INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 2, ISSUE 10, OCTOBER 2013 ISSN 2277-8616

On The Conversion Of Gas Oil In Fluid Catalytic

Cracking Risers: Application Of Residence Time
Distribution (Rtd) Concept
H.A. Affum, I.I. Mumuni, G.K. Appiah, S.Y. Adzaklo, M. A. Addo

ABSTRACT: Risers are considered vital parts on fluidized catalytic cracking (FCC) conversion units. It is inside the riser that the heavy hydrocarbon
molecules are cracked into petroleum fractions such as gasoline and liquefied petroleum gas (LPG). A simplified kinetic flow model in combination with
the Tank-In-Series model was used to predict the conversion response of an FCC riser to changes in feed temperature, feed flow rate as well as riser
diameter and height. The various Residence Time Distribution (RTD) functions and flow-model parameters are used in the characterization of the
mixing regime of the riser and the degree of any non-ideal flow behaviour. Conversion was observed to increase with increasing riser height and
diameter with values of 95.16% and 94.0% at a riser height of 80 m and diameter of 1.0 respectively. Conversion also increased with increasing feed
temperature. A feed flow rate of 10 m 3/s is converted at 95.95 % whiles a feed flow rate of 40 m 3/s is converted at 89.72%, indicating an inverse
relationship between conversion and feed flow rate. The simulation also revealed that the riser reactor is equal to approximately 1–2 perfectly stirred
tanks in series as conversion started to decrease after an N, the number of tanks in the Tank-In-Series model, of 1.5.

Index Terms: Conversion, gasoline, residence time distribution, risers, Tank-In-Series model
————————————————————

1. INTRODUCTION This study, therefore, aims at developing a mathematical

Fluid catalytic cracking (FCC) is one of the most important model which simulates non-ideal flow in a continuously
and complex processes in petroleum refining which is used stirred riser reactor (CSTR) and to demonstrate the
to upgrade heavy petroleum gas oils into gasoline and other relevance and application of the concept of residence time
valuable products. The FCC process comprises mainly 2 distribution (RTD). An appropriate mixing model in
parts: (1) a riser reactor where high molecular weight combination with the kinetics of the reactions occurring in
hydrocarbons come into contact with a catalyst and crack to the riser is developed to predict the reactant conversion at
lower molecular weight products with the simultaneous various operating conditions. To account for the non ideality
deposition of coke on the catalyst surface and (2) a in the reactor, the tank-in-series model is assumed to
regenerator where the coke on the catalyst is burnt with air estimate the exit age distribution (E) curve.
and the catalyst is returned to the riser for the next run of
cracking [1]. There have been several simulation studies 2. MATERIALS AND METHODS
focused on FCC due to its importance [2, 3, 4] with special
attention on the riser reactor which is probably the most 2.1 Riser Design and kinetic model
important equipment in an FCC plant. All cracking reactions The complexities of gas oil mixtures which are typical FCC
and fuel formation occur in this reactor. Therefore, a feedstocks make it extremely difficult to characterize and
mathematical model can be a valuable tool to design describe the inherent kinetics at a molecular level. To allow
modifications or operational changes that give higher yields for easy studies, similar components are grouped into few
from the process. This explains why many of the models cuts or lumps [9]. The study of the reactions involved in the
found in the literature describe the riser reactor [5, 6, 7, 8, catalytic process has therefore followed the lumping
9]. However, most of these simulations assume idealized methodology. The number of lumps of the proposed models
mixing conditions within the riser reactors. for catalytic cracking reactions has been consecutively
increasing to obtain a more detailed prediction of product
distribution [10]. Since the simple models, with just a few
slumps are usually more suitable for specific simulations
and for reactor design purposes, the 4- lump kinetic
reaction scheme will be adopted in the present study. The
four–lump mechanism suggested by Juarez et al in [11] has
been selected. It is shown in Fig. 1.

