Crude Unit Process Operations
Group: Economics & Commerce (E&C)
Prepared for
NGHISON REFINERY AND PETROCHEMICAL
VIETNAM (NSRP)
Prepared by
PETROVIETNAM MANPOWER TRAINING
COLLEGE
Vung Tau, 2015
Training material
Training module:
Crude Unit Process Operations
Duration:
3 days
Delivery:
L
Prepared for:
NGHI SON REFINERY & PETROCHEMICAL LLC
Prepared by:
PETROVIETNAM MANPOWER TRAINING COLLEGE
Issued by PVMTC
Version
Prepared
(date)
Checked
(date)
Approved by TMC/ NSRP
Approved
(date)
Page 1 of 60
Reviewed
(date)
Approved
(date)
Table of content
1. Crude distillation ....................................................................................................4
a. Introduction ...........................................................................................................4
b. Unit processing objectives.....................................................................................5
c. Process description ................................................................................................5
2. Operation factors of crude distillation units ........................................................9
a. Fractionation ..........................................................................................................9
b. Cut points ............................................................................................................11
c. Degree of fractionation ........................................................................................12
d. Over-flash ............................................................................................................12
e. Column pressure ..................................................................................................13
f.
Overhead temperature .........................................................................................13
g. Pre-flash columns and crude column capacity ....................................................14
3. Crude oil desalting ................................................................................................14
a. Types of salts in crude oil....................................................................................15
b. Desalting process .................................................................................................16
c. Desalter operating variables ................................................................................18
4. The pumparound ..................................................................................................20
a. Pumparound heat removal ...................................................................................20
b. Purpose of a pumparound ....................................................................................22
c. Pumparounds fractionate .....................................................................................24
5. Vapor flow .............................................................................................................25
a. How top reflux affects vapor flow ......................................................................25
b. Reflux effect on vapor molecular weight ............................................................27
6. Fractionation .........................................................................................................28
a. Improving fractionation .......................................................................................28
b. Flooding the fractionation trays ..........................................................................30
7. Condensers and tower pressure control .............................................................31
a. Effect of subcooling ............................................................................................32
b. Mechanics of subcooling .....................................................................................32
c. Pressure control ...................................................................................................40
8. NSRP’s refinery CDU ...........................................................................................49
a. General information: ...........................................................................................49
Page 2 of 60
b. Feed and products ................................................................................................49
c. Crude desalting ....................................................................................................50
d. Unit description: ..................................................................................................52
e. Some notes for unit operation .............................................................................58
Page 3 of 60
1. Crude distillation
a. Introduction
Crude distillation unit (CDU) is at the front-end of the refinery, also known as
topping unit, or atmospheric distillation unit. It receives high flow rates hence its
size and operating cost are the largest in the refinery. Many crude distillation
units are designed to handle a variety of crude oil types. The design of the unit is
based on a light crude scenario and a heavy crude scenario. The unit should run
satisfactorily at about 60% of the design feed rate. Seasonal temperature
variation should be incorporated in the design because changes in the cut point
of gasoline can vary by 20o C between summer and winter.
The capacity of the CDU ranges from 10,000 barrels per stream day(BPSD) or
1400 metric tons per day (tpd) to 400,000 BPSD (56,000metric tpd). The
economics of refining favours larger units. A good size CDU can process about
200,000 BPSD (that is the case of NSRP refinery). The unit produces raw
products which have to be processed in downstream unit to produce products of
certain specifications. This involves the removal of undesirable components like
sulphur, nitrogen and metal compounds, and limiting the aromatic contents.
Typical products from the unit are:
 Gases
 Light straight run naphtha (also called light gasoline or light naphtha)
 Heavy gasoline (also called military jet fuel)
 Kerosene (also called light distillate or jet fuel)
 Middle distillates called diesel or light gas oil (LGO)
 Heavy distillates called atmospheric gas oil (AGO) or heavy gas
oil(HGO)
Page 4 of 60
 Crude column bottoms called atmospheric residue or topped crude.
b. Unit processing objectives
The objective of the CDU is to provide primary separation of crudes to produce
straight run blendstocks of distillate products (after suitable downstream
treatment processes) and feedstocks for other downstream process units.
c. Process description
The process flow diagram of a typical crude distillation unit is shown in Figure
1. Crude oil is pumped from storage tanks where it is freed from sediments and
free water by gravity.
It goes through a series of heat exchangers where it is heated with hot products
coming out from the distillation column and by the exchange with heat from the
pumparound liquid streams. The temperature of the crude feed can reach 120–
150oC.
The crude oil contains salt in the form of dissolved salt in the tiny droplet of
water which forms a water-in oil emulsion. This water cannot be separated by
gravity or through mechanical means. It is separated through electrostatic water
separation. This process is called desalting. In the electrostatic desalter, the salty
water droplets are caused to coalesce and migrate to the aqueous phase by
gravity. It involves mixing the crude with dilution water (5–6 vol%) through a
mixing valve.
The crude is further heated in product heat exchangers. The preheating of the
crude using the hot products cools down the products to the desired temperature
for pumping to the storage tanks. This is essential for the economics of the unit
in terms of energy conservation and utilization.
Page 5 of 60
Figure 1. A typical atmospheric crude distillation unit.
Page 6 of 60
Of course, preheating is not enough, as the crude has to be partially vaporized to
the extent that all products, except for the atmospheric residue have to be in the
vapor phase when the crude enters the atmospheric distillation column. Thus a
furnace is required to boost the temperature to between 330 and 385oC
depending on the crude composition.
The partially vaporized crude is transferred to the flash zone of the column
located at a point lower down the column and above what is called the stripping
section. The main column is typically 50 m high and is equipped with about 30–
50 valve trays.
The vapor goes up in tremendous amounts and at a high flow rate, necessitating
a large diameter column above the flash zone. At the bottom of the stripping
section, steam is injected into the column to strip the atmospheric residue of any
light hydrocarbon and to lower the partial pressure of the hydrocarbon vapors in
the flash zone.
This has the effect of lowering the boiling point of the hydrocarbons and causing
more hydrocarbons to boil and go up the column to be eventually condensed and
withdrawn as side streams. As the hot vapors from the flash zone rise through
the trays up the column, they are contacted by the colder reflux down the
column. In the overhead condenser, the vapors are condensed and part of the
light naphtha is returned to the column as reflux. Further reflux is provided by
several pumparound streams along the column. In the distillation tower, heat
required for separation is provided by the enthalpy of the feed. For effective
separation heat has to be removed from the tower, in this case, by the overhead
condenser and several pumparound streams along the tower length.
The pumparound stream is a liquid withdrawn at a point below a side stream
tray that is cooled by the cold crude feed as part of the preheat exchangers train.
It is then returned to the column a few trays above the draw tray. This
pumparound cooling accomplishes a number of tasks.
Page 7 of 60
First, the cold liquid condenses more of the rising vapors thus providing more
reflux to compensate for the withdrawal of products from the column. Second,
heat is removed from the column at higher temperatures. This is in addition to
the heat removal from the condenser which takes place at relatively lower
temperatures, thus the thermal efficiency of the column is improved and the
required furnace duty is reduced.
Third, pumparound streams reduce the vapor flow rate throughout the column.
Therefore, the required column is smaller than what would otherwise be
required if pumparound streams where not there. The drawback to using more
pumparound streams is that they tend to reduce the fractionation because a more
fractionated liquid is mixed after cooling with a less fractionated liquid a few
trays above.
The side draw products are usually stripped to control their initial boiling point.
The strippers contain several trays and the stripping is done using steam at the
bottom of the stripper or reboiler type side stream strippers. The end boiling
point of the side stream is controlled by the flow rate of the side stream product.
Table 1. Tray distribution in a crude distillation unit
Zone
Number of trays
Overhead product to kerosene
10
Kerosene to light gas oil
8
Light gas oil to heavy gas oil
6
Heavy gas oil to flash zone
6
Flash zone to atmospheric residue
6
Pumparounds
3–4
The overhead vapor is condensed at the top of the tower by heat exchange with
the cool crude coming into the unit and by air and cooling water. The liquid
product is called light straight run naphtha. Part of this product is returned to the
Page 8 of 60
column as an external reflux. Down the column, other products are withdrawn,
such as heavy straight run naphtha, kerosene or jet fuel, LGO and HGO. All of
these products are withdrawn above the feed tray. The atmospheric residue is
withdrawn from the bottom of the column.
The main column is equipped with between 30 and 50 valve trays. Typical
designs have the trays distribution between products as shown in Table 1.
2. Operation factors of crude distillation units
The CDU can be looked at from the point of view of a process engineering as a
multicomponent distillation column. Indeed, the commercial process simulation
program models CDU as a case of multicomponent distillation with undefined
pseudo-components instead of the normally encountered defined components.
