Accepted Manuscript

Title: System Optimization of Turbo-expander Process for

Natural Gas Liquid Recovery

Authors: Yajun Li, Feiyu Xu, Canteng Gong

PII: S0263-8762(17)30320-9
DOI: http://dx.doi.org/doi:10.1016/j.cherd.2017.06.001
Reference: CHERD 2703

To appear in:

Received date: 4-12-2016

Revised date: 25-4-2017
Accepted date: 2-6-2017

Please cite this article as: Li, Yajun, Xu, Feiyu, Gong, Canteng, System Optimization
of Turbo-expander Process for Natural Gas Liquid Recovery.Chemical Engineering
Research and Design http://dx.doi.org/10.1016/j.cherd.2017.06.001

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
System Optimization of Turbo-expander Process for Natural

Gas Liquid Recovery

Yajun Lia,*, Feiyu Xua, Canteng Gonga

a
Key Lab of Heat Transfer Enhancement and Energy Conservation of the Ministry of Education, South

China University of Technology, Guangzhou, China

*Corresponding author. Tel.: 86-020-87112044.

Address: SHAW Engineering Building, South China University of Technology, Wushan RD., Tianhe

District, Guangzhou, Guangdong Province, P.R.China

Postal code: 510640

E-mail address: [email protected]

Highlight
 A certain NGL recovery system is optimized.
 An optimization model is developed considering all the influential factors.
 A proper optimization strategy is put forward based on SQP algorithm.
 Optimal parameters and maximum overall profit are obtained by Aspen and
Excel VBA.

Abstract: Turbo-expander process is a complex technology to recover the light

hydrocarbon in natural gas where energy is highly integrated and some important
parameters such as demethanizer pressure and precooling temperature interact with
each other, which also codetermine both the recovery and energy consumption of the
system. Based on an NGL recovery engineering of 60×104 Nm3/h pipeline gas, this
paper focuses on the optimization for NGL recovery system. The overall profits of
whole process are chosen as objective function and optimization model is developed.
Considering all the influential factors and restrictions such as CO2 freezing, a proper
optimization strategy is put forward which is based on SQP (Sequential Quadratic
Programming) along with constraint boundary research. The optimization model is
solved by Aspen Plus combined with Excel VBA to determine the optimal value of
1
parameters which satisfy both the constraints of the process and the maximum
economic benefits of the system. After optimization, it turns out that the overall
profits have been improved by 13.5%. The optimization model and the corresponding
solution algorithm can provide theoretical guidance for process design and production
practice of the turbo-expander process.

Keywords: NGL recovery; turbo-expander process; CO2 freezing; optimization

modeling; SQP; algorithm

NOMENCLATURE

E ——the benefits of the system, $/hour

B ——the profits of product and by-product
CM——material cost, $/hour
CU——utilities cost including water, electric and steam, $/hour
E1——the profits per hour, $/h
PNGL——the price of NGL
FNGL——the flow of NGL (not including CO2)
PH——the price of unit calorific value, $/MJ
HBOT——the calorific value of bottom NGL, MJ/Nm3
FBOT——the flow of NGL, Nm3/h
WE——the electricity consumed by the whole process, kW

PE——the unit price of electricity, $/ (kW h)