r2, k2 light gases

————————————————
r4, k4
 H.A. Affum is a senior research scientist at the Gas oil r1, k1 gasoline
Ghana Atomic Energy Commission and is currently r5, k5
pursuing a PhD program in Applied Nuclear Physics r3, k3
at the University of Ghana in collaboration with the
International Centre for Theoretical Physics (ICTP), Coke
Trieste, Italy
 E-mail: afmyn79@yahoo.com Figure 1: A Schematic diagram of the four-lump scheme

168
IJSTR©2013
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 2, ISSUE 10, OCTOBER 2013 ISSN 2277-8616

According of this scheme, gasoil is converted into gasoline, reactor. The tracer must not disturb the flow pattern of the
light gases and coke. A part of the gasoline is also system. The analysis of the output concentration with time,
converted to light gases and coke. It is assumed that gives the desired information about the system and helps to
cracking of gas oil is a second- order reaction whiles that of determine the residence time distribution function E (t) [15,
gasoline is a first-order reaction, and that the reactions take 16]. The RTD curve can be used as a diagnostic tool for
place only in the gas phase. Based on the riser flow ascertaining features of flow patterns in reactors. These
characteristics and dynamics which involves turbulence and include the possibilities of bypassing and/or regions of
mixing, the riser is considered as a CSTR. The assumption stagnant fluid (i.e., dead space). Since these
is also made that the riser is a constant density system. For maldistributions can cause unpredictable conversions in
a constant density fluid flowing in a system of volume V at a reactors, they are usually detrimental to reactor operation.
flow rate Q, the mean residence time () of fluid is According to Levenspiel in [13], the application of the RTD
theoretically defined as  = V/Q [12]. It is also assumed the to the prediction of reactor behavior is based on the
catalyst and gas have a same temperature along the riser, assumption that each fluid (assume constant density)
instantaneous vaporization occurred in entrance of riser behaves as a batch reactor and that the total reactor
and that there is no radial and axial dispersion in the riser. conversion is then the average of the fluid elements, that is:
The ideal steady-state mixed flow reactor is one in which
the contents are well stirred and uniform throughout. Thus,
(3)
the exit stream from this reactor has the same composition
as the fluid within the reactor. For this type of reactor,
According to the tank in series model, the RTD is given by
mixing is complete, so that the temperature and the
composition of the reaction mixture are uniform in all parts
of the vessel and are the same as those in the exit stream
[13]. According to Levenspiel, Equation (1) represents the (4)
performance equation for mixed reactors.
Where

(1) = mean residence time = 

Therefore, for the second order reaction for the cracking of t = time
gasoil in an ideal mixed flow reactor, the performance
equation is given by: N = number of tanks

Therefore, combining equations (3) and (4) results in:

(2)

Where (5)
k = rate constant in the rate equation Equation (5) is the performance equation for a non ideal
mixed flow riser with A representing gas oil.
= mean residence time
The following are the reactions occurring in the riser:
CA0 = inlet/initial concentration of A (gas oil) into reactor

CA = final or concentration of A at reactor outlet

However, real equipment always deviates from these

ideals. Deviation from ideal flow patterns can be caused by
channeling of fluid, by recycling of fluid, or by creation of
stagnant regions in the vessel. This is because elements of (6)
fluid taking different routes through the reactor may take
different lengths of time to pass through the vessel. The Where
time an atom spends in a reactor is called the residence
time. The distribution of these times for the stream of fluid r1 = the rate of reaction of gas oil to form gasoline at rate
leaving the vessel is called the exit age distribution E, or the constant k1
Residence Time Distribution (RTD) of fluid [14]. Therefore,
the E curve is the distribution needed to account for the r2 = the rate of reaction of gas oil to form light gases at rate
non-ideal flow [13]. The RTD is determined experimentally constant k2
by injecting an inert chemical molecule or atom called
tracer, into the reactor at some time t = 0 and then r3 = the rate of reaction of gas oil to form coke at rate
measuring the tracer concentration C, in the effluent stream constant k3
as a function of time. The tracer usually possesses physical
properties similar to those of the reacting mixture, so that its r4 = the rate of reaction of gasoline to form light gases at
behavior will reflect that of the material flowing through the rate constant k4
169
IJSTR©2013
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 2, ISSUE 10, OCTOBER 2013 ISSN 2277-8616

r5 = the rate of reaction of gas oil to form coke at rate Table 2: Industrial Riser operating parameters
constant k5
Parameter Value
Cgo=concentration of gas oil

