You are on page 1of 5

Effect of reactor inlet temperature

in a hydrotreater
Solving a problem of pressure drop in a naphtha hydrotreater led to fresh
insights into the role of reactor inlet temperature

AMIT KALYAN CHANDRA


Indian Oil Corporation Limited

I
n a petroleum refinery, a cat-
alytic reforming unit (CRU) is
set up to increase the octane Low COT Stripper
number of naphtha through offgas

reforming reactions. A CRU is First Second


always accompanied by a naph- reactor reactor

tha hydrotreater unit (NHDT)


High Normal Reflux
∆P ∆P pump
tasked with preparing the feed, Recycle gas
compressor
primarily to remove sulphur
from naphtha feed as sulphur Stripper
Feed column
acts as a poison for CRU cata- heater
Low reactor
effluent temp.
Make-up
lyst. A process flow diagram hydrogen

of a typical NHDT is shown in


Recycle gas
Feed/effluent rich in
heat exchanger hydrogen
Figure 1.

Process description Reboiler


The process can be broadly cat- furnace
Feed High pressure
egorised into a high pressure Naphtha pump separator
reaction section (the HP sec- storage tank Desulphurised
naphtha to CRU
tion) and a low pressure strip-
per section (the LP section).
Figure 1 Process flow diagram of a naphtha hydrotreater unit
High pressure reaction section
The naphtha feed (90-140 cut) reaction to yield desulphur- cled to recycle gas compressor
mixed with recycle gas, pre- ised naphtha and hydrogen sul- suction and partly purged to
dominantly hydrogen, is phide (H2S): the refinery fuel gas header.
heated in a series of feed efflu- Low pressure stripper section
R-SH + H2 = R-H +H2S (1)
ent exchangers before being fed The liquid stream from the sep-
to the fired heater where the arator bottom, rich in dissolved
feed mixture is heated to the The reactor effluents are fed H2S, is fed to a stripper col-
required reaction temperature. to a high pressure separator. umn where the H2S is stripped
There are two reactors The vapour stream from the out. The stripper bottom des-
arranged in series where the separator top mainly consti- ulphurised naphtha is fur-
feed mixture undergoes an exo- tutes unreacted hydrogen. The ther processed in the catalytic
thermic hydrodesulphurisation vapour stream is partly recy- reforming unit.

www.digitalrefining.com/article/1001546 PTQ Q3 2018 1


6

First reactor ∆P, kg/cm2 5

0
1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301
Time, days

Figure 2 Differential pressure drop across the NHDT first reactor catalyst bed

Problem faced void spaces and increasing the • The reaction temperature was
In the present case, the pressure pressure drop across the bed. only 290°C
drop across the NHDT first These particles could be gener- • The reaction was taking
reactor would increase sharply ated as a result of erosion/corro- place in the gaseous phase in a
after the plant was operated for sion in connecting pipelines and hydrogen atmosphere.
a few months (see Figure 2). the feed storage tank. Thus there was no explana-
This increased pressure drop Surprisingly, no such pres- tion as to how the feed naphtha
would hamper normal unit sure drop was observed across could be cracking at a temper-
operation. It was a recurring the magnetic filter (which is ature of 290°C in a gaseous
problem and the unit had to be supposed to trap such particles) phase reaction carried out in a
shut down to unload and skim in the feed line. Even down- hydrogen atmosphere and yet
the catalyst. stream of the filter, no line thin- generating coke.
ning of connecting pipelines More evidence further con-
Suspected reasons for pressure (that would have suggested tradicted the coke formation
drop corrosion) was observed, thus hypothesis. In previous runs,
Gum formation discarding the theory of foreign the NHDT reaction tempera-
Often the reason for pressure particles plugging the catalyst ture was maintained at 315°C,
drop in a naphtha hydrotreater bed. but the problem of increased
is attributed to the tendency of pressure drop across the reactor
olefins to react with oxygen to Coke formation was not encountered.
form gum. In the present case, It was suspected that cok- The catalyst of the NHDT
the gum formation theory was ing might be responsible for unit had been changed and the
discarded because: no cracked increased pressure drop across new catalyst required reduc-
feed was processed and hence the catalyst bed. On opening tion of the NHDT reaction tem-
the olefin percentage of the feed the reactor bed during catalyst perature to 290°C. The new
was minimal; and a floating unloading and skimming, car- catalyst’s guidelines pointed
roof feed tank was being used, bon deposits were found on the out that increasing the reac-
so there was no opportunity for catalyst, so supporting the the- tion temperature can lead to a
reaction with oxygen. ory of coking. recombination reaction (mer-
However, the following captan recombination), which
Foreign particles factors argued against this would affect the quality of the
It was suspected that foreign analysis: product:
particles might have deposited • The feed used was straight
on the catalyst bed, plugging the run naphtha (90-140 cut) R-H +H2S = RSH + H2 (2)

