RETROFIT OF CRUDE PREHEAT TRAIN WITH MULTIPLE
TYPES OF CRUDE
Kitipat Siemanond
1,2
and Supachai Kosol
1,2
1
The Petroleum and Petrochemical College, Chulalongkorn University, Bangkok 10330, Thailand
2
National Center of Excellence for Petroleum, Petrochemicals, and Advanced Material
Chulalongkorn University, Bangkok 10330, Thailand
Keywords: Heat Exchanger Network (HEN), Pinch analysis, Mixed integer linear programming, Stage model, Retrofit.
Abstract: This study explores the retrofitting of the crude preheat train of a crude distillation unit (CDU) processing
two types of crude--light and heavy--for a period of 200 and 150 days per year, respectively, with the aim of
finding the optimal design that would yield the highest net present value (NPV). A mathematical
programming model using GAMS software of heat exchanger network (HEN) called stage model (Zamora
and Grossmann, 1996) is applied to carry out the retrofit. The base case CDU is simulated by PRO II
software. Using pinch analysis, the composite curves show the retrofit potential of base cases with light and
heavy crude. The 10-stage model generates six retrofit designs--Designs 1, 2, 3, 4, 5, and 6--of which
Designs 1, 2, and 3 are suitable for light crude and Designs 4, 5, and 6 are suitable for heavy crude. Using a
graphical technique of searching for optimization with maximized NPVs of all designs, it is shown that
Design 2 is the optimal retrofit design processing both types of crude, yielding the highest NPV of
$11,529,511 for a 5-year lifetime and resulting in furnace duty saving of 32%.
1 INTRODUCTION
The crude distillation unit (CDU), as shown in
Figure 1, is one of the largest energy-consuming
units in a refinery. It has a complex heat exchanger
network (HEN) of crude preheat train which
transfers heat from hot-product and pump-around
streams to preheat crude before it enters the CDU,
resulting in energy saving in crude furnace and
coolers of CDU. For this study, PRO II software is
used to simulate base case CDU operated under
Arabian light (light crude) and Bacha quero (heavy
crude) with different distillation curves (Figure 2).
The volumes of crude products from CDU of light
and heavy crude are found in Table 1. CDU of light
and heavy crude of 5000 barrels/hr consumes
different steam and condenser duties (Table 2).
This work focuses on retrofitting the base case crude
preheat train of light and heavy crude by using a
graphically searching technique with n-stage model.
2 LITERATURE SURVEY
In the 1970s, pinch technology, or process heat inte-
Figure 1: Crude distillation unit.
gration, which aids the design of an efficient HEN
by the use of composite curves (T-Q diagram), as
shown in Figure 3, was developed. This technology
has enabled a theoretical approach to design an
optimal HEN and find retrofit potential of the
process. The composite curves consist of hot and
303
Siemanond K. and Kosol S..
RETROFIT OF CRUDE PREHEAT TRAIN WITH MULTIPLE TYPES OF CRUDE.
DOI: 10.5220/0003577203030308
In Proceedings of 1st International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH-2011), pages
303-308
ISBN: 978-989-8425-78-2
Copyright
c
2011 SCITEPRESS (Science and Technology Publications, Lda.)
Figure 2: Distillation curves of crude.
Table 1: Products from light and heavy crude.
Table 2: Steam and condenser duties of CDU.
cold composite lines presenting the relationships
between temperature (T) and heat content (Q) for
heat sources and sinks in the system. A pinch point
of two lines indicates a heat recovery approach
temperature (HRAT) or a thermodynamic constraint
on heat exchange. Shifting the cold composite curve
to the left improves heat recovery, or energy saving,
by increasing the heat-exchanger area.
The retrofit technique by Tjoe and Linnhoff
(1986) using pinch technology or thermodynamic
method applies targeting procedures to energy-area
tradeoffs which subsequently translate into
investment savings plots. Yee and Grossmann
(1990) proposed assignment-transshipment models
for structural modifications and a two-stage
approach. Ciric and Floudas (1988) proposed a
retrofit strategy using a decomposition method.
Briones and Kokossis (1998) used the hypertargets
or conceptual programming approach for retrofitting
industrial heat exchanger networks.
