Apr-2023

Minimising unwanted coke formation in FCC operations

Design and operating practices to minimise the amount and effects of unwanted coke formation in FCC operations.

Warren Letzsch
Warren Letzsch Consulting PC

Article Summary

Coke deposition in an FCC unit has been an issue since the inception of the process. In small amounts, it is just a consequence of the operation and is handled during a normal turn-around (TAR) for refurbishing the unit. Excessive coke formation, however, poses serious problems for the operation. It may cause a premature shutdown of the unit, which incurs high costs and lost production. Any time an unscheduled shutdown happens, there is a chance that equipment can be damaged, which will further add to the downtime and costs.

If the unit is put on standby and cools and then is restarted, there is a real possibility that coke will spall off surfaces and may plug diplegs and/or slide valves. Coke build-up can also be costly since it can have a negative impact on the pressure balance. Higher pressure drops through the overhead vapour line, the riser, or the cyclones may reduce the unit’s ability to run at full capacity. Entering any vessel that contains coke can be dangerous due to the risk of fire and falling coke. The unit must be cooled below the coke ignition temperature prior to entering. All of the coke should be removed if possible, which may lengthen the TAR time and cause damage to the refractory.

Coking in the FCC unit has been reported since the early days of catalytic cracking. Locations of coking areas found in FCC units are shown in Figures 1 and 2. Coke has been reported by the feed nozzles, on the feed riser walls, inside the primary and secondary cyclones and outlet tubes, in the reactor vessel, on top of the reactor cyclones, in the plenum chamber, in the overhead vapour line, in the main fractionator, and the slurry circuit. In short, coke can occur nearly everywhere in the feed side of the unit and the heavy oil fractionating circuit in the gas recovery section.

Early units
Coking in FCC units can be related to the feedstocks being processed. Early FCC units operated with regenerators running between 1050-1200ºF and reactor temperatures between 880-950ºF and had feed end points of 925-1000ºF. Reactors contained a lot of catalyst with a typical design WHSV of 2. The catalyst in the reactor adsorbed any coke precursors in the feed or oil droplets that were not vaporised in the feed riser.

The bed crackers had hydrocarbons and steam leaving the bed and going to cyclones. The top of the reactor vessel was usually cooler than the bed because cracking reactions that occur in the dilute phase are endothermic, and catalyst concentrations are low. Any polymerisation or alkylation reactions that occur can lead to condensation of the heavier products formed in the reactor. This then leads to coking on the roof of the reactor, the tops of the cyclones, and possibly the plenum chamber. Another factor was the size of the unit and how the recycle was processed. Some units had separate risers for these streams.

FCC feed coking tendencies
When zeolites were introduced in the 1960s, they revolutionised catalytic cracking. Being more than an order of magnitude more active than the amorphous catalysts they replaced meant that recycle was no longer necessary to increase conversion. Reactor beds were replaced with feed risers and a rapid catalyst/vapour separator, and the catalyst needed to be regenerated to a coke level of 0.3 wt maximum. Most units aimed for 0.10-0.20 wt% carbon on catalyst.

Higher regenerator temperatures were needed, and the complete burning of CO in the regenerator bed replaced the downstream CO boilers. Feed end points for gasoils rose to 1050-1100ºF. Even higher end points occurred when the vacuum towers were pushed past design. Coking capacity increased, and coker gasoils were sent to the FCC. The coking tendencies of the FCC feed increase with aromatic content.

Large molecules can react with the olefins and diolefins generated by the coker’s thermal cracking reactions and condense as a liquid. This can cause coke to form in the FCC reactor and all the downstream equipment leading to the main column. Coker gasoils that were sent to storage picked up oxygen which caused deposits when fed to the cracker. The reactions of the diolefins with other molecules also added to the coke formation. Hydrotreating the feed removes a lot of the coke precursors and greatly reduces coking issues.

Coking in resid cracking
In the 1970s and 1980s, resid cracking was introduced as a method of disposing of an unwanted product. Coking in the feed side of the process started showing up due to non-vaporised feed. Two methods to combat the coking problem were used. First, catalyst coolers were employed to increase the catalyst/oil ratio so the coking occurred on the catalyst. The second was to vaporise as much feed as possible to minimise the oil droplets. Calculations for residual feeds indicate that as much as 25% of the feed will not be vaporised with conventional feed nozzles.

When vaporising feed, the injectors must reduce the oil droplets to the size of the catalyst particles, minimising coking seen throughout the reactor system. Coking around the feed nozzles and/or the riser walls is usually due to wet dispersion steam, which can also damage the nozzles. Uneven feed sprays into the riser can cause coke and will have a negative impact on yields. Any cooling of reactor vapours can also cause coke. This can occur if the insulation is insufficient or damaged. Start-ups can be a time when this is more likely to happen since reactor temperatures may be too low when feed is introduced.

Riser design may also be a problem. If the residence time of the oil is too low, bottoms cracking may be incomplete, and the cracking that occurs in the vapour line or the bottom of the fractionator will cause coking to lay down in the affected area. Coking in the reactor cyclones has been reported on the back of the outlet tubes (see Figure 3), where a dead space may be present. Some refiners have put refractory on that portion of the outlet tubes to minimise coking.

Unit configuration-based coking
The metallurgy used might be a factor. Stainless steel could give better performance than the carbon steels typically used in this service. Uneven flow of the catalyst to the reactor cyclones may also produce coke in the cyclones or on the walls of the reactor vessel. Again, the unit configuration may be a factor contributing to the coking problem. The close-coupled reaction systems create a space above the cyclones where there is little catalyst traffic. Coke forms on the reactor cyclones because fine catalyst will deposit on the cyclones and absorb hydrocarbons, which leads to the coke build-up.

Dome steam is required to prevent this coke laydown on top of the reactor cyclones. Too much dome steam can cool the reactor vapours and form coke. Wet steam must be avoided in this application since it will cause coke formation and may damage any metallurgy it contacts. The temperature drop at the top of the vessel should be limited to about 10ºF. Steam rates are low, and the superficial velocity is only about 0.01 ft/sec. The distance between the steam distributor and the tops of the cyclones needs to be properly sized. Each vender has recommendations for their design.

Any large changes in reactor temperature can cause any coke that has formed in the process to spall off, possibly causing cyclone malfunctions or flow problems if it blocks a slide valve. Coking in the overhead vapour line is frequently due to liquid condensing on the walls. Liquid droplets caused by a feed system that poorly atomises or does not evenly distribute the feed are the main source of this coke. Poor insulation or cooling of the product stream can also result in condensation reactions that form liquids.

Pressure drop increases in the overhead vapour line cause higher pressures in the reactor and regenerator. This can reduce the capacity of the unit if it is at blower capacity and will increase costs since more horsepower will be required at the higher pressure. Higher reactor pressure can reduce olefinicity and increase delta coke. Slide valve differentials may change, and a rebalancing of the pressure balance will be needed if the excessive pressure drop is too large.