See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/321305151

Economic Evaluation of NGL Recovery Process Schemes for Lean Feed

Compositions

Article  in  Chemical Engineering Research and Design · November 2017

DOI: 10.1016/j.cherd.2017.11.027

CITATIONS READS

7 506

2 authors, including:

Chunhe Jin
Seoul National University
3 PUBLICATIONS   9 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Chunhe Jin on 31 January 2019.

The user has requested enhancement of the downloaded file.

 chemical engineering research and design 1 2 9 ( 2 0 1 8 ) 297–305

Contents lists available at ScienceDirect

Chemical Engineering Research and Design

journal homepage: www.elsevier.com/locate/cherd

Economic evaluation of NGL recovery process

schemes for lean feed compositions

Chunhe Jin a , Youngsub Lim a,b,∗

a Department of Naval Architecture and Ocean Engineering, College of Engineering, Seoul National University, 1
Gwanak-ro, Gwanak-gu, Seoul 08826, Republic of Korea
b Research Institute of Marine Systems Engineering, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul

08826, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: The economic analysis of lean feed conditions for natural gas liquids (NGL) recovery is
Received 18 July 2017 required to meet recent increases in the demand for lean gas reservoirs. This paper com-
Received in revised form 9 pares the economic performance of various NGL recovery schemes with consideration of
November 2017 the LNG HHV specification under various lean feed compositions. The ISS and IPSI schemes
Accepted 18 November 2017 are typical processes in NGL recovery, especially in the normal and rich fields, whereas HHC
Available online 26 November 2017 separator and scrub column schemes have advantages for lean feeds. All the schemes were
built by HYSYS simulation software and optimized using a genetic algorithm. The profitabil-
Keywords: ity analysis indicates that the gross profit and payout time give a similar trend compared to
Lean feed compositions the previous studies. However, the results give different outcomes when considering total
NGL recovery annualized cost (TAC), especially in terms of lean feed condition. The results show that scrub
Economic analysis column scheme shows the best economic performance compared to other schemes under
LNG specification lean feed compositions.
Higher heating value (HHV) © 2017 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1. Introduction economical way to achieve NGL separation. The industry-standard

Single Stage (ISS) process is a representative process and uses a turbo-
Natural gas is a globally important energy source for which is increasing expander instead of a Joule–Thomson (JT) valve for greater cooling
rapidly, especially in East Asia (IGU, 2016). Lean natural gas production (Campbell and Wilkinson, 1981). Nevertheless, this process is limited
has increased since 2005 with the growth in production from shale by relatively low NGL recovery and problematic carbon dioxide freez-
gas reservoirs. The increase in unconventional shale gas reservoirs in ing (Lynch et al., 2002). Thus, many advanced process schemes have
Australia, East Africa, and the U.S., and the global demand for cleaner been developed and introduced, such as the gas sub-cooled process
energy resources is expected to lead to increased exploitation of lean (GSP), cold residue reflux (CRR) process, and IPSI process (IPSI is also
gas reservoirs in the next few years (Smith and Doong, 2016; Brussol the company name which owns the process). The heavy hydrocarbon
and Gadelle, 2013). (HHC) separator and scrub column concepts have also been introduced
Due to the higher value of heavier hydrocarbons as well as the in some international conventions specific to lean natural gas feed
necessity of maintaining the pipeline specification for natural gas conditions.
transportation, natural gas liquids (NGL) is often extracted from the The basic ISS process (Fig. 1) normally involves cooling the inlet
original feed gas. Moreover, to avoid icing of heavy hydrocarbons such feed gas using a turbo expander and expansion valve. After passing
as benzene, toluene and xylene isomers (BTX) during liquefaction and a liquid–vapor separator, the vapor portion is further expanded with a
plugging inside cryogenic heat exchangers are also the critical reasons turbo-expander, while the liquid portion is expanded to operating pres-
for extracting heavy hydrocarbons. Previous research and industrial sure with an expansion valve before introducing to the de-methanizer
experience have shown that cryogenic processes are normally the most (Campbell and Wilkinson, 1981). Many previous studies have used the

∗
Corresponding author at: Department of Naval Architecture and Ocean Engineering, College of Engineering, Seoul National University,
1 Gwanak-ro, Gwanak-gu, Seoul 08826, Republic of Korea.
E-mail address: [email protected] (Y. Lim).
https://doi.org/10.1016/j.cherd.2017.11.027
0263-8762/© 2017 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
298 chemical engineering research and design 1 2 9 ( 2 0 1 8 ) 297–305

Fig. 1 – The ISS process scheme.

