flow rate set point must be lowered and
the light gas oil set point must be raised in the same proportion. In this case
the cut points of the other products remain unchanged.

5.1.2.2 Modifying Fractionation Quality

The fractionation quality between two consecutive products is given by the
liquid-vapor traffic in the zone under consideration for a fixed number of trays
in the zone. Increased traffic enhances product separation. For a given column
feed inlet temperature, the liquid-vapor traffic in the different fractionation
zones is distributed by means of the amount of heat extracted by the different
pumparounds. To affect separation quality between two consecutive products
positively or negatively, it is necessary to act in the same way as in modifying
the cut point, i.e. on two parameters simultaneously. For example to improve
separation between the two gas oils, it is necessary to reduce the amount of
heat extracted by the heavy gas oil pumparound (to let the vapors pass) and
increase the heat extracted by the light gas oil pumparound (to condense the
vapors). If this added amount of heat is not extracted at the light gas oil level,
it will automatically be drawn off by the column’s condenser. As a result all the
products lighter than heavy gas oil will be enhanced, since the liquid-vapor
traffic will have increased in this whole zone in the column. This is why an is6
lated action on the kerosine pumparound is sufficient for the separation
between naphtha and kerosine. Finally, it should be noted that the column’s
feed inlet temperature regulates the overflash flow rate and therefore the sep-
aration quality between the heavy gas oil and atmospheric residue. When
pumparounds are fixed, the feed inlet temperature can be used to act on all the
fractionations. In conclusion, a cut point modification and a fractionation
adjustment between products requires action on three or four relevant vari-
ables simultaneously, thereby prompting the development of tools to control
and help manage the refinery (advanced control: see Section 5.1.9).
228 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM /NOUSTRY

5.1.3 Calculation and Simulation of Topping Columns

The considerable development in industrial data processing in the area of col-
umn simulations in the nineteen seventies and eighties has freed the engineer
of long, tedious and often approximate manual calculations for these compli-
cated columns.

5.1.3.1 Background
In the field of manual calculation, the basic references belong to Packie (1941),
completed by Prater and Boyd (1955) and Watkins (1969).
This method uses overall enthalpy and material balances by zone and
empirical correlations to calculate the number of trays. It is described in great
pedagogical detail in the second book by Watkins (1979). The book is a source
of interesting experience that provides a genuine understanding of how these
complex columns work.
The reader can also refer to the books by Nelson(1958), Van Winckle (1967)
and Wuithier (1972), where he will find examples of manual column calcula-
tion.
At the beginning of the nineteen sixties the first correlations of liquid-vapor
equilibria for heavy petroleum fractions were published by Edmister (1961).
The correlations, supplemented by the physical properties of heavy fractions
found in the book by Maxwell (1965) and in the publications of the American
Petroleum Institute (MI TechnicalData Book), can be used to calculate the col-
umn tray by tray.
The simultaneous development of information technology and the appear-
ance on the market of specialized software have made rigorous column calcu-
lation possible.
Nowadays a simple labtop can be used to calculate a column in a few min-
utes. The time it takes to enter data is longer than the time for calculation,
which is carried out by means of extremely powerful algorithms. The software
offers a choice of thermodynamic models used both for liquid-vapor equilibria
and for the products' physical properties. They are generally linked to calcu-
lation methods for tray geometry (that can be found in the publications of the
Fractionation Research Institute, for example) and the column diameter can
consequently be directly estimated.
The software is used in the plant sizing stage or to check the operation of
existing units.

5.1.3.2 Using the Packie Curves (1941)

The curves proposed by Packie establish a direct relation between the offset
of the 95% and 5% ASTM points (gap or overlap) of two consecutive offtakes
on a topping tower and the product of the reflux ratio times the number of
trays in the section separating the two offtakes.
They were established for lower offtake stripping ratios of 0 to 24 kg/m3 (kg
of steam injected per m3 of stripped product corrected to 15OC).
 Chapter 5. DISTILLATION, ABSORPTION
AND STRIPPING IN THE PETROLEUM /NDUSTRY 229

The proposed relation is as follows (Fig. 5.3):

gap or overlap = f
(?)
with:
gap or overlap algebraic difference between the 5%ASTM D 86 point of the
lower cut Si+land the 95%ASTM D 86 point of the higher cut Si(vapor
+ liquid distillate for the overhead product/upper distillate separa-
tion)
R number of actual fractionation trays between the two cuts under con-
sideration

Ri reflux flow rate under the upper offtake (m3/h at 15°C)

CSi sum of offtakes that are lighter than the lower offtake (m3/h at 15'C)
For the overhead product/upper distillate separation, Ri corresponds to
the internal reflux flow rate under the overhead tray and C Si to all the liquid
and gaseous overhead products.
If a pumparound exists at the top of the section under consideration, Ri cor-
responds to the internal reflux flow rate under the pumparound offtake tray.

I CSi

I
L
Figure
5.3 Definitions used in the Packie method.
230 Chapter 5. DISTILLATION,
ABSORPTION
AND STRIPPING IN THE PETROLEUM INDUSTRY

F = nr,
100
80
60

40

20

10
8
6
4
30 -20 -10 0 10 20 30 40 50 60
4 u T

Overlap ("C) Gap ("C)

Separation between overhead product
and first draw off

-
Figure
5.4 Estimating gaps and overlaps with the Packie method.
 Chapter 5 DISTILlATlON, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY 231

k is a corrective factor. Because of the better efficiency observed in the

fractionation zones for the lightest cuts (distilling above approximately 250°C
on TBP), a corrective factor is proposed to be included in the original method.
The factor is taken equal to 0.75 when the cut point between the two relevant
offtakes is less than 250°C.
The curves presented in Figure 5.4 were plotted versus the difference in
temperature between the 50% volume point on TBP of the lower offtake (&+J
and the 50% volume point on TBP of the sum of the lighter offtakes.
When the stripping ratio ranges between 0 and 24 kg/m3, the results can be
interpolated linearly.

5.1.3.3 Computerized Calculation:

Choosing the Thermodynamic Model
Simulation software gives access to a variety of thermodynamic models for
refining units, allowing the liquid-vapor K values , the enthalpy and entropy
values as well as the liquid and vapor densities (see Vol. 1, Chapter 4) to be
calculated.
The different models can be associated with each other (one for K, another
for enthalpy values, a third for densities). Generally speaking the software
manufacturer proposes the most common associations or packages. Table 5.2
presents the most widely used for topping units.

I I

Liquid and gas

K value density

Braun Curl-Pitzer Ideal-AP1

Grayson-Streed Curl-Pitzer Redlich-Kwong-API
~ Soave-Redlich-Kwong Soave-Redlich-Kwong Soave-Redlich-Kwong
Peng-Robinson Peng-Robinson Peng-Robinson-API
Chao-Seader Lee-Kesler-Johnson Riedel-API
Table
5.2 References used in choosing thermodynamic models.

These models use a number of very diverse tools such as:

equations of state (Peng and Robinson),
empirical or semiempirical correlations (Braun/Grayson),
the law of the corresponding states (Lee and Kesler).
Each of them has its proponents and choosing among them involves know-
how.
Documented deviations between two extreme models go up to 5% on
the amount of vapor generated in the flash zone or the thermal load of the
232 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM /NOUSTRY

overhead condenser. Deviations of 3 to 5°C are possible on offtake tempera-

tures and variations of 20% of the overflash have been seen.
Each year new models appear and their evolution had not yet come to an
end in 1994.

