Troubleshooting a FCC unit

The source of a fouling problem in a FCC CO boiler was identified by systematic

analysis of fresh catalysts, additives and equilibrium catalysts

Chiranjeevi Thota, Shalini Gupta, Dattatraya Tammanna Gokak, Ravi kumar Voolapalli, P V C Rao
and Viswanathan Poyyamani Swaminathan  Bharat Petroleum Corporation

F
luid catalytic cracking (containing CO and CO2) holds CO boiler stack during cleaning.
(FCC) is one of the major an appreciable amount of Subsequent to startup after
processes in the refining energy. In most units, the flue cleaning, the CO boiler started
industry for converting heavier gas is routed through a steam to experience fouling again.
hydrocarbons to useful middle generating boiler (referred to as There was also a steady decline
distillates. Such a process a CO boiler), where the carbon in activity of the e-cat
demands the continuous addi- monoxide in the flue gas is inventory.
tion of fresh FCC catalyst and burned as fuel to provide steam
results in the generation of for use in the refinery as well as Methodology
tonnes of spent equilibrium to comply with any applicable The problem of fines deposition
FCC catalyst. To understand the environmental regulatory limits in the CO boiler stack section,
FCC unit’s operation, the proc- on carbon monoxide emissions. as well as catalyst activity loss,
ess can be divided into six FCC units generally experi- was investigated systematically
sections: feed preheater, reactor, ence catalyst-related problems by analysing particle size, attri-
regenerator, main fractionator, such as circulation, catalyst loss tion index, surface area and
gas plant and treating facilities. and activity decline. The FCC chemical composition in terms
The cracking reactions occur in units studied were encountering of metals, to arrive at a possible
the reactor zone and this leads low yield problems due to a solution to enable smooth oper-
to coking. Coked catalyst is decrease in catalyst activity. ation of the plant.
circulated back to the regenera- During the same period, it was Samples of e-cat, fresh cata-
tor to burn the coke at high also observed that the CO boiler lysts and CO boiler samples
temperature in the presence of became fouled externally, and a from FCC units were obtained
air. The regenerator hot flue gas lot of fines were found in the from refineries, and these

Particle size distribution of fresh catalysts and additives (ASTM D4513)

Sample code, wt% -180 -150 -106 -90 -75 -63 -53 -45 -32 -20 (APS) µ
Fresh Cat-1 99 98 83 69 34 37 25 12 5 1 72
Fresh Cat-2 100 98 87 75 61 44 31 20 10 3 69
Fresh Cat-3 99 98 88 77 61 41 24 15 3 0.1 69
Fresh Cat-4 100 99 90 75 58 44 25 16 3.5 0.27 71
Fresh Cat-5 99 99 90 77 62 44 28 19 5.7 0.6 67
Fresh Add-1 98 97 81 53 33 19 12 6 1.2 0.1 88

Table 1

www.digitalrefining.com/article/1000354 Catalysis 2012 1

 Attrition index data (ASTM D5757)
78

74 Catalyst Attrition loss, wt%

Fresh Cat-1 5.0
70 Fresh Cat-3 6.9
Fresh Cat-6 9.3
66 Fresh Add-1 10.0
APS

E-cat-1 4.7
62 E-cat-2 5.7
E-cat-3 7.1
58
Table 2
54
45’ increased from 12 to 20 wt%.
50 Particle size distribution was
04

04

04

04

04

04

05

05
determined in accordance with
20

20

20

20

20

20

20

20
the ASTM D4513-97 method
ar

ay

ov

b
Ju

Ja
Au

Se

Fe
M

N
using a sieving procedure.
8
22

27
14

30
11

19
3

Typically, about 50g of FCC

catalyst sample was sieved
Figure 1 APS variation with respect to time through a test sieve column
comprising 180, 150, 106, 90, 75,
samples were analysed in the was carried out by measuring 63, 53, 45, 32, and <20µm sieves
laboratory. The experimental properties such as particle size for 30 minutes using an auto-
techniques used to characterise distribution, attrition index and sieve shaker. After sieving,
the catalyst samples and the magnesium metal concentration fractions collected over each
results are discussed in this for fresh FCC catalysts, additive sieve were weighed and PSD
article. and e-cat samples collected at calculated as per the ASTM
different time intervals. method.
Results Average particle size
In order to understand the Particle size analysis decreased from 72 microns to 67
reasons for CO boiler fouling in In Table 1, it is clear that the microns. This drop corresponds
the FCC unit, a systematic study fraction of the catalyst coded ‘- to an increased amount of fines
in the catalyst. In order to
substantiate this further,
e-cat samples collected at differ-
28
ent time intervals were also
24 characterised for particle size
(see Figures 1 and 2).
20 The average particle size
−45 fraction

distribution of fractions ‘-45’