However, because we are dealing with a mixture of thousands of compounds
and due to the limitation of any distillation column in terms of its capacity to
fractionate these components, there are specific operational aspects which
characterize the CDU operation. In addition, there are some practical aspects in
meeting the required specifications and boiling range of the required
transportation fuels. In this section, the factors which affect the design and
operation of the unit are explored.
a. Fractionation
The degree of fractionation in a crude unit is determined by the gap or overlap
between two adjacent side stream products. Hence we can talk about the gap or
overlap in the boiling point range between kerosene and LGO for example.
In the ideal case there would be no overlap between these products and the end
boiling point of kerosene would be the initial boiling point of the LGO.
However, if we compare the ASTM distillation boiling points, and since ASTM
distillation does not give perfect fractionation, the ASTM end point of kerosene
Page 9 of 60
is higher than the initial ASTM boiling point of LGO. This is called
fractionation overlap.
Since determining the initial and end point on the laboratory test is not always
possible or accurate, the fractionation gap is defined as the difference between
the ASTM 5% boiling point of the product and the 95% point of the lighter
product. When this difference is positive, we have a gap indicating good
fractionation.
Figure 2. GAP and OVERLAP
Page 10 of 60
A negative difference is called an overlap indication that some of the light
product is still in the heavier product and vice versa. Figure 2 shows the gap and
overlap concept. By controlling the cut point of any two consecutive products
we can affect the degree of fractionation.
b. Cut points
The cut points in the CDU are controlled by the overhead vapor temperature
which determines how much vapor goes to the condensers to produce light
naphtha and by the flow rate of the various products straight from the column or
the side stream strippers. The atmospheric residue level control inside the
column determines its flow rate and thus its initial cut point. The amount of light
naphtha is determined by the dew point of the naphtha at its partial pressure,
which is close to the overhead temperature. Changing the draw off rate of any
product affects the cut points of the heavier product below it. For example,
lowering the kerosene flow rate will lower its end point (make it lighter), but
will also modify the initial cut points of the LGO and HGO and the initial cut
point of the atmospheric residue. The residue flow rate, the internal reflux rate,
the draw off temperatures and the pumparounds are also affected.
Therefore, if the cut point of one stream is changed through a change in its
withdrawal rate, the flow rate of the heavier product next to it should be changed
in the reverse and by the same amount in order to make the changes in the
desired stream only. For example, if the end point of kerosene is lowered by
decreasing the kerosene flow rate by a certain amount, the flow rate of LGO has
to be increased by the same amount. If this action is taken, only the cut point of
kerosene is affected and the cut points of the other products remain unchanged.
The side stream rate also affects the temperature at the withdrawal tray and
lowers the internal reflux coming out of that tray. The internal reflux rate affects
the degree of fractionation. It can be increased by increasing the heater outlet
temperature, and by lowering the pumparound duty in the lower section of the
Page 11 of 60
column. When less heat is removed by the lower pumparound, more vapors will
be available up the column and more internal reflux is produced as the vapors
are condensed.
c. Degree of fractionation
The fractionation quality between two consecutive streams is affected by several
factors such as the vapor and liquid flow rates in the column zone between these
two streams, the number of trays, and the heat extracted by the pumparound.
Fractionation quality is formulated in terms of gap or overlap of the products.
For perfect fractionation, zero gap and overlap are required. This means that the
EBP of the light cut would be the IBP of the heavier cut and so on.
d. Over-flash
In order to fractionate the crude oil into the various products, it has to be heated
to a temperature between 330 and 385oC, depending on the crude composition.
The partially vaporized crude is transferred to the flash zone of the column
located at a point lower down the column. The furnace outlet temperature should
be enough to vaporize all products withdrawn above the flash zone plus about
3–5 vol% of the bottom product. This over-flash has the function of providing
liquid wash to the vapors going up the column from the flash zone, and
improving fractionation on the trays above the flash zone, thereby improving the
quality of the HGO and reducing the overlap with the bottom products below the
flash zone.
This necessitates that there must be few trays in the region between the flash
zone and the HGO draw off. The over-flash provides heat input to the column in
excess to that needed to distill the overhead products. It also prevents coke
deposition on the trays in the wash zone. The furnace outlet temperature is
controlled to keep coking inside the furnace tubes and in the column flash zone
to a minimum. However, the composition of the crude plays a part in
Page 12 of 60
determining the maximum temperature allowed. Paraffinic crude oils cracks
more readily than an aromatic or asphalt-base crude. Therefore, the furnace
outlet temperature for paraffinic crude oils is lower than that for other crude
types.
e. Column pressure
The pressure inside the CDU column is controlled by the back pressure of the
overhead reflux drum at about 0.2–0.34 bar gauge. The top tray pressure is 0.4–
0.7 bar gauge higher than the reflux drum. The flash zone pressure is usually
0.34–0.54 bar higher than the top tray.
Figure 3.Tower pressure profile.
f. Overhead temperature
The overhead temperature must be controlled to be 14–17oC higher than the dew
point temperature for the water at the column overhead pressure so that no liquid
Page 13 of 60
water is condensed in the column. This is to prevent corrosion due to the
hydrogen chloride dissolved in liquid water (hydrochloric acid).
g. Pre-flash columns and crude column capacity
The crude flow rate to the CDU determines the capacity of the whole refinery. A
crude column is typically designed for 80% loading, which means that the unit
can be operated at 20% throughput more than the design value. The capacity of
the column is limited by the vapor flow rate with a velocity between 2.5 and 3.5
ft/s (0.76 and 1.07 m/s). The vapor flow rate increases as the vapors rise from
the flash zone to the overhead. To keep the vapor velocity within the limits
mentioned above, the pumparounds, which are installed at several points along
the column, extract heat from the column.
This results in condensing the rising vapors and reducing the vapor velocity. To
expand crude capacity, the most used technique is to introduce a pre-flash
column before the crude heater. The crude oil after preheating in the hot
products and pumparound heat exchangers is flashed into a column where the
lightest products are removed. The bottoms from the pre-flash column are
introduced into the crude heater and then to the crude column.
The amounts of the light ends in the crude are now less, and this reduces the
vapor loading up the column. Although the unit throughput is increased, the
furnace duty is not increased, since the crude rate going to the furnace is not
affected due to the removal of the light ends. Pre-flash columns are also
introduced in the original design of the CDU when the crude oil is light, and
when it contains a lot of light ends in the naphtha range.
3. Crude oil desalting
When the crude oil enters the unit, it carries with it some brine in the form of
very fine water droplets emulsified in the crude oil. The salt content of the crude
Page 14 of 60
measured in pounds per thousand barrels (PTB) can be as high as 2000.
Desalting of crude oil is an essential part of the refinery operation.
The salt content should be lowered to between 5.7 and 14.3 kg/1000 m3 (2 and 5
PTB). Poor desalting has the following effects:
 Salts deposit inside the tubes of furnaces and on the tube bundles of heat
exchangers creating fouling, thus reducing the heat transfer efficiency;
 Corrosion of overhead equipment; and,
 The salts carried with the products act as catalyst poisons in catalytic
cracking units.
a. Types of salts in crude oil
Salts in the crude oil are mostly in the form of dissolved salts in fine water
droplets emulsified in the crude oil. This is called a water-in-oil emulsion, where
the continuous phase is the oil and the dispersed phase is the water. The water
droplets are so small that they cannot settle by gravity. Furthermore, these fine
droplets have on their surfaces the big asphaltene molecules with the fine solid
particles coming from sediments, sands or corrosion products. The presence of
these molecules on the surface of the droplets acts as a shield that prevents the
droplets from uniting with each other in what is called coalescence. The salts
can also be present in the form of salts crystals suspended in the crude oil. Salt
removal requires that these salts be ionized in the water. Hence, wash water is
added to the crude to facilitate the desalting process as will be explained later.
Going back to the subject of salt types, these are mostly magnesium, calcium
and sodium chlorides with sodium chloride being the abundant type. These
chlorides, except for NaCl, hydrolyze at high temperatures to hydrogen chloride:
CaCl2+ 2H2O  Ca(OH)2+2HCl
MgCl2+ 2H2O  Mg(OH)2+2HCl
Page 15 of 60
On the other hand, NaCl does not hydrolyze. Hydrogen chloride dissolves in the
overhead system water, producing hydrochloric acid, an extremely corrosive
acid.
b. Desalting process
To remove the salts from the crude oil, the water-in oil emulsion has to be
broken, thus producing a continuous water phase that can be readily separated as
a simple decanting process. The process is accomplished through the following
steps:
 Water washing: Water is mixed with the incoming crude oil through a
mixing valve. The water dissolves salt crystals and the mixing
distributes the salts into the water, uniformly producing very tiny
droplets. Demulsifying agents are added at this stage to aide in
breaking the emulsion by removing the asphaltenes from the surface of
the droplets.