Cgl=concentration of gasoline Riser diameter (m) 0.8

Riser height (m) 33

= decay of catalyst activity due to coke

If kgo is the overall gas oil reaction rate constant, then: Gasoil molecular weight
333
(kg/kmol)
(7)

For the purposes of this study, only reaction rates 1, 2 and Feed vaporization
698
3 were considered. The reaction rate constants were temperature (K)
calculated by using Arrhenius type semi-empirical equation
given by: Gas oil Feed rate (kg/s) 19.95
(8)
Gas oil density (kg/m3) 835
Where
Gas constant (kJ/(kmol K) 8.314
ki = rate constant for reaction i
(Source: [17])
kio = pre-exponential constant/factor for rate constant ki
Table 3: Calculated parameters
E = activation energy
Rate constant for overall
R =Gas constant and 0.0969
cracking of gasoil
T = temperature
Gas oil volumetric flow rate,
0.024
Kinetic parameters for the FCC riser reaction and used for (m3/s)
the simulation are provided in Table 1. Other useful riser
parameters are given in the Table 2 below. Mean residence time (s) 691.24

3. RESULTS AND DISCUSSION Initial gas oil

2.508
concentration,(kmol/m3)
3.1 Riser simulation
The riser flow is simulated using the plant data given in Riser volume (m3) 16.59
Table 1 reported by Ali et al. in [8]. The values of other
parameters used in the simulation are listed in Table 2. In order to study the effect of changing one independent
Table 3 also presents parameters which were calculated variable on the gas oil conversion, all others must be held
with data from Tables 1 and 2 and also used for the constant. Therefore to study the effect of riser height on
simulation. Using Eq. (5), an Excel Spreadsheet was conversion, the flow rate was kept constant at 19.95 kg/s,
developed for the gasoline conversion. Riser flow was as well as the feed temperature at 698 K and riser diameter
simulated for various riser operating conditions such as at 0.8 m. Holding these parameters still constant with a riser
feed rate, riser diameter and height, feed temperature and height of 33 m, the effect of riser diameter on conversion
number of tanks in the one parameter Tank-In-Series was simulated. All other conditions being roughly equal,
model. riser height correlates with residence time and for that
matter gas oil conversion [18]. It can be observed from
Table 1: Kinetic Parameters for cracking kinetics in riser Figs. 2 and 3 that a positive correlation exits between
conversion and riser height and with riser diameter. As riser
Reaction k0 E (kJ/mol) height and diameter increase, residence time also
Gasoil to gasoline 1457.5 57.36 increases and hence conversion. This suggests, therefore
Gasoil to light that the design of shorter and small diameter risers is not
127.0 52.75
gases economical.
Gasoil to coke 1.98 31.82
Catalyst
8.38 x 104 117.72
deactivation
(Source: [17])

170
IJSTR©2013
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 2, ISSUE 10, OCTOBER 2013 ISSN 2277-8616

sharp increase in conversion from N=1 to N=1.5 after which

the conversion declines.

