ontr

dP, UK

Article history:Received 3 August 2010Received in revised form 17 January 2011Accepted 18 January 2011Available online 3 February 2011

Keywords:HydrodenitrogenationHydrodemetallizationTrickle bed reactorMathematical modelling

capacity for increasing the distillates production and to removethe impurities such as sulfur, nitrogen, metals (Ni and V) andasphaltenes [2].

The presence of nitrogen compounds in crude oil or oil fractionshas a detrimental effect for rening industries. Nitrogen com-pounds are responsible for catalyst poisoning and reducing catalystactivity. Furthermore, nitrogen compounds have toxic effects on

residue.The metallic compounds in crude oil have also been of great

interest to researchers in this area because of the problems causedby these compounds. The existence of metallic compounds incrude oil and its fractions has harmful effects. These compoundshave a very bad inuence on the HDT efciency, plug the poresof catalysts used, cause rapid deactivation for the hydroprocessingcatalyst, where they tend to deposit on the catalyst, and seem toact to reduce HDT activity by decreasing catalyst surface area[68]. Also, the presence of vanadium and nickel in addition to ironand copper affects the activity of cracking catalysts and causing an

Corresponding author. Fax: +44 (0)1274 235700.

Fuel 90 (2011) 21652181

Contents lists availab

ue

.eE-mail address: I.M.Mujtaba@brad.ac.uk (I.M. Mujtaba).1. Introduction

The technologies for upgrading petroleum fractions are some ofthe most important processes in the rening industry because ofthe growing market demands for different crude oil derivativesand decreasing availability of light oils [1]. Therefore, it is essentialto increase the productivity of distillates with high quality. Amongthese technologies, hydrotreatment operation, which has the

the storage stability of oil products and affect the colour of oilproducts [3]. Andari et al. [4] have shown the impact ofnitrogen and sulfur compounds through their studies of Naphtha,Kerosene and Diesel oils derived from Al-Kuwait crude oil and theyproved that these compounds showed unwanted inuence on thestability of fuel in addition to the environmental pollution. Kaern-bach et al. [5] conrmed that the nitrogen compounds signicantlyaffect the catalyst activity through their works on the vacuumParameter estimation0016-2361/$ - see front matter 2011 Elsevier Ltd. Adoi:10.1016/j.fuel.2011.01.025One of the more difcult tasks in the petroleum rening industries that have not been considered largelyin the literature is hydrotreating (HDT) of crude oil. The accurate calculations of kinetic models of the rel-evant reaction scheme are required for obtaining helpful models for HDT reactions, which can be con-dently used for reactor design, operating and control. In this work, an optimization technique is employedto evaluate the best kinetic models of a trickle bed reactor (TBR) process utilized for hydrodenitrogen-ation (HDN) and hydrodemetallization (HDM) that includes hydrodevanadization (HDV) and hydrode-nickelation (HDNi) of crude oil based on pilot plant experiments. The minimization of the sum of thesquared errors (SSE) between the experimental and estimated concentrations of nitrogen (N), vanadium(V) and nickel (Ni) compounds in the products is used as an objective function in the optimization prob-lem to determine the kinetic parameters.A series of experimental work was conducted in a continuous ow isothermal trickle bed reactor, using

crude oil as a feedstock and the commercial cobaltmolybdenum on alumina (CoMo/c-Al2O3) as a cat-alyst.A three-phase heterogeneous model based on twolm theory is developed to describe the behaviour

of crude oil hydroprocessing in a pilotplant trickle bed reactor (TBR) system. The hydroprocessing reac-tions have been modelled by power law kinetics with respect to nitrogen, vanadium and nickel com-pounds, and with respect to hydrogen. In this work, the gPROMS (general PROcess Modelling System)package has been used for modelling, simulation and parameter estimation via optimization. The modelsimulations results were found to agree well with the experiments carried out in a wide range of thestudied operating conditions. The model is employed to predict the concentration proles of hydrogen,nitrogen, vanadium and nickel along the catalyst bed length in three phases.

2011 Elsevier Ltd. All rights reserved.a r t i c l e i n f o a b s t r a c tKinetic model development and simulatiand hydrodemetallization of crude oil in

Aysar T. Jarullah, Iqbal M. Mujtaba , Alastair S. WooSchool of Engineering, Design and Technology, University of Bradford, Bradford BD7 1D

F

journal homepage: wwwll rights reserved.of simultaneous hydrodenitrogenationickle bed reactor

le at ScienceDirect

l

l sevier .com/locate / fuel

el 9Nomenclature

a dimensionless numberaL gasliquid interfacial area, cm1

aS liquidsolid interfacial area, cm1

AC surface area, cm2

A0j pre-exponential factor for reaction j, (mol/cm3)1n (cm3/g s) (mol/cm3)m

API American Petroleum InstituteCLH2 concentration of hydrogen in the liquid phase, mol/cm

3

CL concentration of i compound in the liquid phase, mol/

2166 A.T. Jarullah et al. / Fuincrease in the level of coal deposited. Also, the presence of thesecompounds, especially vanadium in the fuel used in the highpower machines as gaseous turbines lead to the formation of somesediment on the turbine, which can lead to the change in balance[9,10]. Furthermore, the ash resulting from the combustion of fuelscontaining sodium and particularly vanadium reacts with refrac-tory furnace linings to lower their fusion points and hence causetheir destruction [3].

The process of crude oil hydrotreating is a new challenge andnew technology which has not been considered previously, whereall hydrotreating processes are carried out on each oil cuts sepa-rately, and not on the full crude oil (i.e. after the separation ofcrude oil to its derivatives, such as gasoline, kerosene, light andheavy gas oil). This means that a large amount of the impurities,

icm3

CSH2 concentration of H2 in the solid phase, mol/cm3

CSi concentration of i compound in the solid phase, mol/cm3

Dei effective diffusivity of i compound in the pores of cata-lyst, cm2/s

dc diameter of cylindrical catalyst particle, cmDLH2 molecular diffusivity of H2 in the liquid, cm

2/sDK Knudsen diffusivity, cm2/sDLi molecular diffusivity of i compound in the liquid, cm

2/sDR reactor diameter, cmds diameter of spherical catalyst particle, cmEAj activation energy for j process, J/molGL liquid mass velocity, g/cm2 shH2 Henrys coefcient for hydrogen, MPa cm

3/molkj reaction rate constant for j reaction, (mol/cm3)1n (cm3/

g s) (mol/cm3)m

KLH2 gasliquid mass transfer coefcient for hydrogen, cm/sKSH2 liquidsolid mass transfer coefcient for H2, cm/sKSi liquidsolid mass transfer coefcient for i compound,

cm/sL length of particle, cmLc length of cylindrical catalyst particle, cmLHSV liquid hourly space velocity, h1

mj order of reaction of hydrogen in reaction jnj order of reaction of i compound in reaction jMw molecular weight, kg/kg molep reactor total pressure, psiaPGH2 partial pressure of hydrogen, MPar particle radius, cmrg pore radius, cmrj chemical reaction rate of j reaction per unit mass of the

catalyst, mol/g s1

R universal gas constant, J/mol KSg specic surface area of particle, cm2/gSp total geometric external area of particle, cm2

SSE sum of square errorsSp.gr15.6 specic gravity of oil at 15.6 CT reaction temperatureTmeABP mean average boiling point, RTBR trickle bed reactorug velocity of the gas, cm/suL velocity of the liquid, cm/sv volume, cm3

Vg pore volume per unit mass of catalyst, cm3/gVH2 molar gas volume of H2 at standard conditions, Nl/molVp total geometric volume of catalyst, cm3

z reactor bed length, cm

0 (2011) 21652181namely, sulfur, nitrogen, metals, aromatics and asphaltenes willbe deposited at the bottom of the atmospheric and vacuum distil-lation column. In addition, hydrotreating process each section sep-arately is fairly easy due to the ability to control the reaction, theknowledge of physical and chemical properties, kind of reactionand its condition. Hydrotreating of crude oil is regarded as a bigand difcult challenge since crude oil involves a lot of compoundsand multiple phases, in addition to difcult structures. Addition-ally, hydrotreating of crude oil in the existence of asphaltenes thatcontain a large amount of these impurities, especially metals thatclose the active sites on the catalyst is one of the more difcultand signicant problems. The expected benets of directly hydro-treating crude oil are increasing of middle distillates productivitydue to conversion of heavy compounds and long molecules that

Greek lettersqB bulk density of the catalyst particles, g/cm3

qL liquid density at process conditions, lb/ft3

q15.6 density of oil at 15.6 C, g/cm3

q20 density of the oil at 20 C, g/cm3

qo density of oil at 15.6 C and 101.3 kPa, lb/ft3

qp particle density, g/cm3

gj catalyst effectiveness factor j reactione void fraction of the catalyst bedlL liquid viscosity at process conditions, mPa.stLC critical specic volume of liquid feedstock, cm

3/moltiC critical specic volume of i compound, ft

3/moltL molar volume of liquid feedstock, cm3/molti molar volume of i compound, cm3/molkH2 solubility coefcient of H2, Nl kg

1 MPa1

DqP pressure dependence of liquid density, lb/ft3

DqT temperature correction of liquid density, lb/ft3

/i Thiele Modulus i compoundh particle porositys tortuosity factor

Superscripts0 degreeG gas phaseH2 hydrogenL liquid phase or gasliquid interfaceS solid phase or liquidsolid interfacei compound (crude oil, H2, N, V or Ni)

Subscripts0 at the rst reactor lengthc cylindricalf at the nal reactor lengthg gasH2 hydrogeni compound (N, V, Ni or H2)j reaction (HDN, HDV or HDNi)L liquids spherical

been developed for describing the behaviour of pilot-plant tricklebed reactors applied to the HDN, HDV and HDNi of crude oil. The

el 9model based on two lm theory and includes correlations for cal-culating mass-transfer coefcients, oil density, Henrys coefcients,solubility of hydrogen, oil viscosity, diffusivity, molar volume, spe-cic surface area, etc. under the operating conditions, using infor-mation presented in the literature [1315].

