Developing a Steady-state Kinetic Model for Industrial Scale Semi-Regenerative Catalytic Naphtha...

8/10/2019 Developing a Steady-state Kinetic Model for Industrial Scale Semi-Regenerative Catalytic Naphtha Reforming Proc

1/6

R. S. MOHADDECY and S. SADIGHI: Developing a Steady-state Kinetic..., Kem. Ind. 63 (5-6) 149154 (2014) 149

Introduction

The catalytic reforming of heavy naphtha (heavy straight rungasoline or HSRG) is a favourite process in petroleum rener-

ies due to producing high-octane gasoline.1

The semi-regen-erative naphtha reformer is the oldest type where reactionsare carried out in three or four adiabatic xed-bed reactorsin series, each of which is equipped with a pre-heater. Thisplant usually operates at temperatures between 450 C and520 C, total pressure between 25 and 35 atm, and hydro-gen-to-hydrocarbon amount ratios between 3 and 7.1,2,3

Usually, the feed of the catalytic reforming process is HSRG,including four hydrocarbon groups i.e. parafns, olens,naphthenes and aromatics (PONA) with number of carbonatoms between 5 and 10.

The main reforming reactions occurring through the cata-

lytic beds are dehydrocyclization, hydrocracking, isomeriza-tion, dehydrogenation and cyclization. Some of these reac-tions are desired for increasing octane number of gasoline,whereas others are undesired because they decrease it. Forparafns, increment of octane number is the result of reac-tions increasing the number of branches, such as cyclizationand aromatization. Therefore, normal parafns conversionto isoparafns, naphthenes, and aromatics can increase theoctane number.3

The catalytic reforming process is often modelled based on:1) the number of reactive species, and 2) the type of theused kinetic model. However, the presence of many com-

ponents as reactants or products causes numerous reactions.Therefore, the situation is extremely sophisticated for pro-cess modelling. To decrease these complications, reactantsin the mixture are classied into limited groups called pseu-

do-components or lumps. The number of selected lumps inthe mixture is a determinant factor for designing the reform-ing model. Arrhenius and LangmuirHinshelwood kineticsare widely used for catalytic reforming models.

In the eld of catalytic naphtha process modelling, a sim-ple and rst model was suggested by Smith4in which naph-tha reforming was considered as a combination of onlyfour reactions. Then, in 1997, Taskarsuggested a model forthe catalytic reforming reaction that consisted of 35 pseu-do-components in the reaction network and 36 reactions.5Following the use of Arrhenius kinetics, a well-known modelwas proposed by Padmavathi6 in 1997 in which 26 pseu-

do-components, such as alkyl cyclohexane (ACH), alkyl cy-clopentane (ACP), normal parafns (NP), isoparafns (IP),aromatics (A), hydrogen (H2) and light hydrocarbons (C1toC5) were used in the network.Ancheyta et al.

7developed akinetic model for the naphtha catalytic reforming process.This model utilized a lumped mathematical model, pre-senting the reactions ranging from 1 to 11 carbon atoms forparafns, and from 6 to 11 carbon atoms for naphthenesand aromatics. In 2003 Rahimpour et al. presented a kinet-ic model for industrial scale catalytic naphtha reformers,including deactivation of the catalyst. The impact of inlettemperature, operating pressure and catalyst mass distri-bution on the performance of the reactors was examined.

The results indicated an increase in aromatic yield with in-creasing inlet temperature. However, manipulating the op-erating pressure had no appreciable effect on the gasolineyield. Additionally, this model estimated catalyst deactiva-

* Corresponding author: Reza Seif Mohaddecy, email: [email protected],Tel. +98-21-48255051, Fax. +98-21-44739709, P.O. Box 14665-137

Developing a Steady-state Kinetic Model forIndustrial Scale Semi-Regenerative CatalyticNaphtha Reforming Process