 Heating: The crude oil temperature should be in the range of 48.9–
54.4oC since the water–oil separation is affected by the viscosity and
density of the oil.
 Coalescence: The water droplets are so fine in diameter in the range of
1–10 mm that they do not settle by gravity. Coalescence produces
larger drops that can be settled by gravity. This is accomplished
through an electrostatic electric field between two electrodes. The
electric field ionizes the water droplets and orients them so that they
are attracted to each other. Agitation is also produced and aides in
coalescence. The force of attraction between the water droplets is
given by:
𝑑
F = K 𝐸 2 𝑑 2 ( )4
𝑠
where E is the electric field, d is the drop diameter and s is the distance between
drops centres and K is a constant.
Page 16 of 60
 Settling: According to Stock’s law the settling rate of the water
droplets after coalescence is given by
where r is the density m is the viscosity, d is the droplet diameter and k is a
constant.
Description of Desalter
A typical desalter contains two metal electrodes as shown in Figure 4. A high
voltage is applied between these two electrodes. For effective desalting the
electric fields are applied as follows:
Figure 4. Simplified flow diagram of an electrostatic desalter
Page 17 of 60
Figure 5. Two-stage desalting
 A high voltage field called the ‘‘secondary field’’ of about 1000 V/cm
between the two electrodes is applied. The ionization of the water
droplets and coalescence takes place here
 A primary field of about 600 V/cm between the water–crude interface
and the lower electrode is applied. This field helps the water droplets
settle faster.
The desalter of this design achieves 90% salt removal. However 99% salt
removal is possible with two-stage desalters as shown in Figure 5. A second
stage is also essential since desalter maintenance requires a lengthy amount of
time to remove the dirt and sediment which settle at the bottom. Therefore, the
crude unit can be operated with a one stage desalter while the other is cleaned.
c. Desalter operating variables
For an efficient desalter operation, the following variables are controlled:
 Desalting temperature: The settling rate depends on the density and
viscosity of the crude.
 Since increasing the temperature lowers the density and viscosity, the
settling rate is increased with temperature based on the crude gravity,
typical desalting temperature can vary between 50 and 150oC.
Page 18 of 60
 Washing water ratio: Adding water to the crude oil helps in salt
removal.
Hence, increasing the wash water rate increases the coalescence rate. Depending
on the desalting temperature, a minimum value should be used. For example,
Kuwait crude (31.2 API) requires 7–8 vol% water addition relative to the crude
rate.
 Water level: Raising the water level reduces the settling time for the
water droplets in the crude oil, thus improving the desalting efficiency.
 However, if the water level gets too high and reaches the lower
electrode, it shorts out the desalter. Since the primary electric field
depends on the distance between the lower electrode and the water–
crude interface, it is always better to keep the level constant for stable
operation.
 Washing water injection point: Usually the washing water is injected
at the mixing valve. However, if it is feared that salt deposition may
occur in the preheat exchangers, part or all of the washing water is
injected right after the crude feed pump.
 Demulsifier injection rate: Demulsifiers are basic copolymers with one
end being hydrophilic (loves water and attaches to the surface of the
water droplet), and the other end being hydrophobic (loves the oil and
is directed to the oil side).
 When these compounds are adsorbed on the droplet surface, they
stabilize the droplet. The demulsifier is added to the crude after the
feed pump or before the mixing valve at levels between 3 and 10 ppm
of the crude.
 Type of washing water: Process water in addition of fresh water is
used for desalting. The water should be relatively soft in order to
prevent scaling. It should be slightly acidic with a pH in the range of 6.
It should be free from hydrogen sulphide and ammonia so as to not
Page 19 of 60
create more corrosion problems. Therefore, distillation overhead
condensates and process water from other units can be used after
stripping.
 Pressure drop in the mixing valve: Mixing the washing water with
crude oil is necessary in order to distribute the water and dissolve any
suspended salts crystals. The pressure drop across the mixing valve
determines the mixing efficiency.
 On the other hand, the mixing process produces finer (smaller
diameter) droplets which tend to stabilize the emulsion and make
water separation more difficult. Therefore, there is a compromise in
selection of the appropriate pressure drop across the mixing valve. A
pressure drop between 0.5 and 1.5 bar (7.4 and 22 psi) is used.
One variable which is not mentioned above is the desalter pressure. The
operation of the desalter requires that the crude be in the liquid phase during
desalting. A typical pressure of 12 bar is necessary to achieve this purpose.
When the process control variables are properly adjusted, a 90% salt rejection
(2–5 PTB of salts in the desalted crude relative to the raw crude) can be
achieved. With a two stage operation the salt rejection can reach 99%. Any
remaining salts are neutralized by the injection of sodium hydroxide which
reacts with the calcium and magnesium chloride to produce sodium chloride.
CaCl2+ 2NaOH  Ca(OH)2+ 2NaCl
NaCl does not hydrolyze to the corrosive hydrogen chloride.
4. The pumparound
a. Pumparound heat removal
Figure 7 shows an alternate method, called circulating reflux or pumparound, to
remove heat from a tower. Hot liquid, at 500°F, is drawn from tray 10, which is
called the pumparound draw tray. The liquid pumparound is cooled to 400°F.
Page 20 of 60
The cooled liquid is returned to the tower at a higher elevation onto tray 9. It
appears from Fig. 7 that the cold 400°F pumparound return liquid is entering the
downcomer from tray 8. This is often good design practice. Tray 9 is called the
pumparound return tray.
Figure 6. A pumparound or circulating reflux.
The purpose of the pumparound is to cool and partially condense the upflowing
vapors. The vapors to pumparound tray 10 are at 600°F. The vapors from the
pumparound return tray 9 are at 450°F.
There are two pumparound trays (9 and 10) in the column. This is the minimum
number used. A typical number of pumparound trays is 2 to 5.
Page 21 of 60
b. Purpose of a pumparound
The circulating pumparound is cooling the vapor flowing through tray 10 from
600 to 450°F as it leaves tray 9. The tower-top reflux flow is controlling the
tower-top temperature. If we were to reduce the pumparound circulation rate,
less heat would be extracted from trays 9 and 10. More, and hotter, vapor would
flow up the tower. The top reflux temperature control valve would open. The top
reflux rate would go up. The vaporization of reflux on the top tray would
increase. The overhead condenser duty would increase.
The decrease in the heat duty of the pumparound heat exchanger would equal
the increase in the heat duty of the overhead condenser. The heat balance of the
tower is preserved. Some of the heat that was being recovered to the cold fluid,
shown in Figure 8, is now lost to cooling water in the overhead condenser. This
shows the most important function of pumparounds: recovering heat to a process
stream that would otherwise be lost to the cooling tower.
The cooling-water outlet temperature from the condenser was 140°F. This is
bad. The calcium carbonates in the cooling water will begin to deposit as waterhardness deposits inside the tubes.
It is best to keep the cooling-water outlet temperature below 125°F to retard
such deposits. Increasing the pumparound heat removal will lower the coolingwater outlet temperature.
Another purpose of the pumparound is to suppress top-tray flooding. If tray 1 or
2 in Figure 7 floods, the operator would observe the following:
 The tower-top temperature would increase.
 The distillate product would become increasingly contaminated with
heavier components. If this were a refinery crude fractionator, we would
say that the endpoint of the naphtha overhead product would increase.
 The pressure drop across the top few trays would increase.
Page 22 of 60
 The liquid level in the reflux drum would increase.
Figure 7. Two ways of removing heat from a tower.
If the operator increases the reflux rate to reduce the tower-top temperature, the
top temperature will go up rather than down. This is a positive indication of toptray flooding. The correct way to suppress top tray flooding is to increase the
pumparound duty. This can be done by increasing the cold-fluid flow through
the pumparound heat exchanger, or the pumparound flow itself could be
increased. Either way, the flow of vapor flowing up to tray 8 will decrease. The
flow of vapor through trays 1 to 7 will also decrease. The low vapor velocity
will reduce the tray pressure drop. The ability of the vapor to entrain liquid will
be reduced. The height of liquid in the downcomer will be reduced, and tray
flooding will be suppressed.
Page 23 of 60
Increasing pumparound heat duty will unload the overhead condenser. This will
cool off the reflux drum. A colder reflux drum will absorb more gas into the
distillate product. Less gas will be vented from the reflux drum, and this is often
desirable.
