WHOC11-588

Mapping Reactor Operating Regimes for Heavy Gas Oil Hydrotreating

MUGUREL CATALIN MUNTEANU AND JINWEN CHEN
CanmetENERGY, Natural Resources Canada

One Oil Patch Drive, Devon, AB, T9G 1A8, Canada

This paper has been selected for presentation and/or publication in the proceedings for the 2011 World Heavy Oil Congress [WHOC11].
The authors of this material have been cleared by all interested companies/employers/clients to authorize dmg events (Canada) Ltd., the
congress producer, to make this material available to the attendees of WHOC11 and other relevant industry personnel.
high residue content. Heavy oil hydrotreating is operated under
high temperatures (320–420°C) and pressures (30–120 bars), and
Abstract in the presence of catalyst and hydrogen. Under such conditions,
the reaction system (hydrogen and hydrocarbons) is at vapor-
Hydrotreating (HDT) is one of the most important processes liquid equilibrium (VLE), with both the vapor and liquid phases
used in refineries to improve the quality of distillate fractions by containing hydrogen and hydrocarbons. With increases in
removing sulfur, nitrogen, oxygen, and metals, and saturating temperature and hydrogen flow rate, and decrease in pressure,
olefins and aromatics. HDT is usually operated at temperatures of more hydrocarbons are vaporized into the vapor phase, which can
350–400°C and pressures of 50–100 bars in a fixed bed. Under considerably alter the fluid flow rates and hydrodynamics, as well
such conditions, HDT typically takes place as a gas-liquid-solid as the concentrations of various reacting species in both phases.
process (trickle bed) due to the presence of high-boiling-point For pilot plant studies where a relatively small reactor is used,
fractions in the feedstocks. In the trickle bed, hydrogen gas and VLE becomes especially important since any departures from the
liquid oil flow through the solid catalyst fixed-bed. The three- desired ideal operating regime (plug flow, full catalyst wetting,
phase process involves complex couplings between the reactions, absence of reactor wall effects) could result in faulty and
and mass and heat transfer. While efforts to model trickle-bed unrepresentative data for use in commercial scale-up and kinetics
reactors have made considerable progress in combining transport studies. Therefore, it is essential to study, understand, and
phenomena and heterogeneous reactions, variations in flow rate correctly predict VLE in hydroprocessing reactors and,
and composition for each phase due to the volatilization of liquid additionally, to predict the operating regimes under which pilot
oil are often ignored because no vapor-liquid equilibrium (VLE) plant studies will generate reliable, repeatable, and representative
data are available for the HDT system. results.
In this research, a calibrated flash calculation program was The literature considering VLE in three-phase reactors is very
used to predict VLE for a typical heavy distillate feed in pilot limited1-5 Even fewer studies account for VLE effects in
plant hydrotreaters under various operating conditions. The petroleum hydroprocessing.6,7 Systems consisting of hydrogen and
results of the flash calculation were further used to predict the a single hydrocarbon or a simulated hydrocarbon mixture are
flow hydrodynamics in a pilot plant trickle-bed reactor for HDT studied more often8-12 than systems consisting of hydrogen and
of heavy distillate feeds in order to map operating conditions actual feedstocks13,14. The lack of research on VLE during
under which the desired operating regimes can be maintained. petroleum hydroprocessing is probably due to the significant
difficulty in experimental measurements (or simulations) and the
poor definition of the problem due to the complexity of the
system. In addition, severe operating conditions, and sample
Introduction characterization and analysis, add to the difficulty in the
experiments.
Fixed-bed reactors operated under trickle-flow conditions In this research, VLEs in heavy distillate pilot plant
(TBRs) are widely employed for hydrotreating processes in hydrotreaters were predicted under various operating conditions
petroleum refineries, and for many gas-liquid-solid reactions in using a flash calculation program calibrated in-house. The
chemical/petrochemical plants. During heavy oil upgrading and calculated VLE results were used to predict flow dynamics in the
petroleum refining, the objective of hydrotreating bitumen and hydrotreaters to investigate if they were being operated under the
bitumen-derived feedstocks is to reduce sulfur, nitrogen, and desired operating regime (plug flow, full catalyst wetting, and
metal contents to acceptable levels. Bitumen hydrotreating is absence of reactor wall effects). Detailed descriptions and
limited by the nature of the feedstock: high molecular weight and definitions of these hydrotreater operating regimes are discussed