There are three phases in the reactor: gas phase (hydrogen fre-quently), liquid phase (feedstock oil) and solid xed particles (cat-alyst-bed), where the reactions take place. Trickle bed reactorprocess is marked by the simultaneous existence of gas and liquid,over and through a third catalyst solid phase in a cocurrent owmode [1618].

Mathematical modelling of HDT process is a hard task due tothe complex physical and chemical changes that the feed under-goes along with the mass transfer phenomena in the reaction sys-tem. Kinetic aspects are a major factor of reactor modelling, but inthis case, the conversion of a large amount of nitrogen, vanadiumand nickel compounds made it a huge problem. The followingassumptions were used to create the mathematical models forHDN, HDV and HDNi processes using TBR:

No radial concentrations gradients. Steady-state operation of the reactor. One-dimensional heterogeneous model. Isothermal and constant pressure operation of the reactor.is concentrated in heavy fractions to light compounds as a result ofhydrotreating of crude oil before distillation process. In contrast toconventional processes that are carried out for each fraction sepa-rately, which means that the heavy compounds and long moleculeswill be deposited at the bottom of the atmospheric and vacuumdistillation column, which is difcult to hydrotreating them usingnormal operations and conditions.

Furthermore, the mathematical modelling of the hydrotreatingof crude oil is a hard task in view of the intricate physiochemicalchanges that are undergone in the feed together with transportphenomena and mechanisms of catalyst deactivation in the reac-tion system, the major challenge being the evaluation of the kineticmodels accurately, which can accurately predict the product com-pounds at different process conditions. For HDN, HDV and HDNireactions, the development of such kinetic models is a hard taskaccording to the great variety of structures. Thus, this paper is fo-cused upon calculating the parameters of kinetic models applied tothe hydrodenitrogenation, hydrodevanadization and hydrodenick-elation of crude oil based on detailed experimental data. The mod-els utilized have taken from the literature and the kineticparameters are estimated via minimizing sum of the squared errorbetween experimental data and model prediction. Finally, themodel is used for simulation of the HDN, HDV and HDNi processescarried out using gPROMS software [11].

2. Mathematical model of TBR for HDN, HDV and HDNi reactions

A mathematical model is a set of variables and a set of equa-tions that build relationships among the variables for describingsome aspects of the behaviour of the system under investigation.Process models are very protable. It has been employed for oper-ator training, safety systems design, design of operation as well asoperation control systems designs. The improvement of fastercomputer and advanced numerical methods has enabled modellingand solution of the whole process [12].

In the present study, a threephase heterogeneous model has

A.T. Jarullah et al. / Fu Complete wetting of catalyst. No change of catalyst activity with time (thus the effect of cat-alyst deactivation on kinetic parameter is negligible).i = H2, N, V and Ni.

2.2. Chemical reaction rate

Development of kinetic models for hydrotreating of crude oilreactions is a difcult task owing to the complexities of crude oilcomposition and its analysis. Heteroatoms are found in more thanone form in crude oil, for instance, metals compounds can be foundas porphyrine, vanadyl and non-vanadyl, whereas nitrogen com-pounds are occurred aspyridine, quinoline, isoquinoline, pyrrole, in-dole and carbazole [19,20]. Each form is described by its ownThe solution of Eqs. (1)(3) need surface concentrations ofhydrogen, nitrogen, vanadium and nickel. At steady-state, the com-pounds transported between the liquid phase and the solid phaseare consumed or produced through the chemical reaction. Byequating the liquidsolid interfacial mass transfer of H2, N, V andNi components with their reaction rates, we get the followingequations:

kSH2aS CLH2

CSH2

qBX

gjrj For H2 4

kSi aS CLi CSi

qBgjrj 5

i = N, V and Ni and j = HDN, HDV and HDNi.The above equations can be solved using the boundary condi-

tions at z = 0 as follows:

PGH2 z 0 PGH2

initial 6

CLi z 0 CLi initial 7The required data and available tools with the assumptions formodelling and simulation processes for crude oil hydrotreating aretabulated in Fig. 1.

2.1. Mass balance equations

Mass balance equations in the trickle bed reactor for HDN, HDVand HDNi processes are described with the following set of differ-ential and algebraic equations.

(i) Gas phase

Hydrogen :dPGH2dz

RTug

kLH2aLPGH2hH2

CLH2 !

1

(ii) Liquid phase

Hydrogen :dCLH2dz

1uL

kLH2aLPGH2hH2

CLH2 !

kSH2aS CLH2

CSH2 " #

2The above equations represent the mass balance equations for

the gaseous compounds (H2), while the mass balance equationfor the liquid compounds (nitrogen, vanadium and nickel) can bewritten by equating their liquid-phase concentration gradients totheir mass transfer between the liquid-phase and the solid phase.The mass balance equations can be written as:

dCLidz

1uL

kSi aS CLi CSi

i N; Vand Ni: 3

(iii) Solid phase

0 (2011) 21652181 2167reactivity and complex reaction ways, which are specic to eachfeed. In order to estimate for such complexity of feed, the rate ofchemical reaction is usually lumped into a single power law

el 92168 A.T. Jarullah et al. / Fureaction [21]. TheHDN,HDVandHDNi reactions are irreversible un-der usual operation conditions. HDN, HDV and HDNi reactions aremodelled by the power law models with respect to the concentra-tion of nitrogen, vanadiumandnickel andwith hydrogen as follows:

rj KjCSi njCSH2 mj 8

i = N, V and Ni and j = HDN, HDV and HDNi.Reaction rate constant for HDN, HDV and HDNi reactions (Kj)

can be determined for each reaction using the Arrhenius equationas follows:

Kj A0j exp EAjRT

9

j = HDN, HDV and HDNi.

2.3. Reactor performance

The trickle bed reactor contains a number of parameters: masstransfer coefcients, oil density, oil viscosity, solubility of hydro-gen, diffusivities, effectiveness factor and others. These parametersare estimated using the correlations presented in the literature[15,16,2233] as follows.

The equation used to determine the gasliquid mass transfercoefcient for H2 is:

KLH2aL

DLH2 7 GL

lL

0:4 lLqLD

LH2

!0:510

Fig. 1. Required data and available tools for modeling0 (2011) 21652181GL qLuL 11

The liquidsolid mass transfer coefcients can be calculated bythe Van KrevelenKrekels equation as follows:

KSiDLi aS

1:8 GLaSlL

0:5 lLqLD

Li

!1=312

i = H2, N, V and Ni.In order to determine the liquidsolid and gasliquid mass

transfer coefcients, it is necessary to know the molecular diffusiv-ity of H2, N, V and Ni in the liquid. The diffusivity can be calculatedby a Tyn-Calus equation:

DLi 8:93 108t0:267Lt0:433i

TlL

13

i = H2, N, V and NiThe molar volume of crude oil (L), H2, N, V and Ni can be calcu-

lated by the following equation:

ti 0:285 tiC 1:048 14

i = crude oil, H2, N, V and Ni.The critical specic volume of liquid (crude oil) is estimated by

a RiaziDaubert correlation:

tLC 7:5214 103TmeABP0:2896q15:60:7666

Mw 15

and simulation of HDN, HDV and HDNi reactions.

To determine the values of gi, the following equation is used

el 9Henry coefcients of H2 can be obtained from solubilitycoefcients:

hH2 VH2kH2qL

16

Korsten and Hoffmann [34] have presented the following equa-tion for the solubility of hydrogen in hydrocarbon mixtures:

kH2 0:559729 0:42947 103T 3:07539 103Tq20

1:94593 106T2 0:835783q202 !

17

The oil density (qL) as a function of temperature and pressure isestimated by the Standing-Katz equation:

qL q DqP DqT 18

DqP 0:167 16:181 100:0425qh i

P1000

0:01

0:299 263 100:0603qh i

P1000

219

DqT 0:0133 152:4q DqP2:45h i

T 520

8:1 106 0:0622 100:764qDqP h i

T 5202 20

Glasos equation has used as a generalized mathematical equa-tion for oil viscosity. The equation has the following form:

lL 3:141 1010T 4603:444log10APIa 21

a 10:313log10T 460 36 447 22

API 141:5sp:gr15:6

131:5 23

The surface area of the particles per unit volume of the bed isdescribed as:

as AC1 em

24

For cylindrical particle

as 2prLpr2L

1 e 2r1 e 41 e

d25

The bed void fraction of the catalyst (e) is calculated by the fol-lowing equation. This equation has been developed for packed bedof spheres:

e 0:38 0:073 1DRds

2

2DRds

2264

375 26

For cylindrical particles, the equivalent spherical diameter is gi-ven by the equation:

ds dcLc d2c

2

!" #1=227

The terms of the catalyst effectiveness factor (g) are usually re-ferred to internal diffusion limitations [35]. It was observed that anincreasing in the particle size, the chemical reaction rate decreases.In the literature, the effectiveness factor has been found to be inthe range of 0.00571 [36,37]. Thieles Modulus (/) is utilized for

A.T. Jarullah et al. / Fucalculating the catalyst effectiveness factor (g) because of the par-ticle size of catalyst is small [38]. The generalized Thiele Modulusfor nth-order irreversible reaction is:[37,39,40]:

gj tanh/i

/i35

2.4. Kinetic parameters of the models

The accurate estimation of kinetic parameters of the relevantreactions scheme are required in order to obtain a useful model,which can be condently used for reactor design and process opti-mization. In the model presented above, the reaction orders ofnitrogen, vanadium and nickel compounds (nj), hydrogen com-pound order (mj), reaction rate constants (Kj), activation energies(EAj) and pre-exponential factors A

0j

parameters of Eqs. (8) and

(9) are such signicant parameters for the HDN, HDV and HDNiprocesses. The major focus of this paper is to accurately calculatethese parameters.