Heat recovered in the pumparound heat exchanger is often a valuable way to
recover process heat. Heat not recovered in the pumparound exchanger is lost to
cooling water in the overhead condenser.
c. Pumparounds fractionate
The process design engineer typically assumes that a pumparound is simply a
way to extract heat from a tower. The engineer does not expect the trays used to
exchange heat between the hot vapor and the cold liquid to also aid in
fractionation. In practice, this is not what happens.
Figure 8. Pumparound trays do fractionate
Let’s refer back to Fig. 7. Note that the vapor temperature leaving tray 9 is
450°F. The temperature of the liquid leaving tray 10 is 500°F. This sort of
temperature difference shows that fractionation is taking place across the
pumparound trays. The temperature difference between:
Page 24 of 60
∆T = (temperature of liquid leaving a lower tray) minus (temperature of vapor
leaving a higher tray)
is a measure of the amount of fractionation. The bigger this temperature
difference, the more the fractionation that is taking place across the trays.
As the pumparound rate is increased, tray efficiency is improved. However, at
some point, the pumparound liquid flow becomes too great. Probably, at this
point the downcomers start to back up. Tray efficiency is impaired because of
this downcomer flooding. The temperature difference between the liquid leaving
the pumparound draw tray minus the temperature of the vapor leaving the
pumparound return tray becomes smaller. This point is called the incipient flood
point for the pumparound trays.
5. Vapor flow
a. How top reflux affects vapor flow
Let’s assume that the vapor flow into a tower is constant. Figure 9 shows such a
situation. Both the pounds per hour and temperature of the vapor flowing up to
tray 9 are constant. This means that the heat flow into the tower is constant.
Now, let’s increase the reflux rate. Certainly, the result will be:
 The tower-top temperature will decrease.
 The gasoline overhead product flow will also decrease.
When we increase the reflux rate, the tower-top temperature drops - let’s say
from 300 to 240°F. Actually, the temperature of the vapor leaving all the trays in
the tower will decrease. The effect is bigger on the top tray, and gradually gets
smaller as the extra reflux flows down the tower. If the top-tray temperature has
dropped by 60°F, then the vapor temperature leaving tray 9 might drop by only
5°F. Let’s assume that the extra reflux causes the temperature of the vapor from
tray 4 to decrease by 40°F. The sensible-heat content of the vapor has decreased.
Page 25 of 60
Figure 9. Effect of top reflux on vapor flow.
Sensible heat is a measure of the heat content of a vapor due to its temperature.
If the specific heat of the vapor is 0.5 Btu/[(lb)(°F)], then the decrease in the
sensible-heat content of the vapor when it cools by 40°Fis 20 Btu/lb.
A small portion of this 20 Btu is picked up by the increased liquid flow leaving
tray 4. The main portion of this heat is converted to latent heat. This means that
some increment more of the liquid on the tray turns into a vapor. But where does
this extra liquid, which vaporizes on the tray, come from? It comes from the
extra top reflux.
The vaporization of the extra reflux cools the tray. The extra vapor generated
adds to the vapor flow from the tray. This increases the vapor flow from the
tray. Even though the heat flow into the tower is constant, increasing the top
reflux does increase the pounds of vapor flowing up the tower.
Page 26 of 60
b. Reflux effect on vapor molecular weight
Let’s assume the following, for the tower shown in Fig. 9:
 Constant vapor flow to tray 9
 Tower-top reflux flow increased
 Pumparound duty constant
As a result of the increased reflux rate, the
 Tower-top temperature drops
 Gasoline flow drops
It will result in an increase in the mass flow of the vapor through trays 3, 4, and
5. But what will happen to the molecular weight of the vapor? The vapor’s
molecular weight will decrease as the reflux rate is increased. The vapor leaving
each tray is in equilibrium with the liquid. This means that the vapor leaving
each tray is at its dew point and the liquid leaving each tray is at its bubble point.
As the top reflux rate is increased, all the trays are cooled. The vapors leaving
trays 3, 4, and 5 are cooled. As a vapor at its dew point cools, the heavier
components in the vapor condense into a liquid. The remaining vapors have a
lower molecular weight because they are lighter. As the heavier components in
the vapor condense into a liquid, they give off heat. This heat is called the latent
heat of condensation. This latent heat is picked up by the liquid flowing across
the tray. This liquid flow is called the internal reflux. This latent heat promotes
extra vaporization of the internal reflux. Naturally, the lighter, lower-boiling
point components preferentially vaporize from the internal reflux.
These lighter components have a relatively low molecular weight.
The uncondensed vapors flowing from the tray below, plus the newly vaporized
vapors from the reflux, flow to the tray above. The combined molecular weight
Page 27 of 60
of vapors is thus reduced. As the molecular weight decreases, the volume of
each pound of vapor increases.
As the molecular weight of the vapor decreases, the density of the vapor
decreases. As the density of a vapor is reduced, each pound of vapor occupies
more volume. It is true that cooler vapors do occupy less volume per pound than
do warmer vapors of the same composition. That is, gases and vapors tend to
contract on cooling and expand on heating. However, this is a small effect
compared to the increase in the volume of vapors, due to the decrease in the
vapor’s molecular weight.
6. Fractionation
a. Improving fractionation
A refinery crude distillation tower is producing gasoline, truck diesel, and a gas
oil. The diesel is contaminated with the gas oil. Also, the gas oil is contaminated
with the lighter diesel. As shown in Figure 10, the vapor flow into the tower is
constant. Our job is to improve the degree of fractionation between diesel and
gas oil. Our objective is to remove the relatively heavy gas oil from the diesel
and to remove the lighter diesel from the gas oil. We could reduce the amount of
diesel product from the tower.
That could wash the heavier gas oil out of the diesel. But it would also increase
the amount of diesel in the gas oil. Increasing the heat removed in the
pumparound would have a similar effect: less gas oil in diesel, but more diesel
in gas oil.
Page 28 of 60
Figure 10. Decreasing pumparound improves fractionation
Reducing the pumparound heat-removal duty increases the vapor flow from tray
8 in the column shown in Fig. 10. The extra pounds of vapor flow up the tower,
and raise the tower-top temperature. The reflux control valve opens to cool the
tower-top temperature back to its temperature set point. Then the liquid flow
rates, from trays 1, 2,and 3, onto tray 4, all increase. If the diesel draw-off rate is
maintained constant, the liquid overflow rate onto trays 5, 6, and 7 will increase.
This liquid flow is called the internal reflux. Trays 5, 6, and 7 are the trays that
fractionate between diesel and gas oil. The more efficiently they work, the less
the contamination of the adjacent products.
Page 29 of 60
The way we increase the fractionation efficiency of trays is to make the trays
work harder. The correct engineering way to say this is: “To improve the
separation efficiency between a light and heavy product, the vapor flow rate
through the trays is increased, and the internal reflux flowing across the trays is
increased.”
Again, this improvement in the degree of fractionation developed by trays 5, 6,
and 7 is a result of reducing the amount of heat duty removed by the
pumparound flowing across trays 8, 9, and 10.
b. Flooding the fractionation trays
Reducing the pumparound duty increases the tray loadings on trays 1 through 7.
But in so doing, the trays operate closer to their incipient flood point. This is
fine. The incipient flood point corresponds to the optimum tray performance.
But if we cross over the incipient flood point, and trays 5, 6, and 7 actually start
to flood, their fractionation efficiency will be adversely affected. Then, as we
decrease the pumparound heat-removal duty, the mutual contamination of diesel
and gas oil will increase.
From an operating standpoint, we can see when this flooding starts. As we
decrease the pumparound duty, the temperature difference between the dieseland gas-oil product draws should increase. When these two temperatures start to
come together, we may assume that we have exceeded the incipient flood point,
and that trays 5, 6, and 7 are beginning to flood.
Page 30 of 60
7. Condensers and tower pressure control
The total condensation of a vapor to a liquid is best illustrated by the
condensation of steam to water. Steam flowed from the boiler in the basement.
The steam condensed inside the radiator, and flowed back into the boiler,
through the condensate drain line. This is a form of thermosyphon circulation.
The driving force for the circulation is the differential density between the water
in the condensate drain line and the steam supply line to the radiator.
The bigger the radiator, the more heat is provided to a room. The bigger the
radiator, the faster the steam condenses to water inside the radiator. A larger
radiator has more heat-transfer surface area exposed to the condensing steam.
Unfortunately, the radiator shown in Fig. 11 is suffering from a common
malfunction. Water-hardness deposits have partly plugged the condensate drain
line. Calcium carbonate is atypical water-hardness deposit. Banging on the
radiator often breaks loose the carbonate deposits in the condensate drain line.
The steam condensate, which has backed up in the radiator, now empties. This
exposes more of the interior surface area of the radiator to the condensing steam.