1
for typical heavy distillate feed under various operating conditions especially in a short catalyst bed with a high conversion.
and parameters. Otherwise, the catalyst evaluation results and subsequent kinetics
analysis, normally based on an assumption of plug flow, could
cause discrepancies in reactor scale-up. Mears9 derived a criterion
from the perturbation solution of an axial dispersion model with
Mapping of Trickle-Bed Reactor HDT N-series mixers, which was based on the assumption that the
minimum bed length to ensure freedom from axial diffusion
Operating Regimes should deviate less than 5% from the bed length required for plug
flow. In this equation, an implicit assumption that conversion is
less than 90% was applied. An even more relaxed criterion with
System description 15% deviation from plug flow was used in our computation:
In this study, VLE calculations and determination of operating
regimes in trickle-bed HDT reactors were carried out using an in-
house pilot plant hydrotreater. Such units are commonly used in LB 20n 1
> ln
(1 − α )
.......................................................... (1)
hydrotreating catalyst development, evaluation, and activity
stability studies to generate reliable data for commercial scale-up.
d pe Pe L
The specifications of the pilot plant hydrotreater used in this study
are: The effective particle size (dpe) is calculated based on the
Reactor ID: 2.54 cm volumetric average of the catalyst particle (cylinder shape) and the
Catalyst volume: 100 to 200 mL diluent particle (spherical shape) in the reactor. It should be
Catalyst particle size (average): 1.3x5.0 mm pointed out that the above equation contains the Peclet number
Diluent particle size: 0.2 to 0.4 mm PeL based on catalyst particle diameter. In this work, we used the
Catalyst-to-diluents volumetric ratio: 1:1 Hochman and Effron correlation11
The operating conditions used were typical for commercial heavy
distillates hydrotreating:
Temperature: 350, 360, 370, 380, 390, 400, 410, and 420°C Pe L = 0.034 Re L0.53 ................................................................... (2)
Pressure: 50, 60, 70, 80 and 90 bars
Liquid hourly space velocity (LHSV): 1.5/h in which the particle Reynolds number
Gas/oil ratio: 500, 800, and 1000 NL/h
The properties of the HVGO feed are given in Table 1.
d pe u L ρ L
Table 1. Physical properties and chemical composition of the Re L = ........................................................................ (3)
HVGO feed μL
Density (15.6°C),
g/cm3 (ASTM D4052) 0.9609
Full catalyst wetting criterion
Paraffins, wt% 30.94 In a trickle-bed reactor, it is desirable to have all the catalyst
particles in the reactor completely covered by the flowing liquid
Cyclo-paraffins, wt% 7.86
to ensure maximum catalyst utilization. Any partial catalyst
Aromatics, wt% 61.2 wetting could cause by-passing of the feed and, thus, reduced
catalyst performance. If the liquid superficial velocities are very
Boiling points, °C IBP 201.8 low, which is common in a pilot plant reactor due to the relatively
small volume of catalyst loading, partial catalyst wetting may
10 wt% 292.6 occur. In this work, the criterion for wetting efficiency proposed
by Gierman10 was used:
30 wt% 350.2

50 wt% 386.4 μLuL

> 5.0 × 10 -6 ................................................................ (4)
70 wt% 445.4 ρ L d pe g
2

90 wt% 529.6

FBP 702.6 Absence of Reactor Wall Effects Criterion

Reactor wall effects on reactor performance are not normally a
Operating Regime Mapping concern for a commercial hydrotreating reactor, since the reactor
In trickle-bed pilot plant hydrotreating studies, in order to diameter is much larger than the catalyst particle size. However, in
generate consistent and reliable experimental data, it is critical to a pilot plant reactor, a concern arises when a small-diameter
maintain the reactor operation in the desired regime based on the reactor and relatively large catalyst particles are used. To keep
criteria described below. These criteria are discussed in detail in a wall effects to a minimum, a trickle bed reactor should meet the
previously published paper on hydrotreating of various petroleum following condition:12
middle distillates14.
DR
Plug Flow Criterion > 25 ....................................................................................... (5)
In a pilot plant trickle-bed reactor, attention should always be d pe
paid to the axial dispersion or backmixing in the liquid phase,

2
 In our study, the calculated DR/dpe = 49.7 >> 25. Therefore, 0.8
reactor wall effects were negligible and are not discussed further Maximum Diluent Particle Size Required for PFR
in this paper. 0.7 Operating Regime
-1
(HGO, LB=63 cm, LHSV=1.5 h , Gas/Oil=800 NL/kg)