Figure 3: Composite curves.
3 N-STAGE MODEL
The stage model developed by GAMS software is
based on the stage-wise superstructure representa-
tion proposed by Zamora and Grossmann (1996), as
shown in Figure 4. Within each stage of
superstructure, possible exchanger between any pair
of hot and cold streams can occur. Heater and
coolers are placed at the end of cold and hot streams,
respectively. The objective function of the model is
to minimize the duties of heater, cooler and number
of exchangers under the constraint functions of
energy balance, thermodynamics, logical, and
retrofit constraints. The target temperatures and flow
rates of hot and cold streams are fixed and the stage
model will design HEN into n stages with the
minimum utility usages and number of exchangers
for fixed EMAT (Exchanger Minimum Approach
Temperature).
Figure 4: n-stage model structure.
SIMULTECH 2011 - 1st International Conference on Simulation and Modeling Methodologies, Technologies and
Applications
304
Generally, the number of stages in the
superstructure is set equal to the maximum
cardinality of the hot and cold sets of streams,
although sometimes it is necessary to increase the
number of stages to allow designs with minimum
energy consumption. The purpose of the retrofit
model is to minimize the number of exchangers
under constraint functions of energy balance,
thermodynamics, logical constraint and retrofit
constraint. The retrofit constraint is shown in
equation (1):
∑∑∑
===
111
1
ijk
Zijk
(1)
where Zijk is a binary variable of existing exchanger
matches between hot (i) and cold (j) streams at stage
k. This constraint helps retrofit HEN by keeping
base case exchangers in the same location in the
retrofit design.
4 RESULTS AND DISCUSSION
4.1 Base Case
This case study focuses on retrofitting a base case
crude preheat train of light and heavy crude for 200
and 150 days per year, respectively. The base case
consists of eight hot product streams (PA1, PA2,
PA3, Naphtha, Kerosene, Light Gasoil, Heavy
Gasoil, and Residue), three cold crude streams (J1,
J2, and J3), as seen in Figure 1, and eight process
exchangers (E1, E2, E3, E4, E5, E6, E7, and E8)
with an area of 5621.97 m
2
, as shown in Figures 5
and 7. The base case crude preheat train is operated
under light and heavy crude for 350 working days
per year.
Figure 5: Base case HEN with light crude feed.
4.1.1 Base Case of Light Crude
The structure of base case crude preheat train
operated under light crude for 200 days per year for
the lifetime of 5 years is shown in Figure 5. It
consumes furnace and cooler duties of 95.162 and
44 MMW, respectively, at HRAT = 79.3
o
C.
The composite curves of this base case, as shown
in Figure 6, show a retrofit potential, meaning
retrofit of this base case to reduce furnace and cooler
duties is possible.
Figure 6: Composite curves of base case of light crude
HEN.
4.1.2 Base Case of Heavy Crude
The structure of base case crude preheat train
operated under heavy crude for 150 days per year for
a lifetime of 5 years is shown in Figure 7. It
consumes furnace and cooler duties of 91.4 and
96.37 MMW, respectively, at HRAT = 113.6
o
C.
Figure 7: Base case HEN with heavy crude feed.
ETROFIT OF CRUDE PREHEAT TRAIN WITH MULTIPLE TYPES OF CRUDE
305
The composite curves of this base case, as shown in
Figure 8, also show a retrofit potential.
Figure 8: Composite curves of base case of heavy crude
HEN.
The retrofit with 10-stage model is applied to the
base-case HEN and gives six retrofit designs:
Designs 1, 2, 3, 4, 5, and 6. Designs 1, 2, and 3 are
suitable for light crude for 200 days per year while
Designs 4, 5, and 6 are suitable for heavy crude for
150 days per year.
4.2 The Optimal Retrofit Case
The base case of light crude is retrofitted by 10-
stage model using GAMS, generating six retrofit
designs at different HRATs, selectively, with
different furnace duty (hot utilities) and cooler duty
(cold utilities), as shown in Tables 3, 4, and 5.
Table 3: Six retrofit designs with exchanger area.
Table 4: Six retrofit designs for light crude.