5.1.3.4 Entering Data and Analyzing Results

a. Introducing Crude Parameters
Based on the existing crude oil analysis, the crude is cut up again (except for
light fractions, for which an analysis is usually available) into narrow fractions
either manually or by computer. These narrow fractions are entered into the
program in the form of hypothetical components or pseudocomponents
defined by two of the following three characteristics:
density,
boiling point temperature,
molar mass.
The crude must be cut carefully in the cut point zones (by steps of 10
to 20°C) and more roughly in the other zones to keep from overloading the sys-
tem (steps of 30 to 50°C for the atmospheric residue above 400°C).
If the cut is effected by computer, the pure substances/elemental cuts con-
nection must be checked. Some software offers the cutting procedure and
automatic data entering directly from the TBP of the crude.
b. Number of Trays
Table 5.3 recapitulates the number of actual trays to be considered:

Zone Number of actual trays

Gasoline/kerosine 8 to 12
Kerosine/gas oil 5 to 10
Light gas oil/heavy gas oil 6
Heavy gas oil/ flash zone 6 to 8
Residue stripping 4 to 8
Kerosine stripping 6 to 10
Gas oil stripping 4 to 8
Table
5.3 Estimate of the number o f actual trays in the differentzones in the column.

The thermal exchange zones have 2 to 4 real trays depending on the

amount of heat exchanged. Their number is not included in the values in the
table.
The pressure drop in the column between the top and the flash zone is
approximately 0.4 to 0.6 bar.
 Chapter 5. DISTILLATION,
ABSORPTION
AND STRIPPING IN THE PETROLEUM INDUSTRY 233

c. Estimating the Overhead Pressure

The overhead pressure can be rapidly estimated by calculating the bubble
point pressure of a gasoline cut precisely cut on the TBP at the cooling tem-
perature accessible in the overhead condenser. The flash zone pressure is then
equal to this pressure plus the estimated pressure drop between the overhead
drum and the flash zone.

d . Amount of Stripping Steam

The average stripping ratios used are as follows (action on the flash point of
the relevant products):
kerosine and gas oils: 15 to 30 kg of steam/m3 of product;
residue: 20 to 30 kg of steam/m3 of residue.

e. Regulating the Overflash

The column feed inlet temperature is chosen to obtain an overflash of 3 to 5%
of the feed, i.e. the amount of vapor from the feed flash plus the revaporized
product from the bottom of the tower is equal to the sum of distilled products
plus the value of the desired overflash.
A prior crude flash at different temperatures and at the pressure of the
flash zone is used to approach the desired value rapidly.

f. Analyzing Results
Separation calculations yield the TBP and ASTM product curves, recalculated
from correlations. The curve beginnings and ends are fairly reliable and it is
not recommended to make the column converge on product specifications
(gap or overlap). Instead it should converge conventionally on the flow rate of
withdrawn products and the amount of heat extracted by the various
pumparounds, leaving the condenser thermal load free. The specifications are
then checked afterward. The Packie method, described in Section 5.1.3.2, is
also a good means of checking.

g. Convergence Difficulties
After making sure that the amount of vapor leaving the flash zone is greater
than the amount of product to be drawn off, convergence of the column cal-
culation can be hastened by letting the heat rise in the condenser, i.e. refrain-
ing from extracting heat by means of the various pumparounds. In this way, all
the internal reflux in the column and the products are condensed at the top.
The pumparounds are then increased little by little in order to adjust the inter-
nal reflux that is strictly necessary for fractionation in each zone.
The amount of heat extracted by means of the pumparounds is the result
of an economic optimum, since the number of trays increases with the amount
of heat extracted.
Lastly the pumparound flow rates are adjusted and the return tempera-
tures set in order to guarantee a proper approach in the heat exchange zones.
234 Chapter 5. DISTILLATION,
ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY

5.1.4 Technology
The technology used in topping units is very conventional. Most equipment is
made of ordinary carbon steel except for hot zones where there are steel
alloys. Part of the topping tower is usually clad with 12% weight chromium
steel. In zones that are exposed to cold corrosion (top of the tower, reflux
drum) more sophisticated materials such as Monel or special coatings are
sometimes used.
The columns are equipped with conventional valve trays. The number of
passes is determined by the magnitude of liquid and vapor traffic in the zone
under consideration. There may be 4-way trays in heat exchange zones, but
there are usually 2-way trays in the other zones. Tray efficiency is good at the
top of the tower and mediocre in the zone located between the gas oil offtake
and the feed inlet (fortunately the most complex fractionation occurs at the
top of the tower). Trays are normally made of steel alloy (12% Cr). In fraction-
ating or wash zones structured packing is today often used (see Section 5.4).
The draw off trays that feed the side strippers and the pumparound pumps are
nowadays often of the total draw off type, thereby allowing absolute control
over the internal reflux in each fractionation zone (chimney tray). This is
already commonly used in vacuum distillation of atmospheric residue.
The main column feed inlet is of the tangential type or axial type with fins
to promote vapor disengagement.
The furnaces are of the cabin or cylindrical type with vertical hanging or
horizontal pipes. They have air preheaters and steam generating or super-
heating tube banks so that they meet the specified output requirements com-
patible with the huge amounts of heat involved.
The air is preheated by exchange with the flue gases coming from the fur-
nace, which can be relatively cool if the fuel used is desulfurized. Low NO,
burners must be used today to minimize discharges. Part of the air can also be
preheated by means of a non-hydrocarbon process stream (condensates, salt
water). Observed efficiency ranges from 80 to 90% of the heat passing over to
the process and 90 to 95%overall efficiency for recent furnaces. The difference
between the two efficiencies gives the heat available for generating and super-
heating steam.
Heat exchangers are of the TEMA type. Their nominal working pressure
does not exceed some 30 bars.
The column diameter varies depending on the zone, top and bottom are
smaller and heat exchange zones are bigger. The average diameter is approxi-
mately 9 meters for a tower processing 1 000 t/h of crude.

5.1.5 Process Variations

5.1.5.1 Double condensation at the Top of the Column
Since energy management is a very important problem in distillation units (see
Section 5.1.8), it has become necessary to exploit the low level thermal energy
 Chapter 5 DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY 235

available at the top of the tower by exchanging heat with the crude. These heat
exchangers, commonly called "crude/overhead vapors", require a tried and
true special technology and particular supervision, since crude leakage from
tubes can directly pollute the naphtha.
The initial condensation is carried out by heat exchange with the crude and
only condenses the amount of product (naphtha) that is strictly necessary for
the reflux. The reflux is then hot and far from the condensation temperature of
steam. Corrosion problems are, as a result, easier to deal with.
The second condensation is performed by air or cooling water, with the
naphtha produced in this way and the sour water recovered separately after
settling.
This second condensation is in some cases carried out by exchange with
the cool crude which contacts all the vapors to be condensed in countercur-
rent flow.
In the same way as for single condensation flow diagrams, double conden-
sation allows total or partial condensation.
Figure 5.5 shows the flow diagram variation with double condensation at
the top of the column.