ISS process as a base for further development because of its simple a separator. Even though they have been introduced in U.S. patents,
configuration and low CAPEX (Getu et al., 2013; Park et al., 2015; Jibril publications comparing HHC separator and scrub column economic
and Idriss, 2006). performance with other NGL recovery processes are rare (Chen et al.,
The IPSI process scheme (Fig. 2) is highly efficient and provides good 2014; Roberts and Brostow, 2010; Brostow and Roberts, 2013).
economic performance, especially in separating propane, propylene, According to the previous NGL recovery studies, conventional NGL
and heavier hydrocarbon liquids, as shown by Yao et al. (1999). Unlike recovery studies that compare recovery performance or economic cost
other NGL processes, IPSI focused on improving the tower stripping analysis are usually based on a normal or rich feed (Getu et al., 2015;
section. A portion of hydrocarbon liquid is withdrawn from one of the Van Duc Long and Lee, 2013; Mehrpooya et al., 2010a,b). The ethane
bottom trays and expanded and heated by inlet gas to produce a two recovery plant in the South Pars gas field was simulated and analyzed
phase, the vapor phase is recycled to the column increasing the light with an advanced exergy analysis without economic analysis under a
hydrocarbon component concentration and the efficiency of separation rich feed condition (Mehrpooya et al., 2016). An energy efficient design
(Yao et al., 1999). The previous research Getu et al. (2013) also main- was proposed based on an analysis of retrofitting an NGL fractiona-
tained that the highest economic performance, considering both the tion process, and it was reported that the methodology could save
total capital cost and operating cost, compared with other NGL recovery considerable operating costs compared to the original case in terms
processes under eight feed compositions. of side reboiler and heat pump hybrid systems (Long et al., 2016).
The heavy hydrocarbon (HHC) separator and scrub column con- The existing NGL recovery plant in Sirri Island was energy and exergy
cepts are integrated NGL recovery process into the liquefaction unit and analyzed with five different normal and rich feed conditions, and it
could consider for lean natural gas feed NGL recovery (Krishnamurthy was reported that an increase in methane reduced the work for com-
and Liu, 2013). Typically, scrub column and HHC separator can be used pressors and the cold heat exchanger showed the highest exergy loss
when the natural gas feed is lean, with low C2+ contents. These pro- (Tahmasebi et al., 2015). This study also not included the economy
cesses have the advantage of lower capital cost than other NGL recovery effect. NGL recovery was maximized using a commercial simulation
schemes, but they have the limitation of low NGL recovery efficiency. HYSYS and also assessed the de-methanizer pressure change effect
Both scrub column and HHC separator schemes remove heavy hydro- in a rich feed condition, where the study did not show the effect of
carbons after the pre-cooling and before the liquefaction processes capital cost increase when maximizing NGL recovery (Abdulrahman
(refer to Figs. 3 and 4). Compared with the HHC separator, the scrub et al., 2015). An integrated NGL/LNG configuration was introduced
column has less C1 and C2 loss because it uses a column instead of and analyzed considering capital cost, the result showed high ethane

Fig. 2 – The IPSI process scheme.

 chemical engineering research and design 1 2 9 ( 2 0 1 8 ) 297–305 299

Fig. 3 – The HHC separator process scheme.

Fig. 4 – The scrub column process scheme.

recovery and considerable liquefaction efficiency under a rich feed according to the IGU report (IGU, 2016), normally have an HHV spec-
composition (Vatani et al., 2013). Another integrated recovery plant was ification between 39.7 and 43.5 MJ/m3 ; a higher minimum value than
optimized using several process parameters under a rich feed condi- U.S. and Europe (Coyle et al., 2007). Getu et al. (2013) reported the eco-
tion, they analyze plant performance with the object function of net nomic performance of various NGL recovery processes for eight feed
profit. (Mehrpooya et al., 2010b). A conceptual process design was pro- compositions but did not consider simpler processes like HHC sepa-
posed that integrated shale gas NGL recovery and LNG re-gasification rator and scrub column concepts for lean feeds. Ghorbani et al. (2012)
configuration for energy saving (Wang and Xu, 2014). The study did not used exergy pinch analysis to optimize NGL recovery plants but limited
present the economic performance effect when design energy saving their investigation to a normal feed. Also, their studies focused only on
process as well. Halvorsen et al. (2016) demonstrated a dividing-wall recovery efficiency or economic performance without considering the
column could replace the de-ethanizer and de-propanizer columns in HHV specification. Park et al. (2015) included both economic evalua-
a NGL recovery process, which enabled weight reduction in the float- tion and HHV specification in their comparison of various patented
ing liquefied natural gas facilities. A hybrid genetic algorithm (GA) NGL process schemes, but considered only one normal feed condition.
method was used to optimize the economic problem of the turbo- In this paper we focused on four different NGL recovery processes
expander (ISS) recovery process (Jang et al., 2005). A distillation system with lean feed conditions and evaluated their economic performance
design was optimized for obtaining good energy efficient separation based on their total annualized cost (TAC). We also considered the LNG
of multi-component mixtures in the NGL processes (Yoo et al., 2016). HHV specification, which is a requirement for East Asian countries,
Heavy hydrocarbon recovery process of the CRR was optimized to maxi- including South Korea and Japan.
mize the ethane recovery at various demethanizer pressure conditions
and compared with GSP and conventional turbo expander processes 2. Process selection and description
(Kherbeck and Chebbi, 2015).
As abovementioned most of the previous studies focused on a nor-
We selected four types of NGL recovery schemes for economic
mal or rich feed condition. However, the increasing production of lean
feeds necessitates a re-evaluation of the economic performance of
evaluation: ISS, IPSI, HHC separator, and scrub column. ISS is
NGL recovery processes based on lean feeds. In particular, the require- a well-known process in NGL recovery that leads to a major
ments of LNG HHV specifications in East Asia are relatively high, which development in this field, from which the GSP, CRR, RSV and
weakens the benefits of high efficiency NGL recovery processes. South IPSI schemes emerged. Hence, ISS is selected as the base
Korea and Japan, whose market share is reported to be more than 47% case. IPSI is chosen because it is a recently reported advanced
300 chemical engineering research and design 1 2 9 ( 2 0 1 8 ) 297–305