FW——the circulating water required, t/h

PW——the price of circulating water per ton, $/t

Tpre——precooling temperature,

TJT——throttling pressure, bar

Introduction
Ethylene yield is an important index that measures the development level of
petrochemical industry of a country. In China, the equivalent consumption of ethylene
shows an increasing tendency over the years and is twice as much as its production
2
capacity. Due to the lack of resource, the supply gap of ethylene is rather large, almost
half of the ethylene depends on import1. Natural gas is a mixture containing not only
hydrocarbon gases such as C1~5 but also non-hydrocarbon gases like N2 and CO2.
There exist abundant high value hydrocarbon resources in natural gas such as ethane
and propane, which is the cost-optimal cracking feedstock and can be cracked to
prepare ethylene. In this way, the highest recovery of ethylene can be achieved for
such obvious advantages compared to the other raw material. So, the light
hydrocarbon in natural gas must be recycled and reused to relieve the dilemma in
China that ethylene is in short supply and has a large external dependency.
Light hydrocarbon separation is a process in which CH4 is separated from the
heavier component such as C2H6, C3H8, C4H10 et al. in demethanizer column.
Cryogenic condition can be achieved by the expansion process of the raw gas of high
pressure through turbo-expander, which is called turbo-expander process. It is the
mainstream technology to recover light hydrocarbon for the advantage that the
process is comparatively simple with less device, higher refrigeration efficiency and
recovery2. C2 in natural gas are the main objective and the recovery of C2, C3 can
exceed 90% when the separation temperature reaches -90℃ or even lower than
-100℃.
On the circumstance of not adding external refrigeration cycle, the cold energy for
turbo-expander process are all from pressure energy of feed gas. In other words, the
energy source of light hydrocarbon recovery is completely from energy integration in
internal system. There are many important processing parameters, such as
demethanizer pressure, precooling temperature, throttling pressure and so on, which
codetermine the energy consumption and recovery of the system, also influencing the
operation of downstream process at the same time. However, the high energy
integration of the process, varieties of process parameters and complicated restrictions
all enhance the optimization difficulties of turbo-expander process. It is getting
increasingly difficult to determine the optimum value of these operation parameters in
designing light hydrocarbon recovery process or practical production, considering the
mutual influence and restrictions between the parameters along with the high
integration and unique restriction of CO2.
If a group of process variables and a series of restrictions that affect both the
recovery and energy consumption can be considered and analyzed comprehensively,
the optimization model on the whole system can be developed and the corresponding
3
solving strategy can be proposed. It helps to have a thorough understanding of the
inherent law of the process, also providing reference and guidance for process design
and production practice.
At present, majority of researches focus on the optimization for the recovery,
energy consumption and economic benefits. There are also researches concerning
about the optimization and improvement on all kinds of turbo-expander process routes,
all of which are meant to achieve energy integration to the largest extent and recover
the light hydrocarbon as much as possible at the lowest energy cost.
Fernandez L.3 has analyzed and discussed the optimization process for light
hydrocarbon recovery process, where the influence that CO2 freezing exerts on the
process has been considered; Mehrpooya M. et al.4 and Pan D.T. et al.5 both made a
simulation for a specific light hydrocarbon recovery device and the optimal operation
parameter is finally determined by optimization, achieving the maximum economic
benefits; Ali Vatani et al.6 made a simulation of ethane recovery plant to separate
C1–C2 and C3+ at the minimum utility cost and optimal operating parameters were
determined by optimization where Shuffled frog leaping algorithm is used; Behnam
Tirandazi et al.7 put forward an exergy analysis of the multistage refrigeration cycle
used for Ethane and heavier hydrocarbons (C2+) recovery plant. An ethane recovery
unit with its refrigeration cycle has been simulated and refrigeration calculations have
been carried out through the effect and sensitivity analysis of pressure drop and
temperature on the exergy efficiency and coefficient of performance; Chebbi R. et al. 8
have made an optimization for C2H6 recovery in turbo-expander process, choosing
demethanizer pressure and feed composition as design variables; Mehrpooya M.9 has
mentioned that it is difficult to research the behavior of the complex system by
changing the operation parameters at the same time. The net profits of the whole
process is chosen as the objective function for optimization; Mehrpooya M. et al.10
also investigate the step-by-step depressurizing of feed pressure in all stages of device
and finally find an optimal pressure distribution, achieving the highest benefits;
Pitman R.N.11 et al introduced the process of cold residue recycle(CRR) and recycle
vapor-split(RSV), both of which can get a higher recovery of light hydrocarbon
compared with the gas subcooled process and overhead gas recycling process. Also,
the component of product can be adjusted flexibly according to the market demand;
Mehrpooya M. et al.12 studied a process configuration for recovery of hydrocarbon
liquids from natural gas in which the required refrigeration is obtained by a
4
self-refrigeration system. Also, it is proposed that the process can work efficiently
with different feeds; Mehrpooya M. et al.13 introduced a new parameter which based
on the recovered refrigeration against the required refrigeration for evaluating the
Integration Degree (IDG) of cryogenic liquid recovery process. The exergy concept
was used in the calculation and the results showed that with increasing IDG the
process performance and integrity of the process increase; Hudson H.M.14 pointed out
that the traditional expansion refrigeration process recycled part of CO2 while
recycling ethane, leading to setting up a decarburization device subsequently. An
improved process has been studied and put forward, which can reduce the amount of
recycling CO2 on premise of not affecting the recovery of ethane. Some researches
started from the structure of recovery process and made designing and optimization
synthesis for the process15-17; Konukman A.E.S.18 pointed out that the optimization for
integrated heat exchange network and topological structure of process along with
parameters must be conducted at the same time; Luyben W.L.19-20 studied the energy
integration of the light hydrocarbon and demethanizer control. Besides, how the
change of the feed composition influence the light hydrocarbon recovery process and
21
designing of demethanizer have been investigated. Huang S Y. studied the
adaptability of CO2 freezing in different ethane recovery processes and investigated
the heat integration of system.
However, there are kinds of shortcomings in these researches. For instance, some
researches choose the optimal parameters just by sensitivity analysis for one or
several variables; some neglect certain important constraints, such as CO2 freezing
and minimum temperature approach of heat exchanger; some just present the idea of
optimization without illustrating the optimization process in detail. Besides, there is
no report on the overall integration and optimization for the light hydrocarbon
separation system whose product is used as cracking feedstock to prepare ethylene.
In this paper, based on the background that the light hydrocarbon in a natural gas
pipeline are recovered to be used as ethylene cracking raw material, the process
optimization for light hydrocarbon system is discussed where cold energy is
self-produced and the energy is highly integrated. With the target of maximizing the
benefits, the corresponding system optimization strategy and model are established
according to the system feature of its own. Based on the simulation of light
hydrocarbon recover process, the process parameters firstly need to be analyzed and
the restrictions in the process should be considered, such as the limitation of CO2
5
freezing, then it is a necessity to organize an effective algorithm to solve the objective
function and parameters of the process with the help of Aspen Plus、Excel VBA,
aiming to determine the optimum operating point at which the whole system reaches
the maximum benefits.
1. Modeling for light hydrocarbon recovery system
1.1 Process introduction
A certain unit for light hydrocarbon recovery is set as research background in this
paper. The cryogenic separation system in which the refrigeration is achieved by
turbine expansion is a complicated system where the energy are highly integrated,
which is shown in Fig.1.
The feed is the natural gas in a pipeline of 9.0 MPa and 30℃and the flow is 60×
104 Nm3/h. The composition is shown in table 1.