2 PTQ Q3 2018 www.digitalrefining.com/article/1001546


240

Feed preheat temperature, ºC 230

220

210

200

190
High feed preheat
Low feed preheat
180
1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301
Time, days

Figure 3 Feed preheat temperature for two cases

Reduction of the reaction moted by an increase in tem- reactor at a reduced reaction


temperature was a welcome perature. However, it was not temperature?
change as it offloaded the clear why, when the reactor The heat of the reactor efflu-
NHDT fired heater. Even at a was being operated at a higher ents was being utilised to pre-
reaction temperature of 290°C, reaction temperature, there was heat the unit feed to heater
the naphtha feed was being no coking, but when the reac- inlet temperature. In the pres-
desulphurised as was evident tion temperature was reduced ent case, the reaction tempera-
from the CRU feed that always keeping all other parameters ture was reduced from 315°C
tested positive for less than 0.5 identical, substantial coke dep- to 290°C. This substantially
ppm sulphur. osition was detected in the reduced the reactor effluent
Cracking, often held respon- reactor. temperature. Hence, the feed
sible for coking and being an So what might have triggered preheat was also substantially
endothermic reaction, is pro- coke deposition in the NHDT reduced by 15-25°C.

5.0
Catalyst bed ∆P, kg/cm2

4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301
Time, days

Figure 4 Differential pressure drop across the catalyst bed with high feed preheat

www.digitalrefining.com/article/1001546 PTQ Q3 2018 3


6
Low feed preheat
5 High feed preheat

4
∆P, kg/cm2

0
1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301
Time, days

Figure 5 Differential pressure drop across the catalyst bed for two cases

The fired heater has been 210-215°C, the other with feed Coke deposition on the cata-
designed to handle vapour feed preheat in the range 225-230°C lyst bed was detected and iden-
but, as a result of the reduc- (see Figure 5). tified as the culprit behind a
tion in feed preheat, the heater From the plot, it can be high pressure drop across the
was forced to accept mixed feed clearly inferred that the rate hydrotreater reactor. The unit
(vapour and liquid). The liq- of rise in differential pressure was being operated at a lower
uid, once exposed to high heat drop with low feed preheat was reaction temperature than
flux on entering the fired heater 0.5 kg/cm2/month and that used to be maintained in pre-
tubes, cracked, forming coke. with high feed preheat this was vious runs when no such prob-
The coke formed was entrained 0.01 kg/cm2/month. lem of coking was faced. It was
with the feed to the reactor, Two important points can be unclear what might have led to
blocking the reactor bed and inferred from this case study. coke deposition on the catalyst
increasing pressure drop across Reaction temperature is popu- bed.
the bed. larly believed to affect the rate It was observed that reducing
and conversion of reactions. reaction temperature did not
Remedial action In the present case, lowering affect the conversion rate, and
The NHDT reaction temper- the reaction temperature did the quality of product was not
ature was increased to 315°C. not lower the conversion rate. compromised, but it did have
This simultaneously increased However, lowering the reaction an adverse effect on the unit
the feed preheat, resulting in a temperature reduced the feed operation. Reducing the reac-
fully vaporised feed being fed preheat which in turn adversely tion temperature subsequently
to the fired heater. The feed affected the unit, being respon- reduced the reactor effluent
preheat temperatures have sible for coke deposition in the temperature. The heat of the
been plotted for both cases (see catalyst bed. reactor effluent was in turn
Figure 3). being used to preheat the feed.
The increasing rate of pres- Conclusion As a result, preheat was sub-
sure drop across the NHDT In a typical NHDT unit, the stantially reduced.
reactor was regulated (see pressure drop across the hydro- The NHDT fired heater had
Figure 4). treater reactor was increasing at been designed to handle fully
In support of this remedial a sharp rate, rendering it almost vaporised feed. In this case, it
action, the pressure drop across impossible to maintain normal was being fed with mixed feed.
the catalyst bed has been plot- unit operations. The unit had The liquid part of the feed,
ted for two different cases, one to be shut down to unload and when exposed to high heat flux
with feed preheat in the range skim the catalyst. in the heater, cracked, leading

4 PTQ Q3 2018 www.digitalrefining.com/article/1001546


to coke formation in the heater only affects the rate and con- from Jadavpur University and is an
tubes. This coke, when carried version of reactions, it also Associate Member of the Indian Institute
to the reactor bed, plugged the plays a critical role in affecting of Chemical Engineers and the Institution
of Chemical Engineers.
void spaces and increased the pressure drop across the reac-
pressure drop across the reac- tor bed.
tor bed.
Increasing the reaction tem- Amit Kalyan Chandra is an Assistant LINKS
perature regulated the rate of Manager with Indian Oil Corporation
Limited. With more than five years
increase in the pressure drop More articles from the following
of operational experience in crude categories:
across the reactor bed and the
distillation, naphtha hydrotreating, Fired Heaters, Furnaces and
problem was solved. The pres- catalytic reforming and kero- Boilers
ent case study revealed that hydrodesulphurisation, he holds a Catalysts and Additives
reactor inlet temperature not bachelor of chemical engineering degree

www.digitalrefining.com/article/1001546 PTQ Q3 2018 5

You might also like