process with high recovery performance for NGL recovery, 12 is expanded via turbo expander (K-101), leading to efficient
which is also reported to have a good economic performance power generation for use by compressor (K-102), to a tempera-
compared to various other schemes (Getu et al., 2013). HHC ture at −40 ◦ C before being fed into the column. The tower top
separator and scrub column schemes were selected because product stream 15 is heated to 24 ◦ C by heat exchangers (E-
they are expected to have better economic performance in 105 and E-100) then compressed via compressors (K-102 and
lean feed compositions. All the selected process models were K-103) to a pressure of around 60 bar. Stream 19 is heated by
developed using the commercial software Aspen HYSYS, and compression and must be cooled to 30 ◦ C (E-106) in order to
the Peng–Robinson equation of state was selected for use with meet NG specifications.
HYSYS simulation for its good ability to make predictions in The major improvement of IPSI over ISS or GSP is its
hydrocarbon mixtures (Mehrpooya et al., 2010a). advanced utilization of the column bottom pump around
streams. The pump-around streams PA1 and PA2 from the
2.1. Industry single-stage process (ISS) T-100 bottom trays were used to cool stream 5 and 4, after
that the warmed PA1 and PA2 streams were returned to the
The ISS scheme process flow diagram is depicted in Fig. 1. column, respectively. These two pump-around streams not
After the removal of acid gas and water, the treated feed gas only reduce the column (T-100) reboiler duty but also provide
stream, which has a pressure of 60 bar and a temperature of refrigeration to the feed stream as they get warmer by cool-
30 ◦ C, passes the heat exchanger (E-100) and it is cooled down ing streams 5 and 4. However, column pump-around streams
to about −33 ◦ C by gas stream 6, which is from the top product can also be applied in ISS and other NGL recovery schemes.
stream of the column (T-100). Stream 1 is then flashed in the The IPSI scheme achieves its considerable enhancement by
separator (V-100) to vapor stream 2 and liquid stream 3. Vapor using the PA3 stream, which is divided into stream 20 and 21
stream 2 is depressurized by the turbo-expander (K-100) to after being stripped from the T-100 bottom tray. Stream 20 is
column top pressure and introduced to the column top stage, warmed in heat exchanger E-101 after cooling feed gas stream
where the energy generated by the expander is used to run the 2 and then mixed with stream 27 before being returned to the
compressor (K-101). Liquid stream 3 from the flash separator column bottom stage. On the other hand, stream 21 reduces its
(V-100) is expanded using a JT valve (VLV-100) to column pres- pressure via expansion valve and heat exchange with stream 3
sure and introduced to the column (T-100). The column top before entering the flash separator (V-101). The flashed liquid
product stream 6 is used to cool the feed gas stream and after stream 24, which contains heavier hydrocarbons, is pumped
passing the heat exchanger (E-100) the temperature rises to and mixed with the column (T-100) bottom stream and later
about 23 ◦ C. Residual gas stream 7 is compressed by the com- becomes the NGL product. However, vapor stream 25 is com-
pressor (K-101 and K-102) to meet the specified natural gas pressed in K-100 to the column pressure and cooled by air
pressure of about 60 bar, then cooled to the specified temper- cooler, then mixed with stream 28 before returning to the col-
ature of 30 ◦ C by the heat exchanger (E-101). umn bottom tray. Stream 28 can decrease the required reboiler
The liquefaction section is not the focus of this study; we duty because the PA3 stream gets warmer. Additionally, the
instead focus on the economic performance of the NGL recov- relatively high concentration of light components in the PA3
ery section. Therefore, it was assumed that the total capital stream means that the temperature profile inside the column
expenditure is a constant value in each scheme and does is reduced, which increases its heat integration ability (Getu
not affect the economic evaluation of NGL recovery. The NG et al., 2013).
stream enters the liquefaction process and its temperature
is reduced to about −160 ◦ C to produce liquefied natural gas. 2.3. Heavy hydrocarbon (HHC) separator & scrub
Stream 10 is further expanded to about 1 bar by JT valve. After column process
expansion, stream 11 may contain a small portion of vapor
phase, which is separated through the end flash while the Unlike NGL recovery processes that are independent of the
liquid is sent to LNG storage. liquefaction process, the treated feed gas in the HHC sep-
arator scheme directly enters the liquefaction process, and
2.2. Enhanced NGL recovery (IPSI) NGL recovery is performed during the precooling section in
the main cryogenic heat exchanger. The HHC separator and
IPSI is a recently patented NGL recovery scheme that uses scrub column process flow diagrams are shown in Figs. 3 and 4.
the column side streams to improve ethane recovery with- Stream 1 is stripped from the cryogenic heat exchanger and
out additional refrigeration systems. Fig. 2 shows the process adjusted by E-100 exchanger to achieve optimum temperature,
flow diagram of IPSI. The cleaned feed gas stream is divided before entering the V-100 flash separator. The liquid stream is
into stream 1 and stream 2. Stream 1 is first sent to the heat the recovered NGL stream, which contains relatively heavier
exchanger (E-100) and cooled by residue gas stream 16 before hydrocarbons. On the other hand, the remaining vapor stream
being mixed with stream 6 and entering the cold separator NG is reinjected to the liquefaction process to produce the LNG
(V-100). Stream 2 is successively cooled via heat exchangers (E- product. The scrub column process is similar to the HHC sep-
101, E-102, E-103, and E-104) by the column side pump around arator process but is modified by the use of the column (T-100)
streams PA1, PA2 and PA3 then mixed into stream 8 at a tem- instead of the separator (V-100). The E-100 exchanger in the
perature of about −35 ◦ C. Stream 8 then moves to the separator HHC separator and Scrub column schemes is a temperature
(V-100) and is flashed into a vapor stream and liquid stream 9. controller for simulation purpose, to find the most econom-
Liquid stream 9 is expanded via JT valve to column pressure ical temperature for separating NGL when optimization. In
and introduced to the column, while the flashed vapor stream a real plant, after finding the optimal value, it will not need
is divided into two streams, with 30% going into stream 11 the exchanger except for plant control. The E-101 exchanger
and 70% to stream 12. Stream 11 is further cooled in the heat is used for case studies to find the optimal temperature for
exchanger (E-105) by the column top product stream 15 and reducing end flash gas. In this study, we have fixed the same
introduced as top feed stream 14 into column T-100. Stream temperature of outlet stream 4. Both the HHC separator and
 chemical engineering research and design 1 2 9 ( 2 0 1 8 ) 297–305 301