Figure 11. The surface area of the radiator
Page 31 of 60
It rather seems that 40 percent of the surface area of the radiator in Figure 11 is
submerged under water. If the water is drained out, the rate of steam
condensation will increase by the same 40 percent.
− Subcooling, Vapor Binding, and Condensation
− Subcooling
a. Effect of subcooling
When steam condenses at atmospheric pressure, it gives off 1000 Btu/lb of
condensing steam. This is called the latent heat of condensation of steam.
When water cools off from 220 to 120°F, it gives off 100 Btu/lb of water. This
heat represents the sensible-heat content of water between 220 and 120°F.
It takes less of the radiator’s surface area to condense 1 lb of steam at 220°F
than to cool off 1 lb of water from 220 to 120°F. And this is true even though
the condensation of steam generates 10 times as much heat as the cooling of hot
water. It is a lot easier to condense steam than to cool water. This also explains,
then, why condensate back up reduces the rate of heat transfer and condensation.
b. Mechanics of subcooling
As the condensed steam flow out of the radiator is restricted, the surface area of
the radiator available to cool the hot water increases. Hence, the water
temperature leaving the radiator decreases. To summarize, the effect of
restricting the condensate flow from a radiator or condenser is to:
Build water level in the radiator
Reduce the rate of latent-heat transfer from the steam
Increase the rate of sensible-heat transfer from the condensate
Reduce the overall heat-transfer duty from the radiator
Page 32 of 60
Incidentally, the correct way to remove hardness deposit is by chemical
cleaning. Violent banging on radiators is considered bad form in South
Brooklyn.
 Air Lock
Vapor binding, or air lock, is another common cause of house hold radiator
malfunction. Often, the vapor accumulating in the radiator is CO2, rather than
air. The CO2 originates from the thermal decomposition of carbonates in the
boiler. Regardless, air and CO2 form a noncondensable vapor in the radiator.
These noncondensables mix with the steam in the radiator. The noncondensables
then reduce the concentration of the steam, by dilution. The diluted steam has a
lower partial pressure than pure steam. The lower the partial pressure of the
steam, the more difficult it is to condense. As the rate of condensation of the
steam drops, so does the heat radiated by the radiator.
To restore the efficiency of a radiator suffering from the accumulation of
noncondensables inside its condensing coils, the noncondensable gases have to
be removed. The air vent shown in Figure 11 serves this purpose.
To summarize, the two most common malfunctions of a steam condenser (or
radiator) are:
 Condensate backup
 Noncondensable accumulation
And these two malfunctions are also the most common problems we encounter
in the design and operation of shell-and-tube heat exchangers used in total
condensation service.
 Condensation and Condenser Design
Page 33 of 60
Condensation in Shell-and-Tube Heat Exchangers
A steam reboiler has the same problems and works on the same principles as a
process condenser. The only difference is that a steam reboiler’s heat is removed
by the shell-side process fluid, and a process condenser’s heat is removed by
cooling water.
Figure 12 is a sketch of a depropanizer overhead condenser. Let’s make a few
assumptions about this shell-and-tube condenser:
− The propane is totally condensed as it enters the reflux drum.
− There is no vapor vented from the reflux drum, but there is a vapor-liquid
interface in the drum.
Figure 12. Total condensation below the reflux drum.
− The reflux drum is elevated by 20 ft above the top of the condenser.
− We are dealing with pure (100 percent) propane.
If the pipe to the condenser maintained a liquid level, then the shell side of the
condenser would be full of propane.
Page 34 of 60
But if the shell side of the condenser were really liquid full, the tubes would not
contact the vapor. If the tubes do not contact the vapor, then the rate of
condensation is zero. Perhaps a small amount of heat transfer would take place,
as the liquid propane became subcooled. But none of the propane vapor would
condense.
Therefore, the liquid level in the overhead condenser would have to be
somewhere in the condenser’s shell. But then the liquid in the condenser would
be below the reflux drum. How, then does the liquid get from the lower
elevation of the condenser to the higher elevation in the reflux drum? But for
now, we can say that most reflux drums are elevated 20 or 30 ft above grade to
provide net positive suction head (NPSH) for the reflux pump. Also, most shelland-tube condensers are located at grade, for easier maintenance during unit
turnarounds.
Subcooling in a Shell-and-Tube Condensers
Figure 13 is the same propane condenser shown in Figure 12. Let’s assume that
the pressure drop through the shell side is zero.
Figure 13. Condensate backup in a shell-tube heat exchanger
Page 35 of 60
The inlet vapor is at its dew point. That means it is saturated vapor. Under these
circumstances, the outlet liquid should be saturated liquid, or liquid at its bubble
point. As the inlet dew-point temperature is 120°F, the outlet bubble-point
temperature should be 120°F. But, as can be seen in Fig. 13, the outlet shell-side
liquid temperature is 90°F, not 120°F.
The reason is condensate backup. The condensate backup causes subcooling;
that is, the liquid is cooled below its bubble point or saturated liquid
temperature. Perhaps a rat has lodged in the condensate outlet pipe. The rat
restricts condensate drainage from the shell side. To force its way past the dead
rat, the propane backs up in the condenser. The cold tubes in the bottom of the
shell are submerged in liquid propane. The liquid propane is cooled below its
bubble-point temperature.
Note that the propane vapor is still condensing to propane liquid at 120°F. The
condensed liquid is in intimate contact with the propane vapor as it drips off the
outside surface of the colder condenser tubes.
The saturated propane vapor condenses directly to saturated propane liquid at
120°F. The saturated, or bubble-point, liquid then drips from the condensation
zone of the condenser into the subcooling zone of the condenser. This is the
zone where the tubes are submerged in liquid.
Try running your hand along the outside of such a condenser. Feel for the point
on the surface of the shell where there is a noticeable drop in temperature. The
upper part of the shell will be hot. The lower part of the shell will be cold.
The transition point corresponds to the liquid level of condensate in the shell.
The condensate level will always be higher toward the shell outlet nozzle.
Again, this all applies only to condensers in total condensing service.
Effect of Condensate Backup
Page 36 of 60
When the condensate level in an exchanger increases, the area of the condenser
devoted to subcooling the condensate increases. But the area of the exchanger
available for condensing decreases. When the area of the exchanger available for
condensing is reduced, the ease of condensation is also decreased. Depending on
circumstances, one of two unfavorable things will now happen:
1. If the supply pressure of the condensing vapor is fixed, the rate of
condensation of the vapor will fall.
2. If the condensing vapor flow rate is fixed, the condensation pressure will
increase. Heat removed by condensation is easy.
The heat-transfer coefficient U for condensation of pure, clean, vapors may be
400 to 1000 Btu/hr/ft2of heat exchanger surface area, per °F of temperaturedriving force. The U value for subcooling stagnant liquid may be only 10 to 30.
Condensate backup is the major cause of lost heat transfer for heat exchangers in
condensing service.
Reflux Drum Elevation Increase Promotes Subcooling
A rat entered the condenser outlet pipe shown in Fig. 14. The condenser had
been off line for cleaning. The rat, having crawled up the riser pipe to the reflux
drum, got its head stuck in the drum’s inlet nozzle.
Your author, unaware of the rodent’s predicament, put the exchanger back into
service. The condensed butane now flowed across the rat. The rat died. Well, we
all must come to that end eventually, although perhaps not quite that exact end.
Such is the way of all flesh.
This rat is called a “20-lb rat.” Not that the rat weighed 20 lb. The 20 lb refers to
the pressure drop of 20 psig that the liquid encountered as it flowed across the
rat’s now-lifeless body. Before the introduction of this pressure restriction, the
butane entering the reflux drum was at its bubble point.
Page 37 of 60
The term “to flash” is used to denote partial vaporization of the butane. Before
the rat became stuck, the liquid entering the reflux drum did not flash or partly
vaporize.
We can also be sure that at steady state the butane liquid did not flash after the
introduction of the rat because no vapor was vented from the reflux drum.
Figure14. Elevation increase of reflux drum increases tower pressure.
But let’s assume that in the first microsecond after the introduction of the rat the
liquid did vaporize.
The vapor so generated would be trapped in the reflux drum. The pressure in the
drum would increase.
Not by 20 psig, but just a little. The small increase in pressure in the reflux drum
would push up the liquid level in the condenser. The surface area of the
Page 38 of 60
condenser available to subcool the liquid would increase. The liquid temperature
would be reduced. As the subcooled liquid flowed across the dead rat, its
pressure would drop. The liquid’s pressure would fall to exactly that pressure
that corresponds to the vapor pressure of the butane at the temperature in the
reflux drum.
The backup of butane liquid in the condenser would continue until the butane
leaving the condenser was cold enough so that it would not flash as it flowed
across the rat—that is, until equilibrium conditions had been reestablished in the
reflux drum.