Keeping constant the riser diameter of 0.8 m, height of 33

m and the feed temperature at 698 K, the effect of feed flow
rate and number of mixing tanks (N) on conversion was
investigated. It was observed that a negative correlation
exits between flow rate (or the residence time) and
conversion. Table 4 shows in general that a decrease in
residence time will decrease conversion. This observation
is corroborated by [18, 19, 20]. Generally, conversion was
also found to decreases as the number of tanks increased
in the model. It is interesting to note a deviation from the
general trend as conversion increased sharply from N=1 to
N=1.5 tanks and declines thereafter. The small value of N
depicts that, the riser characteristics turns are those of a
single CSTR or perhaps 2 CSTRs in series rather than a
plug flow reactor [16]. As the value of N increases, plug flow
is approached tanks have no mixing in the direction of flow
and thus conversion is reduced 12.

Table 4: Effect of feed flow rate and number of tanks (N) on

conversion

Feed flow
Feed flow
Feed flow rate rate
rate
N = 10kg/s = 40 kg/s
= 19.95 kg/s
With a feed flow rate of 19.95 kg/s, riser diameter of 0.8 m, ( τ =1385.245s) (τ =
riser height of 33 m, the effect of feed temperature on (τ =691.24s)
346.311s)
conversion was simulated. It can be seen from Fig. 4 that X
gasoil conversion increased linearly with increasing feed
temperature. This observation is supported by Dasila et al 1 95.95 92.99 89.72
in [19]. Increased temperature means increased kinetic rate 1.5 96.37 93.64 90.86
constant, thereby increasing rate of reaction and hence
1.6 96.06 93.57 90.78
conversion because the cracking of gasoil to gasoline is
endothermic. Increasing temperature can help bring 1.9 95.93 93.04 90.07
gasoline conversions up as shown in Fig 4. However, the 2 95.84 92.73 89.68
heat balance in whole FCC system and limits on operating
temperature in the regenerator can limit the increase 5 95.51 92.59 89.68
possible in operating temperature in the riser [18]. Though 20 94.88 92.58 89.68
generally conversion is observed to decrease with 100 94.70 92.58 89.68
increasing number of mixing tanks, there seem to be a

171
IJSTR©2013
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 2, ISSUE 10, OCTOBER 2013 ISSN 2277-8616

The RTD curves for a gas oil flow rate of 40 kg/s as shown the riser for conversion to take place which also explains
in Fig 7, exhibit long tails, an indication of the presence of why the conversion is lower.
dead volumes or stagnant volume with little exchange in the
riser at those conditions. A short residence time of 346.3 s From the shapes of the curves, it can be seen that mixing is
may mean that the reactants do not spend adequate time in best in Fig. 5, better in Fig. 6 and poor in Fig. 7. This
explains why the conversion also follows a similar pattern.

172
IJSTR©2013
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 2, ISSUE 10, OCTOBER 2013 ISSN 2277-8616

4. CONCLUSION Cracking Process, Catalysis Today, Vol. 127, No. 1

The concept of residence time distribution curves has been , pp. 31-43,(2007).
used to simulate the flow in a riser reactor. Parametric
studies were done on the effects of temperature, feed flow [11]. Juarez, J. A., Isunza, F. L., Rodriguez, E. A. and
rate and FCC riser height and diameter (to vary residence Mayotga, J. C. M.; A strategy for Kinetics
time) on gasoil conversion. Results indicated that Parameter Estimation in the Fluid Catalytic
shortening the residence time and decreasing gas oil feed Cracking Process, Ind.Eng.Chem. Res., 36, 5170-
temperature may significantly reduce overall conversion. 5174, (1997)