respectively. The compressor, turbine and pump efficiency

Table 1 – Feeds composition (mole %).
were fixed at 80%. The pressure drop between heat exchangers
Component Feed 1 Feed 2 Feed 3 Feed 4
was assumed to be 0.5 bar and 3 ◦ C of the minimum temper-
C1 92.97 91.57 91.22 90.67 ature approach was applied. The heavy hydrocarbons (C5+)
C2 4.18 5.18 5.21 5.23 concentration in the LNG product should not exceed 0.1 mol%
C3 1.23 1.5 1.68 1.98 due to the likelihood of freezing during the liquefaction pro-
iC4 0.54 0.5 0.55 0.45 cess (GPSA, 2012). The addition of ethane or propane at the
nC4 0.54 0.48 0.52 0.65
end of liquefaction to meet the HHV specification is not desir-
iC5 0.04 0.15 0.16 0.25
nC5 0.04 0.1 0.11 0.23
able for economic and technical reasons (Coyle et al., 2007).
C6 0.03 0.01 0.01 0.02 Therefore, it is preferred to adjust the HHV in the NGL recovery
C7 0 0.01 0.01 0.03 process. The HHV of LNG product was specified to be mini-
C8 0 0 0 0.02 mum value of 40.50 MJ/m3 (Park et al., 2015).
C9 0 0 0 0.01
C10+ 0 0 0 0.01
N2 0.43 0.5 0.53 0.45 4. Economic analysis
CO2 0 0 0 0
GPM value 1.84 2.2 2.3 2.5
4.1. Capital cost estimation

Table 2 – Common process key parameters and The major equipment of a distillation column, including the
constraints. separator, plate-fin type heat exchanger, shell and tube heat
LNG products 2 MTPA exchanger, compressor, turbine and pump were considered in
Plant inlet gas pressure 60 bar the cost estimation. The sizing of each equipment is required
Plant inlet gas temperature 30 ◦ C
to estimate capital cost, and some parameters were obtained
Compressor/turbine/pump efficiency 80%
from HYSYS simulation results.
Pressure drop across the heat exchanger 0.5 bar
Heat exchangers minimum temperature 3 ◦C Column sizing requires the diameter and height to be cal-
approach culated. Eq. (1) was introduced by Towler and Sinnott (2012)
HHV (LNG) min. 40.50 MJ/m3 for calculation the column diameter and is based on the
C4 (LNG) 2 mol% max. well-known Souders and Brown equation, which gives the
C5+ (LNG) 0.1 mol% max. maximum vapor velocity and the column diameter.
Property fluid package Peng Robinson


