Bagajewicz2001 PDF
Bagajewicz2001 PDF
Bagajewicz2001 PDF
The problem of designing crude fractionation units is not only a distillation design. It has the
added complexity that these units should be able to process different types of crude, sometimes
from heavy to light. Important heat exchange also takes place, and the energy efficiency is related
to the column design parameters. Part I of this two-part series presents a rigorous targeting
methodology to design this multipurpose plant, which can be implemented using a commercial
simulator. Part II deals with the heat exchanger network design.
energy-integrated distillation schemes,6 they are not explored, these columns are merged into a single
directly applicable to crude fractionation for the follow- complex column.
ing reasons: (1) The number of components in the crude The major advantage of the procedure presented by
is too large to handle. Usually around 30-40 pseudo Liebmann et al.9 is that it couples the column perfor-
components are used, while available studies on se- mance with heat recovery goals. However, the procedure
quencing seldom addressed systems containing over 5 is not able to assess the trade-off between steam
components. (2) Previous separation sequencing studies injection and distribution of heat removal in pump-
assumed that the products were pure; however, products around circuits. Finally, addressing the column as a
in crude distillation are mixtures. whole, which is the alternative we present in this paper,
Packie7 pioneered the field of crude fractionation is more convenient and straightforward. It does not rely
design. In his method, 5-95 gaps and 50% distillation on any special rules of thumb for reflux ratios, and it
temperature differences are used as separation criteria. helps determine better the relationships with column
The term gap refers to the difference between the 5% variables and heat integration.
ASTM D86 distillation temperature of a heavier product Sharma et al.10 proposed a method for calculating the
and the 95% ASTM D86 distillation temperature of an maximum pump-around heat removal. First, a practical
adjacent lighter product. When the distillation curves minimum reflux ratio for each column section is deter-
of the two products overlap, a negative gap appears. mined using Packie’s empirical diagram. Then, the heat
Packie used empirical charts to express the relation removal in the upper part of the column is calculated
among the 5-95 gap, the reflux ratio, and the number using a heat balance. The upper part may start from
of trays in the section under consideration. However, an arbitrary tray and end with the condenser. Next, the
the empirical nature of these charts results in inac- upper part is extended tray by tray, and heat surplus
curacy and prevents optimal designs. Furthermore, in is calculated for each tray. The resultant heat surplus
his design procedure, Packie considered the column data are used to construct a column grand composite
design and the heat integration separately. The heat curve. Finally, the maximum heat removal for each
integration does not start until the column design is section is determined using the column grand composite
finished. curve. A major advantage of this method is that the
Watkins2 pointed out that “optimizing the crude maximum heat removal can be estimated quickly with-
preheat-tower cooling heat-exchange train is the heart out the need of simulation. However, as Packie’s dia-
of crude unit design, and each case must be studied on gram is empirical and the effect of the stripping steam
an individual basis in order to arrive at the most is not included, the heat removal calculated is not
economical processing scheme.” However, Watkins did accurate. In contrast, the procedure presented in this
not present a specific methodology to perform this paper is based on rigorous simulations and can capture
design. the relationships between the column variables and the
Liebmann et al.8,9 recently proposed an integrated heat integration opportunities.
design procedure. The design procedure starts with a In this paper we present a new procedure for crude
sequence of simple columns that are generated by distillation design. The major features of this procedure
decomposing the crude main tower. The total number are the following: (1) The design objective is to process
of trays is assumed to be the same as that in Watkins’ several crudes at optimal conditions. (2) The design
design, and the number of trays for each column is calculations are rigorous. (3) Heat demand-supply
calculated with the assumption that the R/Rmin values diagrams are used as a tool guiding the design. The
are approximately the same for all columns. There is major advantage of the heat demand-supply diagram
no thermal coupling between these initial columns. is that the contribution of each process stream or pump-
Next, reboilers and thermal coupling are introduced in around to the total utility consumption is shown explic-
order to reduce utility consumption. The grand compos- itly. This feature helps determine necessary changes
ite curve is used to assess the proposed modifications. leading to lower energy consumption. (4) The interaction
After all of the possible design modifications have been between steam stripping and pump-around duties is
Ind. Eng. Chem. Res., Vol. 40, No. 2, 2001 619
∑ Fsi
steam) are chosen using the guidelines offered by
VW W
j ) VFZ + (3) Watkins with one exception: No pump-around circuits
i∈III
are included at this point.
Replacing VO W O Step 2. The simulation is performed next. Usually,
FZ, Vj , and Vj in eq 1, one obtains
the column is not difficult to converge, as the liquid
∑ Qk ) LO(hL
k∈III
O
- hO
FZ) + ∑
i∈III
Fpi(hpi - hO
FZ) +
reflux ratio is large.
Step 3. The heat demand-supply diagram is con-
structed.
