Mp

YYa

b

c

d

a

ARR2AA

KHAWSPR

1

imgaiaf

1

lc

jf

h0

Computers and Chemical Engineering 86 (2016) 160–170

Contents lists available at ScienceDirect

Computers and Chemical Engineering

j our na l ho me pa g e: www.elsev ier .com/ locate /compchemeng

odeling and simulation of a multistage absorption hydration hybridrocess using equation oriented modeling environment

ou Lia, Xingang Lia,b, Hong Lia,b,c,∗, Luhong Zhanga,b, Feng Xina, Jingyan Lianb,d,onghong Lia,b

School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, ChinaNational Engineering Research Center of Distillation Technology, Tianjin 300072, ChinaCollaborative Innovation Center of Chemical Science and Engineering, Tianjin 300072, ChinaSchool of Chemical Engineering and Technology, Tianjin University of Technology, Tianjin 300072, China

r t i c l e i n f o

rticle history:eceived 10 October 2015eceived in revised form1 December 2015ccepted 25 December 2015vailable online 2 January 2016

a b s t r a c t

Separation of light hydrocarbon mixtures is one of the most important topics in chemical engineeringresearch. With development of theories on hydrate equilibriums and kinetics, researchers are trying toapply hydration based separation technology to industrial applications. It is increasingly important todevelop the corresponding simulation strategies for process design purposes. In this work we use anequation oriented modeling environment, named Aspen Custom Modeler® (ACM®), which enables rapidmodel development and provides powerful simulation solvers. With the help of ACM®, a multistageabsorption hydration hybrid process (AHHP) for refinery dry gas separation is modeled and simulated.

eywords:ydratebsorptionater in oil emulsion

eparationrocess simulationefinery dry gas

Sensitivities of key parameters such as water content and absorbent flow rate, are analyzed. Featuresof the multistage AHHP are discussed. For comparison, based on an industrial data, a butane absorptionprocess is established and simulated. Economic evaluation shows that the multistage AHHP is competitivecompared to current absorption process.

© 2016 Elsevier Ltd. All rights reserved.

. Introduction

Separation of light hydrocarbon mixtures is one of the mostmportant topics in chemical engineering research. Conventional

ethods for separating low boiling point gas mixture include cryo-enic distillation and alkane absorption. New technologies suchs reactive absorption, membrane separation and adsorption arenvestigated (Huang et al., 1999; Padin and Yang, 2000; Adhikarind Fernando, 2006). Recently, hydrate is proposed as a new toolor light hydrocarbon separations (Eslamimanesh et al., 2012).

.1. Literature review on hydrate based separation experiment

Formed at high pressure and low temperature, hydrates are cageike, non-stoichiometric compounds composed of water and hydro-arbons (Sloan, 2003). When a mixture of hydrocarbons forms

∗ Corresponding author at: School of Chemical Engineering and Technology, Tian-in University, Tianjin 300072, China. Tel.: +86 022 27404701;ax: +86 022 27404705.

E-mail address: lihongtju@tju.edu.cn (H. Li).

ttp://dx.doi.org/10.1016/j.compchemeng.2015.12.021098-1354/© 2016 Elsevier Ltd. All rights reserved.

hydrate, some components are more easily captured than others.Large amount of work is performed in hydrate based light hydrocar-bon separation. Ma et al. (2001, 2008) investigated the equilibriumsof CH4 and C2H4. Englezos et al. (1987a, 1987b) studied the kineticsof hydrate formation.

Recent research shows that for industrial purpose, forminghydrate with water as the continuous phase is unviable, becausehydrate formed would float on top of water and thereby impedefurther hydration (H. Liu et al., 2013). Chen and coworkers putforward a new material named water in oil emulsion (w/o emul-sion) and recommended it as the new absorbent (Turner et al.,2009; B. Liu et al., 2013; Ma et al., 2013a). Their observation showshydrate agglomeration can be prevented, since hydrate generatedis dispersed as tiny particles. Meanwhile, due to the large contactarea between oil and water, hydration rate is facilitated (Mu et al.,2014).

Though excellent work is carried out by the scholars, literaturesurvey shows current study has been focused on hydrate formation

kinetics and equilibriums in single reactors. And there is a growingdemand for a process scale modeling and simulation method whichcan provide insights and guidance for the further application ofhydration technology.

Y. Li et al. / Computers and Chemical Engineering 86 (2016) 160–170 161

Nomenclature

SymbolsA Heat exchange area (m2)C Collection of componentsCj Langmuir constants for component jD Vessel diameter (m)f Lj

Fugacity of j in liquid (bar)

f Vj

Fugacity of j in vapor (bar)

Hhy Molar enthalpy of hydrate (GJ/kmol)HL Molar enthalpy of liquid (GJ/kmol)HV Molar enthalpy of vapor (GJ/kmol)HYin Hydrate mole flow rate going in (kmol/h)HYout Hydrate mole leaving (kmol/h)L Vessel length (m)Lin Oil mole flow rate going in (kmol/h)Lout Oil mole flow rate leaving (kmol/h)P Pressure (bar)P0 Operating pressure (bar)Q Heat removed from stage (GJ)T Temperature (◦C)T0 Operating temperature (◦C)Vin Vapor mole flow rate going in (kmol/h)Vout Vapor mole flow rate leaving (kmol/h)Wcmp Working load of compressor (kW)Wj Mole flow rate of j in water (kmol/h)xj

inMole composition of oil liquid going in

xjout Mole composition of oil liquid leaving

yjin

Mole composition of vapor going in

yjout Mole composition of vapor leaving

zjin

Mole composition of hydrate going inj

1

cfwcfap

aptfasm

1

teecaa

Table 1FCC dry gas composition.

Component Composition (mol%)

H2 31.79N2 11.88O2 2.97CO2 2.01CO 0.77CH4 24.65C2H6 12.14C2H4 12.56C3H8 0.15C3H6 1.05

Table 2Operating condition of C4 absorption.