3. Parameter estimation techniques

Parameter estimation is required to ensure accurate model pre-dictions and good model based decision. It is a key problem in theimprovement of process models either steady state or unsteadystate, and hence is an important issue in both process design andcontrol. The features and accuracy of the model utilized estimatethe realism with which the actual process can be represented.Using a suitable model is helpful not only in nding optimum oper-ation conditions of the process and in developing process analysis,but also in the control strategies design for the system at processconditions [41]. For the purpose of process optimization, designof reactor, process control and selection of catalyst, it is importantto develop kinetic models that can accurately predict the concen-tration of product under process conditions. For complex hydrocar-/ VPSP

n 12

KCn1As qP

De

!" #0:528

qP qB

1 e 29

For j reaction and i compound, Thiele Modulus can be stated as:

/i VPSP

n 12

KjC

n1i qPDei

!" #0:530

Dei hs1

1DLi 1DK

0@

1A 31

i = N, V and Ni and j = HDN, HDV and HDNi.The tortuosity factor (s) generally has a value of 27 [35].

Generally, the tortuosity factor is assumed to be 4 according toliterature reports [29,35,38].

Knudsen diffusivity factor (DK) is evaluated as follows:

DK 9700rg TMw 0:5

32

rg 2hSgqP33

h qPVg 34

0 (2011) 21652181 2169bon mixtures (like crude oil), the development of such kineticmodels is a difcult task owing to the existence of a huge varietyof structures [42].

B. Non-linear regression to determine nj, mj, EAj and A0j

simultaneously for each process.

Both approaches use the following objective function based onthe minimization of the sum of squared errors (SSE) between theexperimental concentrations of N, V and Ni compound Cexpi;y

and calculated Ccali;y

, in the products:

SSE XNDatay1

Cexpi;y Ccali;y 2 36

where i = N, V and Ni

3.1. Optimization Problem Formulation for Parameter Estimation

The optimization problem can be described as follows:

The sum of squared errors (SSE).

Subjectto

t epresented as:

Min SSEnj;mj;Kj j s: ;

nLj 6 nj 6L 6

j j j

el 90 (2011) 21652181Parameters estimation is necessary in several elds of scienceand engineering as many physiochemical processes are describedby systems equations with unknown parameters. Recently, thebenets of developing kinetic models for chemical engineers withaccurate parameter calculations have increased owing to thedeveloped control technologies and optimization of process, whichcan apply fundamental models. Estimation of kinetic parameters isan important and difcult step in the development of models,Calculations of unknown kinetic parameters can be achieved byutilizing experimental data. When estimating kinetic parametersof the models, the goal is to calculate appropriate parameter valuesso that errors between experimental and theoretical data (based onmathematical model) are minimized. On the other hand, the pre-dicted values from the model should match the experimental dataas closely as possible [43]. The scope of parameter estimation tech-niques is vast, and some of the parameter estimation techniquesare presented here.

Tatiraju and Soroush [44] have used model inversion for param-eter estimation that includes a left inverse of the process modeland at each time instant calculates least-squared error estimatesof parameters by using readily available on-line measurements.The parameter estimation by state estimation technique, is usuallyutilized in chemical and biochemical engineering and needs a dy-namic model for each of the unknown parameters to be deter-mined [4547]. Kinetic parameters in Calorimetric techniquemethod are determined via simple mass and energy balances[4850]. Optimization technique is employed as well for parametercalculation, where sum of squared errors between the measure-ments and estimated values is minimized [51,52]. This techniqueis very popular and has largely been utilized in the chemical oper-ations to evaluate many kinetic parameters.

In recent years, several optimization techniques have beendeveloped, related to the parameter estimation problems. Amongthem, the regularization methods, the augmented Lagrangianmethods and the level set approaches have been improved in orderto approximate discontinuous parameters and to reduce the sensi-tivities of the optimization schemes [53]. Linear or non-linearregressions are utilized to estimate kinetic rate constant frommea-sured rates and concentrations, and several computer programsare available for this issue [54,55]. Despite the popularity of thelinear programming methods due to their ability to handle severalparameters, they cannot be applied to any operation owing to therequirements of linear objective function in terms of the adjustableparameters [55]. Non-linear optimization is the most popular en-tity and is commonly employed for calculating the best values ofkinetic parameters. Many methods can be employed for non-linearoptimization techniques, such as Stochastic methods [56],maximum likelihood estimation [57], Newton method [58],LevenbergMarquardt method [59,60], Genetic algorithm (GA)[6164], evolutionary algorithm [65], adaptive GA [66], differentialevolution [67] and Successive Quadratic Programming [68] havebeen reported widely to obtain the parameter estimation prob-lems. For HDT processes, the LevenbergMarquardt method andSuccessive Quadratic Programming (SQP) methods are particularlysuitable [33,38,6973].

In order to evaluate the best values of kinetic parameters in thisstudy, two approaches have been employed depending on thenitrogen, vanadium and nickel content in hydrotreated productsunder varies operating conditions. They are as follows:

A. Non-linear regression to obtain simultaneously the reactionorders of nitrogen, vanadium and nickel compound (n),hydrogen compound order (m) and reaction rate constants

2170 A.T. Jarullah et al. / Fu(Kj), then linear regression with Arrhenius equation to esti-mate activation energy (EAj) and pre-exponential factorA0j

for each process.Table 1Properties of commercial catalyst (CoMo/c-Al2O3).

Chemical specicationNiO (wt.%) 3MoO3 (wt.%) 15Na2O (wt.%) 0.07SiO2 (wt.%) 1.1SO2 (wt.%) 2Fe (wt.%) 0.04Al2O3 Balance

Physical specicationForm ExtrudePore volume (cm3/g) 0.5Surface area (m2/g) 180Mean particle diameter (mm) 1.8mj 6 mjKiL 6 Ki 6nUj inequality constraintsmUj inequality constraintsKiU ; i 1;2;3 inequality constraintst: f z; xzically, using the rst approach, the problem can b

HDN;HDV ;HDNixz;uz; v 0; z0; zf model; equality constraintsMathemaProcess constraints and linear bounds on alloptimization variables in the process.So as tominimize(mj), the activation energy (EAj) and pre-

exponential factor A0j

(for each reaction).Given Reactor conguration, the feedstock, the catalyst,reaction temperature, hydrogen pressure andliquid hourly space velocity

Optimize For the rst approach: the reaction order of N, Vand Ni compounds (nj), hydrogen compound order

(mj) and reaction rate constants K1j ;K

2j ;K

3j

for

each reaction at different temperatures (335 C,370 C, 400 C, respectively)For second approach: the reaction order of N, Vand Ni compound (nj), hydrogen compound orderMean particle length (mm) 4Bulk density (g/cm3) 0.67

Fig. 2. General scheme of the hydrotreating pilot-plant unit.

Table 2Experimental data for HDN of crude oil and simulation data of the pilot plant-TBR using two approaches (linear and non-linear regression).

Operating conditions Experimental results Simulated results

LHSV(h1)

P(MPa)

T(C)

Nitrogen(wt.%) 104

Conversion(%)

Nitrogen (wt.%)-nonlinearregression 104

Conversion(%)

Absoluteerror%

Nitrogen (wt.%)-linearregression 104

Conversion(%)

Absoluteerror%

0.5 4 335 526.9 47.31 519.623 48.04 1.38 515.199 48.48 2.221 4 335 713.2 28.68 698.148 30.18 2.11 695.499 30.45 2.481.5 4 335 783.5 21.65 780.410 21.19 0.39 778.546 22.14 0.630.5 4 370 327.9 67.21 322.029 67.79 1.79 317.414 68.26 3.201 4 370 512.7 48.73 518.731 48.13 1.18 515.319 48.47 0.511.5 4 370 618.9 38.11 629.256 37.07 1.67 626.614 37.34 1.250.5 4 400 203.7 79.63 194.645 80.53 4.44 190.118 80.99 6.671 4 400 362.8 63.72 369.570 63.04 1.87 365.107 63.49 0.631.5 4 400 485.3 51.47 487.651 51.23 0.48 483.610 51.64 0.350.5 7 335 472.5 52.75 462.619 53.74 2.10 461.368 53.86 2.351 7 335 649.0 35.10 650.617 34.94 0.25 650.718 34.93 0.261.5 7 335 738.1 26.19 741.917 25.81 0.52 742.335 25.77 0.570.5 7 370 261.0 73.90 270.200 72.98 3.52 268.583 73.14 2.901 7 370 449.1 55.09 461.001 53.90 2.65 460.959 53.90 2.641.5 7 370 571.9 42.81 575.914 42.41 0.70 576.447 42.35 0.790.5 7 400 155.3 84.47 155.760 84.42 0.29 153.690 84.63 1.041 7 400 305.5 69.45 313.755 68.62 2.70 312.679 68.73 2.351.5 7 400 418.6 58.14 428.905 57.11 2.46 428.413 57.16 2.340.5 10 335 424.9 57.51 425.902 57.41 0.23 426.672 57.33 0.421 10 335 605.3 39.47 618.017 38.20 2.10 620.101 37.99 2.441.5 10 335 687.9 31.21 714.779 28.52 3.91 716.934 28.31 4.220.5 10 370 234.2 76.58 239.307 76.07 2.18 239.331 76.07 2.191 10 370 418.5 58.15 423.793 57.62 1.26 425.896 57.41 1.771.5 10 370 518.3 48.17 540.139 45.99 4.21 542.859 45.71 4.740.5 10 400 130.5 86.95 133.936 86.61 2.63 133.082 86.69 1.981 10 400 274.4 72.56 279.811 72.20 1.97 280.665 71.93 2.281.5 10 400 387.1 61.29 391.446 60.85 1.12 393.177 60.68 1.57

