Mar-2012

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

Article Summary

Fluid catalytic cracking (FCC) is one of the major processes in the refining industry for converting heavier hydrocarbons to useful middle distillates. Such a process demands the continuous addition of fresh FCC catalyst and results in the generation of tonnes of spent equilibrium FCC catalyst. To understand the FCC unit’s operation, the process can be divided into six sections: feed preheater, reactor, regenerator, main fractionator, gas plant and treating facilities. The cracking reactions occur in the reactor zone and this leads to coking. Coked catalyst is circulated back to the regenerator to burn the coke at high temperature in the presence of air. The regenerator hot flue gas (containing CO and CO2) holds an appreciable amount of energy. In most units, the flue gas is routed through a steam generating boiler (referred to as a CO boiler), where the carbon monoxide in the flue gas is burned as fuel to provide steam for use in the refinery as well as to comply with any applicable environmental regulatory limits on carbon monoxide emissions.

FCC units generally experience catalyst-related problems such as circulation, catalyst loss and activity decline. The FCC units studied were encountering low yield problems due to a decrease in catalyst activity. During the same period, it was also observed that the CO boiler became fouled externally, and a lot of fines were found in the CO boiler stack during cleaning. Subsequent to startup after cleaning, the CO boiler started to experience fouling again. There was also a steady decline in activity of the e-cat inventory.

Methodology
The problem of fines deposition in the CO boiler stack section, as well as catalyst activity loss, was investigated systematically by analysing particle size, attrition index, surface area and chemical composition in terms of metals, to arrive at a possible solution to enable smooth operation of the plant.

Samples of e-cat, fresh catalysts and CO boiler samples from FCC units were obtained from refineries, and these samples were analysed in the laboratory. The experimental techniques used to characterise the catalyst samples and the results are discussed in this article.

Results
In order to understand the reasons for CO boiler fouling in the FCC unit, a systematic study was carried out by measuring properties such as particle size distribution, attrition index and magnesium metal concentration for fresh FCC catalysts, additive and e-cat samples collected at different time intervals.

Particle size analysis
In Table 1, it is clear that the fraction of the catalyst coded ‘-45’ increased from 12 to 20 wt%. Particle size distribution was determined in accordance with the ASTM D4513-97 method using a sieving procedure. Typically, about 50g of FCC catalyst sample was sieved through a test sieve column comprising 180, 150, 106, 90, 75, 63, 53, 45, 32, and <20μm sieves for 30 minutes using an auto-sieve shaker. After sieving, fractions collected over each sieve were weighed and PSD calculated as per the ASTM method.

Average particle size decreased from 72 microns to 67 microns. This drop corresponds to an increased amount of fines in the catalyst. In order to substantiate this further, e-cat samples collected at different time intervals were also characterised for particle size (see Figures 1 and 2).

The average particle size distribution of fractions ‘-45’ and ‘-63’ were analysed over a period of one year. Figures 1 and 2 show that average particle size is more or less constant around 64 microns before the addition of additive. After the introduction of additive, the average particle size increased initially (from 64 to 75 microns), then stabilised around 65 micron. This increase in average particle size is an indication of less retention of fines in the unit. The ‘-45’ fraction also fluctuated in average particle size after the introduction of additive, before stabilising. During this period, a lot of fluctuation was noted in the unit with respect to catalyst. Hence, the decrease in particle size leads to less retention of catalyst in the unit and ultimately leads to fouling of the boiler.

Attrition index
Attrition index was measured for fresh FCC catalysts, which were used at different intervals of time, and also e-cat samples and catalyst additive samples (see Table 2). Attrition index was measured as per the ASTM D5757 procedure. It was observed that the additive attrition index was on the high side (10 wt%). Table 2 shows that the attrition loss for fresh catalyst batches was also on the high side. Further, it was observed that the attrition value for the e-cat-1 sample (4.7%) was in the acceptable range, while for the e-cat-2 and e-cat-3 samples attrition losses (5.7 and 7.1 wt%, respectively) were on the high side. This observation was attributed to attrition-prone catalyst and additive usage in the unit at that time.

Surface area measurements
To further understand the source of fines, the surface areas of fresh FCC catalyst fines (<40 micron), additive, attrited fresh catalyst and CO boiler fines were analysed. The surface areas of e-cat, fresh catalyst and fresh additive fines were comparable with those of the original sample. Fresh additive fines and attrited fresh catalyst were heated to 1000°C, and the measured surface area was found to be in the range 30-40 m2/g. A sample from the CO boiler was found to have a surface area of 9 m2/g. The very low surface area of CO boiler fines was due to their exposure to very high temperatures (>950°C), which could have resulted in pore sintering and formation of non-porous material; hence, it could not be attributed to either additive or FCC catalyst fines. Surface area measurements for catalyst samples were carried out using nitrogen adsorption/desorption measurements in an Autosorb-1MP unit. Nitrogen adsorption/desorption isotherms were measured at -196°C after degassing about 50 mg of sample below 10-3 torr at 300°C for three hours. BET specific surface area was estimated by following ASTM method D4365 using adsorption data in a relative pressure range from 0.008 to 0.08 bar. Surface area values for various samples are shown in Table 3 as per ASTM D4365.