Unit operation Pressure (MPa) Temperature (◦C) Tray number

B-1 4.25 18 15B-2 4.25 18–118 20B-3 2.6 18–120 40HX-1 2.6 In 120, out 40 –

discussed and analyzed. Nevertheless, we found there is still room

zout Mole composition of hydrate leaving

.2. Literature review on hydration simulation

A number of theories are proposed for hydrate equilibriumalculation, among which Van-der-waals equations are the mostamous. However, though accurately derived, solving the van-der-aals equations requires global optimization algorithm that is time

onsuming. As an alternative, Chen–Guo equations have a simplerormulation and can provide a quick solution without losing muchccuracy, and thereby more appealing for process simulation pur-oses (Chen and Guo, 1998; Sloan and Koh, 2007).

Naeiji et al. (2014) calculated the performance of hydrate in sep-rating methane and ethane with the help of a phase map. Thehase map can provide an intuitive understanding in the hydra-ion separation mechanism. Yet, there are two problems aheador this method. First, the phase map cannot provide a quick andccurate solution. Second, it can only calculate the hydrate compo-ition formed from pure water. Therefore, more robust simulationethods are needed.

.3. Current process flow chart for dry gas separation

Ethylene is a fundamental raw material in petroleum indus-ry. Refinery dry gas usually contains a significant amount ofthylene (about 10–20 mol%), which, however, is often flared fornergy (Li and Luo, 2015). Moreover, the development in catalytic

racking, which primarily aimed to improve the light oil gener-tion, results in ethylene content increment in dry gas (Rahimind Karimzadeh, 2011; Sadeghbeigi, 2012). Therefore, it becomes

HX-2 2.6 In 40, out 18 –HX-3 2.6 In 40, out 18 –

increasingly important to recover ethylene from dry gas. A list ofdry gas composition can be observed in Table 1 (Lei Si, 2013).

Admittedly, cryogenic distillation system is known to producehigh quality products. However, at the same time, it also demandshuge initial investment. As one of the most expensive units in cryo-genic process, cold box is usually used to chill gas at cryogenictemperatures. To save investment in cold box, before cryogenic dis-tillation, dry gas should be preprocessed, as is the case of Sheng Lirefinery (Lei Si, 2013).

The C4 absorption flow chart for dry gas pretreatment can beobserved in Fig. 1. The absorption flow chart can be divided intothree parts: absorption (B-1), stripping (B-2) and desorption (B-3).Feed gas is absorbed in the absorption section. In stripping section,part of methane and nitrogen are driven away from the absorbent.Desorption column extracts enriched dry gas from butane. Accord-ing to industrial data (Lei Si, 2013), the operating conditions arelisted in Table 2. Pressure of enriched dry gas is 2.6 MPa. Beforegetting into cryogenic distillation, the enriched dry gas shouldbe compressed up to 3.3 MPa, which is a typical condition formethane/ethylene separation (Salerno et al., 2011).

1.4. Contribution of this work

In this work, we build the models in Aspen Custom Modeler®

(ACM), an integrated development environment for chemical pro-cess modeling (Aspen Technology, 2012). ACM can greatly reducethe amount of labor in coding by providing with the built in flashfunctions and property calculation procedures. The models are thenexported as unit operation blocks to Aspen Plus® for integratedsimulation and optimization.

Naeiji et al. (2014) proposed a multistage separation strategy forhydrate based separation. Following their footprints, we proposeand simulate a multistage absorption hydration hybrid process(AHHP) for dry gas separation applications. Sensitivity of the keyparameters are studied. Then we point out important featuresand evaluate some economic issues. For comparison, a C4 absorp-tion process is also simulated and evaluated. It is found that theproposed AHHP is more economical. Reasons for the savings are

for improvement provided that the refrigeration is reused.To summarize, most previous work on hydration experiment

and simulation is on unit operation scale. In this article we put

162 Y. Li et al. / Computers and Chemical Engineering 86 (2016) 160–170

Feed dry gas

B-1

B-2 B-3

Enriched dry gasMX-1

Off gas

HX- 3

Fig. 1. Absorption process flow chart (B-1: absorber; B-2: stripper; B-3: deso

Table 3Operating condition of AHHP.

Unit operation Pressure (MPa) Temperature (◦C) Tray number

RCT 4.25 0 –Absorber 4.25 0–10 11R-1 3.3 40 –R-2 0.1 40 –

fstcsipSE

2

tihihtrt

TTffliofIT

CMP outlet 3.3 – –HX-1 4.25 In 40, out 0 –HX-2 4.25 In 40, out 0 –

orward a powerful tool, which is especially suitable for largecale process calculation. Then the suggested approach is appliedo model a hydration based dry gas separation system. This arti-le is organized as follows. Section 2 introduces the multistagechematic. Section 3 explains the overall solution strategy includ-ng hydration modeling, mathematical formulation of optimizationroblem as well as the corresponding optimization algorithm.ection 4 shows, explains and discusses the calculation results.ventually, Section 5 states the conclusions.

. Multistage AHHP schematic

The conceptual design of multistage schematic can be found inhe work of H. Liu et al. (2013) and Naeiji et al. (2014). As is shownn Fig. 2a, feed gas and water forms hydrate in the reactor. Afterydration, equilibrium gas phase is treated as off gas, while hydrate

s controlled to dissociate. Gas released from hydrate then formsydrate again. There is a defect in the design: though composi-ion can be updated through pressure and gas/water ratio control,ecovery of desired component is low. This is because gas that failedo enter hydrate is all treated as off-gas.

To tackle with the above issue, we redesigned the flow chart.he new design is shown in Fig. 2b and is named multistage AHHP.he operation condition can be found in Table 3. Refinery dry gas iseed to Stage-n after chilling in HX-2 (n is the stage number). Gasows from Stage-n to Stage-(n-1). The absorbent, which in this case

s the w/o emulsion, goes the opposite way. The final liquid phase

ut of Stage-n is transported to R-1. The cryogenic process thatollows AHHP operates at 3.3 MPa. Hence, R-1 operates at 3.3 MPa.n R-1, the temperature is raised to 40 ◦C for hydrate dissociation.he liquid phase out of R-1 is flashed at atmosphere pressure in

Pump HX- 2HX- 1

rption column; HX-1, 2, 3: heat exchanger 1, 2, 3; MX-1: mixer tank).