Model prediction

P(MPa)

T(C)

Nitrogen content-wt.% 104 (using 2nd approach) Conversion%

1.0 10 385 347.610 65..240.75 10 400 211.025 78.891.25 4 370 580.986 41.900.5 7 385 206.803 79.321.5 5.5 335 759.106 24.09

A.T. Jarullah et al. / Fuel 90 (2011) 21652181 2171

Using the second approach, the problem can be expressed as:

Min SSEnj;mj; EAj;A

0j j HDN;HDV ;HDNi

s:t: f z; xz; xz;uz; v 0; z0; zf model; equality constraintsnLj 6 nj 6 nUj inequality constraintsmLj 6 mj 6 mUj inequality constraintsEALj 6 EAj 6 EA

Uj inequality constraints

A0Lj 6 A0j 6 A

0Uj inequality constraints

f z; xz; xz;uz;v 0, represents the process model presented inSection 2, where z is the independent variable (length of the bedreactor), x(z) gives the set of all differential and algebraic variables,xz denotes the derivative of differential variables with respect tolength of the bed reactor, u(z) is the control variables, and v repre-sents the design variables (length independent constant parame-ters). The length interval of interest is [z0, zf] and the function f: isassumed to be continuously differentiable with respect to all itsarguments [74,75].

The solution method of optimization employed by gPROMS is atwo-step method known as feasible path approach. The rst stepperforms the simulation to converge all the equality constraints(described by f) and to satisfy the inequality constraints. The sec-ond step performs the optimization (updates the values of thedecision variables such as the kinetic parameters) [76]. The optimi-zation problem is posed as a Non-Linear Programming (NLP)

problem and is solved using a Successive Quadratic Programming(SQP) method within gPROMS software.

4. Experimental work

4.1. Materials

Iraqi crude oil has been used as a feed for hydrotreating studies.It contains 2.0 wt.% of sulfur, 0.1 wt.% of nitrogen, 26.5 ppm ofvanadium, 17 ppm of nickel and 1.2 wt.% of asphaltene. The cata-lyst used for the HDN, HDV and HDNi processes in this work wasthe commercial cobaltmolybdenum on alumina (CoMo/c-Al2O3) type catalyst. The properties of the catalyst used are listedin Table 1.

4.2. Equipment and procedure

A schematic diagram of the hydrotreating pilot plant is shownin Fig. 2. Generally, the pilot plant can be divided into four sec-tions: the feed section, the reactor section, the products sectionsand gases section.

The feed supply module primarily includes a liquid feed tankand a feed pump. A cylindrical tank with a capacity of 2 l of thefeedstock is the feed tank. In order to introduce the feed oil intothe reactor, a high-pressure dosing pump has employed for thispurpose. The feedstock and hydrogen passes through the reactorin a concurrent ow mode. The length and the diameter of thereactor were 2 cm and 65 cm, respectively, and the reactor tube

Table 3Experimental data for HDV of crude oil and simulation data of the pilot plant-TBR using two approaches (linear and non-linear regression).

Operating conditions Experimental results Simulated results

LHSV P T Vanadium Conversion Vanadium(ppm)-nonlinear Conversion Absolute Vanadium (ppm)-linear Conversion Absolute

0.5 10 335 3.84 85.51 3.9996

M

V

2172 A.T. Jarullah et al. / Fuel 90 (2011) 216521811.0 10 3850.75 10 4001 10 335 9.42 64.45 9.70131.5 10 335 13.23 50.00 13.35930.5 10 370 1.98 92.53 1.88831 10 370 6.63 74.98 6.31051.5 10 370 10.27 61.24 9.88070.5 10 400 0.891 96.64 0.89081 10 400 3.95 85.10 4.03291.5 10 400 6.89 74.00 7.1805

P(MPa)

T(C)(h1) (MPa) (C) (ppm) (%) regression

0.5 4 335 8.33 68.57 8.69971 4 335 14.34 45.89 14.87271.5 4 335 17.79 32.87 17.94050.5 4 370 5.54 79.09 5.42271 4 370 11.70 55.85 11.50111.5 4 370 15.50 41.51 15.03540.5 4 400 3.21 87.89 3.32301 4 400 8.40 68.30 8.74681.5 4 400 12.10 54.34 12.43260.5 7 335 5.60 78.87 5.66211 7 335 11.80 55.47 11.77331.5 7 335 15.10 43.02 15.27850.5 7 370 3.15 88.11 3.01281 7 370 8.66 67.32 8.26861.5 7 370 12.50 52.83 11.95140.5 7 400 1.59 94.00 1.58741 7 400 5.46 79.40 5.70261.5 7 400 9.13 65.55 9.19821.25 4 370 10.5 7 3851.5 5.5 335 15.0913 80.792.3578 91.10(%) error% regression (%) error%

67.17 4.44 8.7776 66.88 5.3743.88 3.71 14.9471 43.60 4.2332.30 0.84 18.0027 32.06 1.1979.54 2.12 5.4468 79.45 1.6856.60 1.70 11.5347 56.47 1.4143.26 2.99 15.0676 43.14 2.7987.46 3.52 3.3159 87.49 3.3066.99 4.13 8.7436 67.00 4.0953.08 2.75 12.4320 53.09 2.7478.63 1.11 5.7592 78.27 2.8455.57 0.23 11.8885 55.14 0.7542.34 1.18 15.3831 41.95 1.8788.63 4.35 3.0503 88.49 3.1668.80 4.52 8.3341 68.55 3.7654.90 4.39 12.0203 54.64 3.8494.01 0.16 1.5954 93.98 0.3478.48 4.44 5.7279 78.38 4.9165.29 0.75 9.2308 65.17 1.1084.91 4.16 4.0972 84.54 6.7063.39 2.99 9.8389 62.87 4.4549.59 0.98 13.4929 49.08 1.9992.87 4.63 1.9251 92.73 2.7776.19 4.82 6.3899 75.89 3.6262.27 3.79 9.9718 62.37 2.9096.64 0.02 0.9009 96.60 1.1184.78 2.10 4.0692 84.64 3.0272.90 4.22 7.2312 72.71 4.95

odel prediction

anadium content-ppm (using 2nd approach) Conversion%3.4918 49.092.2116 91.656.4829 37.80

was made of stainless steel. The length of the reactor has beendivided into three sections. The rst section, having a length of20 cm, was packed with inert particles (glass beads with 4 mmdiameter). This entrance section has been used to heat up the mix-ture to the required temperature, to ensure homogeneous ow dis-tribution of gas and liquid and to avoid end effects. The followingsection with a length of 27.8 cm contained a packing of 60.3 g cat-alyst. The bottom part (17.2 cm) was packed with inert particles toensure to serve as disengaging section. The reactor was operated inisothermal mode by independent temperature control of ve zoneelectric furnaces, which provided an isothermal temperature alongthe active reactor section.

The product part includes of low and high gasliquid separatorand products storage tank. The reactor outlet is led to the highpressure separator where the liquid and gas are separated. Finally,the gases section where the gas is exiting from is passed through agas ow meter before being released.

Before starting up any run, a leak test must be conducted. Theleak test is done with nitrogen (N2) at 13 MPa for 12 h. Once theunit passes the leak tests, the catalyst presulding process (see fur-ther details on experimental procedure and presulding in Jarullahet al. [77]) will start. Calibration has been carried out on all pilotplant equipments (such as pump, instrumentation and control)for ensuring the measurements accuracy.

4.3. Experimental runs

The main hydrotreating reactions in this work are hydrodeni-trogenation (HDN) and hydrodemetallization (HDM) that includes

hydrodevanadization (HDV) and hydrodenickelation (HDNi) reac-tions. The data obtained from these experiments were used inthe development of models that can represent the HDN, HDV andHDNi reactions to determine kinetic parameters and to validatethe model under different operating conditions.

Note, all analytical techniques that have been used for the spec-ications of the feedstock and the products were accurate, fast andrepeatable. Product analysis has been repeated twice for each runat each operating condition for ensuring the accuracy of the results.Average results have been taken into considerations for each sam-ple with maximum deviation of 2% among all runs. IP-285 methodwas used to calculate the vanadium and nickel content in the feed-stock and the products samples. While the nitrogen content in thefeedstock and products were estimated by using ASTM: D-4629method.

5. Results and discussions

5.1. Experimental results

Experimental works for hydrodenitrogenation, hydrodevanadi-zation and hydrodenickelation of crude oil were carried out underthe following operating conditions:

Reaction temperature: 335400 C. Liquid hourly space velocity (LHSV): 0.51.5 h1. Hydrogen pressure: 410 MPa. H2/Oil ratio: 250 l/l.