Due to the demand for high octane gasoline as a transportation fuel, the catalytic naphtha reform-er has become one of the most important processes in petroleum reneries. In this research, thesteady-state modelling of a catalytic xed-bed naphtha reforming process to predict the momen-tous output variables was studied. These variables were octane number, yield, hydrogen purity,and temperature of all reforming reactors. To do such a task, an industrial scale semi-regenerativecatalytic naphtha reforming unit was studied and modelled. In addition, to evaluate the developedmodel, the predicted variables i.e.outlet temperatures of reactors, research octane number, yield

of gasoline and hydrogen purity were compared against actual data. The results showed that thereis a close mapping between the actual and predicted variables, and the mean relative absolutedeviation of the mentioned process variables were 0.38 %, 0.52 %, 0.54 %, 0.32 %, 4.8 % and3.2 %, respectively.

Key words: Catalytic naphtha reforming, modelling, simulation, kinetics

R. Seif Mohaddecy*and S. Sadighi

DOI: 10.15255/KUI.2013.009KUI-9/2014

Received April 11, 2013Accepted June 4, 2013

Catalysis and Nanotechnology Research Division, Catalytic Reaction Engineering Department, ResearchInstitute of Petroleum Industry, West Blvd. of Azadi Sports Complex, Tehran, Iran
mailto:[email protected]://dx.doi.org/10.15255/KUI.2013.009http://dx.doi.org/10.15255/KUI.2013.009mailto:[email protected]


2/6

R. S. MOHADDECY and S. SADIGHI: Developing a Steady-state Kinetic Model..., Kem. Ind. 63 (5-6) 149154 (2014)150

tion where the corresponding parameters were estimatedusing the plant data.8In 2006 Hou et al.presented a new18-lump kinetic-based mathematical model for an industrialcontinuous catalytic reforming plant. In this model, reactiontemperature and concentration proles of all reactors, heat-er duties, catalyst deactivation, recycle gas composition andoctane number for different feedstock or operating condi-tions could be predicted.9In the next effort in 2009 Araniet al.developed a lumping procedure to obtain kinetic andthermodynamic parameters of catalytic naphtha reformerusing a model consisting of 17 lumps ranging from C6 toC8+ hydrocarbon, and including 15 reaction pathways.

10Inthe same year Fatemi et al.developed a mathematical modelfor a commercial naphtha catalytic reformer which includedthree sequencing catalytic xed-bed reactors at steady-statecondition. They used a detailed kinetic scheme involving 26pseudo-components connected by a network of 47 reac-tions. The output variables of the reformer, such as RON (re-search octane number) and yield of gasoline showed goodagreement with actual data obtained from the under studyreforming unit.11Recently Ziaoonpresented a detailed ki-netic model including 24 components, 1 to 11 carbon atomsfor parafns, and 6 to 11 carbon atoms for naphthenes andaromatics, forming 71 reactions.12

In this paper, in order to model an industrial scale catalyt-ic naphtha reforming process, a steady-state kinetic modelincluding 9 pseudo-components and 4 reactions was devel-oped. In comparison to the complex models proposed forthis process, this model can also show low average absolutedeviation against actual data.

Process description

A commercial xed-bed catalytic naphtha reforming unit,called Platformer, licensed by Chevron research coopera-tion was chosen as a case study. The feed of the plant prior

to entering the catalytic reformer should undergo hydrode-sulphurization (HDS) reaction in the hydrotreatment unit.Then, the produced naphtha, called Platcharge, is intro-duced to the reforming process. The most commonly usedtypes of catalytic reforming units have three or four reactorseach having a xed catalytic bed.3,4

As shown in Fig. 1, Platcharge is rst preheated by the rstfurnace (H-1), and then it enters the rst reactor (R-1) wherethe naphthenes are dehydrogenated to aromatics. The pro-duct stream from the rst reactor then passes through thesecond reactor (R-2), and enters the third reactor (R-3). Theoverall reforming reactions are endothermic; therefore, apreheater (H-1, H-2, and H-3) should essentially be provid-ed before each reforming reactor.