R-2. Gas released from R-2 should be recompressed to 3.3 MPa. Theemulsion is reused for continuous operation.

The stages in Fig. 2b stand for stirred tanks in which the phasesare at equilibrium. In unit operation text books, these connectedstages are represented as columns (Seider et al., 2009). However,to avoid hydrate agglomeration, stirring is required in each tank.For this reason, and because refrigeration device is needed for tem-perature control, making inner structure differ a lot from columntrays, in this work the stages cannot be simply represented by acolumn.

3. Solution strategy

3.1. Modeling and simulation environment

Instead of FORTRAN or phase map, in this work we buildthe hydration models in an equation oriented modeling environ-ment, which has user friendly interface and whose code is easyto implement and maintain. There are a number of options, suchas gPROMS® and SystemModeler®. And we choose Aspen CustomModeler® due to the following reasons. First, ACM® has powerfulbuilt in functions, property packages which greatly lighten the bur-den in coding. Second, the models built in ACM® can be exported toAspen Plus® for process integration purposes. Last but not least, forcomplex flow charts, we must tune the parameters through opti-mization, which is perfectly handled by the built-in optimizationmodular in Aspen Plus®.

3.2. Multiphase equilibrium modeling

In the proposed AHHP, each of the stage in Fig. 2b is assumedto be at multiphase equilibrium. We will introduce the calculationprocedure in four aspects including structure determination, massand energy balance equations, equilibrium correlations and phasedetermination.

3.2.1. Hydrate structure determination

As is known, three types of hydrate structures have been rec-

ognized, namely SH, SI and SII (Subramanian et al., 2000; Murshedet al., 2010). An important aspect in hydration modeling is structuredetermination. Structure of hydrate is influenced by a number of

Y. Li et al. / Computers and Chemical Engineering 86 (2016) 160–170 163

Hydrati onP0,T0

Feed gas and water

Hydrate

Off Gas 1a

b

Diss ociation

Product Gas

Water

Hydrati onP1,T1

Off Gas 2

Dissociation

Water

Product GasWater

CMP

Gas

HX-2

R-2

Emulsion reuse

R-1

Enriched dry gas

Hydrate in oil Slurry

Stage-n Stage-(n -1)

Slurry

Off-gasStage-(n-2 ) Stage-1

...

Feed dry gas

char

fpaadK

3

e

T

P

a

f

t

(a

Fig. 2. (a) Conceptual design of hydration process. (b) AHHP flow

actors including pressure, temperature, additives as well as com-osition if gas mixture is concerned. With the existence of propanend since composition of ethane and ethylene in the gas phase isbout 3–25 mol%, it is found that the SII type is preferred for thery gas separation system (Sun et al., 2007; Watanabe et al., 2011;ondo et al., 2014).

.2.2. Equilibrium correlationsTemperature and pressure are controlled at certain value to

nsure hydrate formation.

= T0 (1)

= P0 (2)

The vapor phase and liquid phase are considered at equilibriumnd thereby we have identical fugacity for vapor and liquid.

Lj = f V

j , j ∈ C (3)

Peng–Robinson EOS (Equation of State) is employed to calculate

he fugacity of components in oil and vapor phase.

Chen–Guo equations are used for hydrate-vapor equilibriumChen and Guo, 1998; Ma et al., 2013b). The occupancy of smallnd large hydrate cages are calculated separately.

Pump HX-1

t (CMP: compressor; HX: heat exchanger; R-1, 2: releaser 1, 2).

For small cages,

�j =f Vj

Cj

1 +∑

ifVi

Ci

(4)

Cj = Xj · exp

[Yj

T − Zj

](5)

Cj is the Langmuir constant. For large cages,

z∗j =

f Vj

f 0j

[1 −

∑i�i

]˛ (6)

∑j

z∗j = 1 (7)

f 0j = f 0

Tj · exp

[ˇP

T

]· a−1/�2

w (8)

f 0Tj = exp

[−∑jAij�j

T

]·[

A′j exp

(B′

j

T − C ′j

)](9)

1 ical Engineering 86 (2016) 160–170

˛

z

z

tf

hn

trace

3

V

c

W

tw

N

Ts(lco

H

�F8

V

l

�

wdiei

Table 4Design specifications and parameter ranges.

Parameters Specifications and ranges

Ethylene recovery 0.9Final product methane/(methane + ethylene) ≤0.05Absorption pressure (MPa) =4.25Emulsion flow rate >0

64 Y. Li et al. / Computers and Chem

= �1

�2(10)

j =z∗

j+ ˛�j∑

j(z∗j

+ ˛�j)(11)

∗j

denotes mole fraction of hydrocarbon component j in large cavi-ies. zj stands for the mole fraction of component j in hydrate phase.Vj

is the fugacity of component j in the gas phase. f 0j

stands for theydrate-phase fugacity of component j when the small cavities areot occupied by hydrocarbon molecules. f 0

Tjdenotes the effect of

emperature on f 0j

. For SI hydrate, ̨ = 1/3. For SII hydrate, ̨ = 2. aw

epresents the activity of water. A′i, B′

i, C ′

i, Xi, Yi, Zi and binary inter-

ction coefficients (Chen and Guo, 1998; Sun et al., 2007). Aij areonstants. The parameters values can be found in the work of Mat al. (2013a,b).

.2.3. Mass and energy balance equationsThe mass balance equations are shown as follows,

inyjin

+ Linxjin

+ HYinzjin

+ Wjin

= Voutyjout + Loutx

jout + HYoutz

jout

+ Wjout (12)

Hydrocarbons dissolved in water phase are so trivial that theyan be neglected in the calculation (Chen and Guo, 1998).

jout = 0, j ∈ C and j /= w (13)

A hydrate unit is a fundamental brick with which a hydrate crys-al particle is build. The additional mass balance equation correlatesater molecules and number of units (NOU) in hydrate.