∑
i∈II
Fpi(hVO - hO
j
W W W
FZ) + VFZ(hV - hFZ) +j Step 4. The maximum amount of heat is transferred
to a pump-around circuit located in the region between
∑
i∈III
Fsi(hVW - hsi) + Lj-1(hVO - hL
j j
) (4)
j-1
the top tray and the first product withdrawal tray. The
location of the pump-around circuit withdrawal and the
return temperature are conveniently chosen so that the
There are six terms on the right-hand side of eq 4. energy recovery is maximized. This step is discussed
The first through the fifth terms represent the conden- further when presenting the example.
sation heat of the overflash stream LO, the condensation Step 5. If the product gap becomes smaller than
heat of the products leaving envelope III, the apparent required, the stripping steam flow rate is increased to
heat released by the hydrocarbon vapor VO j , and the
fix the gap. As long as the steam added has a lower cost
apparent heat released by the steam streams. The last than the energy saved, one can continue shifting loads.
term stands for the vaporization heat of the internal Otherwise, it is advisable to stop when a trade-off has
reflux Lj-1. Apparently, when Lj-1 goes to zero, the heat been reached.
removal from envelope III reaches its maximum. By Step 6. If there is heat surplus from the pump-around
including more pump-arounds in envelope III and circuit just added, transfer the heat to the next pump-
Ind. Eng. Chem. Res., Vol. 40, No. 2, 2001 621
Table 1. Feedstock Used for the Design Table 5. Tray Requirements in Watkins Design
density throughput number
crude (kg/m3) (m3/h) separation of trays
light crude 845 (36.0 API) 795 light naphtha to heavy naphtha 6-8
intermediate crude 889 (27.7 API) 795 heavy naphtha to light distillate 6-8
heavy crude 934 (20.0 API) 795 light distillate to heavy distillate 4-6
heavy distillate to gas oil 4-6
Table 2. TBP Data flash zone to first draw tray 3-4
temperature (°C) steam stripping sections 4
maximum. Therefore, the first pump-around has to be Table 7. Effect of Increasing PA2 Duty without
Changing Steam Flow Rates
above the kerosene withdrawal tray. The question is the
number of trays that one should put between the 1 2
condenser and the top pump-around region. We recom- duty of PA2 29.31 MW 33.71 MW
mend that the top pump-around region be adjacent to duty of PA1 32.83 MW 28.43 MW
the condenser. No tray is put in between. This recom- duty of condenser 41.94 MW 42.03 MW
mendation is based on the observation that the trays (5-95) kerosene-naphtha gap 18.49 °C 16.60 °C
below a product withdraw line and above an adjacent (5-95) diesel-kerosene gap 1.63 °C 1.48 °C
(5-95) gas oil-diesel gap 1.22 °C 1.23 °C
pump-around circuit receive little reflux and barely energy consumption 70.59 MW 67.35 MW
contribute to separation. The pump-around stream is
withdrawn from tray 4, cooled in the heat exchangers, of PA1, a second pump-around is installed at the
and returned to tray 2. The return temperature is 104.4 position indicated in Figure 5.
°C, which is optimized after the duty is determined. Two Pump-Around Circuits. The second pump-
The duty of the top pump-around (PA1) is increased around (PA2) is positioned between tray 10 and tray
gradually, and product gaps are examined in each 12, just below the kerosene withdrawal tray. The return
simulation. The kerosene-naphtha gap decreases with temperature is chosen to be approximately equal to the
the increase in the PA1 duty but remains well above withdrawal temperature of PA1. A lower temperature
the specification, while the other gaps are almost would result in a heat surplus in the region of PA1,
unchanged. The heat shift continues without violating whereas a very high return temperature would not alter
the gap specifications until the reflux ratio is around the energy savings but would result in heavier liquid
0.1. Further heat shift would result in liquid drying up traffic in the PA2 region. With the increase in the PA2
on the top tray. Thus, the limit of the heat shifting has duty, the gap between kerosene and naphtha decreases
been reached. The duty of 62 MW represents the total quickly. Table 7 shows the change in the gaps as a
amount of heat one could obtain from all pump-around function of the duty of pump-around PA2.
circuits. The following steps consist of distributing this When the duty of PA2 is larger than 33.7 MW, the
amount of heat properly among several pump-around kerosene-naphtha gap does not satisfy the specifica-
circuits. The main operation variables of the scheme tion. To recover this gap, one could increase the strip-
with one pump-around are shown in Table 6. ping steam flow rate or increase the number of trays in
The major conclusions are as follows: (1) The total the naphtha-kerosene section. The former option is
energy consumption (E) decreases by 7 MW compared used in this work, as the latter might not be sufficient
to the no pump-around scheme. (2) The kerosene- or even practical on its own. The kerosene and diesel
naphtha gap is reduced from 25 to 23 °C, remaining stripping steam flow rates are adjusted with a controller
well above the specification of 16.7 °C. (3) The yield in which the gap specifications are defined.
of naphtha decreases, and the yield of kerosene in- With the help of the stripping steam, it is possible to
creases. This is because some light components of the move more heat from PA1 to PA2. The trade-off between
vapor are absorbed by the cold pump-around stream and increasing the energy recovery and spending more
carried to the kerosene withdrawal tray. Note that the steam is evaluated using eq 5. Heat shifting continues
total yield of the two products remains constant. (4) until the liquid reflux at the kerosene withdrawal tray
Little change takes place below the kerosene withdrawal is small and/or the kerosene-naphtha gap cannot be
tray. recovered even with increased amounts of stripping
We now turn our attention to the resulting heat steam. This is a limit imposed by the separation
demand-supply diagram (Figure 5). The shaded area requirement. The limiting case is shown in Table 9 (first
is the energy savings achieved by adding PA1. The heat column) and should be compared with the second
surplus in the condenser region is greatly reduced, but column of Table 6.
it is still significant. However, it is impossible to shift The major changes from one pump-around to two
more heat from the condenser to PA1. pump-around circuits are: (1) The net energy consump-
The return temperature of PA1 is not important in tion decreases sharply by 32 MW. (2) The flow rate of
terms of energy consumption, because the heat surplus the kerosene stripping steam is nearly doubled. The
is larger than the demand below the PA1 withdrawal large amount of extra steam is used to strip a significant
temperature. To reduce the heat surplus in the region amount of light components in the kerosene withdrawal
Ind. Eng. Chem. Res., Vol. 40, No. 2, 2001 623