Crude

' Produced naphtha

Figure
5.5 Crude oil topping. Double condensation at the top of the column.
236 Chapter 5 DISTILLATION, ABSORPTION
AND STRIPPING IN THE PETROLEUM lNDUSTRY

5.1.5.2 Other Flow Diagrams

Each flow diagram is specific to the site, the target products and the crude oil
supply.
In France after the Second World War port refineries processed relatively
undiversified crudes (Algeria: Hassi Messaoud, Zarzaitine) and few mixtures.
From the nineteen sixties onward, sourcing diversified (Middle East) and
mixture processing appeared on the scene. In the seventies and eighties, s u p
plies started coming from the North Sea and the then Soviet Union. These
developments meant more flexibility in designing and operating units.
At the same time, prompted by constantly growing demand in highly indus-
trialized countries, refining capacity developed from 1 to 3 Mt/year then
5 Mt/year and up to 10 Mt/year (corresponding to columns ten meters in diam-
eter). There are also more modest units installed for the captive market in
countries with limited development (300 000 to 500 000 t/year).
The result today of all this is increasingly systematic integration of the
vacuum column in the crude fractionation operation and flow diagrams theo-
retically highly integrated between the two sections. The gas plant, gasoline
distillation and water stripper are also often integrated into atmospheric dis-
tillation.
As for unit design, with the appearance of desalters, implementation of a
flash drum to vaporize light fractions and especially corrosioncausing water
was the precursor of the flow diagrams incorporating preflash columns that
some refiners use.
Topping units preceded by reboiled or stripped preflash columns have
been built and help relieve the furnace of the gasoline that is evacuated at the
top of these columns. It is a fact that heating products that distill at 150°Cto a
temperature of over 350°C is pointless. Additionally, the light fractions tran-
siting in the main column deteriorate the temperature profile although they do
promote vaporization.
Figure 5.6 shows a simplified distillation flow diagram with a preflash col-
umn.
Others have generalized this concept by incorporating several columns in
the exchanger train, thereby recovering several gasoline cuts directly
(Elf/Technip patent: progressive crude distillation).
Energy recovery has prompted a large number of innovations and heat
exchange trains have always been the ideal area for research (see
Section 5.1.8). Exchanger surface areas were increased considerably (practi-
cally doubled) in the years following the petroleum price shocks of the seven-
ties.

5.1.6 Crude Oil Desalting

Crude oil desalting is an essential operation in the refinery, as it conditions
proper (or less than proper) downstream processing.
 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM /NDUSTRY 237

Gases + gasolines

,
n
7

Crude
@+ Gas oil

Atmospheric

Preflash column )c/ Main tower

- ~

Figure
5.6 Simplified diagram of a crude oil topping unit with a preflash column.

Effectively,poor desalting has direct consequences on the operation of the

atmospheric distillation tower:
fouling of exchangers and the furnace;
corrosion of overhead equipment;
atmospheric residue with a high sodium content, with the following con-
sequences:
- increase in fouling rate of the vacuum distillation furnace;
- shorter cycle length for visbreakers;
- catalyst poisoning in catalytic cracking, particularly when heavy feeds
are processed:
- fouling and corrosion of boiler superheaters.
Lastly, poor desalting can also result in massive hydrocarbon carryover
into the water coming from the desalter, and therefore severe environmental
problems that the refiner will have a hard time dealing with.

5.1.6.1 Crude Salinity. Marine Contamination

Ever since they come on stream most wells produce oil with some degree of
salt. Sodium chloride is the majority salt, but it is always associated with vari-
able amounts of alkaline earth salts. Salinity comes from the reservoir water
or from infiltrations. It depends on the structural position of the well and the
physical characteristics of the reservoir rock. In addition, it changes during
the lifetime of the well.
238 Chapter 5. DISTILLATION, ABSORPTION AND STRlPPlNG IN THE PETROLEUM lNDUSTRY

Furthermore, transportation has an inevitable influence on the salinity of

the crude unloaded at its destination. During the voyage, enrichment in sea
salts may reach from 10 to 80 mg/l of crude on the average.’ There are three
main causes of the increase in salt content:
1. “Inevitable”contamination (pools of water around suction screens and
at the bottom of tanks, salt deposits at the sea water/air interface).
2. “Deliberate”contamination. It should not be forgotten that sea water
legislation mandates that the carrier should risk polluting crude with
water instead of water with crude, so it is now out of the question to dis-
charge deballasting water containing over 100 ppm of crude oil. In this
connection, international legislation has laid down the ”load on top” pro-
cedure that we mention as a reminder, and which alone causes an
increase in salt content ranging from 5 to 50 mg/l of crude.2
3. “Accidental”contamination, finally, due to miscellaneous causes in the
terminal or on board the tanker.
As a general rule, and this is what should be remembered, sea water is pre-
ponderant in the salinity that the refiner will have to deal with (see Table 5.4).
The crude is delivered to the off-loading port after having been partially
desalted on the oil field and then enriched in salts during shipping. It is then
sent to the refineries where it undergoes a desalting operation before frac-
tionation in the topping unit (see general flow diagram in Figure 5.7).

5.1.6.2 m e of Salts
a. Characteristics
The salts in crude oil are mainly chlorides with the following approximate
breakdown by weight per cent:
NaCl 70-80
MgCI2 20-10
CaCI2 10
These salts are either in the form of crystals or ionized in the water present
in the crude.3
By simple settling all the ionized salts should theoretically be eliminated,
but the viscosity of some crudes would require very large storage capacity.
Crystals can be eliminated by washing with water: the crystals are ionized
in the water, then hydrated. The advantage of these hydrated salts is that they
are much more soluble in water. This shows, without a doubt, the importance
of adding water when a crude oil is to be properly desalted.

1. Crude salinity is generally expressed in mg of chloride equivalent per liter.

2. For the new ships, the oil tanks are independent of the ballasts and are not any more
washed with sea water.
3. When the crude enters the refinery, the water content is less than 1% by mass.
 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY 239

Crude oil origin Average content Spot content measured

in loading harbor @pm) in refineries

North Africa
Amna < 30 90
Arzew < 25 62
Brega G 30 155
Es Sider 5 149
Hassi-Messaoud < 10 31
Sarir 60 345
Zarzaitine < 33 40
Sub-Saharan Africa
Gamba < 21 58
Nigerian Bonny Light < 60 136
Mandji 30 115
Emeraude - 220
Middle Fast
Arabian Light < 12 25
Safaniya < 180 40-280
Basrah (Zubair) 3 24
Iranian Light (Agha Jari) 15 25
Iranian Heavy (Gasch Saran: < 18 25
Kirkuk < 10 25
Kuwait 12 35
Miscellaneous
Ural - 20
Boscan < 80 30-60

Table
Salt content of a number of crude oils (before and after transportation).

6. Drawbacks of Salts
Salts foul preheating exchangers. As soon as the salt concentration
exceeds 40 pprn at the desalter outlet a change can be seen in the fouling
resistance, with deposits laid down more rapidly in tube bundles and
therefore overconsumption of utilities.
Alkaline earth chlorides (MgCl,, CaC12) are hydrolyzed with generation of
hydrogen chloride, which causes corrosion in the overhead equipment of
topping units.
It is generally accepted that the concentration in chlorides of overhead
water in the tower should not exceed 10 pprn, otherwise severe corrosion can
occur.
 h)
Corrosion inhibitors P
I 0

Gasoline + LPG + gases

Caustic A Steam 4 Gases ’

Desalted : To hydrotreating
Gasoline and gas plant

+r c Steam
I Water

f I
I- Steam

Atmospheric residue 8
To vacuum distillation 3
Water from desalter P
Mixing to stripper then to 2
e
valve secondary treatment

Crude to Water for desalting

be treated

Figure
5.7 Crude desalting and topping.
 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM /NOUSTRY 241

Besides conventional control methods (injection of caustic in the crude

after desalting and of corrosion inhibitors in the top of the tower (see Fig. 5.7),
desalting is one of the most important anticorrosion weapons.
Poor desalting means that atmospheric residues are high in sodium.
According to the literature, the coking rate in furnaces (vacuum distillation,
visbreaker, etc.) varies as the square of the sodium concentration. In addition,
catalytic cracking catalysts are rapidly poisoned (sodium poisoning is irre-
versible).
These facts illustrate the impact that desalting quality can have on the
downstream processing units.
c. Other Contaminants
Other contaminants are present in crude oils: mineral sediments, rust and
even iron sulfide. The major drawback is that they can generate stable emul-
sions that are (sometimes) hard to break.