After preprocessing like desulfurization, demercuration and dehydration, the feed is

purified, getting rid of partial impurity and water. Then the purified gas enters the cold
box and is firstly precooled in cold box, after which it goes into evaporator in
demethanizer and works as a heat source2. After cooled in the evaporator, the gas goes
into the cold box and is cooled again, then goes through throttling and
depressurization, subsequently enters the flash for preliminary separation. The liquid
phase from flash enters demethanizer after throttling and depressurization again; the
gas leaving the separator is split, with one fraction entering the turbo-expander for
refrigeration after which it enters the demethanizer, and the other fraction subcooled
by the cold box before it is flashed in a valve and fed to the tower as a reflux.
To achieve energy integration to the largest extend and recover the cold energy in
internal system, a side reboiler can be set at middle-lower part of demethanizer, from
which a stream can be drawn out and enters cold box for reheating, then the gas goes
back to the demethanizer. It helps to reduce equipment investment and ensure light
hydrocarbon recovery on premise of not adding external cold source. The temperature

of the overhead product where CH4 takes up 98% is about-90 -100 , which means

that it is the most essential cold energy in the light hydrocarbon separation process.
The overhead product is heated up in the cold box and goes into the pressurizing side
of turbo-expander, getting preliminary pressurized. After been further pressurized in
the compressor, it is sent back to natural gas pipeline network. As for the natural gas
liquids in bottom demethanizer, it is sent to product tank field for preparing ethylene
after further getting rid of CO2 in decarburization device.

6
 Over 90% dry gas is transported to backbone network after light hydrocarbon
separation. The cold energy required in the process is all from the expansion
refrigeration of high-pressure natural gas. The larger expansion ratio is, the more cold
energy can be recovered and the higher light hydrocarbon recovery will be; however,
larger expansion ratio leads to more energy consumption that is required for
compressing the dry gas to pipeline network, which exerts a great influence on the
economy of the whole process. On the other hand, demethanizer pressure determines
the state function of bottom product (temperature, pressure and the content of CO2),
which means that processing property and benefits of downstream decarbonization
process are closely related to demethanizer pressure. However, energy consumption
and recovery interact with each other. It is an important trade-off decision that have a
great influence on benefits of the process. Thus, the demethanizer pressure is the
control parameter that determine the recovery and energy consumption of the whole
process.
In addition to meeting the request of recovery and energy consumption, safety is
the most important factor that must not be neglected in chemical production process.
On premise of ensuring 90% C2 recovery, it is significant to achieve energy-saving
and cost-reducing for the process based on the principle of avoiding CO2 freezing in
pipeline or key device and keeping right temperature approach.
1.2 Optimization objective
Economic benefits of the whole process are set as optimal objective in this paper.
The optimization function comprises the following parts: the profits of product and
by-product, material cost, utility cost, equipment investment cost and so on. The
target in this paper is to seek the maximum of the established objective function by
optimizing the manipulated variables. To make it simplified, equipment investment
cost is not taken into consideration and the objective function is shown as eq.1.
E = B - CM - CU (1)
The function of profits in the whole turbo-expander process is shown in eq.2.
E1 = PNGL × FNGL - PH × HBOT × FBOT - WE × PE - FW × PW (2)
After light hydrocarbon separation, the calorific value of natural gas will
correspondingly decrease and this dry gas will be sold out as by-product. To
compensate for the calorific value loss, the price of the dry gas that goes back to the
networks needs to be adjusted according to unit calorific value price which equals the
price of feed gas divided by feed gas calorific value. In another word, the price of unit
7
calorific value multiplied by its calorific value equals the price of sale-out dry gas.
Thus, price of feed gas minus the price of sale-out dry gas equals the raw material
cost. The utilities in cryogenic separation process are mainly comprised of electricity
and water consumption, which is used for compression and interstage cooling. The
quantity of the circulating water FW is determined by the temperature of returned
recirculating cooling water and the required condensation load.
Here are the restrictions to the objective function:
(1) C2H6 recovery. The process mainly aims to recover ethane, providing sufficient
raw material for subsequent cracking plant at the same time. It is regulated that
recovery of ethane must be higher than 85%.

  C2 H 6   0.85 (3)
(2) CO2 freezing. The temperature in the light hydrocarbon separation process is often

lowered to -90 -100℃. On this occasion, the CO2 in feed gas may convert to dry ice

at cryogenic equipment, leading to CO2 freezing. The top of demethanizer is most

likely to have CO2 freezing since its temperature is the lowest. CO2 freezing will
increase the differential pressure of demethanizer or even force the turbo-expander to
stop working, which is a big threat to the safe, reliable operation and continuous,
stable production for the whole process. In the process, it is a must to guarantee that

there is at least 3 5℃ safe margin away from CO2 freezing point in both vapor and

liquid22. Objective function stipulates that the stage temperature is at least 3℃ higher
than CO2 freezing temperature on every stage where CO2 freezing is likely to take
place.