Table 4Experimental data for HDNi of crude oil and simulation data of the pilot plant-TBR using two approaches (linear and non-linear regression).

Operating conditions Experimental results Simulated results

Con(%)

88.72.

1.5 4 335 6.80 60.00 6.7673 60.0.5 4 370 1.11 93.47 1.1222 93.

81.70.95.86.77.93.80.69.96.33 0.52 0.6183 96.36 0.2787.83 2.42 2.0635 87.86 2.1579.00 1.99 3.5668 79.02 1.9197.91.84.95.85.75.97.91.83.98.94.88.

A.T. Jarullah et al. / Fuel 90 (2011) 21652181 21731 4 370 3.10 81.76 3.20161.5 4 370 4.95 70.88 5.04900.5 4 400 0.73 95.70 0.69481 4 400 2.28 86.59 2.24981.5 4 400 3.85 77.35 3.80100.5 7 335 1.15 93.23 1.15261 7 335 3.22 81.06 3.25961.5 7 335 5.10 70.00 5.11860.5 7 370 0.62 96.35 0.62321 7 370 2.02 88.12 2.06891.5 7 370 3.50 79.41 3.56990.5 7 400 0.36 97.88 0.36731 7 400 1.33 92.18 1.37081.5 7 400 2.48 85.41 2.55080.5 10 335 0.81 95.23 0.79951 10 335 2.50 85.29 2.49181.5 10 335 4.07 76.06 4.14160.5 10 370 0.42 97.53 0.41531 10 370 1.48 91.29 1.50751.5 10 370 2.71 84.06 2.75840.5 10 400 0.24 98.59 0.23801 10 400 0.93 94.53 0.96241.5 10 400 1.87 89.00 1.8978

P(MPa)

T(C)

1.0 10 3850.75 10 400LHSV(h1)

P(MPa)

T(C)

Nickel(ppm)

Conversion(%)

Nickel (ppm)-nonlinearregression

0.5 4 335 2.01 88.18 1.93321 4 335 4.80 71.76 4.68361.25 4 3700.5 7 3851.5 5.5 33584 2.03 0.3680 97.83 2.2294 3.07 1.3784 91.89 3.6499 2.85 2.5667 84.90 3.4930 1.29 0.7847 95.38 3.1234 0.33 2.4649 85.50 1.4064 1.76 4.1103 75.82 0.9956 1.12 0.4123 97.57 1.8313 1.86 1.5046 91.15 1.6677 1.78 2.7583 83.77 1.7860 0.83 0.2388 98.59 0.5034 3.48 0.9690 94.30 4.1984 1.49 1.9127 88.75 2.28

Model prediction

Nickel content-ppm (using 2nd approach) Conversion %

1.2064 92.900.5562 96.73version Absoluteerror%

Nickel (ppm)-linearregression

Conversion(%)

Absoluteerror%

63 3.82 1.9001 88.82 5.4745 2.42 4.6368 72.72 3.4019 0.48 6.7195 60.47 1.1840 1.10 1.1128 93.45 0.2517 3.28 3.1903 81.23 2.9130 2.00 5.0393 70.36 1.8091 4.82 0.6950 95.91 4.7976 1.32 2.2577 86.72 0.9859 1.27 3.8353 77.44 0.3822 0.23 1.1318 93.34 1.5882 1.23 3.2251 81.03 0.1689 0.36 5.0804 70.11 0.384.1712 75.460.4783 97.195.8171 65.78

el 9200

300

400

500

600

700

800

900

gen

cont

ent i

n pr

oduc

t (pp

m)

(a)

2174 A.T. Jarullah et al. / FuThe experimental results for HDN, HDV and HDNi processes intabular form are shown in Tables 24 (also in Figs. 35 with modelpredictions).

It has been observed from the experimental results (Tables 24,also in Figs. 35) that the nitrogen, vanadium and nickel content inall products decreased with increasing in temperature and pres-sure and decreasing in liquid hourly space velocity. Similar attitudehas also been found by several investigations for HDN, HDV andHDNi processes using different oily feedstocks (but not on the fullcrude oil) [6,33,7880].

The increasing in nitrogen and metals removal at high temper-ature may be attributed to several reasons: at high reaction tem-perature, the unreactive nitrogen and metal compounds or thecompounds containing these impurities become activated enough

0

100

200

300

400

500

600

700

800

0.4 0.6 0.8

0

100

0.4 0.6 0.8

LHS

LHSV

LHS

Nitr

oN

itrog

en c

onte

nt in

pro

duct

(ppm

)N

itrog

en c

onte

nt in

pro

duct

(ppm

)

0.4

100.4

200.4

300.4

400.4

500.4

600.4

700.4

800.4

0.4 0.6 0.8

(c)

(b)

Fig. 3. Experimental data (points) and simulated (lines) variation of outlet nitrogen cont(a) 4 MPa, (b) 7 MPa, (c) 10 MPa.T=608K SimulatedT=643K Simulated

0 (2011) 21652181to react with hydrogen. Also, the large molecules are decomposedinto smaller molecules, which can more easily diffuse inside thecatalyst pores and reach the inner active sites where the HDT reac-tions occur. Oil diffusivity increases through the catalyst pores dueto decreases in the oil viscosity. Furthermore, the increase in tem-perature raises the activation energy leading to increase the num-ber of particles of these compounds interacted. As a result, thelong-nitrogen and metal compounds will cleavage and spreadwithin the catalyst [6,81]. As liquid hourly space velocity de-creased, denitrogenation and demetallization increase because ofthe contact time (residence time) increases between the moleculesof reactants and catalyst, and provide sufcient time for the reac-tion process [6,82,83]. Whereas the reason for increasing in nitro-gen, vanadium and nickel removal by increasing in hydrogen

1 1.2 1.4 1.6

T=608K SimulatedT=643K SimulatedT=673K Simulated

1 1.2 1.4 1.6

V (hr-1)

(hr-1)

V (hr-1)

T=673K Simulated

1 1.2 1.4 1.6

T=608K SimulatedT=643K SimulatedT=673K Simulated

ent vs. liquid hourly space velocity at different reactor temperature and at pressure

el 9m c

onte

nt in

pro

duct

(ppm

)6

8

10

12

14

16

18

20

(a)

A.T. Jarullah et al. / Fupressure due to the contact between the hydrogen and hydrocar-bons and the catalyst [79].

It is interesting to note that the conversion of HDNi is found tobe higher than HDV (Tables 3 and 4). These observations althoughnot reported in the public domain, are not uncommon in industries[84,85]. This could be due to varying composition and properties ofdifferent crude oils and due to variation in the actual amount of Niand V present in these crude oils.

5.2. Estimation of kinetic parameters

The kinetic parameters for crude oil hydrodenitrogenation, hyd-rodevanadization and hydrodenickelation presented in the present

Vana

diu

0

2

4

6

8

10

12

14

16

Vana

dium

con

tent

in p

rodu

ct (p

pm)

0

2

4

6

8

10

12

14

16

18

0.4 0.6 0.8 1

0.4 0.6 0.8

Vana

dium

con

tent

in p

rodu

ct (p

pm)

0.4 0.6 0.8

LHS

LHSV

LHSV

0

2

4

(c)

(b)

Fig. 4. Experimental data (points) and simulated (lines) variation of outlet vanadium con(a) 4 MPa, (b) 7 MPa, (c) 10 MPa.T=608K SimulatedT=643K Simulated

0 (2011) 21652181 2175work have been estimated depending on the experimental datausing TBR model.

In the rst approach, the reaction order of N, V and Ni(nj, j=HDN, HDV, HDNi), hydrogen order (mj) and reaction rate constantsKij; i 1;2;3

for each reaction were estimated simultaneously.Linearization process is then used for estimating the activation en-ergy (EAj) and the pre-exponential factor A

0j

for each reaction. To

estimate activation energies and pre-exponential factors for eachreactions, the Arrhenius equation described previously (Eq. (9)) isused for this purpose. The Arrhenius-based dependence of the ki-netic model is demonstrated in Fig. 6 for all processes. Plot of ln Kjversus 1/T gives a straight line with a slope equal to EAj/R andintercept equal to lnA0j . In the second approach, the activation

T=673K Simulated

T=608K SimulatedT=643K SimulatedT=673K Simulated

1.2 1.4 1.6

1 1.2 1.4 1.6

T=608K SimulatedT=643K SimulatedT=673K Simulated

1 1.2 1.4 1.6

V (hr-1)

(hr-1)

(hr-1)

tent vs. liquid hourly space velocity at different reactor temperature and at pressure

atedatedated

el 9icke

l con

tent

in p

rodu

ct (p

pm)

T=608K SimulT=643K SimulT=673K Simul

1

2

3

4

5

6

7

8

2176 A.T. Jarullah et al. / Fuenergies (EAj), pre-exponential factors A0j

, reaction order of N, V

and Ni (nj) and hydrogen order (mj) were determined simulta-neously. The generated kinetic parameters for HDN, HDV and HDNiprocesses are presented in Table 5 for both approaches,respectively.