Next, the product stream from the third reactor enters theash separator (V-1) wherein the produced hydrogen is sep-arated, and recycled to the beginning of the process. Thisrecycled stream is then mixed with the fresh naphtha feed(Platcharge). Finally, the liquid product leaving the separator

is introduced to the gasoline stabilizer in which the LPG andlight gases are separated from the gasoline. So, the vapourpressure of the gasoline can be set according to the marketrequirement.

The specications of feed and catalyst distribution of thestudied catalytic naphtha reforming plant are given in Tables1 and 2. The normal operating conditions of the unit arepresented in Table 3.

The design pressure of the studied catalytic reforming plantwas 34 bar. However, depending on the feed specicationand deactivation of the catalyst, the operating pressure of

the plant could be varied between 27 and 32 bar. So, theeffect of the pressure on reforming reactions was included inthe model, which is discussed later.

F i g . 1 Catalytic reforming owchart (semi-regenerative)S l i k a 1 Dijagram toka poluobnavljajueg katalitikog reformiranja


3/6


T a b l e 1 Specication of feed of catalytic naphtha reformingprocess

T a b l i c a 1 Sirovina za katalitiko reformiranje benzina

VariableVeliina

ValueVrijednost

density / g cm3

gustoa / g cm30.745

distillation method (ASTM D86)destilacija po normi ASTM D86

volume fractionobujamski udjel

boiling point / Cvrelite / C

IBP 88

10 % 99

30 % 105

50 % 113

70 % 123

90 % 137

FBP 156PONA analysisanaliza PONA

componentskomponente

volume fractionobujamski udjel

parafnsparani

55 %

aromaticsaromati

12 %

naphthenesnafteni

33 %

T a b l e 2 Catalyst distribution in reforming reactors

T a b l i c a 2 Raspodjela katalizatora u reaktorima

1streactor1. reaktor

2ndreactor2. reaktor

3rdreactor3. reaktor

catalyst massmasa katalizatora

8 648 kg 14 223 kg 22 452 kg

catalyst mass distributionraspodjela katalizatora

20 % 30 % 50 %

T a b l e 3 Operating conditions in the catalytic reforming ofthe target oil renery

T a b l i c a 3 Uvjeti katalitikog reformiranja u raneriji

Process variableProcesna varijabla

ValueVrijednost

inlet temperature / Culazna temperatura / C

490 515

hydrogen/hydrocarbon amount ratiomnoinski omjer vodika i ugljikovodika

3 7

LHSV / h1 1 2

volume yield / %obujamsko iskoritenje / %

70 85

Development of kinetic model4In the present study, in order to simulate a catalytic reform-ing unit, the naphtha feed is classied into three general

lumps i.e.aromatics, naphthenes and parafns. Also consid-ered were hydrogen, methane, propane, butane, and pen-tane. The reactions within this model are classied into fourgroups, as follows:

Naphthenes to aromatics

Rate constants concerning this reaction are as follows:4

(1)

(2)

Naphthenes to parafns

Rate constants concerning this reaction are as follows:4

(3)

(4)

Hydrocracking of parafns

The rate of parafns cracking and rate constants concerningthis reaction are as follows:4

(5)

(6)

Hydrocracking of naphthenes

The rate of naphthenes cracking and rate constants concern-ing this reaction are as follows:4

(7)

(8)Using the presented rate equations, the mass and energyconservations can be written as follows:

(9)

(10)

(11)


4/6


, (12)

where nis the number of each presumed carbon of pseu-do-components4which is between 7 and 6 for the feed inthe model.

Results and discussions

After developing the kinetic model using data obtained froma pilot-scale catalytic naphtha reforming unit, it should bescaled up to the industrial scale. An optimization subrou-tine was used to determine the coefcients until reaching

a suitable consistency between the industrial data and themodel results. In this subroutine, the LevenburgMarquardtoptimization algorithm was used, and the following targetfunction was optimized:

(13)

The magnitudes of constants are presented in Table 4 for thestudied reforming plant.