scrub column schemes are advantageous when the feed gas is    L − v 0.5 4Vw
uv = −0.17lt 2 + 0.27lt − 0.047 , Dc =
lean because the complex recovery processes of ISS and IPSI v v uv
are rendered inefficient by the low amount of heavy hydro- (1)
carbon contents. Furthermore, the HHC separator and scrub
column concepts can reduce capital cost.
where uv and lt denote the maximum allowable vapor veloc-
ity and plate spacing (range 0.5–1.5), respectively. Vw is the
3. Feed compositions and specifications
maximum vapor rate which can be obtained easily from the
simulation. The distillation column height is calculated based
Feed composition is a crucial factor in selecting an NGL recov-
on HETP (height equivalent to theoretical plate) and a HETP of
ery scheme (Getu et al., 2013). Feed composition can also vary
0.85 m was used here (Towler and Sinnott, 2012). The separator
over time. Four lean feeds with GPM values lower than 2.5 were
diameter calculation uses the gas flow rate Eq. (2), which uses
selected to cover the wide range of available lean feed compo-
the droplet settling velocity (Manning and Thompson, 1995).
sitions, as shown in Table 1. The GPM is defined as the amount
of recoverable liquid, expressed in gallons per 1000 standard
cubic feet of gas at 60 ◦ F. A feed is termed lean when the GPM
 0.5
P Ts 1 l − g
value is lower than 2.5, between the GPM value of 2.5–5 called qs = 67824Ks d2 (1 − M) (2)
Ps T Z g
normal rich and the value higher than 5 considered as very rich
feed (Manning and Thompson, 1991). Additionally, the tradi-
tional NGL recovery process is carried out after the AGRU (acid where qs and Ks stand for gas flow rate and sizing parameter,
gas removal unit) and dehydration processes. Therefore, the respectively. The sizing parameter can be obtained either from
feed compositions shown in Table 1 do not include acid gas the API recommendation or from the droplet-settling equa-
and water components. tion, where we use a constant value of 0.1 in this study. The
It is important to specify the key process parameters, which overall heat transfer coefficient (U value) of 454.24 W/m2 ◦ C is
have a major effect on plant performance, to compare the used to calculate the heat exchange area of a heat exchanger
different NGL schemes. Thus, these parameters are fixed to (Arnold and Stewart, 2012). The compressor, turbine and
certain values and conditions for faring comparisons (Getu pump cost rely on the power consumption of each piece of
et al., 2013). The specified key process parameters are shown equipment, and the power duty is easily calculated from the
in Table 2. The LNG specifications in Table 2 we referenced simulation.
from GPSA (2012) and previous study conducted by Park et al. The Engineering Sciences Data Unit (ESDU) provides the
(2015). The LNG production rate was fixed at 2MTPA (mil- costing method for multi-stream plate-fin-type heat exchang-
lion ton per annum) for each scheme. The feed flow rate was ers, and ESDU 97006 was introduced for the selection and
adjusted to satisfy the constant production rate, and the inlet costing of plate-fin type heat exchangers (ESDU, 2003). The
feed pressure and temperature were kept at 60 bar, and 30 ◦ C, total exchanger volume should be obtained to calculate the
302 chemical engineering research and design 1 2 9 ( 2 0 1 8 ) 297–305

Table 3 – Equipment cost data parameters.

Equipment type Description K1 K2 K3 Capacity, units

Compressors Centrifugal, reciprocating 2.2897 1.3604 −0.1027 Power, KW

Turbines Axial gas turbines 2.7051 1.4398 −0.1776 Power, KW
Heat exchangers U-tube 4.1884 −0.2503 0.1974 Area, m2
Process vessels Vertical 3.4974 0.4485 0.1074 Volume, m3
Pumps Reciprocating 3.8696 0.3161 0.122 Power, KW
Towers Tray and packed 3.4974 0.4485 0.1074 Volume, m3

cost of the exchanger. First, a mean volumetric coefficient Bz turbine, and pump, and the reference unit price of electric-
is obtained from Eq. (3). ity and the natural gas cost were obtained from the EIA (U.S.
Energy Information Administration) recent three years price
· ·
Qz Q n
for reference.
i
= (3) Unlike ISS and IPSI, the HHC separator and Scrub column
Bz ˇi
i=1 schemes should include additional refrigeration cost, extra
. precooling of NGL flow, for fair comparison of total operating
where Qz is the heat transferred in the zone,.
n is the total cost because they get precooling refrigeration from integrated
number of streams involved in the zone, Qi and ˇi denote the liquefaction system. The refrigeration cost is calculated by
amount of heat transferred to the ith stream in the zone and applying the liquefaction efficiency reported from previous
the volumetric film coefficient at ith stream, respectively. Next, references (Barclay and Shukri, 2000; Lim et al., 2014). The
the heat exchanger volume is calculated from Eq. (4). liquefaction efficiency in this study assumes 14 kW/t/day and
. with the electricity cost of 0.1$/kWh, we could obtain the oper-
Qz /Tm,z ating cost 33.6$/t. Then multiply NGL mass flow rate for getting
Vz = (4)
Bz additional refrigeration cost for HHC separator and Scrub col-
umn schemes. In this study, we focused on NGL recovery part
where Tm,z is the logarithmic mean temperature in the zone.
and neglect the heat exchanger cost effect of liquefaction cycle
Finally, the total active volume of the exchanger is calculated
because the recovered NGL mass flow is very small compared
by summing the volumes for each zone with an allowance of
to LNG production (about 1.5% of LNG production) under the
15% for headers and distributers.
lean feed compositions. Thus, we ignore the increasing size of
After sizing the major equipment, the purchase cost is esti-
cryogenic heat exchanger in liquefaction part.
mated by using the below cost Eq. (5), which is commonly
applied in the preliminary design stage (Turton et al., 2012).
4.3. Profitability analysis
2
log10 C0p = K1 + K2 log10 (A) + K3 [log10 (A)] (5)
The profitability analysis is made for each scheme for com-
paring net profit and payout time. The net profit is obtained by
where C0p is the equipment purchase cost. A is the capacity or
gross profit minus tax (the tax rate assumed 30% in this study).
size parameter for equipment. For compressors, turbines and
The gross profit is calculated by products revenue minus oper-
pumps it represents power consumption. For heat exchang-
ating cost (Peters et al., 2002). The payout time, also called
ers and vessels it represents area and volume, respectively.
payback time in some research, is evaluated and compared to
The correlation parameter data for K1, K2, K3 used in the each
each process. Payout time is the total capital cost divided to
equipment as below Table 3 which is taken from Turton et al.
the net profit which is expressed as Eq. (6) (Biegler et al., 1997).
(2012).
The total capital cost (TCC) can then be obtained by multi-
Total capital cost
plying the Lang factor (this study used factor of 4.74 which was Payout time = (6)
Net profit
suggested for fluid processing plant) by the major equipment
purchase costs, as well as considering the time value, which 4.4. Total annualized cost (TAC) optimization
is based on the current chemical engineering plant and cost
index (CEPCI) (Peters et al., 2002). The annualized total capital cost (TCC) is expressed on an
annual basis by assuming an interest rate of 5% over 10 years
4.2. Operating cost estimation for the economic evaluation. In this study, we considered the
TAC to be an object function of optimization. Here, the TAC
Calculating the operating cost is not easy, and it mainly relies is defined as the sum of annualized TCC and, total operating
on energy costs which are may vary with area by area and year cost minus byproduct credits as shown in Eq. (7).
by year. In addition, utility costs like electricity are the main
utilities and directly influenced by the cost of fuel (Getu et al., TAC = Annualized TCC + Total operating cost − Byproduct credits
2013). The cost of utilities, such as electricity and refrigera-
tion, and the feed raw material cost are the main operating (7)
costs considered in this research. The raw material cost in
here represents the cost of feed natural gas. The steam cost is In which the byproducts credits include each component
ignored in this study because the column operating in similar of the NGL (C2+ ) cost.
conditions, and accordingly reboiler temperature and column The genetic algorithm was used by linking Matlab with
bottom NGL mass flow are also no big difference with all the HYSYS optimization to find the minimum TAC. The GA
candidate four schemes which result less effect in total oper- method has been used in some previous studies, and its appli-
ating cost. Electricity is mainly consumed by the compressor, cations have been tested (Mehrpooya et al., 2006). GA is an
 chemical engineering research and design 1 2 9 ( 2 0 1 8 ) 297–305 303