As the butane liquid level in the condenser increased, the area of the exchanger
exposed to the condensing vapors would decrease.
Let’s assume that the tower’s reboiler duty was constant. The vapor flow rate to
the condenser would then be constant. To condense the same flow rate of vapor
with a shrinking exchanger surface area, the pressure of condensation must
increase. The tower pressure would also go up as the condenser pressure rose.
The real story is that the riser pipe connecting the condenser outlet to the reflux
drum was undersized. Nat Taylor, the project engineer, specified a 4-in pipe
when an 8-in pipe was needed. The pressure drop through the pipe was then 32
times higher than the intended ∆P of 0.60 psig (∆P in a pipe varies to the fifth
power of the pipe’s diameter). The resulting riser pipe pressure drop was 20
psig. This frictional loss of 20 psig in the pipe had the exact same effect on the
condenser—and on the tower’s pressure—as the 20-lb rat.
The elevation of the reflux drum was 80 ft above the condenser. The specific
gravity of the butane liquid was 0.59. This means that 80 ft of liquid exerted a
head pressure of about 20 psig:
Page 39 of 60
where s.g. : specific gravity.
This elevation head loss of 20 psig had the same effect on the condenser—and
on the tower’s pressure.
Common Design Error
Refer back to Fig. 14. How can the liquid from the condenser rise to the higher
elevation in the reflux drum without being pumped?
The pressure head of the liquid leaving the condenser is converted to elevation
as the liquid flows up into the reflux drum.
This works fine as long as the liquid leaving the condenser is sufficiently
subcooled.
The liquid leaving the condenser is subcooled. The liquid entering the reflux
drum is saturated liquid at its bubble point. Of course, the temperature of the
liquid is the same at both points. The subcooled liquid is “subcooled” in the
sense that its pressure is above the bubble-point pressure at the condenser outlet
temperature. It is this extra pressure, above the bubble-point pressure, that may
be converted to elevation.
c. Pressure control
 Tower Pressure Control
For total condensers, there are three general schemes for controlling distillation
tower pressure:
 Throttling the cooling water flow to the condenser
 Flooding the condenser
 Hot-vapor bypass around the condenser
Page 40 of 60
Figure 15. Tower pressure control using cooling-water throttling.
Regardless of the method selected, the principal concept of tower pressure
control is the same. We control the pressure in the reflux drum by manipulating
the temperature in the reflux drum. The tower pressure then floats on the reflux
drum pressure. To lower the tower pressure, we must first cool the reflux drum.
This reduces the vapor pressure of the liquid in the reflux drum.
The oldest, most direct method of pressure control is throttling on the coolingwater supply. This scheme is shown in Fig. 15. Closing the water valve to the
tube side of the condenser increases the condenser outlet temperature. This
makes the reflux drum hotter. The hotter liquid in the reflux drum creates a
higher vapor pressure. The higher pressure in the reflux drum increases the
pressure in the tower. The tower pressure is the pressure in the reflux drum plus
the pressure drop through the condenser.
Throttling on the cooling water works fine, as far as pressure control is
concerned. But if the water flow is restricted too much, the cooling-water outlet
temperature may exceed from 125 to 135°F. In this temperature range, waterhardness deposits plate out inside the tubes. Then the heat-transfer coefficient is
permanently reduced by the fouling deposits.
Page 41 of 60
 Hot-Vapor Bypass
Hot-Vapor Bypass Pressure Control
A more modern way of controlling a tower’s pressure is shown in Fig. 15. This
is the hot-vapor bypass method. When the control valve on the vapor bypass line
opens, hot vapors flow directly into the reflux drum. These vapors are now
bypassing the condenser. The hot vapors must condense in the reflux drum. This
is because there are no vapors vented from the reflux drum. So, at equilibrium,
the hot vapors must condense to a liquid on entering the reflux drum. They have
no other place to go.
The latent heat of condensation of this vapor is absorbed by the liquid entering
the reflux drum. The liquid that enters the reflux drum comes from the
condenser. The hot vapor mixes with the condenser outlet liquid and is
condensed by this cooler liquid.
The liquid in the reflux drum is in equilibrium with a vapor space. This liquid is
then at its bubble, or boiling, point. If the liquid draining from the condenser is
colder than this bubble point liquid, it must be subcooled. The tower overhead
vapor, shown in Fig. 16, condenses to a liquid on the outside of the cold
condenser tubes. The liquid drips off the tubes. These droplets of liquid are in
close contact with the saturated vapor in the condenser shell. This means that the
liquid is in equilibrium with the vapor. The condensed liquid is therefore
initially at its bubble-point temperature. This liquid accumulates in the bottom of
the condenser’s shell. The submerged tubes then must subcool this liquid. Part
of the surface area of the condenser is hence devoted to subcooling liquid, and
part is devoted to condensing vapor.
Page 42 of 60
Figure 16. Hot-vapor bypass pressure control. Note condensate backup
Let’s assume that the liquid draining from the condenser is not quite cold
enough to absorb the entire latent heat of condensation of the vapors flowing
through the hot-vapor bypass line. The vapors will then be only partially
condensed. Vapor will start to accumulate in the reflux drum. This accumulation
of vapor will increase the reflux drum pressure by a small amount. The higher
drum pressure will back up the liquid level in the condenser by a few inches.
The higher height of liquid in the drum will submerge additional cold tubes with
the condensed liquid.
The cooler liquid will now be able to absorb more of the latent heat of
condensation of the vapor passing through the hot-vapor bypass line. Eventually,
a new equilibrium will be established.
Page 43 of 60
Leaking Hot-Vapor Bypass Valve
Let’s assume that the hot-vapor bypass valve, shown in Fig. 16, is leaking. It is
leaking 10 percent of the tower overhead flow. A good rule of thumb is then:
Hydrocarbons. For each 20°F temperature difference between the cooler
condenser outlet and the warmer reflux pump suction, 10 percent of the tower’s
overhead vapor flow is leaking through the hot-vapor bypass valve.
Figure 17. Opening pressure-control valve cools condenser outlet
Aqueous systems. For each 20°F temperature difference between the condenser
outlet and the reflux pump suction, 1percent of the tower’s vapor flow is leaking
through the vapor bypass valve.
As the hot-vapor bypass valve opens, the condensate level in the shell side of the
condenser increases to produce cooler, subcooled liquid. This reduces the
surface area of the condenser exposed to the saturated vapor. To condense this
vapor with a smaller heat-transfer area, the pressure of condensation must
Page 44 of 60
increase. This, in turn, raises the tower pressure. This then is how opening the
hot-vapor bypass pressure-control valve increases the tower pressure.
Incidentally, as shown in Figs. 16 and 17, the condenser may be located above
or below the reflux drum. Both configurations require a subcooled liquid
effluent from the condenser. But if the condenser is located below the reflux
drum, additional subcooling to offset the elevation effect, described above, will
be needed.
 Flooded Condenser Pressure Control
In a flooded condenser tower pressure-control strategy, the reflux drum is run
full. Restricting the flow from the reflux pump increases the level in the
condenser. This reduces the heat-transfer surface area available for condensation
and raises the tower pressure. Either the reflux or overhead product may be used
to vary the liquid level in the condenser.
Figure 18. Flooded condenser pressure control: the preferred method.
Once a liquid level reappears in the reflux drum, the condenser capacity has
been exceeded. The level in the condenser will continue falling until the drum
empties and the reflux pump begins to cavitate.
Page 45 of 60
In general, flooded condenser pressure control is the preferred method to control
a tower’s pressure. This is so because it is simpler and cheaper than hot-vapor
bypass pressure control. Also, the potential problem of a leaking hot-vapor
bypass control valve cannot occur.
Many thousands of hot-vapor bypass designs have eventually been converted—
at no cost—to flooded condenser pressure control.
The function of the reflux drum in a flooded condenser design is to:
 Separate water from reflux when distilling hydrocarbons
 Give the operators time to respond if they have exceeded the condenser’s
capacity
 Provide a place from which non-condensable vapors may be vented
One problem with flooded condenser pressure control is related to the need to
occasionally vent non-condensables. This vent valve must not leak when closed.
Since the drum is normally full of liquid, a leaking vent valve will pass liquid.
Many pounds of product can be lost in this way. A butterfly control valve with a
soft, rubberized seat is a good choice for a remotely controlled, non-condensable
vent valve.
Air-cooled condensers seem to work as well as shell-and-tube water coolers in
flooded condenser pressure control service. Some air coolers are sloped toward
the outlet to reduce the ratio of the delta condenser surface area to delta height
change of condensate backup.
 Partial Condensation
If we normally have a situation in which non-condensable vapors appear in the
reflux drum, then there is only one pressure-control option available. This is to
place the tower pressure-control valve on the vapor off-gas as shown in Fig. 19.
