Sep-2018

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

Article Summary

In a petroleum refinery, a catalytic reforming unit (CRU) is set up to increase the octane number of naphtha through reforming reactions. A CRU is always accompanied by a naphtha hydrotreater unit (NHDT) tasked with preparing the feed, primarily to remove sulphur from naphtha feed as sulphur acts as a poison for CRU catalyst. A process flow diagram of a typical NHDT is shown in Figure 1.

Process description
The process can be broadly categorised into a high pressure reaction section (the HP section) and a low pressure stripper section (the LP section).

High pressure reaction section
The naphtha feed (90-140 cut) mixed with recycle gas, predominantly hydrogen, is heated in a series of feed effluent exchangers before being fed to the fired heater where the feed mixture is heated to the required reaction temperature.

There are two reactors arranged in series where the feed mixture undergoes an exothermic hydrodesulphurisation reaction to yield desulphurised naphtha and hydrogen sulphide (H2S):

R-SH + H2 = R-H +H2S             (1)

The reactor effluents are fed to a high pressure separator. The vapour stream from the separator top mainly constitutes unreacted hydrogen. The vapour stream is partly recycled to recycle gas compressor suction and partly purged to the refinery fuel gas header.

Low pressure stripper section
The liquid stream from the separator bottom, rich in dissolved H2S, is fed to a stripper column where the H2S is stripped out. The stripper bottom desulphurised naphtha is further processed in the catalytic reforming unit.

Problem faced
In the present case, the pressure drop across the NHDT first reactor would increase sharply after the plant was operated for a few months (see Figure 2).

This increased pressure drop would hamper normal unit operation. It was a recurring problem and the unit had to be shut down to unload and skim the catalyst.

Suspected reasons for pressure drop
Gum formation
Often the reason for pressure drop in a naphtha hydrotreater is attributed to the tendency of olefins to react with oxygen to form gum. In the present case, the gum formation theory was discarded because: no cracked feed was processed and hence the olefin percentage of the feed was minimal; and a floating roof feed tank was being used, so there was no opportunity for reaction with oxygen.

Foreign particles
It was suspected that foreign particles might have deposited on the catalyst bed, plugging the void spaces and increasing the pressure drop across the bed. These particles could be generated as a result of erosion/corrosion in connecting pipelines and the feed storage tank.

Surprisingly, no such pressure drop was observed across the magnetic filter (which is supposed to trap such particles) in the feed line. Even downstream of the filter, no line thinning of connecting pipelines (that would have suggested corrosion) was observed, thus discarding the theory of foreign particles plugging the catalyst bed.

Coke formation
It was suspected that coking might be responsible for increased pressure drop across the catalyst bed. On opening the reactor bed during catalyst unloading and skimming, carbon deposits were found on the catalyst, so supporting the theory of coking.

However, the following factors argued against this analysis:
•    The feed used was straight run naphtha (90-140 cut)
•    The reaction temperature was only 290°C
•    The reaction was taking place in the gaseous phase in a hydrogen atmosphere.

Thus there was no explanation as to how the feed naphtha could be cracking at a temperature of 290°C in a gaseous phase reaction carried out in a hydrogen atmosphere and yet generating coke.

More evidence further contradicted the coke formation hypothesis. In previous runs, the NHDT reaction temperature was maintained at 315°C, but the problem of increased pressure drop across the reactor was not encountered.

The catalyst of the NHDT unit had been changed and the new catalyst required reduction of the NHDT reaction temperature to 290°C. The new catalyst’s guidelines pointed out that increasing the reaction temperature can lead to a recombination reaction (mercaptan recombination), which would affect the quality of the product:

R-H +H2S = RSH + H2             (2)

Reduction of the reaction temperature was a welcome change as it offloaded the NHDT fired heater. Even at a reaction temperature of 290°C, the naphtha feed was being desulphurised as was evident from the CRU feed that always tested positive for less than 0.5 ppm sulphur.

Cracking, often held responsible for coking and being an endothermic reaction, is promoted by an increase in temperature. However, it was not clear why, when the reactor was being operated at a higher reaction temperature, there was no coking, but when the reaction temperature was reduced keeping all other parameters identical, substantial coke deposition was detected in the reactor.

So what might have triggered coke deposition in the NHDT reactor at a reduced reaction temperature?

The heat of the reactor effluents was being utilised to preheat the unit feed to heater inlet temperature. In the present case, the reaction temperature was reduced from 315°C to 290°C. This substantially reduced the reactor effluent temperature. Hence, the feed preheat was also substantially reduced by 15-25°C.

The fired heater has been designed to handle vapour feed but, as a result of the reduction in feed preheat, the heater was forced to accept mixed feed (vapour and liquid). The liquid, once exposed to high heat flux on entering the fired heater tubes, cracked, forming coke. The coke formed was entrained with the feed to the reactor, blocking the reactor bed and increasing pressure drop across the bed.