Table 4 – Optimization variables by each scheme.

Optimization variables

ISS T1 , P4
IPSI F1 , F11 , F21 , T7 , T3 , T4 , T5 , T6 ,
T14 , P14
HHC separator T2
Scrub column T2

intelligent random search algorithms based on the mechan-

ics of natural evolution, it utilizes historical information and
finally could yield a global optimum by repeating a series of
Fig. 5 – TAC analysis for each feed composition.
population selection and generation for complex optimiza-
tion problems which have many local optimum solutions. The economic analysis from literature also can provide relative
parameter of population uses in this study is 30 and cross over comparison among different processes when design and opti-
rate of 0.1 for GA. The process constraints are given as above mize chemical processes. The effect of the total capital cost is
mentioned Table 2 and optimization variables by each scheme bigger than that of the raw material cost, it means that even
are listed in below Table 4. Where T, P and F stands for tem- though IPSI has the lowest raw material cost, the higher capital
perature, pressure and flow rate, respectively. The subscript cost and the operating cost offset its economic performance.
number represents stream number of each process. The HHC separator scheme has the lowest total capital cost,
but its higher operating cost affects its economic performance.
5. Results and discussion According to the economic analysis shown in Table 5, ISS
has the most favorable utility cost. However, IPSI gives bet-
The results of profitability analysis for the selected four ter results for raw material cost, while the HHC separator has
schemes is shown in Table 5, which gives the average values of the lowest total capital cost due to its simple design concept.
the four different feed types for good comparison. The average Because of this dissimilarity in each scheme, we defined the
value would not affect the total trends of this study since the TAC to facilitate comparison between these different schemes
trend of each feed condition by each scheme is similar. The and perform an overall economic evaluation which was con-
total capital cost result shows that IPSI has the highest value, sidering annualized total capital cost, total operating cost and
whereas the HHC separator scheme has the lowest value of byproduct credit as discussed in part 4.4. The TAC value for
the other schemes due to its simple configuration and fewer each feed is shown in Fig. 5. For feed 1 (GPM 1.84) the scrub
number of process equipment. The most complex scheme IPSI column scheme had the lowest TAC value, and IPSI gave the
gives 29.2% higher capital cost than the ISS scheme because highest TAC. The savings of the scrub column scheme were
IPSI needs additional compressors and some heat exchangers 0.49% and 0.72% over the ISS scheme and the HHC separa-
for improvements to the column’s bottom section. tor schemes, respectively. The scrub column also had the best
The total operating cost is mainly composed of the util- results for feed 2 (GPM 2.2), saving about 0.06% and 0.39% com-
ity cost and raw material cost. The IPSI process scheme has pared to the ISS and HHC separator schemes, respectively.
the lowest raw material cost compared to the other schemes, The economic advantage of the scrub column lessened with
whereas it has the highest capital cost. The raw material sav- increasing feed GPM value. When the GPM value reached 2.3
ing for IPSI is 0.02% compared to ISS, and 0.05% and 1.66% (Feed 3), both the scrub column and ISS schemes had almost
compared to the scrub column and HHC separator schemes, the same TAC value. Both of these schemes had a lower value
respectively. Because this study focused on lean feed condi- than the IPSI and HHC separator schemes by 0.33% and 1.8%,
tions, the high efficiency of the IPSI is not so remarkable. The respectively. However, when the GPM value reached 2.5 (Feed
gross profit of ISS and IPSI are similar and both of them higher 4), the ISS scheme had the lowest TAC value, which was 0.32%
than the other two schemes, whereas the payout time for lower than the scrub column and up to 2.62% lower than the
HHC separator and scrub column schemes are significantly HHC separator scheme.
lower than ISS and IPSI schemes because of simple configu-
ration. The payout time shows the same order of magnitude 6. Conclusions
compared to the previous study by Getu et al. (2013). The pay-
out time is very short because the capital cost estimation Economic analysis was performed for the ISS, IPSI, HHC sepa-
used here is study estimate which may have order of mag- rator and scrub column NGL recovery schemes with various
nitude accuracy. Nevertheless, we believe that the method of lean feeds conditions. We used the TAC (Total Annualized

Table 5 – Profitability analysis for the selected NGL schemes.