Min Tsafe  3℃ (4)

(3) Temperature approach. The cold box is an important device in the turbo-expander
process, where the energy is highly integrated. The plate aluminum is usually chosen
as the material and the minimum temperature approach is usually 2 3℃2. In this
article, for the cold box, the temperature approach at cold side is higher than 2℃and
5℃ at the hot side.

Tcold  2℃
 (5)
Thot  5℃

8
(4) Methane content. In the process of ethylene cracking, the content of CH4 is
commonly used to represent the extent of cracking and CH4 is the impurity in the
cracked gas, which will increase the handling load of subsequent devices23. However,
if more CH4 is separated out in the recovery process, the amount of dry gas that goes
back to pipeline network will increase, which will enhance the profits of the process.
In this article, the mole fraction of CH4 in NGL at bottom demethanizer is restricted to
0.5%.

xCH4   0.005 (6)

The corresponding economic parameters for the objective function are shown in
table 2.
2. Optimization for turbo-expander process
To investigate the character of the process, it is necessary to analyze the main
variables and determine the proper value range for every variable and then seek the
appropriate optimization method. By way of effective algorithm, the optimization
results of the objective function can be get.
2.1 Variable analysis
In the optimization of light hydrocarbon recovery, the key variables chosen in this
paper includes: demethanizer pressure, precooling temperature, throttling pressure,
split ratio, bottom flow.
2.1.1 Demethanizer pressure

Demethanizer pressure in industry is usually kept at 7 32 bar24. The lower the

demethanizer pressure is, the larger relative volatility difference between CH4 and
other hydrocarbon will be, which is better for separation. What’s more, the cold
energy required in the demethanizer system are from the pressure energy of itself. The
lower the separation pressure is, the more the pressure energy can be utilized, which
means more cold energy could be supplied to the recovery system and higher light
hydrocarbon recovery can be achieved. However, on the other side, the recompression
energy consumption for top dry gas will increase with the decrease of demethanizer
pressure and the increased cost of compression energy consumption will counteract
the profits by the increase of C2H6 recovery. At the pressure of 20 bar or even lower,
the recompression work consumption will be too much to ensure the economy of the

process14. Thus, the variation range of demethanizer pressure is chosen as 20 32 bar.

9
2.1.2 Precooling temperature
The variation range of the precooling temperature for feed gas is decided by the
composition and state. In the process, the heat duty for reboiler is supplied by feed gas.
Within the allowed temperature approach range, so long as the precooling temperature
is kept low, not only the feed can get more cold energy for first vapor-liquid
separation, but also the heat duty of rebolier can be guaranteed.
It can be found by calculation that the pressure and temperature at critical point are
respectively 64.5 bar and -61.15℃. The feed gas is in a state of 30℃and 90 bar,
which means that it has already been in the supercritical state. Since the feed will
enter the flash for separation after precooling and throttling to ensure that it enters the
gas-liquid two-phase region, the precooling temperature must be low enough. Judging
from the P-H diagram of feed gas, when the gas goes through isenthalpic throttling,
gas-liquid two-phase flow may form so long as the temperature is lower than 10.7℃,
which is shown in Fig.2. Moreover, to avoid critical region, in this paper, the lower
limit of precooling temperature for feed gas is set at -60.0℃.
Thus, the value range of precooling temperature is

-60.0  Tpre  10.7 (7)

2.1.3 Throttling pressure

The variation range of throttling pressure is closely related to the composition and
state of feed gas together with the precooling temperature. According to Fig.2,
throttling pressure must be low enough if the feed gas were to enter the vapor-liquid
two- phase region which is below the dew point line. On the basis of the relationship
between enthalpy and temperature of feed gas, together with dew point line, the curve
that represents the relationship between upper limit pressure and precooling
temperature can be drawn out, which is shown as curve A in Fig.3.
However, it is not enough to ensure that throttling pressure is lower than the
pressure at dew point. In the recovery process, the gas from top flash is divided into
two parts, which respectively enters the cold box and turbo-expander. If the pressure
is extraordinary high, the gas may go through critical region, where gas-liquid phase
mutation would occur, which is a big threat to the normal operation of the recovery
device. Thus, the throttling pressure must be lower than the critical pressure of the gas.
Under different precooling temperature, the sensitivity analysis on the critical point of
the gas from top flash has been made, which is shown in Fig.4.
10
 As is shown in Fig.4, the critical pressure of gas from top flash decreases with the
decrease of precooling temperature; when the precooling temperature is settled, the
critical pressure of gas from top flash decreases with the decrease of throttling
pressure. So long as the operation point is kept at the left side of isobar, it can be
guaranteed that the pressure of gas from flash overhead is lower than its critical
pressure. Then the curve representing the relationship between precooling temperature
and upper limit of throttling pressure can be drawn out, which is shown as curve B in
Fig4. Curve A and B codetermines the limit of throttling pressure.