It is has been observed from Table 5 that the second approachgives sum of squared errors (SSE) less than rst approach in allreactions. It can be concluded depending on the objective function(SSE) that parameter estimation with the second approach is moreaccurate. In other words, determine of activation energy and pre-exponential factor by linearization process of Arrhenius equationgives higher error in comparison to those obtained with simulta-neous estimation of kinetic parameters via non-linear regression.The best reaction order of nitrogen, vanadium and nickel were

0.4 0.6 0.8

N

0.4 0.6 0.8 1

Nic

kel c

onte

nt in

pro

duct

(ppm

)

T=608K SimulatedT=643K SimulatedT=673K Simulated

0.4 0.6 0.8 1

Nic

kel c

onte

nt in

pro

duct

(ppm

)

T=608K SimulaT=643K SimulaT=673K Simula

0

LHSV

LHSV

LHSV

0

1

2

3

4

5

6

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Fig. 5. Experimental data (points) and simulated (lines) variation of outlet nickel conten4 MPa, (b) 7 MPa, (c) 10 MPa.(a)

0 (2011) 216521811.6723, 1.2514 and 1.6884, respectively, while, the best order ofH2 for HDN, HDV and HDNi were 0.3555, 0.6337 and 0.5667,respectively, which is typical for lumped kinetics. Several investi-gators have extensively studied the reaction kinetics of hydropro-cessing processes of several distillate cuts (not of crude oil which isthe focus of this work) and showed that most HDT reactions fol-lows half to second order kinetics and zero to one order kineticsfor hydrogen [15,30,33,78,80,86,87]. Note, Jarullah et al. [88] hada similar observation in estimating kinetic parameters for hydrode-sulfurization of crude oil.

The established kinetic parameters for HDN, HDV and HDNireactions in the present study, has been applied to simulate theperformance of a pilot plant trickle bed reactor. The partial pres-sure of hydrogen, molar concentration proles of N, V and Ni and

1 1.2 1.4 1.6

1.2 1.4 1.6

1.2 1.4 1.6

tedtedted

(c)

(b)

(hr-1)

(hr-1)

(hr-1)

t vs. liquid hourly space velocity at different reactor temperature and at pressure (a)

13

5

7

9

11

13

1.46 1.48 1.5 1.52 1.54 1.56

1000/T

lnk

(HD

N &

HD

Ni)

lnk

(HD

V)

1

1.5

2

2.5

3

3.5

4

4.5

5

Fig. 6. Linear representation of Arrhenius equa

Table 5Comparison of kinetic parameter values estimated with two approaches for HDN,HDV and HDNi models.

HDN HDV HDNi

First approach (linear)

Kij ; i 1 9.06245 11.6142 51827.2Kij ; i 2 20.0872 17.5322 77768.1Kij ; i 3 34.3759 28.8485 103539m 0.3325 0.6267 0.5633n 1.6302 1.2482 1.6819EAj 69973.95 47172.80 36288.95

A0j 9435596.91 127388.9139 0.683 108

SSE 3.5442 106 0.24521638 0.01000841Second approach (non-linear)m 0.3555 0.6337 0.5667n 1.6723 1.2514 1.6884EAj 71775.5 46181.6 37678.3

A0j 2.85 107 126566.0 1.045 108

SSE 2.8957 106 0.2225156 0.007773

100

200

300

400

500

600

700

800

100 200 300 400

Nitrogen (ppm

Nitro

gen

(ppm

)- Si

mul

ated

Fig. 7. Comparison between experimental an

A.T. Jarullah et al. / Fuel 90 (2011) 21652181 2177H2 in the liquid phase and in the solid phase were calculated usingthe correlations given earlier, which contain a number ofparameters.

A comparison between experimental results and model predic-tion results for HDN, HDV and HDNi of crude oil were plotted inFigs. 35 (using second approach, where the experimental dataare represented in the form of points; while the simulation resultsare represented in the form of curves (each curve representingthree simulated points)), and given in Tables 24 (for both ap-proach). As can be noticed from the results, the model was foundto simulate the performance of the pilot plant TBR very good agree-ment in the range of operating conditions studied between both

1.58 1.6 1.62 1.64 1.66

(K)

0

0.5

tion for HDN, HDV and HDNi of crude oil.concentrations with average absolute error less than 5% by usingsecond approach. It has also been noted from these gures thatthe nitrogen, vanadium and nickel removal increase with increas-ing in temperature, pressure and decreasing in liquid hourly spacevelocity. These increases happened due to the kinetics parametersused to describe HDN, HDV and HDNi processes in this model areaffected by the operating conditions. The reaction temperature of

500 600 700 800

)- Experimental

d calculated concentrations of nitrogen.

the reactor impacts upon the mass velocity of the gases and liquids,diffusivities of the components, mass transfer coefcient at thegasliquid and at the liquidsolid interfaces, solubility and Henryscoefcients of hydrogen in addition to viscosity and density of thecompounds.

The temperature also inuences the rate constant of HDN, HDVand HDNi processes. Increasing the reaction temperature lead to anincrease in reaction rate constants dened by the Arrhenius equa-tions. As a result, the reaction rates of these reactions will increase.These results supported the fact that the operating temperature isvery effective for enhancing the degree of denitrogenation anddemetallization.

Liquid hourly space velocity (LHSV) is also a signicant opera-tional factor that calculates the severity of reaction and the ef-ciency of hydrotreating. As the liquid hourly space velocitydecreased, the quantity of the reactions rates will be signicant.On the other hand, decreasing liquid hourly space velocity de-scribed by liquid velocity, means increasing the residence timeand hence the reaction severity will increase [6,16,83,89].

In addition, the hydrogen partial pressure has an effect on thereactions used in this study. The mechanisms utilized to describeHDN, HDV and HDNi reactions used a kinetic equation with theorder of the hydrogen concentration at the catalyst surface less than1. Therefore, conversion of N, V and Ni compounds increases with

pressure. The effect of pressure above 100 atm can be neglectedowing to the viscosity of the oil feedstock increase with pressure,and the diffusivity and mass transfer coefcient will decrease withthe pressure. Hence, at high operating pressures, the pressure im-pact upon N, V and Ni conversions becomes important [34,83,90].

Figs. 79 show a parity plot of the model for HDN, HDV andHDNi reactions studied in this study (each point represents simu-lated (Y-axis) and experimental (X-axis) values at the same timewith the same operating conditions for each point). The plot be-tween the experimental results and simulated nitrogen, vanadiumand nickel contents in all products appears to be straight line witha slope close to 1.0, indicating very good agreement between themeasurement results and the predicted. The model can be usedto describe the behaviour of the pilot plant trickle bed reactor atdifferent operating conditions for which experimental data arenot available.

5.3. Simulation of the HDN, HDNi and HDV Pilot Plant Reactor

The model developed now can be applied for describing thebehaviour of the pilot plant trickle bed reactor at differentoperating conditions for which experimental data are not available.Tables 24 show model predictions in terms of nitrogen conver-sion, vanadium conversion and nickel conversion at operating

ppm

Vana

dium

(ppm

)- Si

mul

ated

4

6

8

10

12

14

16

18

20

l an

2178 A.T. Jarullah et al. / Fuel 90 (2011) 216521810 2 4 6 8Vanadium (

0

2

Fig. 8. Comparison between experimenta

0 1 2 3

Nic

kel (

ppm

)- Si

mua

lted

0

1

2

3

4

5

6

7

8Nickel (ppm)

Fig. 9. Comparison between experimental a10 12 14 16 18 20)- Experimental

d calculated concentrations of vanadium.

4 5 6 7 8

- Experimental

nd calculated concentrations of nickel.

10

el 9H2 c

once

ntra

tion

(mol

/cm

3 )

6.57E-04

6.57E-04

6.57E-04

6.57E-04

6.57E-04

6.57E-04

6.57E-04

A.T. Jarullah et al. / Fuconditions other than experimental conditions. Using the model,the concentration prole of hydrogen in the gas, liquid and solidphases generated under maximum process condition (T = 400 C,P = 10 MPa and LHSV = 0.5 h1) is presented in Fig. 10. It is noticedthat the hydrogen partial pressure in gas phase decreased alongthe catalyst bed length as a result of hydrogen consumption.Whereas, the concentration prole of hydrogen in the liquid phaseand solid phase increased a long the catalyst bed length. As is wellknown, their forms are estimated by a balance between chemicalreaction and gasliquid mass transfer, decreased initially becauseof the high reaction rate in this part of the reactor and then in-creased substantially along the reactor bed length. This behaviourcan be attributed to the difference in mass transfer rate at gasliquid and liquidsolid, and reaction kinetics. When mass transferat a liquidsolid interface becomes predominant, the H2 concentra-tion decreases in both the solid and liquid phases, and when themass transfer from liquid to gas becomes important, the liquidphase concentration also solid phase concentration increases[15,16,33,91].

0 5 10 15

Reactor bed l

6.57E-04

Fig. 10. Concentration proles of H2 in liquid, so

0 5 10 15

Reactor bed le

N c

once

ntra

tion

(mol

/cm

3 )

0.0E+00

5.0E-07

1.0E-06

1.5E-06

2.0E-06

2.5E-06

3.0E-06

3.5E-06

Fig. 11. Concentration proles of N, V and Ni in liqH2 c

once

ntra

tion

(Mpa

)

Liquid phaseSolid phaseGas phase

9.99994

9.99995

9.99996

9.99997

9.99998

9.9999910.00001

0 (2011) 21652181 2179The concentration proles of N, V and Ni along the catalyst bedlength in the liquid and solid phase at maximum conditions is pre-sented in Fig. 11. As can be seen from this gure, the concentrationprole of these compounds reduced in both liquid phase and solidsurface along the reactor bed length. In addition, there is a concen-tration gradient between both phases. This gradient is governed byliquidsolid mass transfer rate calculated from the equations usedin this model, which is based mainly on the physical properties ofthe liquid, such as density and viscosity, and also liquid massvelocity. Therefore, the feedstock becomes lighter and thus physi-cal properties are improved and mass transfer of liquidsolid willenhance reducing this concentration gradient [15,16].