16
and ‘-63’ were analysed over a
12 period of one year. Figures 1
and 2 show that average parti-
8 cle size is more or less constant
around 64 microns before the
4
addition of additive. After the
0 introduction of additive, the
average particle size increased
04

04

04

04

04

04

05

05
20

20

20

20

20

20

20

20

initially (from 64 to 75 microns),

ar

ay

ov

b
Ju

Ja
Au

Se

Fe
M

then stabilised around 65

N

8
22

27
14

30
11

19
3

micron. This increase in average

Figure 2 Variation in ‘-45’ fraction with respect to time particle size is an indication of

2 Catalysis 2012 www.digitalrefining.com/article/1000354

less retention of fines in the
unit. The ‘-45’ fraction also fluc- Surface area data for fresh catalyst, additive, e-cat and CO boiler samples
tuated in average particle size
after the introduction of addi- S.No Catalyst S.A, m2/g
tive, before stabilising. During 1 Fresh Cat-1 349
3 Fresh Add-1 100
this period, a lot of fluctuation
5 E-cat-4 204
was noted in the unit with 7 CO boiler fines surface area 9
respect to catalyst. Hence, the 8 Additives fresh fines, heated to 1000°C 30
decrease in particle size leads to 9 Fresh Cat-3, attrited fines, heated to 1000°C 40
less retention of catalyst in the
unit and ultimately leads to Table 3
fouling of the boiler.
heated to 1000°C, and the meas- Chemical analysis
Attrition index ured surface area was found to Chemical analysis was carried
Attrition index was measured be in the range 30-40 m2/g. A out on various samples to ascer-
for fresh FCC catalysts, which sample from the CO boiler was tain the source of CO boiler
were used at different intervals found to have a surface area of fines by measuring the magne-
of time, and also e-cat samples 9 m2/g. The very low surface sium content. Additive contains
and catalyst additive samples area of CO boiler fines was due magnesium metal and its
(see Table 2). Attrition index to their exposure to very high concentration should rise with
was measured as per the ASTM temperatures (>950°C), which the inventory build-up in the
D5757 procedure. It was could have resulted in pore unit. A microwave digestion
observed that the additive attri- sintering and formation of non- procedure has been standard-
tion index was on the high side porous material; hence, it could ised for quantifying magnesium
(10 wt%). Table 2 shows that not be attributed to either addi- in different fractions of CO
the attrition loss for fresh cata- tive or FCC catalyst fines. boiler fines. A microwave diges-
lyst batches was also on the Surface area measurements for tion system equipped with 14
high side. Further, it was catalyst samples were carried vessels and designed for pres-
observed that the attrition out using nitrogen adsorption/ sures up to 800 psi was used.
value for the e-cat-1 sample desorption measurements in Each vessel is charged with 0.2-
(4.7%) was in the acceptable an Autosorb-1MP unit. 0.5g of sample, to which is
range, while for the e-cat-2 and Nitrogen adsorption/desorption added 10 ml of nitric acid, 3 ml
e-cat-3 samples attrition losses isotherms were measured at - of distilled water and 1 ml of
(5.7 and 7.1 wt%, respectively) 196°C after degassing about 50 hydrogen peroxide. After about
were on the high side. This mg of sample below 10-3 torr at 15 minutes, when the first vigor-
observation was attributed to 300°C for three hours. BET ous reaction has taken place, the
attrition-prone catalyst and specific surface area was esti- pressure vessels are closed and
additive usage in the unit at mated by following ASTM put into a microwave oven to
that time. method D4365 using adsorption complete the one-stage diges-
data in a relative pressure range tion. Metal content is measured
Surface area measurements from 0.008 to 0.08 bar. Surface using the ICP technique.
To further understand the area values for various samples Various parameters of micro-
source of fines, the surface areas are shown in Table 3 as per wave digestion system were
of fresh FCC catalyst fines (<40 ASTM D4365. optimised for obtaining a clear
micron), additive, attrited fresh
catalyst and CO boiler fines
Microwave system parameters for digestion
were analysed. The surface
areas of e-cat, fresh catalyst and
fresh additive fines were compa- Stage Power RAMP minuets Control pressure, psi Hold time, min
Max %
rable with those of the original I 1200 W 100 30.00 120 30.00
sample. Fresh additive fines and
attrited fresh catalyst were Table 4

www.digitalrefining.com/article/1000354 Catalysis 2012 3

 sium content of the e-cat is
0.8 lower than the expected values.
Mass balance
Actual
The mass balance assumes both
0.7
catalyst and additive are
removed from the system
0.6
(either through cyclone losses
Ma, %