OU = (Wwin

+ HYwin

− Wwout)

NOM(14)

he superscript w denotes water flow rate in each phase. NOMtands for the H2O molecule number in a unit. According to Sloan2003) NOM is determined by the type of hydrate structure. Sincearge cages are always considered fully filled, therefore we can cal-ulate the total amount of light hydrocarbons in hydrate using theccupancy of small cages.

Yout = NOU ·

⎛⎝�1

∑j

�j + �2

⎞⎠ (15)

1 and �2 are the number of small and large cages in a hydrate unit.or SI hydrate, �1 is 2 and �2 is 6. For SII hydrate, �1 is 16 and �2 is.

Energy balance is as follows:

inHVin + LinHL

in + HYinHhyin

+ WinHwin + Q = VoutH

Vout + LoutH

Vout

+ HYoutHhyout + WoutH

wout (16)

Heat generated from hydrate formation is described by the fol-owing equation:

Htotal =∑

j

(�Hj · ncptj

) (17)

here ncptj

is the amount of component j captured in hydrate. In

ry gas separation system, composition of hydrogen and propane

n hydrate is less than 1%, hence they are not incorporated in thenergy balance calculation. For simplicity, oil in water oil emulsions calculated as n-dodecane.

Absorption temperature (◦C) ≥0Water volume fraction (vol%) 10–30

3.2.4. Phase determinationAn issue that deserves attention is that there are two categories

for the multiphase calculation. First, when sufficient water is sup-plied, the system will have 4 phases, namely vapor, oil, water andhydrate. And if we perform calculation using Eqs. (1)–(15) we willhave a positive value for Wout. However, when water is entirelytransformed into hydrate, the system will contain 3 phases namely,vapor, oil and hydrate. In this case, after the calculation using Eqs.(1)–(15), we will obtain a negative value for Wout, which means thatall water is transformed into hydrate. In this situation, the calcula-tion procedure should be carried out again replacing Eqs. (6) and(7) with the following equations

Wout = 0 (18)

z∗i =

f Vj

f 0j

·[∑

i

f Vi

f 0j

]−1

(19)

As shown in Eq. (19), the large cage occupancy is replaced by acompetition mechanism (Zhang et al., 2004; Ma et al., 2008, 2013b).

3.3. Economic evaluation methods

In this work, we evaluate the economic performance of differ-ent process in terms of total annual cost (TAC). TAC is comprised ofannualized utility cost and equipment investment. Utility price isreferred to Seider et al. (2009). Formulas for the equipment invest-ment can be found in Appendix. We choose a payback period of 10years with annualized interest rate of 10%.

3.4. Problem formulation and solution

Economic evaluations in this work are aimed to compare eco-nomic behavior between the proposed AHHP and the C4 absorptionprocess. Since current dry gas recovery system operates at condi-tions where ethylene recovery is 90%, we choose total annual cost(TAC) as the target function with the constraint that ethylene recov-ery equals to 90%. However, the enriched dry gas composition fromdifferent process may not be the same. Therefore, we have to addcost functions for different product gas composition. The functionsare obtained through evaluation of a simplified cryogenic process,which separates methane and nitrogen from enriched dry gas. Theflow chart and operating conditions can be found in Appendix.Eventually, an optimization problem is formulated as follows forthe multistage AHHP.

Min TAC

s.t. Unit operation models;

Specifications on yield and product purity;

Variable ranges.

In the above formulation, TAC incorporates the investment and

operating cost of a dry gas pre-separation process and a simplifiedcryogenic process. The constraints include unit operation mod-els, parameter ranges and design specifications. Specifications andparameter ranges are listed in Table 4. Temperature should be

Y. Li et al. / Computers and Chemical Engineering 86 (2016) 160–170 165

Table 5aOptimized profile for AHHP (ethylene recovery = 90%).

Parameter Optimized value

Absorption temperature 0Water content (vol%) 30Oil flow rate (kmol/h) 105.7Ethylene recovery in AHHP 0.919C1/C2 ratio 0.511Equipment Cost (M$/yr) 0.149Utility Cost (M$/yr) 0.455TAC Cost (M$/yr) 0.604

Table 5bProfile of C4 absorption (ethylene recovery = 90%).

Parameter Value

n-butane flow rate (kmol/h) 638.6B2 reboiler duty (GJ/h) 10.6B3 reboiler duty (GJ/h) 5.67B3 condenser duty (GJ/h) −0.69Ethylene recovery in absorption 0.911C1/C2 ratio 0.401

atgttfl

gTpaocpT

4

4

oprsoasa

adpwr

4

4

T

Table 6Optimized profile for different ethylene recovery.

Ethylene recovery 0.88 0.89 0.9 0.91 0.92

Absorptiontemperature (◦C)

0 0 0 0 0

Water content (vol%) 30 30 30 30 30Oil flow rate (kmol/h) 103.4 104.5 105.7 106.8 108.0Ethylene recovery in

enriched dry gas0.899 0.909 0.919 0.930 0.940

C1/C2 ratio 0.511 0.511 0.511 0.511 0.511Equipment cost

(M$/yr)0.147 0.148 0.149 0.150 0.151

Utility cost (M$/yr) 0.445 0.450 0.455 0.460 0.465TAC (M$/yr) 0.592 0.598 0.604 0.610 0.616

Table 7Optimized profile for different water content (ethylene recovery = 0.9).

Water content (vol%) 10 20 30

Absorption temperature (◦C) 0 0 0Oil flow rate (kmol/h) 164.2 130.7 105.7Ethylene recovery in enriched dry gas 0.917 0.918 0.919C1/C2 ratio 0.350 0.430 0.511Equipment cost (M$/yr) 0.184 0.166 0.149

Equipment cost (M$/yr) 0.195Utility cost (M$/yr) 0.590TAC cost (M$/yr) 0.786

bove 0 ◦C, which is the ice formation temperature. Pressure is seto be the same with C4 absorption. Water content is supposed to bereater than 10 vol% and less than 30 vol%. When water content isoo low, the absorbent cannot reflect the effect of hydration. Whenhe water content is over 30 vol%, the viscosity becomes too largeor efficient stirring if hydrate is formed. Feed gas composition isisted in Table 1. Dry gas flow rate is 7.5 t/h at 10 ◦C, 4.25 MPa.