T a b l e 4 Reaction constants calculated using optimized

approach

T a b l i c a 4 Konstante izraunate optimiziranim pristupom

ReactionnumberBroj reakcije

Reaction nameNaziv reakcije

k0

1aromatic productionnastajanje aromata

18.59 34 807

2parafns productionnastajanje parana

26.74 58 591

3 parafns hydrocrackinghidrokrekiranje parana

42.97 62 857

4naphthenes hydrocrackinghidrokrekiranje naftena

42.97 61 224

Now, in order to evaluate the accuracy of the model, thesimulated data were compared with the actual (industrial)data. Figs. 2 to 4 show comparisons between the outlet re-actor temperatures obtained by the model against the actualvalues. The results showed that AAD of the predicted reactor

temperatures were 0.38 %, 0.52 % and 0.54 %, respectively.From Figs. 2 to 4, close mappings between the measuredand simulated outlet temperatures of reactors can be ob-served.

F i g . 2 Simulated outlet temperature of 1st reactor againstactual values

S l i k a 2 Simulirana ulazna temperatura 1. reaktora u usporedbisa stvarnim vrijednostima

F i g . 3 Simulated outlet temperature of 2nd reactor againstactual values


F i g . 4 Simulated outlet temperature of 3rd reactor againstactual values


Other signicant operating parameters in catalytic reform-ing are product volume yield and RON. Comparisons of thepredicted product volume yield and RON with the actualdata are shown in Figs. 5 and 6. From these gures, closemappings between the measured and simulated productvolume yield and RON can be observed. Moreover, it isfound that the presented model can simulate the RON and

volume yield of gasoline with AAD of 0.32 % and 4.8 %,respectively. These results conrm that the presented ap-proach can reliably be applied by reneries to monitor theoperation of the catalytic reforming plant.


5/6


F i g . 5 Simulated volume yield of gasoline against actual valuesS l i k a 5 Simulirano obujamsko iskoritenje benzina u usporedbi

sa stvarnim vrijednostima

F i g . 6 Simulated RON of product against actual valuesS l i k a 6 Simulirani RON produkta u usporedbi sa stvarnim

vrijednostima

The other signicant parameter in the catalytic reforming

plant is the hydrogen purity of the recycle stream. Fig. 7 pre-sents the comparison between the simulated hydrogen pu-rity and the actual data. These results show that the kineticmodel can predict the purity of the produced hydrogen withAAD of 3.19 %. It is concluded that the presented kineticmodel is also good for predicting the purity of hydrogen pro-duced in the catalytic naphtha reforming unit.

F i g . 7 Simulated hydrogen purity (expressed as volumefraction) against actual values

S l i k a 7 Simulirana istoa vodika (izraena kao mnoinskiudjel) u usporedbi sa stvarnim vrijednostima

Conclusions

The catalytic reforming of heavy naphtha (heavy straight rungasoline or HSRG) is a favourite process in petroleum ren-eries for producing high-octane gasoline. In this research,

signicant process variables of a commercial naphtha cata-lytic reforming plant were modelled using a heterogeneouskinetic model. These variables included the outlet tempera-tures of the rst, second and third reactors, RON of gasoline,product volume yield and hydrogen purity. To evaluate theproposed model, the results were compared against dataobtained from a commercial catalytic naphtha reformer. It

was found that the mean relative absolute deviation (AAD)of the mentioned parameters were 0.38 %, 0.52 %, 0.54 %,0.32 %, 4.8 % and 3.2 %, respectively. Therefore, a closemapping was conrmed between the simulated variablesand data obtained from an industrial-scale reforming plant.These results show that the presented kinetic model can re-liably be utilized to monitor the operation of the catalyticreforming plant.