5.1.6.3 Desalting Mechanism

The crude is washed with water and the washing water is separated out by
electrostatic desalting in order to eliminate all the impurities mentioned
above.
Desalting includes three consecutive steps:
diffusion of the crude’s salts in water (by washing),
coalescence of water droplets (by electrocoalescence),
settling (by gravity).
a . Diffusion of Salts
The objective is to get the salt crystals contained in the crude to pass into the
water. Since all the salt crystals must be affected, the watercrude emulsion
must be sufficiently fine (Fig. 5.8).
The water and crude are normally mixed through a mixing valve placed at
the desalter inlet.
Water is often injected partly at the mixing valve and partly on the dis-
charge end of the feed pump for improved diffusion of the salt crystals in the
water (Fig. 5.7).
b. Coalescence
The watercrude emulsion, a thorough mixture of two non-miscible liquids, is
made up of a continuous phase (the crude) and a dispersed phase (the water
in the form of droplets whose dimensions vary from 1 to 10 pm). Asphaltenes
and finely divided solids (iron sulfide for example) are adsorbed on the water-
oil interface forming a film which can be seen with a microscope. These agents
stabilize the emulsion.
The difficulty in coalescence is therefore closely related to the asphaltene
content of the crude on the one hand and to the presence of foulants or sedi-
ments on the other.
242 Chapter 5. DISTILLATION, ABSORPTION
AND STRlPPlNG IN THE PETROLEUM /NOUSTRY

Watercrude emulsion.

Coalescence is induced by an electrostatic field. It occurs by means of the

forces of attraction among the droplets due to the polarity of water molecules
(they tend to become oriented) and the agitation created by the electrical
field.

F= KEV( 4 a)
The force of attraction F of droplets of identical size in a line is equal to:

with:
r radius of the droplet
d distance between the centers of the droplets
E tension gradient
K constant
It should be noted that rand r/d are independent variables.
Additionally, demulsifying agents are used to complete the action of the
electrical field.
c. Settling
Settling is governed by Stokes’ law:
 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY 243

with:
V , settling velocity (m/s)
r radius of the water drop (m)
p1 density of the dispersed phase (water)(kg/m5)
p2 density of the continuous phase (crude)(kg/m5)
p2 dynamic viscosity of the continuous phase @‘as)
g acceleration of gravity (9.81 m/s?
Desalters are generally calculated for a dynamic viscosity of 2 mPa.s and
for a apparent settling time of 20 to 30 min.

5.1.6.4 Industrial Implementation

and Recent Technological Developments
Figure 5.9 gives a cross sectional view of a desalter with the distribution of the
watercrude emulsion and the fields in the desalter as described below:
a) A field of around 200 V/cm called the “primary field”which acts between
the water-crude interface and the first electrode.
b) A field called the “secondary field” which acts between the first and the
second electrode and whose intensity is much higher (approximately
1000 V/cm).
Figure 5.10a shows the industrial design of an electrostatic desalter ( p r e
cess, control) on a simplified flow diagram.
Depending on the salt content of the crude and the desired performance, a
second in-series desalting stage may be added (see Fig. 5.10b).
This type of scheme offers two advantages:
1. Since desalters require lengthy maintenance periods (periodic cleaning
due to sediment deposits in the bottom of the drum), the twestage solu-
tion allows operation in one stage during maintenance.
2. The water flow rate can be significantly decreased in comparison with a
one-stage desalter.
Optimization of desalting technology has continued until recently and the
following developments should be pointed out:
a) Location of alternative electrical fields: can vary up to highly complex
configurations providing the water drops with movement and settling zones
that make coalescence faster.
b) Emulsion feed to the desalter: a variety of solutions where manufactur-
ers use different methods (Fig. 5.1 1):
either “high feed inlet”, with the manufacturer expecting a considerable
increase in coalescence effects above the crude/water interface level;
or “low feed inlet” going through the layer of settled water, which washes
and breaks up part of the feed emulsion.
244 Chapter 5 DISTILLATION,
A6SORPTlON AND STRIPPING IN THE PETROLEUM INDUSTRY

I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I

- ~

Figure
5.9 Simplified flow diagram of an electrostatic desalter (cross sectional view).
E,: primary field, E,: secondary field (Source: ENSPM-R).

Figure
5.10a Electrostatic desalter with one stage.
 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY 245

Desalted
crude

Stripped water
Figure

High feed inlet

(above the interface)

Watercrude emulsion Feed inlet at the interface

feed to a desalter (difi
ferent schemes).

Q Low feed inlet

(below the interface)
246 Chapter 5. DISTILLATION, A6SORPTlON AND STRIPPING IN THE PETROLEUM INDUSTRY

c) Design of the whole desalting drum: suffice to say that different flow dia-
grams have been proposed to allow onstream removal of accumulated
deposits at the bottom of the drum and accumulated foulants at the interface.
d) Use of several coalescence stages in one single vessel (Fig. 5.12). The
different stages are hydraulically isolated and allow a succession of parallel or
in-series sequences. Each stage is characterized by flat permeable electrodes.
The system is an assembly of multiple electrodes that make up a series of elec-
trical fields, each of which can have its own load distribution. Two- or three-
stage units have been achieved, but more can be imagined. Pumps and piping
feed the coalesced liquid from one stage to the next.
This system can be installed on an existing desalter.

i E4
xf E3

Yc ElE2

Crude
HT: high tension

- -
Figure
5.12 Desalter with three electrodes (Petrolite patent: US 4149958).

5.1.6.5 Desalter Control Parameters

We will now review the degrees of freedom left to the refiner’s choice to get the
best efficiency from the desalter, excluding the design parameters used by
manufacturers in sizing the equipment.
Table 5.5 gives the seven control parameters the operator can resort to in
operating his desalter.
Let us add that the operating pressure is approximately 10 to 12 bar @res-
sure required to operate in the liquid phase).
 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY 247

Table Ratio of washing water

Desalter control parameters. 5.5 Washing water injection point
Type of washing water

a. Level of Water/Crude Interface

The level of settled water represents a 0 potential electrode in the primary
field defined with the lower electrode in the unit. Any significant variation in
water level modifies the primary field and perturbs electrical coalescence.
Consequently, the level should be kept constant according t o the manufac-
turer’s recommendations.
b. Desalting Temperature
The settling criterion: @1- PJ
P2
with: p1 density of water, p2 density of crude, p2 crude viscosity, is a func-
tion of the temperature.
Table 5.6 below shows desalting temperatures where crudes exhibit the
same value of the settling criterion defined previously:
(PI - PJ
= constant
P2

Crude Gravity CAPO Desalting temperature (p9

Sahara 45 (50)
Qatar 41.3 70-80
Zakum 40.1 id.
Murban 39.5 id.
Urnm Shaif 37 100
Iraq Kirkuk 36 120
Bonny blend 32 id.
Es Sider 35.5 id.
Iran Light 33.7 140-1 50
Zubair 34.5 id.
Arabian Light 34.4 id.
Forcados 30.8 id.
Oman 33.5 id.