Maximum size of diluent, mm

0.6
Results and Discussion 0.5

0.4
Plug Flow Operating Regime P=50 bar
The calculated operating regimes that meet the plug flow 0.3
P=60 bar
criterion are shown in Figure 1. Each curve in the figure 0.2 P=70 bar
represents, for a certain temperature, the boundary of pressure and P=80 bar
gas/oil ratio at which the plug flow criterion (1) is just met. The 0.1
zone above the curve represents the operating temperatures and P=90 bar
pressures that meet the plug flow criterion and the zone below the 0.0
curve represents the operating temperatures and pressures that 350 360 370 380 390 400 410 420
cannot meet the plug flow criterion. Normally, at a fixed Temperature, °C
temperature, a higher gas/oil ratio requires increased pressure in
order to ensure plug flow of the liquid phase. At a given gas/oil Figure 2. Plug flow regime mapping, temperature-maximum
ratio, increased temperature requires increased pressure to size of diluent particle-pressure
maintain liquid plug flow.
For some pilot plant studies, the operating conditions and Figure 3 presents the plug flow criteria for various conversions.
catalyst particle size are fixed according to test requirements. Although the conversion was fixed at 90% for the rest of the
Therefore, to maintain plug flow of the liquid inside the reactor, calculation, the actual conversion varied between 80% and 95%.
diluent particle size has to meet a certain value under a particular As the figure shows, higher conversion requires either higher
set of operating conditions. Figure 2 shows the requirement for pressure or lower temperature (or both) in order to meet the plug
diluent size at different reactor temperatures and pressures for a flow criterion. The pressure requirement does not vary much as
fixed LHSV of 1.5/h and gas/oil ratio of 800 NL/kg. The zone the conversion increases from 80% to 90%, but increasing the
under the curve represents the operating conditions that meet the conversion from 90% to 95% requires a more marked increase in
plug flow criterion and the zone above the curve represents the pressure to maintain plug flow.
operating conditions that cannot meet the plug flow criterion. As
seen in this figure, the increased temperature required a reduced 350
diluent particle size in order to ensure plug flow of the liquid. For PFR Operating Regime
a given temperature, decreased pressure requires decreased diluent 300 -1
(HGO, LB=63cm, LHSV=1.5 h )
particle size to maintain liquid plug flow.
α = 80%
250
α = 85%
Pressure, bar

120 α = 90%
200
PFR Operating Regime α = 95%
-1
100 (HGO, LB=63 cm, LHSV=1.5 h )
150
T=350°C
T=360°C
80 T=370°C 100
Pressure, bar

T=380°C
T=390°C 50
60
T=400°C
T=410°C 0
40 T=420°C
350 360 370 380 390 400 410 420
Temperature, °C
20
Figure 3. Plug flow regime mapping, temperature-conversion-
0 pressure
500 600 700 800 900 1000
Gas/Oil ratio, NL/kg
Full Catalyst Wetting
Figure 1. Plug flow regime mapping, gas/oil ratio-pressure- VLE in the pilot plant hydrotreater can significantly change the
temperature liquid phase flow rate due to partial vaporization of the oil into
vapor phase. Such changes in liquid flow rate can result in
incomplete catalyst wetting, leading to bypassing of the feed and,
thus, reduced catalyst performance. The resulting data may also be
unreliable and irreproducible. Figure 4 shows the mapping of full
catalyst wetting at different LHSVs. The zone above the curve
represents the operating temperatures and pressures that meet the
full catalyst wetting criterion, and the zone below the curve
represents the operating temperatures and pressures that do not.
Normally, under fixed LHSV, increased reactor temperature

3
requires increased pressure in order to ensure full catalyst wetting. the HVGO feed requires higher pressure and LHSV, and/or lower
At a given temperature, decreased LHSV requires increased gas/oil ratio and temperature, and/or smaller diluent particle size
pressure to maintain full catalyst wetting. in order to maintain plug flow of the liquid and full catalyst
wetting in the reactor. Higher conversion makes such
160 requirements more stringent.
Full Catalyst Wetting Operation Regime
140 (HGO, LB=63 cm)

120 Acknowledgement
LHSV = 1.00 1/h
LHSV = 1.05 1/h The authors are grateful to the following individuals for their
Pressure, bar

100
LHSV = 1.25 1/h help and support: Dennis Carson for conducting the VLE
80 LHSV = 1.50 1/h experiments and Conrad Gietz for editing this paper. Partial
LHSV = 2.00 1/h funding has been provided by the Canadian Program for Energy
60 Research and Development (PERD). The authors are grateful to
CanmetENERGY’s staff from the pilot plant and analytical lab for
40 their support and help.
20
NOMENCLATURE
0
350 360 370 380 390 400 410 420
BP = boiling point (°C)
Temperature, °C
CAromatics% = aromatic content of feed (wt%)
dpe = effective particle size (cm)
DR = diameter of reactor bed (cm)
Figure 4. Full catalyst wetting operating regime mapping
g = acceleration of gravity (981cm/s2)
LB = reactor bed length (cm)
Calculations were also performed to evaluate the minimum
LHSV = liquid hourly space velocity (h-1)
LHSV needed to meet the full catalyst wetting criterion. Figure 5
n = reaction order
depicts the calculation results. At the lower temperatures (350–
P = pressure (bars)
370°C) there is no significant difference between the LHSVs at
PeL = particle Peclet number
which the full wetting criterion is met. This is no longer true when
ReL = particle Reynolds number
the temperature increases to 400°C and beyond. At certain
T = temperature (°C or K)
temperatures, when the reactor is operated at higher pressures the
uL = liquid superficial velocity (cm/s)
LHSV needed to establish a full wetting regime in the reactor is
α = conversion of component
lower than when it is operated at lower pressures. In other words,
Suffixes
at certain temperatures, lower pressure requires higher LHSV to
L = liquid phase
meet the criterion.
Greek Symbols
3.0
μL = liquid phase viscosity (cP)
Minimum LHSV Required for Full Catalyst Wetting ρL = liquid phase density (g/cm3)
Operating Regime
2.5
(HGO, LB=63 cm, Gas/Oil=800 NL/kg) REFERENCES
2.0 P=50 bar 1. AKGERMAN, A., COLLINS, G.M., and HOOK,
P=60 bar B.D., Effect of Feed Volatility on Conversion in
-1