List of symbols and abbreviationsPopis simbola i kratica

c amount concentration, mol m3

mnoinska koncentracija, mol m3

cp heat capacity at constant pressure, J K1mol1

toplinski kapacitet pri stalnom tlaku, J K1mol1

Ea activation energy, J mol1

aktivacijska energija, J mol1

H enthalpy, J mol1

entalpija, J mol1

Keq equilibrium constant ravnotena konstanta

k rate reaction constant konstanta brzine reakcije

k0 pre-exponential factor

predeksponencijalni faktorN number of atoms

broj atoma

nA amount of aromatics, mol mnoina aromata, mol

nN amount of naphthenes, mol mnoina naftena, mol

nP amount of parafns, mol mnoina parana, mol

nT total amount of substance, mol ukupna mnoina tvari, mol

pA partial pressure of aromatics, atm parcijalni tlak aromata, atm

pN partial pressure of naphthenes, atm parcijalni tlak naftena, atm

pP partial pressure of parafns, atm parcijalni tlak parana, atm

pT total pressure, atm ukupni tlak, atm

R universal gas constant, R= 8.314 J K mol1= 4,619 J R mol1

opa plinska konstanta, R= 8.314 J K mol1= 4,619 J R mol1

rN-crack-ing

naphthenes cracking rate brzina krekiranja naftena

rP-cracking parafns cracking rate brzina krekiranja parana


6/6


T thermodinamic temperature, R

termodinamika temperatura, R

VR volume of reactor, m3

obujam reaktora, m3

A aromatics aromati

AAD mean relative absolute deviation srednja relativna apsolutna devijacija

ACH alkyl cyclohexane alkilirani cikloheksan

ACP alkyl cyclopentane alkilirani ciklopentan

FBP nal boiling point zavrno vrelite

HDS hydrodesulphurization hidrodesulfurizacija

HSRG heavy straight run gasoline teki primarni benzin

IBP initial boiling point poetno vrelite

IP isoparafns izoparani

LHSV liquid hourly space velocity, h1

prostorna brzina kapljevine, h1

LPG liqueed petroleum gas ukapljeni naftni plin

NP normal parafns normalni parani

PONA parafns, olens, naphthenes, aromatics parani, oleni, nafteni, aromati

RON research octane number istraivaki oktanski broj

degree Rankine; conversion formula: 1 R = 59K Rankineov stupanj; formula za pretvorbu: 1 R = 59K

ReferencesLiteratura

1. A. Chang, K. Pashikanti, Y. A. Liu,Renery Engineering, John Wiley& Sons, 2012, http://dx.doi.org/10.1002/9783527666836.

2. M. A. Fahim, T. A. Al-Sahhaf, A. Elkilani, Fundamentals of Petro-leum Rening, Elsevier, 2009.

3. G. J. Antons, A. M. Aitani, Catalytic Naphtha Reform-

ing, 2nd

Ed., Marcel Dekker Inc., 2004, http://dx.doi.org/10.1201/9780203913505 .

4. R. B. Smith, Kinetic Analysis of Naphtha Reforming with Plati-num Catalyst, Chem. Eng. Prog. 55(6) (1959) 7680.

5. U. Taskar, J. B. Riggs, Modeling and optimization of a semi-re-generativecatalytic naphtha reformer, AIChE J. 43(3) (1997)740753, http://dx.doi.org/10.1002/aic.690430319.

6. G. Padmavathi, K. K. Chaudhuri, Modeling and simulation ofcommercial catalytic naphtha reformers, Can. J. Chem. Eng.75 (10) (1997) 930937.

7. J. Ancheyta, E. Villafuerte, Kinetic Modeling of Naphtha Cata-lytic Reforming Reactions, Energy & Fuels 14(5) (2000) 10321307, http://dx.doi.org/10.1021/ef0000274.

8. M. R. Rahimpour, S. Esmaili, G. N. A. Bagheri, A kinetic anddeactivation model for industrial catalytic naphtha reforming,Iran. J. Sci. Technol., Trans. B: Tech. 27(2003) 279290.

9. W. Hou, H. Su, J. Chu,Modeling, Simulation and Optimizationof a Whole Industrial Catalytic Naphtha Reforming Process onAspen Plus Platform, Chinese J. Chem. Eng. 14 (5) (2006) 584591, http://dx.doi.org/10.1016/S1004-9541(06)60119-5.