Page 46 of 60
If we normally have non-condensable vapors in the condenser effluent, then the
following problems we have been discussing do not exist:
Figure 19. Pressure control for partial condensation.
 Condensate backup
 Subcooling of condenser effluent
 Fouling due to low flow of the cooling-water tubes
The natural gas dissolves in the overhead liquid product and typically flashes out
of the product storage tanks. The correct way to control tower pressure in the
absence of noncondensable vapors is to employ flooded condenser pressure
control. If, for some external reason, a variable level in the reflux drum is
required, then the correct design for tower pressure control is a hot-vapor
bypass.
 Slug Flow in Risers
Figure 20 illustrates a common cause of an erratic tower operating pressure.
Page 47 of 60
The condenser effluent is a two-phase mixture of vapor and liquid. In the
summer, the tower pressure is steady.
But in the winter when the cooling water is cold, pressure control is erratic. The
problem is phase separation in the riser line. The vertical velocity in this line is
too low to maintain a froth flow up the riser. Slugs of liquid form. Periodically
the accumulating liquid is blown out by the vapor. This slug flow causes an
erratic back pressure on the tower but not an erratic pressure in the drum.
Figure 20. Slug flow in risers.
This problem only occurs when the drum is elevated above the partial
condenser. The fundamental difficulty is excessive riser pipe diameter. To avoid
slug flow in such a riser a mixed-phase velocity of at least 20 ft/s is needed.
While the 20 ft/sec may be maintained in the summer, more efficient
condensation in the winter may reduce vapor flow. This can cause the riser
velocity to drop below the minimum to prevent phase separation. Throttling the
cooling water will stabilize the tower pressure, but may result in salting up the
exchanger with water hardness deposits.
Page 48 of 60
8. NSRP’s refinery CDU
a. General information:
The design of the CDU is consistent with the overall refinery objective of
maximizing distillates and minimizing residue. The design incorporates
flexibility in distillate production, will allow for minor variations in feed quality,
and will enable consequent flexibility in refinery blending operations.
There are two design cases and the kerosene production is designed to
accommodate the specifications of both aviation and non-aviation kerosene from
the same draw. However, Kerosene product flash point and ASTM-D86 IBP
specifications shall be achieved at KHDS Unit.
Base Case
CDU is designed for TBP cut points and products specifications. The overflash
has been specified at 5% and performance guarantees are based on this case
only.
Minimum Kerosene Case
CDU shall be able to increase wild naphtha production at the expense of
kerosene cut in order to maximize profits in the Aromatics Complex (overflash
has been reduced from 5.0 to 2.3 %).
b. Feed and products
Design case CDU feedstock is 100% Kuwait Export crude. A maximum of 0.5%
by volume water is assumed to be present in the crude. The Crude Distillation
Unit (CDU) is designed to process an average of 9.66 million tons per annum
(MTA) or 200,000 Barrels Per Stream Day (BPSD - dry basis) of crude oil feed.
Page 49 of 60
The design incorporates features for optimizing energy utilization and heat
recovery consistent with product fractionation. The unit produce the following
streams:
 An overheads gas stream routed to the LPG Recovery Unit.
 A full range unstabilized naphtha stream routed to the LPG Recovery Unit
for further processing.
 A kerosene stream routed to the Kerosene Hydrodesulphurisation Unit
(KHDS) or to Slops in case of KHDS shutdown.
 A gas oil stream routed to the Gas Oil Hydrodesulphurisation Unit
(GOHDS) or to GOHDS intermediate tankage.
 An
atmospheric
residue
stream
routed
hot
to
the
Residue
Hydrodesulphurisation Unit (RHDS) for further processing or to
intermediate tankage.
c. Crude desalting
The water-soluble impurities in the crude oil feed shall be removed by electrical
desalting. Since the atmospheric residue, resulting from the atmospheric
distillation of the crude, shall be further processed in a downstream Residue
Hydrodesulphurisation Unit (RHDS), two stage desalting shall be provided to
achieve the required sodium content in the desalted crude. Two-stage desalting
is provided to reduce the salt level of the crude, and hence the atmospheric
residue, to the minimum practicable level consistent with RHDS operation.
Water is recycled within the desalter system to reduce overall water
requirement. The desalter shall be supplied with stripped water from the sour
water stripper (SWS).
Page 50 of 60
Page 51 of 60
d. Unit description:
Crude oil feedstock is preheated against product and pumparound streams before
being routed to a fired heater. Primary fractionation is carried out in the preflash
vessel/ main crude column fractionator and associated side stream strippers.
Overhead naphtha is further stabilized in the naphtha stabilizer column in the
LPG Recovery Unit. Products are cooled and rundown to product blending,
intermediate storage or further processing as appropriate.
Cold Crude Preheat Section
Cold crude oil feedstock from the offsite storage tanks is pumped by Offsite
Crude Charge Pumps to the CDU which is heated against the hot products and
pumparounds before entering the two-stage desalting section. De-emulsifier is
added at the inlet to the cold crude preheat train to aid the separation of any oilwater emulsions.
In cold crude preheat train, the crude is heated in the first exchanger by Top
Pumparound, and then split into three parallel streams under flow control and
the fallback on differential pressure control. The first split (30% of total crude
flow rate) is heated first by Kero Product in the heat exchanger and then by Kero
Pumparound. The second split (45% of total crude flow rate) is heated by Atm.
Residue. The fallback which is under differential pressure control is heated by
AGO Product. All the three splits are then recombined and sent to the Desalting
section. The temperature of the preheated crude is around 150°C before entering
the desalter section.
Crude Desalting Section
Desalter water from the second stage desalter (pumped by Desalter Water
Circulation Pumps) is injected in-line to the heated crude and is then intimately
mixed across a mixing valve which provides a fixed pressure drop and enters the
1st stage desalter of the Desalting Section. Demulsifier is added into the
dewatered crude (in order to aid separation of any oil/water emulsions that can
not be separated by electrostatic action alone; actual operating experience will
determine the basis on which the demulsifier will have to be injected, i. e.
whether intermittent or continuous) from first stage desalter and mixed
intimately with the fresh water (from Desalter Water Surge Drum) across the
second stage mixing valve before entering the 2nd stage desalter. Effluent from
1st stage desalter is first cooled in exchanger by desalter water feed. The effluent
is then air cooled and then passed through heat exchangers under level control
where it is cooled using cooling water. The cooled effluent then flows to the
effluent treatment plant (ETP).
Stripped sour water from the Sour Water Stripping Unit (SWS) and oxygen
scavenged service water (after dosage of bisulphite) under level control provide
the wash water to the 2nd stage desalter. This water is first collected in the
Desalter Water Surge Drum, before being pumped into the second stage of the
Desalter Package, as well as, into the crude stream upfront the Cold Crude/Top
PA Exchanger, by the Desalter Water Makeup Pumps, under ratio controlled
flow. Before entering the second stage desalter, the wash water is preheated in
the Desalter Water Feed/Effluent Exchanger. Fresh water connection is provided
for desludging of both the desalters. This water will also be used for washing
tube bundles of Desalter Effluent Air Cooler. The Desalter Water Circulation
Pumps take water from the Second Stage Desalter, and deliver it to the First
Stage Desalter.
Hot Crude Preheat & Preflash Section
Desalted and dewatered crude from 2nd stage desalter under pressure control
enters the desalted crude preheat train after injection of antifoulant (to reduce
fouling in the preheat train) and caustic (to protect the column overhead system
against corrosion by the carryover of dissolved acid chloride salts). It is first
split into three parallel streams for maximum heat recovery. A portion of the
crude, under flow control (45% of total desalted crude flow rate), is heated first
Page 53 of 60
by Kero Pumparound and then by LAGO Pumparound. Second split of the
desalted crude under differential pressure control, is heated by Atm Residue.
The remaining portion (25% of total desalted crude flow rate) under flow control
is heated by AGO Product in exchanger.
The flashed crude streams at the outlet of control valves at each exchanger trains
are recombined and sent to the Pre Flash Drum. Provision of antifoaming
chemical injection is provided upstream of the preflash drum and should be used
only on occurrence of foaming in the preflash drum. Temperature of the desalted
crude is around 200°C at the outlet of the desalted crude exchanger train. In the
Pre Flash Drum, light hydrocarbons flash off and are sent as flashed vapours to
the flash zone of the Atmospheric Crude Distillation Column. The flashed crude
leaving the pre-flash drum is pumped by Flashed Crude Pumps and is split into
two streams. For maximum heat recovery, one of the flashed crude streams,
under flow control, is heated by LAGO PA & by HAGO PA. The remaining
portion of the flashed crude which is heated under differential pressure control
by Atm. Residue heated flashed crude is then combined with the earlier stream.