Table
Variation in desalting temperature according to differentcrudes, with constant
settling criterion.
248 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY

The table shows that a Sahara crude could theoretically be desalted at

around 50°C, whereas Arabian Light or Forcados need temperatures of some
140 to 150°C. Furthermore, the position of the desalter in the preheating
sequence does not always give the operator much leeway. Finally, the variation
in crude conductivity with temperature can also limit desalting temperature as
shown in Figure 5.13.

t I

15

10

50 100 Temperature ("C) 150

15131 Crude oils. Electrolytic conductivity versus temperature (S is the symbol for
Siemens).
 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY 249

c. Washing Water Ratio

Electrical coalescence is dependent on the ratio of washing water. For heavy
crudes whose API gravity is less than 30, apparent difficulties of gravity set-
tling (parameter @* - p2 )/p2 introduced previously) will be offset by increas-
ing electrical coalescence, i.e. by increasing washing water.
If we consider for example that desalting performance values are satisfac-
tory in usual design conditions (the case of Kirkuk 115°C approximately, 34%
water), it will be necessary to have a variable ratio of washing water depend-
ing on the type of crude and the desalting temperature to get similar results .
This is illustrated by Table 5.7.
In practice, refiners adjust the desalting temperature and water ratio in
relation to the most viscous crude that is usually processed in the unit.
Table 5.8 gives the average conditions in usual practice according to the
API gravity of the crude.

Gravity Minimum washing water ratio

(~01% with respect to the crude)
Crude
CAPO 110°C 120°C 140°C 150°C
Kirkuk 36 3-4 3-4 -
Zubair 34.5 5-6 45 4
Forcados 30.8 4 5 45 3-4
Oman 33.5 6-7 5-6 5
El Morgan 31.8 6-7 6 4 5
Kuwait 31.2 7-8 7 6
Khafdji 27.8 L 10 L 10 8-10 7-8
Safaniya 28.3 L 10 L 10 8-10 7-8
Nigerian medium 24.8 > 10 > 10 8-10 7-8
Bonny
Eocene Ratawi 19.7 To b e fluxed
iT.ble1-
Variation in the water ratio according to the desalting temperature For a
constant theoretical desalting efficiency.

Crude gravity Desalting temperature Minimum water ratio

(“APT) (“9 (vol 0%)

8-10
Minimum 10
I
Table
Average desalting conditions. (If the crude can not be heated, it can always be
fluxed to lower its viscosity).
250 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY

d. Washing Water Injection Point

If the washing water is not very sour it is best to inject all or part of it on the
discharge end of the feed pump and before the preheating exchangers (here
we are referring to the exchangers located upstream from the desalter which
are designed to preheat the crude)(see Fig. 5.7). This possibility becomes
imperative if desalting is carried out at high temperature (serious risk of
deposits in crude preheating exchangers upstream from the desalter).
It should be pointed out that some refiners successfully inject the water at
the cool crude pump intake.
Lastly, a common practice consists in injecting the washing water partly on
the discharge end of the feed pump and partly at the mixing valve.

e. Type of Washing Water

Condensation water from topping and vacuum distillation is used for desalting
after steam stripping the water from the second one to eliminate NH3 and H,S
beforehand.
Stripped catalytic cracking water can also be utilized. Since it has a high
phenol concentration (between 100 and 300 ppm), it does, however, recycle
phenols into the crude.
Soft water can be used as make up, but as it is highly oxygenated, it is bet-
ter used after mixing with process water. This allows the oxygen to be elimi-
nated by reaction with the sulfides present in the process water.
Nota bene: Sea water is not saturated in salts and could therefore theoreti-
cally serve as acceptable washing water. However, corrosion risks are too high
to deliberately advise use of sea water.

E Pressure Drop in the Mixing Valve

It is very difficult to propose strict rules for mixing valve operation, A dramatic
increase in differential pressure might form stable emulsions that would be
hard to break up.
However, theoretically speaking, the finer the dispersion of the water, i.e.
the greater the pressure drop in the mixing valve, the more effectively the
crude is washed. A compromise must therefore be found between:
the pressure drop,
the temperature,
the amount of water available.
In fact, optimum pressure drop is determined experimentally and ranges
from 1.5 bar (on the average) for light crudes to less than 0.5 bar for viscous
crudes.
 Chapter 5. DISTILLATION, ABSORPTION
AND STRIPPING IN THE PETROLEUM INDUSTRY 251

g. Demulsifier Type and Level

Demulsifying agents are used to complete the electrostatic desalting opera-
tion.
Demulsifier formulas generally include two or three bases, the most com-
mon of which are ethylene oxide-propylene oxide sequenced copolymers,
sometimes condensed with phenol-formol resins, organic diacids or ethylene
diamine groups.
Levels range between 3 and 10 ppm of the crude, with the highest ratios
used for the most viscous or the most asphaltic crudes but also for some sour
crudes.
Injection points are located either in the crude upstream from the feed
pump or in the washing water or in both places.
Lastly, before any foreseeable perturbation (change in crude, recycling
highly emulsified recovery oils, etc.) it is recommended to increase the
amount of demulsifier as a preventive measure.

5.1.6.6 Desalters Performances

When the unit’s control parameters are properly adjusted desalting efficiency
(measured by the rate of salt elimination) varies on the average from 85
to 95%.
The water content in desalted crude is usually lower than 0.2% (volume),
except for heavy crudes where it can go up to 0.4 to 0.5%.
The hydrocarbon concentration in the water coming out of the desalter
must not (normally) exceed 200 ppm.
Problems may arise when recovery oils are recycled (see above) as they
often form stable emulsions. Consequently they perturb the operation of the
unit, with a resulting decrease in desalting efficiency and deterioration of
water quality at the exit of the unit.
As mentioned earlier, in these conditions a larger dose of emulsifying addi-
tives is a good way to offset lower desalter efficiency.
Nota bene: In any case, caustic is injected into desalted crude to reduce cor-
rosion at the top of the topping tower. Caustic prevents hydrolysis of cal-
cium and magnesium chlorides in the form of hydrogen chloride, which is
corrosive when it is dissolved in the condensed water at the top of the col-
umn. Caustic neutralizes these salts and transforms them into NaCI.
One of the operator’s priority objectives is, of course, getting maximum
desalting efficiency, thereby reducing the amount of injected caustic. The
atmospheric residue will consequently have a low sodium content and cor-
rosion problems will also be attenuated at the top of the tower.
252 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM /NDUSTRY

Lastly, monitoring and control of the unit's performance are based on five
types of measurements:
salt content in the feed crude and in the desalted crude,
sodium content in the atmospheric residue,
chloride content in topping overhead water,
water content in desalted crude,
hydrocarbon content in the water coming out of the desalter.

5.1.6.7 Special Case of Heavy Oils

a. Definition and Main Characteristics
In addition to conventional crudes, there are also what are called "heavy
crudes", whose processing results in serious problems in the refinery.
Heavy oils generally come from fairly shallow reservoirs and have under-
gone bacterial degradation due to invasion by surface water. This deteriorated
state, which can be variable within the same field, means a significant loss in
light fractions and normal paraffins, and therefore a loss in product value.
The main parameters which characterize heavy oils are the following:
Viscosity: this is the fundamental characteristic of a crude. It distin-
guishes between a bitumen (in situ viscosity higher than 10 000 mPa.s)
and a heavy oil.
Gravity: an oil is said to be heavy if its API degree is lower than 20. The
Venezuelan ministry of energy and mines distinguishes between heavy
oils (10 c "API c 20) and extra heavy oils CAP1 =s10). This was the first
definition criterion used but it is not the only one.
Percentage of light fractions: the residue at 200°C is often approximately
95%.
Percentage of asphaltenes: high. For example, some 11%for Venezuelan
crudes.
Sulfur content: very high, often situated at about 5% by mass.
Table 5.9 gives a number of characteristics of heavy oils by way of an illus-
tration.