ISS IPSI HHC separator Scrub column

Total capital cost ($) 12,505,611 17,662,398 771,518 1,517,911

Total operating cost ($/year) 344,977,737 345,229,183 352,179,175 345,885,252
1. Utility cost ($/year) 754,981 1,089,377 2,240,383 1,561,578
2. Raw material cost ($/year) 344,222,756 344,139,807 349,938,792 344,323,674
By products credits ($/year) 29,705,007 29,672,069 31,070,406 29,508,476
Gross profit ($/year) 364,727,270 364,442,886 358,891,231 363,623,224
Net profit ($/year) 255,309,089 255,110,020 251,223,862 254,536,257
Payout time (year) 0.049 0.069 0.003 0.006
304 chemical engineering research and design 1 2 9 ( 2 0 1 8 ) 297–305

Cost) to compare the overall economic performance of the Getu, M., Mahadzir, S., Samyudia, Y., Khan, M.S., Bahadori, A.,
selected NGL recovery schemes, which included the annual- Lee, M., 2015. Risk-based optimization for representative
ized total capital cost, operating cost and byproduct credits. natural gas liquid (NGL) recovery processes by considering
uncertainty from the plant inlet. J. Nat. Gas Sci. Eng. 27 (Part
The HHV specification was also incorporated to meet the
1), 42–54.
requirement of the regional specifications. The results show Ghorbani, B., Salehi, G.R., Ghaemmaleki, H., Amidpour, M.,
that the scrub column scheme has the best performance when Hamedi, M.H., 2012. Simulation and optimization of
the feed is lean. When the feed GPM value reaches 2.5, the refrigeration cycle in NGL recovery plants with exergy-pinch
traditional ISS scheme gives better performance. The HHC analysis. J. Nat. Gas Sci. Eng. 7, 35–43.
separator and IPSI schemes have the lowest and highest total GPSA, 2012. GPSA Engineering Data Book, 13 ed. Gas Processors
Suppliers Association.
capital cost due to the differences in configurational complex-
Halvorsen, I.J., Dejanovic, I., Marak, K.A., Olujic, Z., Skogestad, S.,
ity. IPSI gives the best performance in terms of minimum raw
2016. Dividing-wall column for fractionation of natural gas
material cost with respect to its high efficiency. Nevertheless, liquids in floating liquefied natural gas plants. Chem. Eng.
its high total capital cost and operating cost offset its economic Technol. 39, 2348–2354.
performance. The HHC separator scheme has the lowest total IGU, 2016. IGU World LNG Report-2016 Edition, LNG 18
capital cost but shows the highest raw material cost due to Conference & Exhibition Edition.
its poor separation efficiency compared to the other schemes. Jang, W.H., Hahn, J., Hall, K.R., 2005. Genetic/quadratic search
algorithm for plant economic optimizations using a process
The consideration of lean feeds meant that the required raw
simulator. Comput. Chem. Eng. 30, 285–294.
material cost shows little deviation among the schemes. These
Jibril, A., Idriss, I., 2006. Simulation study determines optimum
results demonstrate that when the feed is considerably lean turboexpander process for NGL recovery. Oil Gas J. 104, 58–62.
(GPM value lower than 2.3), the scrub column scheme shows Kherbeck, L., Chebbi, R., 2015. Optimizing ethane recovery in
the best economic performance. turboexpander processes. J. Ind. Eng. Chem. 21, 292–297.
Krishnamurthy, G., Liu, Y.n., 2013. Removal of Heavy
Acknowledgements Hydrocarbons from Lean Natural Gas. AIChE Conference.
Lim, W., Lee, I., Tak, K., Cho, J.H., Ko, D., Moon, I., 2014. Efficient
configuration of a natural gas liquefaction process for energy
This work was partially supported by Technology Innovation recovery. Ind. Eng. Chem. Res. 53, 1973–1985.
Program (10060099) funded by the Ministry of Trade Industry Long, N.V.D., Minh, L.Q., Pham, T.N., Bahadori, A., Lee, M., 2016.
& Energy (MOTIE), Republic of Korea, the Program of Uni- Novel retrofit designs using a modified coordinate descent
versity Specialized for Offshore Plant Engineering, funded by methodology for improving energy efficiency of natural gas
the MOTIE, Republic of Korea, and Engineering Development liquid fractionation process. J. Nat. Gas Sci. Eng. 33, 458–468.
Lynch, Joe T., Cairo, C., Hudson, Hank M., 2002. Unique Design
Research Center (N0000990) funded by the MOTIE, Republic of
Challenges in the Aux Sable NGL Recovery Plant. 81st Annual
Korea.
GPA Convention.
Manning, F.S., Thompson, R.E., 1991. Oilfield Processing of
References Petroleum: Natural Gas. Pennwell books.
Manning, F.S., Thompson, R.E., 1995. Oilfield Processing of