The optimization problem can be categorized as nonlinear pro-ramming problem, which in this work is solved by the DMO solver.he DMO solver implements a variant of the successive quadraticrogramming (SQP) algorithm. In Aspen Plus®, the optimizationlgorithms are conveniently embedded in the software. Equationriented mode (EO mode) is used, which is known to facilitate theonvergence of recycled flow charts. Specifications on yield androduct purity is added as equations in Calculator Blocks (Aspenechnology, 2012).

. Results and discussion

.1. Optimized profiles

The optimized profile of AHHP is listed in Table 5a. As is shown,ptimal water content is 30 vol%, which is the upper bound. Tem-erature is also optimized to the lower bound, 0 ◦C. Dodecane flowate is 105.7 kmol/h. In the next section, we will look into the sen-itivities of the key parameters one by one. For comparison, profilef C4 absorption process is listed in Table 5b. The parameters in C4bsorption is also tuned through optimization. Sensitivities in nextection are obtained through fixing the investigated parametersnd optimizing other parameters.

The process simulated is the combination of two parts. One is thebsorption pre-separation part. The other is the added cryogenice-methane part. To illustrate the composition of enriched dry gasroduced from absorption part, C1/C2 ratio is used as an indicator,hich is the ratio of methane to ethane and ethylene. Higher C1/C2

atio means greater methane content in enriched dry gas.

.2. Sensitivity analysis

.2.1. Ethylene recoveryOptimized profiles of different ethylene recovery are shown in

able 6. As ethylene recovery steps from 0.88 to 0.92, optimized

Utility cost (M$/yr) 0.424 0.439 0.455TAC (M$/yr) 0.608 0.605 0.604

absorption temperature and water content are 0 ◦C and 30 vol%,respectively. The oil flow rate in emulsion grows almost linearlyfrom 103.4 kmol/h to 108.0 kmol/h. Yet, variation in C1/C2 ratio isfound negligible.

4.2.2. Water contentOptimized profiles for different water content are listed in

Table 7. Influence of water content is reflected on two aspects.First, with the increase of water content from 10 vol% to 30 vol%,oil flow rate in emulsion decreases rapidly from 164.2 kmol/h to105.7 kmol/h. Second, as water content increase, C1/C2 ratio hasgrown from 0.350 to 0.511, which indicates greater methane con-tent in the enriched dry gas.

It is interesting to find the change in TAC not significant fordifferent water content (less than 0.01 M$/yr). With larger watercontent, cost of compressor duty and equipment investment can bereduced, because gas released from hydrate have a higher pressurewhich enables this portion of gas to bypass the compressor. How-ever, on the other hand, as more hydrate is formed, correspondingrefrigeration duty consumed by hydrate formation will go up. Com-promise of the two cost sources mentioned above has resulted inthe small variance of TAC.

Since water content has little effect on TAC, we now try to deter-mine water content considering other aspects. First, for the sakeof easy operation, less water content is better, which makes thehydrate/oil slurry behave more like liquid and will alleviate therisk of forming plugs. However, as is known (Sloan and Koh, 2007),hydrate dissociation can provide a significant amount of refrigera-tion. It is found that phase change materials (PCM) are promisingrefrigeration recovery methods. During hydration, PCM transformsfrom solid to liquid, providing cold duty and the other way around,stores the refrigeration when hydrate dissociates (Song et al., 2015).Yet, currently no successful application of the above method isreported. But if the energy recovery method were utilized, morewater content would no doubt become more favorable.

In sum, dispute exists in the choice of water content. First, theTAC variance with respect to water content is not significant. Sec-

ond, even if minor water content makes operation easier, however,on the other hand, development in refrigeration recovery technol-ogy may turn larger water content to the advantage. Consequently,

166 Y. Li et al. / Computers and Chemical Engineering 86 (2016) 160–170

Table 8Optimized profile for different absorption temperature (ethylene recovery = 0.9).

Absorptiontemperature (◦C)

0 1 2 3 4

Water content (vol%) 30 30 30 30 30Oil flow rate (kmol/h) 105.7 106.9 108.1 109.3 110.5DM condenser duty

(GJ/h)−0.159 −0.159 −0.158 −0.157 −0.157

Ethylene recovery inenriched dry gas

0.919 0.919 0.919 0.919 0.919

C1/C2 ratio 0.511 0.511 0.510 0.509 0.508Equipment cost

(M$/yr)0.149 0.150 0.150 0.151 0.152

Utility cost (M$/yr) 0.455 0.456 0.457 0.458 0.459TAC (M$/yr) 0.604 0.606 0.608 0.609 0.610

5 6 7 8 9 10 110.890

0.895

0.900

0.905

0.910

0.915

0.920

0.925

0.930

Ethy

lene

reco

very

wp

4

Tdcb

4

figlnrtbn0

4

ctgt

0 1 2 3 4 5 6 7 8 9 10 11 120.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18 ethylene methane nitro gen ethane

Mol

e fr

actio

n in

oil

a

b

Stage

0 1 2 3 4 5 6 7 8 9 10 11 120.0

0.1

0.2

0.3

0.4

0.5

0.6 ethy lene met hane nitrog en ethane

Mol

e co

mpo

sitio

n in

hyd

rate

Stage

Fig. 4. (a) Composition distribution in oil. (b) Composition in hydrate (water freebasis).

0 1 2 3 4 5 6 7 8 9 10 11 12

0.2

0.4

0.6

0.8

1.0

1.2

1.4 Oil Hydrat e

C1/

C2

ratio

Stage numbers

Fig. 3. Effect of stage number on ethylene recovery.

e cannot state a definite conclusion on which is preferable tem-orarily.