6. Desalting Heavy Oils

Because of their specific characteristics, heavy oils pose serious problems in
refineries. Since their gravity is in the neighborhood of 1, desalting and dehy-
dration require well adapted operating conditions.
Their high viscosity means they have to be kept at a very high temperature
throughout the whole processing operation, and this is detrimental from the
standpoint of cost. Additionally, their high asphaltene content has a negative
impact on desalting, which is usually carried out in two or more stages at a
temperature between 130 and 150°C. It is necessary to flux heavy oil either
 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY 253

wscosity
Crude ‘API

-
(m2/s)
..
Boscan
(Venezuela) 10.1 5600 250000 14
Job0 (Venezuela:
(Orenoco belt) 9.2 12
North Battleford
(Canada) 10 14.4 85.6
Lagunillas
(Venezuela) 16.8 22.5 77.5
Qayarah (Iraq)
-
16.4 150 500 15 40 I 45

These are fractions with a boiling point lower than or equal to that of gas oil.
** (%mass).

1 Table
5.9 Examples of heavy oils.

with gasoline or gas oil at relatively high concentration levels, which may go
up to 30%(mass) in some cases.
Demulsifiers are also injected into the oil, the doses are variable but can
sometimes reach 100 ppm. In some cases it is recommended to inject additives
at the interface when there is too much oil-water emulsion.
Finally, residence times are between one hour and one and a half hours.
By working under these conditions, it is possible to desalt heavy oil prop-
erly. Boscan, for example, that can be considered as an extra heavy oil (“API
close to lo), can be desalted to less than 5 mg/l (NaCI equivalent).

5.1.6.8 Conclusion
Besides the overhead equipment in topping units, which must have constantly
improved corrosion protection, the need to produce low sodium atmospheric
residues in order to convert the heavy petroleum fractions economically
entails the following:
control and optimization of desalter regulation parameters,
but also development of further new technologies.
The quality of desalter discharge water is also an important parameter that
will increasingly have to be taken into consideration.
For further information the reader can consult a number of reference works
and publications among which we will list the following:
Bartley (1982); Bron (1981); Cu et al. (1993); Kronenberger and Pattison
(1986); Marty (1987); Mullick (1993); Scherrer (1971); Warren (1990); Witzig
(1987); Harris (1996).
254 Chapter 5. DISTILLATION, ABSORPTION
AND STRlPPlNG IN THE PETROLEUM INDUSTRY

5.1.7 Corrosion and Erosion

The crude that is processed in topping units contains a significant amount of
corrosive agents such as:
chlorides (non-organic and sometimes organic),
organic acids,
non-organic salts,
sulfur compounds.
Among the numerous articles dealing with the subject of corrosion, we give
special mention to the one by Danilov (1981).
We are going to examine consecutively the different types of corrosion that
may be involved in crude topping units, and discuss the causes, consequences
and of course the remedies.

5.1.7.1 Corrosion by Salts

A distinction should be made between:
1. corrosion by the salts themselves:
2. corrosion by HCI (and H2S) stemming from hydrolysis of these salts.
1. Since 1975 the diversification in crude sourcing has led refiners to pro-
cess crudes with a higher salt content than before. Today the salt content fre-
quently ranges from 100 to 300 ppm. The increased salt content, often coupled
with the fact that distillation units are operating at reduced capacity, cause
salt deposits to be laid down in piping and exchangers in the preheating train
upstream from the desalter. Salt deposits cause underlying corrosion by dif-
ferential aeration. The remedy consists in injecting part of the desalting water
upstream from the preheating train to make the deposits soluble.
2. Despite the improved efficiency of desalting, the residual chloride con-
tent at the desalter outlet generates HCl by hydrolysis of the salts at high tem-
perature in the furnace (see Vol. 1, Chapter 8). The amount of HCI produced is
basically dependent on the residual MgCI, content. MgCI2is 95% hydrolyzed at
350°C, whereas CaCl, is only approximately 10%hydrolyzed at the same tem-
perature.
An HCl content of 15 to 20 ppm per processed ton is fairly frequent, it can
however vary between 3 ppm for a North Sea crude and 60 ppm for some
Egyptian crudes, at constant desalting efficiency. The remedy consists in neu-
tralizing with a dilute caustic solution. The caustic can neutralize 90 to 95% of
the acid, with the remainder neutralized by ammonia injection at the top of the
tower, completed by additions of a corrosion inhibitor, injected a little before
the water dew point. The corrosion inhibitor is generally made up of two
bases: a filming base, which prevents physical contact between the metal and
the corrosive water, and a neutralizing base, which completes the chemical
action of the ammonia. Inhibitors soluble in hydrocarbons (injected in the
 Chapter 5. DISTILLAJION, ABSORPTION AND STRIPPING IN THE PETROLEUM /NOUSTRY 255

overhead gasoline circuits) are preferable to inhibitors soluble in water, which

may tend to increase the water/metal contact instead.
Furthermore, the condensation of hydrogen sulfide H,S must be avoided,
otherwise a black sludge (iron sulfide) is formed. This occurs in particular
when the water is slightly basic @H: 8-9).
Acid water @H: 4-5) must also be avoided. The best compromise is to keep
the pH in the neighborhood of 6 & 0.3 (Fig. 5.14).

Water corrosivity (arbitrary units)

-
Figure
5.14 Water corrosivity versus pH.

Caustic must be injected with care at the outlet of the desalting unit to pre-
vent steel embrittlement. In this connection, it is necessary to:
make sure the caustic and crude are properly homogenized,
prevent contacts between caustic and steel (maintain a good circulation
velocity to keep the caustic from settling out).
Caustic is sometimes replaced by sodium carbonate or, when possible, by
spent caustic coming from other refining operations.
Ammonia can be replaced by a neutralizing and filming amine combination.
256 Chapter 5. DISTILLATION, ABSORPTION
AND STRIPPING IN THE PETROLEUM INDUSTRY

Tougher metals or metal alloys can also be used such as:

Titanium
Monel (sensitive to H2S)
Hastelloy
Without going into any more detail on these different solutions, we would,
however, like to point out that when salt water is used as cooling water, brass
(admiralty type) for overhead condensers is incompatible with ammonia.
Whatever solutions may be envisaged, proper monitoring (continuous, if
possible) of the pH and the Fe2+and CI- content of condensed water is indis-
pensable in guarding against corrosion. Additionally, an efficient desalter
makes problems much less serious.
H2S corrosion can be estimated on the basis of Nelson or Porte curves
(Nelson, 1958).

5.1.7.2 High Temperature Sulfur Corrosion

This type of corrosion can prove to be very serious in a certain temperature
range for ordinary carbon steel. The temperature range is situated between
300 and 420°C according to the Nelson curves and the corresponding corro-
sion rate can exceed 1 mm per year. When corrosion is estimated to reach
0.3 mm per year, risks of iron sulfide deposit clogging are to be expected, with
potential pyrophory when the unit is shut down. The remedy consists in using
steel alloy of the 5%Cr-0.5%Mo type (particularly for the furnace tubes) in the
above-mentioned temperature range. At high velocity, corrosion is amplified
and steel with 12%or even 18%Cr can be used.