Since the range for precooling temperature is -60.0  Tpre  10.7 , the relative upper

limit of throttling pressure is described as a section function: when the range

is -16.8  Tpre  10.7 , the curve A is chosen as the upper limit; when the range

is -60.0  Tpre  -16.8 , the curve B is selected as the upper limit.

In addition, since the demethanizer pressure is usually 20 32 bar and the expansion

ratio for units is 2 425, to ensure that turbo-expender possesses sufficient expansion

ratio, the lower limit for throttling pressure should be restricted, which is set as 40 bar.
Thus the precooling temperature is confined to be no higher than 1.14℃.

Therefore, combining all the analysis and Fig.2 4, the variation range of throttling

pressure for feed gas can be determined as follows:

40  PJT <-0.0206Tpre
2
-1.6187Tpre  41.874, 1.14  Tpre  -16.8 (8)

40  PJT <-0.0015Tpre
2
+0.1252Tpre  65.72, -16.8 Tpre  -60.0 (9)

2.1.4 Split ratio

The reflux of demethanizer column is determined by the split ratio of gas from top
flash. The split ratio is denoted as mole ratio that the turbine expander inlet gas flow
accounts for the gas flow from flash overhead. Since the subcooling for the reflux is
achieved by recovering the cold energy of dry gas from overhead, there is a certain
restriction for split ratio considering the cold energy that the dry gas can supply. The
smaller the split ratio is, the less gas will enters turbo-expander. Correspondingly, the
reflux will increase and the separation effect will be better. However, it will increase

11
the risk of CO2 freezing. In this paper, the variation range of split ratio is set as

0.1 0.9.

2.1.5 Bottom flow

In actual operation, when the other condition is settled, bottom flow is commonly
decided by the temperature at bottom demethanizer, which is an important parameter
that helps to ensure the product quality at the same time. When the concentration of
CH4 at NGL recycled exceeds 0.5%, the excess CH4 can be distilled off to meet the
product standard by raising the temperature at bottom. In this paper, bottom flow is
chosen as the control variable that regulates the quality of product at bottom column
and it is adjusted in each calculation to meet the request that the concentration of CH4
reaches 0.5%.
On the other hand, the change of bottom flow will influence CO2 freezing
temperature margin on stage to some extent. When the bottom flow decreases, more
CH4 will be distilled out and enter the updraft in the column. This part of CH4 can
strip more CO2 out of the fluid on stage. It is known that CO2 solubility in vapor is
generally lower than in liquid, thus the safe temperature margin will drop and vice
versa.
2.2 Optimization method
The optimization model in this paper belongs to nonlinear optimization with
restrictions. SQP (Sequential Quadratic Programming) is an algorithm where
nonlinear programming is replaced by quadratic programming at an approximate
solution to obtain a new approximate solution, after which a new quadratic
programming can be reconstructed until the optimal solution is found. It is getting
increasingly widely applied in the field of chemical industry and it is also applied for
the optimization of the manipulated variable in light hydrocarbon recovery process.
There are two main difficulties for the optimization of the objective function
according to the analysis above:
Firstly, according to the analysis above, the upper limit of throttling pressure is
determined by precooling temperature. Essentially, the corresponding correlation
between the value range of throttling pressure and precooling temperature can be
regarded as inequality constraints. Therefore，the value range for some variable is not
unfixed. For SQP, the value range of the variable cannot be adjusted momentarily as