.2.3. TemperatureAs is shown in Table 8, shift in temperature has little effect on

AC and other parameters. The insignificance can provide someesign space for other concerns. For example, for the sake of pro-ess intensification, 0 ◦C is preferred in that lower temperature canoost hydrate formation.

.2.4. Number of stagesThe number of stages has always been an important factor

or absorption process design (McCabe et al., 1993). The effect ofncreasing stage number becomes less noticeable as the numberrows larger. A heuristic is to choose a stage number which is soarge that when we add another stage the recovery of key compo-ent does not evidently change. Therefore, we fix the emulsion flowate and investigate the change of ethylene recovery with respect tohe number of stages. As can be seen from Fig. 3, ethylene recoveryecomes larger with the increment of stage number. When stageumber is larger than 10, the change of recovery becomes less than.1%.

.3. Composition distribution

Fig. 4a and b shows the composition distribution of some key

omponents along the stages in oil and hydrate phase, respec-ively. As is shown in Fig. 4a, ethane and ethylene content in oil hasrown along the direction of emulsion flow, whereas the absorp-ion of methane and nitrogen is not significant. Fig. 4b shows the

Stage

Fig. 5. C1/C2 ratio of oil and hydrate.

composition variance of hydrate, it is found that the increase of

ethane and ethylene content is not as obvious as that of oil. Forbetter illustration, Fig. 5 is drawn to illustrate the C1/C2 ratio inboth oil and hydrate. It is noticed that in stage-1, the C1/C2 ratio inoil is slightly higher than hydrate. However, as the slurry flows to

Y. Li et al. / Computers and Chemical Engineering 86 (2016) 160–170 167

Table 9Optimized profile for different split fraction.

Split fraction 0 0.2 0.4 0.6 0.8 1

Temperature (◦C) 0 0 0 0 0 0Water content (vol%) 30 30 30 30 30 30Oil flow rate (kmol/h) 105.7 113.2 122.5 134.5 151.1 176.9Ethylene recovery in enriched dry gas 0.919 0.917 0.913 0.910 0.906 0.901C1/C2 ratio 0.511 0.472 0.425 0.366 0.288 0.178Equipment cost (M$/yr) 0.149 0.158 0.169 0.183 0.203 0.233Utility cost (M$/yr) 0.455 0.475 0.502 0.539 0.590 0.661TAC (M$/yr) 0.604 0.633 0.671 0.723 0.793 0.894

1 2 3 4 5 6 7 8 9 10 110.80

0.85

0.90

0.95

1.00

Sum of ethane an d ethylene

Cag

e oc

cupa

ncy

Stage

dw

hesttsCc

tpe

4

estsjgoctTr

Table 10aCost distribution analysis of C4 absorption.

Item Cost (M$/yr)/percentage

Absorbent refrigeration 0.129/16%B-2, 3 reboiler utility 0.335/43%Cryogenic refrigeration 0.112/14%Annualized investment 0.195/25%

Table 10bCost distribution analysis of AHHP.

Item Cost (M$/yr)/percentage

Absorption refrigeration 0.220/35%Compressor utility 0.087/14%Cryogenic refrigeration 0.147/24%Annual investment 0.149/25%

Table 11Enthalpy change of hydrate formation.

Component �H (kJ/mol)

N2 65

®

Fig. 6. Sum of ethane and ethylene occupancy in large cage.

ownward stages, C1/C2 ratio in oil drops rapidly from 1.1 to 0.2,hile in hydrate the C1/C2 ratio remains around 1.0.

This phenomena can be explained by the special structure inydrate. As is known, SII hydrate has two cage types. One is largenough to capture ethane or ethylene. The other is comparativelymall, which can accommodate methane and nitrogen, but unableo hold ethylene or ethane (Sloan, 2003). We can see from Fig. 6hat the large cages are already fully occupied even in the firsttage. Thus there is no more room for more ethylene in the hydrate.onsequently, though effective for the oil phase, the multistageonfiguration fails to improve composition of ethylene in hydrate.

The analysis into composition distribution informs us that dueo the special structure, hydrate is suitable to recover certain com-onents from its low concentration mixture but may be not asffective in improving composition profile as oil.

.4. Alternative configuration

An alternative flow chart configuration is considered, whichnables the process to control enriched dry gas C1/C2 ratio. As ishown in Fig. 7, R-3 hydrate is dissociated at 4.25 MPa by raisingemperature. The gas released is split by SP into two streams. Onetream is returned to stage-15 (stage number is increased to 15)ust like a stripping gas for an absorber. The other goes to mix withas produced by R-1 and R-2. The dry gas still feed at stage-11. Theptimized profile for different split fraction is shown in Table 9. Asan be seen, the C1/C2 ratio has fallen from 0.511 to 0.178. However,

he TAC also raises significantly from 0.604 M$/yr to 0.894 M$/yr.o summarize, the alternative process can offer a much lower C1/C2atio but at the expense of higher TAC.

CH4 54C2H6 74C2H4 74

4.5. Cost distribution

Here we analyze the cost distribution of the proposed AHHPand compare it to the C4 absorption process. As we can see fromTable 10a, in C4 absorption, re-boiler duty of B-2, 3 takes 43% of theTAC.

For AHHP, refrigeration and compressor duty add up to 50% ofthe TAC, as is shown in Table 10b. It is found that refrigeration costoriginates from removal of the heat generated during hydrate for-mation, as shown in Table 11. On the other hand, the process has agas stream produced at a lower pressure, which requires recom-pression before further treatment. For this reason and becauseelectricity is comparatively dear (0.04$/kWh), therefore, compres-sor duty becomes the other expense consuming item. It is alsonoticed that equipment cost for AHHP is higher. The main reasonis the compressor (0.95 M$) that recompress a product stream.

From the analysis above, we find that steam utility is the maincost in C4 absorption and therefore we may try to integrate thereboilers with other units in the refinery. For AHHP, as refrigerationand compression is the main cost, refrigeration recovery methodsand pressure equalization methods should be adopted for furtherimprovement.