6. Conclusions

Estimation of the kinetic parameters in trickle bed reactor ap-plied for HDN, HDV and HDNi of crude oil is required to develop

20 25 30

ength (cm)

9.99993

lid and gas phase down through the reactor.

20 25 30

ngth (cm)

0.0E+00

1.0E-08

2.0E-08

3.0E-08

4.0E-08

5.0E-08

6.0E-08

7.0E-08

8.0E-08

9.0E-08

V an

d N

i con

cent

ratio

n (m

ol/c

m3 )

N in liquid phaseN in soild phaseV in liquid phaseV in solid phaseNi in liquid phaseNi in solid phase

uid and solid phase down through the reactor.

el 9an accurate model, so that the model can be effectively used forsimulation, optimization and control.

The kinetic parameters estimations of trickle bed reactor modelfor HDN, HDV and HDNi reactions of crude oil have been calculatedusing pilot plant experimental data and an optimization technique.In order to evaluate the best values of kinetic parameters, two ap-proaches, linear and non-linear regression have been applied forestimating the best values of kinetic parameters of trickle bed reac-tor process for HDN, HDV and HDNi of crude oil. In the rst ap-proach, the reaction order of N, V and Ni (nj), hydrogen order(mj) and reaction rate constants K

ij

for each reaction were esti-

mated simultaneously. Linearization process is then applied toestimate the activation energy (EAj) and the pre-exponential factorA0j

for each reaction. In the second approach, the activation ener-gies (EAj), pre-exponential factors A

0j

, reaction order of N, V and

Ni (nj) and hydrogen order (mj) were determined simultaneously.Based on the objective function (SSE), the parameters estimatedwith the second approach is found to be more accurate and thesimulation results showed very well correspondence with theexperimental data with an average absolute error of less than 5%.

The effect of reactor temperature (T), partial pressure of hydro-gen (P) and liquid hourly space velocity (LHSV) upon the N, V andNi conversion and upon the concentration proles along the reac-tor bed length were studied using the process model. It has beenobserved that the inuence of these operating conditions in HDN,HDV and HDNi of crude oil conrming that high temperature, pres-sure and low liquid hourly space velocity improve the nitrogen,vanadium and nickel conversion. The model can now be con-dently applied to reactor design, operation and control, as well asto predict the concentration proles of any compound at anyconditions.

Finally note, sulfur and asphaltene are also regarded as impor-tant contaminants in crude oil. Studying the effect of these impu-rities were beyond the scope of this paper, but have beenreported elsewhere.

References

[1] Rana MS, Samano V, Ancheyta J, Diaz JAI. A review of recent advances onprocess technologies for upgrading of heavy oils and residua. Fuel2007;86:121631.

[2] Ancheyta J, Betancourt G, Marroqun G, Centeno G, Casta~neda LC, Alonso F,et al. Hydroprocessing of Maya heavy crude oil in two reaction stages. ApplCatal A 2002;233:15970.

[3] Speight JG. The desulfurization of heavy oils and residues. New York: MarcelDekker, Inc; 2000.

[4] Andari MK, Behbehani H, Stanislaus A. Sulfur compound type distribution inNaphtha and gas oil fractions of Kuwaiti crude. Fuel Sci Technol Int1996;14:93961.

[5] Kaernbach W, Kisielow W, Warzecha L, Miga K, Klecan R. Inuence ofpetroleum nitrogen compounds on hydrodesulphurization. Fuel 1990;69:2214.

[6] Abbas AS. Low sulfur feedstock from basrah reduced crude oil for cokeproduction. M.SC. Thesis. Iraq: University of Baghdad; 1999.

[7] Bartholdy J, Cooper BH. Metal and coke deactivation of resid hydroprocessingcatalysts. ACS symposiumon resid upgradingDenver, USA: CO; 1993. p. 39686.

[8] Pereira CJ, Cheng JW, Suarez WC. Metal deposition in hydrotreating catalyst.Ind Eng Chem Process Des Dev 1990;29:5201.

[9] Ali LH, Abdul-Karim E. The oil, origin, composition and technology. Iraq:AlMosul University; 1986.

[10] Gary JH, Handwerk GE. Petroleum rening: technology and economics. 3rded. New York: Marcel Dekker; 1994.

[11] gPROMS. Introductory user guide. Process System Enterprise Ltd (PSE). http://www.psenterprise.com/gproms/; 2005.

[12] Khalfhallah HA. Modelling and optimization of oxidative desulfurizationprocess for model sulfur compounds and heavy gas oil. Ph.D. thesis. UK:University of Bradford; 2009.

[13] Froment GF, Depauw GA, Vanrysselberghe V. Kinetic modeling and reactorsimulation in hydrodesulfurization of oil fractions. Ind Eng Chem Res1994;33:297588.

[14] Jimenez F, Nunez M, Kafarov V. Modeling of industrial reactor for

2180 A.T. Jarullah et al. / Fuhydrotreating of vacuum gas oils Simultaneous hydrodesulfurization,hydrodenitrogenation and hydrodearomatization reactions. Chem Eng J2007;134:2008.[15] Mederos FS, Ancheyta J. Mathematical modeling and simulation ofhydrotreating reactors: cocurrent versus countercurrent operations. ApplCatal A: General 2007;332:821.

[16] Bhaskar M, Valavarasu G, Meenakshisundaram A, Balaraman KS. Petrol SciTechnol 2002;20:25168.

[17] Avraam DG, Vasalos IA. HdPro: a mathematical model of trickle-bed reactorsfor the catalytic hydroprocessing of oil feedstocks. Catal Today 2003;7980:27583.

[18] Shokri S, Marvast MA, Tajerian M. Production of ultra low sulfur diesel:simulation and software development. Petrol Coal 2007;49:4859.

[19] Wauquier JP. Crude oil petroleum products process owsheets. Paris: InstituteFrancais du Petrole; 1995.

[20] Mara A, Fukase S, Al-Marri M, Stanislaus A. A comparative study of the effectof catalyst type on hydrotreating kinetics of Kuwaiti atmospheric residue.Energy Fuels 2003;17:6618.

[21] Lababidi HMS, Shaban HI, Al-Radwan S, Alper E. Simulation of an atmosphericresidue desulfurization unit by quasi-steady state modeling. Chem EngTechnol 1998;21:193200.

[22] Mears DE. The role of axial dispersion in trickle-ow laboratory reactors. ChemEng Sci 1971;26:13616.

[23] Carberry JJ. Chemical and catalytic reaction engineering. New York: McGraw-Hill; 1976.

[24] Shah YT. Gasliquidsolid reactors design. New York: McGraw-Hill Inc.; 1979.[25] Raid RC, Prausnitz JM, Poling BE. The properties of gases & liquids. 4th ed. New

York: McGraw-Hill; 1987.[26] Ahmed T. Hydrocarbon phase behaviour. Houston: Gulf Publishing; 1989.[27] Froment GF, Bischoff KB. Chemical reactor analysis and design. 2nd ed. New

York: Wiley; 1990.[28] Dudukovic MP, Larachi F, Mills PL. Multiphase catalytic reactors: a perspective

on current knowledge and future trends. Catal Rev 2002;44:123246.[29] Macas MJ, Ancheyta J. Simulation of an isothermal hydrodesulfurization small

reactor with different catalyst particle shapes. Catal Today 2004;98:24352.[30] Rodriguez MA, Ancheyta J. Modeling of hydrodesulfurization (HDS),

hydrodenitrogenation (HDN), and the hydrogenation of aromatics (HDA) in avacuum gas oil hydrotreater. Energy Fuels 2004;18:78994.

[31] Shokri S, Zarrinpashne S. A mathematical model for calculation of effectivenessfactor in catalyst pellets of hydrotreating process. Petrol Coal 2006;48:2733.

[32] Mederos FS, Rodrguez MA, Ancheyta J, Arce E. Dynamic modeling andsimulation of catalytic hydrotreating reactors. Energy Fuels 2006;20:93645.

[33] Alvarez A, Ancheyta J. Modeling residue hydroprocessing in a multi-xed-bedreactor system. Appl Catal A: General 2008;351:14858.

[34] Korsten H, Hoffmann U. Three-phase reactor model pilot trickle-bed forhydrotreating in reactors. AIChE J 1996;42:135060.

[35] Sattereld CN. Trickle-bed reactors. AIChE J 1975;21:20928.[36] Scamangas A, Papayannakos N, Marangozis J. Catalytic hydrodesulfurization of

a petroleum residue. Chem Eng Sci 1982;37:18102.[37] Li C, Chen W, Tsai C. Highly restrictive diffusion under hydrotreating reactions

of heavy residue oils. Ind Eng Chem Res 1995;34:898905.[38] Marroquin G, Ancheyta J, Esteban C. A batch reactor study to determine

effectiveness factors of commercial HDS catalyst. Catal Today 2005;104:705.[39] Aris R. The mathematical theory of diffusion and reaction in permeable

catalysts. Oxford: Clarendon Press; 1975.[40] Chang J, Liu J, Li D. Kinetics of resid hydrotreating reactions. Catal Today

1998;43:2339.[41] Gau Ch, Stadtherr MA. Reliable nonlinear parameter estimation using interval

analysis: error-in-variable approach. Comput Chem Eng 2000;24:6317.[42] Khorasheh F, Chan EC, Gary MR. Development of a continuous kinetic model

for catalytic hydrodesulfurization of bitumen. Petrol Coal 2005;47:408.[43] Poyton AA, Varziri MS, McAuley KB, McLellan PJ, Ramsay JO. Parameter

estimation in continuous-time dynamic models using principal differentialanalysis. Comput Chem Eng 2006;30:698708.