10. H. M. Arani, M. Shirvani, K. Safdarian, E. Dorostkar, Lumpingprocedure for a kinetic model of catalytic naphtha reform-ing, Braz. J. Chem. Eng. 26 (2009) 723732, http://dx.doi.org/10.1590/S0104-66322009000400011 .

11. A. Fazeli, S. Fatemi, M. Mahdavian, A. Ghaee,MathematicalModeling of an Industrial Naphtha Reformer with Three Ad-iabatic Reactors in Series, Iran. J. Chem. Chem. Eng. 51 (3)

(2009) 97102.12. Z. M. Shakoor, Catalytic reforming of heavy naphtha, analysis

and simulation, Diyala J. Eng. Sci. 4(2) (2011) 86104.

SAETAKRazvoj kinetikog modela poluobnavljajueg katalitikog reformiranja benzina

u ustaljenom stanju u industrijskom mjerilu

R. Seif Mohaddecy*i S. Sadighi

Zbog velike potranje za visokooktanskim benzinom kao transportnim gorivom, katalitiko refor-miranje benzina postalo je jedan od najvanijih procesa u ranerijama nafte. Da bi se predvidjelevrijednosti vanih izlaznih varijabli, u ovom je istraivanju modelirano katalizirano reformiranjebenzina u ustaljenom stanju s nepokretnim slojem katalizatora. Te su varijable oktanski broj, isko-ritenje, istoa vodika i temperatura u svim reaktorima za reformiranje. Stoga je prouena i mo -delirana jedinica za poluobnavljajue katalitiko reformiranje benzina u industrijskom mjerilu,a vrijednosti izlaznih varijabli usporeene su sa stvarnim podacima. Rezultati su pokazali dobroslaganje stvarnih i predvienih vrijednosti. Srednje relativno apsolutno odstupanje izlaznih tempe-ratura reaktora, oktanskog broja, iskoritenja, istoe vodika i temperature iznosi 0,38 %, 0,52 %,0,54 %, 0,32 %, 4,8 % i 3,2 %.

Prispjelo 11. travnja 2013.Prihvaeno 4. lipnja 2013.

Catalysis and Nanotechnology Research Division,Catalytic Reaction Engineering Department,Research Institute of Petroleum Industry,West Blvd. of Azadi Sports Complex, Tehran, Iran
http://dx.doi.org/10.1002/9783527666836http://dx.doi.org/10.1201/9780203913505http://dx.doi.org/10.1201/9780203913505http://dx.doi.org/10.1002/aic.690430319http://dx.doi.org/10.1002/aic.690430319http://dx.doi.org/10.1002/aic.690430319http://dx.doi.org/10.1002/aic.690430319http://dx.doi.org/10.1021/ef0000274http://dx.doi.org/10.1016/S1004-9541%2806%2960119-5http://dx.doi.org/10.1590/S0104-66322009000400011http://dx.doi.org/10.1590/S0104-66322009000400011http://dx.doi.org/10.1590/S0104-66322009000400011http://dx.doi.org/10.1590/S0104-66322009000400011http://dx.doi.org/10.1590/S0104-66322009000400011http://dx.doi.org/10.1590/S0104-66322009000400011http://dx.doi.org/10.1590/S0104-66322009000400011http://dx.doi.org/10.1590/S0104-66322009000400011http://dx.doi.org/10.1590/S0104-66322009000400011http://dx.doi.org/10.1590/S0104-66322009000400011http://dx.doi.org/10.1016/S1004-9541%2806%2960119-5http://dx.doi.org/10.1021/ef0000274http://dx.doi.org/10.1002/aic.690430319http://dx.doi.org/10.1201/9780203913505http://dx.doi.org/10.1201/9780203913505http://dx.doi.org/10.1002/9783527666836

Developing a Steady-state Kinetic Model for Industrial Scale Semi-Regenerative Catalytic Naphtha...

Documents

Transcript of Developing a Steady-state Kinetic Model for Industrial Scale Semi-Regenerative Catalytic Naphtha...