5. Conclusions

In this work, we put forward ACM for the modeling and sim-ulation of complex hydration process. This approach is found to berobust and convenient. With the help of ACM®, a multistage AHHP

168 Y. Li et al. / Computers and Chemical Engineering 86 (2016) 160–170

CMP

HX- 2

R-2

Pump HX-1

R-1

Enriched dry gas

Stage-n Stage-(n-1 ) Off- gas

Stage-i

Stage-1

...

Fee d dry gas

R-3

...

SP

ydrate dissociation tank, 4.25 MPa, 40 ◦C; SP: gas splitter).

ffkiwtleetstAA

A

NtpSo

A

Table 12Operating condition for cryogenic process.

Unit operation Pressure (MPa) Temperature (◦C) Tray number

CR-HX1 3.3 Dry gas target −29.5 –CR-HX2 3.3 Dry gas target −62.5 –CR-HX3 3.3 Dry gas target −96 –DM 3.3 −98 to −5 60Turbine 0.3 −130 –Refrigerant 1 – −34.5 –

Fig. 7. An alternative configuration for AHHP (R-3: h

or ethylene recovery from dry gas is modeled and simulated. Weound the optimal operating conditions for AHHP. Influence of theey parameters are discussed. It is found that emulsion flow rates controlling the ethylene recovery. Influence of temperature and

ater content on TAC is found insignificant. As a consequence, thewo parameters should be determined by other concerns. By ana-yzing composition distribution along the stages, we can see thatthylene and ethane composition in hydrate cannot improve asffectively as that of oil, which is attributed to the special struc-ure in hydrate. Economic evaluations on AHHP and C4 absorptionhow the former has a lower TAC. The cost distribution indicateshat most cost of AHHP comes from refrigeration and compression.doption of refrigeration recovery methods is recommended forHHP’s further improvement.

cknowledgments

Financial support for this investigation was received from theational Key Basic Research Program of China (No. 2012CB215005),

he 2013 China–Europe Small-Sized and Medium-Sized Enter-rises Energy Saving and Carbon Reduction Research Project (No.Q2013ZOA100002), and the National Natural Science Foundationf China (No. 21336007).

ppendix.

Introduction to cryogenic process:In order to compare different process, we are supposed to make

different process at the same end point. As the two process intro-duced above may have different enriched dry gas composition, asimplified cryogenic process is presented in Fig. 8. By adding sucha cryogenic process the two flow charts can now share the same

finish point, at which methane/ethylene ratio satisfies the require-ment of polymerization (methane/ethylene less than 0.05 mol%).Cost of the cryogenic distillation part is also incorporated in ourevaluation. The flow chart of cryogenic process is a simplified

Refrigerant 2 – −68.5 –Refrigerant 3 – −101 –

one, in which we ignored some unit operations that are consid-ered identical and therefore make little difference in economiccomparison, such as alkaline wash tower and de-NOx tower.

As is shown in Fig. 8, the enriched dry gas is chilled throughseveral stages. After each stage, the mixtures are flashed. The liquidphase is lined to DM and the vapor phase is further cooled in thenext stage. This stage by stage refrigeration pattern is devised forenergy saving. Finally, after the final flash operation, the vaporphase is mixed with gas out of DM condenser. The mixture thengoes to expander to provide refrigeration. Operating conditionsare listed in Table 12.Equipment cost ($):

Vessel investment = 17640 · D1.066L0.802 (20)

Heat Exchanger investment = 7296 · A0.65 (21)

Compressor = 1293 · 517.3 · 3.11 · (0.746Wcmp)0.82

280(22)

Cryogenic Heat Exchange investment = 30000 + 750 · A0.81 (23)

Space between trays are calculated as 0.64 m. As is shown inEq. (23), cryogenic (−60 to −101 ◦C) heat exchangers are moreexpensive than ordinary heat exchangers.

Y. Li et al. / Computers and Chemical Engineering 86 (2016) 160–170 169

Enriched dry gas fee d

F-1 F-2 F-3

DM

Product

Turbine

Off ga s

CR-HX1 CR-HX2 CR-H X3

Refrigerant 1 Refr igerant 2 Refr igerant 3

Fig. 8. Simplified cryogenic process (CR-HX1, 2, 3: cryogenic heat exchanger 1, 2, 3; F-1, 2, 3: flash tank 1, 2, 3; DM: de-methane column).

Table 13Parameter value for Chen–Guo equations.

Gas Antoine constant SII structure

X (bar) Y (K) Z (K) A′ (bar) B′ (K) C′ (K)

H2 5.64E−06 120.775 253.1 1.00E+23 0 0O2 9.50E−06 2452.29 1.03 4.32E+23 −12,505.00 −0.35N2 4.32E−06 2472.37 0.64 6.82E+23 −12,770.00 −1.10CH4 2.30E−06 2752.29 23.01 5.26E+23 −12,955.00 4.08COa

CO2 1.65E−06 2799.66 15.9 3.45E+23 −12,570.00 6.79C2H4

b 3.77E+21 −13,841.00 0.55C2H6

b 3.99E+21 −11,491.00 30.4C H b 4.10E+23 −12,312.00 39

ramet

C

A

i2

R

A

AC

E

E

E

H

K

3 8

a There is currently no CO parameter for Chen–Guo equation in literature. The pab For the blank spaces, the parameter is considered to be 0.

hen–Guo equation parameters

Parameters for Chen–Guo equation can be found in Table 13.

ppendix A. Supplementary data

Supplementary data associated with this article can be found,n the online version, at http://dx.doi.org/10.1016/j.compchemeng.015.12.021.

eferences

dhikari S, Fernando S. Hydrogen membrane separation techniques. Ind Eng ChemRes 2006;45(3):875–81.

spen Technology Inc., 2012. http://www.aspentech.com. Cambridge, MA, USA.hen GJ, Guo TM. A new approach to gas hydrate modelling. Chem Eng J