12
required in the optimization process, which may lead to the problem that the
technological parameter exceeds the restrictions.
Secondly, if the stage temperature is lower than the temperature at CO2 triple point
(-56.6℃), CO2 may freeze in both the vapor and liquid phase on stage. Nevertheless,
when stage temperature is higher than -56.6℃, there is no solution to the equilibrium
model that is used to predict CO2 freezing temperature. The limitation of CO2 freezing
needs to be taken into consideration, leading to numerous inequality constraints,
which may influence the stability of calculation and even leads to no results.
Aimed at the feature of the cryogenic separation process and the main difficulties of
objective function, the process of algorithm is proposed which is shown in Fig.5.
The algorithm is achieved in Excel VBA environment. The variables that need to be
considered in the optimization include demethanizer pressure, precooling temperature,
throttling pressure, split ratio. As the analysis above, demethanizer pressure has a
greatly influence on cryogenic separation process and it determines the
light hydrocarbon recovery and energy consumption of the system. Thus,
demethanizer pressure is a critical parameter on process benefits. The precooling
temperature and demethanizer pressure are separated from SQP while parameters like
C2H6 recovery, CO2 freezing margin, temperature approach of cold box and bottom
CH4 content are all set as variables of SQP problem. Initial value of demethanizer
pressure and precooling temperature are set at the upper limit of value range. Then,
they are gradually lowered and finally go through all the feasible range. Every time
the precooling temperature is adjusted, the value range of throttling pressure can be
determined and corresponding SQP problem is constructed to solve the first main
difficulty.
For complex nonlinear programming with constraints, the result can be calculated
taking no account of constraints. If the result mismatches the constraints, the optimal
value of target may locate on the boundary of part constraints17,26 As long as searching
in the main constraint boundary, the optimal point can be found which meets other
conditions at the same time and this point is the solution of objective function.
Therefore, the SQP problem is calculated by Aspen Plus firstly taking no account of
constraints, and then the result should be judged to ensure CO2 freezing safe margin.
If the solution cannot meet the requirement, the minimum safety margin can be found
by further searching in the boundary constraints, changing variable such as split ratio
and bottom flow rate to meet 3℃ requirement. When other constraints like recovery
13
meet the requirement, the point found by searching is the optimal point and record the
solution to solve the second main difficulty.
For every calculation, Excel extracts results from Aspen plus and then makes the
corresponding judgment and adjustment, after which the adjusted results are sent back
to Aspen plus for recalculation. A series of SQP problem can be solved under different
demethanizer pressure and precooling temperature. Finally the optimal curve can be
drawn out by comparing the optimum point at certain pressures.
The main algorithm of VBA and logical judgment of Excel function, along with the
properties calculation by Aspen Plus, process simulation and SQP optimization block,
together guarantee the smooth realization of the algorithm. Compared with single
SQP algorithm, the established optimization algorithm is stable and effective, which
can solve the problem that single SQP cannot solve although it inevitably inquires
more time for calculation.
PR-BM equation of state is used for the properties calculation in the property
model. Based on the phase equilibrium criterion that fugacity in each phase equals
and using the vapor-solid and liquid-solid equilibrium model of CO2 in natural gas,
the CO2 freezing temperature in demethanizer can be predicted.
3. Results and analysis of optimization
3.1 Optimization results
The whole process is optimized according to algorithm in Fig.5 and the results is
shown in Fig.6.
As is shown in Fig.6, with the decrease of demethanizer pressure, the profits of
process firstly increases and then decreases. At the pressure of 28.9 bar, the objective
function established in this paper reaches its maximum, which is 13066.6 $/h. When
the pressure is lowered to 28.9 bar, the profits start to decrease due to the fact that
lower demethanizer pressure will increase the energy consumption of compressor
although higher C2H6 recovery can be achieved and the by-product dry gas will
decrease. When the increased profits by extra light hydrocarbon cannot compensate
for the cost of rising energy consumption for compressor, the profits will start to
decrease. When the pressure is lowered to 27 bar, considering the influence of CO2
freezing, the profits start to drop comparatively sharply. When the pressure is lower
than 25 bar, 85% recovery of C2H6 and CO2 freezing safe margin cannot be satisfied
at the same time.

14
 The key parameters of process equipment and optimization results are shown in
Table 3 and Table 4.

From the optimization results, it can also be found that the profits is not the best at
the condition where C2H6 recovery is the highest. When the operating pressure is
around 27 bar, the recovery can be achieved as 0.938 on premise of ensuring CO 2
freezing margin. However, the maximum profit is realized at the pressure of 28.9 bar.
CO2 freezing is an essential problem in actual light hydrocarbon recovery process
and the freezing margin of demethanizer is the smallest. In other words, the CO2
freezing risk of demethanizer is the largest. Analysis of CO2 freezing temperature for
demethanizer at the optimum condition is conducted in this paper27.
The simulation for CO2 freezing in demethanizer is also conducted by Aspen plus.
According to the CO2 concentration and temperature on each stage, the operating line
of demethanizer can be drawn out. Also, the freezing line of vapor and liquid phase
can be drawn by data of CO2 concentration and CO2 freezing temperature in both
vapor and liquid phase. Then whether CO2 will freeze in demethanizer can be judged
by whether the freezing line intersects the operating line. The CO2 freezing condition
in vapor and liquid phase are respectively shown in Fig.7 and 8.
In Fig.7, the abscissa represents the temperature on each stage; the ordinate on the
left side represents CO2 mole fraction on each stage, and the other on the right side
represents stages of the column.
It is easy to find that the horizontal distance between A and B is the smallest and

the minimum freezing margin is 3.68℃ while corresponding CO2 concentration on

point C is 5.34%. Then a vertical line (isothermal) can be drawn out and intersect the
temperature profile line at point D. Afterwards, a horizontal line is drawn from point
D to intersect the right y-axis at point E. That is the 15th stage where CO2 freezing is

most likely to take place and the stage temperature is -76.22 .The analysis of CO2

freezing in liquid is the same. Finally it comes to the conclusion that the minimum
freezing temperature margin in liquid is 6.74 °C in the 15th stage and the
corresponding CO2 mole fraction is 16.50%.