[44] Tatiraju S, Soroush M. Parameter estimator design with application to achemical reactor. Ind Eng Chem Res 1998;37:45563.

[45] Stephanopoulos G, San KY. Studies on on-line bioreactor identication. I:theory. Biotechnol Bioeng 1984;26:117688.

[46] Kosanovich KA, Piovoso MJ, Rokhlenko V, Guez A. Nonlinear adaptive controlwith parameter estimation of a CSTR. J Proc Control 1995;5:13748.

[47] Soroush M. Nonlinear state-observer design with application to reactors. ChemEng Sci 1997;52:387404.

[48] Carloff R, Probss A, Reichert KH. Temperature oscillation calorimetry instirred tank reactors with variable heat transfer. Chem Eng Technol1994;17:40613.

[49] Barandiaran MJ, De Arbina LL, De La Cal JC, Gugliotta LM, Ausa JM. Parameterestimation in emulsion copolymerization using reaction calorimeter data. JAppl Polym Sci 1995;55:12319.

[50] Regnier N, Defaye G, Caralp L, Vidal C. Software sensor based control ofexothermic batch reactors. Chem Eng Sci 1996;51:512536.

[51] Muske KR, Rawlings JB. Nonlinear moving horizon state estimation. In: BerberR, editor. Methods of model based process control. Netherlands: Kluwer; 1995.

[52] Robertson DG, Lee JH, Rawlings JB. A moving horizon based approach for leastsquares estimation. AIChE J 1996;42:220924.

[53] Ahn J, Ch-Ki Cho, Kang S. An efcient numerical parameter estimation schemefor the space-dependent dispersion coefcient of a solute transport equation

0 (2011) 21652181in porous media. Math Comput Model 2010;51:16776.[54] Kuzmic P. Program DYNAFIT for the analysis of enzyme kinetic data:

application to HIV proteinase. Anal Biochem 1996;237:26073.

[55] Mendes P, Kell D. Non-linear optimization of biochemical pathways:applications to metabolic engineering and parameter estimation.Bioinformatics 1998;14:86983.

[56] Rinnooy Kan AHG, Timmer GT. Global optimization: a survey. Int Ser NumerMath 1989;87:13355.

[57] Tjoa J, Biegler LT. Simultaneous solution and optimization strategies forparameter estimation of differential-algebraic equation systems. Ind EngChem Res 1991;30:37685.

[58] Fletcher R. Practice methods of optimization. 2nd ed. Chichester: John Wileyand Sons; 1987.

[59] Levenberg K. A method for the solution of certain nonlinear problems in leastsquares. Quart Appl Math 1944;2:1648.

[60] Marquardt DW. An algorithm for least-squares estimation of non-linearparameters. SIAM J 1963;11:43141.

[61] Holland JH. Adaptation in natural and articial systems. Ann Arbor (MI): TheUniversity of Micigan Press; 1975.

[62] Goldberg DE. Genetic algorithms in search, optimization and machinelearning. Reading (MA): Addison-Wesley; 1989.

[63] Alonge F, Dippolito F, Ferrante G, Raimondi FM. Parameter identication ofinduction motor model using genetic algorithms. IEE Proc Control Theory Appl1998;145:58793.

[64] Orlowska KT, Lis J, Szabat K. Identication of the induction motor parametersat standstill using soft computing methods. Comput Math Electr Electron Eng2006;25:18194.

[65] Nangsue P, Pillay P, Conry S. Evolutionary algorithms for induction motorparameter determination. IEEE Trans Energy Convers 1999;14:44753.

[66] Abdelhadi B, Benoudjit A, Nait Said N. Identication of induction machineparameters using an adaptive genetic algorithm. Electr Power Compon Syst2004;32:76784.

[67] Ursem RK, Vadstrup P. Parameter identication of induction motors usingdifferential evolution. The 2993 congress on evolutionary computation, vol. 2.CEC03; 2003. p. 7906.

[68] Tjoa TB, Biegler LT. Reduced successive quadratic programming strategy forerrors-in-variables estimation. Comput Chem Eng 1992;16:52333.

[75] Ekpo EE, Mujtaba IM. Performance analysis of three controllers for thepolymerisation of styrene in a batch reactor. Chem Prod Process Model2007;2:129.

[76] Mujtaba IM. Batch distillation: design and operation. Series on chemicalengineering, vol. 3. London: Imperial College Press; 2004.

[77] Jarullah AT, Mujtaba IM, Wood AS. Improvement of the middle distillate yieldsduring crude oil hydrotreatment in a trickle-bed reactor. Energy Fuels 2011.doi:10.1021/ef101327d.

[78] Maria AC, Martinez MT. Hydroprocessing of a Maya Residue. Intrinsic kineticsof sulfur-, nitrogen-, nickel-, and vanadium-removal reactions. Energy Fuels1999;13:62936.

[79] Ancheyta J, Rivera GB, Sanchez GM, Arellano AMP, Maity SK, Cortez MT, et al.Exploratory study for obtaining synthetic crudes from heavy crude oils viahydrotreating. Energy Fuels 2001;15:1207.

[80] Bhaskar M, Valavarasu G, Sairam B, Balaraman KS, Balu K. Three-phase reactormodel to simulate the performance of pilot-plant and industrial trickle-bedreactors sustaining hydrotreating reactions. Ind Eng Chem Res 2004;43:665469.

[81] Isoda T, Kusakabe K, Morooka Sh, Mochida I. Reactivity and selectivity for thehydrocracking of vacuum gas oil over metal-loaded and dealuminated y-zeolites. Energy Fuels 1998;12:493502.

[82] Kim LK, Choi KS. Hydrodesulfurization over hydrotreating catalysts. Int ChemEng 1987;27:34057.

[83] Al-Humaidan FS. Modelling hydrocracking of atmospheric residue by discreteand continuous lumping. MS.C. thesis. Kuwait: Kuwait University; 2004.

[84] Abdul-Halim A-KM. Private communications. Professor, University ofBaghdad-Iraq; 5 Jauary 2011.

[85] Riyadh HH. Private communications. Manager, North Reneries Company-Baiji, Iraq; 5 Jauary 2011.

[86] Kumar VR, Balaraman KS, Rao VSR, Ananth MS. Performance study of certaincommercial catalysts in hydrodesulfurization of diesel oils. Petrol Sci Technol2001;19:102938.

[87] Cheng Z, Fang X, Zeng R, Han B, Huang L, Yuan W. Deep removal of sulfur andaromatics from diesel through two-stage concurrently and countercurrently

A.T. Jarullah et al. / Fuel 90 (2011) 21652181 2181[69] Tsamatsoulis D, Papayannakos N. Investigation of intrinsic hydrodesulphuri-zation kinetics of a VGO in a trickle bed reactor with backmixing effects. ChemEng Sci 1998;53:344958.

[70] Sanchez S, Rodrguez MA, Ancheyta J. Kinetic model for moderatehydrocracking of heavy oils. Ind Eng Chem Res 2005;44:940913.

[71] Trejo F, Ancheyta J. Kinetics of asphaltenes conversion during hydrotreating ofMaya crude. Catal Today 2005;109:99103.

[72] Resendiz E, Ancheyta J, Rosales-Quintero A, Marroquin G. Estimation ofactivation energies during hydrodesulfurization of middle distillates. Fuel2007;86:124753.

[73] Sanchez S, Ancheyta J. Effect of pressure on the kinetics of moderatehydrocracking of Maya crude oil. Energy Fuels 2007;21:65361.

[74] Morrison KR. Optimal control of processes described by systems ofdifferential-algebraic equations. Ph.D. thesis. UK: University of London; 1984.operated xed-bed reactors. Chem Eng Sci 2004;59:546572.[88] Jarullah AT, Mujtaba IM, Wood AS. Kinetic parameter estimation and

simulation of trickle-bed reactor for hydrodesulfurization of crude oil. ChemEng Sci 2011;66:85971.

[89] Areff HA. The effect of operating conditions on vacuum gas oil hydrotreatingon sulfur and aromatics content. M.SC. thesis. Iraq: University of Tikrit; 2001.

[90] Jimenez F, Nunez, Kafarov V. Study and modeling of simultaneoushydrodesulfurization, hydrodenitrogenation and hydrodearomatization onvacuum gas oil hydrotreatment. Comput Aid Chem Eng 2005;20:61924.

[91] Van Hasselt BW, Lebens PJM, Calis HPA, Kapteijn F, Sie ST, Moulijn JA, et al. Anumerical comparison of alternative three-phase reactors with a conventionaltrickle-bed reactor. The advantages of countercurrent ow for hydrodesulfuri-zation. Chem Eng Sci 1999;54:47919.

Kinetic model development and simulation of simultaneous hydrodenitrogenation and hydrodemetallization of crude oil in trickle bed reactorIntroductionMathematical model of TBR for HDN, HDV and HDNi reactionsMass balance equationsChemical reaction rateReactor performanceKinetic parameters of the models

Parameter estimation techniquesOptimization Problem Formulation for Parameter Estimation

Experimental workMaterialsEquipment and procedureExperimental runs

Results and discussionsExperimental resultsEstimation of kinetic parametersSimulation of the HDN, HDNi and HDV Pilot Plant Reactor

ConclusionsReferences