This combined stream is finally heated by exchanging heat with Atmospheric
Residue.
Crude Charge Heating Section
Preheated flashed crude enters the Crude Charge Heater at around 302°C, where
it is partially vaporized before entering the flash zone of the Crude Distillation
Column. Crude Charge Heater, is an 8-pass, balanced draft, dual fuel (fuel oil &
fuel gas) fired heater. Apart from refinery fuel gas, spent air from the LPG
Treater and RFCC LPG Treater is fired into the heater. To ensure satisfactory
thermal efficiency from the heater, the flue gases from this heater are used to
preheat the combustion air required in the Flue Gas Air Preheater. The total
combined feed is automatically controlled into each one of the parallel passes of
the heater by the pass balancing control, under level control from the Pre Flash
Drum. The heater outlet temperature is controlled through the Burners Control
Page 54 of 60
System, which sends a set point signal to either fuel oil or fuel gas control valve,
depending on the type of fuel being fired. Also, the Burners Control System
receives a signal from the O2 analyzer to ensure that the desired heater efficiency
is achieved by the control of excess air. Emissions such as NOX, SPM
(suspended particulate matter), CO & O2 are monitored after the heater
convection section.
Crude Distillation Column and Side Strippers Section
Partially vaporized crude from each pass of the crude charge heater at around
347-359°C is combined in the common transfer line before entering the flash
zone of the crude distillation column. In the Crude Distillation Column, the
crude is fractionated into four products: Kerosene, Light Atmospheric Gas Oil
(LAGO), Heavy Atmospheric Gas Oil (HAGO) as side stream products and an
atmospheric residue bottoms stream. Reflux between the side streams is
provided by four pumparound circuits: Top Pumparound over trays 1 to 4, Kero
pumparound over trays 14 to 16, LAGO pumparound over trays 23 to 25, and
HAGO pumparound over trays 26 to 28.
The top pumparound is circulated by the Top PA pump, and is heat integrated
with the crude in heat exchanger. The pumparound is further cooled in the Top
PA Return Air Cooler before being refluxed in the column. Crude column top
temperature is controlled by duty control taking input from the crude column top
temperature control, the temperature differential and flow control of the top
pumparound circuit. Corrosion inhibitor is injected in the PA pump suction to
reduce corrosion in the top PA circuit.
The kerosene pumparound is circulated by the Kero PA Pump, and is heat
integrated with the crude in heat exchangers. The temperature of this section of
the crude column is controlled by a duty control between the temperature
differential and flow control of the kerosene pumparound circuit.
Page 55 of 60
The LAGO pumparound is circulated by the LAGO PA Pump, and is heat
integrated with the crude in heat exchangers. The temperature of this section of
the crude column is controlled by a duty control between the temperature
differential and flow control of the LAGO pumparound circuit.
The HAGO pumparound is circulated by HAGO PA Pump, and is heat
integrated with the crude by heat exchangers. After heat exchange, HAGO PA is
returned to the column. Provision is kept for the HAGO product draw-off to be
taken to the AGO stripper.
Continuous pumpback to the column is provided for each of the above
pumparound pumps. Pump back flow-rate is controlled by chimney tray level.
There is also a provision of bypassing the heat exchanger trains in each of the
pumparound circuits except the top pumparound.
Crude Column Overhead Section
The naphtha overhead product, together with stripping steam, is condensed in
the Crude Column Overhead Condenser, and run into the Crude Column
Overhead Receiver, where vapor, oil and water separate. Neutralizer (to
neutralize acid gases) and Corrosion Inhibitor (to control corrosion) are injected
into the Crude Column overhead line. There is also a provision of fuel gas in the
overhead receiver for maintaining column pressure during start-up.
The vapor leaves the receiver under a split range pressure control to the Offgas
Compressor Suction Drum in LPG recovery unit. The oil phase from the
receiver is pumped by the CDU Overhead Liquids Pumps, under level-flow
cascade control, to the LPG recovery unit for further processing. A part of this
overhead liquid can be used as cold reflux and added to the Top PA Return Air
Cooler outlet during startups and for additional temperature control flexibility in
the crude distillation column overheads. The water phase from the overhead
receiver is pumped under level control by the Sour Water Pump, to the Sour
Water Stripping Unit (SWS). Deentrained liquid from LP compressor suction
Page 56 of 60
drum and 1st stage spillback from LP compressor of the LPGRU is added to the
overhead line from the crude distillation column at the upstream of the overhead
condenser.
Product Cooling & Rundown Section
Kerosene product is drawn from the main column tray under stripper level
control to the downstream Kerosene Stripper. The stripper contains 6 valve trays
over which the product is steam stripped using LP Steam. The Kerosene Product
Pump, takes the product from the bottom of the stripper and pumps it to the Cold
Crude/Kero Product Exchanger. The Kerosene Product is then sent to the KHDS
section. When KHDS is down or not able to draw the entire kerosene produced,
the kerosene product will be routed to the slops under backpressure control after
cooling it in Kerosene product air cooler and Kerosene product trim cooler.
AGO product is drawn from the main column tray under stripper level control to
the downstream AGO Stripper. The stripper contains 6 valve trays over which
the product is steam stripped using LP Steam. The AGO Product Pump, takes
the product from the bottom of the stripper, and then transfers the AGO Product
to the exchangers for heat recovery. It is then sent to the GOHDS section. When
GOHDS is down or not able to draw the entire AGO produced, it will be routed
to GOHDS feed storage after further cooling in AGO Product Air Cooler.
The bottom product is steam stripped at the bottom of the Distillation Column
with LP Steam and is pumped by the Atmospheric Residue Pumps (main
bottoms pump is steam turbine driven, whereas motor driven pump will be
normally standby) for heat recovery to the exchangers. It is then sent to the
RHDS section. Alternately, when RHDS is down, it will be sent to the Atm
residue storage section after cooling down in the Atm residue/Tempered Water
Trim Cooler.
Page 57 of 60
e. Some notes for unit operation
Crude Charge Heater Outlet Temperature
The crude oil temperature at the heater outlet determines the proportion of the
feed vaporized and the total heat input to the crude distillation column, therefore
it will affect the attainable cut point between distillate and residue products. For
normal operation the COT (Coil Outlet Temperature) temperature should be
approximately 360°C.
In general, a lower heater temperature results in a lower crude distillate yield.
Also, the attainable separation between the products may deteriorate at lower
heater temperatures, due to reduced over-flash or reduced reflux flows.
Excessively high heater COT temperatures should be avoided, since this initiates
cracking reactions in the feed to form gas, which in turn causes overloading of
the overheads system with non-condensables also Cracking reactions will also
result in premature coking of the furnace tubes.
Crude Distillation Column Pressure
The CDU Column overhead pressure should be maintained as steady as possible
for normal operation. A higher pressure may fail in obtaining the required
vaporization of the feed at maximum furnace outlet temperature, and design
distillate cut point will not be attained. On the other hand, targeting too low a
pressure will result in heavy hydrocarbons going overhead and passing to the
LPG Recovery Unit. This could result in failure to meet naphtha product
specifications. However, minor adjustments may be effective when lower
pressure is necessary to achieve the deeper cut operations if required.
Crude Distillation Column Overhead Temperature
The CDU overhead vapor temperature determines the amount and boiling range
of the hydrocarbons taken overhead. For normal operation the CDU column
Page 58 of 60
overhead temperature will be in the range 109°C to 123°C, base case of 113°C.
A higher temperature will result in a greater amount of heavy hydrocarbons
leaving in the overhead vapor that flows to the LPG Recovery Unit, where
heavy hydrocarbons exceeding the C4 specification will be recovered. To
minimize the C4 in the overheads the overhead vapor temperature may be
adjusted to minimize the quantity of C4 in the overheads. Lower overhead
temperatures normally result in reduced overhead material and can also result in
water condensation in the overheads system causing corrosion.
Distillation Columns’ Pumparound, Reflux, and Wash Oil Circulations
Pumparound, reflux, and recycle flows are critical to maintaining optimum
vapor/liquid rates in the fractionation and wash sections of the columns. The
design will include a top pumparound. This is intended to offload duty from the
overheads condenser. Other pumaround circuits will be provided to maximize
heat recovery whilst maintaining the required fractionation.
Inadequate pumparound heat removal or poor pumparound distribution will
result in high column overhead pressure, poor fractionation and a reduction in
flash zone vacuum. The strainers on the pumparound returns should be regularly
monitored and cleaned to ensure that solids build-up does not impair flow rate or
distribution efficiency.
Excessive pumparound heat removal will result in difficulties in maintaining
some product end points, and may cause internal flooding, entrainment and high
column pressure drop.
Page 59 of 60