15
 It can be found that the CO2 freezing margin in liquid is higher than in vapor on
each stage, which means that CO2 is more likely to freeze in vapor. According to the
analysis, CO2 freezing margin in both vapor and liquid are higher than 3℃, which
indicates that not only CO2 does not freeze in demethanizer, but the safe temperature
margin is enough. In conclusion, the operating condition determined by the above
optimization is safe and stable with good elasticity.
4. Conclusion
4.1 A certain system to recover light hydrocarbon is chosen as research object in this
paper. The overall benefits are the objective function and the optimization model is
established. After considering all the influential factors and variety of restrictions,
optimization algorithm based on SQP is proposed to solve the model, aiming to obtain
the optimal operating parameters and maximize the overall profits of the whole
process.
4.2. The optimization algorithm put forward in this paper which is based on the SQP
and combined with constraint boundary search is achieved by the combination of
Excel VBA and Aspen Plus. Compared with the sole SQP, the algorithm proposed
inevitably spend more time on calculation. However, on the other hand, the
optimization algorithm is stable and efficient, which can solve the problem that single
SQP cannot solve. It is flexible enough and can adjust the value range of variable in
the optimization process to realize optimization calculation smoothly. Combined with
Excel, the data form Aspen plus can be exported and displayed in the program
interface, such as minimum CO2 freezing margin on stage, temperature approach,
product flow, composition, recovery and objective function, which is convenient for
comparison and make the optimization calculation process simplified to largest
extent.
4.3. With the help of optimization model established above and the corresponding
solving strategy, the optimization for the turbo-expander process where light
hydrocarbon is recovered from natural gas of 60×104Nm³/h in pipeline of high
pressure is implemented. It has been found out that at the operating pressure at 28.9
bar, the C2H6 recovery can reach 92.33% and the profits of product and by–product,
the material cost, and utilities cost is 41051.93 $/h, 25247.06 $/h, 2738.27 $/h
respectively. Therefore, the maximum profits of the unit is 13066.6 $/h on premise of
meeting all the restrictions, which testify the feasibility and advantage of the proposed
algorithm. And the ratio of benefits, material cost and utilities cost respectively are
16
3.14, -1.93 and -0.21. It can also provide theoretical guidance for optimization design
and production of turbo-expander process.

17
Reference
(1) Li, Y.J.; Luo, H. Integration of light hydrocarbons cryogenic separation process in
refinery based on LNG cold energy utilization. Chem. Eng. Res. Des. 2015, 93,
632−639.
(2) Yang, W.J; Li, Y; Li, Y.J. Design of light hydrocarbon recovery process from
natural gas and the optimization of operation parameters. Chemical Industry and
Engineering Progress, 2015, 34(10): 3589-3594.
(3) Fernandez, L.; Bandoni, J.A.; Eliceche, A.M. et al. Optimization of ethane
extraction plants from natural gas containing carbon dioxide. Gas Separation &
Purification, 1991, 5: 229-234.
(4) Mehrpooya, M.; Gharagheizi, F.; Vatani, A. An Optimization of Capital and
Operating Alternatives in a NGL Recovery Unit. Chemical Engineering &
Technology, 2006, 29(12): 1469-1480.
(5) Pan, D.T.; Huang, M.Z.; Zhang, X.J.A. static simulation model of overall process
for the NGL recovery unit. Journal of Shandong University (Engineering
Science), 2012, 42(3): 57-62.
(6) Vatani, A., M. Mehrpooya, and H. Pakravesh, Modification of an industrial ethane
recovery plant using mixed integer optimization and shuffled frog leaping
algorithm. Arabian Journal for Science and Engineering, 2013. 38(2): p.
439-455.
(7) Tirandazi, B., et al., Exergy analysis of C 2+ recovery plants refrigeration cycles.
Chemical engineering research and design, 2011. 89(6): p. 676-689.
(8) Chebbi, R.; Al-Amoodi, N.S.; Abdel, J.N.M. et al. Optimum ethane recovery in
conventional turbo expander process. Chemical Engineering Research and
Design, 2010, 88(5-6): 779-787.
(9) Mehrpooya, M.; Vatani, A.; Mousavian, S.M.A. Optimum design of integrated
liquid recovery plants by variable population size genetic algorithm. The
Canadian Journal of Chemical Engineering, 2010, 88(6): 1054-1064.
(10) Mehrpooya, M.; Vatani, A.; Moosavian, S.M.A. Optimum pressure distribution in
design of cryogenic NGL recovery processes. Iranian Journal of Chemistry &
Chemical Engineering, 2012, 31(3): 97-109.
(11) Pitman, R.N.; Hudson, H.M.; Wilkinson, J.D. et al. Next generation processes for
NGL/LPG recovery. Dallas: Gas Processors Association, 1998.
(12) Mehrpooya, M., A. Vatani, and S.A. Mousavian, Introducing a novel integrated
18
 NGL recovery process configuration (with a self-refrigeration system
(open-closed cycle)) with minimum energy requirement. Chemical Engineering
and Processing: Process Intensification, 2010. 49(4): p. 376-388.
(13) Mehrpooya, M., A. Vatani, and S.A. Moosavian, Introducing a new parameter for
evaluating the degree of integration in cryogenic liquid recovery processes.
Chemical Engineering and Processing: Process Intensification, 2011. 50(9): p.
916-930.
(14) Hudson, H.M.; Wilkinson, J.D.; Lynch, J.T. et al. Reducing treating requirements
for cryogenic NGL recovery plants. San Antonio: Gas Processors Association,
2001.
(15) Diaz, S.; Serrani, A.; Bandoni, A. et al. A study on the capital and operating
alternatives in an ethane extraction plant. Computers & Chemical Engineering,
1996, 20: S1499–S1504.
(16) Nawaz, M.; Jobson, M. Synthesis and optimization of demethanizer flowsheets
for low temperature separation processes. Distillation Absorption, 2010: 79-84.
(17) Kherbeck, L.; Chebbi, R. Optimizing ethane recovery in turbo expander
processes. Journal of Industrial and Engineering Chemistry, 2014, 21: 292-297.
(18) Konukman, A.E.S.; Akman, U. Flexibility and operability analysis of a
HEN-integrated natural gas expander plant. Chemical Engineering Science, 2005,
60(24): 7057-7074.
(19) Luyben, W.L. NGL demethanizer control. Industrial & Engineering Chemistry
Research, 2013, 52(33): 11626-11638.
(20) Luyben, W.L. Effect of natural gas composition on the design of natural gas
liquid demethanizers. Industrial & Engineering Chemistry Research, 2013,
52(19): 6513-6516.
(21) Huang, S.Y. Study of ethane recovery technology containing CO2. Chengdu:
Southwest Petroleum University. 2015.
(22) Jiang, H.; He, Y. X.; Zhu, C. A. Forecast model for solid CO2 formation
conditions in a CH4-CO2 system. Natural Gas Industry. 2011, 31 (9), 112−114.
(23) He, X.O. Hydrocarbon steam cracking theory and industrial practice. Ethylene
Industry. 2008, 20(3): 49-55.
(24) Gao, C.Y. Analysis on Domestic Ethylene Market in 2013 and Forecast in 2014.
Coal Processing& Comprehensive Utilization, 2014, (2): 47-50.
(25) Zhao, X. The Significance and Technological Progress of Light Hydrocarbon
19
 Recovery of Natural Gas. Guangzhou Chemical Industry, 2013, 41(9):32-33.
(26) Yao, P.J. Process system engineering. Shanghai：East China University of Science
and Technology press, 2009.
(27) Gong, C.T; Li, Y; Li, Y.J. Application of Highly Accurate Phase-Equilibrium
Models for CO2 Freezing Prediction of Natural Gas System. Industrial
Engineering Chemical Research, 2016, 55, 5780−5787.