Abdulrahman, R.K., Zangana, M.H.S., Sebastine, I.M., 2015. Petroleum: Crude Oil. Pennwell books.
Optimal NGL recovery from natural gas using turboexpander: Mehrpooya, M., Gharagheizi, F., Vatani, A., 2006. An optimization
a case study and simulation. Chem. Technol. Fuels Oils 51, of capital and operating alternatives in a NGL recovery unit.
536–538. Chem. Eng. Technol. 29, 1469–1480.
Arnold, K., Stewart, M., 2012. Surface Production Operations, Mehrpooya, M., Lazemzade, R., Sadaghiani, M.S., Parishani, H.,
Volume 2: Design of Gas-Handling Systems and Facilities, 3 2016. Energy and advanced exergy analysis of an existing
ed. Gulf Professional Publishing. hydrocarbon recovery process. Energy Convers. Manage. 123,
Barclay, M., Shukri, T., 2000. Enhanced single mixed refrigerant 523–534.
process for stranded gas liquefaction. In: 79th Annual Gas Mehrpooya, M., Vatani, A., Ali Mousavian, S.M., 2010a.
Processors Association Convention, Atlanta. Introducing a novel integrated NGL recovery process
Biegler, L.T., Grossmann, I.E., Westerberg, A.W., 1997. Systematic configuration (with a self-refrigeration system (open–closed
Methods for Chemical Process Design. cycle)) with minimum energy requirement. Chem. Eng.
Brostow, A.A., Roberts, M.J., 2013. Integrated NGL Recovery In the Process.: Process Intensif. 49, 376–388.
Production Of Liquefied Natural Gas. Prentice Hall, Old Mehrpooya, M., Vatani, A., Mousavian, S., 2010b. Optimum design
Tappan, NJ (United States) Google Patents. of integrated liquid recovery plants by variable population
Brussol, L., Gadelle, D., 2013. Lean LNG plants-heavy ends size genetic algorithm. Can. J. Chem. Eng. 88, 1054–1064.
removal and optimum recovery of liquid hydrocarbons for Park, J.H., Khan, M.S., Andika, R., Getu, M., Bahadori, A., Lee, M.,
refrigerant make-up. 17th International Conference & 2015. Techno-economic evaluation of a novel NGL recovery
Exhibition on Liquified Natural Gas (LNG 17). scheme with nine patented schemes for offshore
Campbell, R.E., Wilkinson, J.D., 1981. Hydrocarbon gas processing. applications. J. Nat. Gas Sci. Eng. 27, 2–17.
Google Patents. Peters, M., Timmerhaus, K., West, R., 2002. Plant Design and
Chen, F., Luo, X., Ott, C.M., Roberts, M.J., Krishnamurthy, G., 2014. Economics for Chemical Engineers, 5 ed. McGraw-Hill.
Heavy Hydrocarbon Removal From A Natural Gas Stream. Roberts, M.J., Brostow, A.A., 2010. Integrated NGL Recovery And
Google Patents. Liquefied Natural Gas Production. Google Patents.
Coyle, D., Vega, F., Durr, C., 2007. Natural gas specification Smith, T., Doong, S., 2016. Selective C5+ Removal for Lean Feed
challenges in the LNG industry. In: 15th International Gas. LNG-INDUSTRY.
Conference and Exhibition on Liquefied Natural Gas, Tahmasebi, S., Abbasabadi, A.B., Ghasemi, N., Javadian, H.,
Barcelona, Spain, pp. 1–21. Mashhadi, S., Fattahi, M., Arian, Y.R., Maddah, H., 2015.
ESDU, 2003. Selection and Costing of Heat Exchangers Plate-Fin Investigation of various feed conditions on NGL recovery
Type. ESDU. plant energy and exergy performance: a case study. J. Nat. Gas
Getu, M., Mahadzir, S., Long, N.V.D., Lee, M., 2013. Sci. Eng. 22, 83–89.
Techno-economic analysis of potential natural gas liquid Towler, G., Sinnott, R., 2012. Chemical Engineering Design, 2 ed.
(NGL) recovery processes under variations of feed Elsevier.
compositions. Chem. Eng. Res. Des. 91, 1272–1283.
 chemical engineering research and design 1 2 9 ( 2 0 1 8 ) 297–305 305

Turton, R., Bailie, R.C., Whiting, W.B., Shaeiwitz, J.A., 2012. Yao, J., Chen, J.J., Elliot, D.G., 1999. Enhanced NGL recovery
Analysis, Synthesis and Design of Chemical Processes. processes. Google Patents.
Pearson Education. Yoo, H., Binns, M., Jang, M.G., Cho, H., Kim, J.K., 2016. A design
Van Duc Long, N., Lee, M., 2013. A novel NGL (natural gas liquid) procedure for heat-integrated distillation column sequencing
recovery process based on self-heat recuperation. Energy 57, of natural gas liquid fractionation processes. Korean J. Chem.
663–670. Eng. 33, 405–415.
Vatani, A., Mehrpooya, M., Tirandazi, B., 2013. A novel process
configuration for co-production of NGL and LNG with low
energy requirement. Chem. Eng. Process.: Process Intensif. 63,
16–24.
Wang, M., Xu, Q., 2014. Optimal design and operation for
simultaneous shale gas NGL recovery and LNG re-gasification
under uncertainties. Chem. Eng. Sci. 112, 130–142.

View publication stats