Table 5: Six retrofit designs for heavy crude.
The net present value (NPV) is based on future cash
flows for a certain number of years, n, and a specific
annual interest rate. The NPV is calculated as
follows:
(2)
Table 6 shows the NPV for each retrofit design.
Table 6: NPV of six retrofit designs.
The economic data including utility and
investment costs for this retrofit case are as follows.
The lifetime of this retrofit project is 5 years and
the annual interest rate is 10% (350 working days
per year). The cost of hot and cold utilities are
0.4431 and 0.0222 cents per megajoule,
respectively. The maximum exchanger area added to
and removed from existing exchanger shells are 10%
and 40%, respectively. The maximum limit of area
per shell is 5,000 m
2
and one exchanger can contain
up to 4 shells. The constraint of this retrofit case is
that there is no splitting on hot streams. The cost for
stream splitting and repiping is $20,000. The
investment costs of area are shown in equation (3),
(4), (5), and (6).
Exchanger ($) = 26,460 + [389×Area (m
2
)]
(3)
Area addition ($) = 13,230 + [857×Area
added
(m
2
)]
(4)
Area reduction ($) = 13,230 + [5×Area
reduced
(m
2
)]
(5)
N
ew shell ($) = 26,460 + [857×Area
shell
(m
2
)]
(6)
The optimal retrofit design (Retrofit Design 2)
from the graphically searching technique is the one
with HRAT = 27.43
o
C, giving the highest NPV of
SIMULTECH 2011 - 1st International Conference on Simulation and Modeling Methodologies, Technologies and
Applications
306
$11,529,511 for a lifetime of 5 years, as shown in
Figure 9.
Figure 9: Graphical technique for searching optimization.
Details of Retrofit Design 2 are provided in
Table 7 and Figures 10 and 11. It will be applied to
handle light and heavy crude, giving different
furnace and cooler duties.
Table 7: Exchanger details of Retrofit Design 2.
5 CONCLUSIONS
The 10-stage model of HEN generates six retrofit
designs of crude preheat train. Designs 1, 2, and 3
are suitable for light crude for 200 days per year, and
Designs 4, 5, and 6 are suitable for heavy crude for
150 days per year. In comparing the NPV of the six
designs, it is shown that the optimal retrofit design
handling light crude for 200 days and heavy crude
for 150 days per year is Retrofit Design 2, which
gives the optimal NPV of $11,529,511 for a 5-year
lifetime and results in 32% saving at the furnace.
Figure 10: Design 2 with light crude feed.
Figure 11: Design 2 with heavy crude feed.
ACKNOWLEDGEMENTS
The authors would like to express their gratitude to
the Government Budget Bureau, the Petroleum and
Petrochemical College, Chulalongkorn University,
and the National Center of Excellence for
Petroleum, Petrochemicals and Advanced Materials
ETROFIT OF CRUDE PREHEAT TRAIN WITH MULTIPLE TYPES OF CRUDE
307
for funding support. The invaluable assistance of
Prof. Miguel Bagajewicz for educating us in
mathematical programming, GAMS, is also
gratefully acknowledged.
REFERENCES
Briones, V., Kokossis, A. C., 1998. Hypertargets: a
conceptual programming approach for the optimisa-
tion of industrial heat exchanger networks — II.
Retrofit design. Chemical Engineering Science 54,
541-561.
Ciric, A.
R., Floudas, C. A., 1988. A retrofit approach for
heat exchanger networks. Computers and Chemical
Engineering 13, 703-713.
Tjoe, T. N., & Linhoff, B., 1986. Using pinch technology
for process retrofit. Chemical Engineering 28, 47-60.
Yee, T. F., Grossmann, I. E., 1990. Simultaneous
optimisation models for heat integration — II. Heat
exchanger network synthesis. Computers and
Chemical Engineering 14, 1165-1184.
Zamora, J. M., Grossmann, I. E., 1996. A global MINLP
optimization algorithm for the synthesis of heat
exchanger networks with no stream splits. Computers
and Chemical Engineering 22, 367-384.
SIMULTECH 2011 - 1st International Conference on Simulation and Modeling Methodologies, Technologies and
Applications
308