1998;71(March):145–51.nglezos P, Kalogerakis N, Dholabhai PD, Bishnoi PR. Kinetics of formation of

methane and ethane gas hydrates. Chem Eng Sci 1987a;42(11):2647–58.nglezos P, Kalogerakis N, Dholabhai PD, Bishnoi PR. Kinetics of gas

hydrate formation from mixtures of methane and ethane. Chem Eng Sci1987b;42(11):2659–66.

slamimanesh A, Mohammadi AH, Richon D, Naidoo P, Ramjugernath D. Applica-tion of gas hydrate formation in separation processes: a review of experimental

studies. J Chem Thermodyn 2012;46:62–71.

uang HY, Padin J, Yang RT. Comparison of �-complexations of ethylene and carbonmonoxide with Cu+ and Ag+. Ind Eng Chem Res 1999;38(7):2720–5.

ondo W, Ohtsuka K, Ohmura R, Takeya S, Mori YH. Clathrate-hydrate forma-tion from a hydrocarbon gas mixture: compositional evolution of formed

ers of CO2 is used for CO.

hydrate during an isobaric semi-batch hydrate-forming operation. Appl Energy2014;113:864–71.

Lei Si WZ. Summary on running of 110 kt/a catalytic dry gas ethylene recovery unit.Qilu Petrochem Technol 2013;4(41):274–7.

Li Y, Luo H. Integration of light hydrocarbons cryogenic separation process inrefinery based on LNG cold energy utilization. Chem Eng Res Des 2015;93:632–9.

Liu B, Liu H, Mu L, Sun C, Chen G. Compositional analysis and hydrate dissocia-tion condition measurements for CH4 + C2H4 + C2H6 + N2 gas mixtures in dieseloil + water + anti-agglomerant emulsion system. In: Proceedings of the 13thinternational conference on properties and phase equilibria for product andprocess design; 2013.

Liu H, Mu L, Wang B, Liu B, Wang J, Zhang X, et al. Separation of ethylene fromrefinery dry gas via forming hydrate in w/o dispersion system. Sep Purif Technol2013;116:342–50.

Ma C, Chen G, Wang F, Sun C, Guo T. Hydrate formation of (CH4 + C2H4) and(CH4 + C3H6) gas mixtures. Fluid Phase Equilib 2001;191:41–7.

Ma Q, Chen G, Zhang L. Vapor-hydrate phases equilibrium of (CH4 + C2H6) and(CH4 + C2H4) systems. Pet Sci 2008;5:359–66.

Ma Q, Huang Q, Chen G, Wang X, Sun C, Yang L. Kinetic and phase behaviors ofcatalytic cracking dry gas hydrate in water-in-oil emulsion. Chin J Chem Eng2013a;21(3):295–300.

Ma QL, Chen GJ, Sun CY. Vapor–liquid–liquid–hydrate phase equilibrium calcu-lation for multicomponent systems containing hydrogen. Fluid Phase Equilib2013b;338:87–94.

McCabe WL, Smith JC, Harriott P. Unit operations of chemical engineering. New York:McGraw-Hill; 1993.

Mu L, Li S, Ma Q-L, Zhang K, Sun C-Y, Chen G-J, et al. Experimental and model-

ing investigation of kinetics of methane gas hydrate formation in water-in-oilemulsion. Fluid Phase Equilib 2014;362:28–34.

Murshed MM, Schmidt BC, Kuhs WF. Kinetics of methane-ethane gas replacementin clathrate-hydrates studied by time-resolved neutron diffraction and Ramanspectroscopy. J Phys Chem A 2010;114:247–55.

1 ical E

N

P

R

S

S

S

S

S

70 Y. Li et al. / Computers and Chem

aeiji P, Mottahedin M, Varaminian F. Separation of methane–ethane gas mixturesvia gas hydrate formation. Sep Purif Technol 2014;123:139–44.

adin J, Yang RT. New sorbents for olefin/paraffin separations by adsorptionvia �-complexation: synthesis and effects of substrates. Chem Eng Sci2000;55(14):2607–16.

ahimi N, Karimzadeh R. Catalytic cracking of hydrocarbons over modified ZSM-5zeolites to produce light olefins: a review. Appl Catal A: Gen 2011;398(1):1–17.

adeghbeigi R. Fluid catalytic cracking handbook: an expert guide to the practicaloperation, design, and optimization of FCC units; 2012.

alerno D, Arellano-Garcia H, Wozny G. Ethylene separation by feed-splitting fromlight gases. Energy 2011;36(7):4518–23.

eider WD, Seader JD, Lewin DR. Product & process design principles: synthesis,analysis and evaluation. John Wiley & Sons; 2009.

loan ED. Fundamental principles and applications of natural gas hydrates. Nature2003;426(November):353–63.

loan ED Jr, Koh C. Clathrate hydrates of natural gases. CRC Press; 2007.

ngineering 86 (2016) 160–170

Song X, Xin F, Yan H, Li X, Jia H. Intensification and kinetics of methane hydrateformation under heat removal by phase change of n-tetradecane. AIChE J2015;61(10):3441–50.

Subramanian S, Kini RA, Dec SF, Sloan ED. Evidence of structure II hydrate formationfrom methane + ethane mixtures. Chem Eng Sci 2000;55:1981–99.

Sun C-Y, Ma C-F, Chen G-J, Zhang S-X. Experimental and simulation of single equi-librium stage separation of (methane + hydrogen) mixtures via forming hydrate.Fluid Phase Equilib 2007;261(1):85–91.

Turner DJ, Miller KT, Sloan ED. Methane hydrate formation and an inward growingshell model in water-in-oil dispersions. Chem Eng Sci 2009;64(18):3996–4004.

Watanabe S, Saito K, Ohmura R. Crystal growth of clathrate hydrate in liquid

water saturated with a simulated natural gas. Cryst Growth Des 2011;11(7):3235–42.

Zhang LW, Chen GJ, Guo XQ, Sun CY, Yang LY. The partition coefficients ofethane between vapor and hydrate phase for methane + ethane + water andmethane + ethane + THF + water systems. Fluid Phase Equilib 2004;225:141–4.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具