20
 Compressor
Methane

Throttle

Expansion
machine

Compressor

Cold box
Flash drum
Feed
Demethanizer

Buffer tank
Purification processes Reboiler

Bottom NGL to decarburization device

Fig.1 The turbo-expander process for light hydrocarbon separation

21
 Fig. 2 Pressure-enthalpy diagram of raw natural gas

Fig.3 Relationship of throttling pressure upper limit vs. precooling temperature

22
Fig.4 Relationship of separating drum gas critical pressure vs. throttling pressure

23
 Start

Column pressure
P=32bar

Precooling temperature
Tpre=MaxTpre

Aspen Plus SQP

Split ratio and bottom flow

are chosen as variables to N Min ΔTsafe
search the optimum point Tpre=Tpre-ΔTpre
on the restriction ≥3℃？
boundary

Recovery:φ≥0.85？
CH4 mole fraction=0.005
Y Record
et.al

N
N
Tpre≤-60℃？
Y

Draw the
optimal curve

N
P≤20bar？ P=P-ΔP

Draw the optimal

curve

End

Fig. 5 Optimization algorithm

24
Fig.6 Optimization results of cryogenic separation section

25
Fig.7 CO2 freezing in vapor in demethanizer

Fig.8 CO2 freezing in liquid in demethanizer

26
 Table 1 Composition of raw natural gas
Components CH4 C2H6 C3H8 iC4H10 nC4H10 iC5H12 nC5H12 N2 CO2

Contents 0.90036 0.04825 0.01306 0.00555 0.00200 0.0023 0.00047 0.00791 0.0201

27
 Table 2 Economic parameters
Project Price

Light hydrocarbon 599.8 $/t

Feed gas 0.3 $/Nm3

Electricity 0.1 $/KW/h

Cooling water 0.3 $/t

Steam 21.0 $/t

Unit calorific value 0.8$/MJ

28
 Table 3 Key parameters of process equipment
Equipment Key parameters
pressure/bar 28.9

pressure drop/bar 0.2

stages 20
Demethanizer
bottom rate/kmol/h 1500

condenser none
reboiler Kettle
reflux (kmol/h) 5367.75
Flash drum adiabatic flash
Cold box Temperature approach at cold side (℃) 2
isentropic efficiency 0.83
Expander
mechanical efficiency 0.98
Compressor isentropic efficiency 0.7

29
 Table 4 Optimization results
Precooling temperature (℃) -52.12
Throttling pressure (bar) 47.38
Split ratio 0.6701
Temperature approach at cold side (℃) 2
Temperature of reflux (℃) -96.2
Flow of reflux (kmol/h) 5367.75
Expansion ratio 1.63
Outlet temperature of turbo-expander (℃) -90.3
Liquid quantity from outlet turbo-expander (w %) 14.83
Brake horsepower of turbo-expander (kW) 1037.80
Minimum CO2 freezing margin (℃) 3.676
CH4 mole fraction in bottom demethanizer 0.005
C2H6 recovery 0.9233
System profits ($/h) 13066.6

30