0.5
or catalyst purging) at the same
rate. Ideally, the estimated and
0.4
actual Mg concentrations
should be the same. The differ-
0.3 ence suggests that additive
retention is 85% of the main
0.2 catalyst. Perhaps excessive
0 30 60 90 120 generation of fines with addi-
Days tive may be contributing to this
lower retention. Hence, it may
Figure 3 Comparison of actual and calculated Mg content in e-cat be necessary to improve the
attrition resistance of the
solution for sample analysis. balance over a period of four additive.
The optimised program and months was done by analysing Based on magnesium analysis
parameters used in the one- magnesium content at different of CO boiler fines alone, the
stage digestion are shown in time intervals. Figure 3 problem cannot be attributed
Table 4. compares the expected and solely to either catalyst or addi-
To understand additive reten- actual magnesium content of tive, as the fines could have
tion in the unit, an e-cat mass the e-cat. The actual magne- accumulated over a period of
time and the addition of addi-
tive may have started
Mg and RE2O3 analysis of different fractions of e-cat
meanwhile. Magnesium and
rare earth oxide distribution in
different-sized fractions of the
E-cat -2 Mg, wt% RE2O3, wt% e-cat was analysed. Magnesium
32 to 45 µ 0.5850 1.62
45 to 63 µ 0.4855 1.55
and RE2O3 distribution with
63 to 90 µ 0.4384 1.45 different-sized fractions of e-cat
>90 µ 0.2780 1.02 are shown in Table 5. The table
shows that, with respect to
Table 5 particle size, magnesium and
rare earth content decreases,
80
which indicates that a loss of
fines leads to a decrease in
70 activity. The effect is shown in
60 Figure 4.
Systematic analysis of fresh
50
FCC catalyst, equilibrium cata-
Activity

40 lysts and fines collected from

30
the CO boiler at different time
intervals, and analysis of fresh
20 additive samples, was
10 employed to identify problems
and enable corrective actions to
0
Dec May Aug Dec May be recommended for the
smooth operation of the FCC
Figure 4 Activity variation with respect to time unit.

4 Catalysis 2012 www.digitalrefining.com/article/1000354

Conclusions attrition index as well as parti- Bharat Petroleum Corporation. He holds
Systematic studies were cle size analysis of the FCC a MSc in chemistry from Karnatak
conducted on different refinery catalyst and additive would be University, Dharwad, and a PhD from
samples to troubleshoot opera- good practice University Of Baroda.
tional problems in the FCC unit. • Systematic physico-chemical Ravi Kumar Voolapalli is a Chief
The studies led to the following characterisation of fresh catalyst, Manager at Corporate R&D Centre,
Bharat Petroleum Corporation. He holds
conclusions: e-cat and CO boiler fines helped
a BTech in chemical engineering from
• The CO boiler fines problem to diagnose the CO boiler fines
Andhra University, Visakhapatnam, a
confirmed that in-use fresh cata- and fouling problem, and
MTech in chemical engineering from the
lysts and additive are more enabled suitable remedial meas- Indian Institute of Technology, Kanpur,
prone to attrition, so they ures to be suggested to and a PhD in chemical engineering from
contributed more to the fines operations. Imperial College of Science Technology
loss from the unit and ulti- and Medicine.
mately to the fouling of the P V C Rao works on refining processes,
boiler. More fines loss also biofuels, crude assay and compatibility,
affected the conversion loss, Acknowledgement fuel additives and new products
because fines have more zeolite The authors gratefully acknowledge development. He holds a PhD in chemistry
and additive content BPCL’s management for encouragement from Indian Institute of Technology,
• Even though the attrition
and permission to publish the paper. Bombay.
index of additive is on the high Viswanathan Poyyamani Swaminathan 
is a Chief Manager and Head of the
side, which may contribute to
Corporate R&D Centre, Bharat Petroleum
the total fines, the amount of Chiranjeevi Thota is a Deputy Manager
Corporation. He holds a MSc and PhD
additive used is relatively small at the Corporate R&D Centre, Bharat
Petroleum Corporation, India. He holds a in chemistry from Indian Institute of
compared to the catalyst; hence, Technology, Delhi.
the additive alone cannot MSc in chemistry from Andhra University,
Visakhapatnam, and a PhD from Indian
contribute to the total fines
Institute of Petroleum, Dehradun. Email:
generated. This was established chiranjeevit@bharatpetroelum.in Links
by correlating the magnesium Shalini Gupta is a Deputy Manager (R&D)
content of CO boiler fines with with Bharat Petroleum Corporation. She More articles from the following
additive holds a master’s in chemistry.
categories:
• Since both fresh catalyst and Email: guptashalini@bharatpetroleum.in Catalysts & Additives
additive can contribute to fines, Dattatraya Tammanna Gokak is a Senior Fluid Catalytic Cracking
regular monitoring of the Manager at the Corporate R&D Centre,

www.digitalrefining.com/article/1000354 Catalysis 2012 5