Trickle-bed reactors; Industrial & Engineering

LHSV, h

P=70 bar
1.5 P=80 bar Chemistry Fundamentals, Vol. 24, pp 398-401, August
P=90 bar 1985.
1.0
2. SMITH, C.M., and SATTERFIELD, C.N., Some
0.5 Effects of Vapour-Liquid Flow Ratio on Performance
of a Trickle-bed reactor; Chemical Engineering
Science, Vol. 41, pp. 839-843, April 1986.
0.0
350 360 370 380 390 400 410 420 3. LAVOPA, V., and SATTERFIELD, C.N., Some
Temperature, °C
Effects of Vapour-Liquid Equilibria on Performance of
a Trickle-Bed Reactor; Chemical Engineering Science,
Vol. 43, pp. 2175-2180, August 1988.
Figure 5. Full catalyst wetting operating regime mapping,
temperature-LHSV-pressure 4. DE JONG, K.P., Effect of Vapour-Liquid Equilibrium
on Coke Deposition in Trickle-Bed Reactors during
Heavy Oil Processing. 2. Modeling; Industrial &
Conclusions Engineering Chemistry Research, Vol. 33, pp. 3141-
3145, December 1994.
The calculated VLE results obtained in this work were used to
predict the flow dynamics in pilot plant hydrotreaters with one 5. TESSER, R., DI SERIO, M., and SANTACESARIA,
HVGO feed. Further calculations were performed to provide a E., Influence of the Vapour-Liquid Equilibrium (VLE)
map of operating conditions under which the desired operating on the Kinetics in Gas-Liquid and Gas-Liquid-Solid
regimes (plug flow, full catalyst wetting, and absence of reactor Systems; Catalysis Today, Vol. 79&80, pp. 323-331,
wall effects) could be maintained. The calculation results of pilot April 2003.
plant hydrotreater operating regimes showed that hydrotreating

4
6. KOCIS, G.R., and HO, T.C., Effects of Liquid
Evaporation on the Performance of Trickle-Bed
Reactors; Chemical Engineering Research and Design,
Vol. 64, pp. 288-291, July 1986.
7. BELLOS, G.D., and PAPAYANNAKOS, N.G., The
Use of a Three Phase Microreactor to Investigate HDS
Kinetics; Catalysis Today, Vol. 79&80, pp. 349-355,
April 2003.
8. CARBERRY, J.J., and WENDEL, M.M., A Computer
of the Fixed Bed Catalytic Reactor: The Adiabatic and
Quasi-adiabatic Cases; AIChE Journal, Vol. 9, pp. 129-
133, January 1963.
9. MEARS, D.E., The Role of Axial Dispersion in
Trickle-Flow Laboratory Reactors; Chemical
Engineering Science, Vol. 26, pp. 1361-1366,
September 1971.
10. GIERMAN, H., Design of Laboratory Hydrotreating
Reactors Scaling Down of Trickle-flow Reactors;
Applied Catalysis, Vol. 43, pp. 277-286, October 1988.
11. HOCHMAN, J., and EFFRON, E., Two-phase
Cocurrent Downflow in Packed Beds; Industrial &
Engineering Chemistry Fundamentals, Vol. 8, pp 63-
71, February 1969.
12. CHU, C.F., and NG, K.M., Flow in Packed Tubes with
a Small Tube to Particle Diameter Ratio; AIChE
Journal, Vol. 35, pp. 148-158, January 1989.
13. MUNTEANU, M.C., CHEN, J. and FAROOQI, H.,
Flash Experiment Findings; Hydrocarbon Engineering,
Vol. 15, pp. 39-46, March 2010.
14. CHEN, J. WANG, N., MEDEROS, F. and
ANCHEYTA, J., Vapor-Liquid Equilibrium Study in
Trickle-Bed Reactors; Industrial & Engineering
Chemistry Research, Vol. 48, pp. 1096-1106,
September 2008.