5. ACKNOWLEDGEMENTS [12]. International Atomic Energy Agency (IAEA).

The authors are grateful to the International Atomic Energy Radiotracer Residence Time Distribution method
Agency (IAEA) and the Ghana Atomic Energy Commission for industrial and environmental applications, IAEA-
TCS-31(2008), Vienna.
for the technical support.
[13]. Levenspiel, O., 1999. Chemical Reaction
6. REFERENCES Engineering. 3rd Edn., John Wiley & Sons, Inc.,
[1]. Moharir, A.S., Sarae, S.K., Kinetic Modeling of 605 Third Avenue, New York.
Vacuum Gasoil Catalytic Cracking, Chem. Age of
India, 35; 433-440, (1984). [14]. Richardson, J.F. and D.G. Peacock, Coulson and
Richardson’s Chemical Engineering, 3rd Edn.,
[2]. Landeghem, F.V., Nevicato, D., Pitualt, I., Pergamon, Great Britain, 3: 71-80, 102-103.
Forissier, M., Turlier, P., Derouin, C. and Bernard, (1994).
J.R. Fluid Catalytic Cracking: Modeling of an
Industrial Riser. Applied Catalysis A: General, vol. [15]. Smith, J.M., H.C. Van Ness and M. M. Abbott,
138, pp.381-405, (1996) (1996). Introduction to Chemical Engineering
Thermodynamics, 5th Edn., pp: 64-5, 189-91, 604-
[3]. Theologos, K.N., and Markatos, N.C Advanced 8, 831-37, McGraw-Hill, Singapore.
modeling of fluid catalytic cracking riser-type
reactors. AIChE J, 43,486-493, (1993). [16]. Fogler, H.S., Elements of Chemical Engineering
Reaction. 2nd Edn., Prentice-Hall of India Private
[4]. Derouin, C., Nevicato, D., Forissier, M. ,Wild, G.,& Ltd., New Delhi, pp: 708-723, 759-765, (1997).
Bernard, J.R. Hydrodynamics of riser units and the
impact on FCC operation. Industrial and [17]. Ahari, J.S, Farshi, A. and Forsat, K. A.
Engineering Chemistry Research, 36, 4504–4515, Mathematical Modeling of the Riser Reactor in
(1997) Industrial FCC Unit. Petroleum & Coal, 50 (2), 15-
24, (2008).
[5]. Arandes, J. M. and H. I. Lasa. Simulation and
Multiplicity of steady states in fluidized FCCUs, [18]. Chang, S.L., Lottes, S.A., Zhou, C.Q. and Petrick.
Chem.Eng.Sci., 47, 2535-2540,(1992) M A numerical study of short residence time
FCCriser flows with a new flow/kinetics modeling,
[6]. Arbel, A., Z. Huang, I.H. Rinard, R. Shinnar and submitted to International Mechanical Engineering
A.V. Sapre. Dynamics and control of fluidized Congress and Exposition, Nov. 15-20, (1998),
catalytic crackers. Modeling of the current Anaheim, CA
generation FCC.s. Ind. Eng. Chem. Res., 34: 1228-
1243, (1995) [19]. Dasila, P.K., Choudhury, I., Saraf, D., Chopra, S.,
Dalai., A Parametric Sensitivity Studies in a
[7]. Han, I. S. and Chung, C. B. Dynamic modeling and Commercial FCC Unit. Advances in Chemical
simulation of a fluidized catalytic cracking process Engineering and Science, 2, 136-149, (2012).
Part I: Process modeling, Chem.Eng.Sci., 56,
1951-1971, (2001) [20]. Dagde, K.K., and Puyate, Y. T. Six-lump kinetic
modeling of adiabatic plug-flow riser-reactor in an
[8]. Ali, H., S. Rohani and J.P. Corriou. Modeling and industrial FCC unit. International Journal of
control of a riser type Fluid Catalytic Cracking Engineering Research and Applications. Vol. 2,
(FCC) unit. Chem. Eng. Res. Des., 75(A4): 401- Issue 6 , pp.557-568, 557, 2012
412. (1997).

[9]. Coxon, P.G., and Bischoff, K.B., Lumping Strategy.

Introduction To Techniques and Application of
Cluster Analysis, Ind. Eng. Chem. Res., 26, 1239-
1248, (1987).

[10]. Bollas,G.M., Lappas, A.A., Iatridis, D.K. and Va-

salos,I.A. Five-Lump Kinetic Model with Selective
Catalyst Deactivation for the Prediction of the
173
IJSTR©2013
www.ijstr.org