5.1.7.3 Naphthenic Acid Corrosion

Naphthenic acids are compounds of variable molar mass, which are highly cor-
rosive in the 220420°C range. Above this temperature they become dissociated
and corrosion no longer occurs. Crude oil analyses do not usually report the
naphthenic acid content of the crude so experience plays an important role
here in preventing this type of corrosion. Useful indications can be provided by
the crude's acid number which is generally accessible. As soon as the acid num-
ber exceeds 0.5 mg eq. KOH/g, more thorough analyses should be requested.
Corrosion remains slight below 300°C and becomes very severe above
350°C. It is seen in the furnace, transfer line and bottom of the column and is
accentuated by high fluid circulation velocity. An 18/10/3 steel alloy should be
used (at least for furnace tubes) or the crude should be processed mixed with
other less sour crudes.

5.1.7.4 Erosion
Erosion occurs mainly in the transfer line and the overhead line when circula-
tion velocity is high and when mitered elbows are used.
Figure 5.15 recapitulates the precautions that need to be taken to avoid the
different types of corrosion.
 Protect metal against H,S
and HCI attack

Atmospheric
Demulsifying agent tower

$-
Prevent emulsions caused Water
by asphaltic compounds
Desalter r ’ 1 n Steam
b
2
D

81
8 I
I
Desalting water
I
Caustic
Preheating

Transform part of the

Cladding I
- ~
Steam

Steam

remaining salts into

Dissolve inorganic salts stable sodium salts
Wash the crude
I I

I 1
5.15 Summary of precautions to guard against corrosion risks in topping units (Source: ENSPM-flJ
258 Chapter 5. DlSTlLLATION, ABSORPT~ON
AND STRlPPlNG IN THE PETROLEUM /NOUSTRY

5.1.8 Energy Consumption

It is accepted that refineries consume 7 to 9% of the crude they process to
cover their energy needs. Topping and vacuum units account for 20 to 30% of
this figure, with 2/3 for topping and 1/3 for vacuum distillation. They are great
energy guzzlers and particular attention must accordingly be paid to their
design and operation. Their function is fairly simple: supply heat to a cool
product (the crude) in order to vaporize it and recover most of the heat by
means of condensers, pumparounds and product coolants, as products are
cool when sent to storage. The operation is globally speaking irreversible and
part of the heat must be provided by the furnace while some is dissipated in
the air- and water-cooled exchangers. Linhoff (1983) in his many articles and
publications has developed absolutely remarkable tools that can be used to
visualize all the heat exchanges and carry out a detailed analysis.
The basic tool is the enthalpy/temperature diagram, commonly called the
TQ diagram. It is used by many refining professionals and is accessible by
means of simulation software.

5.1.8.1 Building and Using the TQ Diagram

For a given distillation a number of distinct streams with their enthalpy curves
and the inlet/outlet temperatures are available. They are as follows:
streams that are heat sources, called hot streams;
streams that are heat sinks, called cold streams.
The TQ diagram represents two enthalpy curves:
the first cumulating the enthalpies of all the hot streams versus temper-
ature;
the second cumulating the enthalpies of all the cold streams versus tem-
perature (Fig. 5.16A).
With an overall temperature approach fixed between the hot curve and the
cold curve (average minimum admissible approach), the first use of the dia-
gram is to establish the pattern of exchange and the minimum amounts of heat
to be supplied to the process (heating Q 3 and to be evacuated from it (cool-
ing QS> as shown in Figure 5.16B.
The diagram gives the following useful information:
1. The pinch zone is where the temperature between the cold curve and
the hot curve is at a minimum. If an amount of heat dQ on the hot curve
located above the pinch is used to heat cold streams below the pinch, this
heat will have to be provided by the outside hot source in addition to QE and
will be discharged to cooling in addition to Qs. As a result, there is degradation
of heat through the pinch (Fig. 5.17A and B).
2. In contrast, when an amount of available heat underneath the pinch is
transferred above the pinch, heat recovery is improved by a simultaneous
decrease in both QE and Qs. This is the case when a heat pump is used, it
 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY 259

T,

Figure
5.16 TQ diagram. A. Representation of the diagram. B. Determination of amounts
of heat to supply and extract for a given approach (enthalpy flows QEand QS).

absorbs an amount of heat underneath the pinch dQ, consumes work d W, and
discharges an amount dQ + d W above the pinch at a higher temperature. The
amount of heat to be supplied goes from QE to QE - (dQ + d W), and the amount
of heat to be removed from Qs to Qs - dQ. For a heat pump to be efficient, it
must in fact work through the pinch (Fig. 5.18A and B).
3. The temperature approach can vary a lot along the TQ curve. It is then
possible to place utility generation (on the cold curve) in the zone where the
temperature approach is significant, or utility consumption (on the hot curve)
in the zone where the temperature approach is not significant, in order to
260 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY

' QE
Hot curve
i-i
I I
I
I

Q
 Chapter 5 DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY 261

Figure
5.18 TQ diagram. A. Base case. B. Case where an amount of heat taken below the
pinch is used to heat a cold stream above the pinch (heat pump).
262 Chapter 5. DISTILLATION, A6SORPTlON AND STRIPPING IN THE PETROLEUM /NDUSTRY

Hotcurve
1I I
I
I
I
\ I I

I I I
I
I I

Qs + Qz I

/5191 TQ diagram. A. Base case. B. Case with utility consumption on the hot curve.
 Chapter 5. DISTILLATION,
ABSORPTION
AND STRIPPING IN THE PETROLEUM INDUSTRY 263

improve the system’s efficiency by pulling the curves closer. Figure 5.19A
and B represents the latter instance. The utilities are generally steam, hot oil,
etc. This in fact is the same as supplying part of the heat at a thermal level that
is lower than the furnace (and therefore at a more attractive cost), or generat-
ing utilities at a low thermal level that can be used elsewhere.
4. More elaborate calculations based on this diagram can help define
exchangers with the most efficient surface area and incorporate split-flows
properly. This also explains why the diagram is used in revamping to improve
the efficiency of old units.

5.1.8.2 Applications to Topping and Vacuum Distillation Units

The topping and vacuum distillation units are the ones that have the most
streams and the largest exchange trains in terms of complexity and area. The
TQ diagram finds its ideal field of application here. By using the diagram, the
exchange surface area can be reduced at constant efficiency. It can also
improve exchange efficiency by incorporating utility generation and con-
sumption in the best places. For example, it is very advantageous to generate
steam at very low pressure (lower than 3 bar rel.) by means of low level avail-
able heat. The steam is used either for stripping in vacuum distillation or for
stripping sour process water in the immediate distillation neighborhood .
The TQ diagram also allows capital budgeting to be done, for example, the
study of the relative cost of the installed exchange surface (investment) and
the furnace consumption (operating cost).

5.1.8.3 Utilities
Typical utility consumption for a 1 000 t/h (about 150 000 BPSD) crude oil
topping unit is as follows:
Furnaces 100 to 140 thermal MW
Electricity 4to6MW
Cooling water 2 000 m3/h.

5.1.9 Advanced Control. Maintenance. Type of Effluents

5.1.9.1 Advanced Control
The design, operation and maintenance of topping units, vacuum distillation
units, gas plant and gasoline distillation units are economically crucial in
refineries due to their position upstream from other units in the refinery flow
diagram (Mullick, 1993). They must adapt quickly to changes in crude and in
running conditions of downstream units. This is why they are equipped with
advanced control systems which allow faster and more effective action on the
various unit regulation parameters. Available software is listed in the
Advanced Process Control Handbook published by Hydrocarbon Processing.
This software is incorporated in the centralized control system or in separate
computers and can be disabled if need be.
264 Chapter 5. DISTILLATION, ABSORPTION AND STRIPPING IN THE PETROLEUM INDUSTRY

Among the most useful systems, we would like to mention those that can:
Reconcile data.
Control product quality (cut points). Product quality is controlled by
means of various continuous analyzers and acts directly on the cut
points (product flow rates) and stripping steam ratios.
Control the internal reflux by pumparounds. Available heat can be recov-
ered at the most appropriate thermal level for a given fractionation, by
optimizing the heat extracted at different thermal levels. This software
can be connected to the previous software in order to act on the liquid-
vapor traffic governing fractionation.
Monitor heat exchanger fouling. Simply by measuring inlet-outlet
exchanger temperatures, fouled exchangers can be detected and heat
recovery can be optimized in the unit.

5.1.9.2 Maintenance
The latest units are designed to operate for at least 2 or 3 years nonstop.
During this period, corrosion and fouling of the exchanger train are attentively
monitored. Isolation systems can be used to clean an exchanger or repair a
control valve while the unit is running.
For units with a capacity of over 8 million tons a year (over 150000 BPSD),
the exchanger train and desalters are often divided up into two identical trains
followed by two distinct heater cells. This type of set up means that the unit
can operate at reduced capacity while work is being done on the isolated train.

5.1.9.3 Effluents
Topping units produce relatively few effluents and their treatment does not
cause any particular difficulties.
a) Liquid effluents discharged are mainly made up of
desalting water (3 to 8%vol of crude),
sour water coming from condensation of stripping vapors (if it is not uti-
lized directly as make up water for the desalter in order to minimize dis-
charges).
b) Gaseous effluents, the only one being furnace flue gases discharged into
the atmosphere via a stack whose height is calculated depending on prevailing
legislation.

5.2 Vacuum Distillation of the Atmospheric Residue

The purpose of vacuum distillation is to extract distillate cuts from the atmo-
spheric residue coming from the bottom of the topping column. Distillate cuts
are designed to feed conversion units and yield high added value products.
 Chapter 5 DISTILLATION,
ABSORPTION
AND STRIPPING IN THE PETROLEUM INDUSTRY 265

When there are no conversion units downstream these cuts serve to pro-
duce heavy fuel oils, except the lightest cut which can be sent along to the gas
oil pool. The residue can constitute a feed for bitumen manufacture. Lastly,
vacuum distillation of crude can, in some special cases, yield base stocks for
lube oil manufacture.
Vacuum distillation is restricted to fractionation of products whose boiling
point temperature would be too high at a pressure greater than atmospheric
pressure and would cause cracking of the feed. The unit's capacity is 50 to 70%
smaller than the topping unit that supplies it. Depending on the destination of
distillates, vacuum distillation columns may or may not have fractionation
zones. Several technologies are implemented depending on whether steam is
used to lower the partial hydrocarbon pressure:
"Dry" vacuum distillation is without injection of steam. It needs to run at
a very low pressure (10 to 15 mmHg at the top) and requires the use of a
booster ejector before the first condenser as shown in Figure 5.20.
"Wet" vacuum distillation is carried out with injection of steam in the fur-
nace feed and stripping steam in the bottom of the tower. Total pressure
is higher (40 to 60 mmHg at the top). A precondenser is used before the
vacuum system (Fig. 5.21).
"Semi-wet" vacuum distillation is when the only steam is injected at the
bottom of the column. The use of an ejector (called booster ejector) is
often necessary. It is located upstream from the first overhead condenser
and designed to boost process pressure high enough to allow condensa-
tion. Bitumen preparation units are in this category.

5.2.1 Fractionation
Typical cut points for a 350+TBP atmospheric residue are as follows:
vacuum gas oil 350 - 390°C,
vacuum distillate 390 - 550°C.
It should be noted that the race to reduce heavy fuel oil production means
that refiners have to prolong distillates as much as possible to achieve strip-
ping that can exceed 585°C TBP.
A slight fractionation alone is required between the vacuum distillate and
the residue to produce fuel oil, the objective being to get the appropriate flash
point, gravity and viscosity in the final fuel oil product.
The 90 or 95% ASTM D 86 point of LVGO is limited to approximately 360°C
simply to meet the cloud point requirements of commercial gas oils.
The distillate that goes to catalytic cracking must have a limited Conradson
carbon content (see Vol. 3, Catalytic Cracking), and very low metal content,
particularly nickel and vanadium, which are poisons for the catalyst.
The distillates due to go to hydrotreatment/hydrocracking units do not
need to be metal free, but their Conradson carbon content and asphaltene con-
tent must be as low as possible.
 Iu
Power steam Q)
Q,

Booster
Demister . ejector
Purnparound

+-
Heat exchange
zone on LVGO LVGO
b f Noncondensables

pumparoundcooler Overhead drum

Purn around Cracking
on HVGO kerosine
Heat exchange
zone on HVGO HVG: Process water ~

pumparoundcooler
Wash zone
n

Vacuum residue
w

1
Figure
5.20 Flow diagram of a dry vacuum distillation unit (Source: Technip).
 Precondenser
Power steam

3
Demister -d/ 111 ' - Pumparound
-
LVGO
Noncondensables

pumparound cooler

Heat exchange
zone on HVGO - A A A A ~
Pumparound
:
on HVGO
kerosine

Process water

Wash zone __D

pumparound cooler

n
Atmospheric
-------- + Slop cut
Overflash
Steam injected into
feed (dilution steam) Revaporized

Stripping zone

Figure
5.21 How diagram of a wet vacuum distillation unit (Source: Technip).
 Next Page

268 Chapter 5. DISTILLATION, ABSORPTION

AND STRIPPING IN THE PETROLEUM /NOUSTRY

The production of lube oil stocks requires good fractionation between the
different cuts. Bitumens are produced from special crudes.

5.2.2 Examples of Specifications for a Distillate Sent

to Catalytic Cracking
5.2.2.1 Metals
Nickel and vanadium promote coke and hydrogen formation by catalyzing
dehydrogenation reactions. As a result, the olefin content increases while
gasoline yield is lowered. Another result is a gradual deactivation and
increased purging of the catalyst. Copper also causes catalyst activity to slow
down by neutralizing acid sites. When the metal content of the equilibrium cat-
alyst rises from 180 to 3 500 ppm, a drop of 15%is observed in conversion.

5.2.2.2 Conradson Carbon

The Conradson carbon content of distillates sent to cracking is related to the
asphaltene content of the atmospheric residue vacuum distillation feed and to
the presence of carbonaceous residues carried off above the wash zone. These
products tend to attach to the catalyst during the reaction and remain on it in
the regenerator, and this means an increase in the amount of combustion air.
Because of their metal content, these carryover products indirectly influence
catalyst deactivation. Conradson carbon values of less than 0.5% are very
good, and 1 to 2% is acceptable. Not included here are the residue cracking
units which can accommodate a much higher content (see Vol. 3).

5.2.3 Definition of Terms Specific to Vacuum Distillation

These introductory definitions are specific to vacuum distillation:
LVGO: light vacuum gas oil.
MVGO: medium vacuum gas oil.
HVGO: heavy vacuum gas oil.
Cracked hydrocarbons: hydrocarbons produced when the feed is
cracked in the furnace. They are found at the top of the tower with the
noncondensables and the steam injected in the process.
Noncondensables:compounds that can not be condensed by the vac-
uum system. They are made up of
- air entering through leaks;
- noncondensables dissolved in the feed when it has been in transit in an
intermediate storage tank
- light hydrocarbons produced by cracking in the furnace.
Slop cut and overflash: the two together constitute the internal reflux
coming down from the first tray above the feed inlet. Overflash, 3 to 5%
of the feed, is the part sent